Frontiers in Education : Computer Science and Computer Engineering [1 ed.] 9781683925682, 9781601324986

Citation preview

WORLDCOMP’19

PROCEEDINGS OF THE 2019 INTERNATIONAL CONFERENCE ON

PROCEEDINGS OF THE 2019 INTERNATIONAL CONFERENCE ON FRONTIERS IN EDUCATION: COMPUTER SCIENCE & FRONTIERS IN EDUCATION: COMPUTER SCIENCE & COMPUTER ENGINEERING COMPUTER ENGINEERING

Frontiers in Education: Computer Science and Computer Engineering

FECS’19 Editors Hamid R. Arabnia Leonidas Deligiannidis, Fernando G. Tinetti Quoc-Nam Tran Associate Editor Ashu M. G. Solo

U.S. $129.95 ISBN 9781601324986

12995

EMBD-FECS19_Full-Cover.indd All Pages

Arabnia

9 781601 324986

Publication of the 2019 World Congress in Computer Science, Computer Engineering, & Applied Computing (CSCE’19) July 29 - August 01, 2019 | Las Vegas, Nevada, USA https://americancse.org/events/csce2019

Copyright © 2019 CSREA Press

18-Feb-20 5:27:25 PM

This volume contains papers presented at the 2019 International Conference on Frontiers in Education: Computer Science & Computer Engineering. Their inclusion in this publication does not necessarily constitute endorsements by editors or by the publisher.

Copyright and Reprint Permission Copying without a fee is permitted provided that the copies are not made or distributed for direct commercial advantage, and credit to source is given. Abstracting is permitted with credit to the source. Please contact the publisher for other copying, reprint, or republication permission.

American Council on Science and Education (ACSE)

Copyright © 2019 CSREA Press ISBN: 1-60132-498-7 Printed in the United States of America https://americancse.org/events/csce2019/proceedings

Foreword It gives us great pleasure to introduce this collection of papers to be presented at the 2019 International Conference on Frontiers in Education: Computer Science and Computer Engineering (FECS’19), July 29 – August 1, 2019, at Luxor Hotel (a property of MGM Resorts International), Las Vegas, USA. The preliminary edition of this book (available in July 2019 for distribution on site at the conference) includes only a small subset of the accepted research articles. The final edition (available in August 2019) will include all accepted research articles. This is due to deadline extension requests received from most authors who wished to continue enhancing the write-up of their papers (by incorporating the referees’ suggestions). The final edition of the proceedings will be made available at https://americancse.org/events/csce2019/proceedings . An important mission of the World Congress in Computer Science, Computer Engineering, and Applied Computing, CSCE (a federated congress to which this conference is affiliated with) includes "Providing a unique platform for a diverse community of constituents composed of scholars, researchers, developers, educators, and practitioners. The Congress makes concerted effort to reach out to participants affiliated with diverse entities (such as: universities, institutions, corporations, government agencies, and research centers/labs) from all over the world. The congress also attempts to connect participants from institutions that have teaching as their main mission with those who are affiliated with institutions that have research as their main mission. The congress uses a quota system to achieve its institution and geography diversity objectives." By any definition of diversity, this congress is among the most diverse scientific meeting in USA. We are proud to report that this federated congress has authors and participants from 57 different nations representing variety of personal and scientific experiences that arise from differences in culture and values. As can be seen (see below), the program committee of this conference as well as the program committee of all other tracks of the federated congress are as diverse as its authors and participants. The program committee would like to thank all those who submitted papers for consideration. About 44% of the submissions were from outside the United States. Each submitted paper was peer-reviewed by two experts in the field for originality, significance, clarity, impact, and soundness. In cases of contradictory recommendations, a member of the conference program committee was charged to make the final decision; often, this involved seeking help from additional referees. In addition, papers whose authors included a member of the conference program committee were evaluated using the double-blinded review process. One exception to the above evaluation process was for papers that were submitted directly to chairs/organizers of pre-approved sessions/workshops; in these cases, the chairs/organizers were responsible for the evaluation of such submissions. The overall paper acceptance rate for regular papers was 19%; 26% of the remaining papers were accepted as poster papers (at the time of this writing, we had not yet received the acceptance rate for a couple of individual tracks.) We are very grateful to the many colleagues who offered their services in organizing the conference. In particular, we would like to thank the members of Program Committee of FECS’19, members of the congress Steering Committee, and members of the committees of federated congress tracks that have topics within the scope of FECS. Many individuals listed below, will be requested after the conference to provide their expertise and services for selecting papers for publication (extended versions) in journal special issues as well as for publication in a set of research books (to be prepared for publishers including: Springer, Elsevier, BMC journals, and others).    

Prof. Afrand Agah; Department of Computer Science, West Chester University of Pennsylvania, West Chester, PA, USA Prof. Abbas M. Al-Bakry (Congress Steering Committee); University President, University of IT and Communications, Baghdad, Iraq Prof. Emeritus Nizar Al-Holou (Congress Steering Committee); Professor and Chair, Electrical and Computer Engineering Department; Vice Chair, IEEE/SEM-Computer Chapter; University of Detroit Mercy, Detroit, Michigan, USA Prof. Hamid R. Arabnia (Congress Steering Committee); Graduate Program Director (PhD, MS, MAMS); The University of Georgia, USA; Editor-in-Chief, Journal of Supercomputing (Springer); Fellow, Center of Excellence in Terrorism, Resilience, Intelligence & Organized Crime Research (CENTRIC).

        

 

  

         

Prof. Mehran Asadi; Department of Business and Entrepreneurial Studies, The Lincoln University, Pennsylvania, USA Prof. Dr. Juan-Vicente Capella-Hernandez; Universitat Politecnica de Valencia (UPV), Department of Computer Engineering (DISCA), Valencia, Spain Prof. Juan Jose Martinez Castillo; Director, The Acantelys Alan Turing Nikola Tesla Research Group and GIPEB, Universidad Nacional Abierta, Venezuela Prof. Emeritus Kevin Daimi (Congress Steering Committee); Director, Computer Science and Software Engineering Programs, Department of Mathematics, Computer Science and Software Engineering, University of Detroit Mercy, Detroit, Michigan, USA Prof. Zhangisina Gulnur Davletzhanovna; Vice-rector of the Science, Central-Asian University, Kazakhstan, Almaty, Republic of Kazakhstan; Vice President of International Academy of Informatization, Kazskhstan, Almaty, Republic of Kazakhstan Prof. Leonidas Deligiannidis (Congress Steering Committee); Department of Computer Information Systems, Wentworth Institute of Technology, Boston, Massachusetts, USA; Visiting Professor, MIT, USA Dr. Lamia Atma Djoudi (Chair, Doctoral Colloquium & Demos Sessions); Synchrone Technologies, France Prof. Mary Mehrnoosh Eshaghian-Wilner (Congress Steering Committee); Professor of Engineering Practice, University of Southern California, California, USA; Adjunct Professor, Electrical Engineering, University of California Los Angeles, Los Angeles (UCLA), California, USA Prof. George A. Gravvanis (Congress Steering Committee); Director, Physics Laboratory & Head of Advanced Scientific Computing, Applied Math & Applications Research Group; Professor of Applied Mathematics and Numerical Computing and Department of ECE, School of Engineering, Democritus University of Thrace, Xanthi, Greece. Prof. Houcine Hassan; Department of Computer Engineering (Systems Data Processing and Computers), Universitat Politecnica de Valencia, Spain Prof. George Jandieri (Congress Steering Committee); Georgian Technical University, Tbilisi, Georgia; Chief Scientist, The Institute of Cybernetics, Georgian Academy of Science, Georgia; Ed. Member, International Journal of Microwaves and Optical Technology, The Open Atmospheric Science Journal, American Journal of Remote Sensing, Georgia Prof. Byung-Gyu Kim (Congress Steering Committee); Multimedia Processing Communications Lab.(MPCL), Department of Computer Science and Engineering, College of Engineering, SunMoon University, South Korea Prof. Tai-hoon Kim; School of Information and Computing Science, University of Tasmania, Australia Prof. Louie Lolong Lacatan; Chairperson, Computer Engineerig Department, College of Engineering, Adamson University, Manila, Philippines; Senior Member, International Association of Computer Science and Information Technology (IACSIT), Singapore; Member, International Association of Online Engineering (IAOE), Austria Prof. Dr. Guoming Lai; Computer Science and Technology, Sun Yat-Sen University, Guangzhou, P. R. China Dr. Andrew Marsh (Congress Steering Committee); CEO, HoIP Telecom Ltd (Healthcare over Internet Protocol), UK; Secretary General of World Academy of BioMedical Sciences and Technologies (WABT) a UNESCO NGO, The United Nations Prof. Dr., Eng. Robert Ehimen Okonigene (Congress Steering Committee); Department of Electrical & Electronics Engineering, Faculty of Engineering and Technology, Ambrose Alli University, Nigeria Prof. James J. (Jong Hyuk) Park (Congress Steering Committee); Department of Computer Science and Engineering (DCSE), SeoulTech, Korea; President, FTRA, EiC, HCIS Springer, JoC, IJITCC; Head of DCSE, SeoulTech, Korea Dr. Akash Singh (Congress Steering Committee); IBM Corporation, Sacramento, California, USA; Chartered Scientist, Science Council, UK; Fellow, British Computer Society; Member, Senior IEEE, AACR, AAAS, and AAAI; IBM Corporation, USA Chiranjibi Sitaula; Head, Department of Computer Science and IT, Ambition College, Kathmandu, Nepal Ashu M. G. Solo (Publicity), Fellow of British Computer Society, Principal/R&D Engineer, Maverick Technologies America Inc. Prof. Dr. Ir. Sim Kok Swee; Fellow, IEM; Senior Member, IEEE; Faculty of Engineering and Technology, Multimedia University, Melaka, Malaysia Prof. Fernando G. Tinetti (Congress Steering Committee); School of Computer Science, Universidad Nacional de La Plata, La Plata, Argentina; also at Comision Investigaciones Cientificas de la Prov. de Bs. As., Argentina Prof. Hahanov Vladimir (Congress Steering Committee); Vice Rector, and Dean of the Computer Engineering Faculty, Kharkov National University of Radio Electronics, Ukraine and Professor of Design Automation Department, Computer Engineering Faculty, Kharkov; IEEE Computer Society Golden Core Member; National University of Radio Electronics, Ukraine



 

Prof. Shiuh-Jeng Wang (Congress Steering Committee); Director of Information Cryptology and Construction Laboratory (ICCL) and Director of Chinese Cryptology and Information Security Association (CCISA); Department of Information Management, Central Police University, Taoyuan, Taiwan; Guest Ed., IEEE Journal on Selected Areas in Communications. Prof. Layne T. Watson (Congress Steering Committee); Fellow of IEEE; Fellow of The National Institute of Aerospace; Professor of Computer Science, Mathematics, and Aerospace and Ocean Engineering, Virginia Polytechnic Institute & State University, Blacksburg, Virginia, USA Prof. Jane You (Congress Steering Committee); Associate Head, Department of Computing, The Hong Kong Polytechnic University, Kowloon, Hong Kong

We would like to extend our appreciation to the referees, the members of the program committees of individual sessions, tracks, and workshops; their names do not appear in this document; they are listed on the web sites of individual tracks. As Sponsors-at-large, partners, and/or organizers each of the followings (separated by semicolons) provided help for at least one track of the Congress: Computer Science Research, Education, and Applications Press (CSREA); US Chapter of World Academy of Science; American Council on Science & Education & Federated Research Council (http://www.americancse.org/). In addition, a number of university faculty members and their staff (names appear on the cover of the set of proceedings), several publishers of computer science and computer engineering books and journals, chapters and/or task forces of computer science associations/organizations from 3 regions, and developers of high-performance machines and systems provided significant help in organizing the conference as well as providing some resources. We are grateful to them all. We express our gratitude to keynote, invited, and individual conference/tracks and tutorial speakers - the list of speakers appears on the conference web site. We would also like to thank the followings: UCMSS (Universal Conference Management Systems & Support, California, USA) for managing all aspects of the conference; Dr. Tim Field of APC for coordinating and managing the printing of the proceedings; and the staff of Luxor Hotel (Convention department) at Las Vegas for the professional service they provided. Last but not least, we would like to thank the Co-Editors of FECS’19: Prof. Hamid R. Arabnia, Prof. Leonidas Deligiannidis, Prof. Fernando G. Tinetti, Prof. Quoc-Nam Tran, and Associate Editor, Ashu M. G. Solo. We present the proceedings of FECS’19.

Steering Committee, 2019 http://americancse.org/

Contents SESSION: FRONTIERS IN EDUCATION: TEACHING METHODOLOGIES AND STRATEGIES, ADVISING METHODS, TOOLS AND RELATED STUDIES Data-Driven Models to Predict Student Performance and Improve Advising in Computer Science Varick L. Erickson

3

An Introductory Visualization Aid for Cybersecurity Education 10 Gabriel Castro Aguayo, Ulises Morales, Xiaoyu Long, Quamar Niyaz, Xiaoli Yang, Ahmad Y. Javaid Using a Telepresence Robot in an Educational Context Laurent Gallon, Angel Abenia, Francoise Dubergey, Maite Negui

16

A Hands-on Oriented Cyber-Learning Curriculum for Undergraduate Cybersecurity Education Yaswanth Kolli, Ahmad Y. Javaid

23

Cognitive Training Mauro Puebla-Alvarez, Daniel Carrillo, Xiaoli Yang

30

Co-construction of Computer Science Knowledge-to-be-taught in a French Context Timothee Duron, Vanea Chiprianov, Laurent Gallon

33

An Assessment Study on the Teaching of Critical Thinking and Mathematical Proofs in a Discrete Structures Course Hang Dinh

40

Living with Digital Education: The Impact and Power (or Otherwise) of Information and Communication Technology (ICT) and Internet in the life of Norfolk State University Community-An Exploratory Study Samuel BO Olatunbosun, Victoria T. Olatunbosun

47

Data Manipulation and Visualization (DMV): A Case Study Mudasser F. Wyne, Anshu Chaudhary, Dhwani Doshi, Manjusha Gusain

53

Preliminary Review - Universities' Open Source Academic Integrity Policies in the UAE Zeenath Reza Khan, Halim Khelalfa, Jawahitha Sarabdeen, Prinayka Harish, Sanjana Raheja

59

The Development and Deployment of a Mobile Music Application for Literacy Enhancement (M2APPLE) Amal Babangida Sabo, Mathias Fonkam, Abubakar Sadiq Hussaini, Charles Nche

65

Collaboration: Key to Student Success in Computing and other STEM Fields in Hispanic Serving Institutions Meline Kevorkian, Greg Simco

71

SESSION: ACCREDITATION, ASSESSMENT METHODS AND STRATEGIES + CURRICULUM DESIGN AND RELATED ISSUES Design and Development of a Modular K12 Cybersecurity Curriculum Giti Javidi, Ehsan Sheybani, Zacharia Pieri

77

Seeking ABET Accreditation: A Case Study in Outcome Assessment John Carelli

81

Continuous Improvement Model to Systematize Curricular Processes in the Context of ABET Accreditation Carolina Zambrano

88

SESSION: RESEARCH PROJECTS AND CAPSTONE DESIGN PROJECTS An Internet of Things Drone Data Mule 97 Drew Cochrane, Nolan Evans, Michael Lane, Grant Woodbury, Xi Yin, Corey Zrobek, Marcia Friesen, Ken Ferens Capstone - Introducing Students to Research through Application Development in Teams Katia Mayfield, Matthew Perry, Christopher Pounders, Lucas Pruitt, Ory Wigington

108

Involving Multiple Levels of Students in a Software Capstone Project - A Case Study Robert Hatch, Nicholas Setliff

115

SESSION: COMPUTER SCIENCE AND COMPUTER PROGRAMMING, SOFTWARE ENGINEERING, AND WORK PRACTICES Coding VR Games Jiaying Chen, Mehdi R. Zargham, Manibharathi Rajendran, Jie Cheng

123

Increasing Student Engagement by Having Them Run Your Test Data Through Their Programs Donald Schwartz

128

Entry-Level Data Science Work Practices and Environments Bill Hefley, Jason Parker, Sourav Chatterjee

134

To GUI or Not to GUI: On How We Teach Introduction to Programming Antonio Sanchez, Bo Mei

140

SESSION: LEARNING METHODOLOGIES AND COGNITION + DISTANCE LEARNING, ON-LINE EDUCATION, AND RELATED ISSUES Teaching an International Distributed Discussion-Based Course Jeff Offutt, Birgitta Lindstrom, Kesina Baral

149

A Hypothetical Model toward Establishing a Relationship between Cognition and Metacognition in Technology-Enhanced Self-Regulated Learning Minh-Tuan Tran, Shinobu Hasegawa

155

Finding Big Liars: A Computational Laboratory Challenge Robert D. McLeod, Luc L. Livolsi-Merritt , Marcia R. Friesen

159

Developing a Recipe Planning Board Game by Design Thinking Approach Shian-Shyong Tseng, Tsung-Yu Yang, Ai-Chi Lu

163

SESSION: POSTER PAPERS AND EXTENDED ABSTRACTS The Considerations for Training Computer Science Chinese International Students in Acadia 171 University Haiyi Zhang Active Learning in Computer Networking Wen-Jung Hsin

173

Outcome and Satisfaction Analysis for Online Exams in an E-learning Class Neftali Watkinson, Lubomir Bic

175

SESSION: LATE BREAKING PAPERS: STEM EDUCATION, LEARNING STRATEGIES AND CAPSTONE PROJECTS Effective Use of Slack and Short Video to Scale Online Learning Communities Alexandra Mehlhase, Robert Heinrichs, Kevin A. Gary

179

Cybersecurity Capstone Case Study: Closing the Loop on Technology Competency Literacy Lethia Jackson, Velma Latson, Haydar Teymourlouei

186

BATTLE 2018: Preparation of Future Cyber Technologists Clarence Ray, Jesse Bemley

195

Investigating Team Effectiveness Using Discord: A Case Study Using a Gaming Collaboration 199 Tool for the CS Classroom Lisa L. Lacher, Cydnee Biehl Improving STEM Performance by Leveraging Machine Learning Models Mohamed Aly, Mohammad Rashedul Hasan

205

VRSafe with Augmented and Virtual Reality: Toolkit for Harassment Prevention and Deescalation - Sensitivity, Training and Best Practices Lamia A. Djoudi, Indu M. Anand, Antoine Luu, Diego Galar, Ana Samedo, Linda Maisano

212

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

SESSION FRONTIERS IN EDUCATION: TEACHING METHODOLOGIES AND STRATEGIES, ADVISING METHODS, TOOLS AND RELATED STUDIES Chair(s) TBA

ISBN: 1-60132-498-7, CSREA Press ©

1

2

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

3

Data-Driven Models to Predict Student Performance and Improve Advising in Computer Science Varick L. Erickson Computer Science California State University, East Bay [email protected] Abstract—As enrollment in computer science (CS) majors continues to rise, CS departments are facing increased pressure to accommodate larger numbers of students, improve student graduation rates, and reduce time to graduate. As a result, effective student advising is critical to ensure the CS major is the right fit for students and to help students complete the degree in a timely and successful manner. Additionally, advisers play a crucial role in identifying struggling students and ensuring at-risk students either improve or quickly transition to a different major (saving resources both for the student and for the department). However, it is challenging for CS advisers to identify struggling students early. Often, by the time at-risk students are identified, these students have expended a considerable amount of time, effort, and money in the major when they could have been pursuing a major more suitable to their strengths. In the present work, we analyze actual enrollment data and develop models for predicting student performance. By leveraging these data-based models, we can identify at risk students with up to 77% accuracy. Based on these models, we also propose approaches to improve student advising in CS undergraduate programs. Index Terms—Education, Success, Prediction, Classification, Advising, Modeling

I. I NTRODUCTION From 2010-2017, the number of computer science (CS) degrees conferred has increased 50.7% [1]. While many students are interested in pursuing a CS degree, not all students are prepared or suited to the degree. In fact, 28% of students who start in CS will change majors within the first three years [2]. Since CS programs are increasingly overcrowded, it is crucial to help students succeed and graduate in a timely manner. Similarly, students who are having difficulty with the computer science degree need to be offered support in order to quickly improve, or be advised to find another major before investing too heavily into the degree. The sooner a student realizes that CS is not the right fit, the sooner the student can transition to a better fit major, and the sooner the CS department can free up limited space and resources to help additional CS students succeed. Given the increasing enrollment and high student turnover in CS degree programs, identifying and advising struggling students early in the students’ academic careers is invaluable. Many CS students fail to realize that CS is the wrong major for them until several semesters or even years have passed. Then, these students feel they are “too far along” to change majors. While some struggling students are ultimately able to complete the degree, it often takes far more time and repeating classes several times to complete the degree. With low GPAs, these

students may also have difficulty succeeding after graduation. Another common reason for staying in the CS major too long is an issue called identity foreclosure. Identity foreclosure refers to situations where students establish goals without first doing thorough exploration of options and self reflection [3]– [5]. Students experiencing identity foreclosure can identify so strongly as CS majors that they refuse to consider alternate majors, even if these students are failing CS courses or not the right fit for the program. As a result, identity foreclosure causes some students to view the CS major as their only option, even if they are struggling. Having an academic adviser objectively examine each student’s performance and advise appropriately has been shown to make a significant positive impact. For example, advising has been shown to enable persistence rates (retention rates) of 53% among underrepresented students [6], [7]. Therefore, it is critical to flag at-risk students as early as possible. Once these students are identified, an adviser can look more closely at a student’s situation and provide better guidance. Timely and effective adviser support can make the difference between having a struggling student stay in the program and ultimately fail or drop out, helping a student by providing access to support and resources, or helping a student move to a different major. There are numerous advantages to an early major change. Arguably the most important advantage is decreasing the number of repeated classes. Students who stay too long in the CS major often need to repeat courses. This means affected students face a longer time required before graduation and potential financial challenges that could ultimately lead to abandoning an undergraduate degree. Students who repeat courses also tend to have lower GPAs even if they replace a grade or repeat a course. Based on our data, the average CS GPA of students who needed to repeat a CS course one or more times is 2.54 (on a 4.0 scale). In contrast, students who only needed to take CS courses once had an average GPA of 3.00. Having struggling students repeat courses also negatively affects CS departments and successful students. When struggling students repeat a course, fewer seats are available for those taking the course for the first time. Again, having fewer seats available can lead to a longer amount of time spent in school and increased financial burden on successful CS students who want to be in the program. Thus, when advisers are able to

ISBN: 1-60132-498-7, CSREA Press ©

4

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | identify and better support at-risk students, all students and the department can benefit as well. In the present work, we aim to improve student advising by analyzing CS student data, developing predictive models for student performance, and proposing approaches to improve advising. First, the academic data for students at California State University, East Bay was compiled and analyzed. Various models were applied to identify and predict at-risk students. Based on these models, we are able to identify at-risk students with up to 77% accuracy. Based on these results and findings in the data, we also propose additional approaches to help CS departments better advise and support at-risk students. We start with a survey of related work in Section II. We then discuss the choice of variables in Section III followed by details of data collection in Section IV. A preliminary data analysis is performed in Section V that motivates our model selection described in Sections VI, VII, and VIII. We compare these models in Section IX and then show how these models can be used in an advisory setting in Section X. Finally, we discuss future work in Section XI and summarize our findings in Section XII. II. R ELATED W ORK Researchers have studied predictive models for succeeding in CS for over 30 years. Many of these papers focus on predicting CS performance prior to entering a university [8]– [10]. Authors of [8] attempt to predict performance at the high school level before entering a university program. Similarly, authors of [10] look at a broad index of 62 items to determine learning traits described by Kolb and Pask [11] in order to predict how high school students will perform. Unlike these papers, our goal is to help advisers and current CS students predict success based on recent performance in CS courses. In [12], authors use a linear regression approach to predict how well students will perform based on past performance. However, there are notable differences as compared with our approach. The authors found that a linear regression model was able to describe their population, which was CS majors at the U.S. Air Force Academy. In our case, we determined a linear regression approach was not suitable based on our residuals. It is likely that hidden factors are not being captured by our data. Additionally, the authors of [12] use a somewhat cumbersome approach, generating 99 models to capture each valid combination of previously taken classes. Our single model is able to capture all combinations of classes taken by treating variables as categorical data rather than continuous data. In the present work, we also introduce a ”did not take” label to more accurately model cases where students did not complete a given course. The authors of [13] have a similar aim as our paper and focus on identifying struggling students early. To predict student performance, they use an integrated development environment (IDE) extension that continuously logs student coding activities. This IDE classifies each student’s successes and failures and uses this data to dynamically refine a prediction model. The authors also compare their approach to

Simplified Label CS1 CS2 DataS Calc1 Calc2 L

Description First quarter course on programming Second quarter course on programming Data Structures and Algorithms First course in calculus Second course in calculus Linear Algebra

TABLE I: The courses used to predict student success. The rightmost column is the course name used by the university. Note that the courses in the dataset used the quarter system. a similar method described in [14], which also utilizes a similar dynamically adaptive classifier. In the present work, our approach is simpler, only requiring data already collected by universities and registrars without the use of specialized plugins. Despite the added complexity, both methods examined have a maximum of 75% accuracy. In contrast, the models proposed in the present work achieve up to 77% accuracy for identifying at-risk students. III. I NDICATOR AND R ESPONSE VARIABLES In this section, we discuss our choice of indicator and response variables. Our goal is to be able to predict student success in the major early using a model. To define a response variable, we must first start with our definition of success. There are many different possible metrics for success. One possible metric is overall GPA for CS courses. Another could be the time to complete the degree. For this paper, we decided to examine classes most closely associated with CS interviews. In particular, we wish to use student performance in the Data Structures course as our response variable since interviews often use concepts from this course [15]. Also, Data Structures is a key prerequisite to many courses required by the major. Our approach is also applicable to measuring success for Analysis of Algorithms, which is another popular topic for CS interviews. However, since our goal is to flag at-risk students as early as possible, we define success in Data Structures to be our response variable since Data Structures is typically taken earlier (during a student’s second year). From an advisory perspective, our aim is to be able to predict Data Structures performance as early as possible, so we are able to identify and advise students early in the program. Thus, logical choices for indicator variables would be collected from required lower division courses that are taken prior to data structures. TableI shows these courses; for clarity, we define simplified course labels rather than use the official university course numbers. Our hypothesis is that performance in these earlier classes will be able to predict performance in Data Structures. IV. DATA C OLLECTION To measure student success in these courses, anonymous registrar data was collected from 924 CS undergraduate students at California State University East Bay (CSUEB). The courses were taken between 2012-2018 and used the quarter system. Table II shows the demographics of CS students at CSUEB. It is important to note that many of the students

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | Gender Male Female

5

75.6% 24.4% White

Hispanic 12.1% 14.1%

Black Multiple Races Hawaiian

3.5% 3.7% 1.14%

29.4%

Asian

36.0%

Unknown

TABLE II: Student Demographics [16] for CS Department 2017-2018. at CSUEB are first-generation college students, and many of them also work full or part-time while attending university. While the data in the present work captures student courses taken, grades, and GPA, the data set used in the present work does not include any data for factors such as whether a student was working full or part-time, or whether the student was a first-generation college student. V. P RELIMINARY DATA A NALYSIS We start the analysis by examining the performance correlations among the lower division courses. Figure 1, shows the correlation matrix for the first attempt grades for Data Structures (DataS), CS1, CS2, Calculus 1 (Calc1), Calculus 2 (Calc2), and Discrete Math (DMath). The diagonal shows the histogram of the grade for the given course. Interestingly, only a weak correlation is seen between CS1 and Data Structures (0.09). However, we do see a stronger correlation between Data Structures and CS2 (0.27). A surprising observation is that Calculus 2 has a stronger correlation to Data Structures than CS1 or CS2. Perhaps most notable is that even the strongest correlations are fairly weak. Different combinations of interactions were also examined and also showed fairly weak correlations. While these correlations are weaker than one might expect, there are some possible explanations for this result. As mentioned earlier, the data set used in the present work captures only student academic data and does not capture external factors such as whether a student was working full-time while taking classes. Stronger correlations my be possible if more data were available. For our first model, we examine a linear regression model with the Data Structures grade as the response variable and the indicator variables from Table III using only performance from CS1 and CS2. Note that the course grade is represented by a number from 0 to 4, with 4 being an ”A” or 4.0 grade and

Fig. 1: Correlation Matrix of Data Structures (DataS), CS1, CS2, Calculus 1 (Calc1), Calculus 2 (Calc2), and Discrete Math (DMath) with respect to first attempt grades. Here, the first attempt grade is plotted as a number ranging from 0 (”F”) to 4 (”A”). Based on this matrix, we see relatively weak correlations. There is a maximum of correlation of 0.37 between Calc2 and DataS and a minimum correlation of 0.09 between CS1 and DataS. Indicator Last Grade Initial Grade Minimum Grade Maximum Grade

Description Grade last earned Grade initially earned Minimum grade from all attempts Maximum grade earned from all attempts

TABLE III: Performance indicators examined for each class. 0 being an ”F”. Unsurprisingly, none of the variables showed coefficients that were statistically significant. In addition, a plot of the residuals (see Figure 2) indicates that there is likely are one or more hidden indicator variables not being captured that are needed to accurately predict Data Structures grades. This could include variables such as how many hours the student works per week, how far the student has to commute, or other factors not present or easily determined with the existing data. These results show that a linear model is not suitable for predicting letter grades in the data structures course, at least not with the current data set. Intuitively, we expect past performance in classes to offer some predictive measures. Rather than predict grades, we instead look at the probability of passing (achieving a grade of C- or better). Table IV shows several conditional probabilities of passing Data Structures given the student’s performance in past classes. For example, if a student fails CS1 in the past,

ISBN: 1-60132-498-7, CSREA Press ©

6

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | P (DataSpass |CS1fail ) = 0.2768 P (DataSpass |CS2fail ) = 0.2656 P (DataSpass |CS1fail and CS2fail ) = 0.1667 P (DataSpass |Calc1fail ) = 0.2676 P (DataSpass |Calc2fail ) = 0.3219 P (DataSpass |Calc1fail and Calc2fail ) = 0.1951

TABLE IV: Calculated conditional probabilities of passing Data Structures given different scenarios. Linear Regression Residuals 3

2

Residuals (CS1)

1

0

-1

variable is binary indicator whether or not the student would pass Data Structures given their initial grades. One issue to consider is the different combinations of classes that students can take. Depending on the order that the classes are taken, we would need a model specific to the given permutation. For example, if a student takes CS1 only, then the model predicting the performance of Data Structures would be the single indicator variable using the CS1 grade. Similarly, if a student has taken Calc1 and CS1 but none of the other courses, then we need a model with Calc1 and CS1 indicator variables for prediction. While class prerequisites do reduce the number of possible permutations, this multiple model approach would require numerous models to capture each situation. To avoid this issue, we instead treat each variable as a categorical variable with each grade being a different classification label.

-2

VII. L OGISTIC R EGRESSION -3

0

0.5

1

1.5

2

2.5

3

3.5

4

True response

Fig. 2: The residuals plot shows violation of linearity. The linear trend of residuals indicates there are likely hidden variables not accounted for by the linear regression model. we see that the student has a probability of 0.2768 (about 28% chance) of passing Data Structures. We see a similar statistic for CS2 with a 0.2656 probability (about 27% chance) of passing Data Structures if the student previously failed CS1. Unsurprisingly, a student who failed both CS1 and CS2 in the past has an even lower probability of 0.1667 (about 17% chance) of passing Data Structures. Interestingly, Calc1 and Calc2 courses show similar conditional probabilities as compared with the CS1 and CS2 conditional probabilities. Students who fail Calc1 have a 0.2676 probability (about 27% chance) of passing Data Structures. Given a student failed Calc2, the student has a 0.3219 probability of passing Data Structures. If the student failed both Calc1 and Calc2, then the student has a 0.1951 probability (about 20% chance) of passing Data Structures. This initial result shows that early grades can indeed be a useful predictor of student performance in future courses. VI. M ODEL S ELECTION Rather than predict student grades for data structures, we instead choose to predict whether students are able to pass Data Structures. Since we are attempting to predict if a student will pass Data Structures based on past class performance, we can formulate this as a classification problem. In this section, we examine two different strategies for predicting whether a student will pass based on the grades of their previously taken courses. In particular, we examine logistic regression (LR) and support vector machine (SVM) classification models. For our indicator variables, we use the initial grades for CS1, CS2, Calculus 1, Calculus 2, and Linear Algebra. Our response

The first model we will examine is a logistic regression model, which uses a logistic function to predict a binary value. For this model, we define the logistic function l to be, l = β0 + β 1 x1 + · · · + βn xn

(1)

where βi are beta coefficients for the parameters of the model and xj is the value of the indicator variable j for 1 ≤ j ≤ n. Using l, we can calculate the probability p of binary value of a response variable by calculating p=

1 . 1 + e−l

(2)

As previously discussed, instead of treating the class grade as a continuous variable and have a different models for each permutation, we instead interpret the grade as a categorical variable with 13 different categories: F, D-, D, D+, C-, C, C+, B-, B, B+, A-, A, and “did not take”. The advantage of this approach is that it allows for different combinations of classes to be captured by a single model. To construct this model, we create 12 “dummy” variables for each class. Thus, we define the model L as l =β0 + β1 CS1F + β2 CS1D− + · · · + β12 CS1A + β13 CS2F + β14 CS2D− + · · · + β24 CS2A + β25 Calc1F + β26 Calc1D− + · · · + β36 Calc1A +

(3)

β37 Calc2F + β38 Calc2D− + · · · + β48 Calc2A + β49 LF + β50 LD− + · · · + β60 LA where βi are beta coefficients for the parameters of the model and β0 is the intercept. CS1g , CS2g , Calc1g , Calc2g , and Lg are 1 if they received grade g or 0 if they did not. If the student did not take the class at all, then, each value is 0. For example, if a student did not take Calc1, then for each value of Calc1g is 0 for g = F, D−, . . . , A.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | VIII. S UPPORT V ECTOR M ACHINE The next classification model tested is a Linear Support Vector Machine (SVM). This model is a supervised learning method that is able to make a classification to specific group given an input vector. Unlike the logistic regression model, it is a non-probabilistic and does not attach a probability to a classification. For this model, we define our training data as z = (x1 , y1 ), . . . , (xn , yn )

(4)

where xi is a vector of categorical inputs (CS1F , . . . , CS1A , CS2F , . . . , CS2A , Calc1F , . . . , Calc1A , Calc2F , . . . , Calc2A , LF , . . . , LA ) with each position in vector xi a 0/1 indicator whether the student received a given grade for a class. This is essentially the same input vector used by the previous LR model. For example, if the student received an A for CS1, then the value of CS1A is 1 and all other values of CS1g is 0 for g = F, . . . , A−. Again, if the student did not take the class at all, then all the values for that particular class are 0. yi is a 1 or -1 label specifying what class xi belongs for data point i. Using the training data in equation 4, we find a boundary line satisfying w · z − b = 0, (5) where w is a vector normal to the line. This boundary line defined by equation 5 allows us to classify new data points. IX. M ODEL C OMPARISON In this section, we examine and compare each model. In particular, we will be examining the ROC (Receiver Operating Characteristics) curve, the AUC (area under the curve), and the confusion matrix. The AUC and ROC curve are typical performance measurements for classifiers as they show how well the model is capable of distinguishing the different classes [17]; in our case, we are determining how well the models can determine if a student will pass or fail Data Structures. An AUC value of 1 indicates perfect classification. This metric is particularly useful for comparing two different classifiers. The other metric we use is positive predictive value and false discovery rate, which can be derived from a confusion matrix. The positive predictive value (PPV) is the probability that the student who is predicted to fail data structures does indeed fail the class. The false discovery rate (FDR) is the complement of the PPV (F DR = 1 − P P V ). Figure 3a shows the results of the LR model. We see that the AUC is 0.67 and the PPV is 77% for failing and 51% for passing. This indicates that a student predicted to fail data structures will with 77% probability fail the course. Similarly, it show a student predicted to pass will pass with 51% probability. Figure 3b show the results of the SVM model. For this model, we see the AUC is slightly higher

7 with an AUC of 0.73 as compare with the LR model; this indicates the SVM is a better classifier. Interestingly, the SVM model shows the same PPV as the LR model. The SVM is able to predict with 77% probability if students will indeed fail data structures. However, the SVM is better able to predict students passing data structures with a 57% PPV for predicting if students will pass data structures. As our aim is to identify students likely to fail, we see that both LR and SVM have the same prediction probability for predicting a student failing data structures. As the SVM does slightly better classifying students who pass and also has a higher AUC, we consider the SVM a better prediction model. X. M ODEL U TILIZATION In this section, we explore how the SVM prediction model is able to be utilized for advising. We first examine the impact of course repeats on GPA and the seats available for students. Table V shows several statistics of interest. The first column shows the average number of times students attempt a particular course. We see that on average students take CS1 1.143 times. This is significant since it means that approximately 14.3% of the seats in a given semester is occupied by someone repeating the course. We find similar results for CS2 where the average number of repeats is 1.149. Data structures, Linear Algebra, and Calc2 have average repeats of 1.160, 1.160, and 1.172 respectively. The highest average for course repeats is for Calc1 with an average of 1.207. These statistics support that a non-trivial number of course repeats are occurring and reinforce the need to reduce the number of students repeating a course through advising. The second column of Table V shows the average GPA of a course after repeating the course. This is calculated by finding the maximum grade from all attempts and averaging this across all students. The third column shows the average GPA of students who only take the course a single time. We see across all courses that students who need to repeat the course on average will earn a lower grade that students who only take the course once. This implies that even if a student repeats a course, it is likely they will earn a lower grade compared to a student who only needs to take the course once. Our next analysis examines several different strategies for utilizing the models. We also examine how better advising Course CS1 CS2 DataS Calc1 Calc2 L

Average Repeats 1.143 1.149 1.160 1.207 1.160 1.172

Average GPA After Repeat 2.88 2.42 2.40 2.34 2.23 2.29

Average GPA No Repeat 3.09 2.85 2.82 2.69 2.65 2.64

TABLE V: This table shows the average number of repeats and the impact repeating classes has on student GPA. The first column shows the average number of repeats per course. The second column shows the average GPA of students if they choose to repeat a course. The third column shows the average GPA of students when they need to take the course only once.

ISBN: 1-60132-498-7, CSREA Press ©

8

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ROC: SVM

1

1

0.8

0.8

True positive rate

True positive rate

ROC: Logistic Regression

0.6

AUC = 0.67 0.4

0.2

0.6

AUC = 0.73 0.4

0.2

0

ROC curve Area under curve (AUC)

0

0.2

0.4

0.6

0.8

1

0

ROC curve Area under curve (AUC)

0

0.2

False positive rate

0.4

0.6

0.8

1

False positive rate

Confusion Matrix Predicted Pass Fail Pass 51% 23% Fail 49% 77%

Confusion Matrix Predicted Pass Fail Pass 57% 23% Fail 43% 77%

Positive Predictive Value

51%

77%

Positive Predictive Value

57%

77%

False Discovery Rate

49%

23%

False Discovery Rate

43%

23%

(a) The first plot show the ROC plot and AUC value for the logistic regression model. The second shows the confusion matrix for the logistic regression model and the positive predictive value.

could affect the average pass rate of data structures and the number of seats that would become available. The first proposed strategy is instituting a rule allowing students to repeat introductory courses a maximum of two times. In other words, if a student did not pass an introductory course after two attempts, they would be advised to pursue a different major or not to continue the program. Table VI shows the seats used and potentially saved applying this threshold. The first column shows the number of potentially saved seats. The second column shows the total number of seats occupied for a particular class. For CS1, we found 7 students who took CS1 two more or more times meaning 7 seats would be saved in this situation. We found roughly twice as many potential seats in CS2 with 15 saved. Data structures we found 17 potentially saved seats. Calc1, we found the most number of potentially saved seat; we found 18 CS students repeating the class more than twice. For Calc2 and linear algebra, we found 12 and 13 potentially saved seats respectively. The next to last row of Table VI shows that the maximum number of saved seats in total is 153 (or 3.3% of the total number of seats). Note that the maximum number of saved seats is higher than the sum of the classes individually. This is because students who repeat an earlier class would not take subsequent classes. For example, a student who unsuccessfully attempted CS1 twice would not attempt any other prerequisite courses. The last row shows a potential savings of $244,800 dollars in tuition cost over six years assuming an cost of $1,600 per class, which

(b) The first plot shows the ROC plot and AUC value for the SVM model. The second shows the confusion matrix for the SVM regression model and the positive predictive value.

is a savings of approximately $40K per year. While this first strategy does improve the number of potential seats, it also does not address students who consistently repeat courses. For example, a who routinely takes courses twice also indicates a need for advising. Our second strategy is expands the first strategy by also taking into account the total number of repeated courses. Table VII shows the results of imposing different thresholds allowed for maximum number of repeated courses while still imposing a two attempt maximum for any lower division course. The first row shows 137 students had to repeat 2 or more courses. If this threshold were applied, then 137 students would need advising and we would potentially save 369 seats (8% of the total seats) possibly saving $590,400; this is arguably too aggressive a threshold and not practical. The second row shows 40 students had to repeat 3 or more courses. Applying this threshold would lead to potentially saving 184 seats (4.0% of the total seats) and saving potentially $294,400; this rule adds 31 additional seats as compared with the two attempt strategy. This threshold seems to be the most reasonable. Thresholds of 4 only shows 8 students meeting this threshold. We also see in this rule only adds 2 additional seats over the two limit strategy and saving only $3,200 additional dollars. Similarly, only 2 students meet the 5 repeat threshold and no additional seats were gained as opposed to the two attempt strategy alone. Therefore, based on this data, we find that imposing a limit of two repeats per course and three repeats maximum across all

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Course CS1 CS2 DataS Calc1 Calc2 L Potentially Saved Seats Potential Tuition Savings

Number of Students repeating more than twice 7 15 17 18 12 13 153 (3.3%) $244,800

TABLE VI: This table shows how many potential seats can be saved using a limit of two attempts for a class. Compounded, this would save 153 seats (3.3% of the total seats) in classes of students who would otherwise would be attempting a course for a third or more times. This would save approximately $244,800 in tuition over about six years. Repeat Threshold 2 3 4 5

Number of Students 137 40 8 2

Potentially Saved Seats 369 (8.0%) 184 (4.0%) 155 (3.3%) 153 (3.3%)

Potential Tuition Savings $590,400 $294,400 $248,000 $244,800

TABLE VII: This table shows how many potential seats can be saved if we impose restrictions on course repeats in addition to a two attempt per course restriction. Each row shows how many students repeated courses at a given maximum repeat threshold and the potential for saving seats. For example, in row 1, we see 137 students needed to repeat at least 2 prerequisite courses during their degree. introductory courses would be a good first step to help reduce the number of repeated courses. Students at risk of hitting the limits on repeated courses could be flagged in the system and offered additional adviser support. XI. F UTURE W ORK These predictive models offer a useful means to flag struggling students earlier in their academic careers. It is important to note that while the models are able to predict whether a student will fail Data Structures with up to 77% accuracy, the models were less accurate in predicting whether a student will pass Data Structures. These models therefore should be used as tools in combination with other information in order to make the best possible recommendations to students for advising. For example, if a student has a low probability of passing Data Structures given their past performance, an adviser could also look at whether the student is also working full-time, whether the student has access to supporting resources, and other factors in order to provide the best possible guidance. Another interesting observation is that CS students tended to fail and repeat the calculus and linear algebra courses twice as often as the CS courses. As mentioned earlier, students who failed both calculus I and II had only a 20% chance of passing Data Structures. Future work could investigate additional indicators of success and identify approaches to help struggling students better perform in these classes. While the focus of the current work was to predict atrisk students early on by focusing on the Data Structures course, future work could also examine whether a similar

9 approach could be applied to the Analysis of Algorithms course and other upper division courses. Additionally, it could be interesting to attempt to capture other variables in the data set, such as whether a student was working full-time or was a first generation student, to attempt to improve the models. Finally, future work should explore actions advisers and departments can take to better support students based on the predictions and data in this paper. For example, it could be beneficial to explore the impact of offering better access to tutoring or other support, or to implement department policies to more strictly enforce prerequisites and minimum grades for critical courses in CS. By combining the learning in the present work with improved advising and support, CS advisers and departments can make significant progress in improving time to graduate, reducing overcrowding, and improving graduation rates. XII. C ONCLUSION In this paper, we explore two different classification based prediction models for predicting CS success and identifying students in need of advising. We show that we are able to identify students with high risk of failing with 77% accuracy. By utilizing these models and the recommended advising approaches, we show that we could potentially reduce the number of repeated courses by 4.0%, make 184 additional seats available for other students, and save approximately $250K for the 924 students in the study. R EFERENCES [1] Digest of Educational Statistics. National Center for Education Statistics, 2017. [2] K. Leu, Beginning College Students Who Change Their Majors Within 3 Years of Enrollment. National Center for Education Statistics, 2017. [3] L. S. Shaffer and J. M. Zalewski, “Its what i have always wanted to do. advising the foreclosure student,” in NACADA Journal, 2011. [4] J. S. Albright, K. M. Martel, and B. D. Webster, “No more missed opportunities: Using the foreclosure model to advise pre-nursing and nursing students,” 2012. [5] R. K., “Applying career and identity development theories in advising,” 2015. [6] K. Klepfer and J. Hull, High school rigor and good advice: Setting up students to succeed. Center for Public Education, 2012. [7] T. Ross, G. Kena, A. Rathbun, A. KewalRamani, J. Zhang, P. Kristapovich, and E. Manning, Higher education: Gaps in Access and Persistence Study. National Center for Education Statistics, 2012. [8] D. F. Butcher and W. A. Muth, “Predicting performance in an introductory computer science course,” Commun. ACM, 1985. [9] L. Werth, “Predicting student performance in a beginning computer science class,” SIGCSE, 1986. [10] D. Capovilla, P. Hubwieser, and P. Shah, “Dics-index: Predicting student performance in computer science by analyzing learning behaviors,” in LaTiCE, 2016. [11] D. Kolb, “Learning styles and disciplinary differences,” 1981. [12] A. Chamillard, “Using student performance predictions in a computer science curriculum,” in ITiCSE, 2006. [13] C. Watson, F. W. B. Li, and J. L. Godwin, “Predicting performance in an introductory programming course by logging and analyzing student programming behavior,” in ICALT, 2013. [14] M. C. Jadud, “Methods and tools for exploring novice compilation behaviour,” in ICER, 2006. [15] G. L. McDowell, Cracking the Coding Interview. CareerCup, 2015. [16] [Online]. Available: http://www.csueastbay.edu/faculty/senate/files/docs/5year-reviews/17-18/17-18-comp-sci-5-yr-review.pdf [17] T. Fawcett, An introduction to ROC analysis. Pattern Recognition Letters, 2006.

ISBN: 1-60132-498-7, CSREA Press ©

10

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

An Introductory Visualization Aid for Cybersecurity Education Gabriel Castro Aguayo1 , Ulises Morales1 , Xiaoyu Long1 , Quamar Niyaz1 , Xiaoli Yang1 , Ahmad Y Javaid2 1 ECE Department, College of Engineering and Sciences, Purdue University Northwest, USA 2 EECS Department, College Of Engineering, The University of Toledo, USA {gcastroa, umorales, long312, qniyaz, yangx}@pnw.edu, [email protected] Abstract— As the technology keeps overgrowing, the Internet surfing becomes more popular. As a consequence, users tend to use it for social media, shopping, banking or any other online services in which they need to put their personal information. These online activities attract malicious computer users to apply cyberattack techniques to steal other users information. The users become attack victims due to limited understanding of cyberattacks and safety practices. In this paper, we propose a framework development for interactive and engaging cybersecurity education. With the help of the framework, the users will be able to learn different types of cyberattacks and defenses along with the safe cybersecurity practices. We also discuss the current state of the framework and conclude the paper with discussion on limitations and future work. 1 Keywords: cybersecurity education, visualization framework, Unity 3D

1. Introduction With this technology boom, most people tend to be on their smartphone or computer surfing online nowadays. Even though the Internet makes our lives easy by providing online services such as allowing customers to make purchases online faster than actually going to the stores, some misfortunes can occur when we are using those services. As more people rely on the computers, the more vulnerable they become to any attack. Malicious computer users can find ways to steal credit card or personal information by using malware or backdoor. According to Symantec 2018 report [1], 27% mobile apps in lifestyle category were malicious. That is not the end, the number of new malware variants for smartphones have increased. According to an article in [2], the cyberattack victims lost around $1.5 billion in 2017 and more than 300K Internet crime complaints were reported to the Internet Crime Complaints Center in the same year. The same article mentioned that 23% of Americans in a conducted survey reported that credit card information of their family members were stolen by the hackers. There are some good practices that one can use to identify these malicious attacks and be properly secured. One good practice is by being observant of any changes on the device or have security software to assist with some automated scan. Another 1 Disclaimer:

This work uses popular characters (such as Pikachu) ONLY for educational purposes. There is no intention of copyright infringement. Names, characters, businesses, places, events, locales, and incidents are not our copyright and have been only used to attract young kids towards cybersecurity education. Any resemblance to actual persons, living or dead, or actual events is purely coincidental.

practice is to observe for any unusual activity in a smartphone, such as obsessive data usage or even overheating. Besides, smartphones record events such as installation and data usage that can help narrow down the application responsible for any malicious activities. Practicing cybersecurity methods can help reduce the potential of having information stolen from attackers. However, there is a lack of cybersecurity safety practices among the users. Therefore, it is imperative to spread the cybersecurity awareness through different channels and educate the users how to avoid cyberattacks. Having users practice visually can be beneficial to teach them safe practices of using technology and being online. As cybersecurity visualization can be effective for security analysts to prevent a malicious attack, it can also be helpful to engage users in the learning process of cybersecurity threats and defenses [3]. With this motivation, we propose a visualization-based cybersecurity education framework that will help users, especially the teenagers in middle and high schools, understand the cybersecurity issues they may run into. It will allow them to learn about the attacks that they are exposed to when navigating online, so they can prevent these misfortunes. There are two main reasons to focus on users from early ages. First, they tend to be the future of our society, so they might be the next generation of cybersecurity specialists who might be able to prevent more advanced cyberattacks. Second, the increased usage of smartphones and the Internet by teenagers has brought several security concerns for them and their families. With the increased time spent online, teens frequently encounter cyberbullying or unpleasant experiences. According to a survey conducted by McAfee, 34% teens acknowledged that they have experienced cyberbullying. The same study mentioned that 39% teens do not adequately set their privacy settings [4] for different online applications. The developed framework consists of user-friendly graphical interfaces with different environments to provide cybersecurity knowledge and safety practices to middle and high school students. As the primary target is teenagers for this framework, the development of the personal computer (PC) as well as the tablet/smartphone versions will be available to fulfill every user needs. The rest of the paper is structured as follows. In Section 2, we briefly discuss the related work. Section 3 discusses the application design of the framework. In Section 4, we discuss various visualization-based cybersecurity education topics implemented at introductory level in the framework. Finally, we conclude our paper in Section 5.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

11

2. Related Work Learning through visualization technologies using computers are mature and widely used in various domains. Learning using tablet/smartphone devices is a relatively new area. There have been studies that involved creating learning applications for smartphone devices, especially for new languages [5]. Apps like Duolingo, Busuu, and many others teach new languages to any smartphone user who installs and uses the apps. Practically anything can be taught nowadays as long as the targeted app is developed for the common user. There are a variety of teaching apps from learning a new language to learn new software. In addition, there are apps that are being used for distance learning [6]. With every student owning at least a computer either PC and/or a smartphone device, any application can be developed to teach the students a subject. There is a group that created a popular Capture The flag (CTF) game for high school students to learn technical concepts by creating a competitive environment [7]. They introduced it in a GenCyber [8] camp and this motivated the students to continue learning after the camps because of the competition-based game format. With introducing serious game into learning, we bring the concept of visualization and interaction into the design of our framework, as it actively engages users to learn subjects. The learning tool will teach each topic by involving the user in active learning through interactive simulation. In addition, the framework includes assessments to evaluate what the user has learned.

Fig. 1: Menu screen for the topics and sections in the framework

3. Application Design In order to develop the application framework, various software engines were analyzed, and Unity 3D [9] was found as the best fit. Unity has a variety of tools that offer developers a friendly environment to work on. Although the final design tends to be for smartphone platform, the application will first be built for PC platform to test it prior to the final delivery. Fortunately, Unity allows user to change the platform easily. In addition, the engine provides an integrated development environment (IDE) to write code as well as high visual and audio effects. Therefore, Unity is the best choice to develop the application. As previously mentioned, the application will be focused on PCs for this design. It will be modified to accommodate smartphone devices upon completion. The application will have a menu, shown in Figure 1, that will help the user to navigate through different topics and be aware of what topic/section they are on. The menu will be made generic so more topics and subtopics can be added in the simplest way possible. It will have the effect of shrink and expand list to display the sub-topics so that the user is not overwhelmed. Each topic will consist of four stages for the users: i) introduction, ii) interaction, iii) explanation, and iv) assessment. The flow chart can be seen in Figure 2, as to how the flow of the learning module will be for the user for each topic. These tasks are important for the user to understand and evaluate what they are learning. The introduction will be a brief description of what the topic is about; the best way to grab a user’s attention is to have an

Fig. 2: Stages involved in each learning module

animated character with a short story or description. Even with a little interaction, it is enough to keep the user entertained and learning [10]. The animated introductions will be different for each topic so that the user does not become disinterested with the same animations. The interaction is more of a challenge due to the fact that the technical terms and procedures have to be displayed in a way that the user can understand. The display is not the only concern, but the interaction part as well where the user can learn from the stories and get a better understanding of the process and concepts implicitly by interacting with the environment. Moreover, the interactive part should not be long else it will overwhelm the user with more detailed information. With having topics and sub-topics, each interaction will be short and straight to the point, therefore still giving the students the ability to interact and learn the main key points. The explanation task will be more in depth where the key points are being explained. This is where the user will get a complete understanding of what they have learned in the interactive task. Just as the introduction, it

ISBN: 1-60132-498-7, CSREA Press ©

12

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 3: Learning modules, information (top left), interaction (top right), explanation (bottom left), and assessment (bottom right) will involve some animation and very minimal interaction in order to have the users complete attention. The explanation will also go over parts that will be reviewed in the quiz. In the assessment, completion of the quiz will be recorded in order to know the progress of the user. There will be a set of questions for each topic, however only a few questions will be asked at random so the same questions are not repeated each time. When the progress is recorded there will be a progress bar that will demonstrate how much the user has completed successfully. Before releasing the final application, testing will be fundamental to assure that the software can be a great tool to acquire the desired cybersecurity knowledge. Testing will be performed to a certain group of students with basic knowledge of cybersecurity who will analyze the application and provide important feedback. Figure 3 shows an example of the four stages of a learning module.

4. Visualization-based Education Topic There are wide range of topics in cybersecurity. To get started with the framework development, we have considered network and web security. In network security, we have started with TCP and DNS protocols (discussed below) and associated attacks with them. For web security, we focused on cross-site web attacks and phishing.

4.1 Network Security Visualization 4.1.1 DNS Illustration The Domain Name System (DNS) is used to resolve domain name, e.g. www.pnw.edu to an IP address [11]. It is requestresponse protocol that runs on the application layer of TCP/IP Internet stack. DNS is unanimously used for resolving domain names due to its distributed implementation. For DNS visualization, following questions had been considered: • What is DNS and how does it work? • How to make the DNS process easy to explain? • How to visualize the attacks associated with the DNS? These questions are answered in the framework. We chose two typical DNS attacks: amplification and hijacking. In the DNS amplification attack, attackers use open DNS servers to flood a victim’s network with a large number of DNS reply messages [12]. The attacker sends DNS look-up queries to a large number of open DNS servers by spoofing the IP address of the victim as source. As a consequence, all DNS responses are sent to the victim. In DNS hijacking, the attacker sends the domain resolution query to a different DNS server instead of the actual ones. For the introduction stage of DNS module visualization, a daily situation has been taken for illustration. How would you

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 4: Introduction and interaction stages for DNS module

find the location of a new address? A man called Bob has to find the address for Google’s office “1600 Amphitheatre Pkwy, Mountain View, CA 94043” (in an age with no Google Maps). To visit Google, he goes to the local community office to obtain knowledge of the new address; the local community office represents the local server. The local community office gives a part of the address to Bob such as that the Google’s office is in California state. Then, Bob goes to California and asks the state government staffs, city staffs, and Google’s neighborhood as shown in Figure 4. These represent root server, TLD (Top Level Domain) server and domain nameserver, respectively. In the DNS amplification attack, the interaction is based on the DNS concept and developed on top of it. In DNS hijacking interaction, it shows that a dishonest staff will provide wrong information. According to the interaction, users can easily understand the DNS process and how attacks may happen through DNS. 4.1.2 TCP Illustration Transmission Control Protocol (TCP) provides reliable communication between two hosts that is achieved by a “three-way handshake” [13]. There are three steps in this handshake process. First, the client sends a connection request to the server using some initial sequence number, and a window size for the buffer used by the client to store the packets coming from the server. After receiving the request, the server sends a message to the client including its randomly chosen sequence number and window size along with the confirmation of client’s sequence number. After receiving the response form the server, the client returns an acknowledgement message with the confirmation of server’s sequence number and then a TCP connection is established. To make this process more understandable, we made a dialoggame how two cartoon characters make friends shown in Figure 5. In this game-based visualization, Pikachu represents the client and Bulbasaur represents the server. In the beginning, Pikachu will send a friend request (first step). Then, Bulbasaur replies to the request and sends the friend acceptance (second step). After Pikachu confirms the acceptance (third), the friendship between Pikachu and Bulbasaur will be established. We have also added a learning module for TCP SYN Flooding attack.

13

Fig. 5: Game-based visualization to explain TCP connection set-up

4.2 Web Security Visualization 4.2.1 CSRF Attack CSRF stands for cross-site request forgery. In this attack, an attacker creates a forge request for a trusted website. The request appears genuine to the trusted website [14]. One purpose for CSRF attack could be modify information in social media or banking system. As many users are familiar with the social media, CSRF visualization shows the social media website of two users (Sammy and Alice), example adopted from SEED labs [15]. Sammy is trying to become friend with Alice, but she does not accept his friend request. Then, Sammy uses CSRF to generate a request from Alice’ web browser using which she logged-in to the social media server. Sammy create a forge request and embeds it an external website. He sends the link to Alice to open it. Alice, then, get affected by the request. The user will be able to interact with both the characters in order to learn the functionality of CSRF, shown in Figure 6. At the same time, definitions and explanations will be shown. A quiz after every level will be pop up to measure whether the user understood the topic or not. Furthermore, the user will also be able to see how the process of CSRF works in an abstract way. 4.2.2 XSS Attack XSS attack stands for cross-site scripting attack. It is an injection attack that affects web application with vulnerabilities. Attackers can manipulate and inject malicious code in web pages. Once the user accesses the web page, the malicious code executes in his/her system [16]. As mentioned in CSRF visualization, social media is popular among the Internet users around the world. XSS visualization will also show how XSS can be injected in a social media website. For this case, Sammy will inject malicious code under his profile with JavaScript language. So every user who visits Sammy’s profile, his/her profile will be updated due to the malicious code [17].

ISBN: 1-60132-498-7, CSREA Press ©

14

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 6: GUI of CSRF attack visualization

4.2.3 Phishing Attack It is an attack where the user provides information to an attacker unknowingly. The common way this occurs is for the attacker to create an email while imitating to be a known third party, such as a bank or social media customer service [18]. When the user clicks on a link or download an attachment, one of the two things can happen; one, the link will guide the user to a fake website imitating a legitimate log-in to get account information or even credit card information; or two, the downloaded attachment may contain an infection and installing a malware on the user’s computer/mobile device. The design of the interaction part will assist the user to understand what a sample phishing email will look like and how to avoid it shown in Figure 7. A typical phishing attack may not seem as an attack at first, but as a genuine email from a third party. In the interactive part of this attack, the user will be using a mail app to open their incoming email. Within the incoming mails, there is an email from Facebook pretending to be from customer service; when the user goes to the link that the email provided it will redirect the to Facebook’s log-in page. The user will put in their log-in information, however once the user enters it and attempt to sign-in a warning will pop-up notifying the user their information has been stolen. This occurs in everyday life where people’s information gets stolen without warning.

5. Conclusion and Future Work As the usage of the Internet keeps growing, users need to understand the risk they are exposed to when navigating to websites. The main objective of our project is to spread awareness of cybersecurity and educate the users how to detect and avoid cyber attacks at a young age. The framework will not only allow the users to learn the different kinds of attacks that adversaries may use, but the ways they can defend from them. Currently, we have implemented a few security examples in the application. We understand that the framework has several limitations, primarily in terms of content and student assessments. In future, we plan to include more attacks and defense mechanisms in the framework. As this framework application will first

Fig. 7: Phishing Email Visualization

be built and tested for PC platform, another future work would involve creating a mobile-compatible version. We plan to use this framework for delivery of a cybersecurity basics course, assess the effects of this framework on learning and prepare a continuous improvement plan (CIP). This CIP would include improving specific modules that can enhance users’ understanding, based on the user assessment. In order to keep the user engaged in a game-based learning environment, more levels will be developed where the progress and scores will be visible to the users. Finally, we plan to make this framework open source to the academic community after development and assessment.

References [1] “10 Cyber Security Facts and Statistics for 2018.” https: //us.norton.com/internetsecurity-emerging-threats10-facts-about-todays-cybersecurity-landscapethat-you-should-know.html Accessed May 15, 2019. [2] A. Zangre, “50 Noteworthy Cybercrime Statistics in 2019.” https: //learn.g2crowd.com/cybercrime-statistics Accessed May 17, 2019. [3] D. Schweitzer and W. Brown, “Using visualization to teach security,” J. Comput. Sci. Coll., vol. 24, pp. 143–150, May 2009. [4] “Cyberbullying Triples According to New Survey of Teens Online.” https://homeword.com/2014/06/10/cyberbullyingtriples-according-to-new-survey-of-teens-online Accessed May 15, 2019. [5] R. Godwin-Jones, “Mobile apps for language learning,” Language Learning & Technology, vol. 15, no. 2, pp. 2–11, 2011. [6] E. V´azquez-Cano, “Mobile distance learning with smartphones and apps in higher education.,” Educational Sciences: Theory and Practice, vol. 14, no. 4, pp. 1505–1520, 2014. [7] L. McDaniel, E. Talvi, and B. Hay, “Capture the flag as cyber security introduction,” in 2016 49th Hawaii International Conference on System Sciences (HICSS), pp. 5479–5486, IEEE, 2016.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

[8] “GenCyber.” https://www.gen-cyber.com/ Accessed May 17, 2019. [9] “Unity.” https://unity.com/ Accessed May 15, 2019. [10] L. Blasco-Arcas, I. Buil, B. Hern´andez-Ortega, and F. J. Sese, “Using clickers in class. the role of interactivity, active collaborative learning and engagement in learning performance,” Computers & Education, vol. 62, pp. 102–110, 2013. [11] “DNS.” https://en.wikipedia.org/wiki/Domain_Name_ System Accessed May 15, 2019. [12] “DNS Amplification Attacks.” https://www.us-cert.gov/ncas/ alerts/TA13-088A Accessed May 15, 2019. [13] “TCP.” https://en.wikipedia.org/wiki/Transmission_ Control_Protocol Accessed May 15, 2019.

15

[14] “CSRF.” https://www.owasp.org/index.php/Cross-Site_ Request_Forgery_(CSRF) Accessed May 15, 2019. [15] “SEED Labs.” www.cis.syr.edu/˜wedu/seed/labs.html Accessed May 15, 2019. [16] “XSS.” https://www.vpnmentor.com/blog/top-10-commonweb-attacks/ Accessed May 15, 2019. [17] “Sammy Worm.” https://en.wikipedia.org/wiki/Samy_ (computer_worm) Accessed May 15, 2019. [18] “6 Common Phishing Attacks and How to Protect Against Them.” https: //www.tripwire.com/state-of-security/securityawareness/6-common-phishing-attacks-and-how-toprotect-against-them/ Accessed May 15, 2019.

ISBN: 1-60132-498-7, CSREA Press ©

16

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Using a Telepresence robot in an educational context Laurent GALLON, Angel ABENIA LIUPPA - University of Pau and Pays de l’Adour - Pau, France {laurent.gallon,angel.abenia}@univ-pau.fr Abstract— This paper deals with the use of telepresence robots in an educational context. To this day, these uses are often related to the problematic of "homebound students" 1 . The emphatic context associated with these students masks the pedagogical difficulties encountered by telepresence learners. In this article, we first detail the use of the robot by homebound students, highlighting motivational factors. Then we focus on persistence using the robot, once this emphatic context declines. Finally, we describe our attempts to reduce transactional distance by adding connected learning environments. Keywords: telepresence robots, motivation, transactional distance

1. Introduction For the last few years, we have witnessed an intensification of the experimental use of telepresence robots in the educational system. In France, a recent study in the Auvergne Rhône-Alpes French area (AuRA) has reported the findings of a classroom use for students who are homebound1 . Other studies [2] [10] have enforced these results and show that the main added value of telepresence in this context is the preservation of the social link between the students, the school, the teachers and especially the classmates. This preservation is an important guarantee of school continuity and the reduction of the risk of dropping out. However, beyond the psychosociological impact, questions about learning opportunities via telepresence robotics arise. Indeed, several difficulties come to disturb the distant student [3], for example the difficulties of perception of the teacher’s and classmates’ body language (virtual proprioception) due to a wide-angle camera, or the perception of the "Merobotic" and the new form of socialization allowed by this mechanical avatar [4]. The teacher may also be disturbed in the realization of his/her task to take into account the specific "presence" of this student, including the student in the didactic situation and follow the evolution of his/her learning, especially in the case of practical and skill assessments. After the positive impact of the robot during the first uses, a dropout over time can sometimes be noted, due to a too large transactional distance [5] correlated to a lack of practical adapted features in the telepresence device. Several 1 as defined in [11], "homebound students" are students who are not able to attend school for a long time, due to symptoms, treatments, or recovery from illness, (e.g., cancer, heart failure,. . . )

Françoise DUBERGEY, Maïté NÉGUI SAPAD40 - Pupilles de l’Enseignement Public des Landes (PEP40) - Mont de Marsan, France [email protected] projects have begun to explore these difficulties. J. Bell [6] studied the feeling of presence (incarnation) of distant students in hybrid teaching, from videoconferencing to the use of the telepresence robots. It appears that the mobility of the robot is essential to getting closer to a feeling of physical presence. Gleason [7] confirms Bell’s results and emphasizes the need for appropriate pedagogy. Newhart and al. [10] [11] study the feeling of acceptability of the robot from a teachers and administrators point of view. In all these works, the sociological and psychological aspects are taken into account, but the didactic part is little discussed. In addition, pedagogical situations of the use of a robot are limited to simple verbal and visual interactions (course / exercises). The use of a robot in practical work is little or not treated [8]. In this paper, we propose a preliminary analysis of the use in duration (perseverance) of telepresence robots in an educational context. We base our analysis on SDT theory [9] to assess the motivation while using the robot. More precisely, we focus on what happens after several sessions of robot uses in the classroom, that is, once the impact of novelty and empathy for the homebound student fades. We then relate some experimentations done to increase the capacity of the interactions of the robot with humans and distant space, in order to reduce the transactional distance [12], induce the student a perception of success and control, and then keep him/her engaged and motivated.

2. Background 2.1 SDT theory The Self Determination Theory (SDT) [13] is a psychological theory which can be applied to education in order to understand the learning process and how to motivate students and continue to attract their attention. It is based on the innate psychology needs of: autonomy (need of being actor of his/her own life, making his/her own choice according to his/her preference), competence (need of controlling and being competent in a domain in which he/she was interested) and relatedness (need of interacting with other people). Deci and Ryan [13] define student motivation as the combination of two factors : external motivation, which is provided by an outside help (teacher, classmate, ITS, chatbot, . . . ), and internal motivation, which is provided by the perception the student has of his/her ability to accomplish pedagogical

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

tasks. Internal motivation is always stronger than external motivation. On this basis, the whole motivation of a person to accomplish a task can be classified into three levels: global, contextual and situational. Each level describes a different evolution time scale. The global level addresses motivation that comes from the social environment, religion, etc (long time range). The contextual level addresses motivation that comes from the context of the student’s environment, as for example the school context (medium time range). The situational level addresses motivation from the current situation in which the student is at a given moment (short time range). Situational contexts are the strongest (but the shortest) contexts. On the other hand, global contexts are the weakest over a short period of time, but are persistent throughout life. Changing a contextual perception can only be only achieved by repeating many times over situational contexts in which the student has a feeling of success and control (positive perception of his/her skills). In the case of the use of a telepresence robot by a homebound student, the first uses of the robot in the classroom are associated with a strong situational context (short range duration) in which the whole motivation is induced by novelty and empathy. At this time, the student, classmates and teacher are not guided by educational concerns. As described by Newhart in [10], a bridge is created between the teacher, classmates, and the homebound student. This bridge does not only include social aspects, but also the remote spaces. Some rules must be defined to ensure a safe bridge between stakeholders. We address this point in section 3. Gradually, the initial context is replaced by the educational solely (contextual level, weaker but more persistent), in which the questions on how to learn through a robot, how to evaluate knowledge and skills through a robot, and how to insert a robot in pedagogical activities are asked. At this time, interactive capacity of the homebound student through the robot is leading motivation. The more functions are numerous and effective, the more motivation and persistence are present. The adequacy of the robot’s capabilities with teaching tasks is closely related to the transactional distance.

2.2 Transactional distance Transactional distance (TD) is defined by Moore [12] as the degree of psychological distance between the student and the teacher. It can be quantified as a function of three variables : dialogue, structure and learner autonomy. In [13] Zhang proposes to refine TD into three items : transactional distance between student and teacher (TDST), transactional distance between student and student (TDSS) and transactional distance between student and content (TDSC). Anderson [14] suggests that the smaller these measures, the more satisfying the learning experience, and as a result, more substantial erseverance. In online distance learning, traditional TD relies on technological mediation. Weidlich [5] proposes a definition of TDTECH, which assesses the

17

transactional distance of the learner and the interface or technology. He argues that this extent depends on two factors: "(1) the basic proficiency of the student in using the necessary technology, (2) the design and functionality (e.g., usability) of the technology itself, as perceived by the student". TD is perceived as "an interplay of these two factors". In [15], Hung proposes a framework to understand and assess students’ readiness in online learning. He demonstrates that modern students are used to computer-mediated communication, with a high confidence in his/her computer/network skills. More precisely, in the context of online learning, students show a strong readiness in computer/Internet selfefficacy, motivation for learning, and online communication self-efficacy. Self-directed learning and learner control are the weakest items. In [5] Weidlich fits his first TDTECH factor (student proficiency) with Hung’s computer/Internet self-efficacy. In other words, he demonstrates that modern students have a good proficiency in using new technologies in learning. TDTECH mainly depends on the functionalities the robot offers. The course topology has also a major impact. The interactions between the distant student and others humans involved in the pedagogical situation (classmates, teacher), and also between the student and the distant space, are different depending on this topology. For example, in the case of a lecture, the student only needs to listen and take notes. In the case of lesson exercises, the student must be able to interact with the teacher, but also with his/her classmates, in order to ask questions or to show his/her work. Finally, in the case of practical work, the student must also be able to interact with the distant space, carry out practical manipulations, or when this is not possible, play a clearly identified role in the group of students to which he/she belongs. So, oral, visual, document sharing and physical capacities of the robot are primary.

3. Our project 3.1 Overview Our project comes from the collaboration of the Computer Science Lab of the University of Pau, in France, and SAPAD 40 (organization for a pedagogical help of homebound students, department of the Landes (40) in France). These organizations help homebound students during their convalescence, and help to finance home classes. But it is not always easy to find teachers who agree going to a student’s home, because of the geographical distance, or because there is are not enough teacher in the desired specialty. The use of telepresence robots is a solution. Since 2015, 24 students are benefitted from the use of a robot. In the following, we give a feedback of these uses, focusing on the efforts we made to ensure that the robot could be used in the students’ incourses, all through their convalescence.

ISBN: 1-60132-498-7, CSREA Press ©

18

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

The description is chronological. First, we place ourselves in the first context where novelty and empathy induces the stakeholders’s motivation. In this part, two sub parts are differentiated : the first (section 3.2) deals with the actions to prepare the arrival of the robot. The second (section 3.3) refers to the first uses in a classroom, and the creation of a safe bridge between stakeholders. After some sessions of robot use, the initial context fades and is replaced by the educational context. In section 3.4, we relate the possibility of how to continue to be engaged in the use of a robot, by adding a dedicated connected learning environment that reduces the transactional distance, and the risk of dropping out.

3.2 Preparing the arrival of the robot The preparation of the stakeholders at the arrival of the robot is essential. The more effective this preparation, the more involved the various stakeholders in the use of the robot, and the more the first context (the strongest in terms of motivation) lasts in time. On the other hand, poor preparation risks to see the first context quickly disappear and give way to the difficulties of teaching a robot, with a high risk of dropping out. In the remainder of this section, we describe the key points to which we paid particular attention for each stakeholder 3.2.1 Choosing the right robot As we noticed before, the first uses of a robot are accomplished in a very strong context, driven by novelty and empathy. But this context fades quickly, and is replaced by the educational context, in which several pedagogical difficulties appear, and can be a reason for dropping out. At this time, the motivation is clearly impacted by the capacities of the interaction of the robot (TDTECH). Depending on the topology of courses the homebound student takes, the choice of the robot can be a benefit or an obstacle to motivation. For example, if courses are given in a small classroom, with little space between the tables, a cumbersome robot will make the distant student feel that he/she is disturbing the rest of the class from the moment he/she wants to move the distant robot. In the case of a lecture, a good video quality is a real advantage to see what the teacher is writing on the board. So the initial choice of the robot is essential for motivation, depending on the school environment, and the courses the homebound student is taking. We have experimented 3 different robots : Double from Double Robotics, Beam from Awabot/Suitable Technology, and Ubbo expert from Axyn Robotics (Fig. 1). Each robot has different technical characteristics which impact the perception the student has, and so TDTECH. A summary of some of these characteristics is given in Table 1. Our experience shows that connection problems are the main factor of impact on motivation to use the robot.

Fig. 1: Used telepresence robots

Table 1: list of technical capacities of telepresence robots cameras video quality audio quality head height rotating head obstruction zoom weight stability battery life driving configuration/boot network connection issues Impact on TDTECH

Beam 2 ++ ++ fixed no medium x2 - (20kg) ++ 2h easy easy variable ++(SMALL)

Double 2 + + variable no small x2 ++ (7 kg) 4h medium medium not a lot +(MEDIUM)

Ubbo 2 + + fixed yes high x2 + (13kg) ++ 4h medium too coplicated medium +(MEDIUM)

TDTECH is mainly impacted by this factor. The quality of video and sound is also important. This leads us to systematically use 4G connections, instead of Wifi, to avoid problems of sensibility, roaming while moving in the classroom, sharing of bandwidth, and sometimes authentication issues. The weight can also be a problem if the robot has to be carried by students, for example to go from one floor to another. Finally, obstruction can be a problem in some cases, especially in crowded classrooms (handbags on the floor), or in classrooms with little space between tables. Other characteristics seem to be less important. For example, a short battery life is not necessary a problem: homebound students are not connected all through the day long, because of a poor capacity of concentration, medical cares, . . . Furthermore, during a session, the robot rarely moves, which extends its battery life. Finally, a lunch break seems sufficient to recharge the battery. Notice that connection issues do not only occur in school, they also appear at students’ home. This leads us to also adopt 4G connections for the student. Unfortunately, this is not always possible at a hospital, or in some rural areas. In these cases, after one or two attempts, the connection problems become stronger than the need to recover the social link, and the student stops using the robot. As mentioned by Newhart [10], ensuring a good technical bridge between distant spaces (classroom and student location) is essential for motivation.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

3.2.2 Preparing the homebound student and his/her parents The major issue for the family is the fear that the homebound student will be further marginalized by other students in his robotic form. This feeling is not related to technology, but to the social link with classmates. Moreover, they are convinced that the teachers will accept the presence of the robot in the classroom. This feeling comes from the educational background they experienced, where the teacher is seen as a caretaker, somewhat like a parent. It is difficult to make them accept that their child may not attend certain classes because the teacher refuses the robot in his/her class. The first fears of the student are about self-image: some have not seen their classmates for a long time, and illness and treatment may have altered their physical appearance. They are reassured when they learn that they can choose the image they want to show the class (photo for example) or no picture at all. This technological functionality can generate strong motivation in this context, but even if the robot does not have this feature, it is always possible to disable or mask the webcam from the cockpit. We always perform a pre-connection test between the robot and the homebound student. In order not to be disturbed by the desire to recover the social link with the school, for this test, the robot is located in a neutral place (in a room of our university), allowing to focus only on the technical aspects. At the end of this test, it is rare that the student does not want to continue, which shows that the ability to use this technology is strong (as mentioned by Weidlicch [5]). In the rare cases where the student refuses to use the robot after the first test, the reasons are mainly related to problems of self-image, not technical issues. 3.2.3 Preparing teachers The introduction of a robot into the classroom generates a first reflex of mistrust from teachers. The presence of a camera in a classroom is not trivial. During his/her class, the teacher is master, and this moment is "sacred". Opening it to the outside world is not easy and the tool can be considered voyeur. We know perfectly well the excess in recent years of student publications on social networks, degrading videos of teachers in their classes, and the impact that these acts have had on the profession. At this first point comes also the fear of being faced with the reality of the disease. The "world of absence" in the school environment is indeed devoid of all humanity: the disease gives rise to immediate empathy, but the daily routine of the class makes us accustomed to this absence. When the student is not present, he/she may be considered as not existing in the class group, that is to say that one does not measure (and one does not try to measure) the impact of the illness on the student and his/her schooling. Teachers discover the reality of this world of absence on the day the

19

robot arrives in their classroom. But our experience shows that the reality of using the robot is different. Indeed, as we already said above, the disease is far from being omnipresent in the image which is sent by the robot (we only see the face of the student). Its acceptance is all the more facilitated, and very quickly, when teachers behave with the robot as with an "ordinary" student. We have even experimented that a teacher, extremely reluctant to introducing the robot in his/her class, regrets at the end of the year not to having tried the experiment, convinced by feedback from his/her colleagues. Another important point is that the teacher does not want to be responsible for any problems on the robot. This covers both the connection problems, but also potential breakage during his/her course. All these reluctances can be eliminated by discussing with the teachers on the forehand, showing them the merits of the process, and especially by having them driving the robot, to reduce the fear of this new technology, reducing TDTECH from their point of view, and better understand how the student is perceiving the class. 3.2.4 Preparing school board The school board has a key role in the initial context. As we have seen before, at this first stage, the main initial obstacle is the teacher. Obtaining the support of the administration (hierarchy of teachers) is very important to "help" him/her, when necessary, to make the effort to accept the robot in class. But the board must also be reassured on certain points. The robot appears as an expensive technological tool (each robot we use costs approximately 5000$), and the problem of responsibility which arise in case of breakage, theft or breakdown. Secure storage when the robot is not in use can also be a stuck point. 3.2.5 Preparing classmates and choosing referents The role of classmates is central. The homebound student wants to be able to attend classes mainly to leave his/her medical environment, to find a "normal" environment. Classmates also want to understand why their friend is not present. One important thing is identifying in the classroom one or two of the homebound students’ friends who can become referents. To be a referent means to be the guardian of the robot, and thus the guardian of the homebound student. The main tasks that are assigned to a referent are: 1) collecting and returning the robot to its storage point as soon as possible (morning, noon, evening) preventing the battery from discharging 2) solving the technical problems encountered during lessons (sound, camera, disconnections, ...) 3) exchanging with the remote student as soon as the situation requires it (sending pictures of the whiteboard by MMS, specific request when the student does not want to disturb the teacher, ...). The referent is a key element that reassures the teachers, who do not have to deal with the robot during

ISBN: 1-60132-498-7, CSREA Press ©

20

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

their course, and reassures the administration, knowing that the robot is never forgotten in a class. Nevertheless, this task is heavy, and should not become a burden for this referent student. So the best is to find two or three referents in the same class who can alternate, for example every day. Clearly referents are a key point. They mask technical issues to teachers and the homebound student, increasing his/her motivation, and so decreasing TDTECH. 3.2.6 Summary In Table 2 we summarize the main key points which have an impact on the motivation of the homebound student in using the robot at the first stage. Clearly, the most critical stakeholder is the teacher. From a student point of view, the more the teacher seems invested in the insertion of the robot in his/her classes, the more the student is persevering. The use of referent students is a crucial point, which makes it possible to partially discharge the teacher and the distant student from technical problems, and allowing them to focus more on learning. This is the element of our process that significantly decreases the transactional distance at the start of the robot use. Table 2: summary of key points for motivation at the first stage student Illness Student image New technology Responsibility Marginalization Video capture

family

board

classmates Empathy

teachers Fear

Exciting

Fear Fear

Fear Fear Fear Fear

3.3 First uses in classroom The appearance of an ailing student in a classroom in the form of a robot at first generates surprise (see Fig. 2), and quickly empathy and solidarity among students. There is no problem of image or marginalization. It may even happen that the ailing student goes, in the class, from being in a position of weakness (because of his illness) to a position of envied (the "master" of the robot), a paradoxical situation that obviously has a strong impact on the motivation of the homebound student. We must admit that the image returned by a telepresence robot is easier to accept than the image of a student who is bedridden or physically impaired, especially since each student who uses the robot chooses the image he/she will present in front of the camera. This allows he to continue to benefit from the positive aura of the robot, without letting the image of the disease take over. It is undeniable that the presence of the robot causes a "shock" in the class, and partly changes its atmosphere, apparently to us, rather positively. When the robot is present, each student becomes more responsible than normal. For example, students will more easily self-discipline at the level of chatter, to allow "the robot" or rather the homebound student

Fig. 2: Catherine driving the robot for the first time as part of practical work in aeronautical construction classroom to hear well what the teacher says. They are silent when the homebound student wants to ask the teacher a question. Although we have been able to observe these behavorial changes, further study will be needed to determine what the real sociological levers on which the robot has an impact, are. This study has, in our opinion, all its interest in Middle Schools, where we know that relations to others are fragile and in full construction. It should show that the tripartite teacher-student-class social link is the main driver of the homebound student from using the telepresence robot, and that classmates are particularly attentive to this factor. The arrival of the robot has at the start a positive impact on the motivation of all classmates. But quickly, new difficulties appear. Indeed, the interactions between the actors and the remote space are limited through the robot. This inevitably has a negative impact on the quality of the knowledge and know-how accumulated by the distant student. The use of the robot in an educational context is not enough. It must be complemented by interaction tools with the different stakeholders and also the remote physical space to get closer to a physical presence, and reduce the transactional distance. In the following, we describe some solutions we have experimented for that.

3.4 Decreasing transactional distance by adding a connected learning environment to the robot This experiment was conducted at our university, after one of our undergraduates has injured his knee (Terry), and had to convalesce at home for several weeks. We proposed to use a telepresence robot to follow certain important courses (Mathematics (lecture and exercises), English (lecture and examination) and IP telephony (lecture, exercises, practicals and practical examination)), with different typologies: courses, exercises on the course and practical work. We created connected learning environments dedicated to each typology. By connected learning environment, we mean a set of digital tools, remotely accessible by Terry, which allow him to interact with the distant space, to perform the requested manipulations. The robot allows Terry to move

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

21

Table 3: Impact of course topology on TDTECH lessons lesson exercises practicals

Fig. 3: Terry and Sam in the practical works room around the room to collaborate with the other students in his group, and interact with his teacher. A referent student, Sam, helped Terry when he was using the robot (Fig. 3). 3.4.1 Course typology and transactional distance In the case of a lecture, the student needs to see and hear the teacher, sometimes talk to him, and to see the whiteboard well. He does not need any interaction with other learners. Necessary documents can be distributed electronically beforehand. All these needs are covered by the intrinsic capabilities of the robot. It is not necessary, in this typology, to add additional functionalities. The transactional distance arising from the use of the robot is weak. It only has an impact on TDST. Note that, occasionally, having a good view on the whiteboard is not always possible, due to certain conditions such as brightness. An extra camera placed near the whiteboard, may be a good solution. Another sensitive point is the notification of request to speak. Often, students cut their microphone not to interfere with the teacher. Signifying that they want to ask a question then becomes an issue. Some robots offer a led panel that can be cloven to signal his request. For other robots, it is possible to add a small lamp on the webcam of the cockpit of the robot, in order to flash the screen, and report to the teacher. In case of lesson exercises, things are different. In addition to the need for interaction with the teacher, the student must be able to interact with his/her classmates, especially in group work. The student must also be able to show his/her work to the teacher and classmates. This feature exists on some robots in the form of sharing the cockpit screen. However, in that case the image of the student cannot be seen anymore and interactions becomes thus less natural. Moreover, the student must also be able to see the work of classmates. Robots possessing the ability to climb / descend / turn the head, give the student this power. Without these capacities, the use of the robot increases TDTECH, with an impact on TDST and TDSS. A possible remediation is the taking of a photo and sending it using MMS between the referent and the student. The more critical typology is practical work. As in the

TDST X X X

TDSS

TDSC

X X

X

TDTECH small medium high

case of lesson exercises, the student must be able to interact with classmates and the teacher. He/she needs to share documents. But he/she must also be able to perform physical actions in the classroom, at least to collaborate with his classmates. In this typology, TDTECH is high, and impacts TDST, TDSS and also TDSC. Adding functionalities to permit the distant student to interact with the distant space is not always possible, and depends on the subject. The transactional distance can be reduced in some cases by adding an adapted pedagogical connected environment. For example, in computer science practical work, it is quite easy to offer distant connections to computers, servers, IP telephons to the distant student. But the transactional distance can also be decreased by adapting the educational objectives set to the remote student. For example, during chemistry manipulations, if it is impossible for the remote student to handle dangerous products, he can, on the other hand, guide his colleagues in their manipulations, create curves, and search documents. He/she can also film manipulations, so that he/she can analyze it later. In this case, despite the difficulties in handling, the remote student still feels like being part of the group. He/she plays a role and participates in the realization of the training sequence (team cognition). This feeling of controlling his/her pedagogical activity is a mechanism that generates strong intrinsic motivation. 3.4.2 Description of the connected learning environments used In our experimentation, for each lesson typology, we associated a dedicated connected learning environment (Fig. 4) without changing any pedagogical objective. For lesson exercises, we proposed Terry to use a shared whiteboard through a tablet, and a share file storage space for persistent works. These abilities could both be used when interacting with the teacher, or with classmates. Clearly, this solution was great for Terry, but faced another problem: some teachers were not used working with these numeric tools, and refused to use them in their class. TDTECH increased, not because of the student, but because of the teacher.

Fig. 4: Connected learning environments

ISBN: 1-60132-498-7, CSREA Press ©

22

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

For IP telephony practical work, we used a VPN connection between Terry’s house and our university, to allow him to connect to servers, IP telephones and the robot. Notice that Terry made his practicals examination in telepresence, under the same conditions as his classmates. He also took the TOEIC exam1 in telepresence. He had no problems with distance, and passed both exams successfully. 3.4.3 Analysis We conducted interviews with Terry, Sam, and the teachers who participated in this experiment, to get their opinion on the positive or negative impact of the connected learning environments associated to the robot (see [8]). The results of these study are: • knowledge and know-how (TDSC): the opinion is unanimous for knowledge, there is no difference between face-to-face and distance. At know-how level, Terry felt at ease. Teachers recognized that the essential topics were validated. The connected pedagogical environment has therefore played its role well at a TDSC level. • interactions with teachers and classmates (TDST and TDSS): teachers recognized that their interactions in class with Terry were equivalent when using the robot compared to face-to-face. They observed the same qualities and deficiencies in Terry while communicating to his robotic form compared to face-to-face: very little intervention in mathematics, a lot of participation in English. In practical works, telepresence did not prevent him from chatting and laughing with classmates. Obviously, the use of the robot did not change his attitude, nor that of the teachers. Note that the math teacher, initially reluctant to use document sharing tools, acknowledged that this would ultimately be a good thing for her interaction with Terry • interactions with distant space: in practical works, there was a before and after connected learning environment. Before, Terry was an observer, and could only observe the work of his classmates. Using the dedicated pedagogical connected environment, he was able to be active, to realize on his own remote manipulations, and learn by practice, a fundamental element of pedagogy. Comforted in his abilities, he agreed to take the risk of doing his IP telephony exam at the same time as classmates. He scored 12/20 with the same exam compared to others. This study shows that the use of a connected learning environment greatly facilitates significantly the insertion of the telepresence robot into the pedagogical sequences, and reduces the transactional distance. A larger study, with a strong validation, should be conducted to confirm these preliminary results.

4. Conclusion and perspectives In this paper, we tackled the problem of perseverance in using telepresence robots for homebound students. We have shown that once the initial emphatic context has passed, pedagogical difficulties appear when taking into account the student in his robotic form in the teachings. We have described several examples of connected learning environments that reduce the transactional distance. In future work, we are thinking of consolidation our first results through a broader study, taking into account in particular the team cognition dimension, so important in the world of education. We haave started experimenting with a telepresence escape game, to see how the players collaborate in the same physical space, while they are all physically distant from this space. The conclusions of these works should allow us to create telepresence learning at our university, for professionals who can not move physically, or for lifelong learning.

References [1] Coureau-Falquerho, E., Simonian, S., Perotin, C. High school telepreesence Robot in Auvergne-Rhône-Alpes french area. [2] Gallon, L., Dubergey, F., Negui, M. (2017). telepresence robot : a numerical tool used by "Sapad" to make prevented student present. "La nouvelle revue de l’adaptation et de la scolarisation", (3), 157-171. [3] Furnon, D., Poyet, F. (2017). Telepresence Robot: Process of Appropriation through the Evolution of the Modalities of Presence. International Journal of Technology and Inclusive Education (IJTIE), 6(1). [4] Furnon, D. (2018). Co-construction of Tangible, Dispersed and Multisemiotic Spaces through the Use of a Telepresence Robot. Telepresence in Training, 145-162. [5] Weidlich, J., Bastiaens, T. J. (2018). Technology matters-The impact of transactional distance on satisfaction in online distance learning. Int. Review of Research in Open and Distributed Learning, 19(3). [6] Bell, J., Cain, W., Peterson, A., Cheng, C. (2016). From 2D to Kubi to Doubles: Designs for student telepresence in synchronous hybrid classrooms. Int. Journal of Designs for Learning, 7(3). [7] Gleason, B., Greenhow, C. (2017). Hybrid education: The potential of teaching and learning with robot-mediated communication. Online Learning Journal, 21(4). [8] Gallon, L. Abénia, A. (2017). Pedagogical environments for telepresence robots. In 4e Workshop pédagogique Réseaux & Télécoms (french). [9] Ryan, R. M., Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American psychologist, 55(1), 68. [10] Newhart, V. A., Olson, J. S. (2017). My student is a robot: How schools manage telepresence experiences for students. 2017 CHI conference on human factors in computing systems (pp. 342-347). ACM. [11] Newhart, V. Warschauer, M. Sender, L. (2016). Virtual Inclusion via Telepresence Robots in the Classroom: An Exploratory Case Study. Int. Journal of Technologies in Learning. 23. 9-25. [12] Moore, M. G. (1993). Theory of transactional distance. Theoretical principles of distance education, 1, 22-38. [13] Deci, E., Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior. Springer Science Business Media [14] Anderson, T. (2003). Getting the mix right again: An updated and theoretical rationale for interaction. The International Review of Research in Open and Distributed Learning, 4(2). [15] Hung, Min-Ling Chou, Chien Chen, Chao-Hsiu Own, Zang-Yuan. (2010). Learner readiness for online learning: Scale development and student perceptions. Computers Education.

1 https://www.etsglobal.org/

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

23

A Hands-on Oriented Cyber-Learning Curriculum for Undergraduate Cybersecurity Education Yaswanth Kolli, Ahmad Y. Javaid Electrical Engineering and Computer Science Department The University of Toledo Toledo, OH 43606, USA [email protected], [email protected] Abstract— With cyber-attacks on daily news nowadays, the victims to the cyber-attacks are usually the people who lack knowledge of how they happen. It is also known that dissociating oneself from technology in this digital world is not a solution to this problem. On the contrary, learning cyberthreats, how they work and protecting oneself in the cyberworld is a more appropriate solution. This overall practice of enabling cyber-learning through the development of stateof-the-art curriculum is often termed as cyber-education. Integrating such a curriculum in undergraduate studies is critical and needs no convincing. This paper discusses a comprehensive hands-on based cyber-learning curriculum we developed, which can be either offered as a separate course or integrated into existing cybersecurity courses, as relevant. The course offers an outline and insight into the internal mechanisms of a few successful cyber-attacks. We study and prepare hands-on labs for eight different classes of attacks and related open-source tools to help students gain insights on these attacks, and learn how to prevent or safeguard against them. These classes include webattack, vulnerability exploitation, password cracking, social engineering, denial-of-service, forensics, and an “other” category with some general security utilities. In total, 15 tools were studied, that again correlate to the above eight classes. Keywords: Cybersecuirty curriculum, cyber education, cyberlearning, Open-source tools, Hands-on learning

1. Introduction All the data varying from personal information to industrial control systems are being connected and transferred through the networks all over the world, which puts cybersecurity in a vital role of the system. One of the leading cybersecurity solutions firm CISCO defines cybersecurity “as the practice of protecting systems, networks, and programs from digital attacks” [1]. The United States tops the list for most targeted cyber attacks across the world and cybercriminals includes cybercriminals to its most wanted list. The United State government is planning to spend 15 billion dollars for all cybersecurity-related activities across the country in 2019. Identity theft is one of the worst impacts of a data breach

caused due to cyber attacks, which is increasing every year and a total of 60 million Americans claim that their identities have been stolen. With the latest technologies such as the Internet of Things (IoT), every device in the home including door lock is connected to the internet, making it available in cyberspace and prone to cyber attacks [2]. According to a study conducted by IBM, the number of mega data breaches (involving more than 1 million records) rose to 16 in 2017 from just 9 in 2013. To give a financial account of the loses due to the breaches, it is estimated that a breach of 50 million records will result in an approximate loss of 350 million dollars [3]. The above mentions manifest the importance of cybersecurity awareness among people. Being aware of cybersecurity attacks and how they happen, has become a responsibility rather than an option. According to CISCO [4], the most common cyber attacks are, malware, phishing, Man-inthe-middle, Denial-of-service, SQL injection and Zero-day exploit. A recent work discusses the vital role of the employees in protecting the organization from cyber attacks [5]. It states that every employee is responsible for following the security policies of the organization. Despite all the awareness campaigns conducted by US government and nonprofit organizations like the Cyber Security Alliance, there is continuing ignorance among the public and the organizations. According to a survey conducted by CompTIA, only one-third of the organizations in the USA need mandatory cybersecurity awareness training [6]. It is essential that every individual and employees of an organization should receive cybersecurity awareness training which can be achieved through cybersecurity education. Therefore we develop a comprehensive hands-on based cyber-learning curriculum that can be offered as a separate course or integrated into existing cybersecurity courses.

2. Literature Survey The cybersecurity threats are not limited to stealing data, disrupting services they also pose a severe threat to world peace. Sixty cybersecurity predictions for 2019 includes cyber war [7]. A cyber attack objective is to steal information, harm or compromise the computer to use it for later

ISBN: 1-60132-498-7, CSREA Press ©

24

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

attacks. , on the other hand, is the practice of performing computer attacks on rival states to achieve their national interests. In traditional warfare, the nations fight physically, but in a cyber-warfare, there is no direct physical attack. Instead, everything is done in virtual cyberspace leading it to cause indirect physical damage like, manipulating industrial controls, exploiting water plants, derailing a train which has adverse effects on the economy of the targeted nation. Stuxnet worm was designed to compromise Iran’s nuclear program by manipulating the industrial controls without being noticed by displaying normal conditions to the plant engineers. Reports were claiming that the U.S. and Israel governments designed the Stuxnet worm as an act of cyberwarfare. Cyber-Terrorism can be defined as the convergence of terrorism and cyberspace. Cyber-terrorism is the use of computer technology to cause severe disruption, harm, and fear in the economy. It is related to four main aspects, computer generation, political motivation, physical violence, and psychological coercion [8]. A recent work showcases the necessity to build International cooperation on cyber defense and deterrence against cyber terrorism. It showcases the possible ways used by the terrorists to exploit cyberspace and mentions that every country has its cyber laws and incidence response teams. It proposes that there should be a robust, international legal framework under the control of the UN that addresses all the cybersecurity related issues and share the intelligence obtained among the nations across the globe [9]. The practical approach for deterrence of cyber-terrorism is proposed in [10]. It discusses the challenges involved in cyberterrorism deterrence and a robust framework to overcome them. Another popular work surveyed the psychological effects of cyber terrorism. They have conducted field survey experiments that simulated cyber terrorism and interviewed 522 individuals. The results are analyzed on a score of 4 in which cyber terrorism (Lethal) is marked at 3.6, which shows severe physiological impact [11]. The cyber attacks targeted at national infrastructure leads to severe harm, damage to the national economy and threatens global peace. The cyber threats that a nation can face are international terrorism, state-sponsored terrorism, malicious hacktivism, insider threats. The methods to counter those threats using intelligence are proposed in [12]. There are also possibilities of insider cyber-threats to an organization that may be caused accidentally or intentionally by the employees. A recent work analyzes the possible insider-threats, how they rise problems in organizations and discusses the lawsuits that are relevant to various scenarios [13]. A popular survey on types of cyber attacks and how they can be detected, classifies the cyber attacks into 13 types including Denial-of-service, cyber espionage, cyber terrorism, cyber war, active attacks and the various approaches including, agent-based approach, artificial intelligence approach, that can be used in detection and prevention of those

attacks [14]. With the intense research being conducted on machine learning and data mining, researchers are exploring ways to use these two areas into the cybersecurity field to detect intrusions. A survey conducted on discovering the use of machine learning and data mining in cybersecurity intrusion detection is presented in [15]. Roy et al. proposed a game theory model for addressing network security problems and surveyed the game designing techniques and their outcomes [16]. Nowadays everything from a device to a city is being converted into smart entities by integration of computing and communication capabilities to it, which makes them prone to cyber attacks. The smart grids are the next generation power supply and management systems, which are part of the critical national infrastructure. A smart grid is achieved by combining advanced computing and communication technologies, to enhance efficiency in power management. The smart girds are one of the hot targets to the cyber attacks as they may cause severe damage and a considerable loss to the nation. A survey paper by Wang et al. [17] describes the possible cybersecurity challenges faced by the smart grids. The concept of smart cities is defined as the communication of all the devices in the city to solve significant issues like traffic, public safety, emergency response. The work [18] discusses the cybersecurity challenges encountered while operating such a large city of connected devices. Cloud computing can be defined as a platform for storing, managing and processing data located on remote servers, that can be accessed using the internet. Cloud computing aims to reduce the cost of buying and maintaining servers by offering computer resources as services. The simplified analysis on cloud computing provides an insight into its concepts, trends and future directions of cloud computing [19]. Availability of organizations and personal data on the internet(cloud) makes it a tempting target for cybercriminals. Proper security measures should be taken to prevent cyber attacks [20]. Denial-of-Service attacks aim is to bring down the service offered by flooding the server with requests and making it unresponsive to legitimate requests. The cloud computing is based on the server-client model which makes it prone to DoS and DDoS attacks. There is ongoing research on the issues caused because of DDoS attacks in cloud computing and future directions [21]. With the Internet of Things and Cloud computing technologies proliferating, the integration of both techniques has become a necessity, and the security challenges that can be faced must study [22]. The studies mentioned above manifest the importance of cybersecurity education and awareness that helps to prevent cyber attacks in various domains. The are several proposals on the guidelines to design a cybersecurity curriculum. Cybersecurity education requires a more practical and handson based learning approach [23]. An analysis conducted by Fred B. Schneider [24] on the cybersecurity education in the universities states that the course should be developed

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

based on the way how attackers think which helps the students to learn the discipline of the art of building secure systems. A new NSA driven model for designing the cybersecurity curriculum includes the teaching of cybersecurity fundamentals, policies, and legal ethics [25]. The effects of a game based cybersecurity training approaches are discussed in [26] [27]. A popular work [28] applies challengebased learning methodology to cybersecurity education that encourages students to collaborate, ask questions, gain indepth knowledge and try to solve real-world challenges. There are various cybersecurity awareness delivery methods to the users including, instructor-led, online, game-based, simulation-based delivery methods [29]. The sub-disciplines including Network security, web security, usage of tools, forensics, should be covered for effective cybersecurity education [30]. The current and future trends of cybersecurity education are discussed in [31].

25

Fig. 1: Course Labs

•

•

3. Proposed Method 3.1 Motivation Before starting the design of the course, a study has been conducted on cybersecurity courses and their curriculum offered in some of the renowned universities in the USA. Based on this study, while most of these courses include some practical aspects of cybersecurity, they are mostly result based, insufficient. A process-based approach works more efficient, in making the cybersecurity concepts clear to the students [23]. Classroom demonstration of the cybersecurity threats and the process of how the attacks happen will enhance the practical knowledge rather than theoretical aspects. The course is designed with the aim of offering more practical knowledge by exploring the internal mechanisms of some successful cyber attacks. The curriculum of the course comprises of 15 tools, used to demonstrate almost every type of possible cybersecurity attack and forensics techniques, that are known till date.

3.2 Course Structure The course we developed introduces the hacker’s perspective rather than a security personnel’s perspective, which enables the students to learn the mindset of a hacker, as suggested in [24]. Every lab is conducted in a hands-on and experiment based setting, as discussed in [23]. The proposed course requires a basic prerequisite course on cybersecurity which might have an obvious title such as “Fundamentals of Cybersecurity” or “Foundational Computer Security,” that discusses the overall concepts of network fundamentals, encryption techniques, cybersecurity: Types of cyber attacks, perspectives, ethics, laws, etc., as proposed in [25]. Every experiment contains the following sections: • Aim contains the description of the motive of the experiment.

•

About discusses the description of the tools being used in the experiment. Environment Setup discusses the installation and setup of the environment required to perform the attack. It includes tools installation and configuration. This gives better understanding of the tool structure and aids the student to overcome any future challenges while working with the tool. Methodology discusses the practical and detailed stepby-step process of performing the attack using the tools.

3.3 Course Syllabus The course comprises of 15 tools classified into seven categories: Web-based attack tools, Vulnerability Exploitation Tools, Password cracking tools, Social engineering tools, Denial of Service tool, Forensics tools, other category with general security tools. The tools are selected based mainly on 5 different criteria, functionality, ease of use, documentation, cross-platform compatibility and free of cost. Figure 1 shows the overall labs included in the designed course. 3.3.1 Web based attack Tools: Web-based attack tools are used to perform attacks involving networks and website vulnerabilities. Four tools are demonstrated in this course. •

•

•

•

Wget Tool: Wget tool is used to download content iteratively from the web and mirroring the websites for phishing attacks. Zenmap Tool: Zenmap tool is used to scan the systems/servers that are connected to the network to determine their OS type, communication protocol, open ports, network topology, Route, etc. SQlmap Tool: SQLmap is used to hack into a vulner˘ Zs ´ login able website database to gather all the userâA information. Ettercap Tool: Ettercap Tool is used to sniff the user credentials of the victim by using ARP poisoning technique.

ISBN: 1-60132-498-7, CSREA Press ©

26

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

3.3.2 Vulnerability Exploitation Tools: Vulnerability exploitation tools are used to exploit the unpatched vulnerabilities present in the system. Two tools are demonstrated in this course. • Armitage Tool: Armitage tool is used to gain backdoor admin access to the Windows 7 operating system which allows the attacker to modify files, keylogging, taking screenshots and webcam shots, etc. • Metasploit Framework: Metasploit framework is used to gain backdoor access to the Android operating system which allows the attacker to access media files, contacts, take screenshots, remote calling, etc. 3.3.3 Password cracking Tools: Password cracking tools aim to decode, identify the password of the targeted system by using various techniques and algorithms. Two tools are demonstrated in this course. • John the Ripper Tool: John the Ripper tool is used to crack Zip/RAR file, Linux, Hash type passwords. • Mimikatz Tool: Mimikatz tool is used as a postexploitation tool to crack into the user credentials of the Windows operating system which has been compromised using Armitage tool.

3.3.7 Other Tools: •

•

3.4 Student Learning Outcomes It is well known that it is important to define student learning outcomes (SLO) as part of any undergraduate curriculum. Accreditation agencies such as ABET has strict guidelines in terms of what language could be used to define SLOs so that they are quantifiable and measurable. For the proposed hands-on oriented course, we developed the following SLOs: Upon completion of the course, the student will get complete understanding of •

•

3.3.4 Social Engineering Tool: Social Engineering Tools are used to perform social engineering attacks like phishing, Baiting, Tailgating, etc. Social Engineering toolkit is demonstrated in this course. • Social Engineering Toolkit: Social Engineering toolkit is used to perform website phishing attack to steal the user credentials.

•

•

3.3.5 Denial-of-Service Tool: Denial-of-Service Tool is used to bring the server down by flooding it with requests. GoldenEye Tool is demonstrated in this course. • GoldenEye Tool: GoldenEye Tool is used to perform a Denial-of-service attack on the server.

•

•

•

3.3.6 Forensics Tools: Forensics tools are used to perform forensic analysis on a system during the incidence response phase. Three forensic tools are demonstrated in this course. • The Sleuth Kit: The Sleuth Kit is used to perform forensic investigation by analyzing and recovering data from a disk image. • Volatility Framework The Volatility Framework is used to perform digital forensics investigation on the volatile memory (RAM) of the system. • FotoForensics Tool: Fotoforensics tool is used to analyze and detect the properties and modifications/edits performed on the image.

GnuPG Tool: GnuPG Tool is used to implement encryption and decryption of the content shared by the users using the OpenPGP standard. Maltego Tool: Maltego tool is used to discover and map the connections between people, groups on social networks, web domains, and servers, etc. This helps in understanding the whole picture and finding the vulnerabilities in the mapped network.

TCP/IP vulnerabilities, Man-In-The-Middle attacks using ARP poisoning and Denial-of-Service attacks using SYN packets. - Zenmap, GoldenEye, Ettercap Tools Information exposed with open ports, mapping of information gathered on an entity, through various websites including social-networking sites, company websites etc. - Zenmap, Maltego Tools SQL injection, web-site phishing, iterative downloading of web content. - Wget, SQLmap, Social Engineering Toolkit tools Vulnerable exploitation of popular operating systems, Android and Windows. - Metasploit framework, Armitage tool. Pretty Good Privacy(PGP) encryption standard and encryption algorithms. - GnuPG Tool Various password cracking techniques including bruteforce, Post-exploitation technique that captures the encrypted passwords. - John the Ripper, Mimikatz Tools. Image forensics, disk image analysis including RAM. - The Sleuth Kit, Volatility Framework, FotoForensics tool

3.5 Armitage Tool: An Example Lab To provide an insight into the course we designed, one of the labs is discussed in more detail in this section. The design of this lab has been previously published [32]. 3.5.1 Aim In this experiment, Armitage tool is used to gain backdoor admin access to the Windows 7 operating system which allows the attacker to modify files, keylogging, taking screenshots and webcam shots, etc.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

3.5.2 About Armitage plugs into the Metasploit framework and offers a graphical user interface. Armitage tool graphical user interface, visually represents the features, such as finding a host, client and server exploitation, privilege escalation and pivoting. It contains the feature of dynamic workspace environment which allows the attacker to change the targets and the type of attack rapidly. Armitage lets people share the same details and session. It opens shared sessions to teamwork while performing penetration testing. Armitage has tools called bots which can be used to automate different kinds of tasks. Armitage encapsulates, aggregates, and organizes the tools that are present in the Metasploit framework into an interface. Armitage contains a lot of post-exploitation tools that are ˘ Zs ´ machine after it is comproexecuted on the victimâA mised. By using these tools, an attacker can find password ˘ Zs ´ keystrokes, execute command line hashes, note victimâA commands, gain root privilege, record video and more. After ˘ Zs ´ machine is compromised, Armitage helps the the victimâA attacker to use the compromised target to launch another attack. In this experiment, we are going to use the Armitage tool to hack into the windows machine. 3.5.3 Environment Setup a) Windows 7 Installation: • Step 1: In this attack, the operating system used is Windows 7. To download the ISO file click https://softlay.net/operating-system/ windows-7-home-premium-full-versionfree-download-iso-32-64-bit.htmlhere and select the 64-bit version. Now open the virtual box and click on the “New” icon. Enter the name as “Windows 7”, select type as “Microsoft Windows” and version as “Windows 7 (64-bit)”. • Step 2: Now hit “next” and allocate the memory size (RAM) of 1 GB. Then hit “next” and select “create a virtual hard disk now” and click “create.” Select “VDI (VirtualBox Disk Image)” in the next menu. In the next screen select the file size of 40 GB and click “create.” • Step 3: Now click on the “settings” icon to open settings and browse to “Storage Empty.” Click on the disk icon and select “Choose Virtual Optical Disk File” and then select the ISO file downloaded. • Step 4: Now start the “Windows 7 Machine and proceed through the installation process. b) Postgresql Installation: Postgresql is the database used by the Metasploit framework to store and retrieve its pay-

27

loads and exploits. To install Postgresql download first update the packages in the Ubuntu to do that execute the following command. “sudo apt-get update” Now execute the following command to install the Postgresql. “sudo apt-get install postgresql” To check if the Postgresql is running execute the following command “service postgresql status” c) Metasploit framework installation: • Step 1: To download the Metasploit graphical installer click https://downloads.metasploit. com/data/releases/metasploit-latestlinux-x64-installer.runhere. Now open the terminal and run it in root mode but executing “sudo su”. Now make the downloaded Metasploit executable by executing the following command. “chmod +x filename” • Step 2: Now run the installer by executing the following command. “./filename” Follow through the steps of the installer and at the end unselect the “Access Metasploit Web UI” and hit “Finish.” Now restart the Metasploit service by executing the following command. “service metasploit restart” d) Armitage tool installation: • Step 1: To download the Armitage tar file click http://www.fastandeasyhacking.com/ download/armitage150813.tgzhere. Now extract the tar file by executing the following command. “tar -xvzf filename” Where x: extract files from an archive, v: verbosely list files processed, z: filter the archive through gzip, f: use archive file or device ARCHIVE • Step 2: Now move the extracted Armitage folder into the installed Metasploit folder by executing the following command. “mv armitage /opt/metasploit” Change the directory to “armitage” by executing the following command “cd /opt/metasploit/armitage” • Step 3: Before launching the Armitage tool, change the network settings of the Windows machine and Ubuntu to “Internal network” so that they both are connected in the same network and share the same subnet. Now launch the Armitage tool by executing the follow-

ISBN: 1-60132-498-7, CSREA Press ©

28

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 3: Available attacks on the Windows machine (victim)

Fig. 2: Command shell displaying the session is opened ing command “./armitage” ˘ Zs ´ IP address After launching it asks for the systemâA on which it is running. Use the “ifconfig” command to the check IP address and enter it. 3.5.4 Methodology •

•

•

•

•

•

Step 1: Now in Armitage browse to “Hosts- Nmap Scan- Quick Scan (OS detect)” and enter the IP address of the windows machine. Wait for the system icon to appear and right click on it to browse to “HostOperating System - Windows 7”. Step 2: Go to the modules pane in the left and navigate to payload-windows-meterpretermeterpreter_reverse_tcp and open the payload set the LHOST (listener host) to 6565 and output a “.exe” file. Create a folder named “payload” in the path “root/var/www/html/” and store the generated “.exe” file in it. Step 3: Now again select the same payload but this time set the output to “multi/handler” which opens the port 6565 on Ubuntu machine to capture the reverse TCP connection from the target (windows) machine. Step 4: Now download the “.exe” file in the windows machine by opening internet explorer and entering the URL “IP-address of Ubuntu machine/payload” where the file is located. After downloading run the “.exe” file which opens the backdoor to the windows machine. Step 5: The windows system icon in the Armitage tool changes to red representing that it is exploited. The command line window displays the message that the meterpreter session opened and also shows the IP addresses of the systems involved in the session as shown in figure 2. Step 6 Right clicking on the system icon shows the available exploits. For example, you can take a screenshot of the target system, access files, record key logs, etc. as shown in figure 3.

4. Conclusion Now a days every device, from mobiles to industrial systems requires connection to internet to transfer data over networks, which makes cyber security most important aspect in every individual life. With all the on going research in the cyber security field new methods are being developed to detect the cyber attacks by integrating it with other fields like machine learning, data mining, but none of these couldn’t be able to stop the cyber criminals from finding their way to perform cyber attacks. Cyber security awareness plays a major role to help people and organizations defend against the cyber security threats. According to a survey conducted by ponemon institute, 78% of the data breaches in organizations are caused by employee negligence [29]. Reports also show that Unites States Public Sector itself requires approximately 20,000-30,000 qualified Cyber-Security specialists [33]. This proves the necessity of cyber security education which helps in preparing the students to defend themselves and the organizations, from cyber security attacks. The hands-on exposure approach is considered as one of the effective educational methods to teach cyber security education [33]. It is achieved by creating lab-based instruction curriculum which makes every student to perform the experiments by offering the required environment needed as described in the above sections. Thus this course offers an effective implementation of 7 different classes of cyber attacks through a hands-on based approach to the students.

References [1] What is Cybersecurity? https://www.cisco.com/c/en/us/ products/security/what-is-cybersecurity.html [2] 10 cyber security facts and statistics for 2018. https: //us.norton.com/internetsecurity-emergingthreats-10-facts-about-todays-cybersecuritylandscape-that-you-should-know.html [3] IBM Study: Hidden Costs of Data Breaches Increase Expenses for Businesses. https://newsroom.ibm.com/2018-0711-IBM-Study-Hidden-Costs-of-Data-BreachesIncrease-Expenses-for-Businesses [4] What Are the Most Common Cyberattacks? https://www. cisco.com/c/en/us/products/security/commoncyberattacks.html [5] Kirsch, Laurie, Boss, S.,:The last line of defense: motivating employees to follow corporate security guidelines. pp 103, ICIS 2007 proceedings (2007)

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

[6] Lack of cybersecurity awareness linked to CIOs. https: //searchsecurity.techtarget.com/opinion/Lackof-cybersecurity-awareness-linked-to-CIOs [7] 60 Cybersecurity Predictions For 2019. https:// www.forbes.com/sites/gilpress/2018/12/03/ 60-cybersecurity-predictions-for-2019/ #7a16b8f14352 [8] Kenny, M.:Cyber-terrorism in a post-stuxnet world. vol. 59, pp 111128, Orbis (2015). [9] Dogrul, M., Aslan, A., Celik, E.:Developing an international cooperation on cyber defense and deterrence against cyber terrorism. In: 2011 3rd International Conference on Cyber Conflict, pp. 1-15, IEEE (2011) [10] Hua, J., Bapna, S.:How can we deter cyber terrorism? In: Information Security Journal: A Global Perspective vol. 21, pp. 102-114, Taylor & Francis (2012) [11] Gross, M. L. and Canetti, D. and Vashdi, D.:The psychological effects of cyber terrorism. In: Bulletin of the Atomic Scientists, Vol. 72, pp. 284-291, Taylor & Francis (2016) [12] Rudner, M.:Cyber-threats to critical national infrastructure: An intelligence challenge.In: International Journal of Intelligence and CounterIntelligence. Vol. 26, pp. 453-481, Taylor & Francis (2013) [13] Hamin, Z.:Insider cyber-threats: Problems and perspectives.In: International Review of Law, Computers & Technology,Vol. 14,pp. 105113, Taylor & Francis (2000) [14] Raiyan, J.:A survey of cyber attack detection strategies.In: International Journal of Security and Its Applications (2014) . Vol. 8, pp. 247-256 [15] Buczak, A., Guven, E.:A survey of data mining and machine learning methods for cyber security intrusion detection. In: IEEE Communications Surveys & Tutorials, Vol. 18, pp. 1153-1176, IEEE (2016) [16] Roy, S., Ellis, C., Shiva, S., Dasgupta, D., Shandilya, V., Wu, Q.:A survey of game theory as applied to network security In: 2010 43rd Hawaii International Conference on System Sciences, pp. 1-10, IEEE(2010) [17] Wang, W., Lu, Z.:Cyber security in the smart grid: Survey and challenges. In: Computer Networks, Vol. 57, pp. 1344-1371, Elsevier (2013) [18] Elmaghraby, A., Losavio, M.:Cyber security challenges in Smart Cities: Safety, security and privacy. In: Journal of advanced research, Vol. 5, pp. 491-497, Elsevier (2014) [19] Da Cunha Rodrigues, G., Calheiros, R. N, Guimaraes, V. T.,Santos, G. L. D., De Carvalho, M. B., Granville, L. Z., Tarouco, L. M. R., Buyya, [27] Alotaibi, F., Furnell, S., Stengel, I., Papadaki, M.: A review of using gaming technology for cyber-security awareness In: Int. J. Inf. Secur. Res.(IJISR), Vol. 6,pp. 660-666, IJISR (2016)

29

[20] [21] [22] [23]

[24] [25]

[26]

[28]

[29] [30] [31] [32]

[33]

R.:Monitoring of cloud computing environments: concepts, solutions, trends, and future directions. In: Proceedings of the 31st Annual ACM Symposium on Applied Computing, pp. 378-383, ACM (2016) Kaufman, L. M.,:Data security in the world of cloud computing. In: IEEE Security & Privacy, Vol. 7, pp. 61-64, IEEE (2009) Somani, G., Gaur, M. S., Sanghi, D., Conti, M., Buyya, R.:DDoS attacks in cloud computing: Issues, taxonomy, and future directions. In:Computer Communications, Vol. 107, pp. 30-48, Elsevier (2017) Stergiou, C., Psannis, K. E, Kim, B., Gupta, B.:Secure integration of IoT and cloud computing. In:Future Generation Computer Systems, Vol. 78, pp. 964-975, Elsevier (2018) McGettrick, A., Cassel, L. N., Dark, M., Hawthorne, E. K., Impagliazzo, J.: Toward curricular guidelines for cybersecurity. In:Proceedings of the 45th ACM technical symposium on Computer science education., pp. 81-82, ACM (2014) Schneider, F. B.:Cybersecurity education in universities In:IEEE Security & Privacy, Vol. 11, pp. 3-4, IEEE (2013) Conklin, W. A., Cline, R. E., Roosa, T.:Re-engineering cybersecurity education in the US: an analysis of the critical factors. In: 2014 47th Hawaii International Conference on System Sciences, pp. 2006-2014, IEEE (2014) Hendrix, M., Al-Sherbaz, A., Victoria, B.,: Game based cyber security training: are serious games suitable for cyber security training?. In: International Journal of Serious Games, Vol. 3, pp. 53-61, Serious Games Society (2016) Cheung, R. S., Cohen, J. P., Lo, H. Z. Elia, F.:Challenge based learning in cybersecurity education. In: Proceedings of the International Conference on Security and Management (SAM), pp. 1, The Steering Committee of The World Congress in Computer Science (2011) Abawajy, J.:User preference of cyber security awareness delivery methods. In:Behaviour & Information Technology, vol. 33, pp. 237248, Taylor & Francis (2014) McGettrick, A.,:Toward effective cybersecurity education. In: IEEE Security & Privacy, Vol. 11, pp. 66-68, IEEE (2013) McDuffie, E. L., Piotrowski, V. P.: The future of cybersecurity education. In: IEEE Computer, Vol. 47, pp. 67-69, IEEE (2014) Kolli, Y., Mohd, T. K. and Javaid, A. Y:Remote Desktop Backdoor Implementation with Reverse TCP Payload using Open Source Tools for Instructional Use. In:2018 IEEE 9th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON). pp. 444-450, IEEE (2018) Rowe, D. C., Lunt, B. M. Ekstrom, J. J.:The role of cyber-security in information technology education. In: Proceedings of the 2011 conference on Information technology education,pp. 113-122, ACM (2011)

ISBN: 1-60132-498-7, CSREA Press ©

30

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Cognitive Training Xiaoli Yang Electrical & Computer Engineering Purdue University Northwest Hammond IN, USA [email protected]

Daniel Carrillo Electrical & Computer Engineering Purdue University Northwest Hammond IN, USA [email protected]

Mauro Puebla-Alvarez Electrical & Computer Engineering Purdue University Northwest Hammond IN, USA [email protected]

Abstract— Cognitive neuroscience focuses on questions of how cognitive activities are affected or controlled by neural circuits in the brain. The cognitive activities include attention, memory, perception, and more. Over time the cognition of individuals deteriorates which leads to weaker cognitive functions. Virtual reality is an immersive environment that when paired with an electroencephalogram (EEG), in this case the Emotiv Insight, can help users develop and enhance their cognitive skills. EEG data will be recorded from the headset and will be categorized into cognitive and emotional metrics including stress, excitement, engagement and more. This EEG data will be used as input parameters that will adapt the virtual reality environment in order to find the best pace for the development of cognitive skills for the user.

Cognitive Training Program will use a device called the HTC Vive, which utilizes a lightweight headset to be worn and used simultaneously with precise 360-degree controllers. Both devices rely on sensors to track movement of the devices as well as the built-in gyroscope to track the user’s movements and place them in the simulated environment. The set up for the HTC Vive, shown in Figure 1, demonstrates how the user should properly wear the headset as well as how the sensors should be placed in order to properly track the Vive.

Keywords— Virtual Reality(VR), Electroencephalogram (EEG), Cognition, Cognitive Training

I. Introduction Cognitive abilities show a small decline with advanced age in a majority of healthy individuals. The decline in these abilities vary from person to person, but on average, age hinders most cognitive abilities ranging from attention, to memory, to language. The amount of information recalled 30 minutes after hearing a story once is about 75 percent less by a person at the age of 70 compared to a person at the of 18 [1]. With technology advancing more and more every day, it is very easy to get lost mindless in one of the unlimited apps that are available to the public. Getting lost in the virtual space may not be beneficial at all to the development of cognitive skills. This program is meant to combine the best of both worlds, using the latest technologies of virtual reality to immerse the user in a virtual environment that is meant to have a positive effect on the development of cognitive skills. The user will be immersed into multiple environments and each of these environments will contain a game that targets a specific cognitive skill to improve. The games will be user friendly and interactive, allowing people of all ages to be able to adapt quickly. The program uses the Emotiv Insight as an input, personalizing the pace of the game to each individual users depending on EEG metrics that will be read. That is, the games will become more challenging as the user is ready to progress while the opposite is true and the game will become easier for the users that progress at a slower pace. Integrating both technologies, the HTC Vive virtual reality headset and the Emotiv Insight, will allow for the development and enhancement of cognitive skills. II. Virtual Reality As the name implies, virtual reality is a simulated environment commonly accessed by a head-mounted display with small screens in front of the user’s eyes [2]. The

Figure 1. HTC Vive setup [3].

III. Electroencephalogram A test to measure the electrical activity of a brain is called an electroencephalogram(EEG). The EEG is done by placing electrodes on a user’s scalp and the amount of electrodes used or the placement of the sensors can vary from different devices. For the following program, the brain activity is measured using the Emotiv Insight headset, a 5 channel wireless EEG shown in Figure 2.

Figure 2. Emotiv Insight sensors placement [4].

The Insight is a headset worn just like many other devices, but the size allows both the Insight and the Vive to be worn at the same time. There may not be as much sensors as other EEG devices, but the Insight still allows accurate readings of the user with its mobile and bluetooth connection displaying values read from each sensor. Using the International 10-20 system, Figure 3 displays the locations of the electrodes on the scalp where the Insight will get its readings from.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

31 of a user’s performance metrics including stress, engagement, interest, excitement, focus and relaxation.

Figure 4. Emotiv Insight Performance metrics Figure 3. Emotiv Insight International 10-20 sensor placement [5].

The Insight is able to display information needed for the program such as raw EEG data, changes in movement, changes in performance metrics that include stress, engagement, interest, excitement, focus, and relaxation. IV. Cognitive Training Cognition is the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses. There has been some research done focusing on improving the cognitive abilities, as well as applications created on mobile devices to play brain training games. Lumosity is an example of one of these applications, where the user can choose from over 60 different games and choose which area to train from categories that include memory, math, and vocabulary[6]. Another application for mobile devices, Mightier, utilizes a bluetooth heart rate monitor in order to adjust the difficulty of games[7]. Mightier is aimed for children who get upset easily, with the idea of teaching them to control their emotions by having to calm down in order to progress through the game. Both Mightier and Lumosity are both accessed on a mobile device, which is good for using at any location. However, using a phone screen to play through the games allows many distractions to occur depending on the surroundings of the user. The immersion of the mobile games is not as good as a VR game where the user is placed into a 3D environment rather than looking at a screen. Another benefit of the VR Cognitive training is the controls of the program that allow the user to interact with objects around them using the 360-degree tracked controller. The Lumosity and Mightier users are limited to only using a touch screen to complete the games. Using a person’s heart rate is a suitable way to detect if they are stressed, but only relying on stress may not be the best for cognitive training. More performance metrics will be useful in reflecting cognitive abilities. By using the EEG, the program is able to detect the changes in the user’s performance metrics, shown in Figure 4, and adjust the difficulty if their focus is low due to the game being too easy for example. Figure 4 is an example of a real time recording

The first game is the Conveyor game. Figure 5 shows what the user will see when using the Cognitive Training Program and selecting this game. The goal of this game is to enhance the users processing speed cognitive skill. The user will use both remotes programmed with ray tracing from Unity to distinguish where the user is pointing. Depending on the remote where the touchpad is pressed, the user will either teleport in order to navigate around the area for a better view or change the direction of the conveyors at intersections where there is an arrow. There will be a starting area where random colored cubes are dropped and will flow through the conveyor belt. The user will then change the directions of the arrows in order to direct the colored cube into its respective colored bin. The game displays which level the user is on, how many of their three strikes they have remaining, as well as the amount of correct boxes they have to direct to move on to the next level. Using the performance metrics, the game will either become more difficult by increasing the speed of the conveyor belt and increasing the number of cubes or the game will become less difficult by slowing down the speed of the conveyor belt and decreasing the number of cubes that drop. This will help the user progress through their game at their own pace working as fast as they want or as slow as they need.

Figure 5. Dynamically updating Conveyor game

The following shown in Figure 6 is the Mole sequence game. The purpose of this game is to help improve the user’s short term memory cognitive skill. As shown in Figure 7, the game will begin with a 3 by 3 matrix with only a sequence of 3 elements. When the start button is pressed by one of the remotes, the user will be shown moles pop up from the ground in a random sequence. The user will then have to input that same sequence into a wall with buttons. After the first

ISBN: 1-60132-498-7, CSREA Press ©

32

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | level, the game will then use the performance metrics to determine if the game should become more difficult by increasing the number of elements in the sequence or by increasing the size of the matrix. The contrary is true, if the game determines that the game should become less difficult, the game will decrease the size of the sequence and will decrease the size of the matrix. This will help the user progress at their own pace.

V. Conclusion and future work The virtual reality games have been created and tested. The EEG data retrieved from the Insight has been integrated with Unity in order to make the Unity games adapt to the users frustration and excitement. The games will become easier when there is no excitement and very high frustration, but will become more difficult when the stress is low and the excitement is high. With the projects done, testing the program is now the next step. A database will be created to record the progress of the users who will repeat the program throughout a specified time. This database will allow the users to track their progress from the beginning. At the same time, more games can be created to further develop and enhance other cognitive skills. [1]

Figure 6. Matrix that shows random sequence [2]

[3]

[4] [5] [6]

Figure 7. Buttons to input sequence shown

[7]

The Dana Foundation, “Cognitive Skills and the Aging Brain: What to Expect,”. [Online]. [Accessed 4 May 2019]. Available from: http://www.dana.org/Cerebrum/2015/Cognitive_Skills_and_the_Agin g_Brain__What_to_Expect/ B. Laurel, “What is Virtual Reality?”. [Online]. [Accessed 4 May 2019]. Available from: https://www.researchgate.net/publication/301891235_What_Is_Virtua l_Reality (28 Apr. 2016). Allan, Darren. “Thinking of getting an HTC Vive? Check that your room is this size first.” [Online]. [Accessed 4 May 2019]. Available from: https://www.techradar.com/news/wearables/thinking-of-gettinga-htc-vive-check-that-your-room-is-this-size-first-1314450 (5 Feb. 2016). Emotiv Inc. (2015). Emotiv Insight User Manual. San Francisco, CA. Author Emotiv Inc. “Insight-20-10”. [Online]. [Accessed 4 May 2019]. Available from: https://www.emotiv.com/?attachment_id=331261 (22 May. 2018). J. Hardy and M. Scanlon, "The science behind lumosity." San Francisco, CA: Lumos Labs. 2009. “Mightier Scientific Overview” (n.d.). [Online]. [Accessed 4 May 2019]. Available from: http://cdn.neuromotionlabs.com/MightierScientificOverview.pdf

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

33

Co-construction of Computer Science knowledge-to-be-taught in a French context T. Duron1 , V. Chiprianov2 and L. Gallon1 , 1 LIUPPA, University of Pau and Pays de l’Adour, Pau, France 2 IRISA, University of South Britain, Rennes, France Abstract— The French national curricula for elementary and secondary schools introduced teaching Computer Science (CS) concepts as mandatory, beginning the 2016-2017 school year. This also raised questions related to specifically what CS concepts should be taught and how. Several proposals of textbooks, pedagogical kits and other knowledgeto-be-taught have been made; some of them contain apparently surprising and even what seems, at a first glance, scientifically incorrect knowledge, which could prove to be obstacles in pupils’ learning. In this paper we analyze such proposals, and advance explanations based on the Theory of the Didactic Transposition of Knowledge (TDTK). The TDTK considers that the knowledge-to-be-taught is the result of a complex process of various interactions and negotiations between the numerous actors of the educational system. We identify such interactions, which explain the existence of didactic obstacles. Being aware of such caveats may reduce the apparition of this type of obstacles in future construction of similar CS bodies of knowledge-to-be-taught. Keywords: K-12, Computational Thinking, Didactics, curricula design and analysis

1. Introduction The mandatory introduction of Computer Science (CS) concepts in the French elementary (6-10 years old) and middle (11-14 years old) school national curricula, beginning the school year of 2016-2017 [1], has had profound impacts on what and how CS is taught in France. However surprising may this have been for some, it was not totally without precedent: in the 1980s there was a movement of teaching CS concepts (and training teachers accordingly), which was replaced in the 1990s with teaching Digital Literacy; for details on the history of teaching CS in France cf. [2]. Following the international trend on Computational Thinking [3], and reports, such as that of the Academy of Sciences [4], pinpointing France’s lateness in adopting proper CS teaching in elementary and middle school, the political decision was taken to introduce CS concepts in the national curricula for these levels. It should be noted that, due to historical decisions [2], at this point France did not have any primary or middle school teachers that had been formally trained in CS, only in Digital Literacy.

Fig. 1: Excerpt from Mission 3, IniRobot [5] As a consequence of this introduction, a number of resources, in the form of textbooks, (robot) pedagogical kits, recommendation documents, etc. have been proposed. Designing such proposals is important, however, it is only part of the continuous improvement education process. To evaluate the potential of such resources, questions need to be answered, such as: What is the knowledge such resources aim to teach? How is this knowledge actually taught by teachers and learned by pupils? Moreover, some of these resources contain a number of "features" that could, at first glance, be classified as errors, and which could become obstacles in pupils’ learning. How were such "features" introduced and possible in the first place? One such example concerns the IniRobot [5][6] pedagogical kit, using the Thymio1 robot for introducing robotics and programming, especially in elementary2 , but also middle school. IniRobot is distributed under a CC-BY license and has engendered several developments and versions. We focus here on the 2014 version [5]. IniRobot proposes a sequence of 14 missions; of which mission 3 If ... then ... (fr. "Si ... alors ...") introduces, in its own terminology, the concepts of event and action, from event-driven programming. Thymio has 6 preprogrammed behaviors, of which, mission 3 proposes exercises, to discover 4. The connect-the-dots exercise for the behavior called The explorer (yellow) (fr. "L’explorateur (jaune)") is presented in Fig. 1. The "events" (or rather, as we will see below, the conditions on which the events are filtered) are introduced by the word IF (fr. "SI"), and the actions are introduced by the word THEN (fr. "ALORS"). The expected result is the discovery of the explorer behavior (using the IniRobot terminology): "IF Thymio detects an object in front 1 https://www.thymio.org/en:thymio

2 https://dm1r.inria.fr/t/inirobot-descriptif-des-activites-autour-de-larobotique-a-lecole-primaire/23

ISBN: 1-60132-498-7, CSREA Press ©

34

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 2: The didactic transposition process, from [7] of it, THEN it goes back"; "IF Thymio detects an object at its right, THEN it turns left"; "IF Thymio doesn’t detect anything, THEN it advances"; "IF Thymio detects an objects at its left, THEN it turns right". The use of the word IF may come as a surprise, and rightly so, as the concept it introduces is related to an event. The Aseba textual editor for programming Thymio even uses the keyword WHEN, but the Aseba Visual Programming Language (VPL) uses the keyword IF. One would expect the use of the same word in the graphical editor as in the textual one - WHEN (fr. "QUAND"). While this may seem insignificant, there are cases in which pupils are disturbed (c.f. Fig. 3). In what follows we argue that such "features" are not simple "errors", but the result of a complex process of transformation of the scholarly knowledge into knowledgeto-be-taught. In this process, many actors are involved, with different concerns, numerous constraints, whose solutions are sometimes surprising ...

2. Didactic Transposition of Knowledge Knowledge, as it is taught in the School, is not immutable. It is a human construct, so as to fulfill a particular goal. As goals evolve through time and to follow changes in society, the knowledge taught in the School evolves as well. As some of the main goals are related to teaching beginners, the knowledge taught in school differs from the knowledge as it was first created by researchers: it is "simplified", differently structured, overlaps and contradictions have been limited the scholarly, scientific knowledge has been subject to a series of transformations, for it to become knowledge to be taught in School. It has been argued that "[b]odies of knowledge are, with a few exceptions, not designed to be taught, but to be used. To teach a body of knowledge is thus a highly artificial enterprise." [8]. In this paper we investigate how such transformations are happening in the CS curricula that is being defined currently in France. The decisions of which subjects, from the broad CS scientific knowledge, to choose for teaching is influenced by various actors, from high political levels such as ministers and academicians to actors closer to the terrain such as teachers, researchers as collaborators of teachers and as analysts of the system. To analyze this system and how it influences the transformations on the CS knowledge to be taught, we adopt a theoretical background based on the Theory of the Didactic Transposition of Knowledge (TDTK). Didactics is the science of the diffusion of knowledge in any institution (e.g. class of pupils, society at large). More particularly, it is the scientific study (and the knowledge

resulting thereof) of the innumerable actions taken to cause (or impede) the diffusion of such and such a body of knowledge in such and such an institution [9]. The TDTK has originally been proposed in French, in the 1980s, and has achieved a wide spread and acceptance in the Frenchspeaking communities, also in the Spanish-speaking, but much less so in the English-speaking communities, no doubt also because of relatively few and late translations (of which we selected a few in the bibliography of this paper, cf. infra). It is a theory initially proposed in the context of mathematics teaching, but has since evolved to encompass teaching of other science subjects, such as biology and geography. The Theory of the Didactic Transposition of Knowledge (TDTK) [8], [9], [10], studies the "transition from knowledge as a tool to be put to use, to knowledge as something to be taught and learned". Let us note that for us, in this paper, the concept of knowledge, in all of its further declinations, comprises both skills, know-how (fr. savoir faire), as well as "theoretical" knowledge (fr. savoir). The TDTK distinguishes thus several types of knowledge. The scholarly (bodies of) knowledge (fr. savoir savant) denotes "an organized and more or less integrated whole" [8]. It is produced by researchers and scholars, usually integrated in theories, to be found in research articles and scientific books. Taking this type of knowledge as starting point, the TDTK studies how it is transformed into knowledge-to-be-taught (fr. savoir à enseigner), which is the "scholarly knowledge that exists only in contexts than cannot be faithfully replicated within school" [8], usually as it appears in curricula, textbooks and other similar resources. This is further transformed into taught knowledge (fr. savoir enseigné) - "the knowledge which becomes visible, so to speak, in the classroom" [10], as it is presented in the particular context of a class, by a particular teacher, to a particular group of pupils etc. Of course, there is a further difference between the taught knowledge and what is actually learned by pupils, the learned knowledge (fr. savoir appris, connaissances). A synthetic view of the TDTK is presented in Fig. 2. These transformations are the result of interactions between actors of the educational system. The totality of these actors form what is called, in the TDTK, the didactic noosphere - the "sphere" of those who "think" about teaching [10], all those who share an interest in the teaching system. It consists of the agents - those actors who are in charge of knowledge (e.g. teachers), but also of the laity - those standing outside the teaching sphere. Obviously, different people have different positions in respect to education.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Therefore, "it is the task of the noosphere to cope with the demands made by society on the teaching system, by transmuting them into conditions acceptable to both parties - society and its teaching system" [10]. Consequently, the central function of the noosphere is the negotiation with the society, taking into account different conditions, constraints, resulting in compromises. We focus in this paper on analyzing the didactic transposition of CS knowledge related to the introduction of event programming based on IniRobot, as it happens in the context of a research and training project - PERSEVERONS3 . We do emphasize that other actions, knowledge and research are happening in PERSEVERONS, which are out of the scope of this paper. In this context, we have identified as agents of the educational system: the Curricula Superior Council (fr. Conseil supérieur des programmes) - the body which defines the national French CS curricula, CS and Learning Sciences researchers, CS pedagogical counselors (fr. Enseignants référents pour les usages du numérique), primary and secondary school teachers. In the next sections, we analyze how these agents are interacting in a process of negotiation and co-construction - a process of defining together, through interactions more or less direct - of a CS curricula.

3. The co-construction process In the context of our case study based on IniRobot, it seems as a fair assumption that the knowledge-to-be-taught should revolve around robots and their programming. We investigate in this section the transposition of such knowledge, from the definition of scholarly knowledge of what is called event-driven programming (one of the main paradigms used for programming robots), to its transformation into knowledge-to-be-taught, as found in the French national curricula and in the IniRobot pedagogical kit (and one of its declinations, the IniRobot for School), and further into taught and learned knowledge, as observed in regular French schools, while emphasizing the actions of different actors involved in each phase.

3.1 Defining the scholarly knowledge

Providing a complete historical and epistemological study on the CS concepts attached to event-driven programming is out of the scope of this paper. However, we do indicate that the concepts usually considered as in relation with the socalled event-driven paradigm have been defined in several places, in different manners, with different names, and although there are commonalities among these definitions, there are also numerous and important differences. For example, [11] defines a computational event as "anything that happens in the course of a computation [...] both occurrences in the program itself [...] as well as occurrences outside the computation proper". A similar and more precise 3 http://perseverons.espe-aquitaine.fr/

35

definition is that of [12], for which events are "the transitions between states that may appear in a system or in its environment". Please notice that these definitions consider both the external and the internal nature of events. Other authors, like for example Turing-award winner L. Lamport [13], offer a more complete picture, speaking of several concepts: events (which are not formally defined), processes, messages (defined as the means of communication among spatially distinct processes - distributed systems - and whose sending or receiving are considered an event), partial ordering (of the sending and reception of messages), logical and physical clocks (for synchronization), etc. [14] speaks about waiting for an event - in which case the program cannot complete an operation immediately and thus it registers a callback - a function that will be invoked when the event occurs. This waiting is typically done in a loop that polls for events and executes the appropriate callback when the event occurs. To differentiate between events, mechanisms of "filtering [...] by a condition" [12], which provide an event every time a condition is true, are needed. A concept similar to the callback is the event handler - a method "able to respond to one kind of external action (or event)" [15]. An event is defined as an external action. Other researchers make a difference between what it is called threaded (or procedure-oriented) and event-driven (or message-oriented) programming models/systems [14], trying either to show their duality e.g. [16], or the precedence of one over the other (at least in some contexts) e.g. [14][17]. Defining the concepts related to the so called "eventdriven programming paradigm" has therefore been a long process, in which several CS researchers (the only type of actors - and more precisely laity in TDTK terminology - in this phase) have built on or against the work of previous researchers. The resulting scholarly knowledge may have attributes that enable construction of real-world applications (use of knowledge), but it is in itself a rather difficult and specialized body to understand, sort and classify.

3.2 Defining the knowledge-to-be-taught : the curricula In France, a distinction is usually made between 2 types of knowledge-to-be-taught: the national curricula, resulted from the external transposition process of scholarly knowledge, and which serves as a reference document for the second type of knowledge-to-be-taught, which comprise textbooks, teacher training documents and other similar documents. The curricula is itself the result of a negotiation process between several organizations and councils. The Curricula Superior Council is the body which defines the national French curricula for elementary and middle schools. However, it does not work in isolation, and an analysis of its CS 2016 curricula [1] shows a number of influences, from a report of the Academy of Sciences [4], to other organizations such as

ISBN: 1-60132-498-7, CSREA Press ©

36

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

EPI4 and ATIEF5 . Regarding knowledge related to robots and event-driven programming, the French national curricula for middle school for example [1, pp. 365, 380], includes requirements such as: "write a program in which actions are triggered by exterior events", or "triggering of an action by an event". It is important to note that the CS concepts are mentioned in the curricula in 2 distinct parts: Mathematics and Technology. While the requirements evoked above are common to both parts, there are CS concepts specific to only one of them. As such, the Mathematics part also includes: "pupils are introduced to event programming (fr. programmation événementielle)", "programming actions in parallel", "control structures related to events", while the Technology part includes: "sensor, actuator, interface". One can see that the Mathematics part is more focused on the programming side, while the Technology part also includes more pronounced engineering and mechanical elements. One can identify, in the curricula, CS concepts that have been selected from the scholarly knowledge, such as event (which in the curricula seem to concern mainly external events - so messages?), action (probably similar to the concepts of callback/handler), triggering and control structures related to events (probably related to waiting and loop polling), actions in parallel (related to processes? or another name for events?). Why and how have been these concepts selected? A possible answer may come from what seems to be a major source of inspiration for the CS part of the curricula, a report of the Academy of Science [4, pp. 23, 25], which indicates, in relation to event-driven programming, concepts such as: "notion of parallel algorithm", "sensors and actuators", "algorithms [...] control the system by acting on actuators depending on the sensed values", "the retroactive command in a closed loop". However, one can notice that, while the Academy report presents together more mechanical (sensors, actuators) and more programmatic (algorithm, loop) concepts, the curricula segregates them into a Mathematics and a Technology part. How could this be explained? As mentioned in the introduction on the history of teaching CS in France, starting from the 1990s, teachers were trained mainly to Digital Literacy, but not to CS (programming). Therefore, in the 2010s, when the (political) decision to introduce CS teaching was taken, the Curricula Superior Council found itself confronted with the reality of having no teachers properly trained. It seems the retained solution was to separate the CS concepts into 2 subjects and assign their teaching to teachers which were the most likely to have connections to (and hopefully interest in) these concepts. The French national curricula is therefore a good example of how different actors, from political decision makers, the Academy of Science and the Curricula Superior 4 https://www.epi.asso.fr/ 5 http://atief.fr/

Council negotiated (indirectly), according to their purposes and with real-world constraints. One can also notice that the curricula identifies CS concepts to be taught, but does not enter into details. For example, triggering is mentioned, but not filtering by a condition; the concept of event appears, but no discussion of the time implications, on questions related to the partial ordering of sending and/or reception of events. As we shall see, this may have had implications on the further transposition of the knowledge-to-be-taught.

3.3 Defining the knowledge-to-be-taught: the case of pedagogical kits If a number of decisions regarding choices among scholarly knowledge, such as focusing on certain concepts of event-driven programming, as well as a separation, at middle school, into Mathematics and Technology, were taken, as a result of an external transposition process between laity of the educational system such as the Academy of Science and political decision makers, on the one hand, and agents such as the Curricula Superior Council, on the other hand, the identified concepts remain described rather vaguely and a number of choices still need to be made. One such more detailed specification is provided by the IniRobot [6] pedagogical kit. With a version made available in 2014 [5], so somehow in parallel with the development of the curricula itself, mainly by actors which seem to be primarily CS researchers or school teachers, with interests in Education Sciences research, IniRobot has several declared objectives, among which "the acquisition and practical use of a number of fundamental concepts", such as, for example: "sensors", "actuators", "instruction", "algorithm" and "how to use basic concepts of event-based programming", by elementary and middle school pupils [6]. It seems therefore safe to assume the authors of IniRobot followed the principle of the lowfloor (easy to get started) [18], and tried to avoid as much as possible known difficulties. "Programming in explicitly event-driven models is very difficult" [19]. Several causes have been identified for this, among which we find particularly relevant for our discussion the fact that the interactive logic of a program is fragmented across multiple event handlers [14][19]. While sequential programming is structured as a single flow of control, with control structures such as branching (i.e. IF) and loops, event-driven programming requires a series of small callbacks/handlers. The control flow among these handlers is not obvious, because of the inversion of the control [19]: a program merely registers with the execution environment its interest to be resumed on certain events; it is the execution environment which dispatches the events to the event handlers; it is not the program which calls the handlers. Thus, the control flow among handlers is expressed only implicitly, through manipulation of shared state [15].

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

It is maybe to avoid such difficulties related to understanding the fragmented and inverted control flow that the authors of IniRobot proposed the use of the word "IF" to introduce what they call an "event", as we have seen in Fig. 1. This may help "lowering the floor" [18], the entry point for pupils, when first encountering event-driven programming, by approaching the fragmented event-driven control flow to a sequential one. This seems the more likely as, in special (simple) cases, the control flow of an event-driven program may actually be sequential. Therefore, as long as the situations/exercises only require sequential code, pupils having a sequential representation of the event-driven programming does not show to be problematic. However, are they, in this case, really taught event-driven programming? Moreover, the similarity of using logical conditions: from the filtering of an event by conditions when triggering/waiting/loop polling, and, respectively, of the branch condition, increases the similarity between the 2 concepts. Nonetheless, the semantics of filtering of events by conditions (let’s denote it by the keyword WHEN) is different from that of a branch (IF). A filter has to be unique in a program, while branching conditions may appear multiple times. This becomes visible in some cases, such as that presented by the situation introduced by IniRobot in Mission 10, analyzed in more detail in Fig. 3. The 2014 version of IniRobot has evolved, engendering several branches and multiple versions. We focus here on a 2016 version, IniRobot for School [20], the basis chosen for activities in PERSEVERONS. First change to notice is the dimension of the document: from 24 pages for IniRobot, to 80 pages for IniRobot for School. The authors of IniRobot for School are CS pedagogical counselors, whose declared objectives comprise, among others: "working numerous competencies related to mastering the language (oral and writing), mastering mathematical languages and mastering scientific languages" [20, p. 2]. While there still are important objectives related mainly to teaching CS concepts, IniRobot for School uses the activities with Thymio "to practice languages", "to practice scientific and technological procedures" and "to organize the personal work" [20, p.3] as well, in relation with other objectives of the French national curricula (defined for all school subjects) [1]. While IniRobot consists of 6 sessions containing a total of 14 missions, with a total estimated completion time of about 6 hours (possibly longer, depending on audience), IniRobot for School consists of 12 sessions, during which contents from IniRobot missions 1, 2, 3, 4, 5, 6, 7, 8, 10 are partially reused, reordered, developed and enriched, for a total of an estimated completion time of about 12 hours. The authors of IniRobot for School have thus selected, according to their objectives, from IniRobot, contents related to CS concepts found in the curricula, excluded others (e.g. missions 11, 12, 13 from IniRobot) and added other knowledge (e.g. mission 1 on drawing Thymio), not necessarily re-

37

lated to CS, in order to augment the interdisciplinary content of their kit. Let’s also note that, in IniRobot for School, while concepts such as event and action are mentioned several times and exemplified, no mention is made to conditions or filtering/waiting; which is, as we have seen, in accordance with their lack of mention in the curricula.

3.4 Defining the taught knowledge At this phase in the transposition, it is the teachers’ turn to intervene in the process. Based on the curricula, textbooks, pedagogical kits, other resources and the training they (are supposed to) have received, the French educational system expects teachers to define situations to be taught in the classroom, that introduce and argue the need of CS concepts. These are high expectations, difficult to meet, especially by teachers with no initial training in CS. To support the teachers in this rather daunting task, the decision makers, most notably the Prime Minister and the Ministry of Education, have intervened through the e-FRAN6 Spaces for training, research and digital organization call for projects. It has selected 22 projects, for a total of 19.5 Million Euros. The successful projects demonstrated that they federate schools, county local authorities, companies, research laboratories, and other actors, around an innovative project with objectives regarding digital objects - whether to use them as a pedagogical resource, or in relation to new skills to be acquired, or as a research object -, taking place on diverse territories which enable a precise monitoring and evaluation (especially of pupils involved in experiments). As part of one such project, PERSEVERONS, we have the opportunity of observing and analyzing interactions between the numerous actors involved. These interactions have driven the transposition of CS knowledge-to-be-taught, as defined by the curricula, textbooks and pedagogical kits. We have observed several teachers and classrooms, keeping records in the form of notes, videos and photos, following an observation-based, case study, document and artifact-driven content analysis, qualitative research method7 [21]. It should come as no surprise that, even though the teachers used the same main pedagogical kit, IniRobot for School, there were differences in the way the class time was organized, the emphasis on some explanations or situations, the general and CS-specific teacher experience etc. However, some conditions were fairly common, such as pupils working in groups of 2 or 3, with 1 computer and 1 Thymio, in halfclassrooms of similar sizes, between around 10 and 15, for about 1 hour, in 1-2 sessions per week, for 4 - 10 weeks. This allowed observing some recurring phenomena across 6 http://www.education.gouv.fr/cid94346/appel-a-projet-e-fran.html

7 With its characteristic concern for context and meaning, of experiments in naturalistic occurring settings, in which the human investigator is the primary instrument for the gathering and analyzing of data, which is mainly descriptive, data obtained as a result of an emergent study design, and on which inductive analysis is applied.

ISBN: 1-60132-498-7, CSREA Press ©

38

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Fig. 3: Expected solution to programming Thymio’s explorer behavior, from [5] the different classes, among which the one related to the use of the IF keyword, described in more detail hereafter.

3.5 Defining the learned knowledge One phenomenon, a recurrent pupils’ "mistake" observed in several contexts and situations, can be exemplified by Mission 10 from IniRobot, optionally part of Session 10 of IniRobot for School. It asks pupils to program Thymio so that it moves about freely, avoiding all obstacles (corresponding to its explorer behavior, cf. Fig. 1); optionally, it asks to add instructions so that Thymio changes its color to red if it detects an obstacle, and to green if not. The expected solution (cf. Fig. 3 [5, p. 23]) consists of 4 instructions (grey blocks), each corresponding to an event (at the left of the ":" sign), and 2 actions (at the right ":"). In IniRobot’s terminology, their respective semantics would be [5, p. 14]: if Thymio does not detect anything with its front sensors, advance and change color to green; if Thymio detects something at its left, turn right and change color to red; if Thymio detects something in front of it, back away while turning a little and change color to red; if Thymio detects something at its right, turn left and change color to red. One "mistake" pupils make (cf. Fig. 4) is that, instead of adding a second action to the right of the ":" sign, for the same "event", they add a second instruction, with the same "event". This is signaled by the compiler as an error (in red at the top of Fig. 3) of the type: the events are the same; of course, for the program to be deterministic, the event callback/trigger has to be unique. While there may be several explanations for this "mistake", such as understanding that each "event" has only one corresponding action/instruction (while actually, at the right of the ":" sign, there is a block of instructions, which may contain an action of each type). However, one possible explanation, on which we focus here, is that the pupils constructed a representation of the event concept which is close to a sequential one, and which therefore allows them to verify the same condition (if all of Thymio’s frontal sensors do not detect anything, in the example of Fig. 3) several times. It would seem that constructing a representation of the event concept that includes time aspects (with a semantics of when (each and every time) all of Thymio’s frontal sensors do not detect anything, in the given example), would help in understanding

the uniqueness of the type of event and of its attached actions. While approaching the event-driven control flow to a sequential one may "lower the floor", a more appropriate representation needs to replace it as soon as possible, if "real" event-driven programming is to be taught. While this is a fairly straightforward example, its analysis helps understanding difficulties encountered at each phase of the didactic transposition co-construction process, from (1) the esoterism of the scholarly knowledge on event-driven programming (in which time considerations are made using concepts such as partial ordering, and conditions are defined as part of a filtering of events, in a waiting/triggering/loop polling mechanism), defined by sometimes competing, sometimes continuing CS researches, passing through (2) the selection of concepts to-be-taught, selection influenced by political decision makers, academic institutions (laity in the sense of TDTK), national curricula definition bodies (which in our case study seem to leave out these time and conditions considerations), (3) the redefinition and reorganization into textbooks, pedagogical kits, or similar resources, by researchers, teachers and pedagogical counselors mainly concerned with education issues, with easing their understanding by pupils (and thus, in our case, seemingly approaching the event-driven flow of control to a sequential one by using keywords like "IF" instead of "WHEN"), (4) the development of classroom lessons by teachers and finally (5) the continuous evolution of pupils’ representation of the taught concepts, who by "mistakes" experiment/test and construct this representation (in our case, e.g. assuming that conditions/events are not unique).

4. Conclusion While most current research related to Computational Thinking (CT), especially in a K-12 context, has focused mainly on designing activities for teaching CT concepts, such pedagogical documents need to be analyzed and experimented in classrooms, so that they can be validated and improved. In this paper we adopted a qualitative approach, as a first step to identifying phenomena that appear recursively in classrooms. This allowed us to identify didactic obstacles to the learning of pupils. Further social and epistemologicaldriven analysis revealed causes of such didactic obstacles in

Fig. 4: Typical pupils’ mistake

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

the complexity of the didactic transposition process involving numerous actors, with various constraints and objectives. More specifically, we focused on event-driven programming, observing and analyzing during 4 years, how the IniRobot for School pedagogical kit using the Thymio robot, was used to teach and learn concepts related to the eventdriven paradigm. We thus pinpointed that concepts of eventdriven programming form the main focus of CS knowledgeto-be-taught in elementary and middle school (other fully arguable choices include robotics, intelligent (cooperative) (autonomous) agents, ethics etc). In an epistemological approach, we reviewed scholarly, research-introduced definitions of concepts that currently are considered to be related to what is called event-driven programming, showing their relations and underlining their complexity (e.g. external vs. internal events, messages vs. processes). We analyzed how some of these concepts were selected and partially redefined (e.g. event, action) in the French national curricula and identified how the lack of mention of certain concepts (e.g. filtering on conditions), may contribute to further difficulties. We found that these selections are indeed reflected in the pedagogical kits that implement the curricula, and are compounded with didactic and pedagogical concerns of "simplifying" the taught concepts (e.g. presenting the fragmented event control flow as a sequential one). This analysis (in particular the review of the scholarly knowledge) may serve as a basis in future designs of event-driven programming pedagogical proposals. In this work, we mainly focus on didactic obstacles. While, for example, the difficulty of understanding the fragmented control flow of event-driven programming is an epistemological obstacle, the didactic choice of the "IF" keyword (while intended precisely to reduce this epistemological obstacle), with its unforeseen side effects, introducing didactic obstacles (sequential representation of the occurrence of events), created by the very redefinition (out of didactic concerns) of some event-driven concepts. While they are probably unavoidable in any didactic transposition process, it is important to recognize such obstacles, be aware of them, and try to address them further along the line. This analysis can therefore be used to draw the attention, in particular, to curricula and pedagogical documents (textbooks, kits, etc.) designers, to pay singular attention to how they redefine scholarly concepts, as these "new" representations - the knowledge-to-be-taught, may introduce its own obstacles. Moreover, this analysis may serve researchers and analysts of the knowledge-to-be-taught (be it curricula, textbooks, pedagogical kits) as caution to drawing conclusions too hastily, and instead exercise a deeper consideration of the wider context and the complex interactions between the involved actors (be they agents of the education system or laity), which may influence decisions seeming, at a first glance, "weird" or simply "erroneous". It thus enlightens the fact that the "final" result of the taught knowledge is

39

the product of a process of co-construction, in which many actors intervene at different stages.

References [1] M. de l’Éducation Nationale de l’Enseignement Supérieur et de la Recherche MENSR, “Programmes d’enseignement du cycle des apprentissages fondamentaux (cycle 2), du cycle de consolidation (cycle 3) et du cycle des approfondissements (cycle 4),” Bulletin Officiel no 11 du 26 novembre, 2015. [2] G.-L. Baron, B. Drot-Delange, M. Grandbastien, and F. Tort, “Computer science education in french secondary schools: Historical and didactical perspectives,” ACM Transactions on Computing Education (TOCE), vol. 14, no. 2, p. 11, 2014. [3] J. M. Wing, “Computational thinking,” Commun. ACM, vol. 49, no. 3, pp. 33–35, Mar. 2006. [4] A. de Sciences AdS, “L’enseignement de l’informatique en france. il est urgent de ne plus attendre,” 2013. [5] T. Guitard, D. Roy, P.-Y. Oudeyer, and M. Chevalier, “IniRobot. Activités robotiques avec Thymio II pour l’initiation a l’informatique et a la robotique,” 2014. [Online]. Available: https://dm1r.inria.fr/t/inirobot-les-documents-a-telecharger/141 [6] D. Roy, G. Gerber, S. Magnenat, F. Riedo, M. Chevalier, P.-Y. Oudeyer, and F. Mondada, “IniRobot : a pedagogical kit to initiate children to concepts of robotics and computer science,” in RIE 2015, Yverdon-Les-Bains, Switzerland, May 2015. [7] M. Bosch and J. Gascón, “Twenty-five years of the didactic transposition,” ICMI Bulletin, vol. 58, pp. 51–65, 2006. [8] Y. Chevallard, “On didactic transposition theory: Some introductory notes,” in International Symposium on Research and Development in Mathematics, Bratislava, Czechoslavakia, 1988. [9] ——, “Readjusting didactics to a changing epistemology,” European Educational Research Journal, vol. 6, no. 2, pp. 131–134, 2007. [10] ——, “A theoretical approach to curricula,” Journal fuer Mathematikdidaktik, vol. 13, no. 2-3, pp. 215–230, 1992. [11] D. Jusak and J. Hearne, “Language and runtime support for event programming in a distributed system,” in [1990] Proceedings. Second IEEE Workshop on Future Trends of Distributed Computing Systems, Sep 1990, pp. 514–519. [12] P.Caspi and N.Halbwachs, “An approach to real-time systems modeling,” in Int. Conference on Distributed Computing Systems, 1982. [13] L. Lamport, “Time, clocks, and the ordering of events in a distributed system,” Commun. ACM, vol. 21, no. 7, pp. 558–565, July 1978. [14] F. Dabek, N. Zeldovich, F. Kaashoek, D. Mazières, and R. Morris, “Event-driven programming for robust software,” in Proceedings of the 10th Workshop on ACM SIGOPS European Workshop, ser. EW 10. ACM, 2002, pp. 186–189. [15] B. Chin and T. Millstein, “Responders: Language support for interactive applications,” in Proceedings of the 20th European Conference on Object-Oriented Programming, ser. ECOOP’06. Berlin, Heidelberg: Springer-Verlag, 2006, pp. 255–278. [16] H. C. Lauer and R. M. Needham, “On the duality of operating system structures,” SIGOPS Oper. Syst. Rev., vol. 13, no. 2, pp. 3–19, 1979. [17] R. von Behren, J. Condit, and E. Brewer, “Why events are a bad idea (for high-concurrency servers),” in Proceedings of the 9th Conference on Hot Topics in Operating Systems - Volume 9, ser. HOTOS’03. Berkeley, CA, USA: USENIX Association, 2003, pp. 4–4. [18] M. Resnick, J. Maloney, A. Monroy-Hernández, N. Rusk, E. Eastmond, K. Brennan, A. Millner, E. Rosenbaum, J. Silver, B. Silverman, and Y. Kafai, “Scratch: Programming for all,” Commun. ACM, vol. 52, no. 11, pp. 60–67, Nov. 2009. [19] P. Haller and M. Odersky, “Event-based programming without inversion of control,” in Proceedings of the 7th Joint Conference on Modular Programming Languages, ser. JMLC’06. Springer-Verlag, 2006, pp. 4–22. [20] J. Sagné, E. Page, and C. Lefrais, “Séquence IniRobot scolaire. «langages et robotique »,” 2016. [Online]. Available: http://tice33.acbordeaux.fr/Ecolien/Langagesetrobotique/tabid/5953/language/frFR/Default.aspx [21] D. Ary, L. C. Jacobs, C. K. S. Irvine, and D. Walker, Introduction to research in education. Cengage Learning, 2018.

ISBN: 1-60132-498-7, CSREA Press ©

40

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

An Assessment Study on the Teaching of Critical Thinking and Mathematical Proofs in a Discrete Structures Course Hang Dinh Department of Computer and Information Sciences Indiana University South Bend [email protected] Abstract— Critical thinking and mathematical proofs are two learning outcomes of focus in our Discrete Structures course. In this paper, we study how these two learning outcomes affect each other. Our study is based on assessment data from two editions of our Discrete Structures course, with mathematical proofs being taught before critical thinking in the first edition and taught after critical thinking in the second edition. Results of our data show that students in the first edition performed significantly better in understanding proofs and better in critical thinking than students in the second edition did. However, the difference between the mean scores on understanding proofs of the two editions is not statistically significant. These results suggest that teaching/learning mathematical proofs may help students understand critical thinking better, while teaching/learning critical thinking may not help students understand proofs better. Keywords: critical thinking, mathematical proofs, teaching methods, assessment, discrete structures

1. Introduction Critical thinking is one of the essential learning outcomes for students in the 21st century, according to LEAP [1]. The importance of critical thinking of course also applies to Computer Science students. In particular, the ACMIEEE Computer Science Curriculum Guidelines 2013 [2] recognized critical thinking as one of the skills that students need to develop during their undergraduate career. Different institutes may have different approaches to include critical thinking into their computer science curriculums. For example, one [3] may design a critical thinking-focussed course that is tailored for engineering or computer science students, while others [4], [5] may embed critical thinking exercises in various Computer Science courses. At our liberal art institute, critical thinking is a required component in the General Education curriculum. As such, our Computer Science majors would usually take a critical thinking course from the Philosophy department. Because of the need to reduce students’ total required credit hours, we proposed to include a Critical Thinking Module (CTM) in our Discrete Structures course, making this course satisfy the critical thinking requirement of our campus-wide General Education

curriculum. Details about adding the CTM to our Discrete Structures course have been published in our previous work [6]. Since Fall 2014, our Computer Science majors no longer need to take a critical thinking course from another department. The Discrete Structures course is a core course in our Computer Science curriculum. It covers all the Core-Tier1 topics listed in the first three parts of the Discrete Structures (DS) Body of Knowledge that is specified in the ACM-IEEE Computer Science Curriculum Guidelines 2013 [2, p.77-78]. Those three parts are: • • •

DS/Sets, Relations, and Functions DS/Basic Logic DS/Proof Techniques.

Of these three parts, the DS/Proof Techniques part is usually the most challenging to teach to Computer Science students. In our experience, students in our Discrete Structures courses often struggle the most with understanding and constructing mathematical proofs. We believe that many other instructors of Discrete Structures or Discrete Mathematics also share similar experience since proof is a notoriously difficulty concept for even math majors [7], [8]. For example, Gries and Schneider [9] from Cornell pointed out that “Generally speaking, mathematicians and computer scientists are not satisfied with the level at which college students understand math. Students have difficulty constructing proofs, their reasoning abilities are inadequate, and they don’t know what constitutes rigor.” Since the Critical Thinking Module is added to our Discrete Structures course, we observe some connections between critical thinking and mathematical proofs. In particular, both critical thinking and understanding mathematical proofs require skills for deductive reasoning, reconstructing extended arguments, recognizing unstated assumptions, and distinguishing non-argumentative language from argumentative language. These connections naturally led us to wonder whether teaching critical thinking would help improve students’ understanding of proofs, or vice versa?. Initially motivated by this question, we conducted our study in two consecutive editions, offered in Fall 2016 and Fall 2017, of our Discrete Structures course. In the first edition, we covered mathematical proofs before covering the Critical

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Thinking Module. In the second edition, we switched the two topics around, by covering the Critical Thinking Module before mathematical proofs. Details of the module order in these editions of our Discrete Structures course are given in Table 1 below. We used essentially the same course materials and assessments in both editions, only changed the order of certain lecture notes and assessment items to match the change in the order of topics. We then collected and analyzed data about student performance on the tests regrading critical thinking and mathematical proofs. Data about student performance on the logic part were also collected, as logic is the foundation on which critical thinking and mathematic proofs are taught. Table 1: Order of major topics in our Discrete Structures course Module No. 1. 2. 3. 4.

Fall 2016 Basic Logic and Sets Mathematical Proofs Relations and Functions Critical Thinking

Fall 2017 Basic Logic and Sets Critical Thinking Mathematical Proofs Relations and Functions

Our data show that students from the Fall 2016 cohort performed better on both understanding proofs and critical thinking than those from the Fall 2017 cohort, although the Fall 2017 cohort started out with slightly better performance in logic. While the average score on understanding proofs of the Fall 2016 class is 4% higher than that of the Fall 2017 class, the two-sample two-tailed t-tests indicate that the difference between the average scores on understanding proofs is not statistically significant. These results suggest that teaching mathematical proofs may help improve students’ critical thinking skills, but teaching critical thinking may not help improve students’ understanding of mathematical proofs. In addition to data on student test results, we also conducted a survey embedded in the course evaluation for both Fall 2016 and Fall 2017 classes. The survey consists of a question that asks students if they feel that critical thinking helped them understand proofs better. Student responses to this question, however, are neutral with roughly half of responses from each class either agreeing or strongly agreeing that critical thinking has helped them understand mathematical proofs better. The remaining parts of this paper are organized as follows: In Section 2, we will describe the details of our Discrete Structures course that are relevant to our study, including the specific topics we covered as well as assessment items used in the course. The next section, Section 3, contains description of how our data were collected and our analysis of the data. The conclusions and our thoughts drawn from our data results will be discussed in the last section.

41

2. Discrete Structures Course Details 2.1 Course Topics Our Discrete Structures course content consists of four modules corresponding to four major topics shown in Table 1. The first module, “Basic Logic and Sets,” covers basic knowledge that set up the foundation for the three remaining modules. In particular, we covered the following topics in the Basic Logic and Sets module: a. Sequential/Propositional logic b. Variables and sets, basic operations on sets, power sets c. Quantificational/Predicate logic d. Inference rules and formal proofs. Our covering of sets and formal proofs in the Basic Logic and Sets module is compatible with Gries and Schneider [9]’s equational treatment of logic in their approach to teaching Discrete Mathematics, which helps introduce rigor in proofs to students. The Mathematical Proofs module consists of two submodules with detailed topics as follows: 1) Proof Techniques a) Disproof by counterexamples. Direct proofs. Deduction method. b) Proofs by contraposition and contradiction c) Proofs involving quantifiers 2) Mathematical Induction and Applications a) Weak induction b) Recursion c) Strong induction. In particular, the topics in our Mathematical Proofs module closely match the Core-Tier1 topics in the DS/Proof Techniques part of the ACM-IEEE Computer Science Curriculum [2, p.78]. The Critical Thinking module covers the following topics: a. Reintroducing formal logic, argumentation, and verification in a broader context b. Critical analysis and identifying arguments. Definitions. Stated and unstated assumptions. Evidence (e.g. argument premises) evaluation c. Case study (includes multiple points of view). The last module “Relations and Functions” covers the Core-Tier1 topics about relations and functions in the DS/Sets, Relations, and Functions part of the ACM-IEEE Computer Science Curriculum [2, p.77]. Note that the topics about sets in this part are covered in our Basic Logic and Sets module.

2.2 Course Assessment The two main instruments for assessment in our Discrete Structures course are homework assignments and tests. Prior to Fall 2016, homework assignments in our Discrete Structures course were graded for accuracy or correctness. This

ISBN: 1-60132-498-7, CSREA Press ©

42

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

would result in inaccurate assessment of students performance since many students would copy work of each other even though the homework was meant to be done individually. Thus, in Fall 2016, we changed our grading method for homework assignments in order to encourage students to do homework by themselves honestly. In particular, we would grade homework for completion and efforts instead of correctness. Because of this nature of homework grading, we exclude homework scores from our assessment study. The tests in our Discrete Structures course include five quizzes and three exams (two midterm exams and one final exam). All questions in the quizzes and most questions in the exams are multi-choice (MC) or true-false (TF). Each exam has one or two free-form questions that ask students to construct a solution. The set of MC and TF test questions given in Fall 2016 is the same as that given in Fall 2017 except for one modification: the answer choices of every multiple-choice question given in Fall 2016 tests were shuffled when the question was given in Fall 2017 tests. Almost all free-form questions given in Fall 2016 also match those given in Fall 2017, except for points possible awarded to each question (See Table 5 below). Test questions are in levels 2 through 5 in Bloom’s Taxonomy: (2) Comprehension, (3) Application, (4) Analysis, (5) Synthesis. Thus, all the tests were open notes. For our study, we collected data about students’ scores on the tests related to three modules in our Discrete Structures course: Basic Logic and Sets, Mathematical Proofs, and Critical Thinking. We regrouped questions in these tests into four groups, named Logic, Critical Thinking, Understanding Proofs, and Constructing Proofs. Questions in the Logic group are all the multiple-choice (MC) and true-false (TF) questions related to the Basic Logic and Sets module as shown in Table 2. Note that there is only one free-form question related to this module, which was left out of our data analysis for two reasons. First, this free-form question is worth 10 points which is only a small fraction of the total points possible (103 points) for all questions related to the Basic Logic and Sets module. Second, grades of free-form questions are always subjective, and thus are less accurate for assessment than grades of MC and TF questions. Although some of the questions in this group are about basic sets, we call this group Logic because the majority of questions in this group are about logic and even questions about sets are also related to logic. Questions in the Critical Thinking group are all the questions related to the Critical Thinking Module, as shown in Table 3. Questions in the Understanding Proofs group are all the multiple-choice and true-false questions related to the Mathematical Proofs module as described in Table 4. Questions in the Constructing Proofs group are all the freeform questions related to the Mathematical Proofs module as shown in Table 5.

Table 2: Test items related to the Basic Logic and Sets module Fall 2016

Fall 2017

Question Points Typea Possibleb MC, 8 TF MC, 9 TF

Quiz 1

Quiz 1

Quiz 2

Quiz 2

Quiz 3

Quiz 3

MC, TF

10

Exam 1, Questions 1-35 Exam 1, Question 36

Exam 1, Questions 1-35 Exam 1, Question 36

MC, TF

35

Free Form

10

Exam 2, Questions 1-15 Exam 2, Questions 23-26 Final Exam, Questions 110 Final Exam, Questions 39-40

Exam 2, Questions 1-15 Exam 2, Questions 29-32 Final Exam, Questions 110 Final Exam, Questions 33-34

MC

15

TF

4

MC

10

TF

2

Detailed Topics Propositional logic Logical equivalence, variables, and sets Inference rules and formal proofs Basic Logic and Sets Truth table, tautology, contradiction Inference rules and formal proofs Inference rules and formal proofs Basic Logic Basic Logic and Sets

a MC denotes multiple-choice questions, and TF denotes true-false questions. b Each MC or TF question is worth one point.

Table 3: Test items related to the Critical Thinking module Fall 2016 Final Questions Final Questions

Fall 2017 Exam, 26-38 Exam, 49-50

Exam 2, Questions 16-28 Exam 2, Questions 33-34

Question Type MC

Points Possible 13

TF

2

Table 4: Test items for Understanding Proofs Fall 2016

Fall 2017

Quiz 4

Quiz 4

Exam 2, Questions 16-22

Final Exam, Questions 26-32

Exam 2, Questions 27-30 Final Exam, Questions 11-14 Final Exam, Question 41

Final Exam, Questions 43-46 Final Exam, Questions 11-14 Final Exam, Question 35

ISBN: 1-60132-498-7, CSREA Press ©

Question Points Type Possible MC, 7 TF MC 7

TF

4

MC

4

TF

1

Detailed Topics Proof techniques Disproof, proofs by contraposition & contradiction, deduction methods Proof basics Mathematical induction & recursion Proof basics

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 5: Test questions for Constructing Proofs. All of these questions are free-form. Fall 2016 Question Points Possible Exam 2, 3 Question 31a Exam 2, 4 Question 31b

Exam 2, Question 31c Final Exam, Question 51a Final Exam, Question 51b

6 2 8

Fall 2017 Question Points Possible Exam 2, 3 Questions 35a

Exam 2, Question 35b Final Exam, Question 47

3

Final Exam, Question 48a Final Exam, Question 48b

2

5

7

Objectives

43

 
 

 
$

 
 


Develop scratch work for a proof involving quantifiers

Write a proof involving quantifiers Develop scratch work for a proof by induction Write a proof by induction

    
 

We use both direct and indirect methods of assessment in our study. Our direct method includes students’ test results, while our indirect method includes a survey embedded in the official course evaluation.

3.1 Test Results

Logic Fall 2016 Fall 2017 Difference

73.12% 74.71% -1.59%

Critical Thinking 67.41% 65.4% 2.01%

Understanding Constructing Proofs Proofs 71.74% 67.7% 4.04%

62.92% 61.57% 1.35%

The average scores as percentage for all four test groups (Logic, Critical Thinking, Understanding Proofs, and Constructing Proofs) are summarized in Table 6 for convenient comparison. Although the scores for the Constructing Proofs group are subjective and there are discrepancies in the points

 
 


   

'! * "! "$ "* "" "%!* # #!%$* $ $##'* $& %! * $' %"'* % %# %* % %$!"* % %%" * % %%" * %! %&"'* %" %'#%* %" %'#%* %# &$#* %# &$#* %% & &* %& &!&%* %& &!&%* '#!",

"" & !# ) !$ !% ! !! ' ( % # ! !) "! $ " !"

!!& ,

# $ &     %   '        

   
   

 

  
  
  
 

* "* #!!!* %!!!* $$$%* $$$%* &* "$$%* %!!!* $$$%* $* &* $$$%* $$$%* %!!!* &* %!!!* $$$%* &* &'$!,

! ' # % # !  &  & # & # & & " $ & '

!! ),

* !'!* $# * %!'* $# * #$# * '!* %& $* &$'$* %& $* $# * %& $* $# * %& $* %& $* $&%* $'#%* %& $* & $* '!'$,

! !# #%# '# % !# "# % & ! ## $#   '%#   ! 

!!)',

* #&%* #* &"%&* %!'* #&%* $!"* %!'* %& $* #$# * $%!'* %%"* "%&!* "%&!* &#&%* '!* "%&!* #$# * "!"&* &")", !&&',

 
 

 
%

3. Assessment Data

Table 6: Summary of average scores from Figures 1 and 2



 

Fig. 1: Test results from the Discrete Structures course in Fall 2016

 
 


Students’ test results for all the aforementioned four test groups are shown in Figure 1 (for the Fall 2016 class) and Figure 2 (for the Fall 2017 class), in which the lists of students are sorted by their scores for the Logic group. The students’ actual identification had been removed before we began compiling the data. The student numbers shown in the first columns of these two figures were randomly assigned between 1 and the class size. The class sizes of the Fall 2016 class and the Fall 2017 class are 22 and 27 students, respectively. Students who missed a test from which data is collected were also removed from the data collection. Thus, in the final data, there are only 18 students in the Fall 2016 class and 21 students in the Fall 2017 class.


 

    
 


 

! !% !& "" "$ " ' $ & ) ( "& "! !# "# % !( !! "' !' !)

 
 


   
   



  
  
  
     

'! * "% ##"* #$ $ * #' $!""* #' $!""* #' $!""* $ $##'* $" $&& * $$ %'%* $& %! * $' %"'* % %$!"* % %$!"* % %%" * %! %&"'* %" %'#%* %$ &% * & &$ * & &%* & &%* &! &' #* &' '#%* '$'!,

# * & #!!!* & #!!!* % "$$%*  $$$%* ' $* & #!!!* ' $*  $$$%*  %!!!* & #!!!*  $$$%* & #!!!* " '!!!* & #!!!*  %!!!* ' $*  $$$%* & #!!!*  &* ! &$$%* # * &%$ ,

! ) ,

!$" ,

! $  $ # &  #  & $ % % $ %  $ ' &   %

* $'#%* "!"&* $'#%* $# * !"%&* # %* $# * # %* %& $* $'#%* %!'* !"!* $'#%* %!'* '!* $'#%* & $* %& $* &$'$* '!* %!'* &'' , !&#%,

 "  ' #%# & # ## # !# '# # & % # $ # '  !# "# $ # # &

 

  * * ##* "#* %'* "* %&* #!* $&* "&* #&* "* !$* &* #&* '#* ##* $&* %!* &* %$* '* &!%', !((%,

Fig. 2: Test results from the Discrete Structures course in Fall 2017

possible for questions in this group between the two classes, we include them in our data for reference. We will be focusing on analyzing the data for the other three test groups for the scores in those groups are objective. These data show that the average score on Understanding Proofs of the Fall 2016 class is 71.74%, which corresponds to a C- grade, while that of the Fall 2017 class is only 67.7%, which corresponds to a D+ grade. This means that an average student in the Fall 2016 class would get 4% higher, which is a letter grade higher, on Understanding Proofs than their Fall 2017 counterpart. The average score on Constructing Proofs of the Fall 2016 class is also slightly better (about 1.3% higher) than that of the Fall 2017 class. The superior performance of the Fall 2016 class on proofs is more significant considering that they started out more poorly in Logic than the Fall 2017 class did. Specifically,

ISBN: 1-60132-498-7, CSREA Press ©

44

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

while the the Fall 2017 class scored only about 1.6% higher on average than the Fall 2016 class in Logic, its distribution of scores on Logic is much better than that of the Fall 2016 class, as shown in the histograms in Figure 5.  


 


 


  





 


 




 









 







 


 


 


 
 


 


 


 


 


 


 


 


 


 


 


 


Fig. 3: Relation between Understanding Proofs scores and Logic scores The chart in Figure 3 shows the relation between students’ scores on Understanding Proofs and their scores on Logic. The trend line for the Fall 2016 class in this chart lies above the trend line for the Fall 2017 class. This indicates that for students with the same score on Logic, those in the Fall 2016 class are more likely to have higher scores on Understanding Proofs than those in the Fall 2017 class. However, this trend is less profound for students with higher scores on Logic, based on the angles of the two trend lines. In particular, the higher score on Logic, the less gap between the Understanding Proofs score on the Fall 2016 class trend line and the Understanding Proofs score on the Fall 2017 class trend line. 








  
 






































 


 


 


 


 


 


 


 


 


 


 


  


Fig. 4: Relation between Critical Thinking scores and Logic scores

The Fall 2016 class also performed better on Critical Thinking than the Fall 2017 class did. However the gap between the the two classes on Critical Thinking is much less significant than that on Understanding Proofs. More specifically, the Fall 2016 class’ average score on Critical Thinking is 2% higher than the Fall 2017 class’ corresponding average score. On the chart in Figure 4 which shows the relation between Critical Thinking scores and Logic scores, the trend line for the Fall 2016 class also lies above the trend line for the Fall 2017 class, although these two lines are closer to each other than the two trend lines in the chart (in Figure 3) showing Understanding Proofs scores with respect to Logic scores. It remains to investigate whether the differences in the average scores between the two classes are statistically significant or by chance. For this purpose, assuming that the Fall 2016 (resp. Fall 2017) class is a random sample from the population of students taking our Discrete Structures course with mathematical proofs taught before (resp. after) critical thinking, we will perform the two-sample two-tailed t-testsfor each of the three groups: Logic, Critical Thinking, and Understanding Proofs. For each of these three groups, we calculated two p-values associated with the t-test: one with the assumption that the scores in the given group from the two populations have unequal variances, and one with the opposite assumption. The p-value is the probability of making the mistake of concluding that there is a difference between population means when there is actually no difference. The calculated p-values are shown in Table 7, in which the p-values with the unequal variances assumption are shown in the first row and those with the equal variances assumption are shown in the bottom row. Note that since the standard deviations for the Logic group are almost equal (11.6% for Fall 2016 and 10.9% for Fall 2017), the two p-values for the Logic group are equal up to a constant at most 0.01. For the Critical Thinking and Understanding Proofs groups, the two p-values for each of the group are also close to each other since the difference in the corresponding standard deviations is small. Before drawing any conclusion from these p-values, we need to make sure that the samples meet the requirements for a t-test, that is, either each sample must have normal (or reasonable symmetric) distribution or sample sizes are sufficiently large (> 20). Since one of our sample sizes is not sufficiently large, we need to examine if our samples have approximately normal distribution, which can be seen though histograms. The histograms for the Logic scores in Figure 5 show that the one from Fall 2017 class is reasonably symmetric while the one from Fall 2016 is not. For the Critical Thinking group, none of the histogram, shown in Figure 6, are reasonably symmetric. Only the Understanding Proofs group has both of the histograms that are reasonably symmetric, as shown in Figure 7. Thus, the p-values are valid for only the Understanding Proofs group.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

45

Table 7: P-values for two-sample (or unpaired) two-tailed t-tests

P-Value, assuming unequal variances P-Value, assuming equal variances

0.67

Critical Thinking 0.63

Understanding Proofs 0.39

0.67

0.64

0.40







Logic







  










  











Since the p-values for the Understanding Proofs group (0.39-0.4) are higher than significant level α = 0.05, we will retain the null hypothesis that “The two samples, Fall 2016 and Fall 2017, come from the populations with the same mean score on Understanding Proofs”. This doesn’t mean that we have proven this null hypothesis, but rather that we don’t have sufficient evidence to reject it. To say that there is a difference between the population means on Understanding Proofs scores is taking a 39% (or 40%) risk of being wrong. One may draw similar conclusions with respect to the Logic and Critical Thinking groups. However, such conclusions may be unvalid because data for these two groups are too skewed. 


!


!


!


!


!


!


!


Fig. 7: Histograms for the scores in the Understanding Proofs test group

understand mathematical proofs betters.” Students responses to this questions from the Fall 2016 class and the Fall 2017 class, as appeared in the official course evaluation, are shown in Figure 8 and Figure 9, respectively. $IWHUOHDUQLQJWKHFULWLFDOWKLQNLQJPRGXOHLWKHOSVPHXQGHUVWDQGPDWKHPDWLFDOSURRIV EHWWHU 4XHVWLRQHQWHUHGE\+DQJ'LQK

Fig. 8: Student survey result from the Discrete Structures course in Fall 2016












 







 






$IWHUOHDUQLQJWKHFULWLFDOWKLQNLQJPRGXOHLWKHOSVPHXQGHUVWDQGPDWKHPDWLFDOSURRIV EHWWHU 4XHVWLRQHQWHUHGE\+DQJ'LQK









 


 


 


 


 


 


Fig. 5: Histograms for the scores in the Logic test group Fig. 9: Student survey result from the Discrete Structures course in Fall 2017 










  
 







  
 


























Fig. 6: Histograms for the scores in the Critical Thinking test group

3.2 Student Survey To learn about students’ perception regarding the effect of critical thinking on understanding mathematical proofs, we added the following question in the course evaluation at the end of semester for each of our Discrete Structure classes: “After learning the critical thinking module, it helps me

Both classes had similar total number of responses, with 17 responses in Fall 2016 and 16 responses in Fall 2017. Nearly half (47% to be exact) of the responses in Fall 2016 and exactly half of the responses in Fall 2017 are either agree or strongly agree with the statement in the given question. If we assign a score between 0 and 4 to each of the five level of responses, with 0 = Strongly Disagree, 2 = Neutral, and 4 = Strongly Agree, then the mean scores of the Fall 2016 class and the Fall 2017 class are 2.18 and 2.56, respectively (see Table 8). In other words, the average responses from both classes tend to be neutral.

4. Conclusions and Discussion Before starting this study, we expected that teaching the Critical Thinking Module before mathematical proofs would help improve the students’ performance on mathematical

ISBN: 1-60132-498-7, CSREA Press ©

46

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 8: Summary of Student Responses from Figure 8 and Figure 9

Fall 2016 Fall 2017

Strongly Agree (4) 2 4

Agree Neutral (3) (2) 6 4

4 5

Disagree Strongly (1) Disagree (0) 3 2 3 0

Mean 2.18 2.56

proofs in our Discrete Structures course. However, the data we have obtained and analyzed from two editions of our Discrete Structures course do not support this expectation. Instead, our data show that students in the Discrete Structures class where mathematical proofs were taught before the critical thinking module scored higher on both understanding proofs and critical thinking tests than those in the other class where critical thinking was taught before mathematical proofs. However, our data don’t have enough evidence to conclude that the difference between the average scores on understanding proofs of the two classes is statistically significant. A reason for these results may be because learning mathematical proofs help students learn critical thinking better, but learning critical thinking does not necessarily help students understand mathematical proofs better. A limit of our study, however, is having small class sizes, which make us unable to test if there is statistical difference between population means on Logic and Critical Thinking scores. To overcome this limit, we may want to replicate this assessment study with future editions of our Discrete Structures course that are taught by the same instructor, who is also the author of this paper, as the editions used in this study.

[7] K. Weber, “Student’s difficulties with proof,” Mathematical Association of America, Research Sampler 8, June 2003. [Online]. Available: https://www.maa.org/programs/faculty-and-departments/ curriculum-department-guidelines-recommendations/ teaching-and-learning/research-sampler-8-students-difficulties-with-proof [8] A. M. Recio and J. D. Godino, “Institutional and personal meanings of proof,” Educational Studies in Mathematics, vol. 48, no. 1, pp. 83–99, 2001. [9] D. Gries and F. B. Schneider, “A new approach to teaching discrete mathematics,” June 2001. [Online]. Available: https://www.cs.cornell. edu/fbs/publications/GSprimus.ps

References [1] A. of American Colleges and Universities, The LEAP Vision for Learning: Outcomes, Practices, Impact, and EmployersÕ Views. Washington, DC: AACU, 2011. [Online]. Available: https://leap.aacu. org/toolkit/wp-content/uploads/2010/12/LEAP-Vision_Summary.pdf [2] J. Association for Computing Machinery (ACM) and IEEE Computer Society, Computer Science Curricula 2013: Curriculum Guidelines for Undergraduate Degree Programs in Computer Science. New York, NY, USA: ACM, 2013. [3] L. M. Pereira and L. Krippahl, “On teaching critical thinking to engineering students,” in The 13th International Conference on Thinking Norrköping; Sweden June 17-21; 2007, ser. Linköping Electronic Conference Proceedings. Norrköping, Sweden: Linköping University Electronic Press; Linköpings universitet, 2007, pp. 153– 158. [Online]. Available: http://www.ep.liu.se/ecp/article.asp?issue= 021%26article=021%26volume=1# [4] B. Fagin, J. Harper, L. Baird, S. Hadfield, and R. Sward, “Critical thinking and computer science: Implicit and explicit connections,” J. Comput. Sci. Coll., vol. 21, no. 4, pp. 171–177, Apr. 2006. [Online]. Available: http://dl.acm.org/citation.cfm?id=1127389.1127423 [5] M. R. K. Krishna Rao, “Infusing critical thinking skills into content of ai course,” SIGCSE Bull., vol. 37, no. 3, pp. 173–177, June 2005. [Online]. Available: http://doi.acm.org/10.1145/1151954.1067494 [6] M. R. Scheessele, H. Dinh, and M. Ananth, “On adding a critical thinking module to a discrete structures course,” Journal of Computing Sciences in Colleges, vol. 30, no. 6, pp. 97–103, June 2015. [Online]. Available: http://dl.acm.org/citation.cfm?id=2753024.2753045

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

47

Living with Digital Education: The Impact and Power (or Otherwise) of Information and Communication Technology (ICT) and Internet in the life of Norfolk State University Community – An Exploratory Study. Samuel BO Olatunbosun Computer Science Norfolk State University Norfolk, VA, USA

Abstract

-

Information and Communication Technologies (ICT) of various kinds have been used in many sectors of life for quite a while. The education sector, in particular, has grown in leaps and bounds in the array of technology tools being used to support students' learning. ICT has transformed teaching, research, and learning at all levels empowering educators and learners and thus making amazing contributions to the development of education worldwide. The combination of ICT and the Internet has ensured that students can spend less time visiting the library and can be anywhere learning and researching with their portable ICT devices. Students anywhere in the world using Internet power can enroll in an asynchronous online class and complete their degree program in less than four years without stepping into the four-walls of a traditional classroom. This study investigated the significance of ICT usage by staff and faculty members at Norfolk State University, (NSU).

Keywords – ICT, ICT and Education, Internet Use

I.

Introduction

"Information and Communication Technologies (ICT) is a term which refers to technologies that are used for collecting, storing, editing and passing on information in various forms. A personal computer, PC is an example of a commonly used type of ICT device in education. In the educational sector, ICT has transformed teaching, research and learning at all levels empowering educators and learners and thus making astonishing contribution to the development of education worldwide (Brush, Glazewski and Hew, 2008). Inextricably linked to ICT is the Internet; another technology that is so pervasive and has remained one of the most important sources through which anyone with a network-connected computer can easily access information, remain in touch with friends, colleagues, and interact with family members irrespective of where they are located on planet Earth (InfoDev, 2005). The combination of ICT and the Internet have made it possible to complete everyday tasks without leaving our homes; from shopping to reading, to

Victoria T. Olatunbosun Biological Sciences Norfolk State University Norfolk, VA, USA watching movies, playing music, and a countless number of activities one can imagine (Castro Sánchez and Alemán, 2011). The Internet reach and coverage of pretty much anything under the Sun is so amazing that even visiting the library or making early morning rounds to newsstands are not just fast becoming old-fashioned, but also a thing of the past (Jacob & Isaac (2008). Students worldwide can enroll in online classes and complete a degree program in four years or less without stepping into the four-walls of a traditional classroom. Electronic textbooks (e-books) are increasingly being substituted for the paper-based versions. Since e-books are now readily available online, students can gain the same knowledge through electronic means with animated materials that are packaged with their e-textbooks, but at a much cheaper cost (Brush et. al., 2008). Additionally, the list of the so-called ‘social media' outlets have been growing that are created to cater for the needs of a variety of people based on their demographics, religious beliefs, or other personal interest. Not surprising; all of these have been made possible because of Internet presence and the array of affordable ICT tools people have access to. Studies have provided useful data on the overall picture of ICT use and Internet benefits in general. Through personal observation, our campus, NSU has not been doing badly in the area of ICT and Internet adoption and utilization for the benefit of our predominantly African-American students. The evidence is abundant to see everywhere on our campus. The multi-million-dollar facility recently added to the Computer Science department located at the McDemmond Center for Applied Research, (MCAR) building for studying cybersecurity programs is a testament to the University's commitment to investing huge resources into modernization and truly making NSU a center of excellence! The life of Faculty and staff of NSU have also revolved around the use of Internet-based technologies in one form or the other. These have all been possible with the advent of the Internet.

Study Goal

This exploratory study sought to analyze trends in ICT and Internet comfort and assess the satisfaction level

ISBN: 1-60132-498-7, CSREA Press ©

48

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

among NSU faculty and staff and to some degree, help to understand more succinctly the culture of ICT and Internet use at the University.

x x x x

Research Questions:

How skillful and comfortable are FS at NSU using the various ICT tools on campus? What are some of the positive and negative impacts of ICT/Internet Technology use among the FS groups? In what ways has the deployment of ICT enhanced or facilitated the educational objectives of NSU students and FS over the years? What future prospects lie ahead with ICT tools and deployment on campus?

II – Literature Benefits of Using ICT in Education

In related studies, several authors have described the many benefits of ICT and Internet use in education. In the work of Jo Shan Fu of the National Institute of Education, Singapore as published in the International Journal of Education and Development using Information and Communication Technology 2013, the author outlined some of the following statements as some of the key benefits of ICT in education.

A. “Assist users in accessing digital information efficiently and effectively As Brush, Glazewski and Hew (2008) have stated, ICT is used as a tool for users to discover learning topics, solve problems, and for educators to provide solutions to the problems in the learning process. ICT makes knowledge acquisition more accessible, and concepts in learning areas are understood while engaging all users in the application of ICT” (Jo Shan Fu, 2013).

B. “Support student-centered and self-directed learning

ICT develop learners’ new understanding in their areas of learning (Chai, Koh and Tsai 2010). ICT provides more creative solutions to different types of learning inquiries. For example, in a reading class, e-books are commonly used in reading aloud activities. Learners can access all types of texts from beginning to advanced levels with ease through computers, laptops, personal digital assistants (PDAs), or iPads. More specifically, these e-books may come with some reading applications, which offer a reading-aloud interface, relevant vocabulary-building activities, games related to reading skills and vocabulary acquisition, and more. Therefore, ICT involves purpose designed applications that provide innovative ways to meet a variety of learning needs” (Jo Shan Fu, 2013).

.

D. “Promote collaborative learning in a distance-learning environment

Koc (2005) mentioned that using ICT enables users to communicate, share, and work collaboratively anywhere, any time. For instance, a teleconferencing classroom could invite learners around the world to gather together simultaneously for a topic discussion. They may have the opportunity to analyze problems and explore ideas as well as to develop concepts. They may further evaluate ICT learning solutions. Learners not only acquire knowledge together, but also share diverse learning experiences from one another in order to express themselves and reflect on their learning” (Jo Shan Fu, 2013).

E. “Offer more opportunities to develop critical (higher-order) thinking skills Based on a constructive learning approach, ICT helps learners focus on higher-level concepts rather than less meaningful tasks (Levin and Wadmany, 2006). McMahon’s study (2009) showed that there were statistically significant correlations between studying with ICT and the acquisition of critical thinking skills. A longer exposure in the ICT environment can foster learners’ higher critical thinking skills. Thus, schools are strongly advised to integrate technology across all of the learning areas and among all learning levels. Where this is done, students are able to apply technology to the attainment of higher levels of cognition within specific learning contexts” (Jo Shan Fu, 2013).

Users are now more frequently engaged in the meaningful use of computers (Castro Sánchez and Alemán 2011). They build new knowledge through accessing, selecting, organizing, and interpreting information and data. Based on learning through ICT, users are more capable of using III - Methods information and data from various sources, and critically A. assessing the quality of the learning materials” (Jo Shan Fu, B. A. Participants and Procedures 2013). Data was collected through the use of an online survey service portal, Survey Monkey at surveymonkey.com via an C. “Produce a creative learning environment anonymous online questionnaire that was created and developed. The sample questionnaire was developed to target mainly NSU faculty and staff (FS) including administrators.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 2: Which of the following Information and Communication (ICT) technology devices do you use on the NSU campus or at other locations? (Check all that apply).

200.00% 100.00% Oth…

Pers…

Chr…

Sma…

Responses

iPad…

0.00%

Des…

The sample attracted 70% females and 29% males FS respondents whose ages range from 25 to 75 years. The mean age was 20.33, and 33% of them were between the 55 to 64 age bracket. 74% of the respondents have a graduate level degree and with 25% earning an income between $50,000 and $75,000 per annum depending on their employment status and classification. We also assumed all faculty and staff had access to at least, a campus or home computer with Internet access. We also assumed some possess other portable devices such as a tablet that could be used to participate in the survey questions. Any NSU faculty or staff meeting the preceding criteria was deemed qualified to participate in the survey. The survey was conducted from October 01 to October 20, 2018.

49

Table 3: How frequently do you interact with or use the following ICT devices?

200.00%

IV. Result Samples

100.00% Des… iPad… Pres… Non… Blac…

0.00%

Does not use or Not Applicable

Table 4: For each source of help listed below, rate how likely you are to use it when you experience computing technology problems (for example, issues with software, network, desktop computer).

Weighted Average

Figu…

Call…

Ask…

A…

0

Goo…

5

My…

Approximately 500 of NSU regular faculty and staff were targeted for this survey. A set of twenty-two (22) survey questions were developed as the research instrument for the purpose of the study focusing on 1) the ICT skills of the FS, 2) their perception and representation of the impact of ICT at NSU, 3) their perception of risks and challenges typically encountered by them using ICT and the Internet, 4) other vital data considered pertinent to the study. These questions were designed to uncover the typical ICT usage, the patterns, and to help understand the culture, competency levels, and how ICT devices have benefitted the community at large. The questions were also designed to identify some of the key weaknesses of the existing infrastructure and how future improvements can be made. Of the 22 items that composed the questionnaire, only one was an open-ended question, while the remaining were closed-ended questions made up of a single answer, multiple choice, or Likert scales types. The open-ended question became necessary to enable the participant to self-identify the department at NSU where they belong.

ICT…

C. Instrument

Table 5: Indicate how strongly you agree or disagree with each of the following statements regarding ICT use.

100.00%

Totally disagree

50.00%

Unsure or No strong opinion

I can…

I…

Using…

0.00%

Disagree

I enjoy…

The following are some of the feedback from the twentytwo questions examined through the survey data and presented as follows just to demonstrate a few of the responses received from the study:

Agree

Table1: What is your age? Table 6: Indicate how often you use the Internet for the following purposes.

40.00% 20.00% Responses

4

ISBN: 1-60132-498-7, CSREA Press ©

Play…

Dow…

Wat…

Buy…

Ad…

0

Res…

2 Soci…

0.00%

Weighted Average

50

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

50.00%

Does not use or Not Applicable

be in abundant supply at NSU apparently as part of growing investment in Technology on campus. Email communication and text messaging are the most common task or routine undertaken on campus. Internet search engine – Google is the most popular search tool. ALL participants also claim to use one or more ICT devices when on or off campus.

0.00%

Rarely used

Utility Software - The most popular software used by FS is

Table 7: How frequently do you interact with or use the following ICT devices?

150.00%

Desktop… iPad/Oth… Presentat… Non NSU… Blackboar…

100.00%

WORD PROCESSING. Since Microsoft Office is usually pre-installed on NSU computers before distribution to users, we presume MS WORD is the most popular version of word processing software in use. Presentation software (presumably MS PowerPoint) is the second most popular application software in use. Timekeeping also appears to be important to FS. From the survey, a substantial number (60%) also responded that they make use of one-time keeping (scheduling software) or the other regularly to keep track of meetings and other tasks for the day, week, or months ahead.

Table 8: Please rate your skill level doing the following:

3.2 3.1 3 2.9 2.8 2.7 2.6 2.5 2.4 2.3

Weighted Average

V. Discussion Demographics - Respondents comprised of 53% of faculty

members, administrative staff members 25%, other staff members 20%, while an additional 1.6% identifying as not FS. In addition, 70% were females while 29% self-identified as males. Respondents came from various offices at NSU with evidence of good representation from several departments including the presidency and provost's office. Participants' age range is between 25 and just above 75 but the predominant age range is between 55 and 64 years of age. The dominant race was Black or African Americans but this is expected since NSU is an HBCU classified institution. The other races including Caucasian Whites and Asians also actively participated. The median income range of the participants was between $50K - $75K per annum. As expected others make more or less depending on their job classification.

Technology use - Desktop PC is the most commonly used tool on campus, Laptop computers followed closely while Tablet computers are also used very often especially by those FS who are constantly on the move. These devices appear to

ICT Use and Comfortability - FS appears to be ICT savvy in all respects. In addition, when there is a technology problem or related issue, many are comfortable trying to solve the problem on their own. If it cannot be solved, they also have good knowledge of where or how to get help, from using NSU ICT helpdesk to asking a friend or colleague, or carrying out a Google search on how to fix the problem. If every attempt to fix the problem or contacting their unit/department's IT desk. In another related question asked in the survey, FS are comfortable and keen to learn new application software especially if it is key to getting their job done in the most efficient manner, act as a guide for others when researching on a subject/topic, happy integrating technology continuously into their daily lives, and using technology to analyze issues and make informed decisions. When it comes to FS experience with interacting with their common everyday ICT devices, respondents do not feel intimidated or threatened by their devices and will readily accept to use any ICT device available. They are confident using the devices on campus, at home or elsewhere there is the need to use one. FS also mostly agree that the culture of ‘know-how' should continue to be extended to students and that ICT tools are capable of being used to learn and develop a whole array of other educationally related works of life. ICT/Internet Common Usage at Home and Elsewhere -FS also use their ICT devices at about the same frequency

both on campus and home. Their devices appear to be used pretty much anywhere there is Internet access especially with folks with portable devices. This observation suggests people are constantly using their devices round the clock except perhaps when they are sleeping or engaging in other activity where the use of ICT is restricted. When prompted to address for what purpose they use their ICT tools for and how often, FS overwhelmingly acknowledged administrative, faculty

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

and staff tasks is the number one purpose and followed closely by research work especially by faculty members. Other popular purpose includes social networking, buying/selling products, playing computer games, and watching movies in their break times.

Perception of ICT/Internet Usage and Impacts on Students With many negative reports in the press associated with social networking, we were curious to know what FS think about students' interaction with the various ones in cyberspace. The findings show many FS are still in favor of continued interaction and perhaps perceived as good educational tools for NSU students. Popular social media favored include YouTube, LinkedIn, Facebook, Twitter, Instagram, Snapchat and WhatsApp. From the study also, faculty appears to have the culture of using the Internet to help develop lesson plans/ideas, to prepare tests for students, manage student performance and grades, and also for personal research work. They also typically utilize smart boards, digital video cameras, projectors, and several other assistive technologies both in their classrooms and offices. FS is also very comfortable using the provided learning management software, Blackboard for managing everyday affairs and interaction with students. When asked for FS perception on the impacts of ICT tools on our student body, FS perception differs on the many subjects raised. There seems to be more agreement on the positive impacts of ICT on NSU students than the negatives. For example, FS believe ICT enables students to collaborate more with others for learning and development. They believe that inappropriate websites create safety risks for students who visit them but students are nevertheless more academically motivated when they are in possession of ICT devices. On the negative impacts, FS believe that the abundance of unreliable sources on the Internet diminishes students critical thinking skills, and that with ICT, plagiarism and other unethical behaviors among students are major problems on the campus. They are also of the opinion that too many inappropriate websites are frequently visited by students.

Perception of ICT/Internet Future in Education -

Considering FS perception of ICT present and future usage on NSU campus, FS agrees that Technology has provided the means for useful collaboration with others especially when building school units or departmental plans, that Technology will continue to improve the ability of faculty to teach more effectively, and that Technology has significantly impacted the way that faculty teach and/or interact with students. In addition, FS are split or unsure whether Electronic media will replace printed text materials within the next few years even though there is ample evidence of increasing use of electronic textbook materials on campus. Finally, FS believe that existing ICT on campus are reliable and helping to facilitate meeting their daily job goals, targets, and expectation.

51

ICT Improvement, Challenges, and Efficiency - On NSU ICT practice and possible improvement, FS think more technical support could be provided to keep the campus computers and applications running well, that more options are needed for professional development in the areas of technology use, and that more access to technology tools are needed to integrate them into the classroom for teaching and instruction. Other areas needing improvement include more training in the use of technology in general before being implemented as well as having faster access to the Internet and wireless campus. Most FS are dissatisfied with accessing NSU Internet resources when off-campus. This appears to be the area of most concern. On FS perception of what typically make them work most efficiently on campus, FS voiced adequate training and prior instruction in the use and function of information technology as the number one reason that brought about efficiency. Technology support via NSU help desk, or interdepartmental help was considered to be the second most important reason that make FS efficient, while Training/instruction in the use of technology as it relates to teaching pedagogy (for example, instruction on using Blackboard) and Technology infrastructure (network, central system) are equally considered to be the third most important reason that aid their efficiency. On FS opinion on the state of NSU ICT infrastructure and other perceived impacts on our campus community, FS believe that the ICT infrastructure benefits and impact the lives of the student population in many positive ways than many of the reported negatives in the media, that the ICT infrastructure has created a much better learning environment and has enabled good connection with the students. However, a sizeable number of FS did not give an opinion or disagree on the question of whether NSU ICT infrastructure state is very strong and adequate for the learning environment, or comparably better than many other institutions across the country, or that the campus ICT infrastructures are located at the right places making them easily accessible to make a difference in the learning of the students.

VI. Conclusion This exploratory study sought to determine the impact and power (or otherwise) of Information and Communication Technology (ICT) and Internet use in the life of NSU Community. It is obvious that as we progress far into this century, ICT and the Internet will continue to play a huge role in how our educational community operates and how the infrastructural systems as provided will continue to impact the typical NSU student's learning. In particular, this study discovered that the provision of ICT infrastructure at NSU has impacted the educational practices of the University in quite a number of positive ways and the relevance of ICT use will continue to grow considerably in years to come. ICT and the Internet infrastructure at the University will continue to

ISBN: 1-60132-498-7, CSREA Press ©

52

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

become a strong agent for change among the many educational uses on campus. Extrapolating current activities and practices, the continued use and development of ICT with faster Internet access at NSU will continue to have a strong impact on 1) what the students learn, 2) how they learn, 3) when and where their learning takes place, and 4) who is learning and who is teaching a subject matter. All respondents in this survey provided answers to the entire survey questionnaire. The findings focused on the following areas: ICT competence level, expertise, and training, the frequency of ICT use and access to ICT resources (remotely and on-campus), ICT use for teaching, administration, professional development, and personal use; and the influence of ICT on students' learning. The results from this study have helped to understand the perceptions of NSU FS on their interaction and practices with ICT devices installed on campus or used elsewhere. We were able to determine that FS have no resistance to using ICT devices in any way or form and that the community seems fairly comfortable with the campus technology infrastructure as currently deployed. Going forward, ICT on the NSU campus is here to stay, and people will continue to yearn for newer and latest technologies to be deployed. Because access to digital tools, applications, and networks will continue to grow nationwide; and various educational media increasingly available in digital forms, ICT use in education can be expected to grow exponentially not just at NSU, but other comparable institutions across America.

g.

h.

i.

j.

k.

l.

Jo Shan Fu, ICT in Education: A Critical Literature Review and Its Implications. National Institute of Education, Singapore. International Journal of Education and Development using Information and Communication Technology (IJEDICT), 2013, Vol. 9, Issue 1, pp. 112125. King, R. (2012). Mobile devices have positive impact on education, survey says. Retrieved from http://www.zdnet.com/blog/btl/mobile-deviceshave-positive-impact-on-education-survey-says/68028 Koc, M. 2005., Implications of learning theories for effective technology integration and preservice teacher training: A critical literature review, Journal of Turkish Science Education, vol. 2, pp.2-18. Levin, T. and Wadmany, R., 2006. Teachers’ beliefs and practices in technology-based classrooms: A developmental view, Journal of Research on Technology in Education, vol. 39, pp.417-441. McMahon, G., 2009. Critical thinking and ICT integration in a Western Australian secondary school. Educational Technology and Society, vol. 12, pp.269– 281. Sarwar, M., & Soomro, T.R. (2013, March). Impact of smartphones on society. European Journal of Scientific Research, 98 (2), 216-226. Retrieved March 18, 2014, from http://www.europeanjournalofscientificresearch.co m/

VII. References a.

b.

c.

d. e. f.

Brush, T., Glazewski, K. D. and Hew, K. F., 2008. Development of an instrument to measure preservice teachers’ technology skills, technology beliefs, and technology barriers. Computers in the Schools, vol. 25, pp.112-125. Castro Sánchez, J. J. and Alemán, E. C., 2011. Teachers’ opinion survey on the use of ICT tools to support attendance-based teaching. Journal Computers and Education, vol. 56, pp.911-915. Chai, C. S., Koh, J. H. L. and Tsai, C.-C., 2010. Facilitating preservice teachers’ development of technological, pedagogical, and content knowledge (TPACK). Educational Technology and Society, vol. 13, pp.63-73. Font, S. (2013). How smartphones narrow the achievement gap in education. Retrieved 23 March 2014, from http://mobileworldcapital.com/en/article/78 InfoDev (2005), “Knowledge Maps: ICTs in Education”, November, www.infoDev.org. Jacob, S.M. and Isaac, B. (2008). The mobile devices and its mobile learning usage analysis. Proceedings of the International Multi-conference of Engineers and Computer Scientists, Hong Kong, Vol. 1, March, 19-21, pp. 782-87.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

53

Data Manipulation and Visualization (DMV): A Case Study Mudasser F. Wyne, Anshu Chaudhary, Dhwani Doshi, and Manjusha Gusain School of Engineering and Computing, National University, San Diego, USA Abstract - Academic Program Directors (APDs) at various academic institutions are required to manage a specific program with the understanding that APD will not only maintain currency of the courses in the program but will also provide academic advising and manage student enrollment in the program. However, because of larger enrollments and frequent offering of the program, at times, it becomes very difficult to complete and manage all of the responsibilities. The advances in technology offer the efficient use of large amount of data for more objective and better decisions. In this paper, we investigate the design for information visualization through an interactive application using Tableau. The use of Data Manipulation and Visualization (DMV) helps APD to keep track of student progress and their enrollment history through user-friendly reporting and data visualization. The paper also presents additional tools for analyzing data and visualizing interaction histories that supports data analysis and communication of findings. Keywords: Tableau, Data Manipulation, Data Visualization.

1

Introduction

Higher education sector has increasingly begun to pay more attention to academic leadership, since it plays important role in offering good quality and sustainable programs. Among academic leadership one group who has a significant role to play in offering high quality programs is the Academic Program Director (APD). Academic Program Directors in their role are responsible for coordinating and managing degree courses in the program [1]. The challenge for the APDs is that, in addition to having academic credibility they must lead and manage the faculty teaching in the program, in most of the cases, without having any management authority. Program management can be difficult and stressful for everyone involved, but a successful APD can ultimately help the institution thrive by not only managing good quality programs but also students in the programs [2]. Academic Program Directors (APDs) at various academic institutions are asked to manage a specific program with the understanding that APD will not only maintain currency of the courses but will also provide the academic advising to the students in the program. In general, APD will ensure that curriculum in the program is current, program is offered at regular intervals and that the students are taking courses in a sequence following a predetermined prerequisite structure. However, because of large enrollments and frequent offerings of the program at

times it becomes very difficult to fulfill and manage all assumed responsibilities. In recent years, use of management systems have become an important tool for an effective leadership and data-based decision making [3] that could certainly help APDs. At times, academic institution uses different systems to manage student and program related data, leading to various sources of data that make managing analysis and reporting requirements across various systems very challenging. APDs would also need data management and visualization tool to further improve decision making in all aspects of program management. Our objective for the case study is to design a system that eases this task by allowing anyone to perform sophisticated education analytics and share their findings. In order to make a quick and learned decision the use of Tableau is explored in this case study. Tableau is a data analytics and visualization tool used widely in the industry today, many businesses even consider it indispensable for data-sciencerelated work. The use of Tableau does not require any prior programming experience and its ease of use comes from the fact that it has a drag and drop interface and being userfriendly it is suitable for any kind of user, ranging from data scientist to manager. Tableau is also compatible with multiple data sources, including Excel, SQL Server, and cloud-based data repositories which makes it an excellent choice for Data Scientists [4]. Tableau visualizations are interactive and shareable, helping user to get answers promptly. It also allows Excel users to keep their spreadsheets while greatly enhancing their ability to analyze their data while delivering simple to build, simple to read visualizations that convey information clearly. In addition, Tableau as stated before, also has a very intuitive graphical user interface with drag-and-drop option which makes it possible to create charts, tables and other visualizations with just a few mouse clicks. Moreover, Tableau requires no additional packages for mapping and statistics as these features are built in the software.

2

Existing Setup

The existing system available for APD support although is computer based but requires lot of manual work in order to retrieve any useful information that can later be used for making any program related decision. Following is the list of some of the shortcomings of the existing system; x APDs having difficulty in getting status of program pre requisite courses.

ISBN: 1-60132-498-7, CSREA Press ©

54

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

x Hard to maintain admission and enrollment data of students, because of multiple offerings of the program per year. x Difficulty in managing compiled information for students in the program. x Endless delays in making an educated decision that is based on complete analysis of the situation. x Absence of result-based monitoring tools to track student progress within a program Existing system and the process used by APDs has few loop holes that will result in students to slip through the cracks and take courses without completing prerequisites, among various other issues. APD may get contacted in this regard towards the end of program when students apply for graduation, thus making the situation very complicated for APD. The program used for this case study is a graduate program, that is offered three times a year and each offering is called a String. In some cases, students are registered and start the program in one String and somehow, student moves from one string to another string and take courses according to their own convenience, rather than following recommended sequence. The program also has three program prerequisite courses for students who do not meet admission criteria. In addition, the way the existing system is set up, APD needs to log in to different university portal to look at an individual student’s enrollment summary to determine if he/she has taken the required program and/or course prerequisites. In the case a student skipped prerequisite APD must first search for the students’ admission advisor and then contact him/her to get the details as to why student skipped prerequisites course(s). There are many possible reasons why a student may not be required to register in the program perquisite courses, such as, student satisfies admission requirements so do not have to take prerequisites and can therefore directly enroll in the core classes. It is also possible that sometimes a student transfer from other institute and hence receives equivalent transfer credits for prerequisites. On the other hand, student may have skipped prerequisite classes because of taking and successfully passing challenge exam for these courses. Therefore, in order to ensure all the students who, need program prerequisites and other students are following the courses sequence, the APD from time to time must complete the following timeconsuming tasks: x Check student enrollment history for all students registered in the first core course in the program to confirm that program prerequisite requirements have been satisfied. x Check student enrollment history for all students in the program to ensure students are following the required prerequisite. x At times APD would also like to ensure that all scheduled classes are staffed. For this APD will need to check the status of each class in the program one by one.

3

DMV System Design

Trying to use spreadsheets for advanced, responsive analytics or analyzing large volumes of data, is using the wrong tool for the job. We also understand that in such situations too often, mistakes are made at the expense of efficiency and accuracy in addition to hours of lost time. In response to address above stated concerns the design of DMV system is based on a reporting application that will assist APD or any other user in decision-making process by visualizing data and/or generating reports. However, for creating reports there are guidelines that need to be followed, since generating a unique report may be complicated so it is a necessity that all the relevant requirements are fully understood [5]. We have developed a prototype, close to be the final release product, as an initial model with an objective to evaluate our design [6]. The proposed DMV system consists of an interactive user interface that functionally allows APD to generate various interactive reports in order to keep track of student progress, student grades, and their enrollment history in different strings of the program. In addition, DMV will also list name of the faculty member and the name of the course that he/she is assigned to teach as well as list of courses that are not yet staffed with a faculty. Tableau filter utility can be used for generating reports by changing the contents of the data that may enter a Tableau workbook, dashboard, or view. Since Tableau has multiple filter types so each type can be used for a different purpose. It is important to know that one can change these filters and the order in which each of these filters is executed. The reports also provide interactive features like highlights to differentiate among specific scenarios. Following is a list of reports that APD can generate using DMV system. Enrollment Status Report: A list of all students in the program can be seen in this report. A click on the student name reveals enrollment and grade history of that particular student. The screenshot in figure #1 shows all the students and classes which includes prerequisite as well as core classes. The highlighted box indicates prerequisites courses details, the term in which the class was taken. When you want any specific student’s detail then click on his/her ID and it will reveal student progress report, described below. Student Progress Report: This report shows student progress in the program where APD can analyzes whole enrollment history of a specific student with just one click. This report shows all the courses that a particular student has taken, term for each class and the grades for each course. The grades for each course are shown as color codes. The name of the faculty assigned for each is also shown however, for confidentiality reasons we are showing Null for each faculty name. Figure #2 shows progress report for a student ID 579.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

55

Core Classes Status Report: This report lists core classes offered in all terms, as shown in figure #3. A click on any core class will reveal a list of registered students. In addition, it will also provide information about the faculty assigned to this class. Bottom of the screen you can find pre-requisite worksheet button. Core classes status report in fact, shows all the classes in one report for all the students.

Figure #1: Enrollment Status Report with Term

Figure #3: Core Class Status Report

Figure #2: Student Progress Report with Grades Strings Status Report: In this report APD can identify one of the strings for report generation using various options (Core classes, or students or prerequisite classes).

Figure #4: Course selection from Core Class Report

ISBN: 1-60132-498-7, CSREA Press ©

56

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Moreover, from core classes status report we can also analyze student grades and terms of the classes from just one screen. ID in this report represents all student Ids, term represent the term (year and month) when student took the core class. Description represents the class name. A click on the term for example 1802, will reveal detailed information such as class name, student ID, and official grade for that student in this class.

according to our needs. For instance, figure #6 shows information for four students for analysis such as ID 57, 63, 149 and 530.

Different filters can also be used to get specific information for a class. APD may want details about one class only for example, Modern Operating Systems which was offered in January 2018. In this case through the use of filters these details can be obtained. Just clicking on the term 1801 and one can get detailed information such as class name, student ID, and official grade for that student, shown in figure #4. Program Pre-Requisite Status Report: This report lists pre-requisite classes, identifying students who are registered in these classes as well those who are not registered, as shown in figure #5. The program pre-requisite status report is where one can visualize all pre-requisite class in one screen. ID represents student Ids, Term represent the term (year and month) when student took the prerequisite class. Description represents the class name. Distinctive shading boxes indicate students took the classes or not. Scrolling the courser on the box, one can get detailed information such as course name, student ID and official grade for the student if he/she took the class.

Figure #5: Program Prerequisite Classes Status Report with Student Grades. When APD wants to analyze multiple reports together, dashboards are the solution in Tableau. Here, we have created this dashboard to analyze prerequisite and core classes reports together. Moreover, filters are available to select different data

Figure #6: Program Prerequisite and Core Classes Report From multiple report dashboard this dashboard one can analyze multiple data at the same time. For example as shown figure #6, student ID 57 did not take the first and third prerequisite classes named Introduction to Programming Concepts and Programming in Java and started taking core classes. In another example studentID 579 did not take any prerequisite and directly registered core classes, although there are many reasons for student not taking prerequisite classes. Some time students have bachelor’s in computer science so, they do not have to take prerequisites. Hence, they directly enroll in core classes, some time students transfer from other schools, and have already passed prerequisites. On the other hand, a student may pass the challenge exam for the prerequisite class(es) and thus do not need the prerequisite classes so can directly register in the core classes. Grade Report: This report lists students who have received a particular grade for example A or A-, as shown in figure #7. The report shows student ID, Class Name, term for the class and faculty name teaching the class, for confidentiality reasons we are showing Null for each faculty name. The report can also be generated for all students and their respective grades as shown in figure #8. In order to analyze multiple reports together, dashboard is the solution provided in Tableau. For example, this dashboard can show all the details about progress report of multiple students at the same time. Similarly, we can analyze performance of all the students in a specific term as well as name of faculty assigned to each course. In addition, for any subject performances of all students can also be analyzed.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

57

Term Report: This report lists students in all classes for a specific term. The report shown in figure #9 is for November 2017 term, has student ID, the class name and the term. The grades for classes are shown as color codes for the said term, if the class is not yet completed then no grades are shown for that class.

Figure #9: Term Report

4 Figure #7: Grade Report for a “A” Grade

Benefits of DMV System

Our proposed system can easily be adopted for any academic institution to visualize data and create various reports at the same time. DMV system will work for any data source, APD is however required to provide details about the program including prerequisite structure. This system has all the potential to be the academic program director’s most reliable asset. Some of the major benefits for APDs using DMV system are listed below: x It will reduce the report generation time so that APD can make decisions in a timely manner. x Get current status of all the students in the program by tracking student progress in real time. x Track student case exception for students who have not followed the proposed course and program prerequisite structure.

5

Figure #8: Grade Report for all students

DMV System Users

Analytic users are not all the same; in most organizations, there are a number of users at different levels with varying distinct needs. User persona for DMV system will be any user who wants to analyze and visualize data in addition to generating interactive reports. The primary user and

ISBN: 1-60132-498-7, CSREA Press ©

58

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

authorized user may be able to add and modify data and information. Following are possible DMV System users; x Academic Program Directors (APD): One of the main users that holds a position that is responsible for managing all aspects of a program. In practice, APD tends to work with highly diverse problems and cannot always predict the nature of the problems that will need addressing. Therefore, APD places a premium on being able to use a wide variety of analytic methods, techniques and tools.

7

Based on the test results from the use of the prototype additional functionality can be incorporated in the system. Followings are the add-ons for the future version of the DMV; x Current system is only accessible on desktops and laptops. To make this system handy to analyze the report anywhere any time thus mobile and tablets access is planned for the future design. x Current system works with a single data source. However, in case a user wants to connect system to different data sources and generate reports the system needs to be more flexible. This aspect is also planned for the future design.

x Program Faculty: Similar in many respects to the Academic Program Director. Share interest in “Easy to Use” tools, and a desire to engage at a granular level with the data. Faculty believe that knowing their students well is fundamental to effective instruction. x Students: Can utilize dashboards to keep track of their progress in the program as well as future class enrollment data.

6

8 [1]

Test and Conclusion

The goal of testing the prototype was to engage a sufficient number of potential end users with varying backgrounds and skill levels. These users helped to determine the level of ease of navigation and the overall effectiveness of the user interface designed for DMV system. Instructions for each task to be performed were given with minimal details to encourage the participants to use their own judgement to complete each task. The overall usability of the completed prototype as determined by the analysis of the usability test results, was measured at a very high level. The ease of use, clarity of steps as well as ease of locating the specified filters and analyzing proper data was reported by 85% of the users. On the other hand, 80% of the users rated overall system navigation as either “easy” or “very easy”. The test users were also kind enough to provide constructive feedback regarding areas of further improvement to further enhance the ease of use. Some of these suggestions were implemented in the revised prototype version. However, other recommendations required extensive changes and hence were postponed for future implementation.

Future Add-Ons

[2] [3]

[4]

[5]

[6]

References Ladyshewsky, R. and Flavell, H., “Transfer of Training in an Academic Leadership Development Program for Program Coordinators”, Educational Management Administration & Leadership, 40(1), pp. 127–147, 2011. Hope, J., “Address Challenges of Managing Enrollment during a Program Prioritization”, Enrollment Management Report, 20(12), pp. 6-7, February 2017. Balu, I. and Presser, O., “e-Leadership of school principals: Increasing school effectiveness by a school data management system”, 44(6), British Journal of Educational Technology, pp. 1000-1011, 2013. Backaitis V. “An introduction to tableau: what it is and how it can provide insight for your business”, Retrieved from CMSwire. https://www.cmswire.com/analytics/an-introduction-totableau-what-it-is-and-how-it-can-provide-insight-foryour-business/. 2018 Shneiderman, B., and Plaisant, C., “Designing the user interface: Strategies for effective human-computer interaction”, 5th, Edition, Boston, MA: Addison Wesley, 2010. Blackwell, A., and Manar, E. Prototype, “UXL Encyclopedia of Science”, 3rd Edition, U X L, 2015.

This case study investigated program and student management by Academic Program Director (APD) through a Data Management Visualization (DMV) System. The results showed that successful use of the DMV consequently improved management of student enrollment and their progress in the program. This is realized through data-driven decision making; monitoring student registration, student performance and student activity in the program.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

59

Preliminary Review - Universities’ Open Source Academic Integrity Policies in the UAE Zeenath Reza Khan, Halim Khelalfa, Jawahitha Sarabdeen, Priyanka Harish and Sanjana Raheja University of Wollongong in Dubai, Dubai, UAE Abstract - In this paper, we conducted investigation of the state of academic integrity policies in universities accredited by the UAE Ministry of Higher Education. For practical reasons, we focused on universities whose websites provided an open access to their policies of academic integrity. The accessible policies were collected, then assessed based on eight identified categories mapped to the five core elements as defined by the Academic Integrity Standards Project conducted in Australia. This project found that of the seven universities whose policies were available as open access online from sample population, highlighted disparities including unclear purpose of the policy, missing complete definitions of academic integrity, misconducts, reporting systems and so on. These findings lay the foundation for a more comprehensive review and analysis of all the universities’ policies, not only those with open-access policies to assist stakeholders in developing shared understanding of academic integrity across all institutions in the country. Keywords: Policy, Academic Integrity, Review, Baseline Standards, UAE “Regular Research Paper”

1.

Introduction

Academic Integrity is defined as “honesty, trust, fairness, respect, responsibility, and courage” no matter what the situation may be in an academic setting [1]. Academic misconduct is any behavior, intentional or unintentional, minor or major, by students that “do not meet [a university’s] expectations of academic integrity in the work they prepare or submit for assessment tasks, including exams” [2]. It has been a common area of concern among academics and researchers over the last few decades [3] [4]. Studies have presented consistent recorded cases of various types of cheating ranging from 65% to as high as 80% self-reported instances on campuses [5] [6] [7] Over the past decades, there have been discussions on instances of student cheating, collusion, plagiarism, data fabrication, contract cheating and a variety of other behavior considered as academic misconduct [8] [9]. Given the gravity and frequency of cheating and e-cheating, some studies have suggested a variety of strategies that some academic institutions apply to address academic dishonesty [10]. However, there is a growing need to adopt a holistic approach to address the issue of academic dishonesty among students in universities [8]. Tennant and Duggan produced research in 2007 towards benchmarking misconduct among students [11]. East highlighted the need to align academic integrity policies,

practices, teaching and learning in higher educational institutions [12]. Grigg studied the inconsistencies in academic integrity policies for plagiarism in 2010 [13]. In particular, the Australian Learning and Teaching Council project titled Academic Integrity standards: Aligning policy and practice in Australian universities [14] that analyzed policies and practices across 39 Australian universities over two years is of significance to this study.

2.

Importance of Policies

Various concepts are defined here to avoid any ambiguity. They concern the difference that exist between a policy, a procedure, a standard, and a guideline. Whittman and Mattord [15] illustrated the schematic definition of these concepts by first establishing the success of top-down approach by organizations that include planning from top management who support and drive the production, implementation and acceptance of policies, procedures and processes: Policies are mandatory and often elaborated and supported by executive management and sometimes inspired by existing practices. They provide general directions and rules that mandate what is supposed to be acceptable behavior. Policies feed into standards. Standards are also mandatory but are more detailed than policies; they provide more specific directions for the compliance with the relevant policy and are often inspired by industry and government benchmarks. Standards feed into guidelines next. Guidelines are not mandatory and are usually used before a particular set of standards get formally adopted within a given sector and are generally inspired by regulatory sector exemplars. Finally guidelines feed into procedures. Procedures on the other hand, provide step-by-step guidance on how to apply standards. Although Whittman and Mattord [15] book focuses on information security, they spend considerable time explaining the importance of policy and the difference between policy and other terminologies which may help us here. As they have explained, policies help to dictate acceptable and unacceptable behavior within an organization. Universities use academic integrity policies to set the tone as well as emphasize the importance of the academic values of integrity, honesty and ethics. Policies are the cornerstone of any course of actions. A major goal of policies is to be complete and comprehensive. Yet, very often, this goal is not achieved because more than one policy is addressing the matter and these policies are much too confusing, convoluted and complicated for application and enforcement.

ISBN: 1-60132-498-7, CSREA Press ©

60

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Universities emphasize to their student and staff bodies the value of academic integrity through their definitions, roles and actions against academic dishonesty that they lay out in their policies. The policies dictate the practices and ultimately ‘affect the way it is taught and embedded in curricul[a]’ [16]. But inconsistencies can raise serious question on quality of degrees conferred. Tennant, Rowell and Duggan [11] found in an extensive study of United Kingdom universities that the range of penalties as described in policy documents implemented by higher education institutions varied substantially and raised concern over consistency and management of academic misconduct. Similar studies conducted in Australia showed also that policies defining plagiarism among universities in that country were varied and not consistent [13]. Research has shown that it is imperative that policies and practices are inter-twined with teaching and learning practices. In the UK the study of Academic Misconduct Benchmarking by Independent Adjudicator for Higher Education highlighted the need for consensus on policies and procedures [11] [17]. In Australia, the need to establish consistency in academic integrity among universities came in 2010 when AUQA produced a report to that effect [18]. A twoyear project was then initiated by the Australian Government Office for Learning and Teaching to review policies and produce good practices [16]. Quoting importance of consistency in application of penalties, reforms, practices in application of academic integrity information and testing, research highlighted the need for such consistency to exist in any given educational sector [18].

3. Relevance to Education Sector in the UAE The United Arab Emirates (UAE) is a developed nation in the Middle East, a metropolitan country housing about 200 nationalities, ranking number one among all Arab countries and top 10 in the world for most competitive economies according to the IMD World Competitiveness Report 2017 [19]. In accordance with the UAE Vision 2021, the country has become an attractive education hub for students from around the world in recent years with over 25 international branch campuses from USA, UK, Australia, Canada, India, Iran, to name a few, and hundreds of other institutions housing hundreds of students from over 80 nationalities [20]. As a young sector, the education sector in the UAE is yet to have established sector standards and best practices that can be found across campuses in the UAE for academic integrity. Some universities have academic integrity policies, some don’t. Those that do have policies vary in their definitions, penalties and actions. Furthermore, although different universities may have differing curricula, and teaching and learning style, it is believed that the core principles of academic integrity apply to all students because these principles are multi-dimensional in nature. According to the Academic Integrity Standards Project: Aligning Policy and Practice in Australian Universities 2010-

2012, a project funded by the Asia Pacific Forum on Educational Integrity, “…it is critical that [academic integrity] is dealt with consistently by staff and taught to students [because] it is fundamental to all assessment practices” [21].

4.

Research Objective

To create a foundation for producing sector good practices, establishing exemplars and developing teaching and learning resources that can be used by government, nongovernment institutions, staff and students alike, this project aims to review existing open-access academic integrity policies for sample population of universities in the UAE to lay the foundation for a more comprehensive review and analysis for better shared understanding. The rest of the paper is organized in the following manner: the next section lays out the methodology that has been followed to conduct this pilot study. After the methodology, the next section presents the results and discussion of the study along with limitations and ultimately the conclusion.

5.

Methodology

To review and analyze the policies, we first identified accredited universities within the UAE that are recognized by the Federal Government and local Emirates. This list is readily available on the Commission for Academic Accreditation (CAA) web portal [22]. As a pilot study, we adopted random sampling to select the universities from the complete list. Of the total 73 accredited universities, 13 universities were identified as sample population for the pilot study by the research team, spread across the seven Emirates (or states) of the UAE. Strict manner of anonymity of universities was maintained through coding system that was put in place by research assistants who signed confidentiality to maintain anonymity of universities they coded so the principal investigators would not be aware of the names. Of the 13 universities, the universities with open access policies were then targeted as part of the pilot study. This process was expected to be cumbersome, time consuming and requiring assistance based on literature review of similar study [16] particularly as universities are currently assumed to not use, document or record same/similar terminologies and policies under similar headings. Bretag et al. (2011) found in the course of their project that in fact universities often had academic integrity policies spread across other policy documents and not always updated [16]. Our research team decided to follow the [16] methodology of searching for and identifying only the main policy document. We did this by using the following search words: ‘academic integrity’, ‘academic misconduct’, ‘plagiarism’ or ‘student cheating’. We did not, at this preliminary stage, follow any further documents that may be related to the policy or have links within the policy, unlike what [16] did. The rationale for the team was that with almost double the number of accredited universities, we realized we would begin small to make a case for a more in-

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

depth case at a later stage. This rationale also played a role in [16] not following Grigg [13] style of comprehensive search where that study included all related documents across policy databases for each university. Once identified, the policies were divided among researchers and research assistants to code the policies for analysis. Researchers referred to [16] and own experiences to code the policies based on categories identified such as: x Access level including number of clicks to policy x Name of Policy/Identity x Purpose of policy* x Definition of AI* x Scope of Policy* x Definition of Misconduct * x Reporting System* x Record Keeping* It is important to mention here that although the [16] study identified and used 22 categories, we decided to use only eight of those at this stage. The rationale for this was that this study, unlike that one, was only a preliminary study and aimed to identify and establish existence of marked inconsistencies in policies among universities to pave way for a more comprehensive study -which we believe these eight categories will be able to highlight, if there exist any inconsistencies. The baseline standards used were developed for this study using the Exemplary Elements of Policy Activity as proposed by Academic Integrity Standards Project. This activity defined and used five core elements that were identified by [16]. Our team discussed and decided that based on the five core elements mentioned, that is, Access, Approach, Responsibility, Detail and Support, the baseline standards were developed for use after mapping our eight categories to the five elements as illustrated in Table 1. The team used a scale of 0, 1, or 2 to categorize the eight identified criteria based on the level of standards met as discussed by the team. The categorizing as illustrated in the table above allowed the team to review the identified policies to explore and note shared or inconsistent understanding of

61

policies and lay foundation for a more comprehensive review of closed policies in the future. The core elements as descried by [16] allowed for a broad explanation of what was expected for our categories. The task we had was to place the categories into appropriate boxes and found some of them overlapped. For our category Scope of Policy, we found it suited best across both Approach and Responsibility exemplary elements. We defined the scores such that if any policy did not have a scope at all, it would score ‘0’, while just summarizing the policy would award the policy a ‘1’ and if the policy scope summarized the policy while highlighting who was affected and how, the policy would earn a ‘2’ score. So, the team decided to colour-code the scale in order to allow for faster and convenient reference model when analysing the policies as illustrated in the Table 1.

6.

Results and Discussion

This preliminary study found only seven out of 13 universities that had their academic integrity (AI) policy on open access. This was surprising to the team as our background study showed most universities including likes of Harvard in USA, Oxford in UK, and those in Australia had their academic integrity policies and procedures available on open access. This practice helps to create a learning environment that is inclusive [23] and transparent. However, we believe due to the fierce competition existing among the universities in the UAE for a small pool of potential students, policies may be protected and therefore beyond open access. We hope to look into this practice in detail in a future-study to understand if this is the rationale behind not having the policies on open access. Applying the baseline scoring model, as shown in Table 2 below, we found that out of the seven, only two were easily accessible by one click and so scored a ‘2’, and the highest number of clicks to the policy was counted as four, scoring ‘0’. Universities with open access policies probably either did not test accessibility of the policy when developing the site or preferred to keep it as obscure as possible by ensuring users had to really hunt for it.

ISBN: 1-60132-498-7, CSREA Press ©

62

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 1: Mapping categories to Five Exemplary Elements and developing Baseline Standards

Exemplary Core Elements [16]

Our Chosen Category

0 (Not Meet Standard at all)

1 (Just Meet Standard)

2 (Fully Meets Standard)

0 = difficult, with clicks more than 2

1 = not so difficult, with 2 clicks

2 = not difficult at all with only 1 click

0 = not available or not direct

1 = available but obscure

2 = available and includes all or relevant part of ‘Academic Integrity’ terminology

Definition of AI

0 = not available or incomprehensible as per ICAI definition

1 = available but missing completeness as per ICAI

2 = available and captures complete essence of ICAI definition

Purpose of Policy

0 = not available or obscure

1 = statement of purpose available but not educative/available through out

2 = statement of purpose exists, educative, and available through out

Scope

0 = not available

1 = available and summarizes policy

Access level

The policy is easy to locate, read, concise, comprehensible (Access)

There is statement of purpose in the policy with educative focus up front and all through policy (Approach)

Baseline Standard Scaling (Score 0 to 2)

Name/Identity

2 = available and summarizes policy plus highlights who is affected, and how

Scope The policy details responsibilities for all stakeholders (Responsibility)

The policy has an extensive but not excessive description of breaches, outcomes and processes (Detail)

The policy includes proactive and embedded systems to enable implementation of the policy (Support)

Reporting/ Recording

Definition Academic Misconduct

2 = available and clear stakeholder responsibilities including reporting/recording and clear processes

0 = not available

of 0 = not available

1 = available but not extensive (or) available and excessive

Reporting/ Recording

Reporting/ Recording

1 = available and includes implementation including reporting/recording

ISBN: 1-60132-498-7, CSREA Press ©

2 = available and not excessive, more educative

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 2: Baseline Scoring Relative Frequency for Identified Categories Score Scale (0-2) per criteria Score Score Score Criteria 0 1 2 Access level 57.14 14.29 28.57 % % % Name/Identity of Policy 42.86 28.57 28.57 % % % Purpose of policy (PP) 57.14 42.86 0.00% % % Definition of Academic 71.43 28.57 0.00% Integrity (DA) % % Scope of Policy(SP) 57.14 28.57 14.29 % % % Definition of Misconduct 42.86 42.86 14.29 (DM) % % % Reporting System (RS) 42.86 42.86 14.29 % % % Record Keeping (RK) 71.43 14.29 14.29 % % % For category Name of the policy/identification of the policies from among the seven open-access policies, two identified the policy as ‘academic integrity’ or ‘academic misconduct’ or ‘academic honour code’, earning a score of ‘2’ and two were considered obscure, not referring directly to academic integrity at all earning a score of ‘1’ and three scored a ‘zero’ for either not having a definition or not referring to it as ‘Academic Integrity’ or part-thereof. This could indicate a vast gap in understanding of concepts relating to the field of academic integrity among the universities or even how universities value or see academic integrity. For categories Purpose of Policy and Definition of Academic Integrity, none of the seven policies managed to score a ‘2’, and in fact five of the policies had no Definition of Academic Integrity. Similar result can be seen for category Record Keeping with five policies that did not have any details of record keeping in their policy document. Our team converted Table 2 into a radar-graph (see Figure 1) in order to visualize the results. As can be seen in Figure 1, most of the open-access policies reviewed either did not meet the baseline standards (green) or just met them (red). For instance, for the criteria ‘Purpose of Policy’, four out of the seven universities did not meet the baseline standard, while three just met with a scoring of ‘1’. This means that majority of the policies reviewed did not have a Purpose of Policy available in their policy document. Purpose of Policy is a section of a policy document that determines the tone/ideology of the policy, that is, whether it is educative or punitive in nature. If the policy does not include this section, it is missing out on a significant opportunity to strengthen the policy. One university’s policy scored a ‘2’ for the Scope of Policy, two just met the standards while four did not meet the standards. Similarly, one university met the standard on Definition of Misconduct, three just met and three did not meet the

63

standards at all. For Reporting System category, we found one policy that met the standards, three just met and three did not meet at all and finally for Record Keeping category, we found one met, one just met and five did not meet at all. Bretag, et al. (2011) have stated that the definitions and explanations of academic integrity in university policies affect the way the universities teach about academic integrity. Based on the results above, the review highlights serious and significant inconsistencies in all identified categories against which the policies where reviewed. From definitions of academic integrity to the scope of project and other criteria, clearly different universities have different approaches to setting up policies on academic integrity. While having differing approaches is expected, the possibility of different understanding of the concepts behind the policies maybe of concern for the education sector as a whole in the UAE. For each of the eight categories, three or more policies identified scored ‘0’ in our baseline standards scale. This is very significant to this study because we believe these results set the precedence for the next, comprehensive review process where we aim to use all 22 categories from Bretag et al. (2011), include a further scoring level of ‘3’ to identify parts of extended population of policies that may be exemplar by exceeding the baseline standards we have set.

Legend: Green - Score 0 | Red - Score 1 | Yellow - Score 2 Figure 1: Radar Graph illustrating Percentage of Scores for Identified Categories

7.

Conclusion

We conducted a pilot study on seven of the 13 sample universities which were chosen using random sampling out of 73 accredited universities in the United Arab Emirates. The seven universities were found to have open access policies that

ISBN: 1-60132-498-7, CSREA Press ©

64

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

could be found online through web searches. We found there exists differences in policies of different universities based on eight criteria identified and considered against baseline standards. As the intent of the project is not to ‘name and shame’ nor to recommend some kind of standardization, the findings highlight the need for a more comprehensive review of existing policies across universities within the UAE, to understand how different universities understand, educate and respond to issues of academic dishonesty in order to develop exemplar good practices that can lead to shared understanding of academic integrity. It further paves way for a future study by acquiring ethics clearance and approaching universities to gain approval to access the closed-access policies in order to make the review more indepth and more comprehensive for decision makers and stakeholders of higher education.

References

Australia,” 2010. [Online]. Available: http://repository.unimelb.edu.au/10187/8971 . [Accessed 18 Aug 2014]. [14] T. Bretag, R. Walker, M. Green, M. Wallace, J. East , C. James, U. McGowan and L. Partridge, “Academic integrity standards: Aligning policy and practice,” Successful proposal to the Australian Learning and, Australia, 2010. [15] M. Whitman and H. Mattord, Principles of Information Security, Cengage Learning, 2017. [16] T. Bretag, S. Mahmud, M. Wallace, R. Walker, C. James, M. Green , J. East, U. McGowan and L. Partridge, “Core elements of exemplary academic integrity policy in Australian higher education,” International Journal for Educational Integrity, vol. 7, no. 2, pp. 312, 2011. [17] P. Tennant and F. Duggan, “Academic Misconduct Benchmarking Research Project: Part 2.,” The Higher Education Academy and JISC, UK, 2008. [18] Australian Universities Quality Agency (AUQA), “Audits: Universities,” 2010. [Online]. Available: http://teqsa.gov.au/auditreports. [Accessed 10 09 2016]. [19] UAE Government, “Statistics,” 2018. [Online]. Available: https://government.ae/en#/.

[1] Centre for Academic Integrity, “The Fundamental Values of Academic Integrity pp5-9,” 2012. [Online]. Available: www.academicintegrity.org/icai/assets/FVProject.pdf. [Accessed 15 09 2016]. [2] University of Wollongong, “Academic Integrity Policy,” 13 March 2018. [Online]. Available: https://www.uow.edu.au/about/policy/UOW058648.html. [3] S. Davis, C. Grover, A. Becker and L. McGregor, “Academic Dishonesty; Prevalence, Determinants, Techniques, and Punishments,” Teaching of Psychology, vol. 19, pp. 16-20, 1992. [4] K. Blankenship and B. Whitley, “Relation of general deviance to academic dishonesty,” Ethics & Behaviour, vol. 10, no. 1, pp. 1-12, 2000. [5] W. J. Bowers, Student dishonesty and its control in college, New York: Bureau of Applied Social Research, Columbia University, 1964. [6] J. Sheard and M. Dick, “Computing student practices of cheating and plagiarism: a decade of change,” 2011. [7] Z. r. Khan, “Developing a factor-model to understand the impact of factors on higher education students’ likelihood to e-cheat,” UOW, Sydney, 2014. [8] R. MacDonald and J. Carroll, “Plagiarism - a complex issue requirin a holistic institutional approach,” Assessment and Evaluation in Higher Education, vol. 31, no. 2, pp. 233-245, 2006. [9] W. Sutherland-Smith, “Retribution, deterrence and reform: the dilemmas of plagiarism management in universities,” Journal of HIgher Education Policy and Management, vol. 32, no. 1, pp. 5-16, 2010. [10] J. Goosney and D. Duda, “Avoiding the plagiarism pitfall: preventing plagiarism in undergraduate research,” 2009. [Online]. Available: http://ojs.acadiau.ca/index.php/AAU/article/viewFile/74/41. [Accessed 24 July 2012].

[20] M. Swan, “UAE among emerging education hubs,” 2014. [Online]. Available: http://www.thenational.ae/uae/education/uae-amongemerging-education-hubs. [Accessed 17 09 2016]. [21] Asia Pacific Forum on Educational Integrity, “What is educational integrity?,” 2010. [Online]. Available: http://www.apfei.edu.au/. [Accessed 15 09 2016]. [22] CAA, “Active Programs,” 2018. [Online]. Available: https://www.caa.ae/caa/desktopmodules/instprograms.aspx. [23] National Federation of the Blind, “Higher Education Accessibility Online Resource Centre,” 22 June 2018. [Online]. Available: https://nfb.org/higher-education-accessibility-online-resource-center. [24] UAE Higher Education Scientific Research, “The UAE Higher Education Factbook,” 2014. [Online]. Available: http://www.mohesr.gov.ae/En/ServicesIndex/Documents/UAEfactbook24Feb-en-CDversion.pdf. [Accessed 18 09 2016]. [25] B. Nelson, “Higher Education at the Crossroads: An Overview Ministerial Discussion Paper,” 2002. [Online]. Available: http://www.voced.edu.au/content/ngv13176. [Accessed 18 09 2016]. [26] S. M. Graves, “Student Cheating Habits: a predictor of workplace deviance.,” Journal of Diversity Management, vol. 3, no. 1, 2008. [27] The Young Vision, “A Straight Talk on Higher Education with Dr Ayoub Kazim,” 2016. [Online]. Available: http://www.theyoungvision.com/a-straight-talk-on-higher-educationwith-dr-ayoub-kazim/. [Accessed 18 09 2016]. [28] The Centre for Academic Integrity, “The Fundamental Values,” 2014. [Online]. Available: http://www.academicintegrity.org/icai/assets/Revised_FV_2014.pdf. [29] Georgia Institute of Technology, “Policy Scope Statement,” 21` June 2018. [Online]. Available: https://policylibrary.gatech.edu/policy_development__policy_scope_statement.

[11] P. Tennant, G. Rowell and F. Duggan, “AMBeR Project. Joint Information Systems Commitee,” 2007. [Online]. Available: https://www.jisc.ac.uk/news/jisc-and-the-higher-education-academyset-up-new-academic-integrity-service-23-jan-2008. [Accessed 16 09 2016]. [12] J. East, “Aligning policy and practice: An Approach to integrating integrity,” Journal fo Academic Language and Learning, vol. 3, no. 1, pp. A38-A51, 2009. [13] G. Grigg, “Plagiarism in higher education: Confronting the policy dilemma, Unpublished PhD. University of Melbourne. Melbourne.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

65

The Development and Deployment of a Mobile Music Application for Literacy Enhancement (M2APPLE) Amal Babangida Sabo, Mathias Fonkam†, Abubakar Sadiq Hussaini†, & Charles Nche† † School of IT & Computing, American University of Nigeria, Yola, Nigeria.

Abstract We present in this paper the design, implementation, and broad scale deployment and use strategy of a Mobile Music Application for Literacy Enhancement (M2APPLE) that leverages the fun activity of Karaoke music as a motivational and complementary tool learning English as a second or even third language. English language illiteracy is a particularly endemic problem for a vast majority of youth in Northern Nigeria who are consequently disadvantaged since English is not just the language of governance but is the lingua franca for commerce and much else across the entire country. M2APPLE leverages the full potential and portability of HTML 5 for a complete client-side application that can runs seamlessly on mobile and non-mobile technologies alike. The explosive growth of the mobile phone industry in Nigeria and Sub-Sahara Africa, coupled with the growing integration of web and mobile content on smart phones presents a very real, viable and affordable opportunity to begin bridging the knowledge barrier as a key foundation to socio-economic development. M2APPLE puts the learner in the driver’s seat and not only allows them to pick and choose songs of their liking to play and learn from but they can also interact with the system as a song is being played; to learn words/phrasal translations and/or to get a better handle on word phonetics. The overall purpose of M2APPLE is to help increase the learner’s vocabulary pool in order to increase their level of confidence. This is part of a wider intervention measure to English illiteracy in Northern Nigeria supported by the Unites States Agency for International Development (USAID) program. Keywords: Karaoke music, illiteracy, hypermedia, transcription

1. Introduction Music has been widely researched and employed as an important tool for, and complement to learning, especially within informal settings. Music is widely used in early learning of phonetics, word recognition, counting, etc.; in general, as a motivator for the learner to improve their level of literacy, build their vocabulary base and gain some level of confidence [8]. Illiteracy has been identified as a major hindrance to the socio-economic development of much of the developing or so “third” world. Such is the case for Nigeria, in spite of speculations and indicators that Nigeria is on a fast trajectory to emergence as an economic power house on the African continent. Illiteracy, and especially youth illiteracy, which is particularly endemic in the densely populated Northern parts of Nigeria, becomes a major impediment to this economic development. It also adds to the security challenges facing this region of the country [12], [2]. The

insurgency in the North-Eastern region of Nigeria, largely perpetuated by Boko Haram, easily breeds on illiterate youth. As a development University, and in recognition of the added danger posed by the Boko Haram insurgency, the American University of Nigeria (AUN) in Yola has mounted many initiatives to help improve the level of literacy in the State of Adamawa and beyond. One such is TELA (Technology Enhanced Learning for All) which leverages such technologies as Radio and mobile tablets to impact on this problem. Our research effort, part of which is reported in this paper, adds a dimension to the TELA initiative and focuses on developing a mobile application, code-named, M2APPLE (Mobile Music Application for Literacy Enhancement) that leverages the fun activity of Karaoke music as a motivator and complementary tool to learning and building the vocabulary of illiterate youth in Northern Nigeria. M2APPLE leverages the full potential and portability of HTML 5 and related technologies for a complete clientside application that can run seamlessly on mobile and non-mobile technologies alike. It is built as a progressive web application [7], that is, as a modern, offline-capable, cross-platform mobile web application. The explosive growth of the mobile phone industry in Nigeria and Sub-Sahara Africa, coupled with the growing integration of web and mobile content on smart phones presents a very real, viable and affordable opportunity for bridging the knowledge barrier that has simply not been there before for the developing world. On the pedagogy front, research has shown time and again that learnerdirected learning is far more effective than instructor-led [20] and developing life-long learners is becoming a common learning outcome for many an educational institution. M2APPLE puts the learner in the driver’s seat and not only allows them to pick and choose songs of their liking to play and learn from but they can also interact with the system as a song is being played to learn words/phrasal translations and/or to get a better handle on word phonetics. The direct relationship between a learner’s vocabulary bank and their level of confidence has been known for some time now [], [15]. The overall purpose of M2APPLE is to help increase the learner’s vocabulary pool in order to increase their level of confidence. The second author of this research has experienced first-hand the motivational force of music in learning a new language and believes he was able to quickly build and grow his vocabulary base in the Portuguese language because of his connection with and influence of Brazilian music he employed to quickly learn the Portuguese language while on his first major post-PhD job teaching at a University in Brazil. He recounts that his love for a song that was being played provided the impetus to invest time in finding out the meaning of the

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

ISBN: 1-60132-498-7, CSREA Press ©

66

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

words behind the music, further expanding his vocabulary pool and with that his confidence in speaking the language. The rest of this paper is organized as follows: in the next section – section 2, we introduce the problem of illiteracy and its specific manifestation and impacts in Nigeria. In part 3, we review the literature on the use of music as a motivator to learning a new language. We look at Karaoke music in particular, and how it has been employed to improve the learning of a new language. In part 4, design and implementation, we first present important design and implementation considerations or objectives for the system and then elaborate on the enabling technologies supporting our M2APPLE system and how these have been strung into a system to meet the objectives. In Part 5, we critically evaluate our implementation and deployment for use, outlining some important contributions and some pointers for future work.

2. Background Since independence in 1960, Nigerian leaders, scholars, and institutions have initiated many efforts aimed at improving the level of literacy of Nigerian youth. Unfortunately, most of these efforts have not had the impact hoped for, especially in the North. Today, youth and adult literacy in the English Language remains a critical determinant for improving the livelihood of individuals, their families and the country at large. According to [13], ‘before we can conquer poverty, ignorance and disease, we must first conquer illiteracy; because illiteracy is the most serious handicap for economic, political, social and individual development’. [18] cites illiteracy as one of the major causes of the Boko Haram insurgency. The United States Institute of Peace listed ignorance of religious teachings, poverty and unemployment and high levels of illiteracy as linked to youth radicalization and extremism. Before the UN’s effort, education experts monitoring the activities of UNESCO, United Nations International Children's Emergency Fund (UNICEF), Action AID, British Council & United States’ Agency for International Development (USAID) had made two sanctions: (1) the need to establish a literate and learning society using innovative approaches (Tahir, 2005), some of which include Each-One-Teach-One, Literacy Shops, distance learning, etc.; (2) the need to expand our current educational system to allow adult men and women access to education. Too many adult Nigerians, especially in the Muslim North, are disadvantaged for religious or financial reasons, and so are deprived of their right to basic education which is achieved mostly through classroom learning at a young age. Illiterate children (whether rich or poor) grow up with low selfesteems; when they go to hospitals or banks or court houses, they cannot communicate with the nurses, doctors, cashiers, etc. so they are forced to find interpreters or to resort to other ways to solve their problems. To reduce this inherent disadvantage associated with high illiteracy rate in the Northern Nigeria especially, it is necessary to develop innovative approaches that are self-applicative and can be used at any time and from anywhere. This research proposes to implement a mobile application code-named M2APPLE: Mobile Music APPlication for Literacy Enhancement, with the broader goal being to expand access to education and command of the dominant English Language by leveraging mobile technology and promoting lifelong learning [13].

2.1

Broad Goals of the Research

As a Development University, the American University of Nigeria (AUN) has led many initiatives to alleviate many social ills in Yola and the wider region of Adamawa State. These include the Adamawa Peace Initiative (API), Student Empowerment Through Language, Literacy and Arithmetic (STELLAR) and STEM Projects. Another initiative also aimed at improving the literacy level across all age groups, especially the young and adolescent, was code-named TELA (Technology-Enhanced Learning for All). TELA is a program that proposes to offer basic reading and math lessons through radio and mobile technology to some 20,000 Nigerian children and adolescents, orphaned, displaced, homeless or at-risk [3]. However, TELA targets mostly children within the ages of 5 to 15, even though it shares the same broad purpose of ‘learning for all’. M2APPLE is an offshoot of TELA that seeks to leverage the fun-activity of Karaoke music on mobile devices to help illiterate youth improve their spoken and written English by building up their vocabulary. The M2APPLE application aims to employ popular, catchy local, national and even international songs and their lyrics to improve their spoken and written vocabulary in the English language. It specifically targets Nigeria Youth, some of whom are already fluent in other languages (Hausa in particular). The widespread availability of smart mobile phones with internet access and sufficient computational power to run the intended application means it should be more widely available to just about anyone with some appreciation of music (teenagers and above), and eliminates having to train anyone on how to use it, since most of them already use smart phones. The mobile domain also provides huge opportunities to the various institutions interested in this problem. Ease of use, portability and scalability are important objectives of the system.

3. On the Use of Karaoke Music in Learning Karaoke is a musical entertainment activity that originated in Japan. Typically, a singer on stage sings along into a microphone the lyrics of a song displayed on the screen, while the instrumentals are played in the background. The background music could also be vocalized depending on the singer’s preferences and confidence level. Karaoke has mostly been seen as a fun activity and stereotypically happens in social settings such as at parties. However, web and mobile technologies have helped redefined what karaoke represents, where it is done and who does it – it can happen anywhere and everywhere today! When used to support learning, it has been seen to be highly motivating for learners of all age groups [8]. It is an important complement to learning a new language, especially for music lovers and for a language and culture of music such as English with some of the best music in the world.

3.1

Music & Literacy

Most countries, especially Nigeria, have come to the realization that literacy must not necessarily only be achieved through formal education but also through distance learning; so the delivery medium to achieve this must be ‘inclusive’ of the many, not ‘exclusive’ of the few [1]. This ‘delivery medium’ must, however, be multi-faceted; and

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

what better medium than the use of Information and Communication Technologies (ICTs). In [9]’s words, “to tech or not tech education is not therefore the question, the real question is how to harvest the power of ICTs to make education relevant, responsive and affective for school settings and lifelong learning for anyone, anywhere, anytime”. However, in [1], it is noted that until such a time as we begin to remember that ICTs are not just limited to computers, we probably would not fully leverage ICTs as a channel to enhance literacy, especially ‘distance learning’. He reminds us that ICTs include radios, mobile phones and tablets, televisions, and prints, all of which are cheap and readily available in our homes, and more importantly have better penetration, culturally and geographically speaking, than our precious computers; although computers bring a certain degree of ‘interaction’. He discusses the three learning revolutions the world has seen so far. First was the discovery of the ‘written language’, the second was the expansion to moveable books and types, and the third ICTs. He calls them ‘revolutions’ and opines that ICTs alone have the ability to prepare us to become life-long learners because of their ever-changing nature. This, he describes as a move towards the ‘constructive learning theory’. The ‘interaction’ mentioned earlier regards [9]’s view that ICT applications have not only succeeded in making learning socially interactive, but also take into consideration the learner’s abilities and needs. For instance, most literacy apps have different ‘levels’ categorized typically as novice, advanced beginner, competent, proficient, and expert following [5]. Because ICTs as literacy tools, especially mobile phones and computers have to some extent penetrated the geographical barrier, [9] has interestingly split them into two. The first category, technologies in location, includes “digital notepads, mobile phones, printed materials, CDs, films & videos, scanners, slides, etc.” The second category, which comprises “correspondence, radio, television, web pages, web internet, webcasts, etc.”, are technologies of distance. They both integrate audios, videos with tools such as emails and chat rooms to promote synchronous and asynchronous interactions amongst learners”. The general use of Karaoke music in teaching and learning as in the works of [10], [16] and [4] fall in the first category. Both authors employed Karaoke music as an entertaining, motivational and complementary tool for learners in the classroom to improve the pronunciation and communication skills of the learners and as a medium that encourages more practice and collaborative learning amongst the students since students can take turns to sing and sing in groups too. Qualitative assessments of past works concluded that karaoke did indeed improve students’ pronunciation challenge. Even better, it boosted their confidence greatly because it was observed that the learners interacted more freely during class exercises; all the while communicating in English and trying to accomplish classroom tasks together. Our system shares the same goal with previous works to advance the communication skills of the learner in the English Language in order to improve the learner’s economic welfare. It targets youth who are already conversant in another language such as Hausa, or Fulfulde or Yoruba, but are practically illiterate in the dominant English Language - the national language of commerce and governance in Nigeria. M2APPLE qualifies as innovative because it meets the two criteria sanctioned by Nigerian

67

education experts [17]: first, it proposes to use technology to enhance learning and secondly, it ‘includes’ every English illiterate Nigerian: not just youth, or children. The latter translates to distance learning. Distance learning is advantageous according to [1] for 2 reasons; it uses ‘multimedia’ to deliver educational services, and it ensures access equity for all, even though M2APPLE mostly targets illiterate youth who can already reason for themselves. M2APPLE can be used locally within a class or at distance by learners doing their own self-applicative learning.

4. Design & M2APPLE

Implementation

of

We have stated already an important design consideration for M2APPLE as a mobile application to support learning for all. To achieve this, two important design considerations are ease-of-use and broad accessibility to the majority of learners and learning environments. We add to this scalability or the ease with which a solution can be made widely accessible to a larger pool of users as the need arises. Implementation as a web application that is portable across device types should guarantee both the ease of use requirement as well as go some way to fulfilling the accessibility need. Guaranteeing portability and deployment as a mobile app will go a long way to meeting the latter goal. A solution and deployment strategy that can be easily replicated should address the goal of scalability. As this system targets youth who can already speak a second language, an important end-user goal is to provide more active user interactivity with the system than is available with the typical Karaoke session where the actor simply reads off a screen. An important design objective is support for an interactive word and phrase dictionary that the user can explore to find out the meanings of words and song lines as they learn English.

ISBN: 1-60132-498-7, CSREA Press ©

68

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Cordova

Mobile app for Android or iOS Fig 1: Architecture View of System Components, their Inputs, Outputs &

file using as input the time-stamped text-file of the lyrics, its translation file, and the name of the audiofile. The output of the integrator is an html file that serves as a launcher for the player or Karaoke system for the given song.

M2APPLE is made up of the following key components, technologies and functions. 1.

2.

3.

A Creative Studio component made up of the Transcribe Software System used for audio-to-text transcription. It takes as input an audio file of a song and its lyrics as text within the text area of its window and spits out a time-stamped text or word file of the lyrics with start and end times for each line, or word of the song. We delimit on lines not words. A good part of this involves some user action in delineating the start and end parts of the lines of a song. A standard Player + template HTML file: the player (jPlayer) is a Javascript audio player deriving from the hyperaudio project [23]. Hyperaudio permits to weave audio into normal HTML enabling for audio similar capabilities to normal text such as search, indexing etc. An Integrator written in Ruby is a software module that serves to build or flesh out the template HTML

4.

Mobile App Generator: this is really an extension of the Creative Studio in 1. aimed at generating a mobile version(s) of the application for the different platforms, principally Android and iOS systems. The main technologies employed for this generation are the open-source Cordova system – an off-shoot of Phonegap from Adobe.

Fig 1 is a sketch of a high-level System Architecture showing how these components fit together, their inputs, outputs and the technologies driving each component. Inputs into each component of the system are shown both to the left and right of the input arrow while the outputs of any module are the ones shown to the left of the output arrow from that module.

5. Assessment of the Implementation & Deployment Strategy

M2APPLE is a cross-platform implementation and adaptation of Karaoke music to support, motivate and

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

advance the learning of English by Youth in Northern Nigeria as a second or even third language. It is principally intended to be run on mobile devices. However, by employing open standard and open-source technologies the application can be run on any platform supporting a web browser. We set out 4 design goals for M2APPLE – easeofuse, accessibility, scalability and enhanced user interactivity. To a large extent our implementation has addressed all four. Ease of use and accessibility are assured by implementing the system as a cross-platform web application that can run seamlessly on mobile devices including mobile phones and personal computers. To support and advance the University’s TELA illiteracy intervention measure and deliver a system that can scale to support an even larger group of learners we not only stuck to open source and open standard web technologies (HTML, CSS3 and Javascript) but we automated the generation of a single launcher HTML file that embeds the text lyrics of a song and its translation dictionary. All other components of the system, namely the Javascript player and CSS files remain the same and can be easily replicated. On the final design goal of interactivity, not only do we provide a fully integrated phrase dictionary but the user can access this while a song is being played by simply hovering the mouse over a line of the song for a display of the translation of that line into the learner’s native language. For our first implementation, Hausa served as the main language we translated into given its very broad adoption in Northern Nigeria, and in fact across much of Sub-Sahara Africa. However, it is an easy matter to translate to other Nigerian languages and employ the same application in other parts of the country. One of our design goals was ease of scalability of an implementation so that it can easily be made available where needed. To meet this goal, we opted for a simple design with just one HTML file carrying the lyrics and its translation. This file can be easily edited by hand to change the translated text to another language. To further facilitate this process, we defined an Integrator module that will take as input the time-stamped transcribed lyrics, its translation into another language and a template HTML file and produce an HTML launcher file for the application. This should greatly ease scaling of the application. We have so far received very positive feedback from key stakeholders on the TELA project, principally the field instructors who employ tablets for teaching. Most of this is anecdotal though. An important remaining part of the assessment will come after actual deployment and use by the Youth themselves.

6. REFERENCES [1]

Aderinoye, R. (2008, November). Literacy and Communication Technologies: Distance Education Strategies for Literacy Delivery. International Review of Education, 54(5/6), 605-626,.

[2]

Akpan, D. A. (2015, April). Youth’s Unemployment and Illiteracy: Impact on National Security, the Nigerian Experience. International Journal of Arts and Humanities , 4(2), 62-71.

[3]

American University of Nigeria, A. (2016). https://www.aun.edu.ng/aclead/portfolio-

69

item/stellarproject/. https://www.aun.edu.ng/.

Retrieved

from

[4]

De Cristofaro, G. (2011, March). Languages Learning through Songs, Karaoke and Internet. International Journal of Communication Research, 1(1), 37-40.

[5]

Dreyfus, S. E. (2004, June). The Five-Stage Model of Adult Skill Acquisition. Bulletin of Science, Technology & Society, 24(3), 177-181.

[6]

Garrett, T. (2008). Student-Centered and TeacherCentered Classroom Management: A Case Study of Three Elementary Teachers. Journal of Classroom Interaction, 43.1(0749-4025), 34 - 47.

[7]

Gatsbyjs.org. (2019, April 8). https://www.gatsbyjs.org/docs/progressive-web-app/. Retrieved April 28 2019, from https://www.gatsbyjs.org/.

[8]

Gupta, A. (2006). Karaoke: A tool for Promoting Reading. The Reading Matrix, 80-89.

[9]

Haddad, W. (2007). ICTs for Education: A Reference Handbook.

[10] Israel, H. F. (2013). Language Learning Enhanced by Music and Song. Literacy Information and Computer Education, 2(1), 1269-1275. [11] Jacquith, D. B., Hathaway, N. E., & Fahey, P. (n.d.). Empowering student-directed learning: developing creative thinking skills through art. [12] Ojong, F., & Ejar, B. F. (2018, August). Mass Literacy as a Tool for Curbing Security Challenges in Nigeria. International Journal of Strategic Research in Education, Technology and Humanities, 5(1), 160171. [13] Omolewa, M. (2006). Educating the "Native": A study of the Education Adaptation Strategy in British Colonial Africa, 1910-1936. African American History, 267-287. [14] Pierson, H. D., & Ekbatani, G. V. (Eds.). (2012). Learner-directed Assessment in ESL. New Yorl: Routledge. [15] Rengifo, A. R. (2009, February). Improving Pronunciations Through the Use of Karaoke in an Adult English Class. Profile 11(1657-0790), 91-105. [16] Sigurðardóttir, V. D. (2011). Language Learning through Music. [17] Tahir, G. (2005). Nigerian National Council for Adult Education in the Current Millenium: A Rear View Mirror. Adult and non Formal Education in Nigeria: Emerging Issues, 6.

ISBN: 1-60132-498-7, CSREA Press ©

70

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

[18] Tella, C. M. (2015). Insecurity in Northern Nigeria: Causes, Consequences and Resolutions. International Journal of Peace and Conflict Studies, 23-36. [19] Turnbull, D., Gupta, C., Murad, D., Barone, M., & Wang, Y. (2017). Using Music Technology to Motivate Foreign Language Learning. International Conference on Orange Technologies, (pp. 218-221). IEEE. [20] Terry Anderson (2008). “Theory & Practice of Online Learning” 2nd Edition, AU Press, Athabasca University, ISBN 978-1-897425-08-4 [21] Bobby Hobgood, “Becoming an Online Teacher” -http://www.learnnc.org/lp/pages/665 [22] Barry Richmond, “Systems Thinking: Critical Thinking Skills for the 1990’s & Beyond”, System Dynamics Review Vol. 9, no. 2, 1993 [23] http://happyworm.com/blog/2013/04/05/thehyperaudio-pad-next-steps-and-media-literacy/

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

71

Collaboration: Key to Student Success in Computing and other STEM Fields in Hispanic Serving Institutions Meline Kevorkian#1, Greg Simco #2 #2

College of Engineering and Computing, Nova Southeastern University 3301 College Avenue, Fort Lauderdale, FL 33314, USA [email protected] #2 College of Engineering andComputing, Nova Southeastern University 3301 College Avenue, Fort Lauderdale, FL 33314, USA [email protected]

The 2019 World Congress in Computer Science, Computer Engineering, and Applied Computing Abstract—As a Hispanic-Serving Institution (HSI) at the forefront of national and regional efforts to increase academic and career success among diverse student populations, Nova Southeastern was selected by the National Science Foundation (NSF) to host one of four inaugural conferences to support HSIs. This paper summarizes the findings from the conference attendees to shed light on the important to collaboration in overcoming the challenges and share best practices facing STEM fields with a special emphasis on computing disciplines. This work is based on on a grant sponsored by the National Science Foundation through NSF Award #1748199, HSI3: Hispanic STEM Ideas, Inspiration, and Innovation. Keywords: Computer Science, Collaboration, and Hispanic Student Success

I.


INTRODUCTION

A recent study by Excelencia in Education noted that HSIs represent only 12% of US colleges and universities, yet enroll 60% of all Latino undergraduates. Although over 80% of HSIs are concentrated in five states and Puerto Rico, the majority of states have at least one Emerging HSI (defined as having 15-24% FTE enrollment). Latino degree attainment nearly doubled between 1995 and 2014, however persistence and completion rates at HSIs still lag behind overall rates (Santiago, et.al. 2016). Between 2000 and 2013, the share of baccalaureate degrees in science and engineering awarded to Hispanics increased from 7% to 11% still far below the 45% of Hispanic freshmen who declare an intent to major in science and engineering, indicating a significant unmet need in retention and completion. In addition, attrition in STEM fields at community colleges was much higher than at four-year institutions (69% and

48%, respectively). A task force formed by the Hispanic Association of Colleges and Universities (HACU) released a preliminary report indicating that the rapid growth of the Hispanic population will play a key role in the economic and social development of the United States for decades to come. Knowledge gained through professional development will lead to increases in the number and diversity of undergraduates recruited and retained in Computer Science and other Science, Technology, Engineering, and Math (STEM) programs. Collaboration among HSIs will expand opportunities for underserved students and broaden the use of our nation’s human resources to produce a stronger and larger STEM workforce. Nova Southeastern University’s conference was called HSI3 because the topics were carefully selected to cover a balance of the latest STEM curriculum content, instructional approaches, and learning support strategies focusing on Hispanic students. The topics were specifically designed to be most relevant and motivating to HSI faculty who would be most likely to develop high quality ideas responsive to NSF’s broader mission and interests. Additionally, participants were encouraged to transform their thoughts into inventive actionable solutions to be submitted to the new NSF HSI grant program in FY2018, as well as to other public and private funding opportunities. The conference had two specific goals: (1) Establish collaborations, resource discovery and sharing, professional development, and expanded participation in diverse thought, experiences, and approaches with respect to underrepresented STEM students, faculty, and research; and (2) Provide a framework for collaborative research, education, and dissemination of effective pedagogical interventions and academic, government and industry opportunities for

ISBN: 1-60132-498-7, CSREA Press ©

72

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

HSIs to prepare for diverse student success in STEM. The conference sessions specifically targeted topics addressing connections between two-year and four-year institutions, industry and education, and faculty and students. Conference sessions covered different aspects of strategic enrollment, retention, graduation, and articulation interventions, focusing on reducing impediments to learning, and building a flexible, supportive, and effective academic pathway, making STEM programs a realistic degree option for more Hispanic students. Keynote presentations from NSF and the founding President of Excelencia in Education, an NSF-led conference session focusing on the new HSI grant program and a mock panel review workshop, and a session to facilitate grant development knowledge fully engaged attendees in support of future institution-level grant work. The conference agenda intentionally included ample opportunities for attendees to network, discuss issues, and ask questions. Specific times in the conference agenda were allocated for Small Group Discussions and a Networking Reception. Table 1 provides attendee demographic information. II.


# Institutions Represented

% Two-Year Institutions

% By Region

Southeast Midwest Southwest Northeast MidAtlantic

47.5% 21.3% 14.8% 11.5% 4.9%

% By State

FL

47.5%

TX

11.5%

CA

8.2%

IL

4.9%

NY

4.9%

CO

3.3%

CT

3.3%

PA

3.3%

AZ

1.6%

DC

1.6%

IN

1.6%

KS

1.6%

MA

1.6%

MO

1.6%

NJ

1.6%

NM

1.6%

DEMOGRAPHICS OF CONFERENCE PARTICIPANTS

Table: 1 Conference Attendance Data Conference Data Collected # Registered Attendees % Female Attendees

# States Represented

Evaluation Notes and Observations 61-22% above the target of 50 63.9% - a high percentage considering that females are underrepresented in STEM 16- the top 9 states with the largest Hispanic populations were represented (2015 Hispanic population rates and percentage growth 2000-2015): 1.
California (15.2%) 2.
Texas (10.7%) 3.
Florida (5.0%) 4.
New York (3.7%) 5.
Illinois (2.2%) 6.
Arizona (2.1%) 7.
New Jersey (1.8%) 8.
Colorado (1.2%) 9.
New Mexico (1.0%)

III.


32 – an average of 2 representatives for each institution in attendancea significant number positively contributing to the diversity of conference conversations and degree of conference impact 49.2% - statistically meets the goal of having an equal balance between two-ear and four-year institutions Based on the conference’s physical location in Florida, conference planners expected a large percentage of attendees to come from institutions in the Southeast, although the conference was marketed nationally. Based on the conference’s physical location in Florida, coupled with Florida being ranked 3rd in Hispanic population, conference planners expected a large percentage of attendees to come from within the state, but were very pleased with the broad geographic diversity of states represented. This wide scope diversified professional networking opportunities and increased attendee awareness of the depth and breadth of interest throughout the nation in strengthening STEM education for Hispanic students.

CONFERENCE PANEL RESULTS

Nova Southeastern University electronically disseminated a conference follow-up survey four work days after the conclusion of the conference. The survey, easily accessible online using SurveyMonkey, was sent to all registered attendees. A total of 32 completed surveys were tallied as of the targeted return date (52.5% of all attendees). Table 2 provides the results from the survey.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Table 2: Survey Results Statements During the conference, there was sufficient opportunity to network with other attendees The presenters were knowledgeable about their topics During the sessions, there was sufficient opportunity for questions and comments from the attendees Overall, this conference met my expectations My knowledge about NSF grants increased My knowledge of/commitment to offer undergraduate research increased Questions from the audience were answered well The conference was appropriate for someone in my position My knowledge of/commitment to the importance of connections between twoyear and four-year institutions increased My knowledge of STEM industry perspective was broadened My knowledge of Hispanic student perspectives was broadened IV.


Strongly Agree/Agree Responses 90.63%

90.63%

73

prepared underrepresented computer scientists (Gates, et. al., 2016). Research supports the need for greater numbers of underrepresented and female Computer Science PhDs (Zweben & Bizot, 2018). Computer science educators need to work together and collaborate to support the needs in the workforce and in education. Collaborating together, especially among HSIs could result in major gains for all students but particularly those that are underserved.

81.25%

80.00% 75.01% 74.20% 74.20% 70.96% 68.75%

59.38% 53.13%

ACTIONS FOR IMPROVEMENT

Attendee feedback revealed five primary suggestions to strengthen learning at HSIs: (1) Identify more ways to let students hear from industry and alumni (especially those who are closer in age to students) who are working and/or are in higher-level programs; (2) Increase student understanding of what math and science have to do with STEM careers (relevance); (3) Strengthen academics expand modes of instruction, active learning, interdisciplinary approach to learning, authentic research experiences incorporated into classes and labs (campus based), more individualized, increase faculty professional development/training; (4) Strengthen support services faculty and peer mentoring, STEM specific advising, connect with professional and student organizations, more follow up with students through their programs and after, help starting college for low income and first generation students; and (5) Increase collaboration, including inter-institutional coordination of higher education resources and programs, and partnerships with K-12. Table 3 presents the results from attendees for future events. This includes more collaborations among HSIs and similar conferences.

Table 3: Recommendations for Future Events 1.
Incorporate time for attendees to dig into/brainstorm/draft grant writing work. 2.
Incorporate opportunities for small group work to come up with novel solutions, discuss specific “how tos” that build on the larger group sessions. 3.
Conduct the simulated NSF grant review in smaller groups. 4.
Present smaller student and industry panels, so discussions can be more in-depth. 5.
Incorporate content specific to math (e.g., undergraduate research in math, math initiatives involving placement testing, co-requisites, studio courses). 6.
Expand presenters to include faculty from more HSIs. 7.
Conduct regional conferences. V.


REFERENCES

1.
Santiago, D.A., Taylor, M., and Calderón Galdeano, E.C. (May 2016). From Capacity to Success: HSIs, Title V, and Latino Students. Washington, DC: Excelencia in Education. 2.
National Science Foundation. Science and Engineering Indicators 2016. 3.
Gates, A.Q., Thiry, H. and Hug, S. (2016). Reflections: The Computing Alliance of HispanicServing Institutions. Retrieved August 23, 2018 from https://dl.acm.org/citation.cfm?id=3010823 4.
Zweben, S. and Bizot, B. (May 2018). 2017 CRA Taulbee Survey: Another Year of Record Undergrad Enrollment; Doctoral Degree Production Steady While Master’s Production Rises Again. Computing Research News, Vol. 30, No. 5. Retrieved March 18, 2019, from https://cra.org/crn/2018/05/2017-crataulbee-survey-another-year-of-record-undergradenrollment-doctoral-degree-production-steady-whilemasters-production-rises-again/

This work is extremely important to the field of computer science as the role of HSIs in increasing PhD-

ISBN: 1-60132-498-7, CSREA Press ©

74

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

SESSION ACCREDITATION, ASSESSMENT METHODS AND STRATEGIES + CURRICULUM DESIGN AND RELATED ISSUES Chair(s) TBA

ISBN: 1-60132-498-7, CSREA Press ©

75

76

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

77

Design and Development of a Modular K12 Cybersecurity Curriculum G. Javidi1, E. Sheybani2, and Z. Pieri3 Technology, University of South Florida, Sarasota, Florida, USA ʹInformation Systems and Decision Science, University of South Florida, Sarasota, Florida, USA ͵Social Sciences, University of South Florida, Sarasota, Florida, USA ͳInformation

Abstract - In this paper the authors describe their innovative methodology to expose high-school teachers to deeper learning for engagement that seeks to connect wireless and internet technology leadership through inspiring inner creativity and critical thinking in cybersecurity. This is to address the need for qualified technology teachers, student access to wireless and web-based cybersecurity experiences, and the role of leadership in technology in response to the evolving digital ecosystem. The idea is to train the teachers to serve as advisors to students by developing strategies for fun-learning activities (e.g. Hackathon, Shark Tank, and Maker-Space) and integrating cybersecurity technology in the curriculum. Using wireless and internet technologies, they learn strategies to encourage students to solve social problems with creative business-based solutions. As a result, students will build their online brand, discover new career directions, and enhance their tech skills. Their network will be augmented by associations with tech and industry leaders. The objectives are to represent K12 students with options in cybersecurity career fields, deliver online resources to distribute curriculum, create tech visionary model for business thinking, and encourage a supplemented network model for cybersecurity learning in K12. Keywords: K12, Curriculum, Cybersecurity, Modular, STEM

1

Introduction

The web-based and wireless cybersecurity skills deficiency is reaching widespread magnitudes. The agreement in the STEM community is that the issue starts at K12 schools with limited students studying STEM topics. In the field of cybersecurity, this issue becomes more trivial since organizations can’t afford to underestimate their exposure to cyber-attacks and the importance of cybersecurity. Increasing cybersecurity professionals is a necessary solution to the cybersecurity skills crisis in the technology industry. Training more K12 teachers will ensure a bigger pipeline into cybersecurity. Trained teachers will not only teach cybersecurity in their schools and assimilate cybersecurity ideas in their classrooms but also promote IT safekeeping as an nice-looking professional path. The gateway to our next, best generation of tech-savvy workforce is training high-quality

teachers since they are the most important fundamentals in influencing and controlling a student's educational track. To prosper in the world of rapidly changing technology and the required technological skills, the proposed curriculum will also incorporate leadership knowledge and soft skills as necessary elements for any STEM worker. Our approach is to develop a curriculum that will combine all those necessary skills to provide a meaningful opportunity to engage students with the real-world cybersecurity issues that industry grapples with. This curriculum will start an effort in the cybertechnology education in K12 and cultivate leadership talents with creative thinking, and practical ability. By introducing important elements of change in K12, there is an opportunity for innovation in computing, which drives economic growth and underlies many scientific advances.

The huge shortage of personnel with sufficient knowledge, skills and abilities in cybersecurity field is already taking a toll locally and globally. Part of this deficiency is due to absence of minorities and women in technical fields (population issue). The other contributor to this problem is the lack of encouragement, interest, and engagement at the K12 for students to find careers in challenging fields such as cybersecurity (pipeline issue). To remedy this, the authors have taken an innovative approach to design and develop a modular K12 cybersecurity curriculum. The innovative techniques and modules used in the proposed curriculum to incorporate cybersecurity concepts include, but are not limited to: 1. Taking a holistic approach to make cybersecurity awareness the minimum requirement for K12 college-bound students, 2. Using a train-the-trainer model to train K12 teachers to integrate cybersecurity concepts in their classrooms, teach cybersecurity, and endorse cybersecurity as an attractive professional path, 3. Introducing creative, modular, and hands-on activities to teach programming, networking, and other cybersecurityrelated topics in a fun and attractive manner, 4. Developing case studies, real-world examples, and projects that connect cybersecurity concepts to terrorism and radicalization, social, political, and financial hacking, military strategic missions, recruitment of extremist and religious groups, etc.,

ISBN: 1-60132-498-7, CSREA Press ©

78

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

5. Adding entrepreneurship and leadership skills to curriculum to prepare students for practical complications. 6. Assessing the dissemination approaches, pacing, timing, format, and outcomes configuration to offer a starting point for future research.

2

Research and Methodology

Design and development of a modular, scalable, and novel cybersecurity curriculum will help train the future workforce. This will be done in coordination, integration, and collaboration with other/existing programs to leverage and/or expand STEM educational research. The design and development of modules in the proposed curriculum was done considering the following aspects:

2.1 Taking a holistic approach to make cybersecurity awareness the minimum requirement for K12 college-bound students Although considerable growth in computing jobs is expected in the next decade, students are not majoring in computing in sufficient numbers to meet this demand. Cybersecurity is a serious economic and national security threat facing our nation today. According to the Bureau of Labor and Statistics, the rate of growth for jobs in cybersecurity is projected at 37% from 2012-2022, higher than the average for all other occupations [1]. Hence, the demand for cybersecurity experts is spiraling. Addressing the achievement gap by leveraging an emerging group of K12 students is a key strategy to consider.

2.2 Using a train-the-trainer model to train K12 teachers to integrate cybersecurity concepts in their classrooms, teach cybersecurity, and endorse cybersecurity as an attractive professional path The proposed curriculum creates a partnership with local K12 teachers to help them build knowledge, skills and confidence in cybersecurity, who in turn, can help engage students in STEM learning and prepare them for success in STEM careers. By helping more teachers become proficient in teaching cybersecurity, at an earlier stage, more young people will become aware of the career opportunities and prepare themselves for technology-related education at the university level, which, ultimately, will help tackle the technology workforce deficit in the United States. Sustaining a “STEMinterested” student body can potentially have substantial longterm implications. In today's competitive work environment, the required technical skills are not enough for future employers. This curriculum will equip students with the requisite combination of technical and soft leadership skills

needed to excel in the field of cybersecurity. For example, teamwork is an important expertise in cybersecurity as organizations often consist of tens or hundreds of information technology or cybersecurity members. Teamwork is required for digital investigation, implementing technical controls, incident handling, enforcing policies, and many other everyday tasks essential for a cybersecurity expert. The curriculum also shadows the rudimentary philosophies of leadership education including encouraging students to challenge the status quo, thinking outside the box, pushing for a deeper understanding of learning, understanding the marketing behind a solution, building strong relationships through networking and collaborating with others. Improved innovation, development and research on leadership education advance the essence of students and are vital to plummeting the number of educated unemployment [2, 3, 4, 5, 6, 7].

2.3 Introducing creative, modular, and hands-on activities to teach programming, networking, and other cybersecurity-related topics in a fun and attractive manner Robotics, video games, music, and movies are some of the fun and attractive ways in which creative, modular, and hands-on activities in programming, networking, and other cybersecurity-related topics can be taught. Computers have become indispensable in robotics, video game, music, and movie making, distribution, performance, and consumption. Students need to learn how to use computing to explore powerful and creative ideas. Introductory computer science courses at each level of K12 is needed to give students a broad perspective of computing and its impact. These courses should also be designed in a way to attract diverse populations including females and underrepresented minorities.

2.4 Developing case studies, real-world examples, and projects that connect cybersecurity concepts to terrorism and radicalization, social, political, and financial hacking, military strategic missions, recruitment of extremist and religious groups, etc. For a curriculum to be current and applicable, it would have to have case studies and real-world examples. Today’s technology world has many cases that can be presented as valid projects for a cybersecurity curriculum. While the rise of social media sites such as Facebook, Twitter, and YouTube have been heralded as innovative platforms that serve to connect individuals, communities, and ideas across the globe, they have also been exploited by violent extremists to radicalize and recruit individuals to their causes [8]. A prime example of this has been Islamic State’s ability to master social media in effective ways to reach vulnerable populations, and despite social media operators' efforts to block terrorists from

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

influencing their networks, each new terror attack spawned more chatter, imagery, and propaganda [9]. Social media allowed Islamic State to amplify their message, giving them access to spaces and audiences without consideration for physical borders. Indeed, recruits to the Islamic State based in Syria and Iraq included people from over 100 countries, and their relative youth was a further signal that the use of social media was important in their radicalization and recruitment [10]. Content has included propaganda videos exalting the virtues of the given group’s narrative, but it also afforded more personalized interaction between extremists and those being targeting through message boards, and the creation of specialized online training sites [11]. Since Facebook has around 1.6 billion user accounts, YouTube around 1 billion active accounts, and Twitter with 336 million regular monthly users, removing such content is a challenge. One aspect that governments and security agencies are investing in is that of countering violent extremism online, and social media companies are being urged to develop and implement automated technologies to identify extremist content that can then be combated. This is something that is already being implemented with some degree of success, for example many ISIS affiliated accounts on Twitter have already been closed down [12]. Automatically being able to identify and delete extremist content from sites is one method of dealing with the situation, though, perhaps, a more effective one is that of “taking on” that content and developing resonant arguments to challenge it. With regards to Islamic State, Islamic scholars can be engaged to develop counter arguments with solid Qur’anic justifications that can then be presented as counter-arguments in online forums and social media sites [13].

2.5 Adding entrepreneurship and leadership skills to curriculum to prepare students for practical complications While K12 students go through numerous courses that teach them basic search engine usage, typing, and computer programing, they are not being taught the soft skills associated with these technologies. While much has been reported on increasing participation in computer science and other STEM fields, little has been reported on the specifics of leadership in cybersecurity. There are also substantial misunderstandings about professions in cybersecurity and the experiences of value to this field [14]. The proposed curriculum will also provide technical as well as non-technical knowledge and skills necessary for cybersecurity, awareness about cybersecurity issues and tools for prevention and various resources to build foundational knowledge and skills for a career in cybersecurity. Teachers are provided with many curriculum materials and programs to help recognize the skills and knowledge students need to become members of the growing cybersecurity community.

79

2.6 Assessing the dissemination approaches, pacing, timing, format, and outcomes configuration to offer a starting point for future research The proposed curriculum has a built-in assessment instrument. The mixed-method evaluation has two main goals: (1) formative evaluation of curriculum planning and implementation; and (2) summative evaluation of learning outcomes for participants. Although the main interest would be in comparing pre-to post-test scores, interim assessments will permit the evaluators to assess whether there are differences between continuing participants and those who drop out. The constructs measured by the assessment survey, with items borrowed from pre-existing, validated instruments whenever possible includes the following: demographics; attitudes toward computer and technology tools; participation in IT-related showcases; knowledge of STEM careers; interest in and understanding of app design; and understanding of educational and social effects in programming and app generation.

3

Results and Discussion

Through focused activities, adopted into an environment where a dearth of Tech-Ed (or STEM Ed) currently exists, there is genuine potential for dramatic results. The proposed curriculum aims at preparing K12 teachers (and as a result K12 students) for the rigorous tech/STEM learning, specifically in the cybersecurity field. It also prepares them to seek leadership skills and apply their knowledge to real-world problems. This project provides insights into training the trainers whose students can (1) innovate as technologists, and (2) act as social change leaders. In an effort to instill an leadership mindset, the curriculum incorporates industry partners so that the teachers will gain a better understanding of the commercial side of innovation and how it links to a targeted market. This opens a door to bringing the power of the latest in education and leadership to the field of cybersecurity. During the course of this study, there is a likelihood of a substantial jump towards a more advanced cybersecurity educational methodology using the proposed model.

4

Conclusion and Future Work

The proposed curriculum focuses on the analysis of current methodologies for cybersecurity education and leadership in K12 and creating a new model. There is demand for curriculum that extends beyond basic cybersecurity and leverages novel learning science and blended learning delivery methods that can reach more people with the same or better level of learning outcomes. Leveraging a standard set of assessments will further ensure quality as this program is

ISBN: 1-60132-498-7, CSREA Press ©

80

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

scaled to other agencies or communities. The goal is to change the way K12 teachers think about technology and leadership by linking them to industry partners, and providing a new model for technical innovation coupled with leadership in cybersecurity. The authors will develop the curriculum tools and learning activity templates that will be tailor-ready for adoption in K12 across the nation, as well as a scalable solution shared with other universities and the public/private educational sectors.

5

References

[1] Bureau of Labor Statistics. (2014, January 8). Occupational Outlook Handbook. Retrieved from Information Security Analysts: http://www.bls.gov/ooh/computer-andinformationtechnology/informationsecurity-analysts.htm

/RR400/RR453/RAND_RR453.pdf, (accessed: 05.11.2018). [9] Berger, J. and Morgan, J. 2015. ‘The ISIS Twitter Census: Defining and describing the population of ISIS supporters on Twitter’, The Brookings Project on U.S. Relations with the Islamic World. March. No. 20. [10] Benmelech, E. & Klor, E. 2016.’What Explains the Flow of Foreign Fighters to ISIS?’ The Kellogg School of Management, Northwestern University, April, https://www.kellogg.northwestern.edu/faculty/benmelech/htm l/BenmelechPapers/ISIS_April_13_2016_Effi_final.pdf (accessed: 05.11.2018). [11] Awan, I. 2017. ‘Cyber-extremism: ISIS and the Power of Social Media’, Society, 54(2): 138-149.

[2] Bell, R., & Bell, H. (2016). Replicating the networking, mentoring and venture creation benefits of entrepreneurship centres on a shoestring: A student-centred approach to entrepreneurship education and venture creation. Industry and Higher Education, 30(5), 334-343.

[12] Alfifi, M., Kaghazgaran, P., Caverlee, J. & Morstatter, F. 2018. ‘Measuring the Impact of ISIS Social Media Strategy’, Texas A&M University, http://snap.stanford.edu/mis2/files/MIS2_paper_23.pdf (accessed: 11.05.2018).

[3] Hynes, B., Kennedy, N., & Pettigrew, J. (2016). The Role of Business Schools in Framing Entrepreneurial Thinking Across Disciplines: The Case of Allied Health Professions. In Innovative Business Education Design for 21st Century Learning, Springer International Publishing, pp. 75-91.

[13] Mandaville, P. & Nozell, M. 2017. ‘Engaging Religion and Religious Actors in Countering Violent Extremism’, Unites States Institute of Peace, Special Report 413.

[4] Mason, C. (2011). Entrepreneurship education and research: emerging trends and concerns. Journal of Global Entrepreneurship, 1(1), 13-25.

[14] Warwick, A. Cybersecurity Skills Shortage need urgent attention, says DoHS: Retrieved on July 23, 2013 from: http://www.computerweekly.com/news/2240178584/RSA2013-Cyber-security-skills-shortage-needs-urgentattentionsays-DoHS, February 26, 2013.

[5] Nielsen, S. L. & Gartner, W. B. (2017). Am I a student and/or entrepreneur? Multiple identities in student entrepreneurship, Education + Training, 59(2), pp.135-154. [6] Nyello, R., Kalufya, N., Rengua, C., Nsolezi, M. J., & Ngirwa, C. (2015). Effect of Entrepreneurship Education on the Entrepreneurial Behaviour: The Case of Graduates in the Higher Learning Institutions in Tanzania. Asian Journal of Business Management, 7(2), 37-42. [7] Zeng, Z., & Honig, B. (2016). How Should Entrepreneurship Be Taught to Students with Diverse Experience? A Set of Conceptual Models of Entrepreneurship Education. In Models of Startup Thinking and Action: Theoretical, Empirical and Pedagogical Approaches, Emerald Group Publishing Limited, pp. 237-282. [8] Behr, I., Reding, A., Edwards, C. & Gribbon, L. 2013. ‘Radicalisation in the Digital Era: The use of the Internet in 15 cases of Terrorism and Extremism’, Rand, https://www.rand.org/content/dam/rand/pubs/research_reports

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

81

6((.,1*$%(7$&&5(',7$7,21$&$6(678'@7KLVLVDVRSSRVHGWRWKH ILYH OLVWHG HDUOLHU LQ WKLV GRFXPHQW )RU FRPSXWHU VFLHQFH SURJUDPV WKHUH ZHUH WZR DGGLWLRQDO 62¶V ZKLOH IRU LQIRUPDWLRQ WHFKQRORJ\ SURJUDPV WKHUH ZHUH ILYH DGGLWLRQDO 7KXVZKHQWKLVHIIRUWEHJDQSULRUWRWKHWRWDOQXPEHU

RIRXWFRPHVWKDWQHHGHGDVVHVVPHQWLQERWKSURJUDPVZDV  'HYHORSPHQWRI3HUIRUPDQFH,QGLFDWRUV  7KH DSSURDFK WDNHQ ZDV WR GHILQH 3HUIRUPDQFH ,QGLFDWRUVIRUHDFKRIWKH$%(76WXGHQW2XWFRPHV,QLWLDOO\ WKHUH ZHUH WZR FULWHULD IRU HDFK RXWFRPH /DWHU WKLV ZDV H[SDQGHGWRDQDYHUDJHRIWKUHH7KHUHGXFWLRQSRVWLQ WKH QXPEHU RI RXWFRPHV ZRXOG KDYH VLPSOLILHG WKLV EXW DV WKH SURFHVV ZDV ZHOO XQGHUZD\ XVLQJ WKH H[SDQGHG OLVW RI 62¶V WKDW H[LVWHG SULRU WKH HIIRUW QHHGHG WR EH UHH[DPLQHG DQGUHWDUJHWHGWRWKHQHZ62¶VGHOD\LQJWKHRYHUDOOSURFHVV  7KHGHYHORSPHQWRISURSHU3HUIRUPDQFH,QGLFDWRUVLVD FULWLFDOHOHPHQWLQDVVHVVPHQWDVWKHHQWLUHSURFHVVKLQJHVRQ PHDVXULQJSHUIRUPDQFHDJDLQVWWKHP$VVXFKLWLVLPSRUWDQW WKH\EHFOHDUDQGZHOOWDUJHWHG2QHKHOSIXOUHVRXUFHLQWKLVLV WKHXVHRI%ORRP¶V7D[RQRP\>@'HWDLOVDUHEH\RQGWKH VFRSHRIWKLVSDSHUEXWJHQHUDOO\WKH7D[RQRP\LVDVHWRI KLHUDUFKLFDO PRGHOV XVHG WR FODVVLI\ HGXFDWLRQDO OHDUQLQJ REMHFWLYHV LQWR WKUHH EURDG GRPDLQV 2QH RI WKHVH LV WKH FRJQLWLYH RU NQRZOHGJHEDVHG GRPDLQ ZKLFK GHILQHV VHYHQ OHYHOV RI REMHFWLYHV GHDOLQJ ZLWK YDULRXV DVSHFWV RI JDLQLQJ DQG XVLQJ NQRZOHGJH $ E\SURGXFW RI WKLV PRGHO LV D OLVW RI YHUEVDQG³DFWLRQZRUGV´WKDWFDQKLJKOLJKWVSHFLILFPHDQLQJV DQG REMHFWLYHV DW WKH YDULRXV OHYHOV 8VLQJ WKH 7D[RQRP\ DQG PRUH VSHFLILFDOO\ LWV OLVW RI ³$FWLRQ 9HUEV´ KHOSV WR FODULI\ FRQFHSWV DQG VHOHFW DSSURSULDWH DQG XQDPELJXRXV ODQJXDJHZLWKZKLFKWRZRUGWKH3,¶V  7KH SURFHVV IRU JHQHUDWLQJ 3,¶V DOVR QHHGV WR LQFOXGH FULWHULDIRUPHHWLQJWKHPDVZHOODVDPHDVXUHPHQWSODQ7KH

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

85

PHDVXUHPHQW SODQ DGRSWHG OHYHUDJHG WKHVDPHLGHDWKDWZDV XVHG LQ FRXUVH DVVHVVPHQW LQ ZKLFK  DFKLHYHPHQW OHYHOV ZHUH GHILQHG IRU WKH VWXGHQWV 7KHVH DUH LGHQWLILHG DV ([HPSODU\ 6DWLVIDFWRU\ 'HYHORSLQJ DQG 8QVDWLVIDFWRU\ 7KHFULWHULDIRUHVWDEOLVKLQJZKLFKRIWKHIRXUOHYHOVDVWXGHQW DFKLHYHG OHG WR WKH GHYHORSPHQW RI DQ HYDOXDWLRQ UXEULF IRU HDFK$JDLQ%ORRP¶V7D[RQRP\LVDXVHIXOWRROIRUJXLGLQJ FUHDWLRQRIWKHUXEULFV  ,WZDVGHFLGHGWKDWSURJUDPDVVHVVPHQWZRXOGEHGRQH XVLQJ VWXGHQW GDWD FROOHFWHG LQ FRXUVHV SHUKDSV D QDWXUDO FKRLFH JLYHQ WKDW FRXUVH DVVHVVPHQW KDG EHHQ GRQH SUHYLRXVO\ 7R IDFLOLWDWH WKLV D FXUULFXOXP PDS RU PDWUL[ ZDVFUHDWHGZKLFKLGHQWLILHGZKLFKFRXUVHVDGGUHVVHGHDFKRI WKHLGHQWLILHG3,¶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

  7KH DVVHVVPHQW SURFHVV IRU D JLYHQ 3, LV IRU WKH LQVWUXFWRU WR FROOHFW WKH VSHFLILHG PDWHULDO DQG XVLQJ WKH UXEULFHYDOXDWHWKHVWXGHQWVDVVLJQLQJHDFKDVFRUHIURPWKH IRXU OHYHO VFDOH GHVFULEHG HDUOLHU 7KLV LQIRUPDWLRQ ZKLFK FRQVWLWXWHV WKH UDZ DVVHVVPHQW GDWD LV VDYHG LQ D UHSRVLWRU\ WRJHWKHU ZLWK WZR UHSUHVHQWDWLYH H[DPSOHV RI HDFK OHYHO LI WKH\ H[LVW $Q H[DPSOH RI D UHSRUW XVLQJ IDEULFDWHG GDWD LV VKRZQLQ7DEOH)RUSULYDF\UHDVRQVQRVWXGHQWQDPHVDUH DWWDFKHGWRHLWKHUWKHVFRUHGDWDRUWKHH[DPSOHZRUNQDPHV DUHUHPRYHGLIWKH\DUHSUHVHQW   (YDOXDWLRQDQG&RQWLQXRXV,PSURYHPHQW  2QFH WKH GDWD DUH FROOHFWHG IRU DOO RI WKH FRXUVHV LQ D JLYHQDVVHVVPHQWF\FOHWKH\FDQEHDQDO\]HG7KLVLVGRQHE\ DJJUHJDWLQJDOORIWKHVWXGHQWGDWDDQGJHQHUDWLQJUHSRUWVIRU HDFK 3, VKRZLQJ WKH SHUFHQWDJH RI VWXGHQWV ZKR UHFHLYHG HDFKVFRUHIRUWKDW3,$QH[DPSOHXVLQJIDEULFDWHGGDWDIRU &6,7FRPSXWHUVFLHQFHVWXGHQWVLVVKRZQLQ7DEOH 

 $VVHVVPHQW&\FOHV  :LWK3,¶VUXEULFVDQGGDWDFROOHFWLRQWDUJHWVLGHQWLILHG WKH DFWXDO DVVHVVPHQW SURFHVV FRXOG EH XQWDNHQ 7KLV EHJDQ E\ LGHQWLI\LQJ D VHW RI 62¶V WR WDUJHW IRU HYDOXDWLRQ RYHU D VSHFLILHGSHULRGRIWLPHRUF\FOH$URWDWLQJVFKHGXOHZRXOG EH HVWDEOLVKHG LQ ZKLFK D VXEVHW RI 62¶V ZRXOG EH DVVHVVHG HDFK F\FOH ZLWK D UHTXLUHPHQW WKDW HDFK 62 EH YLVLWHG RQFH HYHU\WKUHH\HDUV  :LWKLQDF\FOHWKHDVVHVVPHQWIRUDJLYHQ62LQYROYHV WKHFROOHFWLRQRIGDWDIRUWKHFRXUVHVLQGLFDWHGLQWKHUHOHYDQW 3,¶V 7KLV LV GRQH E\ LQVWUXFWRUV RI WDUJHWHG FRXUVHV IRU 3,¶V VFKHGXOHGWREHDVVHVVHGLQDJLYHQVHPHVWHU

ISBN: 1-60132-498-7, CSREA Press ©



86

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

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

 ,PSOHPHQWDWLRQ  $W WKLV ZULWLQJ RXU LQVWLWXWLRQ LV LQ WKH SURFHVV RI VHHNLQJ$%(7DFFUHGLWDWLRQ,QRUGHUWRUHFHLYHDFFUHGLWDWLRQ LWLVUHTXLUHGWKDWDQLQVWLWXWLRQPXVWKDYHFRPSOHWHGWZRIXOO DVVHVVPHQW F\FOHV IRU HDFK 62 *LYHQ WKDW RXU SUHYLRXV HIIRUWV IRFXVHG RQ FRXUVH DVVHVVPHQW DQG QRW 62¶V DQG WKH IDFW WKDW WKH SODQQHG URWDWLQJ VFKHGXOH ZRXOG UHTXLUH VL[ \HDUVWRFRPSOHWHWKHWZRQHHGHGF\FOHVLWZDVQHFHVVDU\WR H[DPLQHDOWHUQDWLYHV 

8VHRI&RXUVH$VVHVVPHQW'DWD

 2QH SRVVLELOLW\ ZDV WR DWWHPSW WR PDNH XVH RI WKH VLJQLILFDQW TXDQWLW\ RI H[LVWLQJ FRXUVH DVVHVVPHQW GDWD E\ PDSSLQJ WKH FRXUVH OHYHO REMHFWLYHV WR WKH $%(7 62¶V WKHUHE\ FUHDWLQJ LQ UHWURVSHFW DQ $%(7 F\FOH 7KHQ WKH VHFRQG F\FOH FRXOG EH GRQH XVLQJ WKH QHZO\ GHYHORSHG SURFHVV GHVFULEHG SUHYLRXVO\ SRVVLEO\ RQ DQ DFFHOHUDWHG VFKHGXOH  $Q HIIRUW ZDV XQGHUWDNHQ WR H[DPLQH WKH IHDVLELOLW\ RI VXFKDPDSSLQJ,WSURYHGWREHGLIILFXOWWRILQGFRXUVHOHYHO REMHFWLYHV WKDW SURYLGHG VDWLVIDFWRU\ LQGLFDWLRQV RI FRPSOLDQFHZLWK$%(762¶V7KHFRXUVHREMHFWLYHVDUHDELW WRR WDUJHWHG WR PDS FOHDQO\ WR $%(7¶V PRUH JHQHUDO 62¶V ZKLFK SHUKDSV LV JRRG VXSSRUW IRU $%(7¶V GHFLVLRQ WR QRW UHTXLUHFRXUVHDVVHVVPHQW  $QRWKHU FRPSOLFDWLQJ IDFWRU ZDV WKDW VWXGHQW¶V PDMRUV ZHUH QRW LGHQWLILHG ZLWK WKH FROOHFWHG GDWD ,Q RUGHU WR GR D SURSHUDVVHVVPHQWIRUDJLYHQSURJUDPLWLVQHFHVVDU\WRRQO\ FRQVLGHU GDWD IURP VWXGHQWV PDMRULQJ LQ WKDW SURJUDP ZKHQ SHUIRUPLQJWKHDVVHVVPHQWIRUWKDWSURJUDPZKLFKLVZK\WKH VWXGHQW¶VPDMRULVQRZSDUWRIWKHFROOHFWHGGDWD7KHLQDELOLW\ WR VHSDUDWH VWXGHQWV E\ PDMRU RU HYHQ WR UHPRYH VWXGHQWV IURP RXWVLGH RI HLWKHU PDMRU ZKR KDSSHQHG WR WDNH WKH FRXUVH ZRXOG LPSDFW WKH DVVHVVPHQW LQ DQ XQNQRZDEOH PDQQHU  7KHHQGUHVXOWLVWKDWDOWKRXJKDQDWWHPSWZDVPDGHWR PDS FRXUVH DVVHVVPHQW GDWD WR $%(7 62¶V DQG JHQHUDWH DVVHVVPHQW GDWD IRU UHYLHZ LW PXVW EH YLHZHG DV VXSSOHPHQWDO ,W LV HYLGHQFH WKDW DQ DVVHVVPHQW HIIRUW ZDV

XQGHUZD\SULRUWRFRQIRUPDQFHZLWK$%(7UHTXLUHPHQWVEXW ZLOO QRW EH XVHG GLUHFWO\ LQ DSSOLFDWLRQ IRU $%(7 DFFUHGLWDWLRQ 

$VVHVVPHQW6WDWXV

 *LYHQ WKDW WKH XVH RI SULRU DVVHVVPHQW GDWD KDG QRW SURYHQ IHDVLEOH RXU GHSDUWPHQW ZDV IDFHG ZLWK QHHGHG WR FRPSOHWHWZRIXOODVVHVVPHQWF\FOHVEHIRUHDSSOLFDWLRQFRXOG EH PDGH WR $%(7 IRU DFFUHGLWDWLRQ 5DWKHU WKDQ GHOD\ DSSOLFDWLRQ ZKLOH WZR \HDU DVVHVVPHQW F\FOHV ZHUH FRPSOHWHG LW ZDV GHFLGHG WR FRPSUHVV WKH VFKHGXOH IRU WKH ILUVW WZR DVVHVVPHQW F\FOHV LQWR WZR DFDGHPLF \HDUV DQG DSSO\IRUDFFUHGLWDWLRQXSRQFRPSOHWLRQRIWKRVHWZRF\FOHV 7KXVDOOVHYHQ62¶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

 &RQFOXVLRQV  $ GHVFULSWLRQ RI DQ DVVHVVPHQW SURFHVV WKDW KDV EHHQ GHYHORSHGRYHUWLPHDQGLVQRZLQXVHDWRXULQVWLWXWLRQKDV EHHQ SUHVHQWHG 7KLV SURFHVV KDV EHHQ GHYHORSHG WR EH FRPSOLDQW ZLWK UHTXLUHPHQWV VHW GRZQ E\ $%(7 IRU DFFUHGLWDWLRQ ,QFOXGHG LV D GLVFXVVLRQ RI SULRU DVVHVVPHQW HIIRUWV WKH UHDVRQV IRU PLJUDWLQJ WR WKH SUHVHQW V\VWHP DQG WKH GHYHORSPHQW VWHSV WDNHQ LQ WKDW PLJUDWLRQ 6RPH RI WKH OHVVRQVOHDUQHGLQWKLVSURFHVVDUHSUHVHQWHGQRZLQWKHKRSH WKDW WKLV LQIRUPDWLRQ ZLOO SURYH XVHIXO WR RWKHU LQVWLWXWLRQV ORRNLQJWRDVVHVVWKHHIILFDF\RIWKHLUSURJUDPVDQGRUDSSO\ IRUDFFUHGLWDWLRQ  $V KDV EHHQ SUHVHQWHG KHUH $%(7 KDV DFFUHGLWDWLRQ UHTXLUHPHQWV WKDW IRFXV RQ DFKLHYLQJ D VSHFLILF OLVW RI OHDUQLQJREMHFWLYHVIRUVWXGHQWV,QVWLWXWLRQVPXVWGHPRQVWUDWH WKDW WKH\ DUH QRW RQO\ DFKLHYLQJ WKRVH REMHFWLYHV EXW FRQWLQXRXVO\ LPSURYLQJ DV ZHOO %H\RQG WKRVH FRUH

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

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¶V2IFRXUVHWKLVLVQRWVRPHWKLQJ WKDW DQ LQVWLWXWLRQ KDV FRQWURO RYHU VR QR VSHFLILF UHFRPPHQGDWLRQV FDQ EH PDGH RWKHU WKDQ WR VXJJHVW WKDW ZKDWHYHU SURFHVV LV DGRSWHG EH IOH[LEOH HQRXJK WR DGDSW DV 62¶VHYROYH  :KLOH LW KDV QRW EHHQ VSHFLILFDOO\ DGGUHVVHG WR WKLV SRLQW LW ZDV IRXQG WKDW LW LV H[WUHPHO\ LPSRUWDQW WKDW DOO IDFXOW\ DUH HQJDJHG LQ WKH DVVHVVPHQW HIIRUW 7KLV LV QRW DQ HIIRUWWKDWFDQEHSHUIRUPHGE\RQHGHSDUWPHQWDO³DVVHVVPHQW H[SHUW´ 7KLV LV HVSHFLDOO\ VR LI WKH SODQ LV WR XVH FRXUVH JHQHUDWHGGDWDIRUDVVHVVPHQW7KHFUHDWLRQRIUXEULFVDQGWKH VHOHFWLRQ RI FRXUVHV DQG FRXUVH PDWHULDOV WR EH FROOHFWHG DOO EHQHILW DQG HYHQ UHTXLUH WKH H[SHUWLVH RI WKRVH LQYROYHG LQ WKH LQVWUXFWLRQ RI WKH WDUJHWHG FRXUVHV 2XU GHSDUWPHQW WKURXJKRXW WKLV SURFHVV KDV PDGH WKH LQYROYHPHQW RI DOO IDFXOW\ D SULRULW\ LQ WKLV HIIRUW ,Q VKRUW D ³FXOWXUH RI DVVHVVPHQW´QHHGVWREHDGRSWHG  )LQDOO\ RXU LQVWLWXWLRQ LV SUHVHQWO\ LQ WKH SURFHVV RI VHHNLQJ DFFUHGLWDWLRQ $V VXFK WKH LQIRUPDWLRQ SUHVHQWHG LV IHOW WR EH ERWK WLPHO\ DQG UHOHYDQW WR RWKHU SURJUDPV FRQVLGHULQJ DFFUHGLWDWLRQ IRU WKHLU LQVWLWXWLRQV ,W LV WKH KRSH WKDW WKH H[SHULHQFH JOHDQHG DQG SUHVHQWHG KHUH LQ EULQJLQJ RXU LQVWLWXWLRQ WR WKH SRLQWRIDSSOLFDWLRQIRUDFFUHGLWDWLRQLV XVHIXOWRWKRVHSURJUDPV

87

 5HIHUHQFHV  >@ $%(7$FFUHGLWDWLRQZZZDEHWRUJDFFUHGLWDWLRQ >@ 0XGDVVHU :\QH $OLUH]D )DUDKDQL DQG /X =KDQJ ³$%(7$FFUHGLWDWLRQ/HVVRQV/HDUQHG´3URFHHGLQJVRIWKH ,QWHUQDWLRQDO &RQIHUHQFH RQ )URQWLHUV LQ (GXFDWLRQ S  >@ $%(7 &ULWHULD IRU $FFUHGLWLQJ &RPSXWLQJ 3URJUDPV ZZZDEHWRUJDFFUHGLWDWLRQDFFUHGLWDWLRQFULWHULDFULWHULDIRU DFFUHGLWLQJFRPSXWLQJSURJUDPV >@ /LQGD6XVNLH³)LYH'LPHQVLRQVRI4XDOLW\$&RPPRQ 6HQVH *XLGH WR $FFUHGLWDWLRQ DQG $FFRXQWDELOLW\´ -RVH\ %DVV:LOH\  >@ 0DUWLQ &DYH 6WHSKHQ +DQQH\ 0DU\ +HQNHO DQG 0DXULFH .RJDQ ³7KH 8VH RI 3HUIRUPDQFH ,QGLFDWRUV LQ +LJKHU(GXFDWLRQ´%LGGOHV/WG >@ $%(7 &ULWHULD IRU $FFUHGLWLQJ &RPSXWLQJ 3URJUDPV ZZZDEHWRUJDFFUHGLWDWLRQDFFUHGLWDWLRQFULWHULDFULWHULDIRU DFFUHGLWLQJFRPSXWLQJSURJUDPV*& >@ 0XUSK\ (ODLQD ³$ 5HYLHZ RI%ORRP V7D[RQRP\DQG .ROE V 7KHRU\ RI ([SHULHQWLDO/HDUQLQJ 3UDFWLFDO 8VHV IRU 3ULRU/HDUQLQJ$VVHVVPHQW´ -RXUQDO RI &RQWLQXLQJ +LJKHU (GXFDWLRQ9RO,VVXHS)DOO >@ $GDPV 1DQF\ ( ³%ORRP V7D[RQRP\RI &RJQLWLYH /HDUQLQJ 2EMHFWLYHV´ -RXUQDO RI WKH 0HGLFDO /LEUDU\ $VVRFLDWLRQ9RO,VVXHS-XO >@ $LGDQ %UDGOH\ *URVDV 6KLZDQL 5DQL 5DMX %XUNKDUGW 6LHJIULHG6FKXHWW-R$QQH&KXFNDQG7KRPDV-DPHV0LOODU ³'HWHUPLQLQJ LI DFWLYH OHDUQLQJ WKURXJK DIRUPDWLYH DVVHVVPHQWSURFHVV WUDQVODWHV WR EHWWHU SHUIRUPDQFH LQVXPPDWLYHDVVHVVPHQW´ 6WXGLHV LQ +LJKHU (GXFDWLRQ 9RO ,VVXHS6HS    

ISBN: 1-60132-498-7, CSREA Press ©

88

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

        ! " #$%& $ 
     
  
  


    



    

                                 

                           

                               

                                  

 - !    # ! /42?42@4284





  

  

'&()*+ & )'

, !  # ! -  .  !  # //  !!# / - # , - ## /0!  -## " ##&!# ,  #!/!  !0 0.  ##! 0!    /  -    ! !  0     ,    1234   / -      / /   ! / ! 5  -      !/ ! ", $   /   ! $  %  #  //&!  /, 6$%&7,6000- /7,  ! / !     / ! 1 -!-! # # 0!!  &! $%&  !     !         !! #  ! , ! $%& 8983   :  

  5 ;&!  /  /   ,     #  /  /!" 0!!!  -/  &!  # !    -     #  !      # !  / )! - #  

&!-!  -!//  /        !  / +/        , !   ! 0   1  1   0   /!1-/  284

A 0,!  #      - ! ! / # !   # #/  #/28 4$ ,!!

# !    !   /   0 -!/  #0 .    ! #! 



)'& '+)+  ()B'&)*C

&! -! # !       , 0!!     - 0  ## , 0 #! /!# # 8@&!# !   #      0! !  # ! /,0!! 0!!

#!/   #!-D,0!!   -!# !  #!/  !   # !       #!  


















 

 

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | !
"   #     

 



&! # , ! E  # !            0  - 5 ! -  #   # !  #   ! -  # (/ #! /  &!-  #  #!  # 

#!#  0/5 

& !  / )-D  D !  ) 5 !   !  #! -D # ! /F !,  ! 5  #         #     # ! , 0!!  !    / !  /!   

&   ! C/ )    !  5!  #!0 # !    #   # ! C/ )         -   !  0  # ! /! 0. # !  /      / !  !! #!     &!-  # (/  ( # !-  #  #!  # , 0!! !   #     /0!!$%&    $
"        

 &!   , ! -DE  # !            ! - 5 67 /, 67   /,  67     -   &!/- !  # /0!!$%&  ,0!!!  / 6"$=7 &!,!  /- # !

#!    0!- !!  #. 6$%& )+& ) =,:  97  !    #! /6$%&)+& )8, ,3 >7 &!#  0/ 5 

# !  /C/) 

* !5/, ,",

(-

 !- !  #! ! #!/  

89 &!   ! ! # !  / 

  - !!-// ! #!$" #! ! 0&-8  -! !/

    ! -D G      ! //G  "   , - ! 11,  &- 85 " #    # ! /

 !- !!-//!

#!          ! " # $ % & + !            #!!   $  # !        !# !

#!  $  !  #               / -  -  !!   )/,         0 . /  0 .  0! -D/  $ /    ,! #! # ! /    ! -D # G      ! //G 85  0 ! H >5 !/! !!/   

  -# -  !         !!// # !     !// !./- ! 0!! # !0!##  - 

 

  -# -  !         !!// # !     ! //   !./ -  ! 0   # !  #! 0!     



ISBN: 1-60132-498-7, CSREA Press ©

90

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | &- 5 " #  #  !   / C/ )   '    )     ./ (    *    + ,   -   ' / 

  *   )        + 

    ,  C/  I  #  8 ) 8      '       (            I  # C/    ) ' '                              C/  I  #  8 ) 8      '       (       0     I  # C/    ) ' '    / / #! #! C/) 5

#      I  J K 93 ! /KA/!

#      I  J K 99      I  L K :9 ! / K 

#      I  J K ::     I LK:!/KC 0

#      I  J K =@     I LK8!/KB C 0  %#1 0!!##- !/!  

 F,!-   #     /! #! #!  !   / -  ! /!  ! #   1!  #/!/   &! !- 5 

  # /   / !    #  !  # ! E5 !  ! #   -     !/!0.!    !  ! !//   # ! ! 0!! !M % !11 ,!"#  

  # /   / !      #  !   / C/ ) 5/!! !   / ! #!/  ./ ! /   0! !     ! ! -  / ! $ ,!!/ -#1 0!!## -!/! F,! # ! # !  /C/)  /! /#  0!# !

# !  )  # !  / !  /0!!$%&   

  #!/  /?- / 0! $%&5 / !   !  ## !/ /, ! , 0 . &!# ,    0 .,  0 .-6-=7A# ?-/0!!>$%&/   !0-##-D #! / 

/ /  #! -  !   5 !   0   / !  - // ! 

- -5 87 !   # ! /

  / !      #  !   /C/) , 7  #! /   / !    - !  #!E ,=7 

#!//?- 

   #! #!0!!   #  5 / !  ! / #!   ! #   # !  / 0! !  #  / -/#-.!# ! /-        

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

91

1 +  2 , 1  (   5  





 1 

 '    . 3    +3 ,  - !)-D # !  / # D! /  #  - C/(   #!

 

'        

 1  - 3      / 0!$%& 

  '   -  3 .  

'   - '  -  

    1 

3 .    '7*   

'   3   3     #/  /  #     #/  / # # !   #/

     #/  /?- /0!$%& 

/ /  #!

-!    ! #!0!!  #!

6          

 1        $  

1  (   .  1  3  1/4 '   - 

. 3  

  

5  3    #/  /  #   /!    /    /?-/0! $%&  / 6,,.,",7 

  #$   6!/ ,   7

'(   .



ISBN: 1-60132-498-7, CSREA Press ©

 1    # !  / C/)  

 (-/0!!?   $%&     !! #! /  -    * 6,., ",7   

92

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

                                                          

&-=5" #(- &0 .   -     8 101*. .*-9-5'01* (06(1'- 

          :8

"+   ,

 1  " $+.   ,

 0  $   &   +9 ' ,

3

  -  

    8 

&!         /! #      ! -D

&!     #    ! -D 

&!  /  # #    !-D 



  ;<        ,: 637,?::?9@  284
!#,$,,,% #,T$, 1-, U, $!, ', $1!,  $  )  $ !   / #  $%& $ 5 $   #   #     / $, 8@,>, 8=39=8=33>  2884
%,&!/  #1 1 !     !5 /,8@@?,$  28 4
U, % P,   ,  &! !  / #  -  /, 8V W, == 3 , -/ + , '0 T .6 =7                                        2@4


B
(F('   284
6$%&7,//  8983,  #  $/  /  //  ! + , % , *,% ,*, 89O= =:  2 4
N , P  / #   /, /,   /   !    #  // ,F   #,8@@?F M@? ?!$, 5 8@@?O88>8 8  2=4
N , PO -, O 0, Q, * /   !   /# // P  # //  8@@?, ?>6:7,  =99 =39  2:4
* .,  , &!      #    #  // /5$ #     /8@@>,:687  294
PR/,FO%.S # ,C, //!   #        -   //     P  # //   >,= 6 7,8?88@8  234
(#, AC, ! (N N - &  /$  , @,:  6887,8 8:  2>4
$ 0, &,O $!, $      !   /  //  /  S0 +  P  # //  , 83, :86:7,=3@=>@  2?4


$-P, %,O $$, A  

$%&      ! //  / P  #//  , 8,=9697,:?@99 

ISBN: 1-60132-498-7, CSREA Press ©

94

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

SESSION RESEARCH PROJECTS AND CAPSTONE DESIGN PROJECTS Chair(s) TBA

ISBN: 1-60132-498-7, CSREA Press ©

95

96

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

97

An Internet of Things Drone Data Mule Drew Cochrane, Nolan Evans, Michael Lane, Grant Woodbury, Xi Yin, Corey Zrobek, Marcia Friesen, Ken Ferens Dept. of Electrical and Computer Engineering, University of Manitoba, Winnipeg, MB, Canada Corresponding Contact: [email protected]

Abstract— This paper is the fourth in an annual series that reports on one of the approaches that an Electrical and Computer Engineering department at a Canadian University takes to improve, maintain, and ensure their accreditation in undergraduate programs. In particular, this paper reports on a capstone design project that gives computer engineering students the opportunity to demonstrate the skills (aka, attributes) that they have attained throughout their studies by designing, implementing and testing a solution to a medium complexity Engineering design problem. The skills these particular students demonstrated were the attributes of Knowledge, Design, Problem Analysis, Investigation, Teamwork, Communication, and Impact on Society, which are part of the Canadian Engineering Accreditation Board’s 12 Graduate Attributes. Keywords— Capstone design project, accreditation, CEAB graduate attributes, Drones, Data Mule, Delay Tolerant Networks, Internet of Things, data networking, Internet of Things Architecture.

1. Introduction The Department of Electrical and Computer Engineering (ECE) at the University of Manitoba, Canada, offers a “capstone” design course, ECE 4600 Group Design Project, which is normally taken by students in the final year of their undergraduate ECE program. The project work constitutes a significant design experience based on the knowledge and skills acquired throughout their undergraduate program and gives students an exposure to the concepts of engineering design in a team environment. Students are required to demonstrate, within a fixed time period (7 months, two terms), the ability to conceive a design of a solution to a medium size engineering problem, and to organize, conduct, test, and report on it. The requirements consists of five parts: a project proposal, the engineering log-book, written and oral progress reports, a formal engineering report, and a public oral presentation. This paper is organized as follows: First, a brief overview is given on a chosen design project, followed by

related work in the application of drone data mules in the Internet of Things (IoT) research and industry fields. Next, in depth description of the different subsystems, including the ground node, drone node, and server node is given. This is followed by a description of the unit and system testing experiments and analysis conducted to verify the system requirements. Finally, conclusions and recommendations for future work are given. Overview of the Drone Data Mule IoT Device The following is a brief overview of the drone data mule design project done by undergraduate students at the Department of ECE, U of M, 2019. This design project uses a drone to gather data from remote sensors and send the data up to a cloud for further processing and storage. The primary goal of this project was to design a drone, which computably acted as a data mule, gathering data from remote regions and transferring to areas with established internet infrastructure. The motivation for this project comes from the desire to reduce risk, decrease expense, and increase accessibility to environmental or customer preferred data. The project is intended for regions that are hard to access with no internet connectivity and for data which is delay tolerant. Retrieval of such data tends to have high costs and places workers in unsafe situations. This project was designed as a proof of concept and relies heavily on full system integration between the ground node (environmental sensor), the drone node (data mule), and the server node (internet connected infrastructure). The primary solution is a Printed Circuit Board (PCB) integrated onto an Unmanned Aerial Vehicle (Drone) which is capable of autonomously and wirelessly exchanging data. The PCB solution consists of a microcontroller unit, Bluetooth BLE module, SD card, and power systems. The PCB software was designed to interface and exchange information with both the ground and the server nodes. As a proof-of-concept, the ground node was developed to record, store, and transmit temperature data. The server node was developed to receive, and store data in a cloud based web server and database. This solution was developed at the University of

ISBN: 1-60132-498-7, CSREA Press ©

98

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Manitoba campus, and resulted in a product which wirelessly and autonomously enables data transfer between the ground and server nodes.

2. Related Work The Internet of Things (IoT) is the inter-networking of heterogeneous physical devices which rely on while also supporting networks of data-gathering sensors and cloud computing. IoT applications typically include a network of sensors, a data collection component or function, and a control or feedback component [1]. In the current work outlined in this paper, the IoT is envisioned as a distributed network of sensors collecting information (for example, water quality sensors on lakes; sensors on city infrastructure that may collect temperature or noise or emissions data). The drone, acting as a data mule, regularly travels past each sensor and collects (uploads) the data from that sensor via a Bluetooth connection between the drone and the sensor, and the drone stores these data in its on-board memory. Once the drone has completed its tour of visiting all sensor nodes, it travels back to a base station where the data are offloaded to Internet infrastructure for analysis. An increasingly familiar arena of a growing range of IoT applications is the so-called Smarthome realm. This includes applications such as a refrigerator maintaining an inventory of contents, remote monitoring and control of home entry doors and garage doors, systems that detect and activate home lighting, and personal assistants that control electronic devices such as Google Home and Amazon Alexa. This scope of IoT applications continues to develop rapidly. In industrial sectors, IoT is often part of a strategy to gain operational and financial efficiencies, and those efficiencies are gained through analysis of data captured through IoT applications. Two fields that are prime targets for IoT applications are transportation and environmental monitoring. In the transportation field, IoT applications vary from fleet management (e.g. vehicle location, optimized trip trajectories and multi-node tours, geofencing), complementary way-finding technology (signage, hazards alert, etc.) and fleet analytics. IoT applications in environmental monitoring include but are not limited to air and water quality monitoring. A common example is monitoring surface water quality over time and area. One can envision the application described in this paper as highly suited to gathering surface water quality information from a sensor network in remote locations difficult to access efficiently by other means, for example, in a series of remote lakes. IoT applications in

air quality monitoring can be both local and regional, with the scale of the network of distributed sensors growing with the range. Data can be collected and/or crowdsourced, and big data approaches and machine learning can be applied to assess and map air quality in terms of pollutants and particulates. Data are amenable to time-series representations, as well as integration with other environmental data to analyze dispersion and forecast trends [2]. Environmental monitoring applications of IoT are part of a larger movement towards Smart Cities, with the objective to manage assets and resources for efficiency and performance, ranging from transportation systems, power, water and waste infrastructure, information systems, and public services (e.g. healthcare, social services). These objectives are met by using sensor networks to gather large amounts of data and apply data analytics to better understand user needs, behaviours, and bottlenecks and to respond in real-time [3] [4] [5]. Previous work in the research group has explored IoT applications in these areas as well, including air quality and parking lot utilization [6] [7].

3. System Architecture and Design Ground Node Hardware The purpose of the ground node is to act as a remote temperature sensor, storing data over time to be collected by the drone node. The ground node collects temperature measurements every second, storing these on an SD card. When a BLE connection is formed with the drone node, all temperature measurements collected on the SD card are transmitted wirelessly to the drone node. As shown in Fig. 1, the ground node consist of a microcontroller, temperature sensor, Bluetooth module, and SD card module. The Arduino Uno was chosen to be used as the microcontroller for a variety of reasons. First, the Arduino Uno supports Inter-Integrated Circuit I2C and Serial Peripheral Interface (SPI) communication protocols. These protocols are needed for interfacing with peripheral devices to build the ground node [8]. Secondly, there is a large open source software community supporting the Arduino Uno. Each respective ground node device purchased, such as the TMP102 temperature sensor breakout board, already had wiring schematics and example software for use with the Arduino Uno. Open source support was integral given that the main focus of the project was the drone node; any time saved during ground node development was essential to project success.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Ground node experimental setup. To collect temperature sensor measurements, the TMP102 temperature sensor was chosen. The SparkFun Digital Temperature Sensor Breakout - TMP102 module was purchased due to the small size of the sensor, to alleviate soldering difficulties. The TMP102 breakout board has built in 4.7 k pull-up resistors to facilitate the I2C serial communication supported by the TMP102 temperature sensor. As well, SparkFun provides open source driver software to facilitate I2C communications with the breakout board [9]. This open source software facilitates changing temperature units, the bit accuracy of measurements, and the frequency of distinct temperature measurements on the TMP102 sensor. After a temperature measurement was collected by the TMP102 temperature sensor, the temperature measurement needed to be stored. The Adafruit Micro SD breakout board was chosen to store collected temperature measurements. The newest Micro SD cards are edge triggered and require very square transitions during SPI communication. Further, SD cards require a strict 3.3V supplied to the card; it is recommended level shifters are used to accomplish this strict voltage level [10]. Thus, after preliminary research it made sense to purchase an SD card circuit that was known to work, rather than building a circuit from scratch. The Adafruit Micro SD Breakout Board met the above requirements of having square transitions during serial communication, as it used a Texas Instruments CD74HC4050 level shifter that converted all interface logic to 3.3V [10]. The manual provided with the

99

Adafruit Micro SD Breakout Board also describes in detail how to wire the breakout board to an Arduino Uno. As well, Adafruit provides details on how to use the built in SD card libraries supported by the open source Arduino Community [10]. Choosing a board that supports the use of open source SD card libraries was essential given the complexity associated with SD card interfacing. Temperature measurements stored over time on an SD card is needed to be transmitted wirelessly to the drone node. BLE was chosen early in the project as the wireless protocol to be used to transmit temperature data from the ground node to the drone node. The original plan was to use the same BLE module in both the ground node and drone node. However, issues were encountered when originally attempting to interface the BM71 BLE module to the drone node; it was unclear whether the BM71 BLE module was going to work for the drone node. Thus, to facilitate continued development of the ground node, the Adafruit Bluefruit LE SPI Friend was chosen as the BLE module for the ground node. The Adafruit BLE module uses SPI communication, this was already proven to work with the Adafruit SD card module at this point in the project. Further, one of the lessons learned from the BM71 module was that extensive documentation and examples on how to use a BLE module saved time during development. The Adafruit BLE module has an extensive manual describing its use, as well as countless example software projects compatible with the Arduino Uno. Further, the module had an accompanying Android Application for testing the aforementioned example projects. As well, Adafruit provides driver software for interacting with the Adafruit BLE module [11]. Given successes using the Adafruit SD card module at this point in the project, as well as the above documentation and examples, the Adafruit BLE module seemed like the most logical choice for the ground node. To create a functioning ground node, the sensors described above all needed to be connected to the Arduino microcontroller board. The TMP102 temperature sensor used the I2C communication protocol; it was connected to the hardware I2C pins of the Arduino Uno. However, both the Adafruit SD card module and Adafruit BLE module used the SPI communication protocol. The Arduino Uno supports two methods of SPI communication, hardware SPI and software SPI. Hardware SPI uses a dedicated SPI circuit to facilitate SPI communication, whereas software SPI manipulates general purpose input output pins to enable SPI communication. Given the overhead of the SPI protocol using purely software, as well as greater inaccuracy of General Purpose Input/Output (GPIO) pins, software SPI is significantly slower than hardware SPI

ISBN: 1-60132-498-7, CSREA Press ©

100

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

[11]. Both the SD card module and BLE module on the ground node require high communication speeds in order to ooad SD card data via BLE to the drone node as fast as possible. Consequently, a shared hardware SPI bus was used to ensure the fastest transfer rate of SD card data to the BLE. Originally software SPI was used for the BLE module, however this resulted in significantly slower throughput to a peripheral when transmitting data. With both the SD card module and the BLE module sharing a hardware SPI bus, a throughput of approximately 16 Kbps was achieved to a Samsung Galaxy S8+ BLE peripheral when transmitting temperature data. However, the aforementioned throughput was not attainable when sending temperature data to the drone node, as is described in detail later in Section 3 of the report. Software The ground node software uses the three hardware components, described in the ground node hardware section. These components collect temperature measurements over time, and then transmit these values to the drone node when a BLE connection is formed. First, the ground node software initializes its connections with the three peripheral components, these being the I2C connection with the TMP102 temperature sensor, and the SPI connections with the SD card module and BLE module. The BLE module is then configured with a custom BLE service, which encapsulates custom defined BLE characteristics for the transmission and reception of data. At first, the intention was to have the drone node use BLE notifications on the transparent Universal Asynchronous Receive Transmit (UART) service of the BLE module. A transparent UART service forms a data tunnel between two BLE modules, masking the manipulation of BLE characteristics in the transmission of data between the two modules. Behind each transparent UART service are a transmission and reception characteristic defined within a BLE service; these are used to form the actual data tunnel between the two BLE modules. However, in order for the transparent UART service to work between two modules, the transmission and reception characteristics must have the same Universal Unique Identifiers (UUID) on each BLE module respectively. The drone node BLE module and ground node BLE module use different characteristics for their transparent UART services. Since transparent UART services manipulate a set of characteristics with distinct UUID’s in order to facilitate communication between BLE modules, it then made sense for the drone node BLE module to notify on the ground node transparent UART transmission BLE characteristic.

The BLE server was the BLE module on the ground node, and the BLE client was the drone node BLE module. A BLE client notifies on a BLE server characteristic in order to receive its data. Every time the value of the server characteristic changes the new characteristic value is sent to the BLE client, in this case the drone node. Thus, if the drone node RN4871 BLE module notifies on the ground node transmission characteristic, it would receive all transmission data from the ground node, as well as simplify Application Programming Interface (API) calls in ground node programming by utilizing the ground node transparent UART service which hides underlying ground node characteristic manipulation. BLE characteristics have certain permissions that specify which operations can be performed on them by a peripheral BLE module. The drone node BLE module invoked notifications on the characteristic of a server by writing 0x0100 to the peripheral characteristic it wished to notify on. However, the transparent UART service transmission characteristic of the ground node did not have the write permission defined, and the request to notify was never received. This led to the custom definition of a BLE transmission characteristic on the drone node with the write permission allowed; the drone node BLE module was able to receive data whenever the value of this characteristic was changed after writing the notification message to the ground node transmission characteristic. An associated custom BLE receive characteristic was then defined within the same BLE custom service as the transmission characteristic. The drone node writes values to the receive characteristic to send data to the ground node. While defining custom characteristics did create more work, it cannot be expected that every ground node will share the same transparent UART characteristics as the drone node BLE module, nor allow writing to their transparent UART characteristics. Having the drone node notify on a custom transmission characteristic is a more general solution to communication between the ground and drone node and allows communication between BLE modules of any type. After the definition of custom BLE characteristics, the ground node software moves into a continuous loop. The ground node checks repeatedly whether a BLE connection has been formed by a client with the ground node BLE module. If a connection has not been formed, the ground node reads a temperature from the TMP102 temperature sensor using the I2C libraries of the Arduino Uno. This temperature value is then written via SPI to the SD card inside the Adafruit SD card module using the Arduino SD libraries, preceded by the number of seconds from the start of Arduino Uno operation and delimited by a space. The

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

number of seconds represents a time associated with the temperature measurement. To get an actual time required the interfacing of a real time clock with the Arduino Uno, which was out of project scope. Distinct time and temperature measurement pairs were delimited by a comma. Comma delimiting was performed due to the fact that the Adafruit BLE Arduino libraries interpret newlines to mean the end of a string. A delay of one second occurs between each distinct temperature measurement. If a BLE connection is formed with a peripheral (the drone node), the ground node waits for the drone node to write the value 0xFF to its custom defined BLE receive characteristic. The drone node must notify on the custom defined ground node transmit characteristic in order to receive the characteristic’s updated values; if the value of the ground node transmit characteristic was updated before notification was requested by the drone node this data would not be sent to the drone node. Once the code 0xEE is received from the drone node, the ground node begins reading chunks of logged temperature data from its SD card. The ground node updates the value of its transmission characteristic in 20-byte chunks (as this is the maximum size that can be written to a BLE characteristic at one time). The value of the transmit characteristic is updated 5 times, and 80 bytes of data are sent. The ground node then waits for the drone node to write 0xEE to its receive characteristic again. This signifies another 80 bytes of data can be sent to the drone node. Data must be sent in chunks due to the limited size of the drone node receive buffer. If notifications from the ground node are received too rapidly, the BLE receive buffer on the drone node over flows. Drone Node Hardware The function of the drone node addition to the actual drone (Fig. 2) is to wirelessly collect data from the ground node, store the collected data, and offload the stored data wirelessly to a server.

Drone Node and enclosure.

101

The microcontroller of the drone node facilitates the collection, storage, and offloading of temperature data through serial communication with peripherals, as well as software run on the microcontroller itself. Components of the drone node module are chosen to support or be controlled by the drone node microcontroller. The microcontroller chosen for this project is Microchip Technology’s PIC24FJ256GA702 (PIC24). This microcontroller serially interacts with the RN4871 Bluetooth module and SD card circuit of the drone node, which communicate and store temperature data respectively. One of the most vital requirements of the drone module is that it had to be battery powered, and last for up to eight hours through the light of the drone carrying the module. While the PIC24 microcontroller is not the largest consumer of power, its low power consumption allows for a smaller battery and more configurability to be used in the overall design, aiding in meeting the weight requirement. As per the PIC24 data sheet, the maximum current output of any Vss pin is 300 mA, while the maximum output voltage is 4V. In terms of system throughput, with the 8 MHz internal oscillator being used serial communication throughput exceeds 10 Kbps. Exemplary of this is SPI communication with the SD card supported by the PIC24 with an SPI clock speed of 2 MHz. Lastly, the 16 KB of RAM on the PIC24 microcontroller met the memory requirement of storing 10 KB of temperature data, however the SD card circuit was added into the drone module to greatly enhance the amount of temperature data stored per drone light and storage on the PIC24 became irrelevant. The acquisition of temperature data from ground nodes, as well as the offloading of collected data to the server node is done through Bluetooth BLE services. The RN4871 Bluetooth 4.2 Low Energy Module produced by Microchip Technologies was chosen to perform BLE communications with peripherals from the drone node. The RN4871 module uses UART for serial communication with the PIC24 microcontroller. As described in the RN4871 manual, an ASCII command API is used to configure different RN4871 settings as well as different modes of operation by sending different text strings from the PIC24 to the RN4871 serially. The RN4871 has two roles in the drone node implementation; it must retrieve data from a ground node acting as a BLE client, or send data to the server node as a BLE server. Given the RN4871 can switch between BLE client mode and server mode through the reception of ASCII commands, or through a hardware pin control, it was a valid choice for communication with the ground and server nodes.

ISBN: 1-60132-498-7, CSREA Press ©

102

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

The RN4871 was chosen for its small form factor, builtin ceramic chip antenna, support of UART, and interfaces at an operating range of 1.9 V to 3.6 V. A small form-factor is important to the overall design of the drone node because of strict weight and space restrictions, and the presence of a integrated antenna was necessary because of the challenging nature of antenna design. The RN4871 supports two different roles, BLE server mode and BLE client mode. Both roles are utilized by the RN4871 during flight. The RN4871 goes into BLE client mode to connect to the ground node, which acts as a BLE server, described in detail in section 2 of the report. Being in BLE client mode allows the RN4871 to connect to multiple ground nodes sequentially if their MAC addresses are known. Data is received from ground node BLE servers by notifying on their BLE transmission characteristics, and data is sent to ground nodes by writing to their receive characteristics. The transparent UART service could not be used for this portion of the project, as the Adafruit BLE module of the ground node does not have the requisite properties to allow notification from the drone node. Conversely, the RN4871 acts as a BLE server when connecting to the server node (which acts as a BLE client). The server node accesses data from the drone node by notifying on its transparent UART transmit characteristic. Data is then received from the server node by the drone node by reading from its transparent UART receive characteristic. Having both BLE client and server mode allows the drone node to follow two distinct programming paths, which reduces redundancy in the code and alleviates the necessity of checking which device the drone node is connected to, reducing the size of the state machine. A method for storing temperature data received from the ground node needed to be put in place, as the PIC24 has limited on-board memory. While the PIC24 did have sufficient RAM to meet the memory requirements in the project proposal, more temperature measurement storage on the drone node would make the drone node more applicable in the real world. More than 10 KB of temperature data is likely to be collected by ground nodes between subsequent visits of the drone node. Due to weight constraints placed on the drone node PCB, a micro SD card was the most weight effective storage method. As well, through research, SD cards were a very prevalent method of data storage in embedded data logging applications; there were examples on the Internet showcasing data logging on SD cards with PIC microcontrollers. The drone module power supply system consists of a lithium polymer (LiPo) battery, voltage regulator, and charging circuit. The LiPo battery has a capacity of 900 mAh, and can power the drone module for over 8 hours.

The voltage regulator takes the 4.2V provided by the LiPo battery and converts it to 3.3V that is required by the rest of the components on the PCB. The charging circuit is used to charge the drone module through a USB port and consists of a microUSB input and a charging controller, and utilizes the 3.3V voltage regulator to provide power to the rest of the drone module circuit when charging. When dealing with small, sensitive electronic components such as in this design it is important to have a stable supplied power that is within the range specified by the manufacturer for the components and has no risk of surging or burning out equipment. Drone Node Software Software development on the drone node was performed using the MPLABX IDE with an XC16 compiler. MPLABX IDE provides on board configuration support for Microchip Technology PIC24 microcontrollers in the form of the MPLAB Code Configurator, as mentioned in the microcontroller section above. The PIC firmware can be divided up into three main subroutines: initialization, ground node communication, and server node communication. It is important that any data stored on the drone node SD card cannot be recovered by adversary entities in the event that the drone is lost or stolen. To achieve this, data will be encrypted using the encryption standard (Advanced Encryption Standard) AES-128. AES-128 is deprecated in many large corporations due to a stronger encryption standard AES-256, which has a larger key size and is therefore more robust to brute force attacks. The microcontroller chosen for this project does not have the memory or processing power to effectively implement AES-256; as such AES-128 is used. This comment should not take away from the security granted by an AES-128 implementation. AES-128 continues to be an extremely secure protocol with a negligible probability of a successful brute force attack. According to the best estimate calculations from 2012, it would take about one quintillion years to execute a brute force attack on AES128. Drone Node PCB Due to limitations in weight and size, the initial breadboard prototype could not be mounted to the drone. Instead, a PCB was designed and manufactured. The design of the PCB went through several iterations and needed to follow strict design rules to ensure crosstalk and interference would not be a factor. The final PCB ended up being a two-layered board with a thickness of 1.6 mm, width of 55.88 mm, and length of 73.66 mm. In terms of layout, the most important design rule is to keep all data lines short and isolated. To do this, the SD

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

card mount and RN4871 were kept on opposite sides of the board with the PIC24 microcontroller placed at an approximately equal distance between them. This ensured that these data lines would not be put in a position to interfere with each other, potentially causing errors. As well, a minimum six mil spacing between traces was used to ensure that there would be no interference between lines running in parallel to and from the PIC24. In order to design the PCB, EasyEDA was used to generate the footprints and draw the traces between the components. Gerber files were then generated o of the EasyEDA design and sent to the manufacturer through PCBway. Once the ordered PCBs arrived, the components were soldered to the board using a standard soldering iron and 27 AWG solder. Server Node The purpose of the server node is to receive stored temperature data from the drone node. The server architecture can be broken into three primary components. The edge server, web server and cloud database. These components form the basis of the fog computing architecture. Fog computing describes a hierarchical network designed to store and analyze internet of things data. The most attractive feature of the fog architecture is its scalability. If more data is to be offloaded to the server node in the future, the server node would be capable of scaling to the larger data and computational demands. This criteria made Amazon Web Services (AWS) a perfect solution to host the server node. The general architecture is shown in Fig. 3.

Server node architecture.

103

perfect availability. RDS and the open source Structured Query Language (MySQL) software together form the 20 Gibibyte (GiB) database used for this project. This 20 GiB database falls into the free tier of Amazon’s offerings, allowing creation of a cloud computing database at no cost for the duration of the project. Security was also paramount in deciding which database to use. Despite the relative insignificance of test data collected in this project, security online is one of the most important features of a project to consumers and corporations. RDS provides a very secure architecture as well as the ability to customize access to the database not only by user but also by Internet Protocol (IP) address. Individual user permissions are controlled by RDS as well as from within the MySQL database implemented in this project. This means that to create a user with database edit permissions you must not only have access to the MySQL database command line but also root access to the database through Amazon RDS. Security of data between the Web server and RDS is handled by AWS using the Virtual Private Cloud (VPC) service. The VPC service enables data to be transferred securely between the database and AWS web server preventing eavesdroppers from sniffing unencrypted data packets. The web server has three important functions. The web server must be able to accept data from the edge server, upload data to the cloud, and securely present the data to the user through a website. These features were achieved using the AWS Elastic Compute Cloud (EC2). EC2 provides access to Linux server instances which are completely configurable. The Linux server was set up as a t2.micro instance which means it has a single Central Processing Unit (CPU) and 1 GiB Ram. This instance type was chosen for its sustainable CPU performance and solid state drive speeds. The server is configured in the same VPC as the database allowing secure communication between the two. The open source Apache HyperText Transmit Protocol (HTTP) server software was installed on the instance to serve the web page to viewers. Port 80 was opened to all IP addresses and Secure Shell File Transfer Protocol (SFTP) port 22 is designed to be open only to specific edge node IP addresses. A screen capture of the web site is shown in Fig. 4.

The cloud component of the server node was created using Amazon Web Services (AWS) and provides a location to store all data retrieved from the server node; theoretically many server nodes. AWS contains a suite of linked product offerings specifically designed to be quick to launch, easy to prototype, and simple to scale. Amazon Relational Database (RDS) service provides a relational database in the cloud with many configuration options, capital free hardware resources, security, and near

ISBN: 1-60132-498-7, CSREA Press ©

Web site home page.

104

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

When navigating to the website you will be presented with a login page. The login form was sourced online [12] for the visual appeal only. Behind the scenes is a login and registration system adapted from a secure login tutorial [13]. The decision to utilize a ready-made system mirrors the justification for using public encryption algorithms. This decision addresses the concern of vulnerabilities that arise in non-professionally written security systems. The login system utilizes session variables, JavaScript, PHP, and MySQL to create a secure interface. Passwords are hashed using Secure Hash algorithm 512 in a JavaScript implementation [14]. This means that all passwords are sent, stored, and encrypted in the MySQL database and cannot be compromised by an eavesdropping adversary. The security of the login system could be improved to prevent brute force attacks using Google’s free reCAPTCHA, however this was omitted due to limited project scope. Fig. 5 is a sitemap showing the redirects and PHP decision tree for the login page.

the directory for new files, parsing any new files, submitting the data to the database with an INSERT SQL statement, and finally zips the text file to conserve disk space. The database is currently configured with four tables. The DatamuleDB schema contains a ‘temperature’ table and a ‘nodes’ table in which respective temperature data and node metadata are stored. The ‘node’ table contains the information for each ground node, including the node ID, name, and latitude-longitude. The ‘temperatures’ and ‘nodes’ tables are designed to be joined on node id. To enforce this a foreign key constraint was added on the node id in the temperatures table. A second schema called secure login contains two tables. ‘Members’, stores the users and hashed passwords for the website login and login attempts stores the number of attempts made by each user. Fig. 6 shows the tables in the database.

Website sitemap. On successful login, the user will be redirected to the website home page. The home page plots a sampled set of data from the database. The chart is created using Google Charts. Google charts is programmed in JavaScript and the data is provided by an Ajax call. The Ajax call directs to a PHP script where the database is queried. The returned data is formatted in a proprietary JavaScript Object Notation (JSON) format set by Google Charts. In order to programmatically get data into the database it must be routed through the AWS web server. A directory was set up to temporarily hold text files before the data is inserted into the database. The insertion process is handled by a PHP script which is automatically run by a Cronjob every 15 minutes. The script works simply by searching

Database tables. The Edge Server has two primary jobs, to connect and receive data from the drone node and to transfer the data to the web server. For testing purposes an Apple Macbook Pro was used as the Edge server. Dedicated hardware would be preferred in a production environment. With that said, the macOS app created to function as the edge server is fully functional. Initially a JavaScript (JS) version of the edge server was created. The JS version relied on open source software which proved to be complicated and inefficient. The decision to change to a macOS app was made to gain more reliable control over the system’s Bluetooth hardware as well as a very large throughput increase. To get started with Apple’s Core Bluetooth

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

framework, a sample program [15] was used as a starting point. This sample code supplied the functions for discovering BLE peripherals (the drone node), connecting to BLE peripherals, discovering BLE services, and notifying on BLE characteristics.

4. Unit and System Testing and Results The PCB was manufactured using a standard soldering iron and 27 American Wire Guage (AWG) solder. By using solder this thin, the small form factor components were able to be reliably soldered to the PCB. The soldering process was done incrementally, with each subsystem being added and tested. The first subsystem implemented to the device was the system power. This included the battery connector, power switch, power LED, 3.3V regulator, charging LED, charging regulator, micro-USB connector, and all corresponding resistors and capacitors. After each of these were soldered onto the board, short circuit tests, connectivity checks, and voltage readings were done. All of these resulted in positive tests that indicated that the subsystem was working properly. However, after introducing a 5 V source via the USB, the charging integrated circuit was found to be drawing too much current to safely charge the battery. Fig. 7 shows the relationship between the resistor used in the programming pin of the charger and the current draw. It was determined that the 2 k resistor used was not accurate enough given the sensitivity of the charger. To remedy this, a much larger 4.7 k resistor was used in its place. This allows the battery to be safely charged, albeit at a slower rate.

Charging current draw. The next subsystem to be integrated onto the board was the PIC24 microcontroller. This was done so that there would be no issues regarding the test of the microcontroller affecting other components by setting pins high and low. The microcontroller also proved to be the most difficult component to solder due to the small spacing between the pins for the Surface Mount Device (SMD) form factor. Once it was soldered onto the board, the same tests were

105

conducted to ensure proper connection. As well, basic sample code was programmed to the PIC24 through the PICkit 3 debugger in order to check that the pins were functioning properly by going high and low in a loop. This was verified by using a multimeter on the pins to see the 3.3 V drop to 0 V and back again. The USB-to-UART converter was the next subsystem to be integrated. This includes the MCP2200, 12 MHz resonator, and corresponding resistors/capacitors. Once soldered, all connectivity checks were made, as well as a test to ensure data is being received by the UART. After plugging in the USB, the computer was able to recognize the device and connect to it through CoolTerm. Test data was then sent on monitored lines to ensure proper function. The next section required to be integrated was the RN4871 BLE module. This system includes the RN4871, all status indicator LED’s, the reset button, the mode switch, and corresponding resistors/capacitors. The soldering of this device went smoothly as well, with it passing all connectivity checks on the first pass. It was then tested by using the USB-to-UART converter to update the firmware of the RN4871. This was done using the isupdate software tool provided by Microchip. Through this testing the various indicators lights were monitored with no issues. Once updated, CoolTerm was used to send test commands to the RN4871 in order to see if it responds correctly. This was done by sending "$$$" to the RN4871, expecting to see CMD>be returned. Once that was verified, the PIC24 to RN4871 connection was tested in a similar manner to ensure that commands from the PIC24 were being sent and received properly. Once again, the CMD>was returned showing proper functionality. The final subsystem to be integrated was the drone’s SD card module. This included the SD card mount, 4050 buffer, SD power LED, and corresponding resistors. Once it was fully soldered, connectivity checks were conducted, followed by tests using sample code to write simple files to an SD card. After some issues with mounting, it was determined that the SD card being used was faulty, causing it to be replaced. After this, the SD card was able to mount, read, and write with the PIC24 successfully. As a final check, files from the ground node were transferred to the RN4871, then to the PIC, then to the SD card, then back to the PIC, then to the Cloud server. This was monitored using the MCP2200 showing that the system was successfully integrated. Functional Integration Connectivity between the ground node and the drone node was fully established using the connection procedure defined in the code, whereby the drone node connected directly to the ground node’s Bluetooth MAC address. A topographic survey of connectivity was performed

ISBN: 1-60132-498-7, CSREA Press ©

106

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

between the ground node and drone node. Connectivity between the modules was established in a semi-circle pattern with 45 degree resolution. As shown in Fig. 8, connectivity was tested to 31 feet at each of the desired positions. In these figures the different colours represent the different altitude planes where connectivity was established. Each connection point was validated by performing a full rotation of the drone node around the zaxis to simulate potential approaches to the ground node by the drone itself. Each rotation was also performed in increments of 45 degrees, by connecting and waiting for a disconnect between the ground node and drone node. Connecting to the ground node was reliable with no ambiguity as to the connection status. No disconnections were observed. Further testing was performed past a 31 foot radius from the ground node as the connectivity exceeds the performance metric of 5 meters. This connectivity is supported by Fig. 9 demonstrating the antenna characteristics of the RN4871. We can see that the module radiates strongly in most orientations. Based on our integrated design, connections would be coming from what is labelled as the xy-axis in Fig. 9, where the radiation pattern has high gain of 0 dB to -5 dBi.

Connectivity topology.

RN4871 Antenna Characteristics.

5. Conclusions and Future Work The object of this project was achieved: to develop a proof of concept for a method in which to gather data from sensors located in dangerous or remote environments through use of a drone. A drone was outfitted with a PCB having an onboard microcontroller, communications module, and non-volatile memory, which when hovered within range of either a ground node or a server node, functionally uploaded and offloaded data, respectively, thus acting as a fully functional data mule. The design of this project has met the performance characteristics pertaining to battery life, range, memory, and security. For instance, a battery life of greater than 8.5 hours was achieved, a connectivity of 9.4 meters was measured, an expandable memory of 16 GB was added to the node, and basic security measures were implemented on the server. The total implementation cost of a single drone node PCB was $31 CAD, meaning that this design represents a relatively inexpensive method of transferring data through use of a data mule. The following describes future potential improvements for each component the project. Ground Node The Arduino which the ground node is based on does not have a real time clock. In a real world scenario, not having a date stamp would render any data from the sensor useless. For this reason we have ordered and received a real time clock I2C module that should be implemented prior to demonstration. There is a risk that the Arduino may not have enough available memory for the real time clock implementation which may prevent us from completing this in time. Because the ground node was more of an accessory to our main scope than within, we took the simplest, highest probability of success route by choosing the Arduino. For a production environment we recognize that the Arduino is not a viable solution; a customized PCB would be more efficient. Finally, some stretch goals may include a solar power system, integrating more features such as wind speed and direction detection, and protection from the environment. Drone Node Given the drone node was our primary focus, the potential improvements lie primarily in the software and would bring us closer to or surpass our performance metrics. The encryption function of the drone node is currently incomplete. There is a formalized plan to implement encryption as described in the Drone Node Encryption section prior to demonstration. We believe this is a very achievable goal with less risk compared to the real

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

time clock on the ground node. An important performance metric which was not achieved was the throughput speeds between the ground node and drone node. Falling short of the 10 kbps goal at about 1.2 kbps, there is much room for improvement. The primary factor affecting this speed is the interactions between the ground and drone node’s BLE modules. As they are two different manufacturers they provide two unique transparent UART services. The transparent UART mode is the most efficient method for data transfer and was successful for the server to drone connection at roughly 18 kbps. Choosing these modules to be of the same make would enable these transparent services and speed up the exchange of information. As well, further integration with the drone controls would allow for more autonomy in the usage of this project. Letting the drone node communicate to the drone when files are done transferring would allow for more efficient travel times. Server Node The Server node, much like the ground node did not receive as much attention as the drone node. This leaves it in a state with many potential improvements especially because the scope of the server node is very large. Across the entire server node, the GUI requires an overhaul and addition of end user features. More interestingly, the potential of AWS has hardly been utilized. As the project stands, we are simply storing data in the cloud. A more dedicated machine learning project could built off of the architecture created by this project. The potential use cases for machine learning with large data in the cloud are unlimited. The edge node server is currently written exclusively for macOS. Given the remote nature of the problem, future iterations should include more platforms and most importantly mobile platforms. Mobility devices tend to have the furthest reach and therefore would make a more ideal server node. References

107

[4]

D. McLaren and J. Agyeman, Sharing Cities: A Case for Truly Smart and Sustainable Cities, MIT Press. ISBN 9780262029728, 2015.

[5]

M. Peris-Ortiz, D. R. Bennett and D. P.-B. Yábar, Sustainable Smart Cities: Creating Spaces for Technological, Social and Business Development, Springer. ISBN 9783319408958, 2016.

[6]

M. Grocholski, N. Moschler, R. Debacker, W. Grocholski, K. Ferens and M.R. Friesen, "HomeUnit: An Internet of Things Air Quality Monitor," in World Congress in Computer Science, Computer Engineering, and Applied Computing – International Conference on Frontiers in Education (FECS), LV, 2017.

[7]

P. de Luna, J. Guan, P. Macalalad, G. Suarez, B. Vital, M.R. Friesen and K. Ferens, "Spotter: An Internet of Things ParkingLot-Stall Monitoring System," in World Congress in Computer Science, Computer Engineering, and Applied Computing – International Conference on Frontiers in Education (FECS), Las Vegas, 2018.

[8]

Arduino, "Arduino Uno Rev 3.," [Online]. Available: https://store.arduino.cc/usa/arduino-uno-rev3. [Accessed 21 March 2019].

[9]

A. Wende, "TMP102 Digital Temperature Sensor Hookup," [Online]. Available: https://learn.sparkfun.com/tutorials/tmp102-digital-temperaturesensor-hookup-guide/all. [Accessed 21 March 2019].

[10] Adafruit, "Micro SD Card Breakout Board," 2013. [Online]. Available: https://learn.adafruit.com/adafruit-micro-sdbreakout-board-card-tutorial/introduction. [Accessed 27 March 2019]. [11] K. Townsend, "Introducing the Adafruit Bluefruit LE SPI Friend," 2015. [Online]. Available: https://learn.adafruit.com/introducing-the-adafruit-bluefruit-spibreakout/introduction. [Accessed 2019 March 27]. [12] Aigars, "Login Form v2 by Colorlib," 2018. [Online]. Available: https://colorlib.com/wp/template/login-form-v2/. [Accessed 28 March 2019]. [13] Wikihow, "Create a Secure Login Script in PHP and MySQL," 2017. [Online]. Available: https://www.wikihow.com/Create-aSecure-Login-Script-in-PHP-andMySQL#Configure_Your_Server_sub. [Accessed 28 March 2019].

[1]

R. D. McLeod, M. R. Freisen and K. Ferens, "The IoT: Examples and Trends," in 2015 International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, 2015.

[14] P. Johnston, "A JavaScript implementation of the Secure Hash Algorithm," 2009. [Online]. Available: http://pajhome.org.uk/crypt/md5/sha512.html. [Accessed 28 March 2019].

[2]

L. Xiaojun, "Datasmart," [Online]. Available: http://datasmart.ash.harvard.edu/news/article/how-cities-areusing-the-internet-of-things-to-map-air-quality-1025. [Accessed 9 May 2017].

[3]

V. L. Castellani, B. N. Zanella and M. Zorzi, "Internet of Things for Smart Cities," IEEE Internet of Things Journal, vol. 1, no. 1, pp. 22-32, 2014.

[15] Apple, "CoreBluetooth: Health Thermometer," 2018. [Online]. Available: https://developer.apple.com/library/archive/samplecode/Health Thermometer/Introduction/Intro.html. [Accessed 28 March 2019].

ISBN: 1-60132-498-7, CSREA Press ©

108

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Capstone – Introducing Students to Research through Application Development in Teams Katia Mayfield, Matthew Perry, Christopher Pounders, Lucas Pruitt, and Ory Wigington Department of Mathematical, Computing, and Natural Sciences Athens State University, Athens, AL, USA Abstract--Computer science, as many other fields, is currently requiring students to be better prepared for their entrance in the workforce or for pursuing advanced degrees by getting involved in advanced research and projects. Service learning, where the students apply their knowledge to solve a real life problem instead of a fictitious scenario, has become a usual method of giving the student the required experience. Beyond the application of their individual knowledge is also the requirement of getting acquainted with team work. This study describes the experience of a team of students of different backgrounds working to develop a tool applicable to ongoing research in the cybersecurity area. This project involved the expected software development field stages such as analysis, design, coding and testing, with an extra research component, requiring the student to plan and schedule the necessary readings to be able to understand the problem. This study describes the research background as well as the students planning and execution of the project. A brief description of the resulting tool, which is now part of the full research project, is also shown. Keywords: service learning, cybersecurity, petri nets, CAPEC, capstone project

1. Introduction In today’s society it can be difficult to find a computer science undergraduate program that teaches the same way they did twenty years ago. Before the boom of technology and every piece of information that is now available at one’s finger tips, courses would be taught with assignments and projects given from a textbook, and twenty years ago professors would expect that students ended up doing their own work. Now, with websites like chegg.com and different forums available where students can post any question (including those associated to homework assignments) and have strangers provide them with solutions and vague explanations that do not help the student learn, professors have had to become more creative than what was required once upon a time. Even when coming up with their own homework problems and exams, the cycle of making changes and producing new material has to happen almost on a semester basis as the information will get circled around between students. Most of the time, such material will get posted online, forever, because a student chose to ask a question to a complete stranger instead of asking their professor for help.

Instead of just re-creating problems for students to solve, many professors have started to implement service learning into their classroom. This is what the Computer Science department at Athens State University has done in their Senior Capstone class. Instead of having a made up scenario where students are required to go through a standardized system analysis, design, and implementation, students are put into teams of three to five and assigned to work with a community client. The professors will partner up with community representatives, usually non-profit organizations, which have not always been the case, and discuss a service that can be provided to them by the students. The majority of the projects in the last five years in the authors’ institution have been of a web application system, with a few others completed in the areas of game development, security analysis, and research. Throughout their time at Athens State University, students take several classes with four different concentrations available: computer science, computer networking, information technology, or information security. When students reach their capstone class they may be grouped with students from other concentrations and they must work together to achieve the task that they have been given. This is a good opportunity for students, especially because many of them will not have held an internship during the time of their studies, but they will be able to state that they have work experience with a client in the development of some software. In this study, the focus will be on the research component of the Capstone class, combining development into the mix. In the next section, the research portion of this study will be introduced, followed by the requirements of the capstone project that was assigned. Once the details have been mentioned, the work completed by the team will be discussed, ending with a summary and future works.

2. Background Cybersecurity, cyberattacks, cyberwarfare, cybercriminals, cyber-something, these are the areas of focus from many different organizations. Software and hardware development is no longer about only developing or creating what has been provided from the specifications, but it is also making sure that all of it is secure. However, it is hard to come up with all the

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

possible vulnerabilities because not all test cases can be developed and tested, especially if one may not know what exactly to test for. Due to this, a fairly new knowledge area has been the focus of many studies, focused on the process that an attacker takes to be successful. If developers are able to understand how an attacker behaves then they should be able to safeguard against it. MITRE Corporation thought of this and wrote a report called Common Attack Pattern Enumeration and Classification (CAPEC) [8]. This document lists all known possible attacks and provides as much detailed information as is known, from the knowledge that the attacker must have to begin an attack, to the goals that the attacker will attain when successful. The report has even suggestions of areas of mitigation. The CAPEC report allows developers and others to start the process of becoming familiar with the procedures taken by an attacker. In the area of cyberattacks, one of the best ways to study an attack is through the development of a model. In this study the basis for creating a cyberattack model was through the use of Petri Nets with Players, Strategies and Cost (PNPSC) [6, 7]. A PNPSC is an extension to the well-known Petri nets, consisting of a bipartite graph made up of nodes referred to as places and transitions. A place within the PNPSC is a true/false state while a transition represents an action which the attacker can take in the process of their attack. A single PNPSC represents one form of attack. There is a token that is found in the beginning places of the PNPSC and as the attacker takes action the token travels through the PNPSC until the attacker has either accomplished his attack goal, failed to accomplish his attack goal, or has been detected and blocked by a defender. Figure 1 displays the basic concept of a PNPSC.

109

The formalism of PNPSC, described in “Petri Nets with Players, Strategies, and Cost: A Formalism for Modeling Cyberattacks,” along with the CAPEC report became the basis for the Capstone project that allowed students to develop a web application that would introduce them to newly conducted research from a group of authors collaborating from Athens State University, the University of Alabama in Huntsville, and Lipscomb University [6, 8].

3. Research Background The capstone project assigned to the students was the development of an application. However the application is also associated to an ongoing research project. This research project encompasses the realm of modeling cyberattacks, which is split into four phases, all of which are ongoing and have had several publications associated to them [9]. Before any of the research phases could evolve, the research group first came up with a formalism for PNPSC [6, 7]. Petri Nets with Players and Strategies has been briefly introduced by other researchers not including the concept of cost that could be associated to resources [13]. With the formalism of the model that would be used by the research group specified, the four phases of the project began to unroll: Phase I – Uses CAPEC reports as a basis to design fault trees that represent the attack. Once the fault trees are created, determines the process for creating a PNPSC that not only corresponds to the fault tree but also corresponds to the CAPEC report. A second part to this phase is to compare existing tools to determine the most ideal way of displaying the PNPSC and being able to not only model it but also simulate the attack represented by it [10, 11]. Phase II - This phase builds off of the generated PNPSCs resulting from Phase I of this research. This phase has three different parts to it. The first is to have a cyberattack model repository created, which became the focus of the capstone project. The repository is needed to hold the information associated to the CAPEC reports but also the full model design, details, and metadata of the constructed PNPSCs. The second part of this phase is to determine a way for the PNPSCs to be designed as components, resulting in the basis for the third part where the components are them used to create a composite and more complex cyberattack model [5].

Figure 1 A simple Petri net, basis for a PNPSC Basic components [9]

Phase III - This phase consists of validating the cyberattack models that are designed. At this moment in time the formalism of PNPSC has been validated through a face validation process where comparisons of the PNPSC were done with the details found in the

ISBN: 1-60132-498-7, CSREA Press ©

110

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

CAPEC report. Since the CAPEC report is widely used as a reference for the different cyberattacks, it has been assumed that MITRE Corp. has gone through and verified and validated the information that they have included in the report. There is ongoing work for composing more complex cyberattack models based on the single attack models, therefore the composition of the more complex attacks will still need to be validated. There is also ongoing research into using different methods to conduct the validation of the models and to not only rely on face validation. [2, 3]

simulate a cyberattack and implement machine learning to recognize attack and defense strategies. Being able to have this machine learning component allows for a vulnerability assessment to be developed [1]. Figure 2 shows a graphical representation of the four phases of the research being conducted. In the next section the details of the capstone project to create the repository associated to Phase II will be discussed.

Phase IV - The last phase of this research project is to take the cyberattack models that are created,

Figure 2 Four Phase Research

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

111

4. The Capstone Project

x

The capstone project was originally introduced to students as a two part project, the implementation of a database and the design of a GUI to interact with that database. The team that worked on this project was made up of four students all from different backgrounds within the Computer Science program at Athens State University, therefore each one of them bringing a different level of skill sets to the project. Since this was not an ordinary database that was being created where it would be simple for the students to understand the data that would need to be stored, there was a learning curve that they had to overcome, and it was through this that they were introduced to research.

x

The research component that the students had to focus on before starting the project involved the following: x x

x x

Learning about Petri nets and how a model was used Become familiar with three different cyberattacks (1) SQL injection, (2) cross-site scripting, (3) spear phishing. Understand the formalism of a PNPSC Comprehend the PNPSC models that were designed for the three different cyberattacks.

Once the team familiarized themselves with the research portions of the study, they then had enough knowledge to be able to start working out the specifications of how to design the database and the graphical user interface to work with that database. The scope of their project was to build a database that will store the 508 CAPEC report entries as well as a web interface for the user to be able to lookup as well as enter new records. The deliverable would consist of four parts: 1) the database that will hold the CAPEC records, 2) web interface for searching records, 3) web interface for entering new records, and 4) database table that will store information about PNPSC models. The capstone teams attempt to work as closely as possible, following an agile methodology and one of the first things that were accomplished was to determine the user stories that needed to be completed during the nine week semester. As a user, one wants to be able to: x filter the database by specific attack criteria x enter new attack entries into the database x search for attack patterns with matching purposes x search for attack patterns by level of CIA impact

x x x x x x x

search for attack patterns by level of typical severity search for attack patterns by the likelihood of exploit search for attack patterns by method of attack search for attack patterns by abstraction search for attack patterns by attack motivation store PNPSC models modify PNPSC models enter new PNPSC models generate PNPSC model details (to use with graphing software)

The students had nine weeks to be able to learn about the research and accomplish the implementation of all of their user stories which was agreed upon by student and professor to fulfill the requirement of the course.

5. Implementation With thirteen required user stories to complete on top of becoming familiar with the research, the students had to be sure to plan their time wisely. To do so, a timeline was established to keep them on track. Figure 3 displays the timeline that was followed for the implementation of the project.

Figure 3 Project Timeline The result of the design of the specification is presented below: A. Introduction to the System The Repository of Cyberattacks provides storage capability for the CAPEC attack pattern information and formatting for generating graphical models. [8]. The system allows users to filter through the CAPEC attack patterns as well add, edit, and delete attack pattern model components. The system allows users to view a textual representation of the attack pattern models. Familiarity with cyberattacks and the CAPEC reports is assumed.

ISBN: 1-60132-498-7, CSREA Press ©

112

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

B. System Overview The Repository of Cyberattacks has two primary components: a database component and a graphical user interface component. A set of scripts will run in the GUI generating and submitting queries to the database component for the purposes of storage and reference. The database initialization was accomplished in an abbreviated format using third party software and therefore deserves a refactor but provides enough utility in storing attack pattern information for this project. The output generated by the GUI can be validated by third party software. C. System Architecture The system is broken down into two components, a backend database and a web graphical user interface. The databased is designed to be stored and accessed locally using XAMPP [12]. The reason for this is due to the fact that the repository being created is part of ongoing research which has not all been made public and therefore access to this implementation needed to be kept private. The database is composed of five tables as shown in Figure 4.

options that the user provides. Figure 5 displays the home screen of the graphical user interface where it starts by listing all attacks that are stored from the CAPEC report, and through the drop down option on the top left side allows the user to start filtering the data to narrow down the list of attacks. Figure 6 displays one of the screens associated to modifying, editing, adding information to the database. In Figure 6, a Place is either being added to a model or deleted. Across the top portion of the GUI the user is able to select other options associated to the PNPSC models, or he/she can go back to Filter Attacks further. Lastly, Figure 7 displays the generated information for a PNPSC, in this example that of an SQL Injection. At this point the user would need to copy the text that is displayed on the website and paste it into Graphviz, a third party software, to have the graphical model generated [4].

Four of the tables contain information associated to the stored models, place, transitions, arc, and inhibitor arc information, and one table stores the information drawn from the CAPEC reports to allow for the searching of attack patterns based on specific filters. The user does not access the database directly. There is a web based graphical user interface that was designed to run a set of scripts that provide information on the formatting and design of the PNPSC models. It also has a set of scripts that takes user input and queries the database based on filtering

Figure 4 Database Tables

Figure 5 GUI Home Screen

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

113

Figure 6 Editing PNPSC Model Screen

Figure 7 Generated model information for graphical design

6. Summary Through the requirements of designing a web application, this team of students have now been introduced to three different areas that are associated to ongoing research. They have been introduced to the CAPEC reports that document cyberattacks, a method of graphically representing a model by first becoming familiar with the basic concept of Petri Nets and then studying the formalism of PNPSCs, and lastly they have been introduced to a full four phase research project and how their work in developing this system

will assist in the fulfillment of the overall research project. The system that has been introduced here is a first prototype system. The students were able to complete all of the required user stories and in the process have identified features that are being completed for version two of this prototype. One of the main items that will be worked on is the table that stores all of the CAPEC information. The students used SQLizer to be able to generate an SQL file from the xml file containing the CAPEC information. This made some areas of the table to be duplicated. One of the first actions is to normalize the database and make sure that there are no

ISBN: 1-60132-498-7, CSREA Press ©

114

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

associated anomalies or implications of redundancy. In additional to normalizing the database, additional features will need to be added to keep track of PNPSC models or pieces of models that are broken down into components and then used to compose a larger PNPSC model that would represent a full cyberattack and not just a single attack. In addition to the new information that would need to be tracked, the addition of a plugin or something similar that will allow for the user to not just generate the information for a model but to have the model displayed graphically on the webpage itself instead of just textually. This will eliminate the step of having to use an additional application to view the graphical representation of the model. Lastly, a script needs to be developed to integrate the automatic generation of PNPSC from Phase I of the research project to where all of the model information is uploaded into the database.

References [1]

Bland, J.A., Mayfield, K.P., Petty, M.D., Whitaker, T.S., Cantrell, W.A., (2018). Machine Learning Cyberattack and Defense Strategies, in: Proceedings of the 2018 AlaSim International Conference and Exposition, Huntsville, AL.

[2]

Bland, J.A., Mayfield, K.P., Petty, M.D., Whitaker, T.S.(2017), “Validating Petri Net Models of Common Attack Pattern Enumeration and Classification”, Proceedings of the 2017 AlaSim International Conference and Exposition, Huntsville, AL.

[3]

Cantrell, W.A., Mayfield, K.P., Petty, M.D., Whitaker, T.S., Bland, J.A., (2018). Structured Face Validation of Extended Petri Nets for Modeling Cyberattacks, in: Proceedings of the 2018 AlaSim International Conference and Exposition, Huntsville, AL.

[4]

Graphviz.org. (2018). Graphviz - Graph Visualization Software. [online] Available at: https://www.graphviz.org/ [Accessed 20 Aug. 2018].

[5]

Mayfield, K.P., Petty, M.D., Bland, J.A., Whitaker, T.S., (2018). Composition of Cyberattack Models, in: Proceedings of the 31st International Conference on Computer Applications in Industry and Engineering, New Orleans, LA.

[6]

Mayfield, K.P., Petty, M.D. (2018) Petri Nets with Players, Strategies and Cost: A Formalism for Modeling Cyberattacks, in Proceedings of the 16th International Conference on Security and Management, Las Vegas, NV.

[7]

Mayfield, K.P., Petty, M.D., Whitaker, T.S., Bland, J.A., Cantrell, W.A., (2018). An Extended Petri Net Formalism for Modeling Cyberattacks, in: Proceedings of the 2018 AlaSim International Conference and Exposition, Huntsville, AL.

[8]

MITRE. (n.d.). CAPEC - Common Attack Pattern Enumeration and Classification (CAPEC). Retrieved May 2017, from https://capec.mitre.org/

[9]

Petty, M. D. (2017). Modeling Cyberattacks with Petri Nets: Research Program Overview and Status Report. AlaSim International Conference and Exposition, Huntsville, AL.

[10]

Whitaker, T. S., Bland, J. A., Cantrell, W. A., Mayfield, K. M., Petty, M. D., (2018). Tools for Simulating and Visualizing Petri Nets, , Proceedings of the 2018 AlaSim International Conference and Exposition, Huntsville, AL.

[11]

Whitaker, T. S., Bland, J. A., Cantrell, W. A., Mayfield, K. M., Petty, M. D., (2017). Modeling Cyberattack Patterns in Fault Trees , Proceedings of the 2017 AlaSim International Conference and Exposition, Huntsville, AL.

[12]

Graphviz.org. (2018). Graphviz Graph Visualization Software. [online] Available at: https://www.graphviz.org/ [Accessed 20 Aug. 2018].

[13]

Zakrzewska, A. F. (2011). Modeling Cyber Conflicts Using an Extended Petri Net Formalism. IEEE Symposium on Computational Intelligence in Cyber Security (pp. 60-67). Paris, France: IEEE.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

115

Involving
Multiple
Levels
of
Students
in
a
Software
Capstone
Project
-
A
Case
Study Robert Hatch

Nicholas Setliff

The University of Virginia’s College at Wise One College Ave. Wise, VA 24293

The University of Virginia’s College at Wise One College Ave. Wise, VA 24293

[email protected]

ABSTRACT This paper documents a joint effort by Software Engineering capstone students and Software Engineering survey course students to create an augmented reality application. The goal for the capstone course was that the soon-to-be graduates could exercise knowledge learned, as well as have a taste of management, and have some help in completing their project. The goal for the survey course was to have students working together in teams on a big software project to give insight into what they would encounter in industry. The paper’s focus is on the issues and lessons learned by everyone involved in the project through qualitative feedback.

Keywords Software Engineering, capstone, multiple class collaboration, communication skills

1.

[email protected]

through a software engineering lifecycle model to develop an assigned project over the course of a semester. Struggles occurred along the way, such as the seniors’ ability to communicate effectively with their subordinates, as well as adjusting to the loss of a member of the capstone course midway through the spring semester. This particular iteration of the capstone course also added complexity because students were required to manage other team members. The supervising faculty hoped that the nature and scope of this project would give the capstone students a more realistic view of the actual work environment in the industry upon graduation, if this field is their intended route. The paper is organized as follows: related research; determining the scope and nature of the project; the lifecycle process of developing the AR app; reactions and suggestions from all students working on the project; and ideas for future work and improvements for future project iterations.

INTRODUCTION

The Software Engineering Capstone course at the University of Virginia’s College at Wise represents the application of the sequence of Software Engineering courses. The course is a two-semester sequence; during that time, students can engage in the entire lifecycle of software development - from inception to deployment. A recent capstone project was determined by the students with minimal suggestions from the supervising faculty: an augmented reality (AR) application that would help incoming freshman, as well as visitors, to navigate their way around campus using their smartphones. The supervising faculty felt that the low number of team members (two students enrolled for the sequence) would not be consistent with what the students might encounter upon entering the workforce. The supervising faculty for capstone included the survey Software Engineering course (CSC/SWE 2300) students in Spring 2018 to help the seniors complete their project - the first time that an entire class has been involved in such a deep, supporting role. Students in this course typically work

2.

RELATED WORK

Knowing the theory is not sufficient to solve real problems that a professional would encounter on a daily basis [8]. Additionally, a challenge for the course is finding an appropriate project for the term - “appropriate” meaning that projects need to be complex enough to require a team effort and application of skills learned in class. Krutz [6] suggests that students are interested in a real-world contribution. Buffardi, et al., [3] also note that “software engineering courses commonly incorporate semester-long team projects to approach emulating the real-world software development process and environment.” Beck [1] notes that these types of projects expose students to the “messy reality” of the types of situations and problems they can expect to encounter when they become employed. Students also need a clearly defined project in order to succeed , but students have issues with decisions made prior to their arrival because they want to understand why choices were made [4]. Buffardi, et al.,[3] observe that real software development projects “involve teamwork to design and produce more sophisticated products.” [1] suggests that students should be divided into teams; each team works on some portion of the system, and teams must work together to complete the project. Matthies, et al., [7] also note that agile may be a better route because the participants “adapt to the circumstances and find solutions that work in the given context instead of blindly adhering to a prescribed process.”

ISBN: 1-60132-498-7, CSREA Press ©

116

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | Another consideration is that “while students may enjoy the freedom of defining their own software products, students have recognized that such ‘toy projects’ lack the pressure of delivering well-designed and tested software to real users” [3]. Other major challenges for group work involve how to grade each individual, as well as assuring work is divided fairly within the group [2], [1]. Francois, et al., [5] suggest the use of a “class-wide project,” which engages the efforts of all the students in the design and implementation, which is “out of the reach” of any individual student working independently in the timeframe of the course. Students will often express a significant amount of satisfaction in a project, which in turn contributes to their satisfaction with a course [6]. It helps if a project looks like something a student might encounter in the real world as a professional. Such a project also has a major benefit related to the process from the instructor’s point of view in that it makes the students critically eye the tasks in the project and then analyze them to make decisions about each task [1].

3. GATHERING REQUIREMENTS The format of the two-semester sequence involved meetings on an as-needed basis to discuss questions and issues that the students had concerning the project. This time also provided feedback on students’ progress for a certain week. The first meeting of the semester focused on students as they were beginning to think about prospective projects. The only restriction placed on the class by the supervising faculty was that the project had to provide some facet of service or community service. At the following meeting, the idea for the augmented reality (AR) campus map application was selected for further development. The rationale for the AR campus map application was to provide visitors with a way to navigate the campus. The major functionality would be to provide directions from one building on campus to another building or location. Other functionality included a virtual bulletin board and a step counter. The virtual bulletin board would be utilized at a designated space in each building for a department, and each bulletin board might have announcements and student opportunities. The steps feature would allow the user to select the number of steps desired to walk, and the application would provide a route around campus that would satisfy the user’s step goal. The major functionality of going from one building to the next would best satisfy the needs of visitors on campus, as well as freshmen and transfer students. The additional features would provide a means for students to continue to revisit the application. Two early issues arose when the scope of the project was narrowed. The first was determining which mobile platform to proceed with - Android or Apple’s iOS. Each student had experience with some development on one of the platforms. In either case, students would have to become familiar with either platform’s augmented reality resources and API. The second issue was designing a product that targeted audiences would actually use. Ultimately, the supervising faculty suggested administering a survey to prospective users; it would address the questions of platform preference and also confirm which functionality would be useful to prospective users. The results of the survey found that the 70% of students surveyed (19 out of 27) preferred or used iOS over An-

droid. Respondents were receptive to the additional features of the virtual bulletin board and tracking steps for a day. The students were able to successfully finish and present their initial prototype at the end of the Fall Semester. The supervising faculty began brainstorming ideas on how to help this small development team and decided to use the next semester’s Software Engineering course to assist the capstone students in completing this project; the faculty member was responsible for teaching both courses. Potentially there would be benefits from both classes working on this project. The Capstone students would still exercise concepts learned throughout their academic career at UVaWise; they could experience what it was like to be a manager or a team lead on a project they were invested in. For the Software Engineering survey course students, this experience could serve as a springboard and generate enthusiasm - to see the potential for their own projects in the capstone course, as well as what their workplace environment might be. The survey course is required for Computer Science and Management Information Systems students. Software Engineering majors do not have to take this course because they will learn about the different parts of the process in each course they take within the major.

4. BUILDING THE PROJECT AND MAKING ADJUSTMENTS Prior to the start of the Spring 2018 semester, the project leads determined the tasks they needed to construct a prototype. These tasks included: finding a suitable server to store data points for the application’s map; physically mapping GPS coordinates for locations of entry into buildings, as well as handicapped accessible paths and entrances, and sidewalk paths; constructing front-end prototypes of the application using the Swift programming language; and determining and implementing an algorithm that took the shortest path between two points. The project leads distributed the project tasks incrementally over the course of the semester, in twoto three-week sprints. Students were assigned to teams of three or four for the duration of the project. Each team was given specific tasks to complete within a sprint. Ten students were enrolled in the survey course, all male. Project meetings were initially handled through a discord server and channel, where students could discuss issues and get help where needed. The project leads and the supervising faculty realized that communication was a problem. Students were not participating in the scheduled meetings due to conflicts or because students forgot about meeting times. During this first part of the semester, communication issues also occurred between teams and capstone students. One of the capstone students was out of town and unable to resolve issues. The other capstone student did the best he could to resolve the issues. However, not long after a quarter of the way through the semester, the project lost this second project lead. Communication between this person and the rest of those invested in the project ceased, even after attempts to reach out to this individual. After receiving initial feedback from students which predominantly focused on communication, project meetings were held for the last 15-20 minutes of a class lecture. Students could resolve questions they had face-to-face with the re-

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | maining project lead. When the remaining project lead determined that students fell behind schedule, project work days were added during the second half of the semester. Everyone from the class would meet in a computer lab and have access to the project lead during these project days. Morale (measured through assigned grades in student feedback) and progress picked up in the second half of the semester. The major lesson learned was that while communicating from anywhere can be convenient, face-to-face communication is necessary to make sure the project continues to progress. The basic functionality of the project was completed at the end of the semester, enabling someone to navigate from one spot on campus to another location. The suggested features of a virtual bulletin board and routes to achieve step goals were postponed. The bulletin board was also delayed to consider the time commitments needed from other departments to maintain the virtual space. While the prototype works, there is room for enhancements, including the aforementioned postponed features, as well as improving battery performance.

5. STUDENT FEEDBACK The following subsections describe feedback received from students at different times in Spring 2018. Subsection 5.1 addresses responses to the same question asked on the first two exams - grading each student in his group and justification of grades assigned. Subsections 5.2-5.4 address questions concerning hindsight, suggestions for faculty, and likes/dislikes/did the student feel he benefitted from the project. These questions were asked on the final exam for the survey software engineering course. Subsection 5.5 shows feedback received from the one capstone student who persisted throughout the semester.

5.1

Feedback Via Exam Questions

When the first two exams for the course were administered, students were required to answer a question about their immediate team, as well as the state of the project. The students needed to assign grades to each of their own group members and provide supporting reasons; specific comments were not shared with any students who received negative feedback. Feedback was shared with students in the form of comments written on their graded exams. These students were encouraged to continue participation because their grades depended upon the outcome of the project’s successful launch. The grades assigned by students were used as a guide as to how to grade an individual student’s work on the project at the end of the semester. One of the major themes which popped up on the first exam was communication. According to one student, “So far (our) group (has) lacked communication. We have not discussed any ideas and I am not sure if anyone else has started making paper prototypes.” Students were not afraid to grade each other down, due to a lack of communication. Another student’s response: “I would give (this) member 30 points (out of 100) as he appears to know what is going on in the project thus far but has not actually said a word to me as of yet.” Another student in another group also commented on communication issues, as well as a concern about organization: “To be honest, my team’s project has been stale. The one time we met together as a group, it was just me

117 (a project lead) and (another team member) in the discord channel. Our other member wasn’t there because of miscommunication. We have the will/desire to work, but no organizational skills.” With the first exam, some students were also marked higher by their teammates for their willingness to take on the leadership role for the project. Other students working on the project were graded higher based on their understanding of concepts that were necessary for the project. One other reason for receiving high marks from teammates came from regular attendance at meetings. The students in the survey class also graded their project leads; at this point, the project had two capstone students as project leaders. On the first exam, students were not receptive to one of the team leads because he was unavailable for a period of time while he was traveling. The second project lead received higher marks because this student was the active leader on the project. Issues with this project lead included his occasional “condescending tone,” as well as “miscommunication leading to confusion.” Those negatives were outweighed by students who thought that he “work(ed) so closely and intelligently with our group to make sure we understand what it is we need to do and how we may be able to do it. He also seems to be willing to assist as needed with the work our group does reliably.” A student’s average score earned from his teammates in the first phase was a 72.63 (out of 100). The average for the project lead whom the students felt was available to answer questions was 80.33; the second team lead earned a score of 74.42, with three students abstaining from assigning a grade, due to lack of involvement. After receiving feedback from students, the supervising faculty attempted to get the project back on track. Part of the reasoning behind asking for feedback about the project leads was to gauge how well the two leads were performing from a leadership standpoint. With one student committed at that point, while the other was less certain, the instructor used the student feedback to address the issues which the survey class had in a meeting that took up part of a lecture day when the first exams were handed back. This meeting involved all of the students in the survey Software Engineering course, the two capstone students, and the instructor. The communication issues continued with responses from the second exam. Students gave each other higher marks than on the first exam, but some noted that others did not speak until directly spoken to. Scheduling conflicts due to other priorities, such as work and personal life, also interfered with being present for meetings and reaching milestones. Some students were reported to have missed meetings, but they remained very active and constructive teammates as they completed project milestones. For the project leads, the second capstone student simply disappeared. The other team lead stepped up his role, and the students took notice. One student responded that the first project lead was “effectively the polar opposite of (the second project lead) as he responds to questions in a timed manner and actively seems to be part of the project it-

ISBN: 1-60132-498-7, CSREA Press ©

118

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | self, attempting to help anyone who may be struggling.” A second student noted, “He’s always available to talk to and pretty much always has an answer when asked a question.” A third response, also positive, was: “Since he jumped (back) into the project, (the first lead) has provided everybody with very good communication and documentation of team goals.” Another student also kept in mind how well things changed after the first exam - “strong improvement, specific tasks laid out, and helps with development often when able. Meetings went from never to weekly (just prior to [instructor] involvement).”

earlier...Also fix some of our current project version’s flaws like battery usage and node placement.” Another response stated, “I would start with the explanation of agile. I think we had potential to ’hit the ground running’ on the project if we knew more about agile. We lacked communication and timely meetings.” The last response was also concerned with communication: “I would have communicated better and pushed through the tasks at a more driven pace. Further I might have recommended to (the project lead) that he make the initial tasks smaller and more digestible to get everyone more motivated.”

Students earned an average score of 86.43 (out of 100), with comments providing proof of improvements in morale and work ethic. The first team lead, the one who eventually stopped communicating and working, received a score of 8; the second project lead rebounded, and earned a score of 96.9 from his subordinates.

In sum, tools, process, and communication appeared to have been the major issues with this project.

With some improvement and still some work to do in other areas, the supervising faculty revised what was done in the survey class. Weekly, in-class meetings were allotted after a lecture. The students communicated with the remaining project lead to clear up or address issues they were having about specific tasks assigned. Additionally, project work days were utilized in lieu of a traditional lecture, which allowed students class time to work and ask any further questions of the remaining capstone student. While a priority, grade-wise, the project was given a higher priority in the classroom as well, which drove the project’s continuing progression to a working prototype by the end of the semester.

5.2

Hindsight

Student answers were varied. One student responded, “I would (have) put more emphasis on reaching deadlines on time. I also would try to help...change communication to become more effective. The software process model wasn’t set in stone and we did have to go back to change things but I would have put emphasis on structure.” Another response was, “I would make sure communication was established earlier on...This would have saved more time for proper documentation and testing due to us avoiding a major problem we ran into that made us have to redesign the implementation of (a necessary) algorithm.” A technical concern from another student was, “I would have asked...to use a serverclient model and do all processing off device. I’d also use an object-oriented design...(and actually have used) instead of a massive combined spaghetti pile.” The third student responded that he would have begun work on the project sooner. “The last 2 weeks felt rushed compared to the first 3 weeks.” Another student was also concerned about structure: “I would have made sure that there was a more stable plan moving forward. Make a more effective schedule for tasks. Learned more about the project, Swift, and GitHub in order to be a more effective team member.” Two other students had similar feelings about GitHub and that some instruction might have helped (either by the project lead or the instructor). Other students were concerned about process. As a student remarked, “I would address the issue of communication and implement the task deadline for every two weeks

5.3

Feedback on Faculty

The students’ responses indicated they felt that the instructor’s approach was adequate without any major changes. One student’s suggestion: “I would change the leadership roles by getting to know each individual and basing the position off their experience, knowledge, and understanding of the task at hand. I would not have the professor do anything differently.” A second response said that “the project days allowed for my group to catch up on our task.” To build off that response, another student replied, “Toward the end of the project, we started having weekly meetings with (the project lead) during class time. It would have been nice to have had that earlier in the semester because it helped us all to discuss the project as a whole rather than separate teams.” One student concurred with the project days “maybe have the project work days earlier on in the semester to get people involved and it makes sure there’s a time where everyone can work together and focus.” Another student responded that placing more emphasis on how important documentation is to the “final phases of the project and assign someone to make sure documentation stays relevant and up to date.” A student suggested enforcing some kind of punishment for teams not completing assigned tasks. “The lack of a real incentive killed most motivation unless personally prodded.” Another responder was concerned about grades and the focus of the course: “I would suggest to make the primary focus of the class the project. Another thing I’d suggest is for us to see (how) the project is (affecting) our grade; I know I worked hard, but if (the project lead) or my teammates don’t, then I need to tell you.” A last concern was both with the project, and something that could have been included - a code review: “I wish I had tried a bit more at the beginning as in the end our project isn’t (as) fully complete as I would like...Otherwise the project just needs better documentation which you could have possibly asked for throughout development (code walkthrough).” The students found that the project days they were given toward the end of the semester were invaluable as well as an increased amount of time to communicate with the project lead. The face-to-face communication seems to have been better received by all than having students communicate via discord server whenever they had the opportunity. The project days were added after midterm, based on student

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

119

concerns about completing the term project. Because our Introduction to Software Engineering course is a survey course, critical parts of the theory were selected, instead of continuing to follow the given textbook.

the level of effort necessary which I enjoyed as it helped prepare for real-life applications better. The chaos at the beginning with the project leadership as it caused a decent amount of confusion.”

5.4

5.5

Likes, Dislikes, Did Students Benefit?

The last question on the final asked students about their likes and dislikes on the project, as well as if they benefited from time spent on the project, and if other students taking the course in the future would also benefit. Nine of ten students responded positively, in that they felt like they benefited from working on the project and could also see it as a benefit for capstone students needing help in completing a project. Some of the other benefits the students saw included real-life communication issues, experiencing chaos, working on a large-scale project for (probably) a student’s first time, and getting a taste of what they might see once gainfully employed in the profession. The one student who didn’t feel they benefited had issues involving the Swift programming language. This was the only positive for this student, in the end - picking up knowledge of an additional programming language. One student opined, “I liked the challenge it presented and the experience I gained through working on it. I didn’t like that it was only targeted towards Apple products only because it made it more difficult to build and test with limited resources...I think other students would benefit by testing and reviewing code.” A second student said that a positive was “being able to be a team leader and coordinate people based on their skill level,” while a negative was “people not understanding what (they) can/can’t help.” Another student replied, “Getting to work on an actual, applicable piece of software rather than learning about one that someone else made. I had no prior knowledge of the project material.” Another response: “I liked how unique and beneficial the project was to bringing together all of our knowledge from both previous and current courses. I disliked how, at first, there could be major communication issues delaying the project. This was by far the most effective learning experience for my major I had and I think it could be for others as well.” Another response: “I greatly enjoyed the project, it was challenging, yet rewarding...I got to see what’s in my future here, put something extremely cool on my resume, and had [some] fun on a collaborative project [unheard of].” The next student response: “I enjoyed how it was an actual tangible thing that you can say you built and not just another program that’s been done for years. My least favorite part was definitely the Gitflow, it’s overly complex.” The next student said “What I liked most was that I learned how to work in a team and how important it is to communicate with each other. What I liked least was that we can’t put the app on the App Store. It’s like our efforts meant nothing. I’m not sure about other students but I feel like they feel (the) same way as me. This was a cool project.” Another student was mostly positive, saying, “(What) I liked most was learning how to work on a large team. I have only worked in pairs until this class. I least liked working with Apple products. I wanted more say in the platform because I have never used (A)pple products...I learned so much about software development.” The last response: “The project showed

Project Lead’s Thoughts

The project lead was asked to write a journal of thoughts over the course of the semester about working with the survey course students. His initial thoughts included that the “capstone project is shaping up to be the most memorable, stressful, and rewarding experience of my college years. Although at times it’s been overwhelming; watching the individual software engineering phases merge together has been a learning opportunity that I wasn’t initially expecting. I believe both myself and the SWE class have grown as software engineers and computer scientists as a direct result of this project.” The project lead’s thoughts about the overall scope of the project: “In my opinion, this is the most real-world project I’ve been a part of. Although they usually result in a working application, the different aspects of software development are usually taught over separate courses and less emphasis is placed on anything outside of that course’s domain. The capstone project demands that equal attention is paid to each individual part or things will begin breaking down quickly. As a result, the benefits of a disciplined, systematic approach are more apparent.” Even though everyone seemed to feel great about this project as a learning experience at the end, there were struggles: “As a learning experience, this project has been amazing. However, as a course that I’m ultimately graded on and responsible for, it’s been extremely stressful. I think any group projects developed in college will have to contend with not everyone doing their fair share, but this one has been exceptionally poor in that regard. For the first (two-thirds) of the project, I’ve felt like my only options are to either do something myself or accept that it probably won’t get done. As we began making progress with the application, the morale and initiative picked up quite a bit and students were eager to solve the next problem or implement the next feature.” The project lead also noticed that the tools utilized for the project were also a struggle: “One of the biggest technical issues that held us back during development was familiarity and knowledge of working with Git. Ensuring we were developing on the proper branch, with the most recent code base, was something that didn’t seem to be understood. I attempted to write up a detailed explanation of Git, the GitFlow workflow, and how it all fits together, but it still didn’t lead to widespread understanding. Ultimately, a few individuals in the development team learned how to use Git, and the individual teams always use one of those accounts to initiate pull requests or commit any code. As I believe that Git is absolutely crucial to developing real-world applications with non-centralized team members, it would be great if more emphasis was placed on teaching students how to properly use source control throughout their tech-related college courses.” Swift was also cited as a major issue, from students in the survey course to learn a new programming language together with the topics covered in the survey course.

ISBN: 1-60132-498-7, CSREA Press ©

120

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 | Suggestions for improvement included more guidance from the supervising faculty. The lead notes, “It would’ve been helpful to have someone in a senior leadership role saying, ‘I need the design documentation on that specific class by next Friday’ or ‘be prepared for a technical code review by the end of the month.’ It may not even be an issue with a capstone class consisting of more members (each able to focus on something different) but being by myself I think a little more guidance would’ve helped nudge me along in the right direction.’ The student does note that this project was a valuable learning experience: “All in all, this experience will stick with me throughout my professional career. Just working with a larger development group helped me realize what real-world software engineering must be like. I would like to thank (the supervising faculty) for trying something new with the capstone project, as there’s no way I could’ve done any of this without the help of the SWE class.”

6. FUTURE WORK Supervising faculty hope to return to this project, either in the form of another capstone project or in some other software engineering-related course. The only functionality completed for the capstone was getting users from one point on campus to some other desired point. The suggestions of a digital bulletin board for each department, as well as achieving step goals through suggested routes could be added in a future capstone iteration. The continuation of this project also introduces a concept that students are rarely exposed to, beyond theoretical discussion: software maintenance. Students do not realize that the “glory job,” so to speak, of creating a new product from scratch rarely happens; when students start work at their first job, typically they will maintain a piece of software. Other courses’ content could be exercised as well, such as testing, and the many different types of testing to ensure the product’s correctness. When other students inherit this project’s code, much of what is taught to them early on in programming about readability would also be addressed. Students may not realize that having meaningful comments, if any comments, is crucial because someone else might pick up work on a piece of software after them. In turn, other items such as writing clear code and adopting meaningful names would be reinforced at this point. A capstone project of maintaining this product could be beneficial to students because all of the prerequisite courses could still be hit, but perhaps with a different emphasis (for example, requirements specification for new functionality and testing strategies for exercising the new functionality, as well as any corrective maintenance). A process would need to be employed, and one better than the semi-agile, Scrum-route that was used for this particular project. Additionally, supervising faculty will need to take a more proactive role, suggesting tasks to complete, such as some type of walkthrough or inspection, and maybe even acting as a de facto change review board. One of the other points supervising faculty should consider if they are utilizing and picking up this same project is an expansion into Android

development. Android users also maintain a significant user base for mobile devices. Students might even see the challenges of making the same app look the same on two, totally different platforms.

7. CONCLUSIONS This paper has documented a year-long project involving capstone students. The focus of this paper is to discuss the results of taking a senior project and providing students with their own shop, consisting of those who are enrolled in the survey Software Engineering course. The major caveat in the second semester was communication, and resolution was attempted through more face-to-face communication with the project leads. The students were able to successfully deliver a working prototype, and most of the survey course students felt like this project benefited them.

8. REFERENCES [1] J. Beck. Fair division as a means of apportioning software engineering class projects. In Proceedings of the 39th SIGCSE technical symposium on Computer science education, pages 68–71. ACM, March 2008. [2] J. Beck, V. L. Almstrum, H. J. Ellis, and M. Towhidnejad. Best practices in software engineering class management. In Proceedings of the 43rd ACM technical symposium on Computer Science Education, pages 201–202. ACM, March 2012. [3] K. Buffardi, C. Robb, and D. Rahn. Learning agile with tech startup software engineering projects. In Proceedings of the 2017 ACM Conference on Innovation and Technology in Computer Science Education, pages 28–33. ACM, June 2017. [4] A. Chidanandan, L. Russell-Dag, C. Laxer, and R. Ayfer. In their words: student feedback on an international project collaboration. In Proceedings of the 41st ACM technical symposium on Computer science education, pages 534–538. ACM, March 2010. [5] A. R. J. Francois. Class-wide projects: fostering collaboration and creativity in computer science courses. In Proceedings of the seventh ACM conference on Creativity and cognition, pages 369–370. ACM, October 2009. [6] D. E. Krutz, S. A. Malachowsky, and T. Reichlmayr. Using a real world project in a software testing course. In Proceedings of the 45th ACM technical symposium on Computer science education, pages 49–54. ACM, March 2014. [7] C. Matthies, K. R. Thomas Kowark, M. Uflacker, and H. Plattner. How surveys, tutors, and software help to assess scrum adoption in a classroom software engineering project. In Proceedings of the 38th International Conference on Software Engineering Companion, pages 183–187. ACM, May 2016. [8] M. A. V. Nelson, R. V. Carneiro, and M. R. Costa. Interdisciplinary software projects as an active methodology to practice for the profession. In Proceedings of the 1st International Workshop on Software Engineering Curricula for Millenials, pages 49–54. IEEE, May 2017.

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

SESSION COMPUTER SCIENCE AND COMPUTER PROGRAMMING, SOFTWARE ENGINEERING, AND WORK PRACTICES Chair(s) TBA

ISBN: 1-60132-498-7, CSREA Press ©

121

122

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

123

Coding VR Games J. Chen1, M. R. Zargham1, M. Rajendran1, and J. Cheng2 Department of Computer Science, University of Dayton, Dayton, Ohio, United States 2 Department of Computer and Information Technology, Miami University, Ohio, United States 1

Abstract - This paper represents the progress of our ongoing effort on a new way of teaching an introductory programming course while incorporating a game strategy. Students use a simple environment, called Innovative Coding (IC), to write a code for designing a game or a virtual world. As they code, they see the development of their world in a 3D environment step by step. This paper describes the addition of a new module that allows students to design a game in a breathtaking landscape. Before the game starts, the user (the student who will set the game) will write codes for a set of robots that prevent the player (the individual who will be playing the game) from reaching to the destination. The game starts when the player found himself/herself trapped in an island and needs to find a set of clues to reach the final destination. The robots will attack the player on the way to the destination. Throughout the game, the user will see the effectiveness of his/her code by the degree of the player’s success. Keywords: VR Headset; Coding; Game; Virtual Reality; Education; Programming.

1

Introduction

One of the most effective uses of VR is in the education field. Within the last three years, we have developed a software package, called Innovative coding (IC) that allows students to learn coding in an engaging environment. IC is intended to encourage more female students to pursue computer science as their major in college [1, 2]. IC allows students to write simple codes consisting of several statements to create a world of their choice in a 3D and Virtual Reality environment. VR goes beyond a simple visual stimulus, and allows students to become directly involved, experiencing the code they write in a very tangible, interactive, and expeditious scheme (strategy). This allows them to experience both the successes and failures of writing (creating, developing) correct code versus incorrect code. Students can appreciate both the scale and the scope of the results of their code immediately, which prompts them to become invested in the quality of their coding skills and strive to continually improve their logic, programming, and code writing abilities.

Through this Immersive and Responsive Visual Stimulus Learning, students become more engaged in coding. In general, the idea is to teach the introductory programming course in such a way that follows the projectdriven learning process and encourages students to develop problem-solving and teamwork skills while fostering creativity and logic. This year, we were able to extend the functionality of IC by adding a game development environment. Our goal of designing the game is to help 9- to 13-year-old students to learn coding and problem-solving in an engaging environment. They can see the progress and result of their code in the 3-D environment. Different types of interaction and excitement are possible through different code complexity of the main objects in the game. By following step-by-step instructions and observing the game outcome, students can learn how to develop, test, run and debug a computer program. They learn a disciplined and structured approach to their program development. They will also learn main control structures of procedural programming languages, becoming familiar with the concepts of loops, assignment statements, and condition statements while others are playing their engaging and interactive game. Besides 9 to 13 years old, students in K-12, computer science, electrical/computer engineering, and business, students from other disciplines such as mathematics, physics, chemistry, biology, communication, and arts are also encouraged to utilize IC for learning how to code.

2

Game Environment

There are four islands with different climates: Japanese Garden, Volcano, Athens ruins, and Ice islands. These four islands shown in Figure 1 are connected with bridges allowing the player to travel from one island to the other.

ISBN: 1-60132-498-7, CSREA Press ©

124

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

Figure 1: Overview

Figure 3: Fountain and Fishes

After the volcano island, players will arrive at the Athens Ruins, see Figure 5. The Athens ruins demonstrate a vivid picture of how impressive the ancient Athens once was and now is covered by a tropical forest. Small bricks are the foundation of an ancient temple; stand-alone columns are all the remnants of a building built over 2,000 years ago.

Figure 2: Japanese Garden Red Bridge

Figures 2 to 4 represents the Japanese Garden which is calming and growing. This island is built around an ancient Japanese tradition with Bonsai trees, water, stone, the lantern, and bridges. Volcano Island, shown in Figures 5 to 6, is located on the north side of the Japanese Garden. This island has a sharp contrast with the Japanese Garden. It is an active volcano forming a series of boiling lava flows. The eruption is throwing lava into the sky from the top of the mountain, and molten rock pouring from the vent. The entire island surrounding with gas, heat and is covered by ashes.

Figure 4: Japanese Garden cherry blossom & Tori

Figure 5: Erupting volcano

ISBN: 1-60132-498-7, CSREA Press ©

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

125

3

Game Strategy

Before the game starts, the user can place robots, called enemy robots, anywhere in the landscape and give them life through action commands within the command window. Also, the robots possess intelligence based on the user code that allows them to recognize the player within a specific range and destroy her/him through their personalized (program) attacks. During the game, the player starts from a point in the Japanese Garden and explore the island to find her/his first hint. The hint will be marked with a balloon that floats up and down in the air shown in Figure 9. Figure 6: Lava flow

After the Athens Ruins, Ice Island is the final destination place. Players will experience the endless and mind-blowing snow-capped mountain. Figure 8, represents a deep ice cave and tunnel in the ice mountain where players enable to explore the incredible icy land from the virtual world.

When the player locates the first hint, he/she will shoot the balloon to activate a hint window. The window will represent a programming question to the player. The player needs to answer the question. If her/his answer is correct, the next hint is activated and the player directed toward a shortcut. Otherwise, the player is led toward a complicated path for reaching the next hint. The hints are located in sequential order, one on each island. Therefore, the player will go through all four islands in order while enjoying different perspectives of different climates. On her/his journey to find the hints, the player might come across enemy robots that have been set up and programmed with the user before the game starts. If the player gets into the robots premise, the robot will follow her/him and starts attacking according to the command. At this time, players can either run away to keep far from danger or shoot the enemy robot with a laser gun and earn bonus points.

Figure 7: Athens ruins

Figure 9: Hint balloon

Figure 8: Ice Cave

The end point is located somewhere on the ice island. When the player finds the last clue, he/she celebrates his/her victory by saving the scores. Among different players, the winner will be the one with the highest points in the rank. Points are calculated based on the time consuming to finish

ISBN: 1-60132-498-7, CSREA Press ©

126

Int'l Conf. Frontiers in Education: CS and CE | FECS'19 |

the game, enemy killing level, and question-answer sessions provided by the hint window.

4

//control movement of objects walk float direction float;

Programming //Detect enemy in a certain region

The syntax of our programming language is similar to Java. The language supports statements such as variable declaration and initialization, assignments, loops, condition, arrays, and functions.

findEnemy(distance);

//attack number of seconds

There are two types of loops, “for” and “while” statements defined as below: for (initiation, condition, increment/decrement) { Statement1;

attack int; The user can utilize the above statements to write a controller for each robot. For example, the following code represents a simple controller for an enemy robot.

Statement2; while(1