Religious Identity in U.S. Politics 9781626378162

Citation preview

Religious Identity in US Politics

Religious Identity in

US Politics Matthew R. Miles

b o u l d e r l o n d o n

Published in the United States of America in 2019 by Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.rienner.com

and in the United Kingdom by Lynne Rienner Publishers, Inc. Gray’s Inn House, 127 Clerkenwell Road, London EC1 5DB

© 2019 by Lynne Rienner Publishers, Inc. All rights reserved Library of Congress Cataloging-in-Publication Data Names: Miles, Matthew R., 1976– author. Title: Religious identity in US politics / Matthew R. Miles. Description: Boulder, Colorado : Lynne Rienner Publishers, Inc., 2019. | Includes bibliographical references and index. Identifiers: LCCN 2019000493 (print) | LCCN 2019017630 (ebook) | ISBN 9781626378162 (e-book) | ISBN 9781626378094 (hardcover : alk. paper) Subjects: LCSH: United States—Religion. | Religion and politics—United States. | Identification (Religion) | Identity (Psychology)—Religious aspects. Classification: LCC BL2525 (ebook) | LCC BL2525 .M55 2019 (print) | DDC 322/.10973—dc23 LC record available at https://lccn.loc.gov/2019000493

British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library.

Printed and bound in the United States of America

The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992. 5 4 3 2 1

Contents

List of Tables and Figures Acknowledgments 1 2 3 4 5 6 7 8 9

vii ix

The Role of Religious Identity in US Politics

1

Defining Religious Identity

15

Religious Identity and Political Trust: The Obama Era

51

Religious Identity and Support for Members of Congress Christian Identity and Anti-Atheist Intolerance The Religious Divide

When Religious and Partisan Social Identities Collide How Political Identities Influence Religious Beliefs Is Reconciliation Possible?

Appendixes 1. A Note on Interviews Conducted 2. The National Surveys 3. Descriptive Statistics and Complete Regression Models for Chapter 3 4. Complete Regression Models for Chapter 4 v

35 63 85 99

113

131 141 143 149 155

vi

Contents

5. Complete Regression Models for Chapter 5 6. Complete Regression Models for Chapter 6 7. Complete Regression Models for Chapter 7 8. Complete Regression Models for Chapter 8 References Index About the Book

159 163 167 171 173 185 190

Tables and Figures

2.1

How Threats Influence Individual Identification with the Group Collision of Religious and Partisan Social Identities Matching House Members’ Religious Affiliations with Those of Their Constituents Strength of Religious Identity, by Partisanship, 2013 Survey Strength of Religious Identity, by Partisanship, 2014 Survey Influence of Collision of Religious and Partisan Social Identities on Identity Strength Influence of Collision of Religious and Partisan Social Identities on Attitudes Intermountain LDS Faith Experiment Text Average Effect of Religious and Partisan Attitude Collision, ILF Experiment

Tables 2.2 3.1 5.1 5.2 7.1 8.1

8.2 8.3

3.1

Importance of Religion and Influence of Shared Religious Identity on Congressional Approval Ratings Importance of Religion and Influence of Shared Religious Identity on Presidential Approval Ratings Predicted Trust in President Obama, by Shared Religious Identity and Race Model-Predicted Trust in President Obama, by Shared Religious Identity and the Importance of Religion

24 30

41 69 70

103 117 119

120

Figures 3.2

4.1

4.2

vii

44

46

56 57

viii 4.3 5.1

5.2 5.3 5.4 5.5 5.6 5.7 6.1 6.2 6.3 6.4 7.1 7.2

7.3

8.1 8.2 8.3 8.4

8.5 9.1

9.2

Tables and Figures

Model-Predicted Trust in President Obama, by Partisan Identity and Shared Religious Identity Attitudes Toward Democrat-Affiliated Groups, by Religious Social Identity, 2013 Survey Attitudes Toward Republican-Affiliated Groups, by Religious Social Identity, 2013 Survey Attitudes Toward Democrat-Affiliated Groups, by Religious Social Identity, 2014 Survey Attitudes Toward Republican-Affiliated Groups, by Religious Social Identity, 2014 Survey Average Treatment Effects, 2013 Survey Religious Bias Toward Minority Groups, 2013 Survey The Influence of Religious Social Identity on Willingness to Help with Political Activities, 2014 Survey How Group Identification Influences Probability of Assistance with Semipolitical Activities, by Religious Identity, 2013 Survey How Group Identification Influences Probability of Assistance with Semipolitical Activities, by Religious Identity, 2014 Survey How Group Identification Influences Probability of Assistance with Nonpolitical Activities, by Religious Identity, 2013 Survey How Group Identification Influences Probability of Assistance with Nonpolitical Activities, by Religious Identity, 2014 Survey Graphic Depiction of Religious Groups, Rated by Partisanship Partisan Disidentification Among Strong and Weak Identifiers with Religion Partisan Disidentification Among Members of Unpopular Religious Groups Leniency Attitudes, by Experimental Group and Partisanship Hierarchy Attitudes, by Experimental Group and Partisanship Theological Views, by Experimental Group and Partisanship Test of Statistical Significance: LDS Church Should Be More Lenient Partisanship Determines Whether Behavior Is Sinful Party Identification, by Strength of Religious Identity, Pew 2007 Party Identification, by Strength of Religious Identity, Pew 2014

59

71 72 73

74 79 80

81

91 93 94 95

104

107

110 122 123 124 125 127

136

137

Acknowledgments

generosity of Donald Haider-Markel. As a scholar at the beginning of my career, I had an idea for an experiment that I thought might effectively isolate the influence of religious identity on atheist intolerance in the United States. Don was kind enough to allow me to include that survey experiment in a national survey he was already planning to field. A year later, and every subsequent year, he has graciously allowed me to include a survey experiment in his national surveys. Don is an intellectual mentor and a prolific scholar; his example inspires me to keep looking for answers. In 2014, a paper presenting some of the findings of this survey experiment was accepted for the annual American Political Science Association (APSA) conference. Although the early morning panel I was assigned to inspired fewer than a dozen people to attend, the late Ted Jelen served as discussant and Geoff Layman was in the audience. The affable manner in which they provided stern feedback helped me to see the flaws of my work while encouraging me to continue with the project. They were the first people I met who were members of the Politics and Religion section of APSA. Since then, I have become acquainted with a number of other members of the section who have been equally kind in encouraging my research. As editor of the journal Politics and Religion, Paul Djupe helped guide me through the literature as he nudged my research on descriptive representation toward publication. That research is included in Chapter 3 of this book. David Campbell has always been available to lend his expertise on politics and religion and shared data with me, which helped me resolve a puzzle. Ryan Claasen and Elizabeth Oldmixon read portions of the manuscript and provided constructive feedback.

This book would never have been possible without the

ix

x

Acknowledgments

Although he was busy with his own work, Jason Adkins read my entire manuscript and pointed out flaws in my reasoning and gaps in my understanding of previous research. Neal Carter has an outstanding editorial eye; thanks to his comments, this book is clearer and more concise, especially the portions related to social identity. The College of Faculty Development at Brigham Young University–Idaho has been generous in its support of this project, and I thank the leadership team, especially Sid Palmer and Andrea Radke-Moss, for their continuing support. Thanks as well to Michael Abel, Bradley Harper, Jeremy Lamoreaux, John Thomas, and Tyson Yost for their feedback. Lynne Rienner has been a patient editor, guiding this book through multiple drafts toward final publication. Without Lynne’s encouragement, I would never have written the book; without her constant support, it would not have been published. She selected a group of anonymous reviewers who provided invaluable feedback, to whom I am also indebted. Finally, I thank my daughter Melanie, who informed me that unless I wrote a book, I would never be a true author in her eyes. My wife, Sandra; our other children; and my mother, Becky, showered me with words of affirmation when I needed it most. You are the best!

1 The Role of Religious Identity in US Politics

pulled apart at the seams. Americans who might otherwise get along with each other are increasingly expressing hostility toward those who affiliate with an opposing political party. The alignment of religion with political identities seems to be causing Americans to distance themselves from each other even more—not just politically but socially.1 Americans sorted into political camps not only disagree on the issues but are angry with those on the opposing side and would prefer that their children not intermarry. Tensions abound. An alternative view is that scholars advocating this perspective and the media coverage highlighting it overstate the influence of religion on political and social polarization. The way religion is measured and conceptualized in the politics and religion literature makes it seem as if the division is larger, deeper, and more profound than it might really be. In addition, the treatment of religious social identities in the literature has led to several unresolved questions that a more complete understanding of it can begin to answer. For example, the Pew Research Center reports that between 2007 and 2014, the percentage of Americans that identify as religiously unaffiliated increased by 6 points. Among those Americans with a low commitment to religion, 72 percent identify as religiously unaffiliated.2 Why are Americans disaffiliating with organized religion? From a psychological social identity perspective, changes in the supply of religious denominations or one’s values need not influence one’s identification with a religion.3 Other factors might have a stronger influence on the development or maintenance of group identifications.4 Some argue that political backlash against the Christian right is contributing to the rise of religious disaffiliation in the United States. The

By some accounts, it appears as if the United States is being

1

2

Religious Identity in US Politics

reason both believing and nonbelieving Americans are much less likely to affiliate with an organized religion today than they were in previous decades can be found in two factors. First, Millennials are more individualistic and less trusting of religious, scientific, judicial, and political authorities. Second, progressives are less likely to affiliate with a religion. The unification of the religious right in US political life, combined with distrust of institutional authority among younger generations, causes people with progressive political leanings to become religious independents even when they retain their belief in God and a religiously moral worldview.5 It might be time to reevaluate how the believing, belonging, and behaving paradigm should be adapted to the study of politics and religion in the United States. How do people connect religion with politics if they increasingly worship outside of traditional congregations and do not affiliate with religious groups that conform to the dominant model? In addition, many categorize religious groups into broader coalitions. Some evidence suggests that religious groups espousing a more liberal theology are more likely to cooperate with each other politically than are religious groups that have a more conservative theological perspective.6 Indeed, those who identify with Protestant denominations group themselves into broader (white evangelical, black, white mainline, etc.) denominational families.7 Do members of non-Christian faiths do the same? Does political cooperation among religious groups cause people who identify with these religions to feel an affinity toward each other, despite considerable theological divergence? Do American Hindus have a more positive affect toward atheists who share a partisan group identification than they have toward Christian political opponents, despite a shared belief in a higher power? A related question concerns the role of religion in the rise of self-identified political independents in the United States. Samara Klar and Yanna Krupnikov argue that Americans are leaving political parties because the labels associated with the dominant parties in the United States embarrass them.8 Another possibility is that, as the two major political parties exploit religious divisions to build electoral coalitions, religious Americans choose not to identify with a political party that advocates political positions incongruent with their religious values. Geoffrey Layman and Christopher Weaver document the change in religious affiliation among delegates to the Democratic and Republican National Conventions from 1980 to 2012. In 1980, 40 percent of the delegates to the Republican convention were mainline Protestants; by 2012, only 18 percent of the delegates were mainline Protestant. The dominant religious affiliation at both the Republican and Democratic National Conventions in 2012 was “none”; 30 percent of the delegates to both conventions did not affiliate with any religion.9 Have major US political parties pushed out religious Americans?

The Role of Religious Identity in US Politics

3

We know a lot about how religion influences politics, but we know much less about how politics influences religion. People who attend church regularly are more likely to adopt political views consistent with those of the clergy because when people are in church they receive cues about political issues from both clergy and congregants.10 The message delivered from the pulpit and contact with others in the church help congregants connect their religious values with political attitudes.11 Religious identity can also have an independent influence on political attitudes. People might not fully understand all of the nuances of political issues, but they can easily distinguish between people who politically side with or against them.12 When members of people’s religions support a political position and members of other religions take the competing position, people rely on social group heuristics to develop their own views.13 To date, the religion and politics literature treats this phenomenon in a unidirectional manner. Yet, social identities are two-way streets: If religious identities influence political attitudes, do political identities influence religious attitudes? Either way, one must ask: Is religion good for society? Robert Putnam and David Campbell report several positive associations between religious belief, behavior, and political participation. People active in church communities are more generous, more likely to volunteer in their community, and more politically involved.14 However, these authors also note a growing political divide between those who believe in God and those who do not: the God gap in US society. James Gibson argues that religious beliefs are the primary cause of anti-atheist political intolerance.15 From this perspective, political and social divisions are rooted in differing religious beliefs. What is the source of religious political intolerance: beliefs, values, or identities? The answer to this question is important. If anti-atheist attitudes are caused by diverging beliefs or values, it is unlikely that the God gap will be bridged in a single generation; however, if something else is causing this divide, it might be more easily overcome. In this book, I aim for greater conceptual clarity regarding what it means to identify with a religious group. Religious social identification is more than group membership or affiliation; when someone identifies with a religious group, it becomes an extension of the self. Greater clarity on this concept helps explain why people identify with religious groups even when society has a negative view of them. Although this book does not provide a complete answer to the questions posed, it addresses each of them and provides a framework that can be used in future scholarship on these questions. Although it is possible that religious beliefs, behaviors, or social networks are the foundation of the religious divide in the United States today, social identity theory suggests that the strength of the group identification matters more than the beliefs that underlie that identification. Beliefs might motivate stronger attachment to a religious group, but they

4

Religious Identity in US Politics

are not necessary for religious social identifications to persist. As I will demonstrate, many people retain a sense of religious identification even when they no longer believe the teachings of the religious group. In addition, group identifications are more malleable than beliefs and values. To the extent that divisions in US society are motivated by religious group identification, the path to greater unity might be simpler than if it required people to change their values and beliefs. Before I explain why this is so, it is important to understand a bit about how people juggle the multiple identities that become extensions of how they view themselves. If thinking about religions in terms of social identity is a chair, the first leg of that chair is that social identities are context dependent. People who no longer believe, belong, or behave consistently with a religion can retain an identity with that religion that becomes salient depending on the context. Much of the empirical evidence for partisan and religious division in US society is based on survey data. This limits inference beyond a single point in time in which the survey is being conducted. As such, a survey measures only the influence of identities that are salient while the survey is being conducted. Furthermore, it is erroneous to conclude that attitudes expressed in a survey (even strong attitudes) last moments beyond the time in which they are registered in a survey. When the context changes, new identities become salient, and we should expect them to have their own independent influence on attitudes and behaviors. The following vignettes illustrate how context changes the relevance of various social identities. Paul is a gay man living in urban Washington, DC, with his husband of many years.16 They wed before Obergefell v. Hodges required each state to permit same-sex marriages. Their neighborhood is diverse. Some of their neighbors own condominiums worth millions of dollars, whereas others receive government housing benefits. Although Paul travels across the globe for work, he is a patriotic American and faithfully votes in virtually every election. Paul is also complex. Like most people, he does not have uniform political views.17 His sexual identity informs his political views, but he is a different person today than when younger. Paul was born a seventh-generation member of the Church of Jesus Christ of Latter-day Saints (LDS) in Texas. Paul began elementary school in a small town in rural Ohio, where he spent the remainder of his childhood until he graduated from high school. At nineteen, he went on mission for two years in Ukraine, converting people to the faith of his youth. When he returned to the United States, his parents were again living in rural, conservative Texas. His parents raised him with very conservative religious and political views, and he spent his young adult

Paul

The Role of Religious Identity in US Politics

5

years choosing among Republican candidates in Texas elections. When Paul moved to Washington, DC, many years later, he realized that the real elections took place within the Democratic primary, so he decided to register as a Democrat for the first time in his life. He told me that registering as a Democrat was not easy because of his upbringing. Paul spent eight years in England attending graduate school and preparing for his career. His work as a writer for a prominent magazine, author of several books, and a television personality has led him to 120 different countries and almost every corner of the United States. Regarding these experiences, Paul noticed that although he is the same person in every new situation, the circumstances specific to each situation tend to change the way he presents himself and how others view him. For example, when he is in Japan and Africa, this tall white man cannot hide that he is a foreigner. The biological distinction between him and the local population makes it difficult to conceal his national identity. This motivates him to represent his country well. In this context, Paul reports feeling a strong US national identity. By contrast, when Paul is in northern European countries, he finds that his biological similarity to the local population makes it easier for him to mask his national identity. When he is with Europeans who dislike some aspect of the United States, his national pride motivates him to acknowledge that he is American, but the circumstances incline him to remind his hosts how dissimilar he is to the stereotypical American. For Paul, social situations in the United States cause his other identities to become salient. Because most people in the United States share his national identity, Paul’s other group identifications take on greater meaning. Religiously, Paul is no longer a practicing member of the LDS Church, and he does not regularly attend worship services, but he strongly identifies with that church; it is an integral part of his culture, upbringing, and worldview. Although he might be openly critical of LDS Church teachings and the behavior of its leaders, he gets upset when outsiders attack his church. Paul identifies as a gay man and supports policies that prohibit discrimination based on sexual orientation. This often places him at odds with the LDS Church. When he was younger, Paul was encouraged to attend workshops to help him overcome his same-gender attraction. To this day, Paul is personally offended that the LDS Church uses the word homosexuality as an adjective rather than as a noun. Paul is gay, but the LDS Church describes him as someone who “has same sex attraction.”18 In addition, the LDS Church has openly opposed same-sex marriage and encouraged some of its members to become politically involved in supporting legislation banning the practice. In 2013, more than 80 percent of American lesbians and gays reported that they feel like the LDS Church is “anti-gay.” 19 This creates tension between two important components of Paul’s identity. Yet others find comfort in both their LDS and gay identities, suggesting that the

6

Religious Identity in US Politics

two do not perpetually conflict. Most social situations do not create conflict between Paul’s LDS and gay identities. In fact, some situations reinforce the two. The first time he saw members of the LDS Church marching in support of a gay pride parade, Paul cried because he felt unity between these two identities for the first time in his life. The conflict between Paul’s LDS and gay identities is not substantially different from that of his geographical identities. As a Midwesterner, Paul gets upset when people who live on the coasts refer to his hometown as “fly-over country.” He has a strong affinity for the authenticity, work ethic, and honesty of people who live in the Midwest. He tries to reflect these traits when he associates with his friends from major metropolitan areas. Likewise, he recognizes that his time abroad has changed him relative to his peers who remained in the Midwest. His travels through small midwestern towns heighten his awareness of how different he is from those who live there. Yet, Paul places greater value on some of his identities than others. He might choose to put on the midwestern conservative hat when he is defending his family from the attacks of his urban liberal friends, but he feels a greater affinity toward his current identity than he does to some of those from his youth. There might be cause for him to defend the LDS Church, but if a situation made it impossible for him to retain both his LDS and gay identities, Paul would not be able to abandon his gay identity; it is too strong a part of how he sees himself. Paul is not an anomaly. All people have a different combination of identities that define how they see themselves, but these identities can create either congruence or conflict in different social settings. Martin was raised as a nonbeliever in a small rural community of devout Lutherans and Catholics in Oregon. It was sometimes hard for him to go to school as a young man because most people there were religious and he felt left out of that religious community. Martin immersed himself in a group of likeminded nonbelievers who shared common philosophical, political, and musical interests. At that time, Martin did not feel close to Christians living in his community, and he attributed differences in attitudes on political issues to the Christians’ ignorance or intolerance. Not long ago, Martin converted to a new faith. For three years after he was baptized in the LDS Church, Martin resisted becoming deeply involved in the LDS community. He wanted to figure out the answers to his questions on his own, without too much interference from people active in the LDS Church he attended. Then he decided it was time to figure out if he really wanted to be a Mormon. He decided to move to another state and transfer to the LDS-sponsored uni-

Martin

The Role of Religious Identity in US Politics

7

versity. In part, because he has had faith-promoting experiences and because he is now in a community where his religion is the societal ingroup, he now says that he has a strong LDS religious identity. In fact, his LDS religious identity is currently at its peak, and although he was once a nonbeliever, he currently has strong negative attitudes about nonbelievers in the United States. Susie was raised Lutheran. While she was growing up, religion was an integral part of her life. The foods she eats, how she relates to others, her culture, and her family traditions are all tied into her Lutheran identity. Being Lutheran is in her blood. However, Susie no longer attends the Lutheran Church. She started incorporating Reiki into her worldview before she ever left the Lutheran Church, but she did not begin openly practicing Reiki until a couple of years ago. Today, Susie regularly attends seminars and Reiki groups in which members discuss their experiences with Reiki. Susie considers herself spiritual but not religious, and she is active in her Reiki community. She says that today she does nothing to hide her nonreligious beliefs from others. When her only sister died suddenly several years ago, Susie realized that the number of people from her youth that are still alive is rapidly shrinking, which gave her the impetus to begin being completely open about who she is and what she believes. Here is how she describes her current religious identity: “I mean, it’s like, I’m sixty years old now, and I have that I’m Reiki on my Facebook profile, and if people want to out me I’ll be very happy to tell them: no, I don’t believe in organized religion, and I think it’s a bunch of nonsense made up by man to keep people in line, and no, I’m not going. Enjoy your church, but I won’t be there.”

Susie

Cody (a Democrat) was raised in rural Oklahoma. Dreaming of becoming a world traveler and author, he left his home and attended college in New Orleans and Canada. Eventually, he returned to Oklahoma to attend graduate school and met a girl from Split, Croatia. In time, they fell in love and married. When I first met Cody, he was in Kansas with his Croatian wife, learning advanced Croatian. He often told stories about teaching English in Turkey and visiting his in-laws in Croatia. After completing his studies, Cody moved with his family to Croatia and currently writes and teaches at the University of Zagreb.20 Cody identifies as a Midwesterner, an American, an expatriate, an intellectual, an author, and at times, a Croat. Social circumstances largely determine how strongly each of these identities are manifest and Cody’s response when one of these identities is threatened.

Cody

8

Religious Identity in US Politics

Julie (a Democrat) grew up in a small rural community in Minnesota. Her parents are lifelong Republicans. They are Protestant, evangelical Christians. She studied as an undergraduate at Bethel University in St. Paul, Minnesota, and later earned a master’s degree in theological studies from there. She continued her studies in the Midwest and eventually earned a doctorate degree. Today, she is an evangelical Protestant Christian, currently attending the Nazarene Church, and pushes her evangelical friends to “work” and “think harder” when it comes to politics. Like those described in these vignettes, everyone has multiple social groups with which he or she identifies. Some people’s identities change depending on life circumstances, whereas others do not. Some people have experienced rather large shifts in their identities, whereas others’ identities have not changed much. Context can largely determine both the salience of an identity and how strongly that identity influences individual responses.

Julie

As these examples demonstrate, religious identity is an important component of how people see themselves and others. In addition, religious identities influence people’s social interactions with others, how they perceive others, and their attitudes. To date, the literature on religion and politics has not ignored the importance of religious identity, but considerable conceptual confusion remains, which I hope to clarify in this book. For example, it is now common to hear about how tribalism in US politics leads to biased thinking.21 From this perspective, much of the animosity between partisan Americans is rooted in tribal instincts that pit one group against another in the competition for scarce resources.22 I contend, however, that clear conceptual understanding of the mechanism motivating attitudes and behaviors is a necessary precondition to appropriately diagnosing the problems and developing recommended solutions. Much of what Amy Chua describes as antecedents to tribal instincts are actually consistent with predictions social psychologists have been making for at least sixty-five years. Since the “Robber’s Cave” studies were published in the 1950s, we have known that when groups compete for scarce resources, it creates strong antipathy toward those who identify with the other group and more positive attitudes about those who belong to the same group.23 These studies also found that when members of competing groups are forced to interact with and cooperate with those from the other group, much of the hostility dissipates. Similarly, scholars of religion and politics have long noted the importance of religious affiliation in shaping political attitudes, but too often they conflate affiliation with religious identity. Until recently, the influence of

Religious Social Identity

The Role of Religious Identity in US Politics

9

religious social identity on political beliefs and behaviors has been subsumed within the believing, belonging, and behaving paradigm. On the one hand, religious social identity has elements in common with each of these three Bs. As the preceding vignettes demonstrate, belief is often a precondition of developing a religious identity. Belonging to a religious community and following the prescribed code of conduct often strengthen people’s religious identity. On the other hand, religious social identity is conceptually distinct from each of the three Bs. People retain strong identification with religious groups long after they stop believing in or living in accordance with the teachings of a given faith. In certain contexts, religious identity can still influence the attitudes of individuals who no longer believe, belong, or behave. Think of Susie as an example. Lutheranism is still an important part of who she is, and when she is placed in a context that makes that identity salient, it influences her attitudes. Furthermore, appropriately conceptualizing religious identity is important for understanding the best prescriptions for resolving conflict in society. If the religious divide in the United States is rooted in beliefs, tribal instincts, or intolerant dispositions, bridging the divide seems an arduous task.24 It is difficult to change strongly held beliefs. However, if the divide is rooted in social identities, it might be much easier to bridge the gap between religious and nonreligious Americans. After all, hundreds of studies demonstrate that increased contact between members of opposing social groups can reduce prejudice relatively quickly.25 If so, the God gap might be bridged simply through increased contact between religious and nonreligious Americans. One of the best ways to reduce societal conflict and prejudice is to break down the psychological barriers social identities construct.26 In this book I argue that religious social identity is an important consideration in future scholarship on religion and politics. Appropriate conceptualization and measurement of religious social identity—distinct from affiliation— clarifies another aspect of religion’s influence on the human experience. I develop and test hypotheses and report findings consistent with this argument. In Chapter 3 I affirm that religious identity has a party-independent influence on public attitudes about elected officials. People who share a religious identity with their elected officials are more likely to approve of them even if they have an opposing partisan identification. In Chapter 4 I discuss political trust in the United States. Although some worry that polarization is causing Americans to lose trust in their political leaders and institutions, in this chapter I demonstrate that shared religious identity motivated greater trust in President Barack Obama. The relationship between religious identity and trust in Obama was stronger than gender and as strong as race.

Outline of the Book

10

Religious Identity in US Politics

In this book, I separate religious social identification from other relevant, salient, social identities in US politics. This is not a small feat. As religious groups have aligned themselves with the two major political parties in the United States, it has become difficult for scholars to distinguish partisan effects from religious effects.27 Jonathan Haidt describes religion as a moral exoskeleton. People who live in religious communities are enmeshed in a set of norms, relationships, and institutions that create “intuitions” that motivate subsequent reasoning and behavior.28 People attribute intangible traits and qualities to those who profess belief in God. Religion evolved to facilitate cooperation in large groups because it deterred freeloading by informing group members that an omniscient overseer was monitoring and punishing poor behavior within the group.29 When people believe that a punitive and moralistic god knows their thoughts and behaviors, they are more likely to behave according to the norms and traditions of the social group. In fact, even when members of the group are not in close physical proximity to each other, their shared belief in the same deity motivates people to behave impartially toward each other.30 People are more likely to trust those who share religious beliefs because affiliation with a religion confers intangible qualities upon the adherents that are absent in nonbelievers.31 This might be why religious people distrust atheists: people who do not profess a belief in a higher power are assumed to operate outside of this evolved system of cooperation.32 I think antipathy toward atheists is much simpler. Again, using partisan social identity as an analogy, I argue that what appears to be strong antipathy toward atheists is really the influence of religious social identity. Partisan antipathy motivates strong negative attitudes toward members of the opposing political party among strong partisans. People with strong partisan identities are much less likely to trust opposing partisans or think they might be qualified for a job. They are also more likely to be upset if their child were to marry an opposing partisan.33 In Chapter 5 I demonstrate that much of what the extant literature calls religious hostility is really identitybased antipathy rooted in the makeup of current party coalitions. As the coalitions aligned with the major parties in the United States change, so will the perceived differences between evangelical Christians and atheists. Overall, I provide clear evidence that some religious intolerance toward atheists is rooted in political competition. Isolating the influence of partisan and religious identities allows me to demonstrate that as religious and nonreligious people in the United States have aligned themselves with opposing political parties, attitudes about members of groups who affiliate with these two political parties have polarized. Party competition causes members of opposing parties to be less trusting of cross-party members and even to feel antipathy toward opposing group members. When religious groups become involved in political competition, some of that antipartisan antipathy spills

The Role of Religious Identity in US Politics

11

over into attitudes about religion, but it is not the only source of aversion between religious groups.34 Individuals with strong religious but weak partisan identities are influenced more strongly by religious identities than they are by partisan identities. In Chapter 5 I move the dialogue further by isolating religious identity from partisan identities to measure the extent to which anti-atheist bias is motivated by religious rather than partisan identities. I find that anti-atheist attitudes are motivated both by partisan identities and by religious identities. People with a strong religious identity demonstrate out-group antipathy toward atheists that is stronger than and separate from partisan outgroup antipathy. This offers a competing explanation for the religious divide in the United States. The distinction is important. If religious Americans distrust atheists because they think nonbelievers are untrustworthy and have therefore developed intuitional caution against atheists, it could take years to bridge the divide between religious and nonreligious Americans and might not occur in a single generation. If, however, the religious divide is motivated by religious identity, the divide can be bridged more quickly through increased contact between believers and nonbelievers. Increased contact often overcomes divisions caused by competing social groups.35 In Chapter 6 I present additional compelling evidence that the strength of religious social identity is the primary factor influencing the way religious people interact with atheists in the United States. One of the major themes of this book is that partisan affiliations can threaten religious social identities. In Chapter 7 I demonstrate that when people with strong partisan and religious identification are informed that members of their political party have negative attitudes about people who affiliate with their religion, those with strong religious but weaker partisan identities will disidentify with their political party. In contrast, those with weak religious but strong partisan identities become stronger partisans. Consistent with theoretical expectations, the strength of one identity relative to the other determines the response. Finally, in this book I present the first evidence of partisan identities influencing attitudes about specific religious doctrines. Identity really is a two-way street. People with strong partisan identities are more likely to think their religions should change their teachings to better align with the dominant view of fellow partisans. In Chapter 8 I look at a specific case study within the LDS Church. It is no secret that in 2008, the LDS Church took a public stance opposing same-sex marriage in California. This created dissonance for LDS Democrats who favored same-sex marriage as a political issue but also had a strong LDS religious identity. This case study demonstrates that one way members of that church resolved this attitudinal dissonance was to express a more lenient position on LDS theology. Scholars have long noted that religion influences politics, but in this chapter I

12

Religious Identity in US Politics

present evidence of political identities influencing religious attitudes. The causal arrow logically moves in both directions. If religious identities can influence political attitudes, political identities ought to be able to influence religious attitudes. Recent scholarship notes how political identities can influence religious behaviors; in this book I present the first evidence of political identities influencing religious beliefs.36 I assess the evidence supporting these arguments using several surveys of the US population collected from June 2013 to June 2016. I also include data from a two-wave, representative survey comprising a unique sample of active, less active, and former members of the LDS Church residing in Arizona, Idaho, and Utah and the 2010 Cooperative Congressional Election Survey.37 After further developing and testing hypotheses based on these general expectations, I offer some concluding thoughts in the final chapter. 1. Lilliana Mason, “A Cross-Cutting Calm: How Social Sorting Drives Affective Polarization,” Public Opinion Quarterly 80, no. S1 (2016): 351–377; Mason, “The Rise of Uncivil Agreement: Issue Versus Behavioral Polarization in the American Electorate,” American Behavioral Scientist 57, no. 1 (2013): 140–159; Shanto Iyengar and Sean J. Westwood, “Fear and Loathing Across Party Lines: New Evidence on Group Polarization,” American Journal of Political Science 59, no. 3 (2015): 690–707; Iyengar, Gaurav Sood, and Yphtach Lelkes, “Affect, Not Ideology: A Social Identity Perspective on Polarization,” Public Opinion Quarterly 76, no. 3 (2012): 405–431. 2. Gregory A. Smith and Alan Cooperman, “The Factors Driving the Growth of Religious ‘Nones’ in the U.S.,” in Fact Tank: News in the Numbers (Washington, DC: Pew Research Center, 2016). 3. Roger Finke and Laurence R. Iannaccone, “Supply-Side Explanations for Religious Change,” Annals of the American Academy of Political and Social Science 527, no. 1 (1993): 27–39; Finke and Rodney Stark, The Churching of America, 1776–2005: Winners and Losers in Our Religious Economy (New Brunswick, NJ: Rutgers University Press, 2005); Pippa Norris and Ronald Inglehart, Sacred and Secular: Religion and Politics Worldwide, 2nd ed., Cambridge Studies in Social Theory, Religion, and Politics (Cambridge, UK: Cambridge University Press, 2011); Norris and Inglehart, Sacred and Secular: Religion and Politics Worldwide, 1st ed., Cambridge Studies in Social Theory, Religion, and Politics (Cambridge, UK: Cambridge University Press, 2004). 4. Daniel N. Posner, “The Political Salience of Cultural Difference: Why Chewas and Tumbukas Are Allies in Zambia and Adversaries in Malawi,” American Political Science Review 98, no. 4 (2004): 529–545. 5. Michael Hout and Claude S. Fischer, “Explaining Why More Americans Have No Religious Preference: Political Backlash and Generational Succession, 1987–2012,” Sociological Science 1, no. 9 (2014): 423–447. 6. James L. Guth et al., “Religious Influences in the 2004 Presidential Election,” Presidential Studies Quarterly 36, no. 2 (2006): 223–242. 7. Lyman A. Kellstedt and John C. Green, “Knowing God’s Many People: Denominational Preference and Political Behavior,” in Rediscovering the Religious

Notes

The Role of Religious Identity in US Politics

13

Factor in American Politics, ed. David C. Leege and Lyman A. Kellstedt (Armonk, NY: M. E. Sharpe, 1993), 53–71. 8. Samara Klar and Yanna Krupnikov, Independent Politics: How American Disdain for Parties Leads to Political Inaction (New York: Cambridge University Press, 2016). 9. Geoffrey C. Layman and Christopher L. Weaver, “Religion and Secularism Among American Party Activists,” Politics and Religion 9, no. 2 (2016): 271–295. 10. Robert Huckfeldt, Eric Plutzer, and John Sprague, “Alternative Contexts of Political Behavior: Churches, Neighborhoods, and Individuals,” Journal of Politics 55, no. 2 (1993): 365–381; Kenneth D. Wald, Dennis E. Owen, and Samuel S. Hill, “Churches as Political Communities,” American Political Science Review 82, no. 2 (1988): 531–548. 11. Rebecca A. Glazier, “Bridging Religion and Politics: The Impact of Providential Religious Beliefs on Political Activity,” Politics and Religion 8, no. 3 (2015): 458–487. 12. Ted G. Jelen, “The Political Consequences of Religious Group Attitudes,” Journal of Politics 55, no. 1 (1993): 178–190. 13. Jelen, “The Political Consequences of Religious Group Attitudes.” 14. Robert D. Putnam and David E. Campbell, American Grace: How Religion Divides and Unites Us (New York: Simon and Schuster, 2010). 15. James L. Gibson, “The Political Consequences of Religiosity: Does Religion Always Cause Political Intolerance?” in Religion and Democracy in the United States: Danger or Opportunity, ed. Ira Katznelson and Alan Wolfe (New York: Russell Sage, 2010), 147–175. 16. All names have been changed. 17. Phillip E. Converse, “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. David Apter (New York: Free Press, 1964), 206–261. 18. One example of official LDS language on homosexuality can be found here: https://www.lds.org/topics/same-gender-attraction?lang=eng&old=true. 19. http://www.pewsocialtrends.org/2013/06/13/a-survey-of-lgbt-americans/ #religion. 20. Cody McClain Brown, Chasing a Croatian Girl: A Survivor’s Tale, 3rd ed. (CreateSpace Independent Publishing Platform, 2015). 21. Robert Reich, “The New Tribalism and the Decline of the Nation State,” March 23, 2014, http://robertreich.org/post/80522686347; Amy Chua, “The Destructive Dynamics of Political Tribalism,” New York Times, February 20, 2018. 22. Amy Chua, Political Tribes: Group Instinct and the Fate of Nations (New York: Penguin, 2018). 23. Muzafer Sherif, Intergroup Conflict and Cooperation: The Robbers Cave Experiment, vol. 10 (Norman: University of Oklahoma, Institute of Group Relations, 1961); Sherif, Experimental Study of Positive and Negative Intergroup Attitudes Between Experimentally Produced Groups: Robbers Cave Study (Norman: University of Oklahoma, Institute of Group Relations, 1954). 24. Putnam and Campbell, American Grace; Jonathan Haidt, The Righteous Mind: Why Good People Are Divided by Politics and Religion (New York: Vintage, 2012); Gibson, “The Political Consequences of Religiosity.” 25. Miles Ed Hewstone and Rupert Ed Brown, Contact and Conflict in Intergroup Encounters (Oxford, UK: Blackwell, 1986); Thomas F. Pettigrew, “Intergroup Contact Theory,” Annual Review of Psychology 49, no. 1 (1998): 65–85; Pettigrew and Linda R. Tropp, “A Meta-Analytic Test of Intergroup Contact Theory,” Journal of Personality and Social Psychology 90, no. 5 (2006): 751.

14

Religious Identity in US Politics

26. Pettigrew and Tropp, “Meta-Analytic Test of Intergroup Contact Theory.” 27. Todd Adkins et al., “Religious Group Cues and Citizen Policy Attitudes in the United States,” Politics and Religion 6, no. 2 (2013): 235–263; David E. Campbell, John C. Green, and Geoffrey C. Layman, “The Party Faithful: Partisan Images, Candidate Religion, and the Electoral Impact of Party Identification,” American Journal of Political Science 55, no. 1 (2011): 42–58; John C. Green, The Faith Factor: How Religion Influences American Elections (Westport, CT: Greenwood, 2007). 28. Haidt, The Righteous Mind, 268–269. 29. Ara Norenzayan and Azim F. Shariff, “The Origin and Evolution of Religious Prosociality,” Science 322, no. 5898 (2008): 58–62. 30. Benjamin Grant Purzycki et al., “Moralistic Gods, Supernatural Punishment, and the Expansion of Human Sociality,” Nature 530 (Febuary 18, 2016): 327–330. 31. Will M. Gervais, Azim F. Shariff, and Ara Norenzayan, “Do You Believe in Atheists? Distrust Is Central to Anti-Atheist Prejudice,” Journal of Personality and Social Psychology 101, no. 6 (2011): 1189. 32. Gervais, Shariff, and Norenzayan, “Do You Believe in Atheists?” 33. Iyengar and Westwood, “Fear and Loathing Across Party Lines”; Samara Klar, Yanna Krupnikov, and John Barry Ryan, “Affective Polarization or Partisan Disdain? Untangling a Dislike for the Opposing Party from a Dislike of Partisanship,” Public Opinion Quarterly 82, no. 2 (2018): 379–390. 34. Paul A. Djupe, Jacob R. Neiheisel, and Anand E. Sokhey, “Reconsidering the Role of Politics in Leaving Religion: The Importance of Affiliation,” American Journal of Political Science 62, no. 1 (2018): 161–175; David E. Campbell et al., “Putting Politics First: The Impact of Politics on American Religious and Secular Orientations,” American Journal of Political Science 62, no. 3 (2018): 551–565. 35. Hewstone and Brown, Contact and Conflict in Intergroup Encounters; Pettigrew, “Intergroup Contact Theory.” 36. Djupe, Neiheisel, and Sokhey, “Reconsidering the Role of Politics in Leaving Religion”; Michele F. Margolis, From Politics to the Pews: How Partisanship and the Political Environment Shape Religious Identity (Chicago: University of Chicago Press, 2018). 37. The complete details of these surveys, including recruitment and demographic information, are in Appendix 1.

2 Defining Religious Identity

political attitudes and behaviors, I begin by exploring a parallel situation in scholarship on the intersection of political ideology and partisanship. At present, research on how identities, values, and ideologies combine to influence political attitudes and behaviors points in different directions depending on how scholars conceptualize these ideas. For example, Russell Dalton argues that as people have become better educated and less dependent on others to provide the basic necessities of life, they have developed emancipative values and the cognitive capacity to evaluate politics without relying on “affective, habitual partisan loyalties or external cues.” 1 More Americans are choosing not to affiliate with a political party and instead to think about politics by weighing the platforms of each candidate or party against their own values. Dalton argues that the growth in the number of Americans who identify as independent but lean Democratic or Republican is evidence of a new kind of partisan not blindly loyal to one party but willing to support any party that presents a platform consistent with their personal values.2 Another view is that political ideology has a strong influence on political attitudes. From this perspective, political ideology provides structure to various political ideas and creates an order such that change in one of the political ideas necessitates change in another.3 Elites do much of the legwork involved in structuring the political ideology, and then they sell this packaging to the public in magazines, newspapers, and television programs.4 Proponents of rival ideologies compete by providing and controlling plans for public policy with the aim of contesting or changing the social and political arrangements in a community.5 Political parties are

To frame the dominant view of how religion influences

15

16

Religious Identity in US Politics

coalitions of people who want to win elections and often build constituencies based on idiosyncratic platforms. Parties adopt ideologies when ideological activists within the party demand that elites change positions on some issues to further align their views with those of the ideology. 6 From this perspective, political ideology has a stronger influence on policy attitudes than does partisanship. Elites use the party vehicle to persuade individuals to adopt a coherent political ideology from which the elites then develop a partisan policy agenda. The policy agenda comes after partisans are persuaded to adopt the political ideology. Finally, some see partisanship as a social identity that has an independent influence on political attitudes and behaviors.7 Political parties might be organizationally weaker now than they were a half century ago, but the politically involved have developed a psychological attachment to their partisan identity that strongly influences how they interpret information. Regardless of their party’s political performance in office or elections, Americans retain their partisan identities, suggesting that partisan identities are not abandoned when a party’s performance in office conflicts with one’s political values or ideology. On the contrary, research suggests that when novel issues arise, partisans simply adopt the policy position of elites with whom they share a partisan identity.8 In addition, partisans display considerable out-group antipathy.9 The stronger their partisan identities, the more likely voters are to see elections as a competition and express hostile attitudes toward those who identify with the opposing political party.10 Ultimately, each perspective has important insights into the influence of political values, beliefs, and identities on political behaviors and attitudes toward partisan policies. Each likely describes the manner in which a segment of the population behaves and the underlying reasons for that behavior.11 Much of the scholarship on how ideologies, values, and identities influence support for policies has either focused on a specific policy or one of these mechanisms. Yet the extent to which each of these three mechanisms influence policy support will vary by policy.12 Not until scholars deconstructed the variation in support for multiple issues simultaneously was it possible to see how political values, knowledge, and ideologies influence policy support differently depending on the issue. For example, political ideology has a much weaker influence on support for gay and immigrant rights than it does on support for Medicare and some tax policies.13 The literature on how religion influences political attitudes and behaviors is in a similar state. The dominant view is that religion influences political attitudes and participation via three primary mechanisms: belief, belonging, and behaving.14 In an early approach, Rodney Stark and Charles Glock

How Religion Influences Political Attitudes and Behaviors

Defining Religious Identity

17

argued for five dimensions of religion: ethics, private devotion, beliefs, religious knowledge, and ritual practice.15 Later, Stark and Roger Finke argued that religion involved two forms of commitment: behavior and belief.16 Recent work emphasizes the belief, belonging, and behaving elements of religiosity.17 The belief component emphasizes the extent to which an individual expresses the same beliefs as members of his or her faith, typically measured by individual responses to questions about the veracity of the bible and the existence of an afterlife. Belonging is often operationalized more broadly as group affiliation with its requisite geographical, ethnic, or denominational ties. Behaving is concerned with both private (personal prayer) and public (church attendance) religious behavior. Believing Religious beliefs have both a direct and an indirect influence on political attitudes and behaviors. Directly, people are more likely to support public policy and candidates that are consistent with and reinforce their own religious beliefs. Indirectly, religious beliefs influence their orientations toward major areas of political conflict, such as the role of government in society, the necessity of social welfare, defense/military spending, and various moral/cultural issues. Much like political ideology or partisan identification influence how a person evaluates candidates and political information, religious beliefs influence an individual’s general political orientations.18 Across a wide range of political issues, people who affiliate with different religious traditions express diverse ideas on these issues. For example, most evangelical Protestants oppose abortion, whereas most Jews do not. One of the reasons for this difference is that members of these religious traditions have different beliefs.19 Belonging Kenneth Wald was among the first to articulate the influence of denominational affiliation on political attitudes and behaviors. He demonstrated the importance of distinguishing between different forms of Protestantism in explanations of voter choices; he also pointed out that the influence of religious group affiliation on individual attitudes depends on how integrated the believer is in the religious community. The more integrated individuals are within a religious community, the more reliably their religious affiliation predicts their political attitudes and behaviors.20 Writing on important variables scholars should consider when conducting cross-national research, Philip Converse noted that religious group affiliation has a strong influence on individual attitudes about members of other religious groups. He also mentioned that individuals’ devotion to their religion is significant. People who practice their religion are more likely to conform to model predictions than those who do not.21

18

Religious Identity in US Politics

Stark and Finke argue that social networks are more important to the religious conversion process than religious doctrine is.22 This is because conversion is less about the beliefs associated with a new religion and more about bringing one’s own attitudes into conformity with those of his or her social group. Belonging influences political attitudes and behaviors through two mechanisms. First, religious groups develop religious beliefs associated with political beliefs (e.g., attitudes about abortion); second, religious groups help individuals make the connection between their religious views and the appropriate political views. Religious leaders and religious social networks are the primary means through which belonging translates into political attitudes and behaviors.23 Members of the clergy possess cognitive, physical, and monetary resources that can be used to influence political outcomes. Political activities (voting, registration drives, and petition signing) often occur inside churches. In addition, members of the clergy can donate money or devote time in their weekly sermons to promote political issues. One important role of the clergy is to help church members connect the teachings of the church with modern applications. In so doing, members of the clergy might provide political cues concerning the issues that most closely align with the theological teachings of their churches.24 Social networks within a church also provide political cues. As church members discuss religion and politics at less-structured church gatherings or outside of church when they meet for coffee, the congregants have a direct influence on each other’s political views. Religious social networks not only influence political views through direct discussion of the issues but shape these views as church members observe the social milieu (cue taking).25 Considerable research substantiates the power of religious social networks and members of the clergy to influence the political views and political activity of congregants.26 It is important to note that the preceding authors imply a lengthy process connecting religious social networks to political attitudes or behaviors. First, the social network must formulate some form of consensus on the political issue. Next, the individual must be sufficiently involved in the congregation and have sufficient motivation to learn about the views of other congregants.27 Finally, the social network must exert some form of social pressure to motivate the individual to conform to the group’s view.28 Thus, although religious belonging can help create religious social identities, the direct influence of religious networks on political attitudes and behaviors might not occur as quickly as the influence of religious identities.29 Behaving Scholars consistently find a relationship between church attendance and political attitudes and behaviors. The more regularly people attend church

Defining Religious Identity

19

in evangelical denominations, the more likely they are to be exposed to solicitations from prolife activists, which is why church attendance is a significant predictor of attitudes toward abortion.30 Other scholars view religious behavior as a proxy for religious salience.31 The more people practice the tenets of their faith, the stronger the relationship will be between their religion and some political outcome. In some contexts, behaving conditions the influence of some other variable on a political outcome, whereas in others, behaving has a direct influence on the political outcome of interest. For example, sermons have a stronger influence on the political attitudes of congregants who attend church regularly than on those who do not (indirect conditioning).32 Meanwhile, those who attend church are more likely to give money to charity than those who do not (direct impact).33 In addition, there appears to be a growing political divide in the United States between those who do and do not pray regularly. Half of those who pray daily identify as Republican, and 70 percent of those who never pray identify as Democrats.34 Those who attend church regularly are more likely to trust other people, but they are also less likely to support civil liberties for antireligious people or gays and lesbians.35 The more active people are in their religious community, the more likely they are to support Republican candidates, hold conservative positions on moral/cultural issues, and actively volunteer in civic organizations.36 The weakness in the current state of the religion and politics literature might be the result of two factors. First, although acknowledging the importance of belief, belonging, and behaving (the three Bs), no existing scholarship of which I am aware deconstructs the role of each of the three in comparison with the others. In other words, there are contexts in which belief should have a stronger influence on political attitudes than social networks do. The strength of people’s belief in God and/or an afterlife probably has a stronger influence on attitudes toward policy proposals that widen the separation between church and the state, remove symbols honoring a deity, or alter arrangements for caring for the bodies of deceased persons than does their church attendance or the strength of their social networks. Instead, existing scholarship often conflates the three Bs rather than considering the influence of each independently. For example, Michele Margolis grounds her theory in literature explaining how competing political and religious social identities influence the ability of an evangelical organization to persuade congregants. Yet her analyses use religious beliefs, or denominational affiliation, as a proxy for religious social identification.37 Pazit Ben-Nun Bloom and Gizem Arikan argue that religious social behavior increases support for

Conceptual and Methodological Confusion

20

Religious Identity in US Politics

democracy, which they attribute to the positive influence of religious social networks. However, they measure behavior with a summative index that combines the frequency of church attendance with membership in a religious organization.38 Bethany Albertson measures religiosity by church attendance in her Pentecostal sample, but her general population measure for religiosity is self-reporting on how religious the participants are.39 Christopher Weber and Matthew Thornton combine denominational affiliation and views about the bible, Jesus Christ, and the preservation of Christian beliefs into a single measure that they argue moderates the effects of religious priming. Although they found considerable empirical support for combining these responses into a single measure, there is not a clear conceptual reason for this. 40 Finally, Paul Goren and Christopher Chapp conceptualize religious belief and commitment as “strong predispositions” akin to party identification and attitudes about culture war issues. They demonstrate an interesting phenomenon in which views on culture war issues can have a small influence on subsequent religious commitment. They also report that religious commitment has a strong influence on subsequent culture war attitudes, suggesting more of a dynamic relationship between the two. Yet their study makes it less clear how religious belief, belonging, and behaving influence attitudes. 41 These prominent examples are only a few of the many studies that conflate religious concepts into a single empirical measure. Second, the belonging mechanism really encompasses two distinct concepts: affiliation and identity. Paul Djupe and colleagues also distinguish between affiliation and identity, but they use the terms differently than I use them in this book. They argue that affiliation is specific to congregation, whereas identity is specific to religious tradition.42 Using the language of psychologists, I argue that affiliation occurs when individuals have formally chosen to associate with a religious community, but religious identification exists when their religion merges with their sense of self. Each of the three Bs can work to strengthen people’s identification with a religion, but their religious identities influence their political attitudes differently than do their affiliations, beliefs, behaviors, or religious social networks. Many people affiliate with a religion but only weakly identify with it. People frequently switch between denominations, and many of the religiously unaffiliated often affiliate with a Christian persuasion within a few years.43 In addition, people can retain their social identities with their religion even when they no longer affiliate with that religion. As I discussed in the previous chapter, people who no longer believe, belong, or behave as members of a religion can still identify with that religion. Furthermore, considerable evidence supports the influence of religious social identities on political attitudes. Ted Jelen found that the influence of religious identities on political attitudes operates through a likeability heuristic. Because most people are

Defining Religious Identity

21

uninformed about political processes and know even less about most political actors, one of the ways they make sense of politics is to rely on information shortcuts (heuristics) that easily categorize information.44 People might not know a lot about a particular candidate for office, but they do know which party they prefer. This partisan preference can act as a heuristic in which they perceive everything associated with the preferred party as good and everything associated with the opposing party as undesirable. Religious identity functions similarly. Jelen demonstrates that identification with various Protestant religious movements has a predictable, independent effect on political issue attitudes.45 A similar scholarly perspective holds that when people adopt religious identifications, they behave as they perceive that members of that group should behave. Identification with religious groups causes people to compare members of their group with those in other groups. This diminishes perceptions of diversity within their own religion and exaggerates perceived differences with members of other religious communities. James Penning demonstrates that religious identification has a strong, independent influence on attitudes about Mormons, American Muslims, and Muslims. Evangelical Christians are more likely to express negative attitudes toward members of these religious groups, as are those who think Mormonism and Islam have little in common with their own faith.46 The prevailing view among scholars is that religious identities form through a combination of socialization and exposure to religious doctrine. Religious organizations’ intimate involvement in early and ongoing socialization produces and maintains common ways of thinking. People often continue in their interaction with others in their religious community, which creates and reinforces like-mindedness among members of the congregation. In time, this produces a positive affect toward their own religious group and a negative affect toward other groups.47 Early work on the development of religious identities found that cues from church leaders had an influence on the expression of these identities. For example, people who believed in conservative religious doctrine but were not exposed to the terminology through attending church or listening to televangelists were less likely to develop a fundamentalist, evangelical, or Pentecostal identity.48 Yet this conceptualization of religious identities makes it difficult to distinguish the influence of those identities from the effects of targeted communication.49 After all, if religious identities are constructed and made salient by communication in the church, and those who attend church regularly express stronger identities than those who do not, much of what the literature calls identity effects might be communication effects.50 Religious

How Religious Identities Form

22

Religious Identity in US Politics

identity implies a different form of religious influence than the other social influences in churches. First, social identities have their strongest influence in low-information environments. People are more likely to follow group-based cues when accurate information is less important.51 In other words, information effects are more likely to occur when the answer is important (e.g., social identities rarely influence how people choose to treat cancer because wrong information could be terminal). Identities matter more when the outcome of the choice is not as important. The more important it is for congregants to find accurate information on issues, the more likely they would be to pay attention to the information provided by the members of the clergy in their church. Second, the negative influence of religious social identity on attitudes toward adherents of other religions attenuates with greater contact. The more time people spend with members of out-groups, the less anxious and threatened they feel about members of those groups. In time, this reduces out-group antipathy.52 In the next section, I develop a framework that explains both how social identities influence attitudes and behavior and why religious social identity is conceptually important to the religion and politics literature. This not only helps to clarify the role of members of the clergy and their church teachings in their congregants developing religious identities but also makes clear that many other influences are at play. Humans are social beings and long for interaction with each other. Even those who deeply enjoy being by themselves and prefer spending time alone to spending time with others need some form of social inclusion. People join groups for the pleasure of human company, for safety, and to accomplish goals they could not accomplish alone. Many people belong to groups without developing group identifications. Group membership is a necessary, but not a sufficient, condition for developing group identification. Cognitive psychologists argue that humans develop schemas—vague abstractions based on experience—that guide their processing of new information and retrieval of stored information.53 A self-schema merges information about some domain of human experience with information about the self. The process of consciously blending the mental representation of the self with the cognitive stereotype of a group to which one belongs creates group identification.54 Group identities come from unique aspects of the self, combined with perceived traits shared with others. 55 For example, people derive their social identities from shared categories such as gender, race, and income level that locate people in social space. These categories distinguish people from those who do not share these characteristics and unite

Religious Social Identity

Defining Religious Identity

23

them with those who do. Social identities designate people’s positions in a social structure that goes beyond the present situation.56 Just as they form other group identities, people in a free society join religions for the pleasure of human company, for a sense of safety, and to accomplish goals they cannot accomplish alone. They can also derive personal satisfaction from being affiliated with a high-status group in their community. Many people join religious groups without developing strong religious group identifications. Americans disaffiliating with evangelical religious traditions likely have weaker identifications with their evangelical religions than do those who do not switch faiths. When the Russian prince Vladimir compelled his people by the threat of death to convert to Christianity in 988 A.D., some people did not convert (and were executed), whereas others converted in name only to the new faith but retained beliefs and practices from their prior religion.57 Those who converted to Christianity by force likely had a weaker Christian identification than those who did not. Regardless, after becoming affiliated with a religion, people develop a self-schema that merges information about the self with information about that religion. This process will cause them to develop a cognitive stereotype of the religion and adopt the religious social identification. This religious social identification is based on unique aspects of the self and perceived traits shared with other religious group members. As their religious social identity strengthens, they distinguish themselves from those in other religions and unite themselves with those in their own religion. In addition, the religious social identity designates their position in the societal structure, which involves more than religion. Both advocates of the elaboration-likelihood58 and the heuristic-systematic59 models of persuasion argue that humans process messages by one of two methods. The language used in their models varies, but these scholars agree that people either systematically or automatically evaluate message content. A good example of systematic information processing is when a mother is trying to decide the appropriate remedy for a sick child. She carefully evaluates the content of the suggested remedies, scrutinizing both the source of the information and the quality of the evidence in support of the message. Only when she determines that the arguments in the message are sufficiently sound will she make her decision. Automatic message evaluation occurs when, for example, consumers are deciding which brand of cola to drink. In this case, people do not engage in careful reasoning; rather, they use cognitive heuristics to help them decide. Instead of carefully scrutinizing the facts about the relative flavor, health benefits, price, and long-term effects of consuming one

How Social Identities Influence Attitudes and Behavior

24

Religious Identity in US Politics

beverage over another, people are less careful about this decision. Usually, their choice is determined by a superficial cue such as the attractiveness of the person promoting the product in an advertisement, the packaging, the product’s popularity, or their familiarity with the brand. When a message is supportive of people’s existing worldviews and perceptions, it is processed automatically and is usually adopted without much cognitive effort.60 Most often, social identities influence attitudes formed through automatic thought processes. When people lack the motivation, cognitive capacity, or information to systematically process a new message, they rely on cues to evaluate the situation.61 Because people develop social identities to enhance their self-image, information is often heuristically evaluated on the basis of how it will influence a person’s self-image. Information that could cause society to negatively view the group with which a person identifies can also have a negative influence on a person’s own selfimage. The simplest resolution to the dilemma is to reject the message. When this is not possible, people might adopt the message but reevaluate their commitment to either the relevant social identities or their own selfschemas. Conversely, when a new message reflects negatively on members of the out-group, by extension, it reflects positively on the in-group. In this case, people will readily accept a derogatory message about out-group members because it enhances their perception of the in-group and their own selves. Table 2.1 describes the influence that the strength of social identities—and whether messages threaten individual or group identities—has on how individuals will respond. The first two responses (1. Accuracy/efficiency and 2. Social meaning) describe status-quo conditions for a group in society. Most of the literature on the influence of social identity on individual attitudes and behaviors focus on these two responses. Groups with which individuals have low identification have virtually no influence on how they evaluate information (1. Accuracy/efficiency). In contrast, when individuals strongly identify with a group, the group will have some influence on how they evaluate information. The stronger the group identification, the more the group schema is incorporated into the self-schema and the stronger the group’s predicted influence on infor-

Table 2.1 How Threats Influence Individual Identification with the Group

No threat Individual-directed threat Group-directed threat

Low

Strength of Group Identity

1. Accuracy/efficiency 3. Categorization 5. Value

High

2. Social meaning 4. Exclusion 6. Distinctiveness

Defining Religious Identity

25

mation processing (2. Social meaning). A strong social identification motivates people to express and affirm their group identities. People in these situations tend to endorse group norms, conform to the prototypical group position, and focus on how they are like those in the in-group, much as how Paul behaves when he is in a group of other gay men. The expectations of these two responses lead to the first two hypotheses tested in this book. The In-Group Favoritism Hypothesis When evaluating individuals, people will have more positive attitudes toward those with whom they share religious identities. This is akin to Jelen’s likeability heuristic. In Chapters 3 and 4, I examine the extent to which the likeability heuristic derived from shared religious identities transcends opposition motivated by differing political identities. Obviously, people who share religious and partisan identities with a political leader have more positive attitudes toward that elected official, but what happens when they share religious identification with the elected representative but have opposing partisan identities? The preceding discussion suggests that the relative strength of the two identities influences these evaluations. When people have a strong religious identity, they might more positively evaluate elected officials with whom they share a religious identity even when their partisan identities are not aligned. The Out-Group Antipathy Hypothesis Group identities enhance individual self-esteem through the process of social categorization. Social identities only serve their esteem-building function when others are classified as members of competing groups. When individuals categorize others, they perceive them through the lens of the relevant in-group or out-group prototype. These prototypes describe one group and distinguish it from the other groups. Over time, derogatory attitudes toward and stereotypes of members of out-groups tend to crystallize. Studies show that people perceive members of their own group positively and attribute negative qualities to members of the out-group.62 This is because seeing others through the lens of an out-group depersonalizes those individuals. People from the in-group judge them as more similar to an out-group-based prototype than they probably are. As group identities become salient, there is a tendency to accentuate intergroup similarities and intragroup differences. People interpret out-group success as an anomaly and attribute failure to the internal, stable features of those in the out-group. Individuals who belong to a group tend to internalize the group’s norms and use these as guides for their own attitudes and behaviors.63 If these hypotheses are correct, we should observe social division between those who do and do not belong to religions. Just as we observe racial differences in society, we should observe a tiered structure that

26

Religious Identity in US Politics

defines the distance between members of the in-group and the out-group. Religious out-groups are defined by doctrinal proximity.64 People see members of religious groups who share their beliefs in certain deities or texts as closer to their religious in-group than they do people who do not believe in the same sacred texts or deities. For example, people with strong Christian religious identities categorize those who reject the existence of God as the most distant out-group. In addition, evangelical Christians might feel greater affinity with Jews—because they believe in similar religious texts— than they do with Muslims. The religion and politics literature suggests that because Jews tend to align themselves with Democrats and evangelical Protestants are predominantly Republicans, members of these two groups should have accentuated out-group perceptions of each other.65 In contrast, I hypothesize that the relative strength of religious and political identities structure the social conflict between these two groups. For nonpolitical evangelical Christians, one could expect more positive attitudes and associations with Jews than with atheists or Hindus. The Competing Identities Hypothesis The final hypothesis tested in this book is based on the remaining responses shown in Table 2.1. The next two responses (3. Categorization and 4. Exclusion) describe what happens when identification with a group begins to threaten individuals’ self-esteem or their conception of themselves. For example, when Paul is in another country, and the crowd begins talking negatively about the United States, he often directs the conversation to what an atypical American he is. By accentuating his differences from other Americans, Paul effectively deflects this individual-directed threat away from himself and toward others who are more prototypical Americans. He does this because his national identity is weaker than his other social identity (3. Categorization). In contrast, when someone attacks him for being gay, that strong identity motivates him to find shelter in his gayness (4. Exclusion). The fourth response predicts that when group identification threatens individuals’ sense of self, those with strong social identities will form even stronger attachment to that social group and its norms because higher group esteem compensates for lower self-esteem. This is one reason Paul is offended that the LDS Church does not acknowledge homosexuality as an identity. Throughout his life, Paul has been personally attacked by members of the LDS Church for being gay, to which he responds by developing an even stronger gay identity. When those in the LDS Church do not acknowledge this as a legitimate identity, it threatens his gay identity. One of the basic tenets of social identification theory is that when their social identities are unsatisfactory, individuals will strive either to leave their existing groups and join some more positively distinct groups and/or to make their existing groups more positively distinct.66 Surely this

Defining Religious Identity

27

depends on the ease with which they can abandon one group identity and adopt another. Obviously, those with weak group identifications might abandon an unfavorably viewed group (5. Value). One of the reasons more Americans identify as political independents is most Americans view the two major political parties in the United States negatively. Rather than working to improve the image of their political parties, many Americans with weak partisan identities abandon their political parties to protect their self-image.67 However, some identities are not easily discarded. When people strongly identify with a group and conditions cause that group to move from high status to lower status, how do they respond? The sixth response predicts that they will dig in their heels and become more committed to their group (6. Distinctiveness). Even when there is no realistic chance of improving the group’s status in the moment, the highly committed group members will stick together.68 Typically, this means creating greater in-group distinctiveness, excluding group members who are not fully committed, or refusing to apologize to people in other groups offended by in-group attitudes and behaviors. Generally, people prefer their groups to be distinct from other groups in positive ways, but when people experience group-level threats to strong social identities, they might settle for a negative identity over being regarded positively at the expense of group distinctiveness.69 The strength of religious identity in particular contexts depends on the strength of a person’s religious commitment and the costs and benefits associated with religious identification in given moments. Individuals might not ever attend church yet feel strong identification with their religions because they were fundamental to their childhood or home experiences (like Susie). Similarly, religious identity might weaken among strict adherents to religions in contexts where those faiths do not seem relevant. For example, religious identity is unlikely to influence group work in math courses because religious beliefs are not the most relevant identity in that context (but contrast that with science classes about evolution). Religious social identities are malleable, and as societal conditions alter the status of religious groups in society, people will weaken or strengthen their identification with these groups accordingly. History is replete with examples of people who converted from one religion to another because political leaders threatened to execute those who did not change their religious affiliations. A high court in India decided that Hindu converts to a religion that does not recognize caste lose their caste after conversion.70 After this court ruling, many scheduled caste (known as “untouchable”) Hindus converted to Islam to escape their stigmatized

When Social Identities Collide

28

Religious Identity in US Politics

status in society.71 When the British military showed a preference for Sikh recruits in the Punjab, many Hindus converted to Sikhism.72 Elizabeth Shakman Hurd discusses how attempts by Turkey, the European Union, and the international community to legally define Alevism create problems within the religion itself and made religious identification salient. Alevis’ beliefs are hard to define because they have no central authority, it is a diverse movement, and poorly demarcated boundaries with other groups. The European Union classifies Alevis as a non-Sunni Muslim minority group, which serves to distinguish Alevis from non-Alevi Turkish citizens in religious terms. This classification reinforces the perception that the Turkish Sunnimajority state is the official arbiter of religious identities and practices and produces the perception that Alevis are different from other Sunni Turkish citizens. In addition, the legal creation of an Alevi religion empowers people within the community to speak on behalf of kindred believers. By empowering some voices within the Alevi community, the Turkish government implicitly sanctions some expressions of Alevism as orthodox but marginalizes other practices. This creates discord within Alevism as adherents contest the meaning and boundaries of their faith. Those whose views do not align with those of the Alevi leaders are seen as apostates and insurgents.73 These examples demonstrate how political circumstances can change the status of religious groups in society, which in turn can motivate people to alter the strength of their identification with religious groups. The Alevi case in Turkey is one of many historical examples of people altering their religious identification in response to changing societal conditions.74

Party Identification as a Threat to Religious Social Identities

Political threats to religious social identification can come in many forms. In the United States, many people develop an attachment to a political party that can be just as strong as any other social identity.75 As discussed previously, the political parties in the United States often support policies that reinforce people’s religious views, but political parties might also support policies incongruent with individuals’ religious beliefs. This can threaten their identification with their religious group, their political party, or both. In addition, their standing within a social group can influence the strength of their group identities. People who rank highly in a group are more likely to see that group as an extension of themselves, whereas lower-status individuals often see themselves as distinct from the group.76 The lower-status members of any social group can feel a weaker identification with that group when they are aware of their standing within the group. Political party coalitions are not constant and can change rapidly. Individuals and groups can move from being less valued to highly valued members of the

Defining Religious Identity

29

party coalition and vice versa. When one religious group loses its status within a political party, this could threaten the strength of individuals’ religious or partisan group identifications. When one social identity threatens another, it has the potential to influence attitudes about each of the social identities. Previous scholars largely assume religious social identities are developed and maintained within churches and do not necessarily distinguish between the influence of exposure to political rhetoric and adoption of those political views.77 Like Paul, many people maintain religious social identities long after they no longer attend worship services or believe in the theology of the churches into which they were born.78 Unless these people have stronger political than religious identities, there is no reason to assume that their religious identity would have a weaker influence on their political views. In fact, the logical conclusion from social identity theory is that a nonreligious, noncommitted partisan with a moderate religious social identity would be more likely to adopt the views of their church than they would the views of a partisan candidate. The strength of the competing identities is more important than the beliefs that underlie those identities. In this case, a social identity represents a group level threat to another social identity. Even people with both strong partisan and religious identification place higher importance on one identity than on the other. Different social conditions can cause one identity to be more salient than another in any given moment. I am not referring to this situation. Rather, when something occurs that causes both identities to become salient simultaneously, people ultimately place higher priority on one identity over the other. This is the condition described in the top left square of Table 2.2 (“Individual adopts view of the higher-status group, becomes weaker identifier with other group”). When one strong identity poses a group-level threat to the other, people place higher priority on one over the other. Doing so has two consequences. First, there is a measurable disidentification with the lower-status group. 79 The strength of individuals’ attachment to the group that loses their favor will measurably decline almost immediately. Second, the individuals adopt views consistent with the group that wins their favor. If partisan identification is stronger, the individuals have weaker religious social identification and adopt the views of the party and vice versa. Based on the preceding discussion, I expect religious identity to have a strong, independent influence on how individuals process information. All else equal, those with a stronger religious identity are more likely to accept and agree with information that reflects positively on their own religion, and they are less likely to accept negative information about their own

Empirical Expectations

30

Religious Identity in US Politics

Table 2.2

Collision of Religious and Partisan Social Identities

Strong Religious Identity Weak Religious Identity

Strong Partisan Identity

Individual adopts view of the higher-status group, becomes weaker identifier with other group Individual becomes less religious, adopts view of the party, becomes stronger partisan

Weak Partisan Identity

Individual becomes less partisan, adopts view of the religion, becomes weaker partisan Individual adopts view of the higher-status group, becomes weaker identifier with the lower-status group

religion. Those with a weaker religious social identity are not influenced by how the information reflects on their own or any other religion. Other relevant social identities will have a stronger impact. In addition, the strength of the religious social identity has a stronger influence on attitudes in low-information environments. Heuristics matter more when people are processing information on topics about which they know less. That is, the more people know about the situation, the more likely they will be to use that information to come to an accurate conclusion about the message. As such, religious social identities have a stronger influence on attitudes about new issues or issues on which religious leaders have recently taken a stand for the first time. In this low-information context, the strength of the religious social identity might be the only available heuristic. However, after an issue has been made salient for a time, other aspects of religion (social networks, beliefs, church attendance, and communication effects) will begin to have a stronger influence on attitudes than religious identity. This is because people now have more information, and accuracy motivations become more important than the more fleeting self-enhancement motivations. The distinguishing feature of the competing identities hypothesis is that it isolates conditions under which partisan identities influence religious views and vice versa. In the United States, there has been an alignment of religious and social identities to the point that it is difficult to empirically distinguish the influence of one from the other. If both partisanship and religion are social identities as opposed to strong attitudinal predispositions, as I contend, social identity theory yields some empirical, testable predictions of what occurs when the two collide. First, the in-group favoritism hypothesis is that we should observe religious-based favoritism toward members of the religious in-group. Second, the out-group hypothesis is that people who identify with religious in-groups have negative attitudes toward members of the religious out-group. Finally, the competing identities hypothesis is that when a situation arises that causes another social identity to compete with the religious identity, the relative strength of the two identities has a predictable influence on the subsequent strength of each social identity.

Defining Religious Identity

31

1. Russell J. Dalton, The Apartisan American: Dealignment and Changing Electoral Politics (Washington, DC: Congressional Quarterly Press, 2013), 39. 2. Dalton, Apartisan American. 3. Donald R. Kinder and Nathan P. Kalmoe, Neither Liberal nor Conservative: Ideological Innocence in the American Public (Chicago: University of Chicago Press, 2017), 13. 4. Hans Noel, Political Ideologies and Political Parties in America (Cambridge, UK: Cambridge University Press, 2014). 5. Michael Freeden, Ideology: A Very Short Introduction, vol. 95 (Oxford University Press, 2003), 32. 6. Noel, Political Ideologies and Political Parties in America. 7. Donald P. Green, Bradley Palmquist, and Eric Schickler, Partisan Hearts and Minds: Political Parties and the Social Identities of Voters, Yale ISPS Series (New Haven, CT: Yale University Press, 2002); Angus Campbell et al., The American Voter (New York: John Wiley and Sons, 1960). 8. Gabriel S. Lenz, Follow the Leader? How Voters Respond to Politicians’ Policies and Performance (Chicago: University of Chicago Press, 2013). 9. Leonie Huddy, Lilliana Mason, and Lene Aarøe, “Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity,” American Political Science Review 109, no. 1 (2015): 1–17; Mason and Julie Wronski, “One Tribe to Bind Them All: How Our Social Group Attachments Strengthen Partisanship,” Political Psychology 39, no. S1 (2018): 257–277; Mason, “‘I Disrespectfully Agree’: The Differential Effects of Partisan Sorting on Social and Issue Polarization,” American Journal of Political Science 59, no. 1 (2015): 128–145. 10. Patrick R. Miller and Pamela Johnston Conover, “Red and Blue States of Mind: Partisan Hostility and Voting in the United States,” Political Research Quarterly 68, no. 2 (2015): 225–239; Shanto Iyengar, Gaurav Sood, and Yphtach Lelkes, “Affect, Not Ideology: A Social Identity Perspective on Polarization,” Public Opinion Quarterly 76, no. 3 (2012): 405–431; Iyengar and Sean J. Westwood, “Fear and Loathing Across Party Lines: New Evidence on Group Polarization,” American Journal of Political Science 59, no. 3 (2015): 690–707. 11. David E. Broockman, “Approaches to Studying Policy Representation,” Legislative Studies Quarterly 41, no. 1 (2016): 181–215. 12. Christopher D. Johnston, Howard Lavine, and Christopher M. Federico, Open Versus Closed: Personality, Identity, and the Politics of Redistribution (Cambridge, UK: Cambridge University Press, 2017). 13. Benjamin E. Lauderdale, Chris Hanretty, and Nick Vivyan, “Decomposing Public Opinion Variation into Ideology, Idiosyncrasy, and Instability,” Journal of Politics 80, no. 2 (2018): 707–712. 14. Corwin Smidt, Lyman Kellstedt, and James Guth, “The Role of Religion in American Politics: Explanatory Theories and Associated Analytical and Measurement Issues,” in The Oxford Handbook of Religion and American Politics, ed. Corwin Smidt, Lyman Kellstedt, and James Guth (New York: Oxford University Press, 2009). 15. Rodney Stark and Charles Y. Glock, American Piety: The Nature of Religious Commitment, vol. 1 (Berkeley: University of California Press, 1968). 16. Rodney Stark and Roger Finke, Acts of Faith: Explaining the Human Side of Religion (Berkeley: University of California Press, 2000). 17. Kenneth D. Wald and Corwin E. Smidt, “Measurement Strategies in the Study of Religion and Politics,” in Rediscovering the Religious Factor in American Politics, ed. David C. Leege and Lyman A. Kellstedt (Armonk, NY: M. E. Sharpe, 1993); Stark and Finke, Acts of Faith; Stark and Glock, American Piety.

Notes

32

Religious Identity in US Politics

18. Geoffrey C. Layman, The Great Divide: Religious and Cultural Conflict in American Party Politics (New York: Columbia University Press, 2001), 253. 19. Kenneth D. Wald and Allison Calhoun-Brown, Religion and Politics in the United States (Lanham, MD: Rowman and Littlefield, 2014), 187. 20. Kenneth D. Wald, “Assessing the Religious Factor in Electoral Behavior,” in Religion in American Politics, ed. Charles W. Dunn (Washington, DC: Congressional Quarterly Press, 1989). 21. Philip E. Converse, “Some Priority Variables in Comparative Electoral Research,” in Electoral Behavior: A Comparative Handbook, ed. Richard E. Rose (New York: Free Press, 1974). 22. Stark and Finke, Acts of Faith. 23. Layman, Great Divide, 59. 24. Corwin E. Smidt, Pulpit and Politics: Clergy in American Politics at the Advent of the Millennium (Waco, TX: Baylor University Press, 2004). 25. Paul A. Djupe and Christopher P. Gilbert, The Political Influence of Churches, Cambridge Studies in Social Theory, Religion, and Politics (Cambridge, UK: Cambridge University Press, 2009). 26. Paul A. Djupe and Christopher P. Gilbert, The Prophetic Pulpit: Clergy, Churches, and Communities in American Politics (Lanham, MD: Rowman and Littlefield, 2003); Robert D. Putnam and David E. Campbell, American Grace: How Religion Divides and Unites Us (New York: Simon and Schuster, 2010). 27. Djupe and Gilbert, Political Influence of Churches, 154. 28. Betsy Sinclair, The Social Citizen: Peer Networks and Political Behavior (Chicago: University of Chicago Press, 2012), 137. 29. James M. Penning, “The Political Behavior of American Catholics: An Assessment of the Impact of Group Integration Versus Group Identification,” Western Political Quarterly 41, no. 2 (1988): 289–308. 30. Clyde Wilcox, “Religion and Politics Among White Evangelicals: The Impact of Religious Variables on Political Attitudes,” Review of Religious Research (1990): 27–32. 31. James L. Guth and John C. Green, “Salience: The Core Concept,” in Rediscovering the Religious Factor in American Politics, ed. David C. Leege and Lyman A. Kellstedt (New York: M. E. Sharpe, 1993), 157–176. 32. Djupe and Gilbert, Political Influence of Churches. 33. Putnam and Campbell, American Grace, 449. 34. Putnam and Campbell, American Grace, 372. 35. Putnam and Campbell, American Grace, 486. 36. Layman, Great Divide, 57. 37. Michele F. Margolis, “How Far Does Social Group Influence Reach? Identities, Elites, and Immigration Attitudes,” Journal of Politics 80, no. 3 (2018): 772–785. 38. Pazit Ben-Nun Bloom and Gizem Arikan, “A Two-Edged Sword: The Differential Effect of Religious Belief and Religious Social Context on Attitudes Towards Democracy,” Political Behavior 34, no. 2 (2012): 249–276. 39. Bethany L. Albertson, “Religious Appeals and Implicit Attitudes,” Political Psychology 32, no. 1 (2011): 109–130. 40. Christopher Weber and Matthew Thornton, “Courting Christians: How Political Candidates Prime Religious Considerations in Campaign Ads,” Journal of Politics 74, no. 2 (2012): 400–413; Bethany L. Albertson, “Dog-Whistle Politics: Multivocal Communication and Religious Appeals,” Political Behavior 37, no. 1 (2015): 3–26. 41. Paul Goren and Christopher Chapp, “Moral Power: How Public Opinion on Culture War Issues Shapes Partisan Predispositions and Religious Orientations,” American Political Science Review 111, no. 1 (2017): 110–128.

Defining Religious Identity

33

42. Paul A. Djupe, Jacob R. Neiheisel, and Anand E. Sokhey, “Reconsidering the Role of Politics in Leaving Religion: The Importance of Affiliation,” American Journal of Political Science 62, no. 1 (2018): 161–175. 43. Ryan P. Burge, “Plenty of the ‘Nones’ Actually Head Back to Church,” Christianity Today, February 6, 2018. 44. Michael X. Delli Carpini and Scott Keeter, What Americans Know About Politics and Why It Matters (New Haven, CT: Yale University Press, 1996). 45. Ted G. Jelen, “The Political Consequences of Religious Group Attitudes,” Journal of Politics 55, no. 1 (1993): 178–190. 46. James M. Penning, “Americans’ Views of Muslims and Mormons: A Social Identity Theory Approach,” Politics and Religion 2, no. 2 (2009): 277–302. 47. Lyman A. Kellstedt and John C. Green, “Knowing God’s Many People: Denominational Preference and Political Behavior,” in Rediscovering the Religious Factor in American Politics, ed. David C. Leege and Lyman A. Kellstedt (Armonk, NY: M. E. Sharpe, 1993). 48. Clyde Wilcox, “The Fundamentalist Voter: Politicized Religious Identity and Political Attitudes and Behavior,” Review of Religious Research 31, no. 1 (1989): 54–67; Wilcox, Ted G. Jelen, and David C. Leege, “Religious Group Identifications: Toward a Cognitive Theory of Religious Mobilization,” in Rediscovering the Religious Factor in American Politics, ed. David C. Leege and Lyman A. Kellstedt (Armonk, NY: M. E. Sharpe, 1993), 72–99. 49. Paul Djupe and Brian Calfano, God Talk: Experimenting with the Religious Causes of Public Opinion (Philadelphia, PA: Temple University Press, 2014). 50. Djupe and Gilbert, Political Influence of Churches. 51. James N. Druckman, Erik Peterson, and Rune Slothuus, “How Elite Partisan Polarization Affects Public Opinion Formation,” American Political Science Review 107, no. 1 (2013): 57–79; Rune Slothuus and Claes H. De Vreese, “Political Parties, Motivated Reasoning, and Issue Framing Effects,” Journal of Politics 72, no. 3 (2010): 630–645; Brian J. Gaines et al., “Same Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq,” Journal of Politics 69, no. 4 (2007): 957–974. 52. Thomas F. Pettigrew, “Intergroup Contact Theory,” Annual Review of Psychology 49, no. 1 (1998): 65–85; Pettigrew and Linda R. Tropp, “A Meta-Analytic Test of Intergroup Contact Theory,” Journal of Personality and Social Psychology 90, no. 5 (2006): 751. 53. Susan T. Fiske and Patricia W. Linville, “What Does the Schema Concept Buy Us?” Personality and Social Psychology Bulletin 6, no. 4 (1980): 543–557. 54. Pamela Johnston Conover, “The Influence of Group Identifications on Political Perception and Evaluation,” Journal of Politics 46, no. 3 (1984): 760–785. 55. Michael A. Hogg and Scott A. Reid, “Social Identity, Self-Categorization, and the Communication of Group Norms,” Communication Theory 16, no. 1 (2006): 7–30; Michael A. Hogg and Graham M. Vaughan, Essentials of Social Psychology (New York: Pearson, 2009). 56. Naomi Ellemers, Russell Spears, and Bertjan Doosje, “Self and Social Identity,” Annual Review of Psychology 53, no. 1 (2002): 161–186. 57. Eve Levin, “Dvoeverie and Popular Religion,” in Seeking God: The Recovery of Religious Identity in Orthodox Russia, Ukraine, and Georgia, ed. Stephen K. Batalden (DeKalb: Northern Illinois University Press, 1993), 29–52. 58. Richard E. Petty and John T. Cacioppo, “The Elaboration Likelihood Model of Persuasion,” Advances in Experimental Social Psychology 19, no. 1 (1986): 123–205. 59. Shelly Chaiken, “Heuristic Versus Systematic Information Processing and the Use of Source Versus Message Cues in Persuasion,” Journal of Personality and Social Psychology 39, no. 5 (1980): 752.

34

Religious Identity in US Politics

60. Richard R. Lau and David P. Redlawsk, “Advantages and Disadvantages of Cognitive Heuristics in Political Decision Making,” American Journal of Political Science (2001): 951–971; Redlawsk, “Hot Cognition or Cool Consideration? Testing the Effects of Motivated Reasoning on Political Decision Making,” Journal of Politics 64, no. 4 (2002): 1021–1044. 61. Arthur Lupia, “Shortcuts Versus Encyclopedias: Information and Voting Behavior in California Insurance Reform Elections,” American Political Science Review (1994): 63–76. 62. James L. Hilton and William Von Hippel, “Stereotypes,” Annual Review of Psychology 47, no. 1 (1996): 237–271. 63. Michael M. Franz, Choices and Changes: Interest Groups in the Electoral Process (Philadelphia, PA: Temple University Press, 2008). 64. The reasons for this are more fully developed in subsequent chapters. 65. Iyengar, Sood, and Lelkes, “Affect, Not Ideology”; Iyengar and Westwood, “Fear and Loathing Across Party Lines.” 66. Henri Tajfel and John C. Turner, “An Integrative Theory of Intergroup Conflict,” Social Psychology of Intergroup Relations 33, no. 47 (1979): 40. 67. Samara Klar and Yanna Krupnikov, Independent Politics: How American Disdain for Parties Leads to Political Inaction (New York: Cambridge University Press, 2016). 68. Bertjan Doosje, Russell Spears, and Naomi Ellemers, “The Dynamic and Determining Role of Ingroup Identification: Responses to Anticipated and Actual Changes in the Intergroup Status Hierarchy,” British Journal of Social Psychology 41, no. 1 (2002): 57–76. 69. Ellemers, Spears, and Doosje, “Self and Social Identity.” 70. “Community Status Lapses on Conversion, Rules Madras High Court,” Hindu, June 24, 2013. 71. Joane Nagel, American Indian Ethnic Renewal: Red Power and the Resurgence of Identity and Culture (Oxford, UK: Oxford University Press, 1997). 72. Baldev Raj Nayar, Minority Politics in the Punjab (Princeton, NJ: Princeton University Press, 1966). 73. Elizabeth Shakman Hurd, Beyond Religious Freedom: The New Global Politics of Religion (Princeton, NJ: Princeton University Press, 2015), 102–105. 74. For another example, see Galina M. Yemelianova, “Islam, National Identity, and Politics in Contemporary Kazakhstan,” Asian Ethnicity 15, no. 3 (2014): 286–301. 75. Campbell et al., American Voter; Green, Palmquist, and Schickler, Partisan Hearts and Minds. 76. Naomi Ellemers and Manuela Barreto, “The Impact of Relative Group Status: Affective, Perceptual, and Behavioral Consequences,” Blackwell Handbook of Social Psychology: Intergroup Processes (Oxford, UK: Blackwell, 2001), 325–343. 77. Djupe and Calfano, God Talk. 78. Abby Day, Believing in Belonging: Belief and Social Identity in the Modern World (Oxford, UK: Oxford University Press, 2011). 79. Maykel Verkuyten and Ali Aslan Yildiz, “National (Dis)Identification and Ethnic and Religious Identity: A Study Among Turkish-Dutch Muslims,” Personality and Social Psychology Bulletin 33, no. 10 (2007): 1448–1462.

3 Religious Identity and Support for Members of Congress

group identity with others will be more likely to trust them and have positive feelings about them because groups become an extension of the self. 1 The central argument of this book is that when people develop a religious social identity, it can influence political attitudes and behaviors independently of partisanship. These social identities manifest themselves in both positive and negative ways for society. One positive way religious social identities manifest themselves is in approval of elected representatives. Recent scholarship on Congress emphasizes how political polarization influences US legislative outputs. Both rules changes from the 1970s through the 1990s and the ideological extremity of House leaders have moved congressional policy outcomes further from the preferences of the average citizen and toward the ideological extremes.2 This has created a situation in which most Americans hold nonideological, less extreme policy preferences but receive ideologically extreme policy proposals from their congressional representatives. This might be why public approval of Congress has averaged 20 percent since mid-2008.3 When Congress delivers policies more ideologically extreme than the public demands, this can cause members of the public to think less of the institution. For decades, scholars have believed that congressional performance had little effect on approval ratings of individual House members. Primarily, this is because it is difficult for constituents to sort through the vast amount of legislative activity in the House and attribute the appropriate blame or credit to their representatives.4 However, as members of Congress have become more politically polarized, constituents find it easier to hold their House members accountable for their performance in office.

Social identity theory predicts that people who share a

35

36

Religious Identity in US Politics

David Jones argues that as House members have become polarized, so have approval ratings, particularly for incumbents. In the previous few decades, the relationship between approval ratings of Congress as an institution and the approval ratings of House members from the party in the majority has strengthened. Divergent policy proposals between the two parties make it easier for the public to discern party positions on the issues and select the party with views closest to their own. When a House member is affiliated with a party that has controlled the legislative agenda, the public connects that member with the behavior of the party in the majority and uses that heuristic to guide evaluations of his or her performance.5 Many assume that party activists make it difficult for House members to represent their broader constituencies. Even if House members wanted to support moderate policy proposals, party activists might, in the next primary election, punish cooperation with opposing partisans. 6 Most current scholarship supports the view that party affiliation is the dominant force motivating public evaluations of House member performance. However, many of the findings supporting this view also suggest that partisanship is not the only force motivating support for or opposition to House members. For instance, Geoffrey Layman and Thomas Carsey’s work demonstrating the connection between congressional and public polarization shows that only strongly partisan members of the public altered their views to match those of party elites. Though aware of the partisan divide, politically attentive independents were significantly less likely to merge their social and cultural views with those of elites. Polarization of political leaders’ views also has little influence on the attitudes of weak partisan identifiers unaware of the positions taken by party elites.7 Thus, although partisanship might have a powerful influence on strong partisans’ evaluations of House member performance, it might not have as strong an influence for political independents or for weaker partisans who do not pay attention to politics. If religious identity influences constituents’ attitudes about their House representatives independently of party policies, perhaps partisanship might not be as constraining as the dominant view suggests. If so, House members could act independently of their party without being punished at the polls by those with whom they share a religious identity. The idea that social identities have a stronger influence than partisanship on views about how well people are represented is not new; the religion and politics literature simply has not explored how shared religious identities influence public evaluations of representatives. Most studies of descriptive representation concern institutional arrangements that can enhance minority representation in public policy processes.8

Descriptive Representation

Religious Identity and Support for Members of Congress

37

After minority groups get some of their own elected to legislative bodies, these elected officials can work to advance causes important to the minority groups they represent. In this way, descriptive representation can lead to substantive representation.9 Descriptive representation yields positive outcomes even when it does not lead to substantive representation. An increase in minority representation often leads to diverse kinds of bills being introduced, which alters the policy agenda.10 In addition, descriptive representation influences the issues discussed on the floor. For example, women as a minority in Congress have used their floor time to speak against stereotypes of welfare mothers as irresponsible and incapable of providing for their families and might continue to feel a responsibility to advocate for the interests of poor, minority women. Thus, members of minority groups are more likely to have their voices heard when their representatives win elections. Laws might not change immediately, but those in power are introduced to the concerns of minority groups.11 In turn, this representation influences minority attitudes and behavior. In the United States, increased minority representation is associated with higher public evaluation of the legitimacy of Congress. People are more likely to think their views are being represented in the legislature and to participate in politics when represented by members of their own minority groups.12 In part, this is because minority legislators work to mobilize new minority voters and include them in the political process, and more minority representation increases minority trust in government.13 Combined, these factors lead to greater minority approval of the elected officials from minority groups.14 Representation is more than policy congruence. People who trust their elected representatives are more likely to support them even when they disagree about the particulars of policy.15 Because of this, members of Congress often develop a “home style” orchestrated to create greater trust among their constituents. 16 This style consists of language, apparel, symbols, and mannerisms familiar to constituents in the members’ home districts and is designed to convince them that their representatives are like them. People who perceive commonality between themselves and those who represent them are more likely to trust and support them notwithstanding divergence on policy preferences. Although the dominant view of descriptive representation emphasizes the advancement of minority group interests, the principles are general enough to apply to any group dynamics. Richard Fenno demonstrates that members of Congress spend considerable effort fitting in with large groups in their home district. It makes sense that ambitious politicians would try to elicit approval from any group of voters in society. This allows them to develop support that does not depend on specific policy

38

Religious Identity in US Politics

outcomes. Moreover, it makes sense that the same psychological processes that lead minority groups toward greater trust in elected officials from their social groups would lead to greater public support for anyone from any social group. Descriptive representation functions even when group membership is not readily visible. Even when people can conceal group membership in the closet, members of the social group are more supportive of elected officials who share their groups’ identities.17 For many, their religious identity is less about what they believe and how they practice their religion and more about the churches into which they were born as children.18 As many as three-quarters of North American Christians do not attend church weekly, and in Europe the figure could be as high as 90 percent. For these Christians, their religious identity comes not from shared beliefs or practices but from being raised as members of a church. Many people, when interviewed, described religion as “an ascribed identity from which they could not disassociate themselves.”19 The salience of their religious identity in a particular context depends upon the strength of their religious commitment and the costs and benefits associated with their religious identification at a given moment. They might not ever attend church, yet feel strong identification with their religion because it was fundamental to their childhoods or home experiences. Similarly, religious identity might weaken among strict adherents to particular faiths in contexts where their religions do not seem relevant. In early US political life, it was difficult to be elected to public office without affiliating with a religious group.20 Although religions have a strong history of independently influencing US political outcomes, recently members of religious groups have tended to affiliate with one political party or the other.21 Thus, religious identities are often relevant in US political life. One of the consequences of this is that the public uses religious affiliations as cues for the political ideologies of candidates and elected officials. For example, as American Catholics have become more Republican, the public has come to see Catholic candidates as more conservative; the same is true of evangelical Protestants.22 These effects are strongest among the most religiously committed. The stronger their identification with their religious traditions, the more likely they are to base their evaluation of elected officials’ political ideologies on the officials’ religious identities rather than their voting records.23 Religious identities can have a politically relevant, independent influence on public attitudes about elected officials. In US society, we should expect that most of the time, partisan identities do not conflict with religious social identities.24 As such, it is likely that attitudes motivated by reli-

Religious Social Identities and Representation

Religious Identity and Support for Members of Congress

39

gious social identities will appear similar to those motivated by partisan social identities. However, because the two identities are distinct and can be more or less dominant in different contexts, it is possible to distinguish support motivated by one from support motivated by the other. Given what we know about descriptive representation, partisan identities should influence trust in and approval of elected officials who share those partisan identities. Partisan identities should not influence trust in or approval of elected officials derived from a shared religious identity. These attitudes are motivated by religious social identity. Controlling for partisanship, I expect higher public trust in and approval of elected officials who share religious identities with their constituents. In addition, because strong religious identification makes it more difficult to abandon a religious identity in favor of other socially relevant identities such as partisanship, I expect strength of religious identification to moderate the relationship between shared religious identities and trust in or approval of elected representatives. Shared religious identities are associated with stronger approval of representatives among those for whom religion is important, whereas shared religious identities will matter less in the positive evaluations of those for whom religion is not important. The first set of analyses uses data from the 2010 Cooperative Congressional Election Study (CCES). This specific dataset came from the contribution of thirty teams to a national sample of 55,400 cases. The questionnaires used by these teams comprised half of the questions, according to common content, designed to study congressional elections and representation in a large-scale national survey system. The other half of the surveys contained questions developed by each individual team. The data set used in this research is devoted to those questions of the common content. Two sets of interviews were conducted by the research teams. The first was a pre-election wave conducted in October 2010 that gauged issue preferences, knowledge of the candidates, some demographics, and voter intentions. The postelection wave was conducted during the two weeks following Election Day. The CCES common content data include five parts: sample identifiers (including state and congressional district), profile questions (largely demographic), pre-election questions, postelection questions, and contextual data (including candidate names and parties, election results, and roll call votes).25 The correspondence between the religious identity of members of the House of Representatives and their constituents is the focus of this study. The CCES asked for detailed information about the specific religious identity of each respondent. Members of Congress do not typically provide such

The Evidence

40

Religious Identity in US Politics

detail when describing their own religions. The religion of every member of the House of Representatives was reported by the Pew Research Center in 2008 for the 111th Congress.26 Of the 534 members of the 111th Congress, only 5 did not specify any religious affiliation, one was atheist, and none described himself or herself as religiously unaffiliated, Hindu, or Jehovah’s Witness. Fortunately, the CCES categorized each respondent’s religious identity with the identical codes used by Pew to categorize the religious identities of House members. That is, after asking for specific denominational information, the CCES coders created a variable that matches the Pew study and coded each respondent within a religious category, as displayed in Table 3.1. For instance, a respondent who selected Greek Orthodox, Russian Orthodox, Orthodox Church in America, Armenian Orthodox, Eastern Orthodox, Serbian Orthodox, or Other Orthodox was classified as Eastern Orthodox by both the CCES and Pew. Members of Congress such as John Garamendi who described themselves as nondenominational Christian, as well as those who identified as Christian Science, were coded as “something else.” Members of Congress who refused to identify with a religious tradition were coded as missing values, and those who described themselves as “atheist,” “Unitarian,” or “none” were categorized as “nothing in particular.” Of course, this coding scheme is far from perfect, particularly when it comes to Protestants; evangelical members of Congress are treated the same as Episcopalians. However, the data easily distinguish between Mormons, Jews, Catholics, Buddhists, Muslims, and those without any religious affiliation. The categorized religions of US representatives were then matched with the religions of the respondents in the CCES data. The variable “religion match” is coded “yes” if survey respondents have the same religion as their members of Congress and “no” if not. Table 3.1 shows the raw counts and percentages of respondents in each religious category who happen to share religious identities with their US representatives. Only one Muslim and four Eastern Orthodox respondents happened to have House members that shared their religious identities. Only three members of Congress identified as Unitarian—Pew labeled these “nothing in particular.” Respondents who reported a belief in Native American religions and those who called themselves “spiritual, but not religious” were classified as matching if their members of Congress were Unitarian; there were 100 respondents in this category. Although a strong majority of those who shared religions with their House members identified as Protestant (75 percent), there is quite a bit of diversity among religious traditions shared between House members and their constituents. There is at least as much diversity as one would find in studies of race and descriptive representation. To measure shared partisanship and gender, I follow the same method of matching respondents’ reported party identification and gender with

Table 3.1

Matching House Members’ Religious Affiliations with Those of Their Constituents

Respondent Matches House Member No

Yes

Total

Missing 527 1.52% 59 0.3% 586

Buddhist 330 0.95% 20 0.1% 350

Eastern Orthodox 570 1.64% 4 0.02% 574

Jewish 2,455 7.08% 227 1.14% 2,682

Mormon 1,117 3.22% 382 1.92% 1,499

Muslim 267 0.77% 1 0.01% 268

Nothing in Particular 462 1.33% 100 0.5% 562

Protestant 15,953 46% 14,800 74.54% 30,753

Sources: CCES 2010; Wikipedia; Pew 2008. Note: Entries comprise the raw count of respondents in the category; percentages are calculated by row.

Roman Catholic 11,234 32.39% 4,148 20.89% 15,382

Something Else 1,765 5.09% 114 0.57% 1,879

Total

34,680 100% 19,855 100% 54,535

41

42

Religious Identity in US Politics

those of their House representatives. In addition, I include controls for the dominant demographics that influence political attitudes. Party identification is coded 1–3, with 1 representing “Democrats,” 2 for “Republicans,” and 3 for “Independents”; political ideology is scaled 1–5, with 1 representing “very liberal.” I also include education levels coded 1–5, where 1 represents “less than high school” and 5 “graduate school of some kind.” Race has seven categories, including those who identify as multiracial. I expect that people with a strong religious identity are more likely to approve of House members who share their religious identity than are those for whom religion is not important. I estimate this interaction in later models, so I include it here in the baseline model. It is coded 1–4, where 4 represents a response that religion is “very important.” The dependent variable is the dichotomized response to the question, “Do you approve of the way [House Member’s Name] is doing their job?” Table A3.3 in Appendix 3 displays the results of three probit regression models that predict approval ratings of the US representatives. As I expected, the gender match and party identification match variables are statistically significant and positive. The coefficient for the party identification match variable is much larger than the coefficients for the religion match and gender match variables. Most of the control variables are statistically significant and in my expected direction, and the model-fit statistics suggest that this model accounts for a reasonable amount of variation in the approval ratings of House members. In addition, the religion match is statistically significant and positive. If constituents identify with the same religion as do their representatives, they are more likely to approve of those House members. The second model includes two kinds of interaction central to the overall argument. The first is interaction between the importance of religion and the match on religion between House members and their constituents (hereafter called religion match). In the baseline model, those for whom religion is more important are more likely to approve of their House members, all else being equal. I expect shared religious identity to moderate this relationship. Those with a strong religious identitiy are more likely to approve of House members who share their religious identity than are those with a weak religious identity. In addition, religious and partisan identities have become closely aligned in US politics, which raises the possibility that what appears to be religious descriptive representation might be partisan-based representation.27 To account for this, I include an interaction between respondents’ party identifications and their religion match with House members. In the second model, most of the covariates have identical coefficients and standard errors, as in the first model. Additionally, the model-fit statistics suggest that this model is practically the same as the first. The interaction between religion match and respondent party identification fails to

Religious Identity and Support for Members of Congress

43

achieve statistical significance. Yet the gender, religion, and party identification match variables retain statistical significance. These results suggest that the relationship between shared religious identities and approval of House representatives is not moderated by partisan identity. Partisan representation is distinct from religion-based descriptive representation. However, the interaction between importance of religion and religion match is statistically significant. The relationship between shared religious identity and approval of House members is moderated by the strength of religious identification. The stronger the religious social identity, the more having members of Congress with a shared religious identity influences approval of them. These findings suggest that religious identity has an influence on congressional approval ratings distinct from that of partisan identity. I would like to know something about the size of the effect of religious descriptive representation relative to the influence of gender- or partisanbased representation. Obviously, partisan representation is not strictly identity based, so I expect party identification match between constituents and their US representatives to have a much larger effect. The third model includes the two interactions from the previous model as well as interactions between respondent gender and gender match with House members, along with party identification of respondents and party identification match with their US representatives. The dominant view of gender-based descriptive representation is that female constituents are more likely to approve of their House members when they are female, but gender is not as powerful a motivator of public approval of representatives as race.28 I find some evidence for this in the models. In the baseline model, the coefficient for gender match is statistically significant and positive, suggesting that people are more likely to approve of House members of the same gender. In this third model, however, the interaction between gender match with representatives and gender of respondents is not statistically significant. This result suggests that women might be more likely to approve of female House members, but the same is also true for men. Consistent with my expectations, partisan identity moderates the effect of the influence of party match and approval of House representatives. The relationship between party match and approval depends on the respondents’ partisanship. Democrats and Republican respondents are significantly less likely to approve of their cross-partisan House members and significantly more likely to approve of their sameparty House members than are Independents. These two additional interactions slightly improve model fit but do not cause substantive changes to the model. The coefficient for the interaction between importance of religion and a shared religious identity is significant at the 0.01 level, and the size of the coefficient decreases slightly. Thus, although partisan-based representation is an important explanation for

44

Religious Identity in US Politics

approval of House members, it does not tell the entire story. Religious identity has a significant, independent influence on support of US representatives that changes depending on the strength of the religious identity. Just how large is this effect? With a sample size in excess of forty-seven thousand respondents, statistical significance can be achieved with a small effect. Figure 3.1 plots the predicted approval of House members for each level of religious importance for respondents by shared religious identity. The two largest differences shown in Figure 3.1 are among respondents for whom religion is “not at all important” and those for whom religion is “very important.” Among those for whom religion is “very important,” the predicted probability of approving of their representatives is 0.55 if they share their House members’ religious identity and 0.50 if they do not. By contrast, among those for whom religion is “not at all important,” the predicted probability of approving of their House members is 0.46 if they share religious identity and 0.43 if they do not share a religious identity with their House representatives. The probability of approval of House members decreases by 5 percentage points (from .55 on the right side of the figure to .50 on the left side) if constituents for whom religion is “very important” share a religious identity with those representatives, whereas approval among those for whom religion is “not at all important” increases by 3 percentage points (from .43 on the left side of the figure to .46 on the right side) if they share a religious identity with their House members. This

Probability Probability of A Approval pproval of H House ouse M Member ember

Figure 3.1 Importance of Religion and Influence of Shared Religious Identity on Congressional Approval Ratings

0.55 0.5 0.45 0.4 0.35 0.3

N Not ot aatt all all iimportant mportant

Not Not too too important important

Somewhat Somewhat important important

V Very ery iimportant mportannt

N Not ot aatt aall ll iimportant mportant

Not N ot too too iimportant mportant

Somewhat Somewhat iimportant mportannt

V Very ery iimportant mportant

Ye Yess

No N o ligious Ide Shared S hared Re Religious Identity ntity

Note: Points are the model-predicted probability of approval of House members; lines are the 95 percent confidence interval.

Religious Identity and Support for Members of Congress

45

suggests that the importance of religion is a key indicator of the salience of religious identity to approval of elected representatives. In all, the substantive difference in House member approval ratings attributed specifically to a shared religious identity seems small. However, compared with the size of gender-based descriptive representation effects, the influence of religion-based descriptive representation is large. It would be difficult for any form of identity-based descriptive representation to rival that of partisanship in today’s political climate, but the findings presented here suggest that for many people, religion-based descriptive representation is a reality. Again, I reiterate that the influence of shared religious identities on approval ratings of House members is substantively large and distinct from partisan, or ideological, considerations. How do these findings compare with those regarding presidential approval? Though presidential approval is different from approval of House members, presidents are quite concerned with improving public evaluations of their performance. Presidents have a particular interest in generating public support based on trust, rather than public approval derived from shared policy positions, because it gives them considerable freedom to formulate public policy proposals.29 This is precisely the kind of public support I would expect a shared religious identity to engender, just as demonstrated in the previous analyses. An identical approach was used to examine the relationship between the strength of religious social identities and shared religious identity with President Barack Obama on presidential approval evaluations. In addition, participants in the 2016 survey were asked to identify Obama’s religious affiliation and whether they approved of his performance in office. A Gallup poll conducted during the same week as the 2016 survey showed that 46 percent of Americans disapproved of Obama. In the CCES 2016 survey, 47 percent said they did not approve of Obama, 12 percent had no opinion, and 41 percent said that they approved. Consistent with other surveys conducted at the time, 31 percent of respondents believed Obama to be Muslim, 35 percent correctly identified him as Protestant, and the rest were split among other religious identities.30 I created a variable indicating whether respondents thought they shared a religious identity with Obama. When social identities can be concealed, their influence on positive evaluations depend on what the individuals perceive to be true. If Hindus incorrectly believe Obama to be Hindu (as did three people in this sample), their perception of a shared social identity influences their attitudes even if it is not factually accurate. Table A3.4 (in Appendix 3) displays the full results of two logistic regression models predicting the probability of approval of Obama with the standard control variables and the religion match variable. The second of the two models includes an interaction between the strength of religious identity and sharing that religious identity with the president.

46

Religious Identity in US Politics

For simplicity, I present only Figure 3.2, plotting the predicted probability of the interaction, in this chapter. How did religious identity influence approval of Obama? Respondents for whom religion is “very important” were twice as likely to approve of Obama if they shared his religious identity. Among strong religious identifiers, the probability of approving of Obama was only 0.23 when they did not share religious affiliation with him. If they did share religious identification with Obama, the probability that a strong religious identifier would approve of Obama was 0.46. In contrast, respondents’ shared religious identity with Obama had virtually no effect on approval among those with weak religious identification. The model predicts that respondents with weak religious identification would be less likely to approve of Obama if they had shared a religious affiliation than if they did not, although this difference is not statistically significant. Among weak religious identifiers, shared religious identity had no influence on the probability of their approval of Obama. These results suggest that religious identity can have a significant positive influence on public evaluations of political leaders. Because so many respondents did not know the religious identity of President Obama, and many others were completely misinformed about his religious identity, I am cautious about making too strong a claim with this study alone. However, considering that two studies each found that respondents are more likely to approve of

Figure 3.2 Importance of Religion and Influence of Shared Religious Identity on Presidential Approval Ratings

Source: June 2016 survey. Note: Bars represent the model-predicted probability of approving of Obama; lines are the 95 percent confidence interval.

Religious Identity and Support for Members of Congress

47

their elected officials when they share religious identities with them, I am optimistic that future work will consistently replicate these findings. The preceding analyses have a few major limitations. Some of the findings are based on a survey conducted during the 2010 congressional election. It is possible that both partisanship and reported religious identities were primed by the election environment. It is also possible that people were thinking more about the job performance of their House representatives than they normally do. The size of the effects I present might have been artifacts of the election calendar and could be smaller during nonelection seasons. Moreover, I do not know whether respondents to the CCES were actually aware that their members of Congress might share their religious identities. Given how little most people know about politics, it is possible that respondents would not know the religious identities of their House representatives.31 Finally, religion and politics have become intertwined in US politics. Because some people use religion as a cue for the political ideology of House members, it is possible that these findings are simply functioning as an instrumental variable for political ideology. In Chapter 4 I continue the analysis using a case that overcomes some of these limitations. One view is that religions in the United States are so closely aligned with partisan identities and political ideologies that we can no longer distinguish an independent influence of these religions on public attitudes. The evidence presented here suggests that although the dominant political parties in the United States count on the support of adherents to specific religious traditions, religious identity has a party-independent influence on public attitudes about elected officials. This result supports continued interest in identifying, specifying, and describing the contexts in which religious identity influences institutional, candidate, and policy support in the United States. The descriptive representation model is an appropriate way to think about the influence of religious identity on public support for elected officials. Religions form moral codes for believers that shape how they decide what is and is not acceptable behavior, how society should be organized, and how elected officials should behave. In addition, religious identities have a powerful, independent influence on how believers relate to other people in society. The findings I present in this chapter suggest that religious identities influence how people evaluate the performance of House representatives. Finally, descriptive representation influences attitudes in a variety of social groups. The same psychological processes that influence support for members of minority groups affect the attitudes of members of majority groups. People are more likely to trust others similar to themselves. Shared group identity is one cue that an elected official can be trusted to look out

Conclusion

48

Religious Identity in US Politics

for the needs of group members. When social identities collide, attitudes are influenced by the dominant social identity. The stronger the identification people have with their group identities, the more likely those identities are to influence their evaluations of elected officials who share their group identities. 1. Henri T. Tajfel and John C. Turner, “The Social Identity Theory of Intergroup Behavior,” in Psychology of Intergroup Relations, ed. by Stephen Worchel and William G. Austin (Chicago: Nelson-Hall, 1986), 7–24; Naomi Ellemers, Russell Spears, and Bertjan Doosje, “Self and Social Identity,” Annual Review of Psychology 53, no. 1 (2002): 161–186. 2. Marc J. Hetherington, “Putting Polarization in Perspective,” British Journal of Political Science 39, no. 2 (2009): 413–448. 3. Andrew Dugan, “U.S. Congress and Its Leaders Suffer Public Discontent,” Gallup, http://www.gallup.com/poll/184556/congress-leaders-suffer-public-discontent .aspx. 4. Gary C. Jacobson, The Politics of Congressional Elections, 7th ed., Longman Classics in Political Science (New York: Pearson Longman, 2009). 5. David R. Jones, “Partisan Polarization and Congressional Accountability in House Elections,” American Journal of Political Science 54, no. 2 (2010): 323–337. 6. Robert G. Boatright, Getting Primaried: The Changing Politics of Congressional Primary Challenges (Ann Arbor: University of Michigan Press, 2013). 7. Geoffrey C. Layman and Thomas M. Carsey, “Party Polarization and ‘Conflict Extension’ in the American Electorate,” American Journal of Political Science (2002): 786–802. 8. Susan A. Banducci, Todd Donovan, and Jeffrey A. Karp, “Minority Representation, Empowerment, and Participation,” Journal of Politics 66, no. 2 (2004): 534–556; Lawrence Bobo and Franklin D. Gilliam, “Race, Sociopolitical Participation, and Black Empowerment,” American Political Science Review 84, no. 2 (1990): 377–393; Claudine Gay, “The Effect of Black Congressional Representation on Political Participation,” American Political Science Review 95, no. 3 (2002): 589–617; Sue Thomas, “The Impact of Women on State Legislative Policies,” Journal of Politics 53, no. 4 (1991): 958–976. 9. Grace Hall Saltzstein, “Black Mayors and Police Policies,” Journal of Politics 51, no. 3 (1989): 525–544; Saltzstein, “Female Mayors and Women in Municipal Jobs,” American Journal of Political Science 30, no. 1 (1986): 140–164. 10. Kathleen A. Bratton, “The Effect of Legislative Diversity on Agenda Setting, Evidence from Six State Legislatures,” American Politics Research 30, no. 2 (2002): 115–142; Donald P. Haider-Markel, “Representation and Backlash: The Positive and Negative Influence of Descriptive Representation,” Legislative Studies Quarterly 32, no. 1 (2007): 107–133. 11. Kathleen A. Bratton and Kerry L. Haynie, “Agenda Setting and Legislative Success in State Legislatures: The Effects of Gender and Race,” Journal of Politics 61, no. 3 (1999): 658–679. 12. Banducci, Donovan, and Karp, “Minority Representation, Empowerment, and Participation.” 13. Gay, “Effect of Black Congressional Representation on Political Participation”; Bobo and Gilliam, “Race, Sociopolitical Participation, and Black Empowerment.”

Notes

Religious Identity and Support for Members of Congress

49

14. Banducci, Donovan, and Karp, “Minority Representation, Empowerment, and Participation.” 15. Heinz Eulau and Paul D. Karps, “The Puzzle of Representation: Specifying Components of Responsiveness,” Legislative Studies Quarterly 2, no. 3 (1977): 233–254. 16. Richard F. Fenno, Home Style: House Members in Their Districts (Boston: Little, Brown, 1978). 17. Haider-Markel, “Representation and Backlash.” 18. Abby Day, Believing in Belonging: Belief and Social Identity in the Modern World (Oxford, UK: Oxford University Press, 2011). 19. Day, Believing in Belonging, 80. 20. Alexis de Tocqueville, Democracy in America, 4 vols. (London: Saunders and Otley, 1835). 21. Geoffrey C. Layman, The Great Divide: Religious and Cultural Conflict in American Party Politics (New York: Columbia University Press, 2001). 22. Monika L. McDermott, “Religious Stereotyping and Voter Support for Evangelical Candidates,” Political Research Quarterly 62, no. 2 (2009): 340–354; McDermott, “Voting for Catholic Candidates: The Evolution of a Stereotype,” Social Science Quarterly 88, no. 4 (2007): 953–969. 23. Matthew L. Jacobsmeier, “Religion and Perceptions of Candidates’ Ideologies in United States House Elections,” Politics and Religion 6, no. 2 (2013): 342–372. 24. Layman, Great Divide. 25. Stephen Ansolabehere, “CCES Common Content, 2010” (Harvard Dataverse, 2012). 26. David Masci and Tracy Miller, “Faith on the Hill: 2008,” Pew Forum on Religion and Public Life (2008). 27. Layman, Great Divide; Jacobsmeier, “Religion and Perceptions of Candidates’ Ideologies in United States House Elections.” 28. Kathleen Dolan, “Symbolic Mobilization? The Impact of Candidate Sex in American Elections,” American Politics Research 34, no. 6 (2006): 687–704; Jennifer L. Lawless, “Sexism and Gender Bias in Election 2008: A More Complex Path for Women in Politics,” Politics and Gender 5, no. 1 (2009): 70–80; Lawless, “Politics of Presence? Congresswomen and Symbolic Representation,” Political Research Quarterly 57, no. 1 (2004): 81–99; Emily A. West, “Descriptive Representation and Political Efficacy: Evidence from Obama and Clinton,” Journal of Politics 79, no. 1 (2017): 351–355. 29. James N. Druckman and Lawrence R. Jacobs, Who Governs? Presidents, Public Opinion, and Manipulation, Chicago Studies in American Politics (Chicago: University of Chicago Press, 2015). 30. It might alarm some that so many respondents did not know Obama’s religious identity, so the analyses were conducted with and without those who believed him to be Muslim. It does not change the findings. 31. Michael X. Delli Carpini and Scott Keeter, What Americans Know About Politics and Why It Matters (New Haven, CT: Yale University Press, 1996).

4 Religious Identity and Political Trust: The Obama Era

strong religious social identity have more positive attitudes toward other people who share their religious identity. According to this hypothesis, they should be more trusting of members of their own religions, more likely to associate with them, and less likely to view them suspiciously. Thus, according to this hypothesis, people who share a religious social identity with the US president should be more likely to trust him or her even when they do not share a partisan identification with that president. I discuss in this chapter why, in an era of declining trust in government, this would be beneficial to US society. The US public is much less trusting of their government today than they were a half century ago. When asked about the biggest problem facing the country, respondents named the government number one in 2014 and 2015.1 In 1958, 73 percent of Americans polled reported that they trusted the government in Washington, DC, to do the right thing “most of the time.” However, by February 2014 only 24 percent of Americans surveyed reported the same level of trust in Washington. Partisanship explains some of the variation in reported trust in the federal government. Early in the George W. Bush administration—when Republicans controlled the House, Senate, and White House—trust in Washington among Republicans was close to the 1958 level. Similarly, in April 2010—when Democrats controlled the House, Senate, and White House—trust in Washington among Democrats was near 40 percent.2 Yet partisanship alone cannot explain the decades-long steady decline in trust the public expresses in the US government. Marc Hetherington and Thomas Rudolph argue that political polarization in the US electorate is largely responsible for the decline in political

The in-group favoritism hypothesis is that people with a

51

52

Religious Identity in US Politics

trust. They argue that as the two political parties in the United States have become more ideologically distinct, ordinary people from one party have developed strong, negative feelings toward members of the other party. When a representative from one party gives a speech, people of the opposing party have angry responses to policy proposals from the speech. These negative feelings toward opposing party identifiers motivate partisans to express less trust in government. In addition, these conditions cause the government to be less likely to pass significant legislation, which perpetuates the view that the government is dysfunctional.3 Individuals’ evaluations of government trust are complex. 4 Most scholars agree, however, that trust evaluations reflect a “basic evaluative orientation toward the government founded on how well the government is operating relative to people’s normative expectations.” Policy outputs certainly influence public trust evaluations; the Vietnam War and civil rights legislation eroded public trust in government among white respondents, but civil rights legislation increased public trust among African Americans.5 Institutional performance also influences trust in government. Some argue that public approval of a president strongly influences trust in government, whereas others have noted that both congressional and social service organization performance can influence public trust in government.6 Yet, these theories alone cannot explain why citizens in advanced democracies are becoming less trusting of both government leaders and government institutions.7 Russell Dalton argues that as democracies work to improve the conditions in their countries, public expectations of their governments change. Those least trusting of government tend to be those highest in education and social status. The cognitive skills that accompany greater education and wealth motivate people to expect more from their political systems. Public trust in government might be on the decline because expectations are rising among these populations.8 If so, declining political trust might be a feature of democratic systems of government, and scholars should become comfortable with governments operating without the trust of their constituents. Declining public trust is a problem for elected officials. Modern presidents commit substantial resources to identifying ways they can strengthen public impressions of their personal attributes. When the public expresses trust in the US president, this gives him or her greater latitude in policy negotiations and decisions. Especially when the politically attentive and mobilized public sorts itself into polarized political parties, the president has a harder time cooperating with the political opposition on salient policy matters. However, when the public trusts the president, they pay less attention to policy specifics and support the broader policy agenda.9 Consistent with my findings in Chapter 3 and the general argument of this book, shared religious identity might influence trust in elected officials.

Religious Identity and Political Trust: The Obama Era

53

Among the first survey items shown to predict trust in government was a battery of questions that measured the ethical qualities of elected officials. Respondents who thought elected officials were looking out for the interests of all were significantly more likely to trust government than those who thought politicians were looking out for themselves. In addition, respondents who thought hardly any elected officials were crooked were more likely to trust government than those who thought quite a few people running government were corrupt. For many people, trust in government is not much different from trust of individuals. When people think that another person is dishonest or self-centered, they are less likely to trust that person. Similarly, when trustworthy individuals run government, one might expect trust in government to increase. One need not suppose declining trust in government is a necessary feature of societal development. Perhaps most educated elites distrust governments because they are familiar with those who run the advanced democracies; that is, they distrust these politicians and distrust government for logical reasons. Regardless, those traits that engender trust on the interpersonal level ought to increase trust in government leaders. Recall that religious identity can blind people to the shortcomings of others who identify with their religion. If so, those with a strong religious identity might be more trusting of political leaders with whom they share a religious identity. For these people, political leaders are more than an elected official; they are a member of the religious community. Using a national survey from June 2015, I conducted another study to see how religious identity influences trust in government leaders. Clear Voice Research recruited 2,118 participants in its online panel to participate in a survey about political attitudes; of those, 1,935 completed the entire questionnaire. Respondents were 50.7 percent male, 80.6 percent white, 33.7 percent Democrat, and 24.7 percent Republican. Most of the respondents had completed at least some college (78.1 percent) and 86 percent reported an annual income of less than $100,000. Using language from Gallup surveys, I asked respondents to identify the religious identity of President Barack Obama. Much to my initial surprise, close to 25 percent of the respondents thought Obama was Muslim. By September 2015, surveys conducted following a Donald Trump rally found similar proportions of Americans believing Obama was Muslim, which alleviated my concerns about this survey. In addition, I asked respondents about their own religious affiliations. I match individual responses on these two questions without regard for the accuracy of the information about Obama’s religion. That is, if a Muslim thought Obama was Muslim, I count that as a shared religious identity because descriptive

The Evidence

54

Religious Identity in US Politics

representation is a matter of perception. Table A3.1 (in Appendix 3) shows the number and percentage of respondents whose religious identities matched that of Obama based on their perception of his religious identity. A handful of respondents who identified with minority religions also believed Obama shared their religious belief, but most people thought Obama was Protestant (correct), Catholic, nonreligious, or something else. I asked participants the standard question about how often they think they can trust the government in Washington to do what is right. In the next question I asked, “How much of the time do you think you can trust President Obama to do what is right?” Individuals’ evaluations of how much they trust government were influenced by policy-related concerns, institutional performance, partisanship, and public expectations. Trust in Obama also depended on public expectations of him, his performance, and some policy-related concerns. However, because the question of trust is about an individual president, these evaluations are less likely to be motivated by strict policy concerns and are more likely to be influenced by his personal characteristics. The correlation between responses on these two questions is moderate (r = 0.44), and twice as many people reported trusting Obama (39.66 percent) as they did trusting Washington (20.22 percent). This suggests that the two trust evaluations are distinct and influenced by different considerations. I expected people who shared a religious identity with Obama to be more likely to trust him, all else being equal. I follow the same estimation strategy for this study as I did for the project in Chapter 3. I predict trust in Obama with a multinomial probit regression model, with “never” as the baseline category. In addition, I control for the same demographics. Party identification is coded 1–3, with 1 representing Democrats, 2 for Independents, and 3 for Republicans; political ideology is scaled 1–7, with 1 representing “very liberal.” I also include education levels coded 1–5, where 1 represents “less than high school” and 5 “graduate school of some kind.” Race has six categories, including those who identified as multiracial. The importance of religion question is worded differently, but it is coded 1–4, where 1 represents a response that religion is “extremely important” and 4 indicates it is “not at all important.” The full results of the regression model are in Appendix 4. Because belief that Obama was Muslim was associated with extreme conservatism and opposition to Obama, I run the models both with and without respondents who thought Obama was Muslim, and the results do not change substantially. The reported models use the truncated data because the effect sizes are smaller. The coefficients in the first model are consistent across response categories and in my expected direction. All else being equal, respondents who shared a religious identity with Obama were significantly more likely to say they trusted him “just about always” and “some of the time” than to say they “never” trusted him; the same is true for liberals and

Religious Identity and Political Trust: The Obama Era

55

Democrats. As individuals became more liberal and the strength of their identification with the Democratic Party increased, they were significantly more likely to trust Obama. Older people were significantly less likely to trust Obama, and racial minorities were more likely to trust him than were white people.10 Finally, those for whom religion was important were less likely to say they trusted Obama “just about always”; they were more likely to respond that they “never” trusted him. Just how much did a shared religious identity influence public trust of Obama? Figure 4.1 displays the predicted probability of trusting Obama for racial categories and for shared religious identities. Conventional wisdom holds that African Americans overwhelmingly supported, approved of, and trusted Obama. All else being equal, the probability of African Americans reporting that they trusted Obama “just about always” is 0.22, “most of the time” is 0.49, “only some of the time” is 0.22, and “never” is 0.05. By contrast the probability of white people reporting that they trusted Obama “just about always” is 0.07, “most of the time” is 0.36, “only some of the time” is 0.45, and “never” is 0.13. As expected, racial identity is associated with quite a gap in trust of Obama. Respondents who shared a racial identity with Obama were much more likely to trust him than those who did not. The trust gap based on religious identity is quite substantial. All else being equal, people who shared a religious identity with Obama were 60 percent more likely to report trusting him “just about always” and 32 percent less likely to “never” trust him than were people who did not. Not all of these differences are statistically significant, but the overall pattern is consistent with theoretical expectations. It seems that when people shared a religious identity with Obama, they were more likely to trust him. The strength of a person’s religious identity moderates the relationship between a shared religious identity and trust in Obama. The main effect suggests that respondents who placed a higher importance on religion were less trusting of Obama, but this relationship changes depending on his perceived religious identity. Figure 4.2 displays the predicted probability of each response among those who believed they shared Obama’s religion and those who did not. Among those for whom religion is “extremely important,” the probability of trusting Obama “just about always” is nearly twice as high if the respondent shared a religious identity with him than it is if they did not. Similarly, the probability of trusting Obama “never” is twice as low among this group. In contrast, a shared religious identity did not change trust in Obama among those for whom religion was not an important part of their identity. Among these respondents, the predicted probability of “never” trusting Obama is identical regardless of whether they shared a religious identity with him. If they had the same religious identity as Obama, they were more likely to “always” trust him, but they were 15 percentage points less likely

Predicted Trust in President Obama, by Shared Religious Identity and Race

56

Figure 4.1

Probability Probability of T Trusting rusting O Obama bama

0.60 0.50 0.40 0.30 0.20 0.10 -

N Never ever ligious Ide ntity S Shared hared Re Religious Identity

O Only nly S Some ome of tthe he T Time ime Diffe ligious Ide ntity Different f rent Re Religious Identity

M Most ost of tthe he T Time ime A frican A merican African American

Ju Just ust A About bout A Always lways M Multi-Racial ulti-Racial

Source: June 2015 Survey. Note: Bars are the model predicted probability of trusting Obama; lines are the 95 percent confidence intervals. Total columns less than 1 because of rounding.

W hite White

Figure 4.2 Model-Predicted Trust in President Obama, by Shared Religious Identity and the Importance of Religion

P Probability robability of T Trusting rusting O Obama bama

0.60

D Different ifferent Re Religious ligious Ide Identity ntity

Shared Religious Identity

0.50 0.40 0.30 0.20 0.10 0.00

E Extremely xtremely Im Important portant

N Not ot aatt aall ll Im Important portant

E Extremely xtremely Im Important portant

N Not ot aatt aall ll Im Important portant

N Never ever

0.08

0.14

0.16

0.14

he T ime O nly S ome of tthe Only Some Time

0.34

0.51

0.34

0.40

M ost of tthe he T ime Most Time

0.35

0.26

0.38

0.41

JJust ust A About bout A Always lways y

0.23

0.09

0.12

0.05

Source: June 2015 Survey. Note: Bars are the model predicted probability of trusting Obama; lines are the 95 percent confidence intervals. Total columns less than 1 because of rounding.

57

58

Religious Identity in US Politics

to trust him “most of the time.” The consistency of this pattern of findings, combined with the magnitude of the predicted differences in trust of Obama among those for whom religion was an important part of their identity, provides substantial evidence in support of the argument that religious identity influences trust in public officials. Were there clear partisan differences? Did a shared religious identity moderate the influence of party identification and trust in Obama? Figure 4.3 displays the model-predicted support for partisans who did and did not share a religious identity with Obama. For Democrats, the probability of trusting Obama did not change much when respondents shared a religious identity with him. Among Independents, those who shared a religious identity with Obama were twice as likely to report trusting him “just about always” as those who did not. Among Republicans the difference is substantial. The predicted probability of a Republican trusting Obama “just about always” is 0.12 if they shared a religious identity with Obama, whereas it is 0.02 if they did not. It might seem like a small difference, but these numbers mean that Republicans were six times more likely to trust Obama if they shared a religious identity with him than if they did not. They were nearly half as likely to “never” trust Obama than were Republicans who did not share a religious identity with him. The influence of religious descriptive representation on trust in Obama was much larger among Republicans than among Democrats and Independents. In summary, the findings presented here demonstrate that religious identity influences trust in a president. The dominant view of descriptive representation is that public officials who share demographic characteristics with their constituents can represent these groups even when they disagree about policy specifics. The central question of this chapter is whether religious identities are strong enough to allow for descriptive representation. The second question is whether religion-based descriptive representation is independent of racial and partisan identities. These findings suggest that religious identification has a substantial independent influence on support for elected officials. It is difficult to think of a group that would be less likely to trust Obama in 2015 than Republicans. Yet Republicans were substantially more likely to trust him if they shared a religious identity with him than if not. When presidents share a religious identity with their constituents, it engenders trust that can give them greater latitude and independence in setting policy agendas. Some argue that democratic development is associated with higher public expectations of their system of government.11 When public expectations rise, but government performance does not, people are less trusting of gov-

Conclusion

Figure 4.3

Model-Predicted Trust in President Obama, by Partisan Identity and Shared Religious Identity SharedShared ReligiousReligious Identity

Identity

0.70 S

0.60 Probability Probaability of T Trusting rusting O Obama bama

Probability Probaability of T Trusting rusting O Obama bama

Different iffeRe rentliRe Religious ligious Ide Identity D Different iffereD nt Religious gious Ide Identity ntntitityy

0.50 0.40 0.30 0.20 0.10 0.00

D Democrat emocraat

Inde Independent pendent

Re Republican publican n

D Democrat emocratt

Inde Independent pendent

Re Republican publican

N Never ever

0.03

0.19

0.26

0.05

0.13

0.15

Only Some Time O nly S ome of tthe he T ime

0.23

0.45

0.57

0.24

0.44

0.59

Most Time M ost of tthe he T ime

0.57

0.33

0.15

0.49

0.36

0.13

JJust ust A About bout A Always lways y

0.17

0.03

0.02

0.21

0.06

0.12

Source: June 2015 Survey. Note: Bars represent the model-predicted probability of trusting Obama; lines are the 95 percent confidence intervals. Estimates from models are displayed in Table A4.1 (in Appendix 4).

59

60

Religious Identity in US Politics

ernment because it fails to meet public expectations. This view predicts a global decline in public trust in government as more countries become democratic and the performance of existing democracies improves. Others argue that political polarization is causing the US public to become less trusting of their government. As the two major political parties have become more ideologically distinct, members of each party have developed increasingly negative views of those in the opposing political party. Furthermore, many Americans have strong negative emotional reactions to public addresses by elected officials from the opposing political party. As the major political parties become more ideologically distinct, and the public sorts itself into the political party that best fits their political ideologies, this view predicts declining trust in government officials from the opposing political party. In this chapter I clearly demonstrate the power of religious identity to solve each of these problems. Religious identities not only influence attitudes about government officials separately from partisan identities but they might have a stronger influence than partisanship on trust in government officials. In the 2015 national survey, Democrats who did not share a religious identity with Obama were equally likely to trust Obama to do what was right “just about always” as were Republicans who shared a religious identity with him. In addition, Republicans who shared a religious identity with Obama were much less likely to report “never” trusting him than were Republicans who did not share a religious identity with him. Religions have more positive influences on the political attitudes of Americans than the extant literature suggests. Although religious identities might be the source of political conflict in the United States, they are also associated with some positive elements of political life in the nation. More powerfully than gender and equally influential as race, religious identity shapes public approval of and trust in elected officials. At a time when it seems partisanship, polarization, and conflict are destined to diminish public evaluations of elected officials, religious identity can motivate partisans to express trust in and approval of opposing partisans. 1. Lydia Saad, “Government Named Top U.S. Problem for Second Straight Year” (Gallup, 2016) https://news.gallup.com/poll/187979/government-named-top -problem-second-straight-year.aspx. 2. “Public Trust in Government: 1958–2014,” http://www.people-press.org /2014/11/13/public-trust-in-government/. 3. Marc J. Hetherington and Thomas J. Rudolph, Why Washington Won’t Work: Polarization, Political Trust, and the Governing Crisis (Chicago: University of Chicago Press, 2015).

Notes

Religious Identity and Political Trust: The Obama Era

61

4. Jack Citrin and Donald Philip Green, “Presidential Leadership and the Resurgence of Trust in Government,” British Journal of Political Science 16, no. 4 (1986): 431–453. 5. Arthur H. Miller, “Political Issues and Trust in Government: 1964–1970,” American Political Science Review 68, no. 3 (1974): 951–972. 6. Citrin and Green, “Presidential Leadership and the Resurgence of Trust in Government.” 7. Pippa Norris, Critical Citizens: Global Support for Democratic Government (Oxford, UK: Oxford University Press, 1999). 8. Russell J. Dalton, Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies, Comparative Politics (Oxford, UK: Oxford University Press, 2004); Dalton, “The Social Transformation of Trust in Government,” International Review of Sociology 15, no. 1 (2005): 133– 154. 9. James N. Druckman and Lawrence R. Jacobs, Who Governs? Presidents, Public Opinion, and Manipulation, Chicago Studies in American Politics (Chicago: University of Chicago Press, 2015). 10. Results from this model are in Appendix 4. 11. Dalton, “The Social Transformation of Trust in Government”; Norris, Critical Citizens.

5 Christian Identity and Anti-Atheist Intolerance

is rooted in political competition. In this chapter, I demonstrate that the current alignment of religious groups with the major political parties elicits out-group antipathy from religious Americans with a strong partisan identification that is manifest in intolerant attitudes toward nonbelievers. Yet strong religious identifiers express similar aversion toward members of other groups aligned with the Democratic Party, and those with weaker religious identity are averse to members of groups aligned with Republicans. My analyses do not rule out the plausibility of competing explanations, but they are the first of which I am aware to use measures of religious identification that match the theoretical construct. To the extent that religious social identification is the cause of antipathy toward atheists, intergroup contact theory suggests that greater contact or political cooperation between religious and nonreligious Americans can diminish this religious divide.1 Although religious Americans might be among the most charitable, trusting, civically engaged, and tolerant people in the country, they are also the most distrusting of atheists.2 To the extent that existing explanations acknowledge the role of social identities in tolerance, they do not specify how religious social identity motivates negative attitudes toward nonreligious Americans. Working in the 1940s and 1950s, Samuel Stouffer found that people who attended church frequently were not supportive of civil liberties for communists.3 Although this finding was interesting, it is not clear whether this statistical association is conceptually meaningful. Over the previous half century, scholars have argued about the appropriate measures and conceptualizations of religions to discover whether religions breed political intolerance.4

Much of the religious antipathy toward atheists in US society

63

64

Religious Identity in US Politics

Current scholarship on the relationship between religion and political tolerance emphasizes how religious believing, belonging, and behaving influences tolerance of minority groups in the United States. Marie Eisenstein and April Clark combined General Social Survey (GSS) data from 1987 with the 1987 Freedom and Tolerance Survey to test each competing view of religious intolerance in the United States. When they controlled for socioeconomic variables such as age, education, and income (something Stouffer did not include), they found that religious involvement and religious behavior were not associated with political intolerance. Instead, they argued, the association between political intolerance and religion is explained by belief.5 The relationship between religion and political intolerance completely disappears if a different measure of intolerance is used as the dependent variable. John Sullivan and colleagues argued that the original list of groups used to measure political tolerance primarily contained left-wing targets (atheists, communists, and socialists). When scholars use that list, it biases the measure of political tolerance because people on the left who might be intolerant of other groups in the United States are erroneously classified as politically tolerant. Instead, Sullivan and colleagues advocated for the “least-liked” approach. This approach lists many more groups in the United States and asks respondents to identify groups they dislike. The tolerance questions are asked only about the groups respondents have indicated that they dislike.6 Eisenstein and Clark note that when this measure is used as the dependent variable, “the effect of religion virtually vanishes.” This is because religious Americans vary in their affinities toward other groups in the United States. Stated differently, the source of political tolerance in the United States is the groups with which people identify and the way members of these groups view each other, not their religious beliefs. Eisenstein and Clark reach a similar conclusion, though they lack a good measure for the strength of religious identity.7 In similar analyses, James Gibson concludes that the active ingredient by which religion contributes to intolerance is through particular religious beliefs.8 Gibson is adamant that religious political intolerance comes from greater involvement with religions, not from group-based threats, dogmatism, or lack of support for democracy. Other scholars find that belonging and behaving have a strong influence on political intolerance. Daniel Cox, Robert Jones, and Juhem Navarro-Rivera report that nearly half of Americans who attend church services weekly or more express intolerance of atheists. They find that the strongest predictor of intolerance of atheists is affiliation with a Christian religious denomination. Even when controls for negative affect toward atheists are included in the model, affiliating with a Christian religion has a strong, independent influence on intolerance of atheists. Cox and colleagues conclude that religious

Causes of Religious Political Intolerance

Christian Identity and Anti-Atheist Intolerance

65

Americans’ expression of intolerance toward atheists is not motivated by negative affect; rather, intolerance is the result of competition between religious and nonreligious groups for members.9 Recent scholarship emphasizes how individuals’ values and traits influence their willingness to extend rights to nonreligious people. Pazit Ben-Nun Bloom and Marie Courtemanche argue that invoking feelings of disgust causes people to express intolerant views, though the longevity of the finding is unclear.10 Paul Djupe and Stephen Mockabee argue that when people hold religious values that emphasize distinctiveness, they are more likely to express political intolerance than when they value religious inclusiveness. 11 Joby Schaffer, Anand Sokhey, and Paul Djupe argue that this situation occurs because people with exclusive religious values are more likely to feel threatened by religious out-groups, which in turn motivates greater intolerance toward those groups.12 It is interesting that most of the literature on religious political intolerance uses the language of social identity theory and relies on concepts developed in social identity scholarship, yet none of the research exploring religious Americans’ antipathy toward atheists actually includes measures of religious social identity in the models.13 James Gibson and Amanda Gouws have long noted the important influence social identity has on political intolerance, yet Gibson excludes religious social identity from his models studying religious intolerance toward atheists.14 In the notes to the chapter “The Political Consequences of Religiosity: Does Religion Always Cause Political Intolerance?” Gibson states that sociotropic threat perceptions are the most influential type of threat (in the model), which implies that social identity concerns might play a key role in this process. He writes that the hypothesis was investigated, but the findings were too complicated to present in the chapter.15 I agree that the influence of religious social identity on antipathy toward groups in society is complex, but it provides the ideal theoretical lens through which future scholarship can make sense of the influence religion has on attitudes toward out-groups. Perceptions of threat are consistently the strongest (and least understood) predictors of political intolerance.16 Because threat is a multidimensional concept, Gibson argues that it can be difficult to isolate the conditions under which people perceive a threat. Yet perceived group-level threats tend to be stronger predictors of intolerance than do perceived personalized threats. It is thus logical that perceived threats to religious group identity would be strong predictors of antipathy toward out-group members. This is precisely what I would expect if people hold meaningful religious social identities.

Why Not Religious Social Identity?

66

Religious Identity in US Politics

Adapting Gibson and Gouws, strong religious in-group positive identification leads to strong religious out-group negative identification. This negativity leads to antipathy toward religious opponents, which heightens perceptions of threat from religious opponents and causes religious intolerance.17 If the logic so easily adapts one theory to another, why do scholars of religion and political intolerance adopt the language of social identity theory without including measures of religious social identification in their models? One reason is the methodological concerns to which Gibson alludes in his notes, which can be examined in two parts. First, mainline and evangelical Protestants are increasingly dividing along partisan lines. Nearly two-thirds of evangelical Protestants identify with the Republican Party, whereas mainline Protestants tend to split fiftyfifty between the Republican and Democratic Parties. Among African American Protestants, more than 75 percent identify with the Democratic Party. Among people who identify with minority religions in the United States (Mormon, Muslim, Buddhist, Hindu, Jewish), only Mormons strongly identify with the Republican Party. Now, consider the top five most disliked groups in the nation (Ku Klux Klan, radical Muslims, militarists, atheists, and people who favor abortion). Of these groups, the Ku Klux Klan made it into the top three most disliked by 72 percent of Americans, but no other group was in the top three for more than 42 percent of Americans. This means that most Americans disagree about which groups should be the most disliked.18 The common element in the most disliked groups in the United States is that (with the exception of the Ku Klux Klan) they are all affiliated with the political left. Current measures of intolerance inherently include elements of alternative social identities. If people strongly identify with a mainline Protestant religion, and they strongly identify as a Democrat, their rankings will look different from those of strong mainline Protestants who strongly identify as a Republican. The first reason it is difficult in these analyses to demonstrate the role of religious social identity regarding tolerance is that social identities determine individual responses to the first question (like or dislike), and none of the options provided is a clear out-group to religious identifiers except atheists. How members of religions rank the other groups will depend on the strength of their religious identities relative to their partisan identities. The dominant view suggests that mainly Christians (who hold exclusive religious values) are politically intolerant. Surely, this is more a result of measurement error than of social science theory. Political intolerance cannot be exclusively Christian. The second methodological challenge that leads to the erroneous conclusion that religious beliefs have a strong influence on political intolerance is that national surveys of Americans contain many more Christian respondents than identifiers with minority faiths. It is

Christian Identity and Anti-Atheist Intolerance

67

hard to locate large numbers of identifiers with minority religions in nationwide surveys, which makes it difficult to isolate the effect of identifying with a particular denomination on attitudes toward members of religious out-groups; the confidence intervals around the estimates are enormous. Many alleviate this problem by combining all religious non-Christians into a single category; others do not even include members of non-Christian religions in their analyses. Yet it is logical that strongly identifying with one religion would lead individuals to view members of some groups in the United States more like members of their in-group and other group members more like an out-group. Although people’s lists would be malleable as societal conditions increase intergroup competition, and individuals within each denomination would not have identical lists, it is reasonable to conclude that Hindus would construct similar lists to each other, as would Mormons, Muslims, Jehovah’s Witnesses, and so on. To the extent that this is true, I would expect Hindus, despite holding more egalitarian religious beliefs, to express greater intolerance toward members of groups who rank at the bottom of their lists.

How Partisan Identities Make It Difficult to Isolate the Influence of Religions

Politically, ideological sorting has changed the nature of elections in the United States. The aligning of ideological and partisan identities in US politics strengthens the connection between political social identities and political attitudes and behavior. This increases the salience of elections and their outcomes because for people with strong politically oriented social identities, politics is not simply about who will govern until the next election. Rather, elections become a competitive battle over scarce resources, which can elicit positive and negative emotions.19 Wouter Van den Bos and colleagues were interested in how strongly social identities influence competitive behaviors. In a series of experiments, they examined how strongly monetary outcomes relative to social identities influenced people’s willingness to overbid in an auction. They found that people were willing to incur substantial financial losses to win an auction when competing against a member of a rival group. The drive to win a competition against an opposing social group was more important than the motivation to preserve money by not paying more in an auction.20 Even when people are not directly involved in the competition, rooting for a winning team stimulates the positive emotion regions in the brain. By contrast, losing a competition to rivals stimulates negative emotions. Another study used a functional magnetic resonance imaging (fMRI) machine to observe brain activity while Red Sox and Yankees fans viewed baseball plays. When the rival team failed, participants were more likely to report positive

68

Religious Identity in US Politics

feelings, which were corroborated by corresponding neural activity. However, when the rival team succeeded, participants reported a desire to heckle, insult, throw food at, or threaten the rival team’s fans. Social identification influences both attitudes about and actions toward members of rival groups.21 Elections stimulate competitive partisan social identification. When it comes to political rivals, partisans behave like sports fans. The stronger their partisan identities, the more pleasure people receive at an electoral loss for a representative from the opposing political party, and the greater the sorrow at a loss for their own party.22 Beyond this, partisan loyalties are associated with negative feelings and hostility toward competing partisans. The same emotions that cause people to overbid to win an auction or to harm fans of the opposing team motivate partisans to view opposing partisans with hostility. In part because political parties compete over scarce resources and because competition fuels competitive emotions, Americans whose ideological and party preferences are aligned report greater anger toward and lower evaluations of people in the opposing political party.23 This response to competition is why it is difficult to isolate the influence of religious social identity on people’s views toward other groups in society. Religious Americans who affiliate with one political party have negative attitudes about members of all groups that affiliate with the opposing political party. Religion has nothing to do with this; rather, the effect of partisan identification is simply making it appear as if the strongly religious have negative views toward people in these groups. Partisan identities, not religious identities, are motivating the negative attitudes. In the remainder of this chapter I first demonstrate why the effects of religious social identities are often conflated with partisan social identities. After demonstrating that the distribution of religious identities in national surveys can lead to erroneous conclusions, I discuss how religious social identification explains antipathy toward religious out-groups, and I then present results from a unique series of experiments that isolates the effect. In June 2013 a national sample of nearly sixteen hundred Americans was recruited to participate in a political research study. In June 2014 another national sample of a thousand US adults was recruited to replicate some of the findings from the first study. Participants in each study appear to be representative of the US population demographically. Between 73 and 80 percent of the participants were white, and the proportion of racial minorities in the sample is similar to the proportion in the US population. Just over 41 percent of respondents identify as political Independents, with about 32 percent Democrats and nearly 27 percent Republicans. Appendix 5 contains a more complete description of the respondents.

Partisan Identities and Religious Social Identities

Christian Identity and Anti-Atheist Intolerance

69

Participants were asked several demographic questions. I included a series of items that measured inward commitment to religion. Robert Putnam and David Campbell report many findings related to the political behaviors of religious Americans, which were included in this survey to measure religiosity. Participants were asked how often they attend church, the importance of religion in their lives and their sense of self, the strength of their beliefs, and the frequency of their church attendance. Putnam and Campbell report that the strongest predictor of party identification is how frequently one prays, so it is also included in the survey.24 Consistent with previous findings, answers to these questions show high internal consistency and could be combined into a single measure for analyses; however, doing so is not consistent with the theory of religious social identification. Tables 5.1 and 5.2 display the strength of religious social identification by the strength of partisan attachment. More than half of the respondents to the 2013 survey for whom religion was “extremely important” identified with the Republican Party, whereas just more than a quarter of them identified with the Democratic Party. Conversely, over 60 percent of those who said that religion was “not at all important” to their sense of identity also identified as Democratic, and only 17 percent of them identified as Republican. The 2014 survey has nearly identical proportions as the 2013 survey. A majority of those for whom religion is “extremely important” was Republicans, and 28 percent of them were Democrats. The number of people who strongly identified with both their religion and the Democratic Party was half as large as the number who strongly identify with both the Republican Party and their religions. On the other side of the aisle, the number of people with no religious identification who also identified as Democrats was almost three times as large as the number with no religious identity who also identified as Republicans. Taken together, these findings suggest that the apparent divide

Table 5.1

Strength of Religious Identity, by Partisanship, 2013 Survey (in percent)

Strong Democrat Democrat Lean Democrat Independent Lean Republican Republican Strong Republican

Not at All Important 30 13 18 21 5 8 4

Not Very Important 17 17 16 20 10 12 7

Note: Totals might not add to 100 because of rounding.

Important 17 13 9 22 9 17 14

Extremely Important 15 9 5 19 14 15 25

70

Religious Identity in US Politics

Table 5.2

Strength of Religious Identity, by Partisanship, 2014 Survey (in percent)

Strong Democrat Democrat Lean Democrat Independent Lean Republican Republican Strong Republican

Not at All Important 23 15 17 26 10 6 4

Not Very Important 23 20 12 16 11 13 5

Note: Totals might not add to 100 because of rounding.

Important 16 12 12 23 10 12 14

Extremely Important 12 10 6 19 15 14 24

between religious and nonreligious Americans might be nothing more than partisanship in disguise.

Conflating Partisan Identities with Religious Social Identities

Since the 1960s researchers have been using a scale called the thermometer rating to measure public attitudes about candidates and groups in the United States. Participants are asked to rate a list of individuals and groups on a scale from 0 to 100 based on how warmly and favorably they feel toward the individual or the group. Ratings between 50 and 100 indicate that they feel warmly toward the individual or group, while ratings between 0 and 50 mean that they do not care too much for the individual or group. If they do not feel particularly warm or cold toward an individual or group, they are asked to select the 50 mark. The American National Election Survey has asked respondents to rate presidential candidates and political parties since the 1970s. For comparison purposes, the highest average rating for presidential candidates was in 1976, with a score of 63. The lowest average rating for presidential candidates was in 1972, with a score of 42. For the most part, Americans do not have strong positive views of presidential candidates or political parties. Over the previous four decades, the average thermometer ratings tend to be close to 50.25 I asked participants in the 2013 survey to use the thermometer scale to rate eighteen separate groups in the United States. The group with the highest average score was scientists (71.27). The group with the lowest average score was cigarette smokers (29.05) and the Tea Party (35.23). Most of the other groups had average scores closer to 50. Does religious identity influence attitudes toward members of groups in society? Because respondents with strong religious social identities were disproportionately Repub-

71

Christian Identity and Anti-Atheist Intolerance

licans, I expected those with strong religious identities to express favorable views toward members of groups affiliated with the Republican Party and to have less regard for groups affiliated with the Democratic Party. In the 2013 survey, the groups affiliated with the Democratic Party were Democrats, environmentalists, and gays and lesbians; the groups affiliated with the Republican Party were Republicans, Tea Party members, and gun owners. Figures 5.1 and 5.2 display the support for these groups by the religious identities of the respondents. Figure 5.1 displays the attitudes about groups affiliated with the Democrats by religious social identity. The largest difference in attitudes was toward gays and lesbians. Respondents with a weak religious social identity had much higher regard for gays and lesbians than they did for the Democratic Party. By contrast, those with a strong religious social identity reported negative attitudes toward gays and lesbians. Consistent with expectations, the overall pattern of findings suggests that groups affiliated with the Democratic Party, including Democrats, were viewed more positively by those with weaker religious identities and less positively by those with strong religious identities. What about groups affiliated with the Republican Party? Americans with strong religious social identities viewed groups affiliated with the Republican Party much more positively. Although they did not report warm feelings toward gun owners, the pattern of findings among these groups is consistent with expectations. The stronger people’s religious social identity, Figure 5.1 Attitudes Toward Democrat-Affiliated Groups, by Religious Social Identity, 2013 Survey

80 70 60 50 40

D Democrats emocrats

E Environmentalists nvironmentalists

N Not ot aatt aall ll Im Important portant

Not Not Ve Very ery Im Important portant

Important Im portant

Extremely Important E xtremely Im portant

Not Important N ot aatt aall ll Im portant

Not Not Ve Very ery Im Important portant

Important Im portant

E Extremely xtremely Im Important portant

Not Important N ot aatt aall ll Im portant

Not Not Ve Very ery Im Important portant

20

Im Important portant

30 E Extremely xtremely Im Important portant

Mean Thermometer Rating ting M ean T hermometer Ra

90

G Gays ays aand nd L Lesbians esbians

Note: Points represent the mean thermometer scores; lines are the 95 percent confidence interval of the estimate.

72

Religious Identity in US Politics

Gun Gun Owners Owners

Re Republicans publicans

N Not ot aatt aall ll Im Important portant

Not Not V Very ery Im Important portant

Im Important portant

E xtremely Im portant Extremely Important

N ot aatt aall ll Im portant Not Important

Not Not V Very ery Im Important portant

Im portant Important

E Extremely xtremely Im Important portant

Not Important N ot aatt aall ll Im portant

portant Not Not V Very ery Im Important

Im Important portant

65 60 55 50 45 40 35 30 25 20 15 E Extremely xtremely Im Important portant

Mean Thermometer Rating ting M ean T hermometer Ra

Figure 5.2 Attitudes Toward Republican-Affiliated Groups, by Religious Social Identity, 2013 Survey

T Tea ea Party Party y

Note: Points represent the mean thermometer scores; lines are the 95 percent confidence interval of the estimate.

the higher the favorability rating they gave to the group; the weaker their religious identity, the lower the favorability rating they gave to the group. Aside from gays and lesbians, there is no known religious doctrine about any of these groups. Religious texts certainly do not take a stand on gun ownership and do not have a primary emphasis on environmentalism. Yet the consistent, predictable pattern demonstrates that religious social identities influence attitudes toward these groups in the same way that partisanship does.26 Because these results might simply reflect a particular time and place and not entrenched views, the study was replicated a year later in a different sample of US adults. (See Figures 5.3 and 5.4.) Once again, respondents were asked to rate eighteen groups using the thermometer scale, measuring the strength of their positive or negative attitudes toward the group. Once again, scientists were the group with the highest average rating (70.97), whereas the Tea Party (34.97) and cigarette smokers (27.98) came in last. Both times participants rated their feelings toward atheists, but the 2014 survey replaced environmentalists with “people unaffiliated with any religion.” The religiously unaffiliated made up the largest religious group among respondents who identified with the Democratic Party in 2014, which makes them a reasonable replacement for environmentalists in this test of how religious social identity influences attitudes toward members of groups affiliated with certain political parties.27 The results displayed in Figure 5.3 are consistent with those from the 2013 survey. The stronger the religious social identity individuals reported,

Christian Identity and Anti-Atheist Intolerance

73

Figure 5.3 Attitudes Toward Democrat-Affiliated Groups, by Religious Social Identity, 2014 Survey

M Mean ean T Thermometer hermometer Ra Rating ting

80 70 60 50 40 30 20 Extremely Important Extremely Important Extremely Important Not Very at all all E xtremely Im portant Not xtremely Im portant N Not Very Very Not Not at Not Very Very Not Not at at all all E ot V ery Not Not at at all all E xtremely Im portant Not Important Important Important Important Important Important Important portant portant Important Important Important Im Important Important Important Im Important Im portant Im portant Democrats Democrats

Gays Gays aand nd Lesbians Lesbians

Religiously Re ligiously Unaffiliated Unaffiliated

Note: Points represent the mean thermometer scores; lines are the 95 percent confidence interval of the estimate.

the less favorable the attitudes they expressed toward gays and lesbians, the religiously unaffiliated, and Democrats. Gays and lesbians and the religiously unaffiliated were more strongly disliked than Democrats. Specifically, respondents with the strongest religious social identity reported favorable views of these two groups 50 percent lower than those with a weak religious social identity reported. By contrast, respondents with strong religious social identities gave Democrats only a 27 percent lower favorability score than did those with weak religious social identities. When considered in combination with the analysis of the 2013 survey, these results suggest that using a method that sorts religious and nonreligious respondents into different political parties might influence the reported attitudes and hostility of religious respondents toward those who affiliate with the Democratic Party. Meanwhile, these results also suggest that something more than partisanship is at play. Although the pattern of dislike is consistent, the positive feelings people with a weak religious identity report toward groups who align with them on religious grounds (gays and lesbians and religiously unaffiliated) is substantially higher than for those with whom they cooperate politically (Democrats and environmentalists). Respondents with strong religious identities were just as likely to have favorable attitudes toward gun owners and the Tea Party in 2014 as they were in 2013. Among both strong and weak religious identifiers, attitudes about these groups were virtually the same as in 2013. Respondents with weak religious social identities dislike these groups, whereas those with strong religious identities positively rate these groups. Once again, it is noteworthy that there is no inherent religious prohibition against any of

74

Religious Identity in US Politics

Figure 5.4 Attitudes Toward Republican-Affiliated Groups, by Religious Identity, 2014 Survey

M Mean ean T Thermometer hermometer Ra Rating ting

60

50

40

30

20

10 Extremely Not Extremely Important Not Extremely Important Not Very Not xtremely Im portant Not portant N ot V Not Very Veery N ot aatt aall Not Very Very N ot aatt all all E ery N ot aatt aall ll Extremely Important Important Not ll E xtremely Im Important Important Important Important Important Important Important Important Important portant Im portant portant Im portant Im Im portant Im Im portant Im portant Im portant Important Gun Owners G un O wners

Re publicans Republicans

Tea Party T ea P arty

Note: Points represent the mean thermometer scores; lines are the 95 percent confidence interval of the estimate.

these groups. The positive feelings religious respondents report are not likely attributable to theological convergence. Rather, cooperation between the religiously committed and those who affiliate with these groups within the Republican Party engenders a sense of trust and team loyalty that spills over into a positive affect toward these groups.28 These findings explain both why previous work on religious intolerance uses the language of social identity theory without including measures of religious social identities in the measurement models and why it appears religion is driving a wedge into US society. However, if religious antipathy toward the LGBT community stems not from deeply held religious values or beliefs but from the current makeup of partisan coalitions in the nation, it could be rapidly mended. Recall that social identities are not as stable as beliefs and values, and as people change them in response to societal conditions, antipathy expressed at one point in time can quickly dissolve as individuals adjust the importance of their relative social identities. Susie, a lifelong Lutheran, left her religious community when she felt her religious group was voicing unwarranted antigay views in church. Prejudice is context specific. Even if religious Americans exhibit prejudice toward members of some disadvantaged groups in society, it is unlikely that it represents a general, stable view. Rather, that which elicits fear-motivated prejudice against a group in one context might not elicit the same emotion in another context.29

Christian Identity and Anti-Atheist Intolerance

75

Djupe and Mockabee found no significant difference between religious denominations in their perceptions that atheists represent a threat. The sample size was small, causing large standard errors around the estimates (a common problem) and small t-statistics, which led these scholars to conclude that religious affiliation has no effect on perceptions of atheist threat.30 Another interpretation consistent with their data is that members of majority, minority, Christian, and non-Christian religious groups are all equally likely to feel threatened by atheists. Although not significant, the size of the coefficient for “other religions” in their model is larger than the coefficient for mainline Protestant, which means that respondents who belonged to non-Christian religious groups were just as likely (if not more likely) to feel threatened by atheists as were mainline Protestants. When I asked Kirk (raised Hindu, early twenties) what kind of person would upset his parents the most if he were to date or marry, he said his parents would not care too much about his future mate’s attributes if the partner were open to Hinduism and willing to embrace a Hindu way of life. However, when I pressed further, he stated that an atheist would be harder for his parents to accept than a Muslim or a Christian because it would be hard for them come to terms with someone who “does not even have the concept of a God.” Blake (atheist, twenties) told me that it was easier for his strict Catholic parents to accept that he is bisexual than it was for them to accept his atheism. Charles (Muslim, over fifty) told me that Islam, Christianity, and Judaism are very similar in many aspects, but atheists “are not religious people; they are too distant from us. Atheists are different; I do not believe that they are religious people—they are totally different.”31 These anecdotes are consistent with theoretical expectations and some quantitative evidence. People who identify with a religion feel distant from atheists and others who do not believe in God. They are the common outgroup of nearly every strong religious identifier with whom I spoke, which is why the social identity framework might explain antipathy toward nonbelievers better than existing approaches.32 Members of social identity groups often display sympathetic behavior toward those affiliated with their group and discriminatory behavior to those in opposing groups. Before proceeding with the analysis, it is important to note the differences between the concepts as they have been discussed thus far and the measures used to assess political tolerance, prejudice, and cooperation. Political tolerance is forbearance; it is the restraint of the urge to repress one’s political enemies.33 Prejudice is thinking ill of others without sufficient cause.34 Cooperation is the process of working together toward the same end. Identity-based out-group hostility could be manifest in each of these. The analysis that follows does not always employ the traditional measures for each of these, but the purpose is to demonstrate that strong

Atheists as an Out-Group to the Religious

76

Religious Identity in US Politics

religious social identities are associated with greater aversion toward members of the out-group (atheists), whether that is in the form of intolerance, prejudice, or unwillingness to cooperate. As just discussed, use of a survey sorting religious Americans into one of the two major political parties can make it difficult to isolate the influence of religious social identity from partisan social identification on attitudes toward members of out-groups. In addition, social desirability motivates most people to hide prejudice in surveys. To combat this, scholars have developed a few innovative designs to measure prejudice, which respondents would rather conceal from them. A dominant approach in psychology is to use implicit attitude measures to detect prejudice. This approach is based on decades of attitude research that shows two types of attitudes: implicit and explicit. An explicit attitude is the kind one deliberately thinks about and reports. For instance, you could tell someone whether you like chocolate because that is an explicit attitude. Implicit attitudes are positive and negative evaluations that occur outside of one’s own control or consciousness. You might say that you like chocolate, but you might also have negative associations with chocolate of which you are unaware. Implicit attitude tests ask people to sort pictures, words, and letters into bins. Software measures the speed at which the participants sort these items. People sort words, pictures, and phrases about which they have positive implicit attitudes more quickly than those about which they have negative implicit attitudes. Several implicit association tests are now used to detect negative attitudes toward members of disadvantaged groups.35 The list experiment is another approach. One group of scholars was concerned that Dutch citizens might conceal their prejudice toward Muslim minorities if directly asked to report their attitudes. People might say that they dislike Muslims’ denial of equality to women but not Muslims in general out of a commitment to that equality. However, this might be a socially acceptable means of concealing prejudice.36 One study attempted to measure hostility toward Muslims in the Netherlands. In the control group, participants read a list of five items that might make them angry, only two of which were overtly negative. In the control group, the option “Muslim” was not included on the list. Then, they were asked how many of the items on the list angered them. Participants never actually indicated which items upset them; they simply chose a number. Respondents in the treatment group were randomly assigned to view the same list, with one additional item—Muslims. If respondents were truly angered by Muslims, the number of groups they indicated should be one higher for the group who viewed the list including Muslims than it was for the group whose list did not include

Isolating Atheist Antipathy

Christian Identity and Anti-Atheist Intolerance

77

Muslims. This is precisely what scholars found. Following this experiment, simple statistical analyses revealed the social group identities of participants who had greater disdain for Muslims.37 Of course, the obvious limitation of the Dutch list experiment is that it is impossible to tell which groups were disliked by respondents. All the list experiment demonstrates is the number of groups disliked by the individuals; if people with demographically similar traits in the treatment group disliked an average of one more item than those in the baseline group, it was assumed that the disliked group was the extra item chosen. This is a strong assumption. I can think of many people demographically similar to myself who might dislike a different number of items in a list than I would. It seems unfair to conclude respondents were prejudiced who might not have been, simply because other demographically similar respondents disliked more items in a list. Rather than look for people with implicit bias against members of a different social identity group or rely on a list that would not tell me which specific individuals were biased against others, I developed a different kind of experiment. In this experimental method, it is imperative that everything in the experiment remain constant except the one item under investigation. Because my interest is in how religious identity influences attitudes about other members of US society, the constants in the experiment are a variety of activities with which people might be asked to assist their neighbor. In the first experiment, respondents were randomly assigned to one of four conditions. They were told that the acquaintance asking for help was either Hispanic, obese, an atheist, or openly gay. Participants were asked how likely they would be to help an acquaintance with “the following tasks.” Using a sliding scale ranging from 1 (“completely unlikely”) to 100 (“very likely”), participants were asked to indicate their willingness to help an acquaintance with six tasks: (1) purchasing auto insurance, (2) mailing a package, (3) challenging a speeding ticket in court, (4) obtaining a marriage license, (5) registering to vote, and (6) helping their acquaintance run for city council. This was rescaled from 0 to 1 to simplify the analyses. Obviously, some people are more likely to help their neighbors than are others. I hypothesize that willingness to assist depends on the nature of the respondent, the nature of the task, and the nature of the person asking for help. We all know people who are willing to help just about anyone with just about anything. People like this will be more likely to help with each of the tasks regardless of the nature of the task or the characteristics of the person asking for help. Others will help their neighbors when it is convenient or if the task is not too onerous. These respondents will vary in their willingness to help an acquaintance depending on the nature of the task regardless of who is asking. Finally, some people’s willingness to help

78

Religious Identity in US Politics

depends on who is asking. Whether the task is difficult or easy is less important than the identity of the person asking for assistance. The challenge is to devise a test that controls for these variables—to account fully for each of them would require a series of experiments that carefully manipulates each of the three motivations to help others. Instead, this approach varies the political nature of the tasks and the identity of the person asking for help. The study controls for differences in willingness to help by having a difficult task and an easier task in each of the experiments. That which might make people less likely to help someone mail a package because it seems too difficult should not influence their willingness to help purchase auto insurance. The two tasks require a distinct skill set. Differences in the nature of each task simply contributed to the statistical noise of the experiment. The existing literature is mixed on how religious identity influences tolerance toward minority groups. My central argument in this chapter is that political conflict is the source of perceived intolerance. Religious Americans are equally tolerant toward groups in the United States unless a group is involved in a political struggle against religion. The 2013 study did not have a control group (a group where the acquaintance had no description) and is discussed first. On average, Americans are significantly willing to help an openly gay acquaintance both run for office and register to vote. However, the nature of the experiment might have inflated average responses for those in the “openly gay” group for the political tasks for the following reason. Although the list of tasks included in the experiment had a randomized order, every respondent saw all the tasks on the same page. In 2013, most people who opposed samesex marriage claimed not to be prejudiced against gay people; they simply opposed a redefinition of traditional marriage. Because of this prejudice, if respondents registered reduced support for helping an openly gay acquaintance obtain a marriage license, they might have also reported higher than average support for each of the other acquaintances in the experiment to reduce psychological discomfort. This would make it appear that responders to the “openly gay” group were especially exuberant about helping their acquaintances with these tasks, when in fact they might not have been. The findings in Figure 5.5 demonstrate that although Latino/a Americans might face more political discrimination than atheist, gay, or obese Americans do, the differences are not statistically significant. The average effect of suggests a lack of general bias toward members of these four minority groups among the US population. To the extent that average Americans are unwilling to help their Latino/a acquaintances run for city council, they are also unwilling to help their obese acquaintances run for city council. However, if we look at how religious identity moderates the relationship between the acquaintance attributes and participants’ willingness to

Christian Identity and Anti-Atheist Intolerance Figure 5.5

79

Average Treatment Effects, 2013 Survey

0.9 0.85 0.8

P Probability robability

0.75 0.7 0.65 0.6 0.55 0.5 0.45 0.4 Atheist A theist

Latino L atino

Obese O bese

Register Re gister to to Vote Vote

Openly Gay Openly G ay

Atheist Atheist

Latino Latino

Obese O bese

Openly O penly Gay Gay

City Run for Ci ty Council Council

Note: Points are the predicted probability of helping such an acquaintance; lines are the 95 percent confidence interval of the estimate.

help with political activities, the data tell a vastly different story. Each of the points in Figure 5.6 represents the predicted probability of helping an acquaintance with that identity. The X-axis is the strength of religious identity measured by how important respondents said that religion was to their sense of self. If the acquaintance was described as obese or Latino, there was not a significant difference in willingness to assist by the strength of religious identity for the political activities of registering to vote and running for office. Respondents with weak religious identities were equally likely to help their obese and Latino/a acquaintances as were those with strong religious identities. However, there was a significant difference among the religious in their desire to help atheist neighbors register to vote (t = –3.17). In addition, those with strong religious social identities were much less likely to help their openly gay (t = –3.92) or atheist (t = –7.09) acquaintances run for city council. An increase of one standard deviation from the mean level of religious social identity decreases the probability that participants would help their atheist acquaintance run for city council by 19.53 percent, compared with an 8.18 percent decline in the probability of helping an openly gay acquaintance run for city council. Though the effect is smaller, it is also true that the stronger their religious social identities, the less willing people are to help an atheist acquaintance register to vote. In fact, the only group from the 2013 study religious Americans were less willing to help register to vote was atheists.

80 Figure 5.6

Religious Identity in US Politics Religious Bias Toward Minority Groups, 2013 Survey Run for City Council

0.85 0.8

0.85

0.75

0.8

0.7

0.75 0.7

0.65

0.65

0.6

0.6

0.55

R

R

Register to Vote

0.9

0.5

0.55

0.55

0.5 0.5

0.45

0.45 0.45 0.4

0.4 1

A Atheist theist

1.8 2.5 3.2 Strength S trength of Religious Religious Identity Identity Latino Latino

Obese Obese

Openly Openly Gay Gay

4

1

Atheist Atheist

1.8 2.5 3 ..2 2 Strength S trength of Religious Religious Identity Identity Latino Latino

Obese Obese

4

Openly Openly Gay Gay

Note: Lines are the predicted probability of helping such an acquaintance.

This single experiment has some limitations. First, because the experiment lacked a control group, it is impossible to know how the findings compare with respondents’ general willingness to assist. Second, social desirability bias likely influenced reported desire to assist openly gay friends with all activities other than obtaining a marriage license. Respondents likely felt some remorse about being less willing to help gay friends get married, which motivated them to slightly overstate their enthusiasm to help gay friends with other tasks. Finally, the current alignment of religious and nonreligious Americans with one of two different political parties might account for the differences. This is evident in the differences between the “register to vote” and “run for public office” results. Although people with a strong religious identity were less likely to help an atheist acquaintance register to vote, the predicted probability of helping an atheist register to vote was above 0.5. Religious Americans were more willing to help atheists register to vote than they were to help them run for city council, suggesting that the lack of desire among the strongly religious to assist their atheist neighbors might be based on political partisan competition and nothing else. The 2014 experiment overcomes these obstacles. The second-wave survey is a cross section without any participants from the 2013 wave. Once again, participants were asked to rate their willingness to assist an acquaintance with six tasks. The tasks and the language in the 2014 survey is identical to that of the 2013 survey, with one exception. The second survey had three groups (Republican, Democrat, atheist) and a control group. That is, one-quarter of the respondents were simply instructed, “Imagine an acquain-

Christian Identity and Anti-Atheist Intolerance

81

tance asked you for help with each of the following tasks,” whereas the others were randomly assigned to a treatment group that included either a partisan or a nonreligious description of the acquaintance. This overcomes all the weaknesses of the previous experiment because it allows comparison of the probability of helping an atheist acquaintance to both a control group and to the partisan groups. There is not a significant difference, on average, between participants’ reported desire to assist a Democrat, Republican, or atheist register to vote, compared with the responses of the control group.38 Likewise, participants in the 2014 survey were equally likely to help Democrat and Republican friends run for city council as they were to help “an acquaintance” (see Figure 5.7). However, respondents in the group asked about an “atheist acquaintance” reported significantly greater enthusiasm for helping their friend run for office than those in the groups. The average respondent was more willing to offer political assistance to their atheist neighbors; perhaps atheists do not face political discrimination. Looking at how the strength of religious identity moderates the relationship, I do not find a statistical difference in willingness to help an acquaintance register to vote by strength of religious identity. Those with a strong religious identity were just as willing to help their atheist, Democrat, and Republican friends register to vote as were those with a weak religious identity. Interestingly—and consistent with Putnam and Campbell’s (2010) findings—participants with a strong religious identity were also equally willing to help their atheist, Republican, and Democrat friends run for city Figure 5.7 The Influence of Religious Social Identity on Willingness to Help with Political Activities, 2014 Survey Run for City Council

0.75

Register to Vote

0.9 0.85

0.7

0.8 0.65

0.75 0.7

0.6

0.65

R

0.55 R

0.6

0.5

0.55 0.5

0.45

0.45 0.4

0.4 1

Atheist Atheist

2.5 1 .8 2.5 33.2 .2 Strength Religious Identity S trength of Re ligious Ide ntity Cont Control rol

D Democrat emocrat

4

Re Republican publican

Note: Lines are the predicted probability of helping.

1

Atheist Atheist

1. 8 2.5 1.8 2.5 3.2 Strength S trength of Religious Religious Identity Identity Control Control

Democrat Democrat

4

Re Republican publican

82

Religious Identity in US Politics

council. However, those with weak religious identities were much more interested in helping their atheist acquaintances run for city council than they were in helping other friends. A one standard deviation decrease in religious identity predicted a 10 percent increase in willingness to help an atheist run for city council. Overall, these findings provide clear evidence that some religious intolerance toward atheists is rooted in political competition. Religious Americans have significantly stronger negative attitudes toward members of some minority groups in society. It is also clear that nonreligious Americans have strong negative attitudes toward members of other groups in society. The most parsimonious explanation of these attitudes is in the context of political competition and religious social identity. As the religious and nonreligious in the United States have sorted themselves into opposing political parties, the attitudes about members of groups who affiliate with these two political parties have polarized. Party competition causes members of the two major parties to be less trusting of opposing party members and to feel antipathy toward opposing group members. The strongest evidence supporting this conclusion is the way participants with strong religious identities behaved toward Latino/as. In the 2014 experiment, the probability of helping “an acquaintance” run for city council is the same for those with weak and strong religious identities (about 0.5). The same is true concerning the probability of helping Latino/as run for city council in the 2013 survey experiment. Yet in the 2013 survey the overall probability of supporting Latino/as running for office is higher (0.56) than the probability in the 2014 survey of helping “an acquaintance” run for office (0.5). Not only are Americans with weak and strong religious identities equally likely to help Latino/as register to vote but they are also as likely to help Latino/as run for office as they are to help most others. Because Latino/as were not clearly aligned with either of the major political parties in 2013–2014, they did not trigger partisan competition motivations; they were treated like anyone else asking for help.

Conclusion

1. Thomas F. Pettigrew, “Intergroup Contact Theory,” Annual Review of Psychology 49, no. 1 (1998): 65–85; Thomas F. Pettigrew and Linda R. Tropp, “A Meta-Analytic Test of Intergroup Contact Theory,” Journal of Personality and Social Psychology 90, no. 5 (2006): 751. 2. Corwin E. Smidt et al., “Religious Involvement, Social Capital, and Political Engagement,” in Religion as Social Capital: Producing the Common Good (Waco, TX: Baylor University Press, 2003), 153; Corwin E. Smidt, Pews, Prayers, and Participation: Religion and Civic Responsibility in America (Washington, DC:

Notes

Christian Identity and Anti-Atheist Intolerance

83

Georgetown University Press, 2008); David E. Campbell and Steven J. Yonish, “Religion and Volunteering in America,” in Religion as Social Capital: Producing the Common Good (Waco, TX: Baylor University Press, 2003), 87–106; Robert D. Putnam and David E. Campbell, American Grace: How Religion Divides and Unites Us (New York: Simon and Schuster, 2010); Azim F. Shariff et al., “Religious Priming: A Meta-Analysis with a Focus on Prosociality,” Personality and Social Psychology Review 20, no. 1 (2016): 27–48. 3. Samuel Andrew Stouffer, Communism, Conformity, and Civil Liberties: A Cross-Section of the Nation Speaks Its Mind (New Brunswick, NJ: Transaction, 1955). 4. For an excellent review of this literature, see Paul A. Djupe, Religion and Political Tolerance in America: Advances in the State of the Art, vol. 23 (Philadelphia, PA: Temple University Press, 2015). 5. Marie A. Eisenstein and April K. Clark, “Heterogeneous Religion Measures and Political Tolerance Outcomes,” in Religion and Political Tolerance in America: Advances in the State of the Art, ed. Paul A. Djupe (Philadelphia, PA: Temple University Press, 2015). 6. John L. Sullivan et al., “The Sources of Political Tolerance: A Multivariate Analysis,” American Political Science Review 75, no. 1 (1981): 92–106. 7. Eisenstein and Clark, “Heterogeneous Religion Measures and Political Tolerance Outcomes.” 8. James L. Gibson, “The Political Consequences of Religiosity: Does Religion Always Cause Political Intolerance?” in Religion and Democracy in the United States: Danger or Opportunity, ed. Ira Katznelson and Alan Wolfe (New York: Russell Sage, 2010), 147–175. 9. Daniel Cox, Robert P. Jones, and Juhem Navarro-Rivera, “Nonreligious Tolerance,” in Religion and Political Tolerance in America: Advances in the State of the Art, ed. Paul A. Djupe (Philadelphia, PA: Temple University Press, 2015), 131–150. 10. Pazit Ben-Nun Bloom and Marie Courtemanche, “Religion, Morality, and Tolerance,” in Religion and Political Tolerance in America: Advances in the State of the Art, ed. Paul A. Djupe (Philadelphia, PA: Temple University Press, 2015). 11. Paul A. Djupe and Stephen T. Mockabee, “Religious Worldviews and Political Tolerance,” in Religion and Political Tolerance in America: Advances in the State of the Art, ed. Paul A. Djupe (Philadelphia, PA: Temple University Press, 2015), 117–132. 12. Joby Schaffer, Anand E. Sokhey, and Paul A. Djupe, “The Religious Economy of Political Tolerance,” in Religion and Political Tolerance in America: Advances in the State of the Art, ed. Paul A. Djupe (Philadelphia, PA: Temple University Press, 2015), 151–164. 13. Deborah L. Hall, David C. Matz, and Wendy Wood, “Why Don’t We Practice What We Preach? A Meta-Analytic Review of Religious Racism,” Personality and Social Psychology Review 14, no. 1 (2010): 126–139, report that religious social identification is the primary source of religious racism. 14. James L. Gibson and Amanda Gouws, “Social Identities and Political Intolerance: Linkages Within the South African Mass Public,” American Journal of Political Science 44, no. 2 (2000): 278–292; Gibson and Gouws, Overcoming Intolerance in South Africa: Experiments in Democratic Persuasion, Cambridge Studies in Political Psychology and Public Opinion (Cambridge, UK: Cambridge University Press, 2005). 15. Gibson, “The Political Consequences of Religiosity,” 172. 16. James L. Gibson, “Enigmas of Intolerance: Fifty Years After Stouffer’s Communism, Conformity, and Civil Liberties,” Perspectives on Politics 4, no. 1 (2006): 21–34.

84

Religious Identity in US Politics

17. Gibson and Gouws, Overcoming Intolerance in South Africa, 76. 18. The group “other” was the seventh least popular out of fifteen, and “don’t know” was ranked ninth. 19. Patrick R. Miller and Pamela Johnston Conover, “Red and Blue States of Mind: Partisan Hostility and Voting in the United States,” Political Research Quarterly 68, no. 2 (2015): 225–239. 20. Wouter Van den Bos et al., “Pyrrhic Victories: The Need for Social Status Drives Costly Competitive Behavior,” Frontiers in Neuroscience (October 23, 2013): 31, https://doi.org/10.3389/fnins.2013.00189. 21. Mina Cikara, Matthew M. Botvinick, and Susan T. Fiske, “Us Versus Them Social Identity Shapes Neural Responses to Intergroup Competition and Harm,” Psychological Science 22, no. 3 (2011): 306–313. 22. Miller and Conover, “Red and Blue States of Mind.” 23. Lilliana Mason, “‘I Disrespectfully Agree’: The Differential Effects of Partisan Sorting on Social and Issue Polarization,” American Journal of Political Science 59, no. 1 (2015): 128–145. 24. Putnam and Campbell, American Grace, 371. 25. The American National Election Studies Time Series Cumulative Data File, Stanford University and the University of Michigan, 2010, www.electionstudies.org. These materials are based on work supported by the National Science Foundation under grant numbers SBR-9707741, SBR-9317631, SES-9209410, SES-9009379, SES-8808361, SES-8341310, SES-8207580, and SOC77-08885. Any opinions, findings, conclusions, or recommendations expressed in these materials are those of the author(s) and do not necessarily reflect the views of the funding organizations. 26. See Appendix 5 for the figures replicating these findings with partisanship instead of religiosity. 27. Alan Cooperman et al., 2014 U.S. Religious Landscape Study (Washington, DC: Pew Research Center, 2015). 28. Miller and Conover, “Red and Blue States of Mind.” 29. Gibson and Gouws, “Social Identities and Political Intolerance”; Gibson and Gouws, Overcoming Intolerance in South Africa. 30. Djupe and Mockabee, “Religious Worldviews and Political Tolerance,” 125. 31. Information on these interviews is in Appendix 1. 32. With the exception of Zen Buddhists. Cameron (raised Buddhist, more than forty years old) told me that his religion is the central feature of his life. When I asked him how he felt about people in other religious and political groups, he explained that the path of Zen Buddhism encourages people to break free from identities. It is so important for him to avoid conflict with all people that he accepts people for who they are and does not attach any special meaning to their religious identities. For him, religion (or lack thereof) is one aspect of the whole individual. 33. Gibson, “Enigmas of Intolerance,” 22. 34. Gordon W. Allport, The Nature of Prejudice (New York: Basic Books, 1979), 6. 35. Tony Greenwald, Mahzarin R. Banaji, and Brian A. Nosek, “Project Implicit,” https://implicit.harvard.edu/implicit/index.jsp. 36. Paul M. Sniderman and Aloysius Hagendoorn, When Ways of Life Collide: Multiculturalism and Its Discontents in the Netherlands (Princeton, NJ: Princeton University Press, 2007). 37. Sniderman and Hagendoorn, When Ways of Life Collide. 38. Full results are in Appendix 5.

6 The Religious Divide

socially and politically categorized by their religious affiliation. Mahatma Gandhi once asserted that politics divorced from religion “has absolutely no meaning” in India.1 Conflict in the Middle East is often waged along religious boundaries.2 International organizations and policies promoting religious freedom accentuate the differences between various forms of religious expression.3 In the United States, 77 percent of those surveyed in the Pew Research Center’s 2014 U.S. Religious Landscape Study said that they were affiliated with a specific religious tradition, suggesting that most Americans perceive their own categorization among religious denominations in the United States.4 In addition, the separation of church and state in the United States provides legal protections that benefit some religious institutions more than others. As an example, the Internal Revenue Service (IRS) automatically grants tax exemption to churches but not to religiously based nonprofit groups; these must file for an exemption like other nonprofits.5 This too creates societal religious categorization. People who affiliate with a religious group are expected to behave in certain ways. Most religions have doctrines, dogmas, and practices to which affiliated members should adhere. There can also be a set of beliefs unique to each religion. The primary distinction between most religions is the beliefs and behaviors expected of those who affiliate with the religion. In addition, religions bind people into a moral matrix that glorifies the ingroup and demonizes other groups that do not share this moral view. Members of religious groups in the United States often have political values and norms shared among their group members. Most evangelical Protestants and Mormons affiliate with the Republican Party, whereas most Jews,

Globally, substantial evidence suggests that people are

85

86

Religious Identity in US Politics

African American Protestants, Buddhists, Hindus, and Muslims affiliate with the Democratic Party.6 This engenders political behavioral expectations among those who affiliate with certain religions.7 People attribute intangible traits and qualities to those who profess belief in God. Evolutionary theory of cooperation explains why people committed to a religion develop a strong social identification with that group. Imagine that you lived in a time when the survival of your community depended on people being willing to make considerable individual sacrifices to benefit society as a whole. Each individual risk or sacrifice benefited all group members, without which the group could not survive. These conditions were conducive to self-interested motivations to free ride and receive the benefits of group membership without self-sacrifice. It is easier to monitor this behavior when groups are smaller because freeloaders can be banished from the group. However, as the group grows, social interactions become increasingly anonymous, and it becomes more difficult to monitor people’s behavior. When freeloading is associated with a substantial punitive cost, people are less likely to mooch. Increasing group size makes it more difficult to enforce the unenforceable as people anonymously freeload. Religion evolved to facilitate cooperation in large groups because it deterred freeloading by informing group members that an omniscient overseer was monitoring and punishing poor behavior within the group.8 When people believe that a punitive and moralistic god knows their thoughts and behaviors, they are more likely to behave according to the norms and traditions of the social group. In fact, even when members of the group are not in close physical proximity to each other, a shared belief in a common deity motivates people to behave impartially toward each other.9 An experiment tested the influence of belief in a punitive deity and shared religious identification on people’s willingness to obey the unenforceable and cooperate with members of their own religion.10 The researchers recruited a sample of participants from eight diverse communities around the globe, including hunter-gatherers, horticulturalists, herders, and farmers as well as fully market-integrated populations engaged in wage labor or operating small businesses. The participants adhered to a variety of world religious traditions including Christianity, Hinduism, and Buddhism and reported beliefs in an immense range of local supernatural agents, including spiritmasters, saints, ancestors, animistic beings, anthropomorphic celestial deities, garden spirits, and ghosts. They gave participants thirty coins and asked them to play a game in which the coins were allocated into two cups. If the participants strictly followed the instructions, each cup would have fifteen coins at the end of the game. The researchers told the participants that the money allocated into the cups would be given to the person represented by the cup.11 In one game,

The Religious Divide

87

participants chose between a cup assigned to themselves and a cup assigned to an anonymous person in a distant community who shared their religion. In the other game, participants chose between a cup assigned to an anonymous member of their religion in their own community and an anonymous person from a distant community who shared their religion. All respondents also filled out a survey that measured their level of religious knowledge and the strength of their belief in a moralistic or punitive deity. They found that people who believed in a punitive god were significantly more likely to obey the rules of the game. They were less likely to cheat and put more money in their own cup, and they were more likely to allocate coins evenly when choosing between someone in their own community versus a person in a distant community. Conversely, those with less conviction in the existence of a punitive deity were significantly more likely to allocate more money to themselves or an anonymous member of their religion from their local community. This experiment demonstrates the power of religious belief to motivate people to obey unenforceable social customs and to treat people of other cultures and societies fairly. People are more likely to trust those who share religious beliefs because identification with a religion confers intangible qualities on the adherent, which are absent in nonbelievers.12 People are more likely to obey the rules of a game, even if it requires sacrifice, and allocate resources to people in distant lands and cultures when they believe in a higher power. This might be why the religious distrust atheists—because people who do not profess a belief in a higher power are assumed to operate outside of this evolved system of cooperation.13 Religion creates a tiered system of social trust. Those who share a religious identity are the closest social in-group. Individuals who belong to a different religion but who profess a belief in a higher power might be viewed as members of an in-group, though not as tightly knit. People who do not profess a belief in a deity are viewed as members of the out-group. Scholars have recently noted the rise in negative attitudes that partisans are willing to express about those who identify with the opposing political party. Using the same thermometer ratings discussed previously, survey researchers have noticed a surprising new trend. Democrats and Republicans provide positive thermometer ratings (ca. 70 points) to members of their own party and lower ratings to members of the opposing political party. The proportion of respondents who give the opposing party a rating below 50 has increased since the 1980s. In 1980, the proportion of partisans rating members of the other party below 50 was 40 percent. Also in 1980, 60 percent of partisans gave a neutral or a positive rating to cross-partisans. In the 1990s, 53 percent of partisans rated opposing partisans negatively; in

Partisan Affective Polarization

88

Religious Identity in US Politics

2004, 56 percent of partisans did so; and by 2008, the proportion increased to 63 percent. In this span of thirty years, American partisans shifted from a majority expressing positive attitudes about those in the other political party to a majority expressing negative attitudes about opposing partisans. 14 This antipathy extends beyond affect toward members of the opposing political party. Adopting the language used to identify racism in the past, survey researchers have begun asking respondents how they would feel if one of their children married someone from the other political party. In 2010, 49 percent of Republicans and 33 percent of Democrats stated that they would be unhappy at the prospect of interparty marriage. 15 By 2016, the partisan social divide had widened. Although fewer partisans stated that they would be upset if a family member were to marry someone outside of the party (Republicans 17 percent and Democrats 15 percent), 91 percent of Republicans and 86 percent of Democrats stated that they have an unfavorable view of opposing party members. In addition, half of all partisan respondents said that the opposing party makes them feel afraid. Partisans also made negative attributions about members of the opposing party. Nearly half of the Republican respondents said Democrats are more closedminded, immoral, lazy, and dishonest than other Americans. More than onethird of Democrats stated that Republicans are more unintelligent, immoral, and dishonest than other Americans, and 70 percent of Democrats thought Republicans were closed-minded.16 Furthermore, strongly identified partisans are more likely to wish harm upon members of the other political party. Of the respondents who identified strongly with their political party, 20 percent of them agreed that the opposing party was “downright evil,” and many of them wished someone would injure leaders of the opposing political party. Partisan antipathy was not limited to members of the partisan out-group. People with a strong partisan identity held negative attitudes about members of their own party. When asked, “When a member of your own party votes against the party on a key issue, have you ever wished that they would get sick and die?” About 5 percent of partisans agreed that they would wish death upon a member of their party who voted against it. Why the rise in partisan antipathy? Over the previous several decades, individuals have ideologically sorted themselves into one of the major political parties in the United States. Simply put, there were fewer conservative Democrats and liberal Republicans in 2016 than there were in 1968. When social groups are not aligned with each other, individuals are less biased and more positively oriented toward out-groups because the crosscutting identities undermine the cognitive and motivational bases of in-group bias.17 Conversely, when identities become aligned it has a synergistic influence that strengthens out-group bias and causes people to feel more negatively toward members of out-groups. To the extent that people possess

The Religious Divide

89

strong ideological and partisan identities, the sorting of people into ideologically homogeneous parties has fused both identities into a single, stronger identity, which motivates greater bias against members of the opposing political party.18 The strength of people’s partisan political identity influences more than political attitudes. It influences the way people think and feel about those who affiliate with the other party. In this case, strong out-group antipathy motivates people to express highly prejudicial attitudes about members of the opposing political party. There is no reason the strength of people’s religious social identity would not motivate similarly strong out-group antipathy toward those who affiliate with opposing religious persuasion. Just as partisan identities motivate people to express strong prejudicial attitudes toward those in opposing political parties, those with a strong religious identity might be prejudiced against people who are not religious. To the extent that this prejudice is caused by social identities, there is hope that the divide can be bridged. Greater contact attenuates the negative influence of social identities on prejudice.19 Just as people with strong political identities with close friends who are members of the opposing political party are less prejudiced against opposing partisans, people with strong religious identities who develop friendships with nonbelievers might become less prejudiced toward them.20 In the previous chapter I demonstrate that religious identity influences people’s willingness to help an acquaintance participate in certain political activities, depending on how closely aligned the group with which they affiliate is aligned with political parties. Consistent with the preceding discussion, we might expect religious identity to influence how willing people are to accommodate the request of a neighbor. Social identities carry invisible, intangible connotations people attribute to members of that social group. Moreover, religious identity can be stronger than national identity, and many people attribute both positive and negative features to others based on their religious affiliations.21 This means people exhibit bias toward members of religious and nonreligious communities. One example of this is the call by 2016 presidential candidate Donald Trump for a ban on Muslims entering the United States. Verbally, Trump said he did not know whether these people posed an actual threat, but the nonverbal message was clear: Muslims could not be trusted to combat Islamic terrorism. Given these sets of findings, I expect religious individuals to express antipathy toward those who do not share their religious identity. Indeed, many argue that the distrust between religious and nonreligious Americans prevents the religious from interacting with nonbelievers. Many fear that

Religious Social Identity

90

Religious Identity in US Politics

because people who attend church regularly consistently align with the Republican Party, out-group partisan antipathy is spilling over into nonpolitical interactions between believers and nonbelievers in the United States. However, consistent with the broader argument of this book, I hypothesize that the gap between religious and nonreligious Americans is overstated. Because those who belong to religions that emphasize church attendance as an essential component of worship (evangelical Christians and Mormons) tend to identify strongly with the Republican Party, using church attendance as a proxy for religiosity overstates the size of the gap between religious and nonreligious Americans. Plenty of people who identify strongly with religions also identify as Democrats and Independents, but they often affiliate with religions (Catholicism, Judaism, Buddhism, and Hinduism) that do not place as strong an emphasis on church attendance as do evangelical Christians and Mormons. Focusing on the extent to which people’s religion is important to how they view themselves and their nonpolitical interactions yields a different set of findings. In this context, the God gap is not as alarming as some have supposed. These analyses rely on the same 2013 and 2014 survey experiments I discussed in the previous chapter. Again, a national cross section of US adults was invited to participate in a study of political attitudes in 2013, and a different cross section was recruited for a follow-up study in 2014. We collected information about the strength of religious identity. Those higher on the scale have a stronger religious social identity. We also collected the standard political sociodemographic variables. Each of the participants was randomly assigned to one of four experimental conditions, imagining that an acquaintance approached them and asked for help “with each of the following activities.” Respondents then indicated the probability of helping their acquaintance using a sliding scale that ranged from 0 (“completely unlikely”) to 100 (“very likely”). Each of these responses was scaled to range from 0 to 1 for ease of interpretation. In the 2013 survey, the experimental conditions included four different descriptions of the acquaintance (openly gay, Hispanic, obese, and atheist). In the 2014 survey, there was a control group and three experimental groups (Republican, Democrat, and atheist). The remaining tasks in the survey included two semipolitical tasks (obtaining a marriage license and challenging a speeding ticket in court) and two nonpolitical tasks (mailing a package and purchasing auto insurance). Partisan competition might explain why political opponents would not help each other with strictly political tasks; however, it does not account for unwillingness to assist with nonpolitical tasks. There might have been some

The Evidence

91

The Religious Divide

spillover effect in which a respondent could not distinguish a semipolitical task from a political one, which could cause some residual bias against political opponents who wanted help in court. It is less conceivable that political competition would manifest as unwillingness to help an acquaintance mail a package.

Weak Religious Identifiers and Atheist In-Group Favoritism

I begin with findings from the 2013 survey regarding the semipolitical tasks of obtaining a marriage license and fighting a traffic violation. Because the full statistical results are available in Appendix 6, I will focus on the substantive influence of religious social identity on willingness to help. Because interactions are not normally distributed, a Wald test might be biased in favor of rejecting the null hypothesis. Following Preacher et al., Figure 6.1 plots the simple slopes of the interaction to test for significance. 22 As shown in Figure 6.1, the model predicts a statistically significant negative slope for helping an atheist (t = –4.73) and openly gay (t = –2.49) acquaintance fight a traffic violation. The stronger the respondents’ religious identity, the less likely they were to respond positively to the request of an atheist acquaintance to help fight a speeding ticket, although they were more likely to help an atheist than they were to help a Latino/a, that difference is not statistically significant. Alternatively, it is also clear that the nonreligious

Figure 6.1 How Group Identification Influences Probability of Assistance with Semipolitical Activities, by Religious Identity, 2013 Survey Obtain a Marriage License

0.9

0.8 0.75

0.7

0.7

0.65

0.65

0.6

0.6

0.55

0.55 F

O

0.8 0.75

0.5 0.45

Fight a Speeding Ticket

0.9 0.85

0.85

0.5 0.45

0.4 1

Atheist A theist

2.5 3.2 1 .8 2.5 Strength Religious Identity S trength of Re ligious Ide ntity L Latino atino

O Obese bese

4

Openly Openly Gay Ga y

0.4 1

A theist Atheist

1.8 2.5 33.2 .2 Strength S trength of Religious Religious Identity Identity Latino Latino

Source: Pew Research Center, 2013 U.S. Religious Landscape Study. Note: Lines are the predicted willingness to help for each experimental group.

Obese O bese

4

Openly Gay O penly G ay

92

Religious Identity in US Politics

were more exuberant about helping an atheist acquaintance fight a speeding ticket than they were to help an obese acquaintance. It is noteworthy that although nonreligious Americans were politically aligned with Latinos, they were 30.7 percent less likely to help a Latino/a fight a speeding ticket than they were to help an atheist. This demonstrates that nonreligious Americans have strong in-group favoritism toward atheists. It is not very surprising that religious Americans were less willing to help their openly gay acquaintances obtain a marriage license; many states had outlawed gay marriage in 2013. No such laws forbade atheist marriage, yet respondents with a strong religious identity expressed significantly less willingness to help their atheist acquaintances wed than did those with a weak religious identity. The “obtain a marriage license” model predicts that increases in religious social identity are associated with significantly less willingness to help an openly gay (t = –8.29) or an atheist (t = –5.17) acquaintance get a marriage license. The findings presented here suggest an opposite effect from the “fight a speeding ticket” model. Respondents with a strong religious identity expressed strong out-group bias against atheists and openly gay acquaintances, but those with a weak religious identification do not have strong in-group favoritism toward atheist or openly gay friends. Statistically, those with a weak religious identification were equally likely to help their obese, Latino, atheist, and openly gay friends get a marriage license, but those with strong religious identities were 12.8 percent less likely to help an atheist friend get a marriage license than they were to help a Latino acquaintance. The 2014 data tell a similar story. Using an identical estimation strategy, the model predicts a statistically significant negative slope for helping an atheist (t = –1.78) acquaintance fight a traffic violation. The stronger people’s religious identity, the less likely they were to respond positively to the request of an atheist acquaintance to help fight a speeding ticket, although they were more likely to help an atheist than they were to help a Republican, Democrat, or nondescript acquaintance (see Figure 6.2). Alternatively, it is also clear that those with weak religious identification were twice as exuberant about helping an atheist acquaintance fight a speeding ticket as they were to help the nondescript acquaintance. By contrast, respondents with a strong religious identity were just as willing to help the average acquaintance obtain a marriage license as they were to help an atheist, Democrat, or Republican acquaintance. This is consistent with previous work. People with strong religious identities seem to be equally willing to help members of all groups obtain a marriage license. The only significant slopes in the marriage license condition were those for atheist (t = –1.87) and the control group (t = 1.9) acquaintances. Once again, those with weak religious identification expressed in-group favoritism toward atheists in their strong willingness to help them fight a speeding ticket or

The Religious Divide

93

Figure 6.2 How Group Identification Influences Probability of Assistance with Semipolitical Activities, by Religious Identity, 2014 Survey Obtain a Marriage License

0.9 0.85

0.8

0.8

0.75

0.75

0.7

0.7

0.65

0.65

O

Fight a Speeding Ticket

0.9 0.85

0.6 0.55 F

0.6 0.55 0.5

0.5

0.45

0.45 0.4

0.4 1

A Atheist theist

11.8 .8 2. 5 3.2 2.5 3.2 Strength Identity S trength of Religious Religious Identity Control Control

Democrat Democrat

4

Republican Republican

1

Atheist A theist

1.8 2.5 3.2 Strength S trength of Religious Religious Identity Identity Control Control

Democrat D emocrat

4

Republican Republican

Source: Pew Research Center, 2014 U.S. Religious Landscape Study. Note: Lines are the predicted willingness to help for each experimental group.

obtain a marriage license, compared with the nondescript acquaintance. The findings presented here suggest that Americans with a strong religious identity do not express strong out-group bias against atheists, but nonreligious Americans do have strong in-group favoritism toward atheists. In addition, this bias is clearly distinct from partisan bias. Both religious and nonreligious respondents did not differ in their willingness to help Republicans and Democrats with semipolitical tasks. Moreover, religious respondents were more likely to help an atheist with these tasks than they were to help a Republican or a Democrat. There does not appear to be considerable antiatheist bias from highly religious individuals in the United States. The semipolitical nature of these tasks makes me cautious about reaching too strong a conclusion. The next section examines religion-based social bias in nonpolitical activities. If acquaintances came to you and asked you to help purchase auto insurance or mail a package, would their social identities change your response to them? The 2013 survey tested this. Once again, the base models included only an interaction between religious identities and the experiment in which the respondent participated. An additional model includes the standard demographic variables as controls (see Figure 6.3). In the models predicting willingness to help an acquaintance purchase auto insurance, the base model explains 4 percent of the variation, and the full controls model explains 6.1 percent. Clearly, social identities do not explain all the variation in respondents’ eagerness to help a neighbor purchase insurance, but they do explain

94

Religious Identity in US Politics

Figure 6.3 How Group Identification Influences Probability of Assistance with Nonpolitical Activities, by Religious Identity, 2013 Survey Purchase Auto Insurance

0.9

Mail a Package 0.9

0.8

0.8

0.7

0.7

0.6

P

M

0.6

0.5

0.5

0.4

0.4 1

A Atheist theist

1 .8 2. 5 3. 3.2 Strength Religious Identity S trength of Re ligious Ide ntity Latino Latino

Obese Obese

4

Openly Openly Gay Ga y

1

A Atheist theist

1.8 3.2 2.5 3.2 Strength S trength of Religious Religious Identity Identity L Latino atino

O Obese bese

4

Openly Openly Gay Gay

Source: Pew Research Center, 2013 U.S. Religious Landscape Study. Note: Lines are the predicted willingness to help for each experimental group.

some of it. However, in the case of purchasing auto insurance, the interaction between the experimental group and the religious identities of the respondents is not significant, suggesting that those with weak and strong religious identification were just as likely to help their atheist neighbors purchase auto insurance as they were to help members of any other group. Yet, religious social identities did influence how willing people are to help their neighbors mail a package. The more religious the respondents, the less willing they were to help their atheist acquaintances mail a package (t = –1.91). Though significant, the difference between individuals with weak or strong religious identities in willingness to help was not substantively large. However, the pattern of findings is consistent enough to suggest that religious identity might influence how willing people are to help their neighbors mail a package; although again, this seems to be because of enthusiasm on the part of respondents with a weak religious identity to help their atheist friends mail packages. Once again, this study did not include a control group, so we do not know how likely religious Americans were to help atheists compared with other Americans. The data from the 2014 survey are consistent with the findings from the 2013 survey (see Figure 6.4). On the task of purchasing automobile insurance, the base model accounts for 3.7 percent of the variation, and the full control model accounts for 5.3 percent. The Wald test suggests that those with a strong religious identity were significantly less willing to help

The Religious Divide

95

their Republican and atheist acquaintances purchase automobile insurance. Upon closer examination, it appears that the significant statistical effects were caused by the gap among the nonreligious in willingness to help an atheist purchase auto insurance compared with the control group. Those with a strong religious identity were 8.2 percent more likely to help an atheist purchase auto insurance than their nondescript acquaintance (control), whereas respondents with a weak religious identity were 30.8 percent less likely to help their nondescript acquaintance purchase insurance than they were to help an atheist. Once again, the largest difference in willingness to help a neighbor was among the nonreligious. This also explains the difference for Republicans. The highly religious were only slightly more enthusiastic about helping their nondescript acquaintance purchase auto insurance than they were to help a Republican neighbor (4.5 percent), but the enthusiasm for helping a Republican mail a package was so much greater among those with a weak religious identity (23.3 percent) that the coefficient achieves statistical significance. It appears that religious social identity influences how enthusiastic people are about helping their neighbors mail a package. Consistent with the dominant view, the stronger the religious identity of the respondent, the more likely they are to help an acquaintance mail a package. Consistent with the argument of this book, anti-atheist bias in the United States is contextual. When people are asked to think about helping an atheist with

Figure 6.4 How Group Identification Influences Probability of Assistance with Nonpolitical Activities, by Religious Identity, 2014 Survey Purchase Auto Insurance

Mail a Package

0.9

0.9

0.8

0.8

0.7 M

0.7

0.6

P

0.6

0.5

0.5 0.4

0.4 1

A Atheist theist

1.8 2.5 3. 2 2. 5 Strength Religious Identity S trength of Re ligious Ide ntity Control Control

Democrat Democrat

4

Re Republican publican

1

A Atheist theist

2.5 3.2 1.8 2.5 3.2 Strength Religious Strength of Re ligious Identity Identity Control Control

Source: Pew Research Center, 2014 U.S. Religious Landscape Study. Note: Lines are the predicted willingness to help for each experimental group.

Democrat Democrat

4

Republican Republican

96

Religious Identity in US Politics

everyday tasks, they are more likely to help them than they are to help a Democrat or a Republican. Only those with the strongest religious identity are less likely to help an atheist mail a package than they are a nondescript neighbor. Americans with a weak religious identity are more enthusiastic, but those with a strong religious identity are perfectly willing to help atheists in need of help. On the one hand, religious Americans are among the most kind, generous, and neighborly citizens. Yet, social identities can change that. In some contexts, atheists are less likely to receive help from their religious neighbors than are obese, Latino, openly gay, or nondescript Americans. However, even in these contexts, atheists receive more favorable treatment than would a Democrat or a Republican. Much of the perceived division in US society between religious and nonreligious people is rooted in partisan competition. The size of the gap depends on how one conceptualizes and measures religiosity. The findings I present in this chapter suggest that strong religious identifiers do not hold strong antireligious sentiments and that they are just as likely to help an atheist as they are anyone else. When religiosity is measured in a way that causes religion and partisanship to align (as often happens with church attendance), one can expect to find a strong God gap in the United States. This is not because religion itself is causing out-group antipathy but because partisan affective polarization is making religion seem more divisive than it really is. Compared with the animosity partisans express toward each other, the religious divide in the United States is not that bad. Furthermore, if religious social identity is the root cause of the religious divide, I am optimistic that the chasm can be bridged with greater social contact between believers and nonbelievers.

Conclusion

1. Mohandas Gandhi, “The Ashram of Soul-Force,” in Mahatma Gandhi’s Ideas ed. C. F. Andres (London: Allen, 1929), 129. 2. Marc Gopin, Holy War, Holy Peace: How Religion Can Bring Peace to the Middle East (Oxford, UK: Oxford University Press, 2002). 3. Elizabeth Shakman Hurd, Beyond Religious Freedom: The New Global Politics of Religion (Princeton, NJ: Princeton University Press, 2015). 4. Alan Cooperman et al., 2014 U.S. Religious Landscape Study (Washington, DC: Pew Research Center, 2015). 5. For more on this, see the IRS instructions at https://www.irs.gov/charities -non-profits/annual-exempt-organization-return-who-must-file and https://www.irs .gov/charities-non-profits/churches-integrated-auxiliaries-and-conventions-or -associations-of-churches.

Notes

The Religious Divide

97

6. Cooperman et al., 2014 U.S. Religious Landscape Study. 7. Matthew L. Jacobsmeier, “Religion and Perceptions of Candidates’ Ideologies in United States House Elections,” Politics and Religion 6, no. 2 (2013): 342–372. 8. Ara Norenzayan and Azim F. Shariff, “Origin and Evolution of Religious Prosociality,” Science 322, no. 5898 (2008): 58–62. 9. Benjamin Grant Purzycki et al., “Moralistic Gods, Supernatural Punishment, and the Expansion of Human Sociality,” Nature 530 (Febuary 18, 2016): 327–330. 10. Purzycki et al., “Moralistic Gods, Supernatural Punishment.” 11. This actually happened; there was no deception in this experiment. 12. Will M. Gervais, Azim F. Shariff, and Ara Norenzayan, “Do You Believe in Atheists? Distrust Is Central to Anti-Atheist Prejudice,” Journal of Personality and Social Psychology 101, no. 6 (2011): 1189. 13. Gervais, Shariff, and Norenzayan, “Do You Believe in Atheists?” 14. Shanto Iyengar, Gaurav Sood, and Yphtach Lelkes, “Affect, Not Ideology: A Social Identity Perspective on Polarization,” Public Opinion Quarterly 76, no. 3 (2012): 405–431. 15. Iyengar, Sood, and Lelkes, “Affect, Not Ideology.” 16. Carroll Doherty et al., “Partisanship and Political Animosity in 2016” (Washington, DC: Pew Research Center, 2016). 17. Marilynn B. Brewer, “Multiple Identities and Identity Transition: Implications for Hong Kong,” International Journal of Intercultural Relations 23, no. 2 (1999): 187–197; Sonia Roccas and Marilynn B. Brewer, “Social Identity Complexity,” Personality and Social Psychology Review 6, no. 2 (2002): 88–106; Marilynn B. Brewer and Kathleen P. Pierce, “Social Identity Complexity and Outgroup Tolerance,” Personality and Social Psychology Bulletin 31, no. 3 (2005): 428–437. 18. Lilliana Mason, “The Rise of Uncivil Agreement: Issue Versus Behavioral Polarization in the American Electorate,” American Behavioral Scientist 57, no. 1 (2013): 140–159; Mason, “‘I Disrespectfully Agree’: The Differential Effects of Partisan Sorting on Social and Issue Polarization,” American Journal of Political Science 59, no. 1 (2015): 128–145. 19. Thomas F. Pettigrew, “Intergroup Contact Theory,” Annual Review of Psychology 49, no. 1 (1998): 65–85; Pettigrew and Linda R. Tropp, “A Meta-Analytic Test of Intergroup Contact Theory,” Journal of Personality and Social Psychology 90, no. 5 (2006): 751. 20. Doherty et al., “Partisanship and Political Animosity in 2016.” 21. Maykel Verkuyten, The Social Psychology of Ethnic Identity (London: Psychology Press, 2004); Maykel Verkuyten and Ali Aslan Yildiz, “National (Dis)Identification and Ethnic and Religious Identity: A Study Among Turkish-Dutch Muslims,” Personality and Social Psychology Bulletin 33, no. 10 (2007): 1448–1462. 22. Kristopher J. Preacher, Derek D. Rucker, and Andrew F. Hayes, “Addressing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions,” Multivariate Behavioral Research 42, no. 1 (2007): 185–227.

7 When Religious and Partisan Social Identities Collide

throughout the course of life, social identities are relatively unstable. Some identities persist throughout life, but others are readily abandoned when circumstances change. Previous examples discussed in this book make it clear that similar to any other social identity, religious social identity can remain an important element of people’s sense of self long after they stop attending church, or they can be discarded as soon as societal arrangements make it inconvenient to identify with the religion. In this chapter I explore how people with both partisan and religious social identities respond when informed of how members of their partisan group feel about members of their religious group. Previous research discusses what might happen when religious identities compete with partisan identities, but previous research focuses on attitudes rather than the strength of subsequent religious or partisan identities.1 In October 2016, many Republican Party leaders began to publicly retract their support from nominee Donald Trump. The release of Access Hollywood tapes that recorded Trump speaking lewdly about women was too much for many Republican elected officials to endure and motivated them to withdraw their support from him. At the time, many in the news media wondered if Republican Party supporters would follow. Republican voters held firm in their support of the nominee. Trump publicly disputed several women who claimed he had sexually assaulted them and argued that the “mainstream media” was using these false accusations to “rig” the election. 2 Despite evidence to the contrary, evangelical and religious Republican supporters believed that Trump’s accusers were insincere, politically motivated, or lying.3 Many wondered how a group that so strongly condemned the conduct of President Bill Clinton with Monica Lewinsky and fought so staunchly for a

Compared with other psychological traits that are stable

99

100

Religious Identity in US Politics

stronger moral voice in the public sphere could condone—at least implicitly— such immoral statements as Trump made to Access Hollywood.4 One explanation is that motivated reasoning changes the way people process information. This theory holds that all political thinking is suffused with feeling that arises within milliseconds after exposure to a familiar political object or event. The human brain does not have the capacity to process every bit of information people encounter through an effortful cognitive mechanism. Rather, people use prior attitudes, feelings, and affect to guide their evaluation of new information and events. All concepts evaluated in the past have been charged with either a positive or negative affect that spontaneously (and often unconsciously) springs into action whenever someone is exposed to a name, face, group, or image. Most attitudes are formed through this “hot” unconscious process. That is, most people’s attitudes are formulated without any conscious thought directed toward them.5 The brain organizes long-term memories associatively. It is helpful to think of memories as being stored in nodes connected via associative neural networks. People have thousands of concepts stored in their memories, each of which is charged with affect and linked with other similar concepts. When people encounter a concept in the physical world, it triggers association with other concepts that infuse the existing object with affect from their prior experience. The affect associated with a concept, rather than cognitive reasoning, motivates the kind of attitude formed. When people encounter information incongruent with existing affect, they tend to evaluate the concept in a manner that helps them retain their existing affect toward the concept. Only on rare occasions do people process counterattitudinal information in a manner that would cause them to develop a new affect toward the concept.6 In the case of Trump, Republican voters’ experience with Trump prior to the release of the Access Hollywood video caused them to have either a positive or negative affect toward him. Those who previously had strongly positive feelings about Trump reacted to the information differently than those with prior negative views. Those with negative views perceived the tape as a confirmation of everything negative they had previously associated with Trump. However, those with positive affect toward Trump approached the tapes with skepticism about their veracity. They questioned the timing, authenticity, and way the tapes were released and the credibility of those who later accused Trump of behavior consistent with the language on the tapes. In reality, the facts of the matter were irrelevant to people’s opinions. Most people formulated their attitudes about the tapes through an unconscious, automatic, psychologically motivated process. If facts are irrelevant to negative or positive political affect, what can be done to change people’s minds? People usually overcome politically motivated reasoning when the negative information environment becomes so prevalent that people can no longer ignore or argue against it. Over the

When Religious and Partisan Social Identities Collide

101

course of a fictional campaign season, participants in an experiment were randomly assigned to experience varied exposure to negative information about their preferred candidates. The study found that when people are exposed to a small amount of negative information about the candidate, motivated reasoning causes them to discount this information and leads to stronger positive affect toward their preferred candidate. In contrast, prolonged exposure to negative information about preferred candidates can change people’s views toward the candidates. After more than 28 percent of the total information about the candidates is negative, people will begin to develop negative views about their preferred candidates. When people feel sufficient emotional discomfort, it can motivate them to think long and hard about information they receive and might cause them to change their minds.7 These findings come from a controlled laboratory environment, so it is possible that people never reach their tipping point in the real world. Because people select an information environment consistent with prior attitudes, it is possible that average citizens are never exposed to enough negative information about their preferred candidates to cause them to reconsider a previous affect toward them. Others argue that when two mental concepts collide, it can motivate people to process information through the cognitive process, which takes more effort. Even strong partisans can have positions on issues that differ from those taken by representatives from their political parties. When partisans agree with their parties, positive affect associated with the party reinforces the partisan attitude, which strengthens the attitude.8 By contrast, when disagreement arises, partisan affect conflicts with affect related to the issue. The amount of influence exerted by partisan motivations depends on people’s ability to justify their partisan identity. Those with stronger affect toward the party than toward the issue position will usually change their position on the issue to match that of their party. What happens when a political party takes the opposite position from that of a partisan, and that person has equally strong affect toward both the party and the issue? If we were to crawl inside the mind of a person in this situation, we would see that the issue node is connected to the party node, and each of them reinforce positive affect with the other. When a political party takes a stance on an issue opposite that of a partisan, his or her mind cannot maintain the status quo. Either the party will be infused with negative affect because of this issue position, the issue will become infused with negative affect because of the party position, the two will nodes will no longer be associated with each other, or the person will change his or her views about the one to accommodate the other. Kevin Mullinix varied party positions on salient issues and tested how people responded when their party’s stance on an issue caused internal motivations to collide, and this is precisely what he found.9 When motivations were in competition, the motivation with the greater salience dominated the reasoning process. If the

102

Religious Identity in US Politics

party was more important to respondents than the issue was, they changed their positions on an issue to match that of their political parties. If the two were equally salient, the two motivations offset each other, and people did not engage in partisan- or issue-motivated reasoning. If an issue position can be sufficiently strong to cause people to reevaluate their affect toward their political parties, certainly religious social identity could do the same. As I discuss in the next section, people’s identification with their religions can be strong and is often closely associated with their views of their own selves. As the case of the Access Hollywood tapes illustrates, sometimes affect about someone associated with a political identity conflicts with affect associated with a religious identity. For religious Republicans, the positive affect associated with Trump because he represents their party conflicts with the negative affect associated with supporting someone who defies religious morals and values. Similar to what Mullinix and Eric Groenendyk find concerning attitude conflict, we should expect one identity to have primacy over the other. If so, the reason that some religious Republicans supported Trump to the end is that their partisan identity had primacy over their religious identity. What happens when two identities continually collide, making it impossible to glide between the two easily? Internal commitment to the norms of the social group influence how people’s social identities influence their attitudes. When facing a threat to their social identities from those who belong to another social group with which they also identify, most people will adopt the view consistent with the high-status social group. However, when their internal commitment to the social group being threatened is stronger, it can lead to disidentification from the high-status group. For example, Dutch Muslims with strong social and behavioral Muslim identities were more likely to abandon their national identities when they thought that Dutch society persecuted Muslims.10 I would expect the same psychological tendencies to influence people’s attitudes about political parties and religions in the United States. What happens in the United States when a political party begins to advocate views inconsistent with the theology of some members of that party? My theoretical expectations are displayed in Table 7.1. The stronger people’s commitment to their religious identity, the more I would expect them to disidentify with the political party advocating views inconsistent with their religious views. By contrast, those with stronger partisan identities will become weaker religious identifiers. Just as when people’s strong issue positions conflict with their social identities, the relative salience of the two determines which one prevails.

When Partisanship and Religion Collide

When Religious and Partisan Social Identities Collide Table 7.1

103

Influence of Collision of Religious and Partisan Social Identities on Identity Strength

Strong Religious Identity Weak Religious Identity

Strong Partisan Identity

Individual becomes weaker identifier with other group Individual becomes stronger partisan

Weak Partisan Identity

Individual becomes weaker partisan Individual becomes weaker identifier with the lowerstatus group

Social identity competition has no effect on those who weakly identify with both religions and political parties. Because both partisan and religious identities are not important to these people, there is no conflict. Those who strongly identify with both religions and political parties face a conundrum different from that of weak identifiers. The competing identities will cancel each other out among people for whom religious and political identities are equally salient and strong. However, others will place higher priority on one of the two identities. If one identity is weaker than the other, it has a weaker influence on subsequent attitudes. Ultimately, people weaken their identification with the lower-status, less-salient group. In summer 2016, I devised a social experiment to test these expectations. Every research project that involves human subjects must make certain that research participants are not harmed in any way. Specifically, this means that researchers should not tell participants anything untrue unless it is absolutely necessary. Rather than focus on issues, I created conflict between social identities by informing participants about partisan attitudes toward members of various religious groups. The expectation is that people for whom religious identities were more important than their partisan identities would become weaker partisans if informed that the majority of their party had negative views about members of their religious group. In contrast, those for whom partisan identities took priority over their religious identities would become weaker religious identifiers when informed that many members of their party had negative views about people who belonged to their religions. The Pew Research Center conducted a survey in summer 2014 in which it asked Americans to report their feelings about religious groups. Respondents were told to use a feeling thermometer ranging from 0 to 100, where 0 reflects the coldest, most negative possible rating and 100 the warmest, most positive rating toward various religious groups. The groups

Testing the Expectations

104

Religious Identity in US Politics

with the lowest average ratings were Muslims and atheists (40 points each), with Jews (64.5), Catholics (63.5), and evangelical Christians (62) receiving the highest average ratings. Interestingly, each of these ratings was broken down by political party identification; Figure 7.1 displays the ratings of various religious groups by partisanship. In Figure 7.1, the major religious groups are ordered from least liked to most liked among each of the partisan groups. The religious groups most disliked by Democrats were Mormons, atheists, and Muslims, with Catholics and Jews the most liked of the religious groups. Republicans had considerable disdain for Muslims and atheists and showed high regard for Catholics, Jews, and evangelical Christians. The actual numbers were more extreme among Republicans, meaning that the average Republican had much colder views toward Muslims and atheists than did the typical Democrat. I included this information in a national survey that we conducted in June 2016. Near the beginning of the survey we asked people, “In politics today, do you consider yourself a Republican, Democrat, or Independent?” and used a couple more questions to determine if they were strong partisans

Figure 7.1 Graphic Depiction of Religious Groups, Rated by Partisanship

D

Democrats

44

46

47

Mormons M ormons

A Atheists theists

M Muslims uslims

61

62

Buddhi Buddhists sts

Ca Catholics tholics

Jews

66

67

Ca Catholics atholics

Jews

53

54

57

E Evangelical vangelical Chri Christians stians

H Hindus indus

R

Republicans

33

34

M Muslims uslims

A Atheists theists

47

49

52

Hindus Hindus

Buddhi Buddhists sts

M Mormons ormons

71

Evangelical Evangelical Christians Christians

When Religious and Partisan Social Identities Collide

105

or if they were Independents who leaned partisan. We also asked, “What is your present religion, if any?” Respondents chose from the following options: Protestant, Roman Catholic, Mormon, Eastern or Greek Orthodox, Jewish, Muslim, Buddhist, Hindu, atheist, agnostic, secular, Jehovah’s Witness, or none. Then, using this information, we randomly assigned each participant to one of two experimental groups. In the Democrat group, respondents were shown the image in Figure 7.1 for Democrats and were told: “Earlier you told us that your present religion is [name of religion inserted here]. This graphic displays the attitudes of Democrats toward members of various religious groups, based on a comprehensive national survey conducted in 2014. The higher the score, the more Democrats liked members of the religious group, while lower scores reflect greater Democrat dislike of members of that religious group. On average, Democrats have very positive views of Jews and Catholics and negative attitudes toward Mormons, atheists, and Muslims.” The Republican experiment was nearly identical, except that respondents were shown the Republican portion of Figure 7.1 and told: “Earlier you told us that your present religion is [name of religion inserted here]. This graphic displays the attitudes of Republicans toward members of various religious groups, based on a comprehensive national survey conducted in 2014. The higher the score, the more Republicans liked members of the religious group, while lower scores reflect greater Republican dislike of members of that religious group. On average, Republicans have very positive views of evangelical Christians, Jews, and Catholics and negative attitudes toward atheists and Muslims.” After viewing the image, respondents answered a series of questions about how much discrimination the various political parties show toward members of their own religious groups and how they think members of their religious groups are treated by the political parties. Following these questions, we inquired about the strength of the respondents’ identification with their own religious groups. We also included measures of religious and partisan disidentification adapted from Maykel Verkuyten and Ali Aslan Yildiz’s study of Dutch Muslims: “I sometimes feel proud of the [respondent’s political party]” and “I feel emotionally involved in the [respondent’s political party].”11 We used a seven-point “strongly agree” to “strongly disagree” scale for each of these questions, so higher scores actually mean that respondents disagreed with these statements. This experiment was less than ideal for a couple of reasons. First, it is unlikely that members of various religious groups are completely unaware of their standing among partisans. People with strong Republican identification would be aware that evangelical Christians have a more prominent role in the party than do Muslims. Most Democrats are probably also aware that Jews and Catholics are held in higher esteem among Democrats than

106

Religious Identity in US Politics

are other minority religious groups, particularly Mormons. As such, the experiment probably did not provide respondents novel information. The experiment likely did not introduce anything new to strong partisans or those with a strong religious identity; rather, it likely made people think about information they might not have usually considered. This is one reason to expect a small effect from this experiment. Another is that people with a strong religious identity are found among all religions. Because Catholics and Jews can have a strong religious identity, and they are held in high regard by people in both political parties, I essentially excluded them from my analysis. Let us consider someone who is both a strong Catholic and a strong Republican. Learning that Republicans love Catholics would not create conflict between those two identities. In fact, this information would likely strengthen both identities. Only those with religious and partisan identities who also happened to be members of disliked religious groups would be confronted with conflicting identities. Of the 1,290 respondents who participated in the survey, only 107 of them identified as members of disliked religious groups. This means that I expect small effect sizes and that it takes considerable analysis to tease out the effects of the experiment on this small group of respondents. The statistical analyses are complex and explaining them in detail might be confusing. Appendix 7 contains all of the information about the statistical models and information on how to replicate the results presented here. How do people respond to information about how they are viewed by members of political parties? It depends on the strength of their partisan and religious identities, just as I expected. First, let us look at the big picture. Figure 7.2 displays the model-predicted partisan disidentification for respondents. The model predictions for people in the Republican groups are on the right, and the predictions for those in the Democrat groups are on the left. Higher scores on the Y-axis indicate that respondents had a stronger disassociation with their political parties. For example, someone with a score of seven on the scale did not feel emotionally involved in his or her political party or feel proud of that political party. The labels along the Xaxis show the respondents’ strength of partisan and religious identification as they reported before the experiment. Starting from the left, we can see that Democrats with strong religious identities reported the highest level of partisan disidentification when they were informed of the way other Democrats thought about members of their various religious groups. In fact, comparing the 95 percent confidence intervals shows that Democrats with strong religious identities were significantly more likely to disidentify with their own political party when informed about the relative status of religious groups in the eyes of other Democrats than were strong Democrats with strong religious identities. If we look at how Republicans with strong religious identities responded to the

0

T Treatment reatment (D (Democrat) emocrat) Strong Re ligious Ide ntity Strong Religious Identity T Treatment reatment (Re (Republican) publican) T Treatment reatment (D (Democrat) emocrat)

trong) Re Republican publican (S (Strong)

Re publican Republican

Re publican (L ean) Republican (Lean)

Inde Independent pendent

D emocrat (L ean) Democrat (Lean)

Democrat Democrat

D emocrat (Strong) (Strong) Democrat

Re Republican publican (Strong) (Strong)

Re publican Republican

Re publican (L ean) Republican (Lean)

Inde pendent Independent

D emocrat (L Democrat (Lean) ean)

Democrat Democrat

D emocrat (Strong) (Strong) Democrat

Re Republican publican (Strong) (Strong)

Re publican Republican

Re publican (L ean) Republican (Lean)

Inde pendent Independent

D Democrat emocrat (L (Lean) ean)

D emocrat Democrat

D emocrat (S trong) Democrat (Strong)

Re publican (S trong) Republican (Strong)

Re publican Republican

Re publican (L ean) Republican (Lean)

Inde pendent Independent

D Democrat emocrat (L (Lean) ean)

Democrat Democrat

D Democrat emocrat (Strong) (Strong)

P Party arty D Disidentification isidentification

Figure 7.2 Partisan Disidentification Among Strong and Weak Identifiers with Religion

5

4.5

4

3.5

3

2.5

2

1.5

1

0.5

ligious Ide ntity Weak W eak Re Religious Identity T Treatment reatment (Re (Republican) publican)

Note: The bars show the model-based predicted party disidentification. The solid vertical lines represent the 95 percent confidence interval. The complete regression model is displayed in Table A7.1 in Appendix 7.

107

108

Religious Identity in US Politics

Republican experiment, we see the same pattern of findings. Republicans with strong religious identities were significantly more likely to disidentify with the Republican Party when informed of how other Republicans viewed members of religious groups in the United States. This is consistent with the expectations laid out in Table 7.1. When partisan and religious identities compete, the higher-status identity wins. Americans with strong religious identities and weaker partisan identities become less partisan when they encounter information that suggests other members of their partisan groups have negative attitudes toward members of their religious groups. What happens when weak religious identifiers and weak partisans encounter information that puts their two weak identities in competition? The results in Figure 7.2 predict that political Independents with weak religious identities would become a little less Independent when informed about the status of religious identifiers in the two major political parties, but the differences are not statistically significant. In short, just as the theory of competing social identities predicts in Table 7.1, those with weak political and religious identities chose to adopt the identity perceived as higher status. In this case, they became slightly less committed to their Independent political identity. When respondents who were strong partisans but weak religious identifiers learned of how religions were viewed by those within their political parties, they became stronger partisans. In the Democrat group, strong Democrats with weak religious identities reported closer identification with their party, as did strong Republicans. In fact, the strong partisans with weak religious identities were significantly more likely to report closer partisan identification with their own parties regardless of the experimental treatment they received. Strong Democrats reported a greater affinity to the Democratic Party if they had weak religious identification, and strong Republicans reported greater identification with the Republican Party if they had weak religious identification. If the expectations in Table 7.1 are correct, this is exactly what we should find. Finally, when religious identities collide with partisan identities, people who are strong identifiers with both must ultimately choose between them. In this experiment, strong Democrats with strong religious identities became significantly stronger partisans when they were informed about the views of respondents from their party toward members of various religious groups. They did the same when told about what Republican respondents thought about religious Americans. In fact, strong Democrats with a strong religious identity had the highest levels of post-treatment identification with their political party of any group in the experiments. The next-highest level of post-treatment partisan identification was among strong Republicans with a strong religious identity. Again, strong Republicans reported similar levels of partisan identification after they were informed about what respondents from the political parties thought about members of religious groups in the

When Religious and Partisan Social Identities Collide

109

United States. Once more, the findings from this experiment are consistent with the theory of competing social identities and suggest that when partisan and religious identities collide, strong partisans tend to favor their partisan over their religious identity, especially when both are strong. Of course, I cannot be sure that the participants’ partisan identities were challenged by the information they saw describing the attitudes of party members toward members of their religious groups. As I alluded to previously, members of religious groups viewed favorably by partisans would have both their religious and partisan identities reinforced by that information. After all, they would have seen that they were at the top of two relevant social identities. Meanwhile, those who identified with disliked religions would have had the opposite experience. Assuming that these strong religious identifiers were unaware of the information given to them in the experiment, they would have suddenly found out that most people from their political party did not like members of their religion. The next set of analyses examines the behavior of participants who identified as members of a disliked religious group. The dependent variable in these analyses is the respondents’ partisan identification in a test given after the experiment described previously. The models used strength of religious identity, strength of partisan identity, and a variable indicating which experimental treatment the person received as independent variables. Unfortunately, not enough respondents identified with a disliked religion to explore how each predicted probability in Figure 7.3 compares with the findings in the previous section. Fortunately, enough respondents who identified with disliked religions were randomly assigned to the survey questions of interest to allow a robust test of the competing identities hypothesis. How did members of these religious groups respond to information that members of their political parties had negative views toward members of their religious groups? Beginning with the response political Independents had when informed that members of their religious groups were held in poor esteem by Democrats and Republican respondents, I found no difference between Independents with a strong or a weak religious identity. The probability that an Independent would identify as such in the post-test is 1 on the scale. Consistent with my argument in this chapter, when partisan identities do not conflict with religious identities, it has no effect on partisan identification. When Democrats learned that members of their religious group were not regarded well by other Democrats, the strength of their religious identification had a strong influence on their subsequent partisan identification. Among strong Democrats, the strength of their religious identity did not make much of a difference. Among weak religious identifiers, the probability of identifying as a strong Democrat on the post-test was 0.50, and the probability increased slightly (0.57) for strong religious identifiers. However,

110

Religious Identity in US Politics

P Probability robability of P Partisan artisan n Ide Identity ntity (P (Post-Test) ost-Test)

Figure 7.3 Partisan Disidentification Among Members of Unpopular Religious Groups $" !#," !# ," !#+" !# +" !#*" !# *" !#)" !# )" !#(" !# (" !#'" !# '" !#&" !# &" !#%" !# %" !#$" !# $" !" Not Not Important Im portant

Very Very Important Important

Strong Democrat Strong Democrat

Not Not Important Important

Very Veery Important Important

Democrat Democrat

Not Not Important Important

Very Very Important Important

Independent Independent

Not Not Important Important

Very Very Important Important

Republican Republican

Not Not Important Important

Very Very Important Important

Strong Strong Republican Republican

Democrat Democrat Treatment Treatment

P Probability robability of P Partisan artisan Ide Identity ntity (P (Post-Test) ost-Test)

Inde pendent Independent

L ean Democrat Democrat Lean

D emocrat Democrat

Strong D emocrat Strong Democrat

$" !# ," !#," !# +" !#+" !# *" !#*" !# )" !#)" !# (" !#(" !# '" !#'" !# &" !#&" !# %" !#%" !# $" !#$" !" N ot Not Im portant Important

Very Very Important Important

Strong Democrat Democrat Strong

Not Not Important Important

Very Very Important Important

Not Not Important Important

Democrat Democrat

Very Very Important Important

Independent Independent

Not Not Important Important

Very Very Important Important

Republican Republican

Not Not Important Important

Veery Very Important Important

Strong Republican Republican Strong

Republican Treatment Treatment Republican Independent Independent

L ean Republican Republican Lean

Re publican Republican

Strong Re publican Strong Republican

Note: The bars show the model-based predicted partisan identification in the post-test. The 95 percent confidence interval is omitted for ease of interpretation. See Table A7.2 in Appendix 7 for the complete regression table.

Democrats were not at all likely to become stronger Democrats on the posttest. In fact, Democrats affiliated with minority religious groups became weaker partisan identifiers after learning that respondents from their party did not hold them in high regard. Among weak religious identifiers, the probability of selecting Democrat as their partisan identity on the post-test was

When Religious and Partisan Social Identities Collide

111

0.36 and was four times smaller among strong religious identifiers (0.09). There was a 70 percent chance that Democrats who strongly identified with a minority religion would lean more Democrat after seeing what other Democratic partisans thought of members of their religious community. These findings are consistent with my argument in this chapter and clearly demonstrate how the strength of two social identities influence each other when in competition. Among Democrats, the relative strength of respondents’ religious and partisan identities predicted subsequent identifications with those two groups. When religious and partisan identities collided, stronger partisans were less likely to disidentify with their parties, whereas weaker partisans were more likely to disidentify with their political parties. A similar pattern of findings emerges when I examine the response of those who identified with an unpopular religion and with the Republican Party to information about how Republicans view members of their religion. The probability that strong Republicans would identify more strongly as Republicans was 1.0, regardless of the strength of their religious identity. Those with a weak religious identification selected Republican identity (though not strongly), whereas those with a strong religious identity were just as likely to identify as strong Republicans as they were Republicans. When both identities were strong, the relative importance to the participant of one of these identities predicted the response, but identity conflict did not cause strong Republicans to abandon their political party. The strength of respondents’ religious identities had a stronger influence on Republicans’ subsequent identification with their party. Among weak religious identifiers, the probability of identifying as a Republican was 0.16, compared with 0.00 among strong religious identifiers. The most likely response among Republicans was disidentifying with the Republican Party. Among weak religious identifiers, there was a 50 percent chance that they would lean Republican, compared with a 75 percent chance of this among strong religious identifiers. Roughly 25 percent of Republicans with a strong religious identity completely disidentified with their party and became political Independents. The consistent findings here present compelling evidence that when people with strong religious and strong partisan identities learn information that causes conflict between these two social identities, it forces them to choose one identity over the other. When they face a choice between more strongly identifying with the (low-status) religious group or the (high-status) political party, the relative importance of those identities to the individuals determines the outcome. Strong religious identifiers with weaker partisan identities become less partisan when the two identities conflict. Only strong partisan identifiers retain a stronger partisan identification after learning that members of their religion are held in low regard by other members of their political parties.

112

Religious Identity in US Politics

The social groups with which people affiliate have a strong influence on their attitudes. When the majority of one social group with which an individual identifies has negative attitudes about the members of another social group with which that person associates, it causes psychological dissonance. People resolve this conflict by abandoning their lower-status identity and becoming more strongly associated with the higher-status group. When deeply religious Republican members of disliked religions (e.g., Mormon, Islam, Buddhism, Hinduism) are informed about their status within the Republican Party, they downplay their religions and become emotionally closer to and proud of their affiliation with the Republican Party, and the same is true for Democrats.

Conclusion

1. Michele F. Margolis, “How Far Does Social Group Influence Reach? Identities, Elites, and Immigration Attitudes,” Journal of Politics 80, no. 3 (2018): 772–785. 2. Alex Swoyer, “Donald Trump: This Whole Election Is Being Rigged,” Breitbart.com, October 14, 2016, http://www.breitbart.com/2016-presidential-race/2016 /10/14/donald-trump-this-whole-election-is-being-rigged/. 3. Lauren Markoe, “Poll: Trump Support Remains Steady Among Evangelicals,” Religion News Service, October 11, 2016, http://religionnews.com/2016/10/11/poll -trump-support-remains-steady-among-evangelicals/. 4. Laurie Goodstein, “Donald Trump Reveals Evangelical Rifts That Could Shape Politics for Years,” New York Times, October 17, 2016, http://www.nytimes .com/2016/10/17/us/donald-trump-evangelicals-republican-vote.html?_r=0. 5. Milton Lodge and Charles S. Taber, “The Rationalizing Voter: Unconscious Thought in Political Information Processing,” unpublished paper, Stony Brook University, 2008; Lodge and Taber, The Rationalizing Voter, Cambridge Studies in Public Opinion and Political Psychology (Cambridge, UK: Cambridge University Press, 2013); Patrick W. Kraft, Milton Lodge, and Charles S. Taber, “Why People ‘Don’t Trust the Evidence’: Motivated Reasoning and Scientific Beliefs,” Annals of the American Academy of Political and Social Science 658, no. 1 (March 1, 2015): 121–133. 6. Charles S. Taber and Milton Lodge, “Motivated Skepticism in the Evaluation of Political Beliefs,” American Journal of Political Science 50, no. 3 (2006): 755–769. 7. David P. Redlawsk, Andrew J. W. Civettini, and Karen M. Emmerson, “The Affective Tipping Point: Do Motivated Reasoners Ever ‘Get It’?” Political Psychology 31, no. 4 (2010): 563–593. 8. Eric Groenendyk, Competing Motives in the Partisan Mind: How Loyalty and Responsiveness Shape Party Identification and Democracy (Oxford, UK: Oxford University Press, 2013). 9. Kevin J. Mullinix, “Partisanship and Preference Formation: Competing Motivations, Elite Polarization, and Issue Importance,” Political Behavior 38, no. 2 (2016): 383–411. 10. Maykel Verkuyten and Ali Aslan Yildiz, “National (Dis)Identification and Ethnic and Religious Identity: A Study Among Turkish-Dutch Muslims,” Personality and Social Psychology Bulletin 33, no. 10 (2007): 1448–1462. 11. Verkuyten and Yildiz, “National (Dis)Identification and Ethnic and Religious Identity.”

Notes

8 How Political Identities Influence Religious Beliefs

within a social group is challenged by their identification with another social group, people will disidentify with the lower-status group and identify more strongly with the higher-status group. I also demonstrated that in US politics, those with strong partisan and religious identifications become stronger partisans when their two identities conflict. In this chapter I build on the same theoretical foundation and describe how partisan identity can influence individuals’ attitudes about religious doctrine. Not only does partisanship trump religion in terms of social identity but the process of becoming stronger partisans and weaker religious identifiers leads people to alter their views about doctrinal issues within their religious communities. Because most people lack complete information, they often look to their social groups for cues on how to evaluate political and religious issues.1 What happens when one social group sends one cue, and the other social groups sends a completely different cue? Which has the stronger influence on attitudes? When church teachings disagree with the dominant view of members of a political party, people reevaluate their membership in both and alter their views accordingly. Particularly when the gap between religion and party widens beyond reconciliation, those for whom partisanship is more important than their religious identity might choose to abandon their place of worship and become stronger partisans. Similarly, those for whom religious identity is more important than partisan identitiy will abandon their party and become more devoted to their religion. The challenge for most comes when abandoning their party or religion is untenable. What do

In the previous chapter I showed that when people’s status

113

114

Religious Identity in US Politics

people do when they have such strong identities with both a political party and a religion that leaving one or both is not an option? Most people like to think they are not biased. When people find out credible facts or information, one might assume they could arrive at the same unbiased conclusions. However, when people encounter new information, they unconsciously consider the implications of particular decisions throughout the process of arriving at those conclusions. Ziva Kunda termed this phenomenon “motivated reasoning,” and this concept has changed the way social scientists think about persuasion.2 Like most groups to which humans belong, people derive pride and self-esteem from the political party with which they affiliate. The stronger their attachment to a political party (group), the more things that affect it influence its members’ self-esteem. Because the group is an extension of the self, people try to enhance the status of groups with which they identify. One way to do this is to denigrate competing groups—often called the outgroup.3 During national elections, voters affiliated with political parties view their party as extensions of themselves. The party’s wins and losses are not seen as a distant contest between people with whom the partisans have no connections, relationships, or associations. Even when voters have never had any personal contact with their party’s candidates, partisan identifiers internalize party victories and losses. Something similar happens during rival week in college football. Fans from the opposing sides hurl insults at each other, and those supporting the opposing teams internalize the outcome of the game in their personal lives. Supporters of the winning teams feel pride and accomplishment even though they had nothing to do with the outcome of the games. Supporters of the losing teams feel discouraged and upset even when they do not have personal relationships with any of the players. This is similar to how identifying with a political party influences the way people evaluate political information. People simply accept information that makes their own political party look good, and they even more easily accept information that makes the other political party look bad.4 By contrast, when strong partisans hear something that makes their own party look bad, they employ a variety of cognitive hurdles to avoid believing that information. Typically, the strength of partisan attachment is measured with two questions in a survey. First, people are asked to which party they belong. If they say that they are “Independent” or “do not belong to any party,” they are asked if they “lean” Democrat or Republican. If they say “Democrat,” they are asked if they consider themselves a “strong” Democrat; likewise for those who respond “Republican.” Those who say that they lean Democrat or Republican are called “weak” partisans, and only those who do not lean either way are labeled “Independent.”

How Political Identities Influence Religious Beliefs

115

During Operation Iraqi Freedom, a group of researchers decided to explore how partisanship influenced people’s attitudes about the casualty levels in Iraq. From October 2003 to December 2004, three different groups of people were recruited to answer a series of questions about the conflict in Iraq. Each group participated in three different surveys that spanned two to three months. As part of the studies, each participant was asked to report the number of US casualties in Iraq. The true figure fluctuated quite a bit during the months of the study, but the overall number per month increased over time. February 2004 had the lowest reported number of casualties (23), and November 2004 had the highest (141). The study design allowed the researchers to measure what people thought about the number of casualties during each month of the fourteen-month study. After respondents were asked how many casualties the United States had sustained, respondents were asked if the number of casualties was large or small. What the studies found was astonishing. Most people responded accurately about the true number of US casualties in Iraq. Participants got the facts right. They knew that with time, more and more US service members were dying in Iraq. However, interpretations of those facts were quite dependent on how strongly the respondents identified with their political parties. In October 2003 there was little divergence in participants’ attitudes about the casualty levels in Iraq. About half of the Democrats and half of the Republicans responded that the casualty number was either “large” or “very large.” As the number of casualties grew and the media coverage of the growing death toll in Iraq dominated the airwaves, beliefs about the number of casualties shifted depending on the strength of respondents’ partisan identification. By August 2004, 95 percent of those who identified as strong Democrats reported that the casualty number in Iraq was either “large” or “very large.” However, only 25 percent of those who identified as strong Republicans believed that the casualty number in Iraq was either “large” or “very large.” The divergence of opinion is explained by how people defined the word large. Most Democrats thought 1,000 casualties represented a “large” number, whereas most Republicans thought that was a “moderate” number of casualties.5 When facts confirm people’s existing attitudes and beliefs, they are more accepting of that information. Psychologists call this univalent information. For example, because most Democrats were opposed to the US intervention in Iraq because they did not want to have US soldiers die in a foreign conflict, Democratic respondents perceived the number of casualties to be “large” or “very large.” The casualty numbers confirmed their belief that the war would result in too many US casualties and interpreted this information as evidence that the level of casualties was too high.

Same Facts, Different Interpretation

116

Religious Identity in US Politics

When information conflicts with people’s existing attitudes or challenges the standing of the groups with which they identify, they are more likely to scrutinize the information more skeptically.6 Psychologists call this ambivalent information. Ambivalent information leads people down one of two paths: to alter their existing views or to find fault with the source of the new information. They argue against the quality of the information when they question the messenger (e.g., the president is not trustworthy) or the source of the information (e.g., the study is biased). This is precisely how Republican participants responded to new information about the casualty levels in Iraq. Recognizing that the number of casualties had grown well beyond the anticipated number, Republicans had to decide how to deal with that fact. Admitting that the number was too high would make their party look bad and, by extension, those who identified with the party. Facing the choice between changing their support for the war in Iraq and changing their attitudes about the human cost of the war, Republican respondents opted to reinterpret the definition of “large” or “very large” casualty numbers. The information bias just described is manifest in how human beings process information. People’s values bias the ways in which they interpret facts and data. People encounter a vast amount of information daily, and for them to accurately process it all is a psychologically demanding task. Therefore, people use psychological shortcuts to help them evaluate information and formulate attitudes. For the most part, biased information processing does not do any permanent damage. However, these psychological preferences tend to manifest themselves in harmful ways when people’s identities and beliefs collide. When new information contradicts one’s deeply valued beliefs, one experiences ambivalence. In contrast, univalent ideas are pleasing to the psyche because they reinforce existing belief systems and actions. Univalent interpretations are biased, but people tend to adopt them because they reinforce their sense of self. Ambivalent ideas are displeasing and cause people to exert more effort when processing the information.7 However, people can hold a variety of conflicting attitudes without experiencing ambivalence if they are unaware of the dissonance. But after people become aware that their attitudes conflict, they are willing to exert considerable effort to reconcile them. Most of the time, people can choose the groups with which they want to affiliate. People choose to identify with groups that have similar worldviews and bring prestige to their self-image. Even people who agree with the views of the Ku Klux Klan (KKK) might choose to identify with a different racist group because the KKK carries negative societal connotations. When people identify with two groups that advocate contradictory views,

Social Identity and Information Bias

How Political Identities Influence Religious Beliefs Table 8.1

117

Influence of Collision of Religious and Partisan Social Identities on Attitudes

Strong Religious Identity Weak Religious Identity

Strong Partisan Identity

Individual adopts higher-status group’s view Individual adopts party’s view

Weak Partisan Identity

Individual adopts religion’s view Individual adopts higherstatus group’s view

they generally have one of four reactions, each dependent on how strongly they value the groups and their own views. Table 8.1 lists the expectations. When people strongly identify with political parties, and they strongly identify with religions that teach ideas contrary to the parties’ positions, they tend to adopt the view of the higher-status group. Members of low-status religions adopt the views of their parties, but those who identify with highstatus religions adopt the views of their religious groups. Those with both weak religious and political identities do the same. Individuals with strong religious identification but weak partisan identification also adopt their religious group’s views regardless of the relative status of the two groups. Finally, people with a strong partisan identity and a weak religious identity adopt the views of their party. An interesting implication of this idea that has not been fully explored is that political party positions can influence religious attitudes, especially among those with strong partisan but weak religious identities. In this chapter I use a unique survey of members of the Church of Jesus Christ of Latter-day Saints (LDS) who live in the Utah/Idaho corridor to examine what happens when people’s religious group takes a position on a political issue that contradicts the dominant views of political parties. This was called the Intermountain LDS Faith survey (ILF). In the previous chapter, I show that in matters of politics, partisan identities often take precedence over religious identities. The extant religion and politics literature also demonstrates the power of partisan identities in determining religious Americans’ political issue positions.8 Although political elites can use coded religious language to generate political support, partisan attachments often have a stronger influence on political attitudes than do religious beliefs.9 In the following analyses, I demonstrate that when a religious group takes a position on a political issue, not only do some people distance themselves from that religion but also they express views on doctrinal matters consistent with their political views. In short, partisanship influences

The Evidence

118

Religious Identity in US Politics

individual views about religious doctrine. In the following study, I use one example to explain how. One challenge for members of the LDS Church is the ambivalence that occurs when political party positions conflict with the teachings of the modern prophets. Those who would like to be active in both the LDS community and the Democratic Party experience ambivalence when their party supports political positions upon which the LDS Church takes an opposing position. For example, the LDS Church’s public support of California’s Proposition 8, defining marriage as between one man and one woman, might have created significant ambivalence for active LDS Democrats in 2008. More than 70 percent of California Democrats voted against the measure. Votes for Proposition 8 were often motivated by political ideology and church attendance.10 Thus LDS Democrats with a strong religious social identity faced a dilemma: to support the position dominant among Democrats or the one dominant among LDS Church members. People have multiple social identities, with only one of them salient in each situation; however, this instance placed two social identities in conflict.11 LDS Democrats were forced either to prioritize one of their social groups above the other and adopt the majority view of that group or to try to reconcile two conflicting attitudes. Their choice to value the Democratic Party above the LDS Church would influence future attitudes about the LDS Church and vice versa. Future conflicts likely would be resolved either by weakening their social attachment to the LDS Church or to the Democratic Party. Reconciliation was an untenable proposition. In this instance, LDS Church leaders did not provide any middle ground for those who wanted to support same-sex marriage and be obedient to their church leaders. Moreover, the official church position on same-sex marriage continues to create ambivalence for active LDS Democrats. One of the most consistent predictors of support for same-sex marriage is the belief that homosexuality is genetically determined.12 That is, people who believe that LGBT people are born with the predisposition for attraction to people of the same gender are significantly more supportive of same-sex marriage than those who do not. This might be one reason some people are reluctant to state that LGBT people are not born with a genetic proclivity to be attracted to people of the same gender. If so, we might expect that LDS Democrats who adopt the position of their party over that of their church rationalize that people born gay should not face discrimination. The ILF survey included an experiment to measure how those who identify with the LDS Church process information about the genetic heritability of various behaviors and reconcile the information with the teachings of their religious group. How do Democrats strongly identified with the LDS Church respond if presented factual information about the genetic heritability of homosexuality? The existing religion and politics literature suggests that they have one of two

How Political Identities Influence Religious Beliefs

119

responses. First, LDS Democrats might reject the information and adopt a dominant theological view that people are not born LGBT. Alternatively, they might choose a path of reconciliation. If so, I would expect respondents to the ILF survey to alter their expectations of the LDS Church, perhaps hoping the church will alter its position on the issue. If this is true, LDS Democrats would respond differently to information regarding the genetic heritability of samegender attraction than they do to information about the genetic heritability of other behaviors the LDS Church considers sinful. In fall 2014, the ILF randomly exposed participants to one of the following experiments. Survey participants were informed about the genetic heritability of three behaviors considered sinful by LDS Church leaders. Respondents were randomly assigned to view one of the statements in Table 8.2. The primary concern with the experiment was to find behaviors that are both genetically heritable and considered sinful by the LDS Church. In addition, having two behaviors both considered sinful by Mormons and are more strongly explained by genetic heritability than is same-sex attraction adds credence to the findings of the experiment. If LDS Democrats think differently about how the LDS Church should treat gays than church leaders treat those who are alcoholics or engage in compulsive gambling behavior, it would suggest that partisan social identity is influencing their judgments about LDS doctrine. The information included in each of the experimental conditions is factually accurate in the presentation of the findings. 13 The research into the genetic heritability of same-sex attraction has become politicized in recent decades, so the genetic heritability estimates are much larger in recent publications than in older publications.14 The estimate for heritability used in my experiment averages the genetic heritability differences between men and women. Because alcohol dependence and compulsive gambling have varying degrees of severity, the evidence presented in Table 8.2 is based on the average heritability of all forms of alcohol dependence and each kind of gambling addiction.15 After participants read the information in Table 8.2, they indicated how much they agreed with the following statements on a scale from 1 to 10, Table 8.2

Intermountain LDS Faith Experiment Text

Same-Sex Attraction

Scientists have shown that 30% of whether a person is attracted to people of the same gender is explained by the genes they inherit at birth.

Alcohol Dependence

Scientists have shown that 50% of whether a person becomes alcohol dependent is explained by the genes they inherit at birth.

Compulsive Gambling

Scientists have shown that 56% of whether a person engages in compulsive gambling as an adult is explained by the genes they inherit at birth.

120

Religious Identity in US Politics

with 1 representing “strongly disagree” and 10 representing “strongly agree.” The first statement examined whether participants thought genetic heritability should influence the LDS position: “Since people have little choice in the matter, the [LDS] Church should be more lenient with people who struggle with this problem.” The next statement tested respondent’s views about one of the behaviors considered sinful by church leaders: “The LDS Church should support efforts to allow [same-sex attraction], [alcohol consumption], or [gambling].” The final statement measured if participants thought LDS theology should change based on the scientific information they read: “[Same-sex attraction], [alcohol dependence], or [compulsive gambling addiction] cannot be a sin if people are predisposed to [same-sex attraction], [alcohol dependence], or [compulsive gambling addiction].” Overall, the results suggest that survey respondents did not agree with these statements. Table 8.3 shows that the mean value for each question is well below the halfway mark of the scale (5), indicating that most respondents disagreed that the LDS Church should change its positions. For most of those who participated in the survey, information about the genetic heritability of behaviors considered sinful by LDS leaders did not influence their evaluations of whether these behaviors actually are sinful. However, some participants did agree with the statements. For example, 6.65 percent of respondents to the compulsive gambling statements agreed that it is not sinful behavior because people are genetically predisposed to it. Likewise, 8.41 percent of respondents to the alcohol dependence statements agreed, and 18.68 percent of respondents to the same-sex attraction statements agreed. These differences are statistically significant (F = 5.078, p = 0.007) and suggest that participants responded differently to the different statements. Those responding

Data and Results

Table 8.3

Average Effect of Religious and Partisan Attitude Collision, ILF Experiment

Church should be more lenient Church should support efforts to allow the behavior Behavior is not a sin N

Alcohol Dependence 3.40 (2.63) 2.13 (2.34) 2.02 (2.08) 107

Same-Sex Attraction

Compulsive Gambling

2.95 (2.93) 107

2.10 (1.97) 105

4.01 (2.98) 3.07 (2.91)

3.25 (2.73) 2.04 (2.32)

Note: Entries are the average value on the 10-point scale for each of the experimental groups. The standard deviation is in parentheses. This is the average spread of the values; most of the values are within two standard deviations of the mean. N is the size of the sample in each treatment condition.

How Political Identities Influence Religious Beliefs

121

to the same-sex attraction statements were much more likely to agree that it is not a sin than those asked about the other two behaviors. What do we know about participants whose views do not immediately conform to those of the LDS majority? Recall that 14.4 percent of the respondents were former members of the LDS Church, and 4.7 percent described themselves as not at all active in the church. One simple explanation would be that the less active/former members of the LDS Church are the nonconformists. Luckily, this is an empirical question that can be easily answered with regression analysis. The ILF included several items that controlled for possible explanations of why respondents to the same-sex attraction statements answered differently than those reacting to the other two behaviors considered sinful by the church. LDS social identity was measured with the five-item scale created by David Campbell, John Green, and John Quin Monson, asking how proud the respondents were to be LDS and other similar questions.16 The interitem reliability on the LDS social identity measure in this survey was 0.84, a higher consistency than has been found in other samples measuring this concept. To assess how partisan and religious social identities influence responses to each statement the participants read, I included an interaction term between party identification, LDS religious social identity, and a variable indicating the experimental treatment the respondent received in a regression model. The full model results are displayed in Appendix 8. The results show little support for the idea that people’s level of LDS Church activity is causing some people to respond differently to the samesex attraction statements than to the other statements. Self-reported church activity was only a significant predictor of attitudes about whether the church should be more lenient with homosexuality, but not in the experiments regarding alcohol and gambling. Furthermore, former members of the LDS Church were just as likely as active members of the church to respond that the LDS Church should not be more lenient toward those who might be genetically predisposed to engaging in these behaviors. The experiment for each of the statements includes an interaction between party identification and the statement read. This allows for some postestimation modeling that explores how partisans responded to the statements in each of the experiments. The model-predicted agreement with each of the statements is displayed in the following figures, beginning with Figure 8.1. The pattern of findings is consistent with my overall argument in this chapter. In the alcohol dependence and compulsive gambling responses, partisanship did not have a significant influence on attitudes about leniency. Though Democrats were more likely to say that the LDS Church should be more lenient with people attracted to others of the same gender, the number of points was not significantly higher than that of the predicted Democrat responses about alcohol dependence or compulsive gambling. There was a

122

Religious Identity in US Politics

Figure 8.1

Leniency Attitudes, by Experimental Group and Partisanship

8.00

D emocrat Democrat

7.00

Independent Inde pendentt

Re publican Republican

6.00 5.00 4.00 3.00 2.00 1.00 0.00 Support Support

S Same-Sex ame-Sex Attraction A ttraction 5.71

A Alcohol lcohol Com Compulsive pulsive Dependence Gambling Dependence G ambling 3.25

3.65

S Same-Sex ame-Sex Attraction A ttraction 3.79

A Alcohol lcohol Com Compulsive pulsive Dependence Gambling D ependence G ambling 3.68

3.45

S Same-Sex ame-Sex Attraction A ttraction 3.64

A Alcohol lcohol Com Compulsive pulsive Dependence Gambling D ependence G ambling 3.22

3.06

Source: Intermountain LDS Faith Survey, 2015, wave I. Note: The height of the bars represents the model-predicted value on a ten-point scale; the lines represent the 95 percent confidence interval.

smaller predicted difference in the responses of Republican or Independent LDS Church members. However, when asked if the church should be more lenient with those who struggle with same-sex attraction, Democrats have attitudes two points higher than Republicans. Although this might seem like a small quantitative difference, the qualitative difference is substantial. On the ten-point scale, a person in the middle would have scored a 5, someone who “somewhat” agreed a 6, and someone who “somewhat” disagreed a 4. The model predicted that LDS Democrats would somewhat agree that the church should be more lenient with homosexuals and that Republicans would somewhat disagree. This is the only behavior regarding which Democrats were predicted to agree with the statement. Figure 8.2 shows that LDS Democrats were more likely to agree that the LDS Church should support people with same-sex attraction than were Republicans, but the difference is not statistically or substantively significant. In contrast, Figure 8.3 displays a significant partisan difference in attitudes about whether same-sex attraction is a sin. LDS Democrats who responded to the same-sex attraction statements were substantially more likely to agree that same-sex attraction is not sinful because people are born that way than were LDS Democrats who responded to the alcohol dependence or compulsive gambling statements. Recall that both compulsive gambling and alcohol dependence have an estimated higher genetic heritability than that of samesex attraction in virtually all peer-reviewed scientific publications. If LDS Democrats responding to the same-sex attraction statements had a different

How Political Identities Influence Religious Beliefs

123

viewpoint about whether same-sex attraction is sinful than do Independents or Republicans who responded to the same statements, it was not because of the scientific evidence they read. Rather, it was because identification as a Democrat motivated them to have different attitudes about same-sex attraction than does identification as a Republican or an Independent. LDS Democrat and Republican participants agreed about the sinfulness of the other two genetically heritable behaviors. In fact, all partisan groups were predicted to have roughly the same views regarding the sinfulness of compulsive gambling and alcohol dependence. They all agreed that notwithstanding the high level of genetic heritability of these behaviors, they are nonetheless sinful. This pattern of findings suggests that Democrats and Republicans have consistently different views about how the LDS Church should treat LGBT people despite their concordance on views about those addicted to gambling or alcohol. These analyses are compelling and suggest that when partisanship competes with religion, partisan identification can change attitudes about theology. The LDS Church has a hierarchical structure. Members of the church revere their leaders as prophets, with direct access to God, who guides these leaders in matters of policy and doctrine. That is, when the church leadership speaks, most members accept it as coming directly from God. At the time of the survey, the official LDS Church position on samesex attraction was that the attraction is not sinful, but acting on it is. Thus, responding that same-sex attraction “cannot be a sin” could be consistent Figure 8.2

Hierarchy Attitudes, by Experimental Group and Partisanship

Source: Intermountain LDS Faith Survey, 2015, wave I. Note: The height of the bars represents the model-predicted value on a ten-point scale; the lines represent the 95 percent confidence interval.

124 Figure 8.3

Religious Identity in US Politics Theological Views, by Experimental Group and Partisanship

Source: Intermountain LDS Faith Survey, 2015, wave I. Note: The height of the bars represents the model-predicted value on a ten-point scale; the lines represent the 95 percent confidence interval.

with LDS theology. Yet, the same could be said about alcohol and gambling. It is not sinful to be addicted to alcohol; it is sinful to consume alcohol. Likewise, it is no sin to have a compulsion to gamble; it is wrong to gamble. Yet, LDS Democrats who responded to the same-sex attraction statements were the only participants committed to the idea that the behavior is not sinful because when religious doctrine conflicted with partisan identity, it was less psychologically demanding to change LDS Church doctrine than a political issue position. Do respondents with strong LDS social identities respond differently when their preferred party takes a position that conflicts with their religion? A second set of analyses included a three-way interaction between the experiments, religious identities, and partisan identities to measure how those with strong LDS social identities responded to each of the statements.17 The test of statistical significance illustrated in Figure 8.4 compares predicted support for each statement and estimates their difference. The X-axis moves from no social identification with the LDS Church to strong religious social identification with it. Former members of the LDS Church and those who reported being very active members of the church are found along the LDS social identity scale. If religious social identity is a strong predictor of attitudes about matters of LDS Church policy and doctrine, I would expect to see a strong positive or negative slope in the lines representing model-pre-

LDS Social Identity

How Political Identities Influence Religious Beliefs

125

dicted support. For that difference to be statistically significant, the 95 percent confidence interval of the estimated support from some groups could not overlap with the confidence intervals of other groups. For example, Figure 8.4 shows that the predicted support for the idea that the LDS Church should be more lenient on those who are attracted to people of the same gender decreases as LDS social identity increases. Conversely, support for church lenience on those predisposed toward compulsive gambling increases as LDS social identity increases among Independents. However, because the confidence intervals around those estimates overlap this difference is not statistically significant, which is not surprising given the sample size. As such, I will not make a strong conclusion based on these findings, but there is a clear pattern. Those with a stronger LDS identity are less likely to state that their church should be more lenient with those who do not fit the doctrinal mold, and those with a weaker LDS identity are much more likely to believe that the LDS Church should be more lenient. LDS Republicans who responded to the same-sex attraction statements were predicted to have the same level of support for church leniency (very little) regardless of behavior under consideration. LDS Republicans who felt socially disconnected from the church were just as likely to disagree that the LDS Church should be more lenient as Republicans with strong LDS social identities. The same was true of LDS Democrats. As discussed previously, Democrats were more likely to think that the church should be Figure 8.4 Test of Statistical Significance: LDS Church Should Be More Lenient 10

A Alcohol lcohol D Dependence ependence

9

Compulsive Gambling Com pulsive G ambling

8

Same-Sex Attraction S ame-Sex A ttraction

M Model odel P Prediction rediction

7 6 5

A

4 3 2 1 0

Democrats D emocrats 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Re Republicans publicans

Inde pendent ns Independents 1

0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

S Strength trength of LDS LDS Ide Identity ntity

0.3 0.4 0.5 0.6 0.7 0.8 0.9

Source: Intermountain LDS Faith Survey, 2015, wave I. Note: The sloping lines represent the model-predicted support for each statement.

1

126

Religious Identity in US Politics

more lenient on people attracted to people of the same gender than they were to agree with leniency for those predisposed to alcoholism or compulsive gambling, but the strength of their LDS identity did not change their level of support. In short, LDS Church members’ attitudes about whether their church should be more lenient on people predisposed to behaviors considered sinful by their church were more strongly influenced by their partisanship than by their religious social identification. The predicted support for the idea that these behaviors are not sinful shows a nearly identical pattern of findings. As discussed previously, partisan identification is a powerful predictor of LDS attitudes concerning the sinfulness of these behaviors, and religious social identification does nothing to alter that relationship. Republicans and Independents with virtually no social identification with the LDS Church were statistically just as likely to think that each of these behaviors are sinful as were those with strong social identification with the LDS Church. Among LDS Democrats, those with strong LDS social identification were less likely to agree that same-sex attraction is not sinful than were those with weak LDS social identification; though the difference is not statistically significant. Compared with Republicans with strong LDS identities, Democrats with strong LDS identities were four times more likely to believe that same-sex attraction is not a sin. LDS Democrats were predicted to be more supportive of the idea that same-sex attraction is not sinful than were LDS Democrats surveyed about alcoholism and compulsive gambling. In fact, the slope of the line regarding compulsive gambling is weakly positive, and the predicted support for that behavior not being sinful is nearly the same as the support for samesex attraction not being a sin. Taken together, these two findings suggest that one way LDS Democrats with strong LDS social identities reconcile their views that same-sex attraction is not sinful might be to begin believing that other genetically heritable behaviors are also not sinful. LDS Democrats with a weak religious identity did not have competing social identities pulling them in different directions, so it caused less psychological discomfort to state that compulsive gambling—though genetically heritable—is sinful. Again, these findings were not statistically significant, so I avoid making too strong a causal claim, but the predicted patterns illustrated in Figure 8.5 are exactly what I would expect if the theory of competing social identities is correct. When one social identity pulls views in one direction and another identity pulls them in the other, the stronger social identity has the dominant impact on subsequent attitudes. Why do LDS Republicans and Democrats have different views about the sinfulness of some behaviors? Well, in this case, we can be certain that both

Conclusion

How Political Identities Influence Religious Beliefs

127

Figure 8.5 Partisanship Determines Whether Behavior Is Sinful

Source: Intermountain LDS Faith Survey, 2015, wave I. Note: The sloping lines represent the model-predicted support for each statement.

Republicans and Democrats were presented exactly the same information about the genetic heritability of same-sex attraction. In addition, the information was based on the highest estimate of genetic heritability ever published in a scientific journal. It is unlikely that the participants saw the number in the experiment and decided that the information was inaccurate based on knowledge of the scientific literature. I can also be certain that Democrats and Republicans saw the same information about the genetic heritability of alcohol dependence and compulsive gambling. When asked how the LDS Church should deal with these behaviors, Democrats had identical views as Republicans. Why did LDS Democrats have different views on LDS theology concerning those attracted to people of the same gender than did LDS Republicans? The difference was not because of a difference in the level of genetic heritability. Both alcohol dependence and compulsive gambling are more genetically heritable than same-sex attraction, and participants were made aware of this. Finally, I can rule out a difference in LDS theology. Although people might argue that compulsive gambling is not considered as terrible a sin as homosexuality in LDS theology, alcohol dependence is at least as sinful a behavior as same-sex attraction. Indeed, recent teachings about same-sex attraction suggest that it is not a sin to be attracted to someone of the same gender but only to act on that disposition. Although this might explain some differences in viewing the behavior as sinful, it cannot explain why

128

Religious Identity in US Politics

Democrats were significantly more supportive of leniency for those attracted to people of the same sex than for those with alcohol dependence. The most logical, plausible explanation for these findings is that when people affiliate with a political party that takes a position contrary to that of their own faith leaders, it creates ambivalence and forces them to choose between competing social identities. As displayed in Figure 8.2, sometimes people place their church above their party. When asked if the church should do more to support LGBT people, Democrats are statistically equivalent to Republicans. Sometimes they attempt to reconcile the disparate views; this is apparent in Figure 8.3. Because some LDS Church leaders have made public statements arguing that same-sex attraction is not a sin, LDS Democrats can place themselves somewhere between the position of their party and the majority view of the church. However, as shown in Figure 8.1, sometimes people abandon one identity for another. LDS Democrats and Republicans disagree about how lenient the church should be about same-sex attraction. Yet sufficient theological support can be found for an argument of more leniency toward those who struggle that Democrats were willing to argue against the church on this issue. It is one thing for people to change their views to conform to the majority view of a political party; it is another to expect their churches to do so. Yet this is precisely what happens when political and religious social identities compete. 1. Leon Festinger, “A Theory of Social Comparison Processes,” Human Relations 7, no. 2 (1954): 117–140. 2. Ziva Kunda, “The Case for Motivated Reasoning,” Psychological Bulletin 108, no. 3 (1990): 480. 3. John C. Turner, Rupert J. Brown, and Henri Tajfel, “Social Comparison and Group Interest in Ingroup Favouritism,” European Journal of Social Psychology 9, no. 2 (1979): 187–204. 4. Donald P. Green, Bradley Palmquist, and Eric Schickler, Partisan Hearts and Minds: Political Parties and the Social Identities of Voters, Yale ISPS Series (New Haven, CT: Yale University Press, 2002). 5. Brian J. Gaines et al., “Same Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq,” Journal of Politics 69, no. 4 (2007): 957–974. 6. John T. Jost and Diana Burgess, “Attitudinal Ambivalence and the Conflict Between Group and System Justification Motives in Low-Status Groups,” Personality and Social Psychology Bulletin 26, no. 3 (2000): 293–305. 7. Alice H. Eagly and Shelly Chaiken, The Psychology of Attitudes (New York: Harcourt Brace, 1993). 8. James N. Druckman, Erik Peterson, and Rune Slothuus, “How Elite Partisan Polarization Affects Public Opinion Formation,” American Political Science Review 107, no. 1 (2013): 57–79; Gabriel S. Lenz, Follow the Leader? How Voters Respond to Politicians’ Policies and Performance (Chicago: University of Chicago Press, 2013). 9. Paul Djupe and Brian Calfano, God Talk: Experimenting with the Religious Causes of Public Opinion (Philadelphia, PA: Temple University Press, 2014).

Notes

How Political Identities Influence Religious Beliefs

129

10. Patrick J. Egan and Kenneth Sherrill, “California’s Proposition 8: What Happened, and What Does the Future Hold?” (San Francisco: National Gay and Lesbian Task Force Policy Institute, 2009). 11. Maykel Verkuyten, The Social Psychology of Ethnic Identity (London: Psychology Press, 2004); Verkuyten and Ali Aslan Yildiz, “National (Dis)Identification and Ethnic and Religious Identity: A Study Among Turkish-Dutch Muslims,” Personality and Social Psychology Bulletin 33, no. 10 (2007): 1448–1462. 12. Donald P. Haider-Markel and Mark R. Joslyn, “Beliefs About the Origins of Homosexuality and Support for Gay Rights: An Empirical Test of Attribution Theory,” Public Opinion Quarterly 72, no. 2 (2008): 291–310; Alison Avery et al., “America’s Changing Attitudes Toward Homosexuality, Civil Unions, and SameGender Marriage: 1977–2004,” Social Work 52, no. 1 (2007): 71–79. 13. For example, most evidence suggests that genetic heritability of homosexuality is about 45 percent genetic for men and 55 percent non-home-environmental experiences in life, but for women only 9 percent of homosexuality is explained by genes and 91 percent is explained by both home- and non-home-environmental experiences in life. See J. Michael Bailey, Michael P. Dunne, and Nicholas G. Martin, “Genetic and Environmental Influences on Sexual Orientation and Its Correlates in an Australian Twin Sample,” Journal of Personality and Social Psychology 78, no. 3 (2000): 524. However, examining the genetic heritability of a number of same-gender sexual partners yields a higher estimate for genetic heritability. Niklas Långström et al., “Genetic and Environmental Effects on Same-Sex Sexual Behavior: A Population Study of Twins in Sweden,” Archives of Sexual Behavior 39, no. 1 (2010): 75–80, find that having a same-gender sexual partner is 39 percent genetically heritable for men and 19 percent for women. 14. Thomas Schofield gives an example of this from developmental psychology in “Knowing What We Don’t Know: A Meta-Analysis of Children Raised by Gay or Lesbian Parents,” The Winnower 3 (2016), https://lib.dr.iastate.edu/hdfs_pubs/53/. 15. Wendy S. Slutske et al., “Common Genetic Vulnerability for Pathological Gambling and Alcohol Dependence in Men,” Archives of General Psychiatry 57, no. 7 (2000): 666–673. 16. David E. Campbell, John C. Green, and J. Quin Monson, Seeking the Promised Land: Mormons and American Politics (Cambridge, UK: Cambridge University Press, 2014). 17. Full results are in Appendix 8.

9 Is Reconciliation Possible?

In September 2016, the US Commission on Civil Rights issued its report Peaceful Coexistence: Reconciling Non-Discrimination Principles with Civil Liberties. The chair of the commission, Martin Castro, wrote,

Religious freedom has become a wedge issue in US politics.

The phrases “religious liberty” and “religious freedom” will stand for nothing except hypocrisy so long as they remain code words for discrimination, intolerance, racism, sexism, homophobia, Islamophobia, Christian supremacy, or any form of intolerance. Religious liberty was never intended to give one religion dominion over other religions, or a veto power over the civil rights and civil liberties of others. However, today, as in the past, religion is being used as both a weapon and a shield by those seeking to deny others equality.1

A coalition of religious leaders from Jewish, Islamic, and Christian faiths were appalled by these statements in the report and issued an open letter asking President Barack Obama and congressional leaders to denounce this view of religious freedom. They argued that “no American citizen or institution [should] be labeled by their government as bigoted because of their religious views and dismissed from the political life of our nation for holding those views. And yet that is precisely what the Civil Rights Commission report does.”2 Without a doubt, religion has been used as both a weapon to destroy liberty and as a shield to protect individuals from the full consequences of their negative behavior. It is also true that religious beliefs have been a motivating influence, expanding and protecting human life and liberty. Given the powerful influence religion has had on the political world, it is 131

132

Religious Identity in US Politics

remarkable that scholars know so little about how religious identities are developed and maintained and influence attitudes. This book is an attempt to provide some conceptual clarity that can guide future scholarship and lead to a deeper understanding of how religious identity can engender both positive and negative outcomes. On the one hand, Elizabeth Shakman Hurd argues that “neither religions nor religious actors are singular, agentive forces that can be analyzed or quantified.” Describing the complexity of daily human religious experience, she finds that few people practice a religion identically. People do not practice religion in a binary good/evil, faith/reason manner; rather, they display considerable nuance as they apply their beliefs to religious practice. Hurd concludes that legal protection of religion requires political actors to define and delineate legal boundaries of religion that result in religious categories that do not reflect the actual religions to which adherents belong.3 Many other scholars combine religious belief, practice, and affiliation into a single measure or use the three interchangeably. This creates even more confusion when their empirical findings reduce the complexity of religious experience to a binary religious/not religious continuum. For example, Pippa Norris and Ronald Inglehart measure secularization using an index that combines church attendance; prayer; religious identity; and belief in heaven, hell, and an afterlife.4 Although it is not surprising that these items yield a statistically reliable measure of religiosity or secularism, the consistent mixing and matching of these measures to test hypotheses has resulted in great conceptual confusion.5 Religion and politics scholars have accumulated more empirical evidence than anyone a century ago could have imagined possible about how religion and politics influence each other, but there are considerable gaps in understanding the processes through which religion or the lack thereof motivates human behavior.6 Acknowledging that people will always adapt religious belief to fit their individual needs and that no two people will practice their religion identically, I argue that religion represents a meaningful category that can be quantified and studied not because the law requires it to be but because humans categorize themselves. Systems of government create both formal and informal institutional arrangements that can have a powerful influence on how religion is practiced in society. Certainly, these arrangements influence how individuals practice and develop religious beliefs, attitudes, and behaviors. However, this does not negate the importance of treating religion as a meaningful category for scientific exploration. Rather, it suggests that future research should help to specify the manner through which institutional arrangements influence the development of religious networks, beliefs, behaviors, and identities. People join religious communities—just as they do any other group in society—because belonging to a group gives people self-esteem, status,

Is Reconciliation Possible?

133

meaning, guidance, and social interaction with others. They choose to belong to a particular religious persuasion (or none) because that group offers them something they cannot get in another group. In time, people begin to view the group as an extension of themselves. For some, their religious identity might be their dominant social identity; for others, it might be a weak identity. When something happens to the religious group, the predicted response depends on how strongly people identify with the group.7 When a religious group is attacked, someone with a strong identification with that religious group will perceive that as a personal attack. Someone with a weaker religious identity will view the same event differently. In this way, the quantifiable boundaries of religion are not defined by legal entities but by individuals who voluntarily choose how closely to merge their own sense of self with a religion (or none at all). My first task of this book, accomplished in Chapter 1, is to explain how religious social identity functions in congruence with other social identifications. In Chapter 2, I explain why it makes theoretical sense to think of religion as a social identity and generated some observable predictions of this conceptualization. The first prediction is that people who identify with a religion have more positive views of others who share their religious affiliation. In Chapters 3 and 4 I describe the findings of two nationally representative surveys fielded in 2015 and 2016 that tested this hypothesis. In these chapters I show that shared religious identity between elected officials and those they represent, distinct from partisanship, engenders trust and support in those government officials. As the political parties in the United States become more ideologically distinct and there is less common ground on salient political issues, religion can potentially bridge the divide. Members of Congress who descriptively represent their constituents have more latitude in policy negotiations because the source of constituent approval is not policy dependent. Americans who do not pay much attention to the particulars of policy debates are more supportive of House members who share their religious affiliations because they have higher trust in these representatives. Similarly, people are more trusting of a president with whom they share a religious identity. In addition, those who identify with the opposing political party extend this trust. In summer 2015, it would be have been hard to imagine President Obama doing anything that would have pleased the average Republican. Yet Republicans who perceived a shared religious identity with the President were 500 percent more trusting of him than Republicans who did not. As politics in the United States polarizes, religion has the potential to unite disparate political groups. Another expectation is that people have negative responses to those who do not belong to their religious groups. Evidence from national surveys conducted in 2013 and 2014 suggests that the divide between religious

134

Religious Identity in US Politics

and nonreligious Americans is rooted in political competition for resources. Although willing to help people generally, religious Americans are less likely to help a nonbelieving neighbor with a variety of political and nonpolitical tasks. The nonreligious are equally biased against their religious acquaintances. A simple explanation is that partisan political conflict is spilling over and creating social conflict. Meanwhile, the most religious Americans are more willing to help an atheist than they are to help a Democrat or a Republican. Current party alignments cause some religious Americans to see nonbelievers as both a religious and a political out-group, which often motivates them to express negative attitudes about nonbelievers. Nonbelievers who identify as Democrats express similar antipathy toward Christians. However, religious Americans who do not identify with the Republican Party do not express the same level of antipathy toward nonbelievers in the United States. This is encouraging because it suggests that the divide between religious and nonreligious Americans can be more easily bridged as party coalitions change. Finally, in this book I tested the expectations of the competing identities hypothesis in a couple of survey experiments. When political and religious social identities pull people in opposite directions, the effects are heterogeneous.8 The response of hypothetical Catholic, working-class Germans to Democrats or to Republicans depends on how strongly they identify with the Catholic Church relative to their other social identities. When party and religion collide, those with strong partisan identities put their religion aside and adopt the view of the partisan majority. In Chapter 7 I demonstrate that these people become weaker religious identifiers and stronger partisan identifiers. Moreover, my evidence in Chapter 8 suggests that they even think their religious groups should alter doctrinal positions to align with those taken by their political parties. This could be another reason why people are leaving evangelical churches: these churches may advocate doctrine that incorporates Republican messaging and makes weaker partisans uncomfortable.9 In contrast, weaker partisans with strong religious identities disidentify with their political parties and more strongly identify with their religious groups. Within their religious communities, they advocate for more stringent doctrinal purity on issues that have become political and think people with weaker religious views should be expelled from the religious group. Not only does this perspective help explain why terrible things have been done in the name of religion but also it explains less extreme, yet confusing, political outcomes. For instance, many are surprised at the level of support Donald Trump receives from religious conservatives, who typically support candidates who reflect their religious values. Geoffrey Layman noted that church attendance was inversely associated with Trump support when the Republican primaries began in January 2016. Though Trump was still the most preferred candidate among evangelical Christians who

Is Reconciliation Possible?

135

attended church weekly (35 percent supported him), a majority of those who rarely attended church (53 percent) supported Trump.10 How could religious Americans support a presidential candidate who did not share their religious views or values? Religious identities were simply weaker for some members of the Republican base than they were for others. Those with a strong partisan identity followed their party rather than their religion. This perspective also explains why some people are abandoning the political parties. Scholars have noticed the steady increase in the number of people disidentifying with political parties in the United States. Russell Dalton argues this trend is occurring because the public is becoming more educated and no longer needs the services traditional mass parties once provided. Cognitively sophisticated voters can recruit and select candidates who represent their interests without political parties, so these voters no longer participate in or identify with political parties.11 Samara Klar and Yanna Krupnikov argue that people are leaving parties because the labels associated with the dominant parties in the United States embarrass them.12 An equally plausible explanation is that when political parties advocate positions inconsistent with the dominant views of religious groups, those with strong religious social identities disidentify with their parties and become stronger religious identifiers. The Pew Forum on Religion and Public Life conducted comprehensive surveys of the US population in 2007 and again in 2014. With more than thirty-five thousand participants, these surveys provide the most complete picture of how religion and politics influence the attitudes of Americans. Figures 9.1 and 9.2 show how religious and political identity combined in the United States in 2007 and 2014, respectively. In 2007, only 15 percent of those for whom religion was very important identified as political independents. In fact, these Americans were just as likely to be found in the Democratic Party (17.8 percent) as in the Republican Party (18.9 percent). In 2007, the sharpest religious/political identification divide was found among those with very weak religious identities. Those for whom religion was not at all important were most likely to identify as political independents (2.97 percent), whereas few identified as Republican (0.9 percent). The Republican Party of 2007 was best described not as the party of the religious but as the party the nonreligious avoided. This distinction is important. The distribution of partisanship among religious Americans in 2007 suggests that those with weak and strong religious identities felt equally at home in the Democratic Party but that Americans with a weak religious identity were much less likely to identify as Republican. By 2014, much of this changed. As the number of religious Independents in the United States increased, they chose to align themselves with groups other than the Republican Party. Political Independents showed an increase of 2 percentage points among those for whom religion was not at all important and a 1.3 percentage point increase among those for whom

136

Figure 9.1 20.00%

Party Identification, by Strength of Religious Identity, Pew 2007 18.94% 17.80%

18.00% 14.99%

16.00% 14.00% 12.00% 10.00% 8.00%

9.01%

8.24% 6.10%

6.00% 2.90% 2.41%

4.00% 2.00%

1.58%

3.60%

2.97%

3.24% 1.18%

0.90%

0.40% 0.41%

0.00% Republican Very V ery e Important

Democrat Somewhat Important

Independent Not too Important

No Preference Preference Not at all Important

Source: Data from Luis Lugo et al., 2007 U.S. Religious Landscape Study (Washington, DC: Pew Research Center, 2008). Note: Bars represent the percentage of the total sample in each category.

Figure 9.2 18.00%

Party Identification, by Strength of Religious Identity, Pew 2014 16.79%

16.68% 15.11% 15.11%

16.00% 14.00% 12.00%

9.65% 10.00% 7.07%

8.00% 6.00%

5.22% 5 22 5.22 3.62% 3

4.28%

4.94% 4 94%5.10%

4.00% 2.00%

2.74% 1.55% 1 1.07%

1.00%0.44%0.51%

0.00% Republican Very V eery Important

Democrat Somewhat Important

Independent Not too Important

No preference preference Not at all Important

Source: Data from Alan Cooperman et al., 2014 U.S. Religious Landscape Study (Washington, DC: Pew Research Center, 2015). Note: Bars represent the percentage of the total sample in each category.

137

138

Religious Identity in US Politics

religion was not too important. The number of Democrats among the nonreligious also grew during this seven-year span. Yet religious Americans also shifted their partisan loyalties. Independents gained a .5 percentage point among those for whom religion was somewhat important and roughly 2 percentage points among those for whom religion was very important, whereas the number Republicans who said religion was very important decreased by 2 percentage points. Democrats saw a similar decline in the number of very religious Americans who identified with their party. Though the percentages might seem small, the pattern is clear. By 2014, religious Americans were more likely to identify with no political party than they were to identify as Republican or Democrat. Why are religious Americans leaving the dominant political parties? As moral issues have come to the forefront of the political debate in the United States, religious identities and partisan identities collide with greater frequency. Put simply, it is harder to be a consistent Republican or Democrat and simultaneously maintain religious views consistent with any religious denomination in the United States. Americans for whom religion is a dominant identity find it easier to disengage with either party than to try to accommodate political positions incongruent with their religious identities. It is noteworthy that this pattern predates the rise of Trump in the Republican Party. For decades, the Republican Party has championed conservative positions on salient moral issues. Exploiting the traditionalist-modernist divide in the United States allowed the GOP to move away from perpetual minority status to a dominant political force over the past several decades, but it seems to have come at a cost.13 As traditionally religious activists in the GOP and secular activists in the Democratic Party push their parties and candidates toward more polarized stands on moral and cultural issues, the most religious Americans have begun disassociating themselves from the two parties. Geoffrey Layman and Christopher Weaver document the change in religious affiliation among delegates to the Democrat and Republican National Conventions from 1988 to 2012. In 1980, 40 percent of the delegates to the Republican convention were mainline Protestants; by 2012 only 18 percent of the delegates were mainline Protestant. The dominant religious affiliation at both the Republican and Democratic National Conventions in 2012 was none. Of the delegates to both conventions, 30 percent did not affiliate with any religion.14 Contemporary party activists are those with strong partisan and weaker religious identities who behave consistently with my theoretical expectations. The demise of religious identification within these two political parties has not reduced the influence of religious beliefs on political positions. In fact, losing strong religious identifiers has exacerbated the problem. What happens when religious identifiers challenge the moral value of political parties because the policies they support are inconsistent with religious doctrine?

Is Reconciliation Possible?

139

The theory of competing social identities predicts that those with stronger partisan than religious identities disidentify with their religions, become stronger partisans, and enhance the distinction between religious and political groups by advocating for greater purity in the ranks, which is precisely what Layman and Weaver find.15 They show that strong secularists and strong religious believers are more ideologically extreme, less willing to compromise, and more supportive of extreme policy positions than the average delegate to a national convention is. In other words, the activists who have remained in the political parties two decades into the culture wars are weaker religious identifiers but strongly committed to their beliefs. Beliefs actually play a stronger role in partisan politics, because those who remain loyal to the party are more dogmatic and less willing to compromise. Finally, in this book I present strong evidence that the God gap in US politics is rooted in religious and political identities. At present, the Democratic and Republican Party coalitions exacerbate this divide because those with strong, non-Christian religious identities feel most at home in the Democratic Party. This reduces opportunities for politically minded Christians to have positive interactions with Americans with whom they share divergent religious identities. Yet positive social contact between opposing social groups is one of the best ways to reduce out-group antipathy. My empirical findings suggest that one way to mitigate the negativity between believing and nonbelieving Americans is to facilitate opportunities for them to work together to solve problems in their communities. There are many areas in which believers and nonbelievers can agree there are problems, and if they were to collaborate on positive solutions, the God gap would diminish. All of this evidence demonstrates the power of conceptual clarity in future studies of religion and politics. Future scholarship should eschew statistically reliable measures of religiosity in favor of theoretically grounded concepts. Rather than conclude that religion is immeasurable—as others have—I suggest that most productive scholarship incorporate insights and concepts from established research programs and apply them to the study of religion.

1. Martin R. Castro et al., Peaceful Coexistence: Reconciling Non-Discrimination Principles with Civil Liberties (Washington, DC: US Commission on Civil Rights, 2016), 306. 2. William E. Lori et al. to Barack Obama, Orrin Hatch, and Paul Ryan, letter, October 7, 2016, “Letter Asks Obama, Congress to Disavow Bias Claim in Commission Report,” October 12, 2016, Catholicnews.Com, https://www.catholicnews .com/services/englishnews/2016/letter-asks-obama-congress-to-disavow-bias-claim -in-commission-report.cfm. 3. Elizabeth Shakman Hurd, Beyond Religious Freedom: The New Global Politics of Religion (Princeton, NJ: Princeton University Press, 2015).

Notes

140

Religious Identity in US Politics

4. Pippa Norris and Ronald Inglehart, Sacred and Secular: Religion and Politics Worldwide, 2nd ed. (Cambridge, UK: Cambridge University Press, 2011). 5. Paul Goren and Christopher Chapp, “Moral Power: How Public Opinion on Culture War Issues Shapes Partisan Predispositions and Religious Orientations,” American Political Science Review 111, no. 1 (2017): 110–128; Amanda Friesen and Aleksander Ksiazkiewicz, “Do Political Attitudes and Religiosity Share a Genetic Path?” Political Behavior 37, no. 4 (2015): 791–818; Christopher Weber and Matthew Thornton, “Courting Christians: How Political Candidates Prime Religious Considerations in Campaign Ads,” Journal of Politics 74, no. 2 (2012): 400– 413; Pazit Ben-Nun Bloom, Gizem Arikan, and Marie Courtemanche, “Religious Social Identity, Religious Belief, and Anti-Immigration Sentiment,” American Political Science Review 109, no. 2 (2015): 203–221; Bethany L. Albertson, “Dog-Whistle Politics: Multivocal Communication and Religious Appeals,” Political Behavior 37, no. 1 (2015): 3–26. 6. For more on this, see Geoffrey C. Layman and Christopher L. Weaver, “Religion and Secularism Among American Party Activists,” Politics and Religion 9, no. 2 (2016): 271–295. 7. Sonia Roccas and Marilynn B. Brewer, “Social Identity Complexity,” Personality and Social Psychology Review 6, no. 2 (2002): 88–106; Catherine E. Amiot and Roxane de la Sablonniere, “Facilitating the Development and Integration of Multiple Social Identities,” in The Psychology of Social and Cultural Diversity, ed. Richard J. Crisp (Hoboken, NJ: John Wiley and Sons, 2011). 8. Seymour Martin Lipset, Political Man: The Social Bases of Politics (Garden City, NY: Doubleday, 1960). 9. Paul A. Djupe, Jacob R. Neiheisel, and Anand E. Sokhey, “Reconsidering the Role of Politics in Leaving Religion: The Importance of Affiliation,” American Journal of Political Science 62, no. 1 (2018): 161–175. 10. Geoffrey C. Layman, “Where Is Trump’s Evangelical Base? Not in Church,” Washington Post, March 29, 2016. 11. Russell J. Dalton, The Apartisan American: Dealignment and Changing Electoral Politics (Washington, DC: Congressional Quarterly Press, 2013). 12. Samara Klar and Yanna Krupnikov, Independent Politics: How American Disdain for Parties Leads to Political Inaction (New York: Cambridge University Press, 2016). 13. Geoffrey C. Layman, The Great Divide: Religious and Cultural Conflict in American Party Politics (New York: Columbia University Press, 2001). 14. Layman and Weaver, “Religion and Secularism Among American Party Activists.” 15. Layman and Weaver, “Religion and Secularism Among American Party Activists.”

Appendix 1: A Note on Interviews Conducted

with everyone I could find who identifies with a low-status religion or as religiously unaffiliated. While in Northern California, I took advantage of the vibrant Buddhist communities that exist in the area and began contacting people at Buddhist temples. All of the Buddhists I contacted came from this outreach. I also contacted a few Jehovah’s Witnesses out proselyting, but I was not able to complete an in-depth interview with any of them. Next, I used Facebook to identify groups that specifically target members of religious communities. I became a member of the groups American Muslim, North American Hindu Religious Advisory Community, and Stories of New Muslims. The moderators of these groups allowed me to post a request to interview group members for my project. I posted the interview questionnaire to the group pages and sent personal messages to many members of these groups requesting an interview. I was never able to contact anyone associated with these groups. I also sent a request to the moderator of a group called Sacramento Freethinkers, Atheists, and Nonbelievers. This person was very helpful and put me in contact with most of the nonreligious people I interviewed for this project. Next, I went to a database I maintain with 300 million email addresses. I created a subset of email addresses that belonged to people with the most popular Indian first names and surnames. I created another subset of email addresses that belonged to people with the first name Mohammad. I combined those lists, randomized them, and sent a request via email to these individuals. The interviews I conducted with Muslims and Hindus came from that combined list. The members of the Church of Jesus Christ of Latter-day Saints

I spent the month of August 2017 conducting interviews

141

142

Appendix 1

(LDS, or Mormon Church) interviewed for this project are my acquaintances. I selected them because I knew something about changes that occurred in their lives as adults, but I did not know their complete stories in advance. Table A1.1 provides a brief biographical sketch of each person who participated in the interviews. All the names are pseudonyms to protect the respondents’ identities, but complete audio recordings and written transcriptions of the interviews have been kept on file for data quality verification. Table A1.1

Demographic Summary of Interviewees

Name

Age

Vincent Tyler Marie Sterling

60 20s

Sarah Oliver

50

Susie Brooks

50

Craig Williams

50

Blake Vella

Venkata Bagchi Amar Joshi Murad Sena Zhen Wuhan

Martin Graham Paul

20s 30s

20s 50 40 20 40

Residence

Northern California Virginia Northern California Northern California Central California Northern California Northern California Missouri New Hampshire Northern California Idaho Washington, DC

Childhood Religion

Current Religion

Interview Date

Mennonite Jehovah’s Witness Lutheran

Zen Buddhist Secularist

8/3/17 8/8/17

Lutheran

Catholic

Jesuit Catholic

Hindu

Hindu Muslim Zen Buddhist

Zen Buddhist

Spiritual but not religious Atheist Agnostic/ atheist Atheist

Hindu Muslim Zen Buddhist

8/3/17

8/8/17

8/8/17

8/8/17

8/15/17

8/17/17 8/21/17 8/23/17

Nonbeliever LDS (Mormon) 8/25/17 LDS (Mormon) N/A 6/7/15

Appendix 2: The National Surveys

A US sample of 1,596 subjects was recruited by Research Now to participate in a study measuring political attitudes. The study was fielded from June 18 to June 21, 2013. Participants completed an extensive member profile survey before they were selected to participate in this particular research project. In addition, Research Now employed a digital fingerprinting technology that prevented more than one person from completing a survey from the same computer, and they used Geo-IP validation to ensure that the computer used matched the geographical location of the survey respondent. Participants were recruited using the e-Rewards, invitation-only methodology. Those who did not receive an invitation to participate via email were blocked from participation in the survey. The American Association for Public Opinion Research (AAPOR) response rate to this study was 11.87 percent.

Survey I

A US sample of 993 subjects was recruited by Research Now to participate in a study measuring political attitudes. The study was fielded from June 23 to June 24, 2014. Participants were recruited using the e-Rewards, invitation-only methodology. Those who did not receive an invitation to participate via email were blocked from participation in the survey. Research Now sent out 9,110 invitations through social media and email; 1,189 people began the survey (13.1 percent response rate), 59 opted out of the survey after reading the

Survey II

143

144

Appendix 2

disclosure statement (4.96 percent), and 993 completed the entire survey (83.5 percent). A US sample of 1,939 subjects was recruited by Clear Voice Research to participate in a national political study from June 15 to June 25, 2015. Clear Voice maintained an online panel used solely for research purposes. Participants in the panel were told they would be invited to participate in online research surveys in exchange for various incentives. The initial registration form collected basic fields, including name, email address, postal address, gender, date of birth, and language. After they completed this form, a double opt-in/confirmation email was sent to their email addresses. Only those with double opt-in/confirmed accounts were invited to participate in surveys. Following opting in, panelists were asked to complete their profile so that Clear Voice collected as many data points as possible, which increased their targeting abilities when they sent the member survey invitations. Based on my specifications, a sample was pulled in quota group formats. Simple randomization was used for a representative sample of new and old members within the quota groups. Participants were invited via email to participate in the survey. Clear Voice sent out 51,492 invitations; 2,488 began the survey (4.8 percent response rate), and 1,939 (77.9 percent) completed the entire survey.

Survey III

A US sample of 1,290 subjects was recruited by Clear Voice Research to participate in a national political study from June 18 to June 28, 2016. Please see Survey III for the methodology. For this survey, Clear Voice sent out 58,481 invitations; 1,639 began the survey (2.8 percent response rate), and 1,290 (78.7 percent) completed the entire survey. The demographic characteristics of these panels closely resembled that of the US population in several important traits. Table A2.1 displays the demographics of these samples compared with Amazon’s Mechanical Turk samples, the American Community Survey (ACS) census estimates, and the National Annenberg Election Survey (NAES).1 Mechanical Turk is an online marketplace where people hire laborers for a variety of tasks. Since the mid-2000s, researchers have been offering people money to participate in online survey experiments through Mechanical Turk. Recently, scholars have spent considerable effort trying to determine the quality of the samples obtained through this Amazon service.2 Table A2.1 shows that the Clear Voice Research sample is more representative of the US population on key variables than samples obtained through Mechanical Turk.

Survey IV

Table A2.1

Survey Demographics

Demographics

Female Age (mean years)

Education (% completing some college) White Black Asian Latino/a Multiracial Party identification Democrat Independent Republican N

June 2016 Survey

June 2015 Survey

57.71%

60.31%

53.14% 50.86

75.29% 12.39% 4.26% 5.03% 1.78% 36.95% 36.33% 26.72% 1,291

June 2014 Survey

June 2013 Survey

32.23%

31.43%

49.23% 50

51.56% 46

80.61% 9.13% 3.2% 4.07% 2.27%

73.21% 7.05% 7.65% 7.75% 3.02%

33.75% 41.49% 24.77% 1,939

32.02% 44.11% 23.87% 993

51.65% 50.5

80.9% 5.31% 5.25% 4.95% 2.71% 31.98% 41.18% 26.85% 1,596

Appendix 2

145

ACS Mechanical 2014 Turk Estimates

NAES 2008

60.1% 20.3 –

50.8% 37.4 (median) –

56.62% 50.05

83.5% 4.4% – – –

73.8% 12.6% 5.0% 16.9% 2.9%

79.12% 9.67% 2.53% 6.3% 2.37%

40.8% 34.1% 16.9% 484–551

– – – –

62.86%

36.67% 20.82% 30.61% 19,234

The LDS Church does not allow outside groups to use membership information to conduct surveys of its members. Researchers who would like to study LDS Church members must find a way to access private information without the assistance of the church. Most surveys that report findings about LDS Church members use one of two means to collect their data. In the first approach, researchers gather information from LDS Institutes of Religion and the campuses at Brigham Young University (BYU), BYU– Idaho, and BYU–Hawaii. Because students at LDS-sponsored universities live in the same apartment complexes and have school-sponsored email accounts, anyone familiar with life at one of these institutions can figure out a way to contact these student populations. This is the source of data for the surveys reported in The Shield of Faith study and is currently a popular method for surveying LDS populations.3 The limitations of this approach are obvious. The results of a survey can be used to generalize only about the population from which the sample is derived. If the only population sampled is college-aged students affiliated with a university, the findings are not generalizable to the broader LDS population. In addition, apartment-complex-based sampling preferentially selects people who live there and excludes those who do not. If there are other differences between students who commute to campus and those who live in an apartment complex near campus, they will not be captured in the survey instrument.

Surveys of LDS Church Members

146

Appendix 2

The second approach to surveying LDS Church members is to post a link to an online survey on the internet. The sample size of this type of survey can be large because many more people have access to it. In addition, when such links are posted on websites about topics similar to the survey topic, viewers are more likely to participate. These sampling strengths, however, are also significant weaknesses of this approach. Ideally, a sample will include people with intense interest in the survey topic as well as those who have absolutely no interest in the survey topic. If the survey is posted only on websites frequented by those with an interest in the topic, the sample will not represent the views of a broader population; results will be generalizable only to populations who frequent certain websites. Finally, a few recent surveys of the LDS population have used sampling methods closer to the scientific ideal. David Campbell, John Green, and John Quinn Monson (2014) used a sample provider that matches survey respondents to census demographics and weights their responses accordingly.4 This dominant sampling method provides results that can be generalized to the US population. In 2011 the Pew Research Center used random digit dial for most of its sample but also recontacted self-identified Mormons who had participated in previous studies.5 To date, this Pew sample was the most scientific used in all the surveys that studied Mormons in the United States. The most challenging issue with studies of Mormons in the United States is that they represent such a small percentage of the broader US population. It is difficult to get a large enough sample through random selection. This is because most scientific polls of the US population that use random digit dialing have response rates of less than 5 percent. Even Gallup struggles to get 2 percent of those it reaches by phone to participate in a survey. In the 1950s–1960s, response rates of 60 percent were common, but the recent decline in survey response rates has made it much more difficult to sample smaller subsets of the US population. Though large in numbers, Mormons represent only 2 percent of the US population. This means if researchers call a random number, the probability that the person who agrees to participate in the survey is Mormon is 2 percent of 2 percent, or 0.0004. Thus a survey of the US LDS population that relies on random digit dial will miss a lot of Mormons who do not participate in the survey. I use a different sampling strategy. To collect responses for the Intermountain Latter-day Saint Faith (ILF) survey, I purchased a dataset of 300 million email addresses that belong to people in the United States. Each person in the data set visited a website (such as Netflix or Amazon) and agreed to the terms and conditions on the website. One of those conditions is that the company can sell your email address. Many of the people in the dataset had duplicate entries and bad email addresses. I merged all the duplicate entries, used a program to ping the remaining email addresses, and dropped those with nonfunctional addresses. Then, I created a subset of

Appendix 2

147

addresses for each of the states in the intermountain West and another subset of addresses for the LDS population centers of Arizona. This resulted in a database with 24,273 email addresses in Arizona, 29,739 in Idaho, and 25,897 in Utah. I used these email addresses to recruit participants for my study with an email invitation to complete a survey about religion and politics. The first question on the survey asked people their religious affiliation. Those who chose LDS were then asked about their level of activity in the church. All other respondents were asked if they had ever been a member of the LDS Church and how long it had been since they last attended church as a faithful member. Those who responded that they have never been members of the LDS Church did not continue with the survey. Of those invited to participate in the survey, 3,316 opened the email message. Of those who opened it, 748 chose to participate in the survey, and 257 said that they currently belonged to some faith other than the LDS Church. When asked if they had ever been a member of the LDS Church, 61 of the 257 responded that they had. These participants were invited to continue with the survey along with all respondents who said they were currently members of the LDS Church. A total of 56 of the former members chose to continue with the survey. When combined with the current LDS participants, this resulted in a sample of 357 respondents. Though this is not a large sample, the results from this sample are more likely to reflect the true attitudes of the Mormon population because respondents were selected using scientifically sound random sampling. The survey was fielded in fall 2014, and respondents were recontacted and asked to complete a shorter version of the survey in fall 2015. 1. Adam J. Berinsky, Gregory A. Huber, and Gabriel S. Lenz, “Evaluating Online Labor Markets for Experimental Research: Amazon.com’s Mechanical Turk,” Political Analysis 20, no. 3 (2012): 351–368; Richard Johnston, Kathleen Hall-Jameison, and Diana Carole Mutz, “National Annenberg Election Survey” (Los Angeles: Annenberg Public Policy Center, 2008), https://services.annenbergpublicpolicycenter.org/naes08/phone/index.html2008. 2. Kevin J. Mullinix et al., “The Generalizability of Survey Experiments,” Journal of Experimental Political Science 2, no. 2 (2015): 109–138. 3. Bruce A. Chadwick, Brent L. Top, and Richard J. McClendon, Shield of Faith: The Power of Religion in the Lives of LDS Youth and Young Adults (Provo and Salt Lake City, UT: Religious Studies Center, Brigham Young University/Deseret, 2010). 4. David E. Campbell, John C. Green, and J. Quin Monson, Seeking the Promised Land: Mormons and American Politics (Cambridge, UK: Cambridge University Press, 2014). 5. Luis Lugo et al., “Mormons in America: Certain in Their Beliefs, Uncertain of Their Place in Society” (Washington, DC: Pew Forum on Religion and Public Life, 2012).

Notes

Appendix 3: Descriptive Statistics and Complete Regression Models for Chapter 3

Table A3.1

Match Between Respondents’ Religion and Perceived Religion of President Barack Obama

Respondents’ Perception of Obama’s Religion Protestant

Roman Catholic Mormon

Eastern or Greek Orthodox Jewish

Muslim

Buddhist Hindu

Atheist

Agnostic

Nothing in particular Something else Total

Respondent Shares Religion with Obama

No

336 60.54% 96 64.43% 12 92.31% 5 83.33% 5 71.43% 513 99.61% 13 76.47% 6 66.67% 20 95.24% 17 80.95% 329 79.85% 215 73.38% 1,567 77.65%

Yes

219 39.46% 53 35.57% 1 7.69% 1 16.67% 2 28.57% 2 0.39% 4 23.53% 3 33.33% 1 4.76% 4 19.05% 83 20.15% 78 26.62% 451 22.35%

Total

555 100.00% 149 100.00% 13 100.00% 6 100.00% 7 100.00% 515 100.00% 17 100.00% 9 100.00% 21 100.00% 21 100.00% 412 100.00% 293 100.00% 2,018 100.00%

Source: Author’s data from 2015 Clear Voice Research national survey. Note: Entries are the raw count of respondents who believed their religion did not or did match that of the president; percentages are calculated by row.

149

150

Appendix 3

Table A3.2

Model-Predicted Respondents’ Approval of Their House Members

Religion match

Party ID match

Shared gender with House representative Religious identity Respondent sex Education Age

Race

Party ID Ideology

Religious identity × religion match

(1)

0.046*** (0.015) 1.382*** (0.013) 0.039** (0.017) 0.038*** (0.005) –0.057*** (0.017) –0.005 (0.004) 0.009*** (0.000) –0.004 (0.004) 0.203*** (0.010) –0.014*** (0.005)

–0.102** (0.051) 1.381*** (0.013) 0.039** (0.017) –0.012 (0.014) –0.056*** (0.017) –0.005 (0.004) 0.009*** (0.000) –0.004 (0.004) 0.209*** (0.030) –0.014*** (0.005) 0.040*** (0.010) –0.006 (0.023)

–3.023*** (0.049) 47,573 0.202

–2.839*** (0.078) 47,573 0.202

Party ID × religion match

Party ID × party ID match Gender × gender match Constant

Observations Pseudo R-Square

(2)

(3)

–0.094* (0.052) 1.760*** (0.038) 0.089 (0.055) 0.002 (0.014) –0.000 (0.055) –0.005 (0.004) 0.009*** (0.000) 0.006 (0.004) 0.505*** (0.041) –0.002 (0.005) 0.037*** (0.011) –0.003 (0.023) –0.243*** (0.023) –0.035 (0.035) –3.528*** (0.118) 47,573 0.204

Source: Author’s data from 2015 Clear Voice Research national survey. Notes: Entries are coefficients from a probit regression model. Robust standard errors are in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1, two-tailed test.

The text in Chapter 3 displays the results from probit models with categorical variables treated as numeric to conserve space. However, I also ran logistic regression models with each of the categorical variables treated as such. Tables A3.3 and A3.4 demonstrate that the results do not change whether I use logit or probit or when categorical variables are not numeric.

151

Appendix 3 Table A3.3

Model-Predicted Respondents’ Approval of Their US House Members Dependent Variable

Match (yes) Importance of religion: NA Importance of religion: not too important Importance of religion: somewhat important Importance of religion: very important Male Current gender of representative Current gender of representative: Female Education: four years Education: high school graduate Education: no high school Education: postgraduate Education: some college Birth year (age) Black Hispanic Middle Eastern Mixed Native American Other White Registered voter: no Registered voter: yes Party ID: lean Democrat Party ID: lean Republican Party ID: NA Party ID: not sure Party ID: not very strong Democrat Party ID: not very strong Republican Party ID: strong Democrat Party ID: strong Republican Current House party: Republican Current House party: vacant Political ideology: liberal Political ideology: middle of the road Political ideology: NA Political ideology: not sure

(1)

0.164*** (0.020)

Approval of US House Representative (2) 0.107*** (0.024)

(3)

–0.155** (0.063) 0.212 (0.398) 0.028 (0.042)

0.221*** (0.038) 0.257*** (0.037)

–0.035 (0.023) –14.845 (194.589)

0.0001 (0.023) –14.820 (193.815)

0.005 (0.039) –0.074* (0.038)

0.012 (0.039) –0.079** (0.038)

–0.547*** (0.074)

–0.569*** (0.074)

0.143*** (0.039)

–0.395*** (0.070) 0.064 (0.043) –0.052 (0.038) –0.015*** (0.001) 0.047 (0.086) –0.093 (0.088) –0.435 (0.329) –0.053 (0.108) –0.087 (0.142) –0.007 (0.111) 0.068 (0.082) –0.029 (0.116) 0.684*** (0.110) 1.172*** (0.062) –1.365*** (0.074) 0.531*** (0.111) 0.274*** (0.104) 0.958*** (0.059) 1.670*** (0.055) –1.605*** (0.072) 0.003 (0.056) –0.321*** (0.047) –0.226*** (0.037) –0.372*** (0.128) –0.780*** (0.059)

0.148*** (0.039)

–0.406*** (0.070) 0.071 (0.043) –0.050 (0.038) –0.014*** (0.001) –0.032 (0.087) –0.129 (0.089) –0.473 (0.330) –0.059 (0.108) –0.108 (0.143) –0.025 (0.112) 0.072 (0.082) –0.025 (0.117) 0.687*** (0.110) 1.194*** (0.062) –1.391*** (0.074) 0.506*** (0.112) 0.264** (0.105) 0.957*** (0.059) 1.661*** (0.055) –1.644*** (0.073) –0.006 (0.056) –0.236*** (0.047) –0.193*** (0.037) –0.331*** (0.128) –0.744*** (0.060)

continues

152

Appendix 3

Table A3.3

Continued

Political ideology: somewhat conservative Political ideology: somewhat liberal Political ideology: very conservative Political ideology: very liberal Match × importance of religion: NA Match × importance of religion: not too important Match × importance of religion: somewhat important Match × importance of religion: very important Male × current gender of representative Male × current gender of representative: female Party ID: lean Democrat × Current House party Republican Party ID: lean Republican × current House party Republican Party ID: NA × current House party Republican Party ID: not sure × current House party Republican Party ID: not very strong Democrat × current House party Republican Party ID: not very strong Republican × current House party Republican Party ID: strong Democrat × current House party Republican Party ID: strong Republican × current House party Republican Party ID: lean Democrat × current House party vacant Party ID: lean Republican × current House party vacant Party ID: NA × current House party vacant Party ID: not sure × current House party vacant Party ID: not very strong Democrat × current House party vacant

Dependent Variable

(1)

Approval of US House Representative (2)

(3)

–0.131*** (0.039)

–0.113*** (0.039)

0.037 (0.039)

0.026 (0.039)

–0.274*** (0.046) –0.270*** (0.053)

–0.216*** (0.046) –0.157*** (0.054) –1.204 (0.919) 0.381*** (0.094) 0.249*** (0.079)

–0.057 (121.095) –0.103* (0.056)

–2.096*** (0.085) 2.597*** (0.088)

–0.383*** (0.144) –0.344** (0.135)

–1.313*** (0.078)

0.301*** (0.072)

–0.058 (120.710)

–0.105* (0.057)

–2.096*** (0.085) 2.604*** (0.088)

–0.374*** (0.144)

–0.340** (0.135)

–1.317*** (0.079)

1.462*** (0.089)

1.466*** (0.089)

–2.973*** (0.072)

–2.973*** (0.072)

–1.430 (250.501)

–1.424 (249.775)

–0.835 (410.649)

–0.849 (410.979)

–1.011 (242.565)

–1.026 (242.095)

3.153*** (0.083)

0.893 (261.775)

–0.011 (458.443)

3.162*** (0.083)

0.877 (261.281)

–0.114 (456.930)

continues

153

Appendix 3 Table A3.3

Continued

Dependent Variable

(1)

Party ID: not very strong Republican × current House party vacant Party ID: strong Democrat × current House party vacant Party ID: strong Republican × current House party vacant Constant –0.112*** (0.010)

Observations Log likelihood Akaike inf. crit.

54,535 –37,731.380 75,466.760

Approval of US House Representative (2)

(3)

0.234 (275.108)

0.225 (274.798)

–1.914 (221.789)

–1.900 (221.324)

28.391*** (1.394)

26.692*** (1.405)

1.061 (232.100) 54,535 –29,212.580 58,535.150

1.029 (232.020) 54,535 –29,141.780 58,409.550

Source: Author’s data from 2015 Clear Voice Research national survey. Notes: Entries are logistic regression coefficients; standard errors are in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.

Table A3.4 Model-Predicted Respondents’ Approval of President Obama Religion match (yes)

Religious identity: important

Religious identity: not very important

Religious identity: not at all important Ideology: very liberal

Ideology: liberal

Ideology: somewhat liberal

Ideology: somewhat conservative Ideology: conservative

Ideology: very conservative

Education: less than high school Education: high school

Model 1

0.681 (0.199) –0.217 (0.226) 0.220 (0.284) 0.354 (0.256) 1.499*** (0.332) 1.722*** (0.264) 0.810*** (0.265) –0.518* (0.287) –1.368*** (0.349) –1.051** (0.429) 0.221 (0.670) 0.095 (0.229) ***

Model 2

–0.190 (0.379) –0.630** (0.265) 0.010 (0.315) 0.107 (0.280) 1.552*** (0.335) 1.761*** (0.266) 0.807*** (0.268) –0.529* (0.287) –1.315*** (0.350) –1.022** (0.431) 0.192 (0.671) 0.080 (0.231)

continues

154

Appendix 3

Table A3.4

Continued

Education: bachelor’s degree

Education: graduate degree (master’s, PhD) Race: black (African American)

Race: Asian

Race: Native American

Race: Hispanic or Latino/a

Race: more than one race; multiracial Party identification: Independent

Party identification: Republican Income: $39,999

Income: $74,999

Income: $99,999

Income: $149,999 Income: $199,999

Income: $200,000 or higher Gender: female

Religion match × religious identity: important Religion match × religious identity: not very important Religion match × religious identity: not at all important Constant N Log likelihood AIC

Model 1

0.072 (0.209) 0.298 (0.282) 1.683*** (0.256) 0.796** (0.372) –1.076 (0.846) 0.768** (0.340) 0.525 (0.553) –1.868*** (0.180) –2.853*** (0.290) 0.236 (0.258) 0.552** (0.250) 0.971*** (0.306) 0.638* (0.344) 1.525*** (0.499) 1.815*** (0.556) –0.465*** (0.166)

–0.079 (0.334) 1,291 –499.304 1,056.608

Model 2

0.088 (0.211) 0.256 (0.283) 1.715*** (0.258) 0.743** (0.374) –1.193 (0.841) 0.792** (0.341) 0.480 (0.560) –1.877*** (0.181) –2.902*** (0.294) 0.207 (0.261) 0.514** (0.253) 0.981*** (0.308) 0.609* (0.345) 1.548*** (0.500) 1.863*** (0.561) –0.416** (0.167) 1.456*** (0.484) 0.616 (0.706) 0.937 (0.685) 0.162 (0.346) 1,291 –494.547 1,053.094

Source: Author’s data from 2016 Clear Voice Research national survey Notes: *p < 0.05; **p < 0.01; ***p < 0.001 one-tail test. Entries are logistic regression coefficients; standard errors are in parentheses.

Appendix 4: Complete Regression Models for Chapter 4

models. The complete models are presented here in Tables A4.1 and A4.2.

The text in Chapter 4 plots the results from regression

155

Table A4.1

Trust in President Barack Obama (Complete Regression Model)

Ideology: moderate

Ideology: conservative

Education: high school

Education: some college Education: bachelor’s degree Education: graduate degree Race: Asian

Race: Native American Race: Hispanic Race: White

Race: multiracial Income

–0.055 (0.051) –0.312** (0.156) –1.590*** (0.245) –2.537*** (0.267) –0.165 (0.797) 0.271 (0.791) 0.200 (0.797) 0.569 (0.822) 0.390 (0.587) 0.090 (1.031) –0.590 (0.540) –1.334*** (0.403) –1.387** (0.614) 0.025 (0.039)

Only Some of the Time –0.161*** (0.043) 0.151 (0.131) –0.601** (0.243) –1.226*** (0.252) –0.740 (0.661) –0.149 (0.657) –0.450 (0.663) –0.080 (0.683) –0.560 (0.611) –0.184 (1.005) –0.650 (0.540) –0.507 (0.402) –0.678 (0.582) 0.012 (0.033)

Just About Always

–0.159*** (0.044) –0.733*** (0.216) –2.109*** (0.313) –2.489*** (0.337) –0.792 (0.890) –0.652 (0.882) –0.639 (0.890) –0.354 (0.927) 0.301 (0.664) –1.584 (1.420) –1.051* (0.609) –2.243*** (0.431) –1.354* (0.727) 0.078 (0.052)

Most of the Time

–0.053 (0.051) –0.308** (0.156) –1.591*** (0.244) –2.544*** (0.267) –0.238 (0.793) 0.198 (0.788) 0.124 (0.794) 0.506 (0.819) 0.383 (0.590) 0.009 (1.039) –0.639 (0.541) –1.396*** (0.402) –1.424** (0.613) 0.025 (0.040)

Only Some of the Time –0.159*** (0.043) 0.153 (0.131) –0.604** (0.242) –1.232*** (0.252) –0.792 (0.660) –0.200 (0.655) –0.502 (0.662) –0.130 (0.682) –0.570 (0.613) –0.248 (1.013) –0.682 (0.541) –0.555 (0.403) –0.701 (0.582) 0.011 (0.033)

Just About Always

–0.159*** (0.044) –0.727*** (0.217) –2.081*** (0.314) –2.469*** (0.338) –0.786 (0.893) –0.676 (0.885) –0.660 (0.894) –0.378 (0.930) 0.323 (0.666) –1.523 (1.425) –1.042* (0.610) –2.233*** (0.432) –1.375* (0.731) 0.071 (0.053)

Model 3

Most of the Time

–0.054 (0.051) –0.298* (0.157) –1.559*** (0.245) –2.517*** (0.267) –0.228 (0.792) 0.192 (0.786) 0.118 (0.792) 0.497 (0.817) 0.404 (0.591) 0.033 (1.047) –0.622 (0.541) –1.369*** (0.403) –1.433** (0.615) 0.018 (0.040)

Only Some of the Time –0.159*** (0.043) 0.154 (0.132) –0.613** (0.242) –1.242*** (0.252) –0.805 (0.660) –0.211 (0.655) –0.510 (0.661) –0.137 (0.682) –0.558 (0.613) –0.220 (1.013) –0.676 (0.541) –0.550 (0.403) –0.696 (0.584) 0.009 (0.033)

continues

Appendix 4

Gender: female

–0.161*** (0.043) –0.745*** (0.215) –2.127*** (0.313) –2.462*** (0.336) –0.704 (0.889) –0.580 (0.880) –0.541 (0.889) –0.278 (0.926) 0.251 (0.654) –1.317 (1.405) –1.003* (0.607) –2.172*** (0.432) –1.336* (0.728) 0.078 (0.052)

Most of the Time

Model 2

156

Age

Just About Always

Model 1

Table A4.1

Continued

Party ID: Independent

Just About Always

Model 1

Most of the Time

Only Some of the Time

–3.245*** (0.358) –4.110*** (0.547) –0.725*** (0.264) –0.034 (0.364) –0.695** (0.315) 0.346 (0.468) 1.199* (0.632) 2.835*** (0.806)

Most of the Time

–2.150*** (0.246) –3.312*** (0.318) 0.224 (0.200) 0.587** (0.275) 0.029 (0.236) –0.037 (0.443) 0.993* (0.513) 1.155* (0.694)

Only Some of the Time –0.957*** (0.245) –1.273*** (0.261) 0.174 (0.159) 0.660*** (0.228) 0.156 (0.198) –0.083 (0.463) 0.703 (0.520) 1.173** (0.566)

6.089*** 4.635*** 3.264*** (1.018) (0.930) (0.821) 1,935 1,935 1,935 4,090.288 4,090.288 4,090.288

Just About Always

Model 3

Most of the Time

Only Some of the Time

–3.257*** –2.167*** –0.961*** (0.358) (0.247) (0.245) –4.055*** –3.285*** –1.272*** (0.549) (0.318) (0.261) –0.662* 0.223 0.230 (0.345) (0.231) (0.172) 0.271 0.714** 0.707*** (0.429) (0.305) (0.245) –0.410 0.238 0.178 (0.370) (0.263) (0.216) 0.817 0.319 0.139 (0.598) (0.560) (0.557) 1.206* 1.005* 0.736 (0.634) (0.514) (0.519) 2.730*** 1.034 1.215** (0.809) (0.698) (0.569) –0.280 –0.131 –0.381 (0.591) (0.497) (0.453) –1.110 –0.672 –0.426 (0.855) (0.722) (0.664) –1.151 –1.118* –0.262 (0.729) (0.593) (0.525) 5.961*** 4.547*** 3.256*** (1.028) (0.932) (0.821) 1,935 1,935 1,935 4,097.921 4,097.921 4,097.921

157

Source: Author’s data from 2015 Clear Voice Research national survey. Notes: *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test. Entries are multinomial logistic regression coefficients; standard errors are in parentheses.

Appendix 4

–1.973*** –0.850*** –2.964*** (0.295) (0.219) (0.219) Party ID: Republican –3.242*** –3.147*** –1.110*** (0.400) (0.289) (0.238) Religion: important –0.659** 0.255 0.196 (0.263) (0.199) (0.158) Religion: not very important 0.006 0.601** 0.672*** (0.361) (0.274) (0.228) Religion: not at all important –0.622** 0.052 0.174 (0.313) (0.235) (0.197) Religion match (yes) 1.464*** 0.929*** 0.723*** (0.250) (0.201) (0.180) Party ID Independent × religion match (yes) Party ID Republican × religion match (yes) Religion match (yes) × religion: important Religion match (yes) × religion: not very important Religion match (yes) × religion: not at all important Constant 5.681*** 4.327*** 3.028*** (1.005) (0.921) (0.808) Observations 1,935 1,935 1,935 Akaike inf. crit. 4,091.535 4,091.535 4,091.535

Just About Always

Model 2

Table A4.2

Model-Predicted Trust in President Obama (Truncated Data)

Political ideology Income

Education Race

Gender

Party ID

Importance of religion

Party ID ×religion match Religion match ×religion important Constant Observations Pseudo R-Square

5.281*** (0.611) 1,441 0.171

0.186 (0.144) 0.003 (0.042) –0.759*** (0.100) 0.005 (0.033) 0.147** (0.070) –0.218*** (0.060) –0.164 (0.129) –0.985*** (0.108) –0.003 (0.061)

4.570*** (0.532) 1,441 0.171

Some of the Time

0.234* (0.138) –0.081** (0.040) –0.310*** (0.099) –0.009 (0.031) 0.103 (0.066) –0.024 (0.062) 0.213* (0.123) –0.191* (0.101) 0.063 (0.058)

1.200** (0.518) 1,441 0.171

Just About Always

–0.867* (0.471) –0.082 (0.051) –0.661*** (0.117) 0.035 (0.040) 0.120 (0.085) –0.348*** (0.063) –0.409*** (0.156) –1.485*** (0.171) –0.180** (0.075) 0.772*** (0.254) 5.856*** (0.647) 1,441 0.174

Most of the Time

–0.821* (0.423) 0.006 (0.042) –0.760*** (0.100) 0.004 (0.033) 0.144** (0.071) –0.231*** (0.060) –0.169 (0.130) –1.126*** (0.125) –0.009 (0.062) 0.512** (0.211) 4.940*** (0.555) 1,441 0.174

Some of the Time

–0.447 (0.418) –0.079** (0.040) –0.310*** (0.099) –0.011 (0.031) 0.102 (0.066) –0.033 (0.062) 0.211* (0.123) –0.274** (0.115) 0.059 (0.059) 0.319 (0.194) 1.443*** (0.539) 1,441 0.174

Just About Always

–0.258 (0.610) –0.082 (0.051) –0.648*** (0.118) 0.030 (0.040) 0.118 (0.085) –0.348*** (0.064) –0.394** (0.157) –1.470*** (0.171) –0.102 (0.090) 0.743*** (0.255) –0.244 (0.157) 5.612*** (0.663) 1,441 0.175

Model 3

Most of the Time

–0.254 (0.551) 0.006 (0.042) –0.748*** (0.101) 0.001 (0.033) 0.142** (0.071) –0.230*** (0.061) –0.157 (0.130) –1.116*** (0.125) 0.053 (0.072) 0.484** (0.212) –0.220 (0.134) 4.733*** (0.567) 1,441 0.175

Some of the Time

–0.299 (0.546) –0.080** (0.040) –0.308*** (0.099) –0.012 (0.031) 0.101 (0.066) –0.033 (0.062) 0.214* (0.123) –0.271** (0.115) 0.072 (0.068) 0.308 (0.196) –0.056 (0.129) 1.405** (0.550) 1,441 0.175

Source: Author’s data from 2015 Clear Voice Research national survey (truncated to exclude those who thought Obama was Muslim). Notes: Entries are regression coefficients from a multinomial probit model; standard errors are in parentheses. “Never” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 4

Age

0.506*** (0.167) –0.087* (0.050) –0.659*** (0.117) 0.034 (0.039) 0.124 (0.085) –0.330*** (0.063) –0.404*** (0.156) –1.206*** (0.134) –0.170** (0.075)

Most of the Time

Model 2

158

Religion match

Just About Always

Model 1

Appendix 5: Complete Regression Models for Chapter 5

Table A5.1

Experiment Balance

Female Age (mean years) Education (% completing some college) White Black Asian Latino/a Multiracial Party Identification Democrat Independent Republican N

Experimental Group, 2013

Latino/a

Openly Gay

80.58% 4.85% 5.34% 4.37% 3.88%

82.24% 6.57% 2.43% 3.89% 3.16%

50.49% 50.5 32.77%

27.71% 46.85% 25.44% 397

159

Obese

Atheist

82.00% 4.62% 5.84% 4.62% 2.19%

80.44% 4.40% 7.33% 5.62% 1.96%

49.88% 51.35 31.14%

53.04% 50.94 31.14%

33.92% 40.15% 25.94% 401

32.67% 36.91% 30.42% 401

52.32% 49.4 30.07%

33.58% 40.85% 25.56% 399

continues

160

Appendix 5

Table A5.1

Continued

Female Age (mean years) Education (% completing some college) White Black Asian Latino/a Multiracial Religiosity (mean) Religiosity (sd) Party identification Democrat Independent Republican N

Control

50.40% 47 33.06% 73.79% 6.85% 7.26% 8.06% 2.82% 0.346 0.204

35.89% 40.73% 23.39% 248

Experimental Group, 2014 Democrat

Republican

Atheist

73.68% 6.48% 6.88% 6.07% 4.05% 0.344 0.208

73.17% 9.76% 7.72% 6.50% 2.44% 0.359 0.199

72.22% 5.16% 8.73% 10.32% 2.78% 0.369 0.200

44.94% 46 36.44%

30.77% 48.99% 20.24% 247

52.03% 46 32.11%

35.37% 38.21% 26.42% 246

Source: Author’s data from 2013 & 2014 Research Now national survey.

Table A5.2

Acquaintance: obese

Acquaintance: openly gay Religious Identity

Acquaintance: Latino × religious identity Acquaintance: obese × religious identity

Acquaintance: openly gay × religious identity

Observations R2

26.19% 48.41% 25.40% 252

Complete Regression Models for Figure 5.6

Acquaintance: Latino

Constant

46.43% 43 32.14%

Register to Vote

Run for City Council

1,587 0.030

1,587 0.073

–0.159 (0.053) –0.062 (0.054) –0.0003 (0.052) –0.043*** (0.013) 0.047*** (0.018) 0.042** (0.019) 0.028 (0.018) 0.899*** (0.039) ***

–0.276*** (0.063) –0.240*** (0.065) 0.008 (0.062) –0.114*** (0.016) 0.114*** (0.022) 0.123*** (0.022) 0.054** (0.022) 0.869*** (0.046)

Source: Author’s data from 2013 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Atheist” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 5 Table A5.3

Complete Regression Models for Figure 5.7

Acquaintance: atheist

Acquaintance: Democrat

Acquaintance: Republican Religious identity

Acquaintance: atheist × religious identity

Acquaintance: Democrat × religious identity

Acquaintance: Republican × religious identity Constant

Observations R2

161

Register to Vote

Run for City Council

993 0.016

993 0.036

0.113* (0.062) 0.078 (0.059) 0.070 (0.062) 0.035** (0.015) –0.024 (0.022) –0.033 (0.021) –0.034 (0.022) 0.676*** (0.042)

0.319*** (0.074) 0.139** (0.071) 0.031 (0.074) 0.013 (0.018) –0.078*** (0.026) –0.048* (0.026) –0.012 (0.026) 0.476*** (0.050)

Source: Author’s data from 2014 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Control group” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 6: Complete Regression Models for Chapter 6

Table A6.1

Complete Regression Models for Figure 6.1 Fight a Speeding Ticket

Acquaintance: Latino Acquaintance: obese

Acquaintance: openly gay Religious identity

Acquaintance: Latino × religious identity Acquaintance: obese × religious identity

Acquaintance: openly gay × religious identity Constant

Observations R2

–0.267*** (0.065) –0.239*** (0.066) 0.019 (0.064) –0.078*** (0.017) 0.063*** (0.022) 0.080*** (0.023) 0.039* (0.023) 0.743*** (0.048)

1,587 0.073

Obtain a Marriage License –0.217*** (0.063) –0.176*** (0.065) 0.001 (0.062) –0.083*** (0.016) 0.077*** (0.022) 0.080*** (0.022) –0.043* (0.022) 0.938*** (0.046)

1,587 0.075

Source: Author’s data from 2013 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Atheist” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

163

164

Appendix 6

Table A6.2

Complete Regression Models for Figure 6.2

Acquaintance: atheist

Acquaintance: Democrat

Acquaintance: Republican Religious identity

Acquaintance: atheist × religious identity

Acquaintance: Democrat × religious identity

Acquaintance: Republican × religious identity Constant

Observations R2

Fight a Speeding Ticket

Obtain a Marriage License

993 0.076

993 0.018

0.358*** (0.073) 0.215*** (0.069) 0.223*** (0.072) 0.011 (0.018) –0.044* (0.026) –0.028 (0.025) –0.030 (0.025) 0.395*** (0.049)

0.245*** (0.067) 0.151** (0.064) 0.122* (0.067) 0.031* (0.016) –0.063*** (0.024) –0.043* (0.023) –0.041* (0.024) 0.599*** (0.046)

Source: Author’s data from 2014 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Control group” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 6 Table A6.3

Complete Regression Models for Figure 6.3

Acquaintance: Latino Acquaintance: obese

Acquaintance: openly gay Religious identity

Acquaintance: Latino × religious identity Acquaintance: obese × religious identity

Acquaintance: openly gay × religious identity Constant Observations R2

165

Purchase Auto Insurance –0.145 (0.058) –0.060 (0.059) 0.047 (0.057) –0.039*** (0.015) 0.026 (0.020) 0.030 (0.020) 0.016 (0.020) **

0.827*** (0.042)

1,587 0.044

Mail a Package –0.160*** (0.048) –0.125** (0.049) –0.001 (0.047) –0.046*** (0.012) 0.060*** (0.017) 0.054*** (0.017) 0.031* (0.017) 0.929*** (0.035)

1,587 0.030

Source: Author’s data from 2013 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Atheist” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

166

Appendix 6

Table A6.4

Complete Regression Models for Figure 6.4

Acquaintance: atheist

Acquaintance: Democrat

Acquaintance: Republican Religious identity

Acquaintance: atheist × religious identity

Acquaintance: Democrat × religious identity

Acquaintance: Republican × religious identity Constant

Observations R2

Purchase Auto Insurance

Mail a Package

993 0.042

993 0.015

0.255 (0.067) 0.147** (0.064) 0.174*** (0.067) 0.025 (0.016) –0.036 (0.024) –0.015 (0.023) –0.036 (0.023) 0.551*** (0.046) ***

0.167*** (0.059) 0.098* (0.057) 0.118** (0.059) 0.042*** (0.014) –0.051** (0.021) –0.041** (0.020) –0.051** (0.021) 0.685*** (0.040)

Source: Author’s data from 2014 Research Now national survey. Notes: Entries are regression coefficients; robust standard errors are in parentheses. “Control group” is the baseline category. *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 7: Regression Models for Chapter 7

Table A7.1

Complete Regression Models for Figure 7.2

Partisan Disidentification

Experimental group: Republican

Religious identity: very important

Party identification (pretest): strong Democrat Party identification (pretest): Democrat

Party identification (pretest): lean Democrat

Party identification (pretest): lean Republican Party identification (pretest): Republican

Party identification (pretest): strong Republican

Condition: Republican × very important religious identity Condition: Republican × party ID: strong Democrat

Condition: Republican × party ID: Democrat

Condition: Republican × party ID: lean Democrat

Condition: Republican × party ID: lean Republican

167

–0.221 (0.257) –1.019*** (0.225) –1.913*** (0.260) –0.311 (0.306) –0.518* (0.292) –0.986*** (0.367) –0.582** (0.281) –1.888*** (0.478) 0.787** (0.330) 0.349 (0.370) 0.042 (0.417) –0.082 (0.460) 0.375 (0.521)

continues

168

Appendix 7

Table A7.1

Continued

Condition: Republican × party ID: Republican

Condition: Republican × party ID: strong Republican

Very important religious identity × party ID: strong Democrat Very important religious identity × party ID: Democrat

Very important religious identity × party ID: lean Democrat

Very important religious identity × party ID: lean Republican

Very important religious identity × party ID: Republican

Very important religious identity × party ID: strong Republican Condition: Republican × very important religious identity × party ID: strong Democrat Condition: Republican × very important religious identity × party ID: Democrat Condition: Republican × very important religious identity × party ID: lean Democrat Condition: Republican × very important religious identity × party ID: lean Republican Condition: Republican × very important religious identity × party ID: Republican Condition: Republican × very important religious identity × party ID: strong Republican Constant

Observations R-squared

Partisan Disidentification 0.305 (0.464) 0.813 (0.596) 0.716** (0.329) 0.477 (0.382) 0.466 (0.443) 1.284*** (0.455) 0.677* (0.362) 0.751 (0.529) –1.045** (0.465) –0.858 (0.530) 0.033 (0.681) –1.396** (0.646) –0.454 (0.567) –1.177* (0.669) 4.721*** (0.171)

1,291 0.237

Source: Author’s data from 2016 Clear Voice Research national survey. Notes: Entries are OLS regression coefficients; robust standard errors are in parentheses. Baseline experimental group is “Democrat”; baseline religious identity is “not important”; baseline respondent party identification is “Independent.” *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Table A7.2

Complete Regression Models for Figure 7.3

71.471*** (1.821) –7.180 (7.573) –186.383*** (2.540) –184.239

–114.990*** (0.00000) –56.027 –46.343*** (0.980) 31.098*** (2.540) –71.388*** (0.00000) –41.135

–24.863*** (0.00000) –22.497*** (0.00000) –21.318

Democrat

16.317 (36.772) 68.251*** (11.296) –8.605 (25.493) –83.002*** (0.000) –69.100 (73.532) –27.551*** (0.000) –32.415*** (0.580) –13.791 (36.773) –26.420*** (0.000) 114.631 (73.532) –15.989*** (0.000) –66.082*** (11.301) –39.254*** (0.000)

Lean Republican 5.041 (53.085) –72.898 (52.330) –27.051*** (0.00000) –50.433*** (0.000) 70.256 (51.576) 0.156*** (0.000) 39.434 (52.330) 0.042*** (0.00000) –3.583*** (0.000) 8.687 (53.085) 0.215*** (0.000) 0.841*** (0.000) 0.144*** (0.000)

Republican 39.221 (285.111) –52.087 (46.765) –23.271 (504.785) –50.672*** (0.000) 81.089 (177.436) 54.992 (46.765) –70.820* (40.731) –6.898 (504.785) –9.048*** (0.000) –27.690 (178.945) 88.010** (40.731) 0.672*** (0.000) –1.685*** (0.000)

Strong Republican –61.575 (40.731) –27.952 (23.384) –24.705*** (0.0001) –50.301*** (0.000) 2.933 (23.386) 127.115*** (46.765) 33.774 (40.731) 0.813*** (0.000) –0.398*** (0.000) –33.926*** (0.000) –26.897 (40.731) 0.190*** (0.000) –4.730*** (0.000)

continues

169

–36.704*** (0.787) –8.566 (7.553) –84.036*** (6.716) –200.783*** (0.000) –115.615*** (0.000) –46.594*** (0.000) 62.526*** (0.832) 38.026*** (0.867) –22.866*** (0.000) –38.316*** (0.000) –8.558*** (0.000) 8.096 (7.604) –18.419*** (0.000)

Lean Democrat

Appendix 7

Experimental group: Republican Religious identity: very important Party identification (pretest): Democrat Party identification (pretest): Independent Party identification (pretest): Republican Party identification (pretest): strong Republican Group: Republican × very important religious identity Group: Republican × party ID: Democrat Group: Republican × party ID: Independent Group: Republican × party ID: Republican Group: Republican × party ID: strong Republican Very important religious identity × party ID: Democrat Very important religious identity × party ID: Independent

Strong Democrat

Predicted Party Identification (Post-test)

Continued

AIC

–58.224

–12.484

–3.399*** (0.000) –3.598

–14.134*** (0.000) –0.965*** (0.000) 83.813*** (6.711) 370.350

Predicted Party Identification (Post-test) Democrat

Lean Democrat

Lean Republican

–10.512*** (0.000) –67.919*** (0.000) –13.649*** (0.000)

0.086*** (0.000) 38.853 (52.330) 0.251*** (0.000)

–65.636*** (0.000) –18.624*** (0.000) –8.488*** (0.580)

–178.319*** (0.000) –23.073*** (0.000) 83.996*** (0.580)

83.813*** (6.703) 370.350

7.689 (25.493) 370.350

–6.536*** (0.000) –34.163*** (0.000) –6.899*** (0.000)

Republican

Strong Republican

–0.125*** (0.000) –26.508*** (0.000) –43.216 (40.731)

–0.081*** (0.000) –33.925*** (0.000) 68.586* (40.731)

–74.231 (52.330) 0.321*** (0.000) 0.196*** (0.000)

–129.660*** (0.000) 94.381** (46.765) 0.195*** (0.000)

–14.042 (51.576) 370.350

–23.776 (280.585) 370.350

27.745 (23.386) –23.620 (46.765) 0.195*** (0.000)

–2.725 (23.384) 370.350

Source: Author’s data from 2016 Clear Voice Research national survey. Notes: Entries are multinomial regression coefficients; robust standard errors are in parentheses. Baseline experimental group is “Democrat”; baseline religious identity is “not important”; baseline respondent party identification is “strong Democrat.” *p < 0.1; **p < 0.05; ***p < 0.01, two-tail test.

Appendix 7

Very important religious identity × party ID: Republican Very important religious identity × party ID: strong Republican Group: Republican × very important religious identity × party ID: Democrat Group: Republican × very important religious identity × party ID: Independent Group: Republican × very important religious identity × party ID: Republican Group: Republican × very important religious identity × party ID: strong Republican Constant

Strong Democrat

170

Table A7.2

Appendix 8: Complete Regression Models for Chapter 8

Table A8.1

Complete Regression Models for Figures 8.1–8.3

Church activity: not active

Church activity: active

Orthodoxy

LDS social identity

Experiment: alcohol dependence

Experiment: compulsive gambling Party: Independent Party: Republican

LDS identity × experiment: alcohol dependence LDS identity × experiment: compulsive gambling LDS identity × party: Independent LDS identity × party: Republican

Experiment: alcohol dependence × party: Independent

Church More Lenient 1.231 (0.729) 0.382 (0.632) –7.521*** (1.519) –1.105 (3.367) –0.688 (2.728) –3.009 (3.525) 0.129 (2.975) –1.747 (2.668) –2.030 (4.116) 1.243 (4.970) –2.995 (4.260) –0.925 (3.642) –2.071 (4.019) *

171

Church Support Lifestyle –0.333 (0.734) –0.359 (0.637) –5.072*** (1.529) –0.042 (3.390) –0.566 (2.747) –8.897** (3.549) 3.625 (2.995) 0.070 (2.686) –2.930 (4.144) 8.106 (5.003) –7.221* (4.289) –2.800 (3.666) –3.010 (4.047)

Not a Sin

0.273 (0.609) –0.495 (0.529) –4.576*** (1.270) –3.608 (2.815) –3.167 (2.281) –6.438** (2.947) –1.347 (2.488) –0.398 (2.231) 0.810 (3.441) 5.917 (4.155) –0.119 (3.562) –1.776 (3.045) 1.494 (3.360)

continues

172

Appendix 8

Table A8.1

Continued

Church More Lenient

Church Support Lifestyle

Not a Sin

2.483 (4.782) –1.893 (5.526)

8.382* (4.814) –4.924 (5.564)

5.893 (3.998) –2.797 (4.620)

Experiment: compulsive gambling × party: Independent Experiment: alcohol dependence × party: Republican Experiment: compulsive gambling × party: Republican LDS identity × experiment: alcohol dependence × Independent LDS identity × experiment: compulsive gambling × Independent LDS identity × experiment: alcohol dependence × Republican LDS identity × experiment: compulsive gambling × Republican Constant N R-squared

–2.454 (4.873) –0.004 (3.404) 2.964 (4.055) 6.323 (6.005) 6.840 (6.975)

10.742*** (2.167)

307 0.432

2.663 (4.906) –4.384 (3.427) 5.784 (4.082) 5.407 (6.046) –1.953 (7.022)

8.010*** (2.181)

307 0.313

2.169 (4.074) –2.761 (2.846) 3.380 (3.390) –0.969 (5.020) –1.447 (5.832)

10.173*** (1.811)

307 0.433

Source: Intermountain LDS Faith Survey, 2015, wave I. Notes: “Former LDS Church members,” “same-sex attraction,” and “Democrat” are baseline categories; robust standard errors are in parentheses. ***p < 0.001, **p < 0.01, *p < 0.05, one-tail test.

References

Adkins, Todd, Geoffrey C. Layman, David E. Campbell, and John C. Green. “Religious Group Cues and Citizen Policy Attitudes in the United States.” Politics and Religion 6, no. 2 (2013): 235–263. Albertson, Bethany L. “Dog-Whistle Politics: Multivocal Communication and Religious Appeals.” Political Behavior 37, no. 1 (2015): 3–26. ———. “Religious Appeals and Implicit Attitudes.” Political Psychology 32, no. 1 (2011): 109–130. Allport, Gordon W. The Nature of Prejudice. New York: Basic Books, 1979. American National Election Studies Time Series Cumulative Data File, Stanford University and the University of Michigan, 2010, www.electionstudies.org. Amiot, Catherine E., and Roxane de la Sablonniere. “Facilitating the Development and Integration of Multiple Social Identities.” Pp. 32–61 in The Psychology of Social and Cultural Diversity, edited by Richard J. Crisp. Hoboken, NJ: John Wiley and Sons, 2011. Ansolabehere, Stephen. “CCES Common Content, 2010.” Harvard Dataverse, 2012. Avery, Alison, Justin Chase, Linda Johansson, Samantha Litvak, Darrel Montero, and Michael Wydra. “America’s Changing Attitudes Toward Homosexuality, Civil Unions, and Same-Gender Marriage: 1977–2004.” Social Work 52, no. 1 (2007): 71–79. Bailey, J. Michael, Michael P. Dunne, and Nicholas G. Martin. “Genetic and Environmental Influences on Sexual Orientation and Its Correlates in an Australian Twin Sample.” Journal of Personality and Social Psychology 78, no. 3 (2000): 524. Banducci, Susan A., Todd Donovan, and Jeffrey A. Karp. “Minority Representation, Empowerment, and Participation.” Journal of Politics 66, no. 2 (2004): 534–556. Ben-Nun Bloom, Pazit, and Gizem Arikan. “A Two-Edged Sword: The Differential Effect of Religious Belief and Religious Social Context on Attitudes Towards Democracy.” Political Behavior 34, no. 2 (2012): 249–276. Ben-Nun Bloom, Pazit, Gizem Arikan, and Marie Courtemanche. “Religious Social Identity, Religious Belief, and Anti-Immigration Sentiment.” American Political Science Review 109, no. 2 (2015): 203–221.

173

174

References

Ben-Nun Bloom, Pazit, and Marie Courtemanche. “Religion, Morality, and Tolerance.” Pp. 100–116 in Religion and Political Tolerance in America: Advances in the State of the Art, edited by Paul A. Djupe. Philadelphia, PA: Temple University Press, 2015. Berinsky, Adam J., Gregory A. Huber, and Gabriel S. Lenz. “Evaluating Online Labor Markets for Experimental Research: Amazon.com’s Mechanical Turk.” Political Analysis 20, no. 3 (2012): 351–368. Boatright, Robert G. Getting Primaried: The Changing Politics of Congressional Primary Challenges. Ann Arbor: University of Michigan Press, 2013. Bobo, Lawrence, and Franklin D. Gilliam. “Race, Sociopolitical Participation, and Black Empowerment.” American Political Science Review 84, no. 2 (1990): 377–393. Bratton, Kathleen A. “The Effect of Legislative Diversity on Agenda Setting: Evidence from Six State Legislatures.” American Politics Research 30, no. 2 (2002): 115–142. Bratton, Kathleen A., and Kerry L. Haynie. “Agenda Setting and Legislative Success in State Legislatures: The Effects of Gender and Race.” Journal of Politics 61, no. 3 (1999): 658–679. Brewer, Marilynn B. “Multiple Identities and Identity Transition: Implications for Hong Kong.” International Journal of Intercultural Relations 23, no. 2 (1999): 187–197. Brewer, Marilynn B., and Kathleen P. Pierce. “Social Identity Complexity and Outgroup Tolerance.” Personality and Social Psychology Bulletin 31, no. 3 (2005): 428–437. Broockman, David E. “Approaches to Studying Policy Representation.” Legislative Studies Quarterly 41, no. 1 (2016): 181–215. Brown, Cody McClain. Chasing a Croatian Girl: A Survivor’s Tale. 3rd ed. CreateSpace Independent Publishing Platform, 2015. Burge, Ryan P. “Plenty of the ‘Nones’ Actually Head Back to Church.” Christianity Today, February 6, 2018. Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. The American Voter. New York: John Wiley and Sons, 1960. Campbell, David E., John C. Green, and Geoffrey C. Layman. “The Party Faithful: Partisan Images, Candidate Religion, and the Electoral Impact of Party Identification.” American Journal of Political Science 55, no. 1 (2011): 42–58. Campbell, David E., John C. Green, and J. Quin Monson. Seeking the Promised Land: Mormons and American Politics. Cambridge, UK: Cambridge University Press, 2014. Campbell, David E., Geoffrey C. Layman, John C. Green, and Nathanael G. Sumaktoyo. “Putting Politics First: The Impact of Politics on American Religious and Secular Orientations.” American Journal of Political Science 62, no. 3 (2018): 551–565. Campbell, David E., and Steven J. Yonish. “Religion and Volunteering in America.” Pp. 87–106 in Religion as Social Capital: Producing the Common Good. Waco, TX: Baylor University Press, 2003. Castro, Martin R., Patricia Timmons-Goodson, Roberta Achtenberg, Gail L. Heriot, Peter N. Kirsanow, David Kladney, Karen K. Narasaki, and Michael Yaki. Peaceful Coexistence: Reconciling Non-Discrimination Principles with Civil Liberties. Washington, DC: US Commission on Civil Rights, 2016, 306. Chadwick, Bruce A., Brent L. Top, and Richard J. McClendon. Shield of Faith: The Power of Religion in the Lives of LDS Youth and Young Adults. Provo and Salt Lake City, UT: Religious Studies Center, Brigham Young University/Deseret, 2010.

References

175

Chaiken, Shelly. “Heuristic Versus Systematic Information Processing and the Use of Source Versus Message Cues in Persuasion.” Journal of Personality and Social Psychology 39, no. 5 (1980): 752. Chua, Amy. “The Destructive Dynamics of Political Tribalism.” New York Times, February 20, 2018. ———. Political Tribes: Group Instinct and the Fate of Nations. New York: Penguin, 2018. Cikara, Mina, Matthew M. Botvinick, and Susan T. Fiske. “Us Versus Them Social Identity Shapes Neural Responses to Intergroup Competition and Harm.” Psychological Science 22, no. 3 (2011): 306–313. Citrin, Jack, and Donald Philip Green. “Presidential Leadership and the Resurgence of Trust in Government.” British Journal of Political Science 16, no. 4 (1986): 431–453. “Community Status Lapses on Conversion, Rules Madras High Court.” Hindu, June 24, 2013. Conover, Pamela Johnston. “The Influence of Group Identifications on Political Perception and Evaluation.” Journal of Politics 46, no. 3 (1984): 760–785. Converse, Phillip E. “The Nature of Belief Systems in Mass Publics.” Pp. 206–261 in Ideology and Discontent, edited by David Apter. New York: Free Press, 1964. ———. “Some Priority Variables in Comparative Electoral Research.” Pp. 727–745 in Electoral Behavior: A Comparative Handbook, edited by Richard E. Rose. New York: Free Press, 1974. Cooperman, Alan et al. 2014 U.S. Religious Landscape Study. Washington, DC: Pew Research Center, 2015. Cox, Daniel, Robert P. Jones, and Juhem Navarro-Rivera. “Nonreligious Tolerance.” Pp. 131–150 in Religion and Political Tolerance in America: Advances in the State of the Art, edited by Paul A. Djupe. Philadelphia, PA: Temple University Press, 2015. Dalton, Russell J. The Apartisan American: Dealignment and Changing Electoral Politics. Washington, DC: Congressional Quarterly Press, 2013. ———. Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Comparative Politics. Oxford, UK: Oxford University Press, 2004. ———. “The Social Transformation of Trust in Government.” International Review of Sociology 15, no. 1 (2005): 133–154. Day, Abby. Believing in Belonging: Belief and Social Identity in the Modern World. Oxford, UK: Oxford University Press, 2011. Delli Carpini, Michael X., and Scott Keeter. What Americans Know About Politics and Why It Matters. New Haven, CT: Yale University Press, 1996. Djupe, Paul A. Religion and Political Tolerance in America: Advances in the State of the Art. Philadelphia, PA: Temple University Press, 2015. Djupe, Paul, and Brian Calfano. God Talk: Experimenting with the Religious Causes of Public Opinion. Philadelphia, PA: Temple University Press, 2014. Djupe, Paul A., and Christopher P. Gilbert. The Political Influence of Churches. Cambridge Studies in Social Theory, Religion, and Politics. Cambridge, UK: Cambridge University Press, 2009. ———. The Prophetic Pulpit: Clergy, Churches, and Communities in American Politics. Lanham, MD: Rowman and Littlefield, 2003. Djupe, Paul A., and Stephen T. Mockabee. “Religious Worldviews and Political Tolerance.” Pp. 117–132 in Religion and Political Tolerance in America: Advances in the State of the Art, edited by Paul A. Djupe. Philadelphia, PA: Temple University Press, 2015.

176

References

Djupe, Paul A., Jacob R. Neiheisel, and Anand E. Sokhey. “Reconsidering the Role of Politics in Leaving Religion: The Importance of Affiliation.” American Journal of Political Science 62, no. 1 (2018): 161–175. Doherty, Carroll et al. “Partisanship and Political Animosity in 2016.” Washington, DC: Pew Research Center, 2016. Dolan, Kathleen. “Symbolic Mobilization? The Impact of Candidate Sex in American Elections.” American Politics Research 34, no. 6 (2006): 687–704. Doosje, Bertjan, Russell Spears, and Naomi Ellemers. “The Dynamic and Determining Role of Ingroup Identification: Responses to Anticipated and Actual Changes in the Intergroup Status Hierarchy.” British Journal of Social Psychology 41, no. 1 (2002): 57–76. Druckman, James N., and Lawrence R. Jacobs. Who Governs? Presidents, Public Opinion, and Manipulation. Chicago Studies in American Politics. Chicago: University of Chicago Press, 2015. Druckman, James N., Erik Peterson, and Rune Slothuus. “How Elite Partisan Polarization Affects Public Opinion Formation.” American Political Science Review 107, no. 1 (2013): 57–79. Dugan, Andrew. “U.S. Congress and Its Leaders Suffer Public Discontent.” Gallup, http://www.gallup.com/poll/184556/congress-leaders-suffer-public-discontent .aspx. Eagly, Alice H., and Shelly Chaiken. The Psychology of Attitudes. New York: Harcourt Brace, 1993. Egan, Patrick J., and Kenneth Sherrill. “California’s Proposition 8: What Happened, and What Does the Future Hold?” San Francisco, CA: National Gay and Lesbian Task Force Policy Institute, 2009. Eisenstein, Marie A., and April K. Clark. “Heterogeneous Religion Measures and Political Tolerance Outcomes.” Pp. 83–99 in Religion and Political Tolerance in America: Advances in the State of the Art, edited by Paul A. Djupe. Philadelphia, PA: Temple University Press, 2015. Ellemers, Naomi, and Manuela Barreto. “The Impact of Relative Group Status: Affective, Perceptual, and Behavioral Consequences.” Pp. 325–343 in Blackwell Handbook of Social Psychology: Intergroup Processes. Oxford, UK: Blackwell, 2001. Ellemers, Naomi, Russell Spears, and Bertjan Doosje. “Self and Social Identity.” Annual Review of Psychology 53, no. 1 (2002): 161–186. Eulau, Heinz, and Paul D. Karps. “The Puzzle of Representation: Specifying Components of Responsiveness.” Legislative Studies Quarterly 2, no. 3 (1977): 233– 254. Fenno, Richard F. Home Style: House Members in Their Districts. Boston: Little, Brown, 1978. Festinger, Leon. “A Theory of Social Comparison Processes.” Human Relations 7, no. 2 (1954): 117–140. Finke, Roger, and Laurence R. Iannaccone. “Supply-Side Explanations for Religious Change.” Annals of the American Academy of Political and Social Science 527, no. 1 (1993): 27–39. Finke, Roger, and Rodney Stark. The Churching of America, 1776–2005: Winners and Losers in Our Religious Economy. New Brunswick, NJ: Rutgers University Press, 2005. Fiske, Susan T., and Patricia W. Linville. “What Does the Schema Concept Buy Us?” Personality and Social Psychology Bulletin 6, no. 4 (1980): 543–557. Franz, Michael M. Choices and Changes: Interest Groups in the Electoral Process. Philadelphia, PA: Temple University Press, 2008.

References

177

Freeden, Michael. Ideology: A Very Short Introduction. Oxford, UK: Oxford University Press, 2003. Friesen, Amanda, and Aleksander Ksiazkiewicz. “Do Political Attitudes and Religiosity Share a Genetic Path?” Political Behavior 37, no. 4 (2015): 791–818. Gaines, Brian J., James H. Kuklinski, Paul J. Quirk, Buddy Peyton, and Jay Verkuilen. “Same Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq.” Journal of Politics 69, no. 4 (2007): 957–974. Gandhi, Mohandas. “The Ashram of Soul-Force.” In Mahatma Gandhi’s Ideas, edited by C. F. Andres. London: Allen, 1949. Gay, Claudine. “The Effect of Black Congressional Representation on Political Participation.” American Political Science Review 95, no. 3 (2002): 589–617. Gervais, Will M., Azim F. Shariff, and Ara Norenzayan. “Do You Believe in Atheists? Distrust Is Central to Anti-Atheist Prejudice.” Journal of Personality and Social Psychology 101, no. 6 (2011): 1189. Gibson, James L. “Enigmas of Intolerance: Fifty Years After Stouffer’s Communism, Conformity, and Civil Liberties.” Perspectives on Politics 4, no. 1 (2006): 21–34. ———. “The Political Consequences of Religiosity: Does Religion Always Cause Political Intolerance?” Pp. 147–175 in Religion and Democracy in the United States: Danger or Opportunity, edited by Ira Katznelson and Alan Wolfe. New York: Russell Sage, 2010. Gibson, James L., and Amanda Gouws. Overcoming Intolerance in South Africa: Experiments in Democratic Persuasion. Cambridge Studies in Political Psychology and Public Opinion. Cambridge, UK: Cambridge University Press, 2005. ———. “Social Identities and Political Intolerance: Linkages Within the South African Mass Public.” American Journal of Political Science 44, no. 2 (2000): 278–292. Glazier, Rebecca A. “Bridging Religion and Politics: The Impact of Providential Religious Beliefs on Political Activity.” Politics and Religion 8, no. 3 (2015): 458–487. Goodstein, Laurie. “Donald Trump Reveals Evangelical Rifts That Could Shape Politics for Years.” New York Times, October 17, 2016. http://www.nytimes .com/2016/10/17/us/donald-trump-evangelicals-republican-vote.html?_r=0. Gopin, Marc. Holy War, Holy Peace: How Religion Can Bring Peace to the Middle East. Oxford, UK: Oxford University Press, 2002. Goren, Paul, and Christopher Chapp. “Moral Power: How Public Opinion on Culture War Issues Shapes Partisan Predispositions and Religious Orientations.” American Political Science Review 111, no. 1 (2017): 110–128. Green, Donald P., Bradley Palmquist, and Eric Schickler. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. Yale ISPS Series. New Haven, CT: Yale University Press, 2002. Green, John C. The Faith Factor: How Religion Influences American Elections. Westport, CT: Greenwood, 2007. Greenwald, Tony, Mahzarin R. Banaji, and Brian A. Nosek. “Project Implicit.” https://implicit.harvard.edu/implicit/index.jsp. Groenendyk, Eric. Competing Motives in the Partisan Mind: How Loyalty and Responsiveness Shape Party Identification and Democracy. Oxford, UK: Oxford University Press, 2013. Guth, James L., and John C. Green. “Salience: The Core Concept.” Pp. 157–176 in Rediscovering the Religious Factor in American Politics, edited by David C. Leege and Lyman A. Kellstedt. New York: M. E. Sharpe, 1993.

178

References

Guth, James L., Lyman A. Kellstedt, Corwin E. Smidt, and John C. Green. “Religious Influences in the 2004 Presidential Election.” Presidential Studies Quarterly 36, no. 2 (2006): 223–242. Haider-Markel, Donald P. “Representation and Backlash: The Positive and Negative Influence of Descriptive Representation.” Legislative Studies Quarterly 32, no. 1 (2007): 107–133. Haider-Markel, Donald P., and Mark R. Joslyn. “Beliefs About the Origins of Homosexuality and Support for Gay Rights: An Empirical Test of Attribution Theory.” Public Opinion Quarterly 72, no. 2 (2008): 291–310. Haidt, Jonathan. The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Vintage, 2012. Hall, Deborah L., David C. Matz, and Wendy Wood. “Why Don’t We Practice What We Preach? A Meta-Analytic Review of Religious Racism.” Personality and Social Psychology Review 14, no. 1 (2010): 126–139. Hetherington, Marc J. “Putting Polarization in Perspective.” British Journal of Political Science 39, no. 2 (2009): 413–448. Hetherington, Marc J., and Thomas J. Rudolph. Why Washington Won’t Work: Polarization, Political Trust, and the Governing Crisis. Chicago: University of Chicago Press, 2015. Hewstone, Miles Ed, and Rupert Ed Brown. Contact and Conflict in Intergroup Encounters. Oxford, UK: Blackwell, 1986. Hilton, James L., and William Von Hippel. “Stereotypes.” Annual Review of Psychology 47, no. 1 (1996): 237–271. Hogg, Michael A., and Graham M. Vaughan. Essentials of Social Psychology. New York: Pearson, 2009. Hogg, Michael A., and Scott A. Reid. “Social Identity, Self-Categorization, and the Communication of Group Norms.” Communication Theory 16, no. 1 (2006): 7–30. Hout, Michael, and Claude S. Fischer. “Explaining Why More Americans Have No Religious Preference: Political Backlash and Generational Succession, 1987– 2012.” Sociological Science 1, no. 9 (2014): 423–447. Huckfeldt, Robert, Eric Plutzer, and John Sprague. “Alternative Contexts of Political Behavior: Churches, Neighborhoods, and Individuals.” Journal of Politics 55, no. 2 (1993): 365–381. Huddy, Leonie, Lilliana Mason, and Lene Aarøe. “Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity.” American Political Science Review 109, no. 1 (2015): 1–17. Hurd, Elizabeth Shakman. Beyond Religious Freedom: The New Global Politics of Religion. Princeton, NJ: Princeton University Press, 2015. Iyengar, Shanto, Gaurav Sood, and Yphtach Lelkes. “Affect, Not Ideology: A Social Identity Perspective on Polarization.” Public Opinion Quarterly 76, no. 3 (2012): 405–431. Iyengar, Shanto, and Sean J. Westwood. “Fear and Loathing Across Party Lines: New Evidence on Group Polarization.” American Journal of Political Science 59, no. 3 (2015): 690–707. Jacobsmeier, Matthew L. “Religion and Perceptions of Candidates’ Ideologies in United States House Elections.” Politics and Religion 6, no. 2 (2013): 342–372. Jacobson, Gary C. The Politics of Congressional Elections. 7th ed. Longman Classics in Political Science. New York: Pearson Longman, 2009. Jelen, Ted G. “The Political Consequences of Religious Group Attitudes.” Journal of Politics 55, no. 1 (1993): 178–190.

References

179

Johnston, Christopher D., Howard Lavine, and Christopher M. Federico. Open Versus Closed: Personality, Identity, and the Politics of Redistribution. Cambridge, UK: Cambridge University Press, 2017. Johnston, Richard, Kathleen Hall-Jameison, and Diana Carole Mutz. “National Annenberg Election Survey.” Los Angeles: Annenberg Public Policy Center, 2008. https://services.annenbergpublicpolicycenter.org/naes08/phone/index.html. Jones, David R. “Partisan Polarization and Congressional Accountability in House Elections.” American Journal of Political Science 54, no. 2 (2010): 323–337. Jost, John T., and Diana Burgess. “Attitudinal Ambivalence and the Conflict Between Group and System Justification Motives in Low Status Groups.” Personality and Social Psychology Bulletin 26, no. 3 (2000): 293–305. Kellstedt, Lyman A., and John C. Green. “Knowing God’s Many People: Denominational Preference and Political Behavior.” Pp. 53–71 in Rediscovering the Religious Factor in American Politics, edited by David C. Leege and Lyman A. Kellstedt. Armonk, NY: M. E. Sharpe, 1993. Kinder, Donald R., and Nathan P. Kalmoe. Neither Liberal nor Conservative: Ideological Innocence in the American Public. Chicago: University of Chicago Press, 2017. Klar, Samara, and Yanna Krupnikov. Independent Politics: How American Disdain for Parties Leads to Political Inaction. New York: Cambridge University Press, 2016. Klar, Samara, Yanna Krupnikov, and John Barry Ryan. “Affective Polarization or Partisan Disdain? Untangling a Dislike for the Opposing Party from a Dislike of Partisanship.” Public Opinion Quarterly 82, no. 2 (2018): 379–390. Kraft, Patrick W., Milton Lodge, and Charles S. Taber. “Why People ‘Don’t Trust the Evidence’: Motivated Reasoning and Scientific Beliefs.” Annals of the American Academy of Political and Social Science 658, no. 1 (March 1, 2015): 121–133. Kunda, Ziva. “The Case for Motivated Reasoning.” Psychological Bulletin 108, no. 3 (1990): 480. Långström, Niklas, Qazi Rahman, Eva Carlström, and Paul Lichtenstein. “Genetic and Environmental Effects on Same-Sex Sexual Behavior: A Population Study of Twins in Sweden.” Archives of Sexual Behavior 39, no. 1 (2010): 75–80. Lau, Richard R., and David P. Redlawsk. “Advantages and Disadvantages of Cognitive Heuristics in Political Decision Making.” American Journal of Political Science 45, no. 4 (2001): 951–971. Lauderdale, Benjamin E., Chris Hanretty, and Nick Vivyan. “Decomposing Public Opinion Variation into Ideology, Idiosyncrasy, and Instability.” Journal of Politics 80, no. 2 (2018): 707–712. Lawless, Jennifer L. “Politics of Presence? Congresswomen and Symbolic Representation.” Political Research Quarterly 57, no. 1 (2004): 81–99. ———. “Sexism and Gender Bias in Election 2008: A More Complex Path for Women in Politics.” Politics and Gender 5, no. 1 (2009): 70–80. Layman, Geoffrey C. The Great Divide: Religious and Cultural Conflict in American Party Politics. New York: Columbia University Press, 2001. ———. “Where Is Trump’s Evangelical Base? Not in Church.” Washington Post, March 29, 2016. Layman, Geoffrey C., and Thomas M. Carsey. “Party Polarization and ‘Conflict Extension’ in the American Electorate.” American Journal of Political Science 46, no. 4 (2002): 786–802. Layman, Geoffrey C., and Christopher L. Weaver. “Religion and Secularism Among American Party Activists.” Politics and Religion 9, no. 2 (2016): 271–295.

180

References

Lenz, Gabriel S. Follow the Leader? How Voters Respond to Politicians’ Policies and Performance. Chicago: University of Chicago Press, 2013. Levin, Eve. “Dvoeverie and Popular Religion.” Pp. 29–52 in Seeking God: The Recovery of Religious Identity in Orthodox Russia, Ukraine, and Georgia, edited by Stephen K. Batalden. DeKalb: Northern Illinois University Press, 1993. Lipset, Seymour Martin. Political Man: The Social Bases of Politics. Garden City, NY: Doubleday, 1960. Lodge, Milton, and Charles S. Taber. The Rationalizing Voter. Cambridge Studies in Public Opinion and Political Psychology. Cambridge, UK: Cambridge University Press, 2013. ———. “The Rationalizing Voter: Unconscious Thought in Political Information Processing.” Unpublished paper, Stony Brook University (2008). Lori, William E. et al. to Barack Obama, Orrin Hatch, and Paul Ryan, letter, October 7, 2016. Lugo, Luis, Alan Cooperman, Gregory A. Smith, Erin O’Connell, and Sandra Stencel. “Mormons in America: Certain in Their Beliefs, Uncertain of Their Place in Society.” Washington, DC: Pew Forum on Religion and Public Life, 2012. Lugo, Luis, Sandra Stencel, John Green, Gregory Smith, Dan Cox, and Allison Pond. “2007 U.S. Religious Landscape Survey.” Washington, DC: Pew Forum on Religion and Public Life, 2008. Lupia, Arthur. “Shortcuts Versus Encyclopedias: Information and Voting Behavior in California Insurance Reform Elections.” American Political Science Review 88, no. 1 (1994): 63–76. Margolis, Michele F. From Politics to the Pews: How Partisanship and the Political Environment Shape Religious Identity. Chicago: University of Chicago Press, 2018. ———. “How Far Does Social Group Influence Reach? Identities, Elites, and Immigration Attitudes.” Journal of Politics 80, no. 3 (2018): 772–785. Markoe, Lauren. “Poll: Trump Support Remains Steady Among Evangelicals.” Religion News Service, October 11, 2016. http://religionnews.com/2016/10/11/poll-trump -support-remains-steady-among-evangelicals/. Masci, David, and Tracy Miller. “Faith on the Hill: 2008.” Washington, DC: Pew Research Center, 2008. Mason, Lilliana. “A Cross-Cutting Calm: How Social Sorting Drives Affective Polarization.” Public Opinion Quarterly 80, no. S1 (2016): 351–377. ———. “‘I Disrespectfully Agree’: The Differential Effects of Partisan Sorting on Social and Issue Polarization.” American Journal of Political Science 59, no. 1 (2015): 128–145. ———. “The Rise of Uncivil Agreement: Issue Versus Behavioral Polarization in the American Electorate.” American Behavioral Scientist 57, no. 1 (2013): 140– 159. Mason, Lilliana, and Julie Wronski. “One Tribe to Bind Them All: How Our Social Group Attachments Strengthen Partisanship.” Political Psychology 39, no. S1 (2018): 257–277. McDermott, Monika L. “Religious Stereotyping and Voter Support for Evangelical Candidates.” Political Research Quarterly 62, no. 2 (2009): 340–354. ———. “Voting for Catholic Candidates: The Evolution of a Stereotype.” Social Science Quarterly 88, no. 4 (2007): 953–969. Miller, Arthur H. “Political Issues and Trust in Government: 1964–1970.” American Political Science Review 68, no. 3 (1974): 951–972.

References

181

Miller, Patrick R., and Pamela Johnston Conover. “Red and Blue States of Mind: Partisan Hostility and Voting in the United States.” Political Research Quarterly 68, no. 2 (2015): 225–239. Mullinix, Kevin J. “Partisanship and Preference Formation: Competing Motivations, Elite Polarization, and Issue Importance.” Political Behavior 38, no. 2 (2016): 383–411. Mullinix, Kevin J., Thomas J. Leeper, James N. Druckman, and Jeremy Freese. “The Generalizability of Survey Experiments.” Journal of Experimental Political Science 2, no. 2 (2015): 109–138. Nagel, Joane. American Indian Ethnic Renewal: Red Power and the Resurgence of Identity and Culture. Oxford, UK: Oxford University Press, 1997. Nayar, Baldev Raj. Minority Politics in the Punjab. Princeton, NJ: Princeton University Press, 1966. Noel, Hans. Political Ideologies and Political Parties in America. Cambridge, UK: Cambridge University Press, 2014. Norenzayan, Ara, and Azim F. Shariff. “The Origin and Evolution of Religious Prosociality.” Science 322, no. 5898 (2008): 58–62. Norris, Pippa. Critical Citizens: Global Support for Democratic Government. Oxford, UK: Oxford University Press, 1999. Norris, Pippa, and Ronald Inglehart. Sacred and Secular: Religion and Politics Worldwide. 1st ed. Cambridge Studies in Social Theory, Religion, and Politics. Cambridge, UK: Cambridge University Press, 2004. ———. Sacred and Secular: Religion and Politics Worldwide. 2nd ed. Cambridge Studies in Social Theory, Religion, and Politics. Cambridge, UK: Cambridge University Press, 2011. Penning, James M. “Americans’ Views of Muslims and Mormons: A Social Identity Theory Approach.” Politics and Religion 2, no. 2 (2009): 277–302. ———. “The Political Behavior of American Catholics: An Assessment of the Impact of Group Integration Versus Group Identification.” Western Political Quarterly 41, no. 2 (1988): 289–308. Pettigrew, Thomas F. “Intergroup Contact Theory.” Annual Review of Psychology 49, no. 1 (1998): 65–85. Pettigrew, Thomas F., and Linda R. Tropp. “A Meta-Analytic Test of Intergroup Contact Theory.” Journal of Personality and Social Psychology 90, no. 5 (2006): 751. Petty, Richard E., and John T. Cacioppo. “The Elaboration Likelihood Model of Persuasion.” Advances in Experimental Social Psychology 19, no. 1 (1986): 123–205. Posner, Daniel N. “The Political Salience of Cultural Difference: Why Chewas and Tumbukas Are Allies in Zambia and Adversaries in Malawi.” American Political Science Review 98, no. 4 (2004): 529–545. Preacher, Kristopher J., Derek D. Rucker, and Andrew F. Hayes. “Addressing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions.” Multivariate Behavioral Research 42, no. 1 (2007): 185–227. “Public Trust in Government: 1958–2014.” http://www.people-press.org/2014/11 /13/public-trust-in-government/. Purzycki, Benjamin Grant, Coren Apicella, Quentin D. Atkinson, Emma Cohen, Rita Anne McNamara, Aiyana K. Willard, Dimitris Xygalatas, Ara Norenzayan, and Joseph Henrich. “Moralistic Gods, Supernatural Punishment, and the Expansion of Human Sociality.” Nature 530 (Febuary 18, 2016): 327–330. Putnam, Robert D., and David E. Campbell. American Grace: How Religion Divides and Unites Us. New York: Simon and Schuster, 2010.

182

References

Redlawsk, D. P. “Hot Cognition or Cool Consideration? Testing the Effects of Motivated Reasoning on Political Decision Making.” Journal of Politics 64, no. 4 (2002): 1021–1044. Redlawsk, David P., Andrew J. W. Civettini, and Karen M. Emmerson. “The Affective Tipping Point: Do Motivated Reasoners Ever ‘Get It’?” Political Psychology 31, no. 4 (2010): 563–593. Reich, Robert. “The New Tribalism and the Decline of the Nation State.” March 23, 2014. http://robertreich.org/post/80522686347. Roccas, Sonia, and Marilynn B. Brewer. “Social Identity Complexity.” Personality and Social Psychology Review 6, no. 2 (2002): 88–106. Saad, Lydia. “Government Named Top U.S. Problem for Second Straight Year.” Gallup, 2016. Saltzstein, Grace Hall. “Black Mayors and Police Policies.” Journal of Politics 51, no. 3 (1989): 525–544. ———. “Female Mayors and Women in Municipal Jobs.” American Journal of Political Science 30, no. 1 (1986): 140–164. Schaffer, Joby, Anand E. Sokhey, and Paul A. Djupe. “The Religious Economy of Political Tolerance.” Pp. 151–164 in Religion and Political Tolerance in America: Advances in the State of the Art, edited by Paul A. Djupe. Philadelphia, PA: Temple University Press, 2015. Schofield, Thomas J. “Knowing What We Don’t Know: A Meta-Analysis of Children Raised by Gay or Lesbian Parents.” The Winnower 3 (2016). Shariff, Azim F., Aiyana K. Willard, Teresa Andersen, and Ara Norenzayan. “Religious Priming: A Meta-Analysis with a Focus on Prosociality.” Personality and Social Psychology Review 20, no. 1 (2016): 27–48. Sherif, Muzafer. Experimental Study of Positive and Negative Intergroup Attitudes Between Experimentally Produced Groups: Robbers Cave Study. Norman: University of Oklahoma Institute of Group Relations, 1954. ———. Intergroup Conflict and Cooperation: The Robbers Cave Experiment. Vol. 10. Norman, OK: University of Oklahoma Institute of Group Relations, 1961. Sinclair, Betsy. The Social Citizen: Peer Networks and Political Behavior. Chicago: University of Chicago Press, 2012. Slothuus, Rune, and Claes H. De Vreese. “Political Parties, Motivated Reasoning, and Issue Framing Effects.” Journal of Politics 72, no. 3 (2010): 630–645. Slutske, Wendy S., Seth Eisen, William R. True, Michael J. Lyons, Jack Goldberg, and Ming Tsuang. “Common Genetic Vulnerability for Pathological Gambling and Alcohol Dependence in Men.” Archives of General Psychiatry 57, no. 7 (2000): 666–673. Smidt, Corwin E. Pews, Prayers, and Participation: Religion and Civic Responsibility in America. Washington, DC: Georgetown University Press, 2008. ———. Pulpit and Politics: Clergy in American Politics at the Advent of the Millennium. Waco, TX: Baylor University Press, 2004. Smidt, Corwin, John C. Green, James Guth, and Lyman Kellstedt. “Religious Involvement, Social Capital, and Political Engagement.” P. 153 in Religion as Social Capital: Producing the Common Good. Waco, TX: Baylor University Press, 2003. Smidt, Corwin, Lyman Kellstedt, and James Guth. “The Role of Religion in American Politics: Explanatory Theories and Associated Analytical and Measurment Issues.” Pp. 3–42 in The Oxford Handbook of Religion and American Politics, edited by Corwin Smidt, Lyman Kellstedt, and James Guth. New York: Oxford University Press, 2009.

References

183

Smith, Gregory A., and Alan Cooperman. “The Factors Driving the Growth of Religious ‘Nones’ in the U.S.” Fact Tank: News in the Numbers (online). Washington, DC: Pew Research Center, 2016. Sniderman, Paul M., and Aloysius Hagendoorn. When Ways of Life Collide: Multiculturalism and Its Discontents in the Netherlands. Princeton, NJ: Princeton University Press, 2007. Stark, Rodney, and Roger Finke. Acts of Faith: Explaining the Human Side of Religion. Berkeley: University of California Press, 2000. Stark, Rodney, and Charles Y. Glock. American Piety: The Nature of Religious Commitment. Vol. 1. Berkeley: University of California Press, 1968. Stouffer, Samuel Andrew. Communism, Conformity, and Civil Liberties: A CrossSection of the Nation Speaks Its Mind. New Brunswick, NJ: Transaction, 1955. Sullivan, John L., George E. Marcus, Stanley Feldman, and James E Piereson. “The Sources of Political Tolerance: A Multivariate Analysis.” American Political Science Review 75, no. 1 (1981): 92–106. Swoyer, Alex. “Donald Trump: This Whole Election Is Being Rigged.” Breitbart.com, October 14, 2016. http://www.breitbart.com/2016-presidential -race/2016/10/14/donald-trump-this-whole-election-is-being-rigged/. Taber, Charles S., and Milton Lodge. “Motivated Skepticism in the Evaluation of Political Beliefs.” American Journal of Political Science 50, no. 3 (2006): 755–769. Tajfel, Henri T., and John C. Turner. “An Integrative Theory of Intergroup Conflict.” Social Psychology of Intergroup Relations 33, no. 47 (1979): 40–74. ———. “The Social Identity Theory of Intergroup Behavior.” Pp. 7–24 in Psychology of Intergroup Relations, edited by Stephen Worchel and William G. Austin. Chicago: Nelson-Hall, 1986. Thomas, Sue. “The Impact of Women on State Legislative Policies.” Journal of Politics 53, no. 4 (1991): 958–976. Tocqueville, Alexis de. Democracy in America. 4 vols. London: Saunders and Otley, 1835. Turner, John C., Rupert J. Brown, and Henri Tajfel. “Social Comparison and Group Interest in Ingroup Favouritism.” European Journal of Social Psychology 9, no. 2 (1979): 187–204. Van Den Bos, Wouter, Philipp J. M. Golka, David Effelsberg, and Samuel M. McClure. “Pyrrhic Victories: The Need for Social Status Drives Costly Competitive Behavior.” Frontiers in Neuroscience (October 23, 2013). https://doi .org/10.3389/fnins.2013.00189. Verkuyten, Maykel. The Social Psychology of Ethnic Identity. London: Psychology Press, 2004. Verkuyten, Maykel, and Ali Aslan Yildiz. “National (Dis)Identification and Ethnic and Religious Identity: A Study Among Turkish-Dutch Muslims.” Personality and Social Psychology Bulletin 33, no. 10 (2007): 1448–1462. Wald, Kenneth D. “Assessing the Religious Factor in Electoral Behavior.” Pp. 105– 22 in Religion in American Politics, edited by Charles W. Dunn. Washington, DC: Congressional Quarterly Press, 1989. Wald, Kenneth D., and Allison Calhoun-Brown. Religion and Politics in the United States. Lanham, MD: Rowman and Littlefield, 2014. Wald, Kenneth D., Dennis E. Owen, and Samuel S. Hill. “Churches as Political Communities.” American Political Science Review 82, no. 2 (1988): 531–548. Wald, Kenneth D., and Corwin E. Smidt. “Measurement Strategies in the Study of Religion and Politics.” Pp. 26–49 in Rediscovering the Religious Factor in

184

References

American Politics, edited by David C. Leege and Lyman A. Kellstedt. Armonk, NY: M. E. Sharpe, 1993. Weber, Christopher, and Matthew Thornton. “Courting Christians: How Political Candidates Prime Religious Considerations in Campaign Ads.” Journal of Politics 74, no. 2 (2012): 400–413. West, Emily A. “Descriptive Representation and Political Efficacy: Evidence from Obama and Clinton.” Journal of Politics 79, no. 1 (2017): 351–355. Wilcox, Clyde. “The Fundamentalist Voter: Politicized Religious Identity and Political Attitudes and Behavior.” Review of Religious Research 31, no. 1 (1989): 54–67. ———. “Religion and Politics Among White Evangelicals: The Impact of Religious Variables on Political Attitudes.” Review of Religious Research 32, no. 1 (1990): 27–42. Wilcox, Clyde, Ted G. Jelen, and David C. Leege. “Religious Group Identifications: Toward a Cognitive Theory of Religious Mobilization.” Pp. 72–99 in Rediscovering the Religious Factor in American Politics, edited by David C. Leege and Lyman A. Kellstedt. Armonk, NY: M. E. Sharpe, 1993. Yemelianova, Galina M. “Islam, National Identity, and Politics in Contemporary Kazakhstan.” Asian Ethnicity 15, no. 3 (2014): 286–301.

Index

Access Hollywood (TV Show), 99–100, 102 Acquaintances: for atheists, 92–93; beliefs and, 133–134; in community, 95–96; in culture, 78–79, 79fig; for minorities, 80–81, 80fig Activism, 36, 81–82, 81fig Affective polarization, 87–89 Affiliation, 20, 23–24, 40 African Americans, 52 Age, 55 Antipathy. See Out-group antipathy; Tolerance Approval ratings: for government, 52; for Obama, 45, 53–55, 56fig–57fig, 58; partisanship in, 60; in politics, 43–44, 44fig; for presidents, 46fig Atheists: acquaintances for, 92–93; Christians and, 63, 75–82, 79fig–81fig; Hindus and, 26; in-group favoritism for, 91–96, 91fig, 93fig–95fig; as threats, 64–65; in US, 11 Attitude: behavior and, 76; church attendance and, 118–119; motivation and, 38–39; politics and, 60, 100; for religion, 121–124, 122fig–124fig; in surveys, 101–102 Behavior: attitudes and, 76; beliefs and, 16–19; ideology and, 67; memories and, 100; for religion, 120–121, 120tab; social behavior, 19–20, 68, 87, 89–90, 93fig, 111–112; from social identity, 23– 27, 24tab

Beliefs: acquaintances and, 133–134; behavior and, 16–19; from church attendance, 121–122; of Congress, 37, 39–40, 41tab; in culture, 115–116; data and, 100–101; education and, 27–28, 53; faith in, 35; from ideology, 52; for minorities, 123–124, 123fig–124fig; politics and, 8–9, 126–128, 127fig; prejudice against, 10–11; psychology of, 3–4, 19–20; public policy and, 58; Reiki beliefs, 7; in religion, 113–114, 126– 128, 127fig; in social identity, 3–4, 16–17 Belonging, 17–18 Buddhists, 84n32. See also Religion California, 118 Castro, Martin, 131 Catholics, 6, 106. See also Christians; Religion CCES. See Cooperative Congressional Election Study Christians: atheists and, 63, 75–82, 79fig– 81fig; church attendance by, 38; culture of, 1–2, 66; Jews and, 26; lobbying by, 18; politics for, 8; as Republicans, 4–5; tribalism for, 20–21. See also Religion Church attendance: attitude and, 118–119; beliefs from, 121–122; by Christians, 38; education and, 113–114; ethics and, 18–19 Church of Jesus Christ of Latter-day Saints (LDS), 4–7; data on, 120–124, 120tab,

185

186

Index

122fig–124fig; sexual orientation for, 26; social identity for, 124–126, 125fig; in surveys, 117–120, 119tab; in US, 11– 12 Clinton, Bill, 99–100 Coding, for data, 39–40 Cognitive psychology, 22–23 Communication: for Democrats, 109–111, 122–123, 123fig; of identity, 21–22; of religion, 108–109; for Republicans, 109–111 Communism, 63 Community: acquaintances in, 95–96; of Catholics, 6; data on, 86–87; faith in, 9, 23, 28; politics in, 46–47, 46fig; psychology of, 37–38, 77–78, 86, 132– 133; religion in, 53; social behavior in, 89–90; in social identity, 7, 55; in surveys, 80–81, 81fig; threats for, 24– 25, 24tab, 65 Competing identities hypothesis, 26–27, 30 Congress, 35–40, 41tab, 133–134. See also Politics Conservatives, 66–67, 119–120, 119tab, 134–135, 138. See also Republicans Cooperative Congressional Election Study (CCES), 39–40, 41tab, 42–47, 44fig, 46fig Culture: acquaintances in, 78–79, 79fig; beliefs in, 115–116; of Christians, 1–2, 66; of conservatives, 138; data on, 103– 106, 104fig, 107fig, 108–111, 110fig; of Democrats, 70–71; faith in, 9, 20; government and, 53; hierarchies in, 25– 26, 28–29; of Hindus, 75; information in, 101–102; of liberals, 66–67; media for, 15–16; politics and, 96; of presidents, 99–100; religion in, 99–102; of Republicans, 73–74, 74fig; ritual practice in, 16–17; social identity and, 3, 38–39, 86; in surveys, 64, 77; thermometer rating for, 70; tolerance and, 131; of US, 1, 63, 68, 113 Data: beliefs and, 100–101; coding for, 39– 40; on community, 86–87; on culture, 103–106, 104fig, 107fig, 108–111, 110fig; on identity, 124–126, 125fig; on independents (voters), 110fig; on LDS, 120–124, 120tab, 122fig–124fig; on race, 68–69; for scholarship, 76–77; on sexual orientation, 129n13; from surveys, 4, 39–40, 41tab, 42–47, 44fig,

46fig, 90–91; on trust, 53–55, 56fig– 57fig, 58, 59fig, 60 Democracy, 52 Democrats: communication for, 109–111, 122–123, 123fig; culture of, 70–71; minorities and, 85–86; nationalism for, 7; partisanship for, 43, 54–55; religion for, 58, 70tab, 118–119; Republicans and, 2, 15, 19, 51, 66, 81, 88, 92–93, 106, 107fig, 108, 111–112, 115–116, 125, 135, 136fig–137fig, 138; in surveys, 60, 69–70, 69tab; tolerance for, 105, 134. See also Partisanship Descriptive representation, 37–38, 53–54, 58 Economics, 18 Education, 27–28, 42, 53, 113–114 Elected officials. See Congress; Government; Presidents Elitism, 16, 36 Empirical expectations, 29–30, 39tab Estimation strategies, 54 Ethics, 16–19, 54. See also Sexual orientation Europe, 28 Evidence. See Data Exclusion, 26 Facts. See Data Faith: in beliefs, 35; in community, 9, 23, 28; in culture, 9, 20; in information, 114; psychology of, 86; in public policy, 52–53; in religion, 3; in surveys, 117– 118 Family, 4–5, 8, 38, 75–76, 129n13 Gandhi, Mahatma, 85 Gays. See Sexual orientation Gender, 40, 42, 45 Gibson, James, 65 Globalization, 85, 102, 131–132 Government, 52–54, 60 Group identification. See Social identity Group-based cues, 22 Heuristics, 30 Hierarchies, 25–29, 44–45, 123–124, 123fig Hindus, 2, 26–28, 40, 75. See also Religion House of Representatives. See Congress Identity: affiliation and, 20; for African Americans, 52; age in, 55;

Index communication of, 21–22; competing identities hypothesis, 26–27, 30; data on, 124–126, 125fig; education in, 42; exclusion in, 26; gender in, 45; for Islam, 89; methodology of, 66; for minorities, 75; motivation and, 25, 101– 102; nationalism in, 6; partisanship in, 39, 59fig, 67–74, 69tab–70tab, 71fig– 74fig; Pew Research Center on, 1, 103–106, 104fig, 107fig, 108–111, 110fig; in politics, 54, 113–114, 126– 128, 127fig; prejudice and, 9; psychology of, 8, 23–27, 24tab, 99; public policy and, 17; race in, 55; religion and, 2, 21–22, 47–48, 84n32, 91–96, 91fig, 93fig–95fig, 131–135, 136fig–137fig, 138–139; scholarship on, 16; sexual orientation in, 4–5; social behavior and, 68; on social media, 7; in surveys, 40. See also Social identity Ideology, 15–16, 42, 52, 67, 88–89, 133 Independents (voters), 104–105, 104fig, 110fig; psychology of, 135, 136fig– 137fig, 138; Republicans and, 125–126, 125fig; in US, 42, 114 Information: in culture, 101–102; faith in, 114; globalization of, 131–132; politics of, 100–101; psychology of, 115–116; social identity and, 116–117, 117tab; on war, 116. See also Systematic information processing In-group favoritism, 25, 51, 91–96, 91fig, 93fig–95fig Islam: globalization for, 102; for Hindus, 27–28; identity for, 89; politics and, 53– 54; for Republicans, 54–55; in surveys, 105; in US, 45, 49n30, 76–77. See also Religion Jehovah’s Witnesses, 40 Jews, 26. See also Religion Ku Klux Klan, 66, 116–117, 117tab LDS. See Church of Jesus Christ of Latterday Saints Leadership, 36, 120–121, 131–132, 138– 139 Lesbians. See Sexual orientation Lewinsky, Monica, 99–100 Liberals, 66–67, 71–72, 71fig–72fig, 119– 120, 119tab, 122fig. See also Democrats Lobbying, 18

187

Loyalty, 68 Lutherans. See Christians Media, 15–16, 99–100, 102 Membership, 24 Memories, 100 Methodology, 19–21, 66 Millennials, 1–2 Minorities: acquaintances for, 80–81, 80fig; beliefs for, 123–124, 123fig–124fig; conservatives and, 66–67; Democrats and, 85–86; identity for, 75; leadership for, 120–121; Native Americans, 40; politics and, 36–38, 82, 91–92, 91fig; psychology of, 47–48, 109–111, 110fig; in religion, 66; in US, 90, 102–103 Mormons. See Church of Jesus Christ of Latter-day Saints Motivation, 25, 38–39, 54, 65, 89, 101–102 Muslims. See Islam Nationalism, 5–7 Native Americans, 40 Nonpolitical activities, 93–96, 94fig, 95tab Obama, Barack: approval ratings for, 45, 53–55, 56fig–57fig, 58; faith in, 9; religion and, 58, 59fig, 60; social identity for, 46, 51–53; in surveys, 49n30; trust for, 58, 59fig, 60 Obergefell v. Hodges, 4 Out-group antipathy, 25–26, 88–89 Partisanship: activism and, 81–82, 81fig; affective polarization in, 87–89; in approval ratings, 60; for Democrats, 43, 54–55; in family, 8, 129n13; gender and, 40, 42; in identity, 39, 59fig, 67– 74, 69tab–70tab, 71fig–74fig; ideology and, 15–16, 88–89, 133; psychology of, 27, 90–96, 91fig, 93fig–95fig; public policy and, 116–117, 117tab; religion and, 28–29, 36, 102–106, 103tab, 104fig, 107fig, 108–111, 110fig; for Republicans, 43, 134; scholarship on, 21, 87–88; in social identity, 11–12, 42– 43, 99–102, 103tab, 111–112; as tribalism, 10, 16, 36, 51–52; in US, 4, 28–29, 131–135, 136fig–137fig, 138– 139 Pew Research Center, 1, 85, 103–106, 104fig, 107fig, 108–111, 110fig Polarization, 87–89. See also Tribalism

188

Index

Politics: activism in, 36; of affiliation, 23, 40; approval ratings in, 43–44, 44fig; attitudes and, 60, 100; beliefs and, 8–9, 126–128, 127fig; in California, 118; CCES on, 39–40, 41tab, 42–47, 44fig, 46fig; for Christians, 8; in community, 46–47, 46fig; culture and, 96; economics of, 18; heuristics in, 30; identity in, 54, 113–114, 126–128, 127fig; ideology for, 42; of information, 100–101; Islam and, 53–54; leadership in, 36; liberals in, 66; loyalty in, 68; in media, 99–100; of Millennials, 1–2; minorities and, 36–38, 82, 91–92, 91fig; of nationalism, 5; of Native Americans, 40; nonpolitical activities, 93–96, 94fig– 95tab; “The Political Consequences of Religiosity” (Gibson), 65; psychology of, 75–76, 116–117, 117tab; of race, 56fig; of reconciliation, 131–135, 136fig–137fig, 138–139; religion in, 3, 10–11, 64–65, 85–86; of religious persecution, 27–28; semipolitics, 91–93, 91fig, 93fig; of sexual orientation, 5–6, 11–12; social behavior and, 93fig; social identity in, 10, 91fig; Tea Party in, 72; tolerance in, 64–65, 88; trust in, 51–53; in US, 38; of war, 115 Praying, 19 Prejudice, 9, 10–11, 76–77, 89, 135 Presidents, 45–46, 46fig, 51, 99–100, 133– 134. See also Obama Protestants. See Christians Psychology, 60; of beliefs, 3–4, 19–20; belonging, 17–18; cognitive psychology, 22–23; of community, 37– 38, 77–78, 86, 132–133; empirical expectations, 29–30, 30tab; exclusion, 26; of faith, 86; group-based cues, 22; hierarchies in, 29; of identity, 8, 23–27, 24tab, 99; of independents (voters), 135, 136fig–137fig, 138; of information, 115–116; of minorities, 47–48, 109– 111, 110fig; motivation, 25, 38–39, 54, 65, 89, 101–102; of partisanship, 27, 90–96, 91fig, 93fig–95fig; of politics, 75–76, 116–117, 117tab; religion and, 10, 83n13; of social identity, 27–28, 63; of sports, 67–68; of systematic information processing, 23–24, 29–30. See also Attitude Public policy: beliefs and, 58; descriptive representation in, 37–38; faith in, 52–

53; identity and, 17; partisanship and, 116–117, 117tab Race, 55, 56fig, 68–69, 83n13. See also Minorities; specific races Reconciliation, 131–135, 136fig–137fig, 138–139 Reiki beliefs, 7 Religion: attitude for, 121–124, 122fig– 124fig; behavior for, 120–121, 120tab; beliefs in, 113–114, 126–128, 127fig; communication of, 108–109; in community, 53; for conservatives, 119– 120, 119tab; in culture, 99–102; for Democrats, 58, 70tab, 118–119; faith in, 3; from family, 38; in globalization, 85; hierarchies in, 123–124, 123fig; identity and, 2, 21–22, 47–48, 84n32, 91–96, 91fig, 93fig–95fig, 131–135, 136fig– 137fig, 138–139; in in-group favoritism, 51; leadership in, 120–121, 131–132; minorities in, 66; Obama and, 58, 59fig, 60; partisanship and, 28–29, 36, 102– 106, 103tab, 104fig, 107fig, 108–111, 110fig; Pew Research Center on, 85; “The Political Consequences of Religiosity” (Gibson), 65; in politics, 3, 10–11, 64–65, 85–86; for presidents, 45–46; psychology and, 10, 83n13; for Republicans, 58, 60, 70tab, 85–86; scholarship on, 8–9; sexual orientation and, 73, 73fig, 118–119; social behavior and, 111–112; in social identity, 8–9, 22–23, 44–45, 55, 56fig–57fig, 58, 59fig, 65–74, 69tab–70tab, 71fig–74fig, 89–90; social networks for, 18; in surveys, 86–87; threats to, 76–82, 79fig–81fig; in US, 92, 96. See also specific topics Religious persecution, 27–28 Republicans: Christians as, 4–5; communication for, 109–111; culture of, 73–74, 74fig; Democrats and, 2, 15, 19, 51, 66, 81, 88, 92–93, 106, 107fig, 108, 111–112, 115–116, 125, 135, 136fig– 137fig, 138; independents (voters) and, 125–126, 125fig; Islam for, 54–55; partisanship for, 43, 134; religion for, 58, 60, 70tab, 85–86; in surveys, 69–70, 69tab, 105–106; Tea Party and, 70–71; Trump for, 99–100. See also Partisanship Ritual practice, 16–17

Index Scholarship: competing identities hypothesis, 26–27; data for, 76–77; on identity, 16; in-group favoritism, 25, 51, 91–96, 91fig, 93fig–95fig; methodology of, 19–21; out-group antipathy, 25–26, 88–89; on partisanship, 21, 87–88; prejudice in, 76, 135; on religion, 8–9; on social identity, 35–36; surveys for, 12; on tolerance, 64 Semipolitical, 91–93, 91fig, 93fig Sexual orientation: data on, 129n13; in identity, 4–5; for LDS, 26; for liberals, 71–72, 71fig, 71fig–72fig, 119–120, 119tab, 122fig; politics of, 5–6, 11–12; religion and, 73, 73fig, 118–119; in US, 74, 78, 126–128, 127fig Social behavior, 19–20, 68, 87, 89–90, 93fig, 111–112 Social identity: behavior from, 23–27, 24tab; beliefs in, 3–4, 16–17; community in, 7, 55; in Congress, 38– 39; culture and, 3, 38–39, 86; in Europe, 28; family and, 4–5; family for, 75–76; hierarchies for, 27; information and, 116–117, 117tab; for LDS, 124–126, 125fig; leadership and, 138–139; for Obama, 46, 51–53; partisanship in, 11– 12, 42–43, 99–102, 103tab, 111–112; in politics, 10, 91fig; psychology of, 27– 28, 63; race and, 83n13; religion in, 8–9, 22–23, 44–45, 55, 56fig–57fig, 58, 59fig, 65–74, 69tab–70tab, 71fig–74fig, 89–90; scholarship on, 35–36; surveys on, 103–106, 104fig, 107fig, 108–111, 110fig; as threats, 29, 75–76; tolerance and, 135 Social media, 7 Social networks, 18 Sports, 67–68 Surveys: attitude in, 101–102; Buddhists in, 84n32; community in, 80–81, 81fig; culture in, 64, 77; data from, 4, 39–40, 41tab, 42–47, 44fig, 46fig, 90–91; Democrats in, 60, 69–70, 69tab; descriptive representation in, 53–54; estimation strategies for, 54; faith in, 117–118; hierarchies in, 44–45; identity in, 40; independents (voters) in, 104– 105, 104fig; Islam in, 105; LDS in,

189

117–120, 119tab; Obama in, 49n30; praying in, 19; questions in, 52–53; religion in, 86–87; Republicans in, 69– 70, 69tab, 105–106; for scholarship, 12; on social identity, 103–106, 104fig, 107fig, 108–111, 110fig; thermometer rating in, 70–74, 71fig–74fig; in US, 68, 85, 129n13 Systematic information processing, 23–24, 29–30 Tea Party, 70–72 Thermometer rating, 70–74, 71fig–74fig Threats: atheists as, 64–65; for community, 24–25, 24tab, 65; to religion, 76–82, 79fig–81fig; social identity as, 29, 75– 76 Tolerance: culture and, 131; for Democrats, 105, 134; motivation for, 65; out-group antipathy and, 88–89; in politics, 64–65, 88; scholarship on, 64; social identity and, 135; in US, 63, 78, 82 Tribalism, 20–21; partisanship as, 10, 16, 36, 51–52; in US, 8, 35, 60 Trump, Donald, 99–100, 134–135 Trust: data on, 53–55, 56fig–57fig, 58, 59fig, 60; in government, 60; for Obama, 58, 59fig, 60; in politics, 51–53; social behavior and, 87 United States (US): atheists in, 11; Catholics in, 106; culture of, 1, 63, 68, 113; Democracy in, 52; descriptive representation in, 58; elitism in, 36; Hindus in, 2, 40; independents (voters) in, 42, 114; Islam in, 45, 49n30, 76–77; LDS in, 11–12; minorities in, 90, 102– 103; partisanship in, 4, 28–29, 131–135, 136fig–137fig, 138–139; politics in, 38; prejudice in, 77; presidents in, 51; religion in, 92, 96; sexual orientation in, 74, 78, 126–128, 127fig; surveys in, 68, 85, 129n13; tolerance in, 63, 78, 82; tribalism in, 8, 35, 60 Vietnam War, 52 Voters, 54 War, 115, 116

About the Book

affiliation on political attitudes and behaviors in the United States, a number of puzzling questions remain unanswered. In response, Matthew Miles demonstrates that a more complete conceptualization of religion as a social identity can help to explain many of those puzzles. As he explores the impact, both positive and negative, of religious identity on political attitudes, he also shows that the religion-politics relationship is not a one-way street.

While existing scholarship addresses the influence of religious

Matthew R. Miles is associate professor of political science at Brigham Young University–Idaho.

190