Everyday Eating in Denmark, Finland, Norway and Sweden: A Comparative Study of Meal Patterns 1997–2012 9781350080485, 9781350080454, 9781350080461

Citation preview

Everyday Eating in Denmark, Finland, Norway and Sweden

i

Also available from Bloomsbury: Caviar with Champagne, Jukka Gronow Food, Masculinities and Home, edited by Michelle Szabo and Shelley L. Koch Making Taste Public, edited by Carole Counihan and Susanne Hojlund Taste, Politics, and Identities in Mexican Food, edited by Steffan Igor Ayora-Diaz

ii

Everyday Eating in Denmark, Finland, Norway and Sweden A Comparative Study of Meal Patterns 1997–2012 Edited by Jukka Gronow and Lotte Holm

iii

BLOOMSBURY ACADEMIC Bloomsbury Publishing Plc 50 Bedford Square, London, WC1B 3DP, UK 1385 Broadway, New York, NY 10018, USA BLOOMSBURY, BLOOMSBURY ACADEMIC and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2019 Copyright © Jukka Gronow, Lotte Holm and Contributors 2019 For legal purposes the Acknowledgements on p. ix constitute an extension of this copyright page. Jukka Gronow and Lotte Holm have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as Editors of this work. Cover image: Bread and cheese on rustic table. (© Vladislav Nosick / Alamy Stock Photo) All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. Bloomsbury Publishing Plc does not have any control over, or responsibility for, any third-party websites referred to or in this book. All internet addresses given in this book were correct at the time of going to press. The author and publisher regret any inconvenience caused if addresses have changed or sites have ceased to exist, but can accept no responsibility for any such changes. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. ISBN:

HB: ePDF: ePub:

978-1-3500-8048-5 978-1-3500-8046-1 978-1-3500-8047-8

Typeset by RefineCatch Limited, Bungay, Suffolk To find out more about our authors and books visit www.bloomsbury.com and sign up for our newsletters

iv

Contents Notes on Contributors Preface Acknowledgements

vi viii ix

1

Introduction: Eating in Modern Everyday Life Lotte Holm and Jukka Gronow 2 The Food We Eat Lotte Holm, Drude Skov Lauridsen, Jukka Gronow, Nina Kahma, Unni Kjærnes, Thomas Bøker Lund, Johanna Mäkelä, and Mari Niva 3 The Daily Rhythm of Eating Thomas Bøker Lund and Jukka Gronow 4 The Social Context and Conduct of Eating Lotte Holm, Drude Lauridsen, Thomas Bøker Lund, Jukka Gronow, Mari Niva and Johanna Mäkelä 5 Family Meals on the Decline? Lotte Holm, Thomas Bøker Lund, Drude Lauridsen, and Jukka Gronow 6 The Complexity of Meals Nina Kahma, Johanna Mäkelä, Mari Niva, and Thomas Bøker Lund 7 Eating Out, Having Guests Thomas Bøker Lund 8 Cooking and Gender Lotte Holm, Marianne Pipping Ekström, Sara Hach, and Thomas Bøker Lund 9 Food Insecurity Anita Borch and Unni Kjærnes 10 Eating Practices and Dietary Health Lotte Holm, Thomas Bøker Lund, and Mari Niva 11 Eating Sustainably Mari Niva, Johanna Mäkelä, Nina Kahma, and Unni Kjærnes 12 Conclusions: Continuity and Change in Everyday Eating Lotte Holm and Jukka Gronow

Appendix Thomas Bøker Lund and Mari Niva References Index

1

15 33

57 77 93 109 123 141 159 173 191 201 211 235

v

Contributors Anita Borch holds positions as Head of Research and Research Professor at the Consumption Research Norway (SIFO) at Oslo Metropolitan University. Her scientific publications cover a range of consumer-related subjects, including food-related insecurity in the academic discourse and in the Nordic countries. Marianne Pipping Ekström is a sociologist, docent at the Department of Food and Nutrition and Sport Science at University of Gothenburg, Sweden. She researches food habits, food culture and the division of labour in households, both nationally and internationally, mostly with a gender perspective. Jukka Gronow is a professor emeritus of sociology at Uppsala University, Sweden, and a docent at the University of Helsinki, Finland. In addition to food studies, he has published on the sociology of consumption, history of sociology, social theory and the social and cultural history of the Soviet Union. Sara Hach was research assistant at the Department of Food and Resource Economics at University of Copenhagen, Denmark in the period 2012–13 and was involved in data management and analyses. She works as a consultant at the Danish Evaluation Institute. Lotte Holm is a professor of sociology of food at the Department of Food and Resource Economics at University of Copenhagen, Denmark. She researches food consumption and eating and food related policies, often in large multidisciplinary and comparative projects. Nina Kahma is a university researcher at the Centre for Consumer Society Research at the University of Helsinki. She has studied inequality and consumption from different angles, and her current interest concerns service use and design. Unni Kjærnes is a research professor at Oslo Metropolitan University, Norway, and a docent at the University of Helsinki, Finland. Her research has dealt with the consumption of food, but also other consumption areas, often with a focus on policy related issues. vi

Contributors

vii

Drude Skov Lauridsen is a PhD student at the Department of Food and Resource Economics at University of Copenhagen, Denmark. She is writing a thesis on the maternal responsibility for children’s feeding in relation to the stigmatization of maternal obesity in childhood obesity prevention. Thomas Bøker Lund is Associate Professor of sociology of food and rural sociology at the Department of Food and Resource Economics at University of Copenhagen, Denmark. He has conducted research about political food consumption and food-related poverty, among others. Johanna Mäkelä is Professor of food culture at the University of Helsinki and the Dean of the Faculty of Educational Sciences. She researches meals, future food consumption, responsible food chain, food classifications and practices in everyday life contexts, Nordic eating habits, and fat wars. Mari Niva is a professor of food culture (2018–21) at the University of Helsinki. She studies food choices and practices and the social and cultural aspects of eating. Currently her research focuses on sustainable and political food consumption.

Preface This book reports from a collaboration between sociologists from four Nordic countries, Denmark, Finland, Norway and Sweden, which has stretched over almost twenty years. Results from two studies conducted by the collective of authors with fifteen years apart are presented: A Day of Food in Nordic Countries, A comparative investigation of eating habits in modern everyday life (1997), and Food in Nordic Everyday Life: A comparative survey of change and stability in eating patterns (2012). The idea that eating is a social phenomenon that is core to wider issues of social structure and daily life has motivated these studies and acted as the general guideline in the comparative analysis of eating patterns and their change over time. The study design was devised together by the project partners: Marianne Pipping Ekström, University of Gothenburg; Jukka Gronow, University of Helsinki; Lotte Holm, University of Copenhagen; Unni Kjærnes, National Institute for Consumer Research, Oslo; Thomas Bøker Lund, University of Copenhagen; Johanna Mäkelä, University of Helsinki; and Mari Niva, National Consumer Research Centre/University of Helsinki. Unni Kjærnes coordinated the study in 1997 and Lotte Holm the study in 2012. Results from the first study from 1997 were published in the book: Unni Kjærnes (ed) (2001) Eating Patterns. A Day in the Lives of Nordic Peoples. Oslo: National Institute for Consumer Research.

viii

Acknowledgements The two studies reported in the book were supported financially by the Joint Committee for Nordic Research Councils for the Social Sciences (NOS-S, 1997) and for the Humanities and the Social Sciences (NOSHS, 2012). Several researchers have contributed to the analyses, and the project partners are grateful to Nina Kahma from the University of Helsinki, Anita Borch from the National Institute for Consumer Research, Oslo, and Sara Hach and Drude Schou Lauridsen from the University of Copenhagen. Some of the chapters in this book are slightly revised versions of articles that have been published earlier. The authors and publishers would like to thank the following publishers for permission to reprint: Anthropology of Food for: Kahma, N., Mäkelä, J., Niva, M., and Lund, T.B., 2014. Associations between meal complexity and social context in four Nordic countries. Anthropology of food, S10, pp. 1–22. Elsevier for: Lund, T.B. & Gronow, J., 2014. Destructuration or continuity? The daily rhythm of eating in Denmark, Finland, Norway and Sweden in 1997 and 2012 (2014). Appetite, 82, pp. 143–53; and for: Holm, L., Lauridsen, D., Lund, T.B., Gronow, J., Niva, M., Mäkelä, J., 2016. Changes in the social context and conduct of eating in four Nordic countries between 1997 and 2012. Appetite, 103, pp. 358–68. Gothenburg University for: Holm, L., Lauridsen, D.S., Gronow, J., Kahma, N., Kjærnes, U., Lund, T.B., Mäkelä, J., Niva, M., 2015. The food we eat in Nordic countries – some changes between 1997 and 2012. In K. Bergström et al., eds. Mat är mer än mat. Göteborg: Göteborg Universitet, pp. 227–46. Springer for: Niva, M., Mäkelä, J., Kahma, N., and Kjærnes, U., 2014. Eating Sustainably? Practices and Background Factors of Ecological Food Consumption in Four Nordic Countries. Journal of Consumer Policy, 37(4), pp. 465–84. Springer Nature for: Holm, L., Lund, T.B. and Niva, M., 2015. Eating practices and diet quality: a population study of four Nordic countries. European Journal of Clinical Nutrition, 69(7), pp. 791–8.

ix

x

Acknowledgements

Taylor & Francis Ltd, www.tandfonline.com on behalf of The Association for the Study of Food & Society (ASFS) for: Holm, L., Ekström, M.P., Hach, S., Lund, T.B., 2015. Who is Cooking Dinner? Food, Culture & Society, 18(4), pp. 589–610. The authors would like to thank anonymous reviewers of the original articles for valuable contributions to the text.

1

Introduction Eating in Modern Everyday Life Lotte Holm and Jukka Gronow

Judging by the widespread coverage that food issues enjoy in the mass media today, one might imagine that eating is a major concern of modern men and women. On an almost daily basis, journalists present news of spectacular scientific discoveries about healthy or harmful food items and ingredients. They advise on ‘how to lose weight’ and on other dietary instructions, and inform consumers on the quality and price of food. In reviews, they report back on their visits to new restaurants and exotic places with food cultures of their own, and on famous restaurants, their chefs, and their spectacular achievements. Food and eating are repeatedly being scrutinized as a matter of ethics and sustainability. While some focus on the responsibility of people as consumers, others emphasize the risks and uncertainties that people face as buyers and eaters of food. It is an interesting question why food and eating are so interesting to the public today, and why they are favourite topics in the modern mass media. Food and eating certainly combine many aspects of human life, from care for others to the care of the self, from the future of our globe to the well-being of our bodies. Food habits and taste divide people and bring them together. At times, moral concerns combined with social and economic worries arise over various topics: the dissolution of family life as a consequence of the decline of family meals; the loss of food competence with the disappearance of housewives; and the rapid advance of the fast food industry. Food scares and scandals are also an almost permanent companion to our modern life. At times, it looks as though almost every act of eating, every piece of food we put into our mouth, has political, economic and ethical dimensions. As an object of scientific research in, for example, nutritional science, food consumption has acted primarily as a potential factor in good and poor health, lengthening or shortening our expected life span. Socio-political interest in food 1

2

Everyday Eating

consumption has traditionally concentrated on issues of poverty and hunger, but more recently, and increasingly so, on the impact of food production on the planet – on sustainability, in other words. Over recent decades, a sociological food research tradition has emerged that treats food and eating mainly as phenomena of social life which offer us one of the best means to study the overall organization of daily life in terms of time and place, social relations and rules of conduct, and the development of these over time. The social organization of eating is in many ways related to developments in other social institutions such as family, work and education. Food practices and their historical evolution are, in other words, a good indicator of what is going on in society. The project presented in this book follows this tradition, concentrating on the mundane, or ordinary, aspects of eating and their historical shifts. The aim is to learn more about how daily eating and everyday practices linked to eating change in modern society. In this introduction, we first highlight a number of key issues that have been debated within social research into modern food and eating. We then describe the design of the project this book is about. Finally, we present a brief outline of the book.

Global or national food cultures? Historically, national and regional food cultures are formed by local geographical, economic and political conditions (Johansen 1998; Kisbán 1986; Mennel 2010; Pelto and Pelto 1983; Sobal 1999; Teuteberg 1986; Trygg 1991). In modern times, a development has been identified in which local and national food cultures are being eroded by an increasingly globalized, homogenizing food sector. This development, facilitated by global food companies, is often seen as an expression of our domination by modern American food culture, with its simple and often mass-produced meals (see Ritzer’s concept of ‘McDonaldization’) (Warde 1997; Ritzer 1993). However, a counter-trend has also been identified in which we see the revival and reinvention of local and regional food cultures (Grasseni 2011) and the emergence of countercultures prioritizing local provisioning, authenticity and fresh food. In this counter-trend, eating patterns are characterized by the consumption of less processed and convenience food, more regional foods, and more seasonal variation (Counihan and Siniscalchi 2014; Dowler, Kneafsey, Cox and Holloway 2010). Several researchers have suggested that these contrasting trends coexist, reflecting increasing social and economic distinctions in eating. Further, convergent discourses of health and nutrition

Introduction

3

may be seen as elements of the globalizing trend which introduce more uniform values of food consumption everywhere (Coveney 2006). While nutritional discourses are increasingly conflicting and competing per se – as witnessed by the frequent debates about the health status of different kinds of fat and familiar foodstuffs such as sugar, wheat, gluten and milk – they do have a potential impact on our diet as they mobilize not only public authorities, but also market actors and popular media. To what degree such globalizing trends are dominant in Nordic eating patterns is a matter we shall return to in the analysis we develop.

Rhythm of daily life: Changes in time and place of eating? How, where, when and with whom we eat, is influenced by the social organization of work and family life, and the geographic location of private homes and workplaces, as well as the rhythm and organization of the workday. This is demonstrated in Robert Rotenberg’s (1981) classical analysis of the change from a five-meal to a three-meal pattern following industrialization in Vienna. This change implied that the location of daily meals shifted: it had included the home, the workplace (which was situated close to or in the home), and public cafés, but came to include only the home and the workplace, now often placed at a distance from the home. The change also meant that the social company enjoyed in daily eating changed: it now included only family members and colleagues, no longer the personal friends or old school mates whom many men would formerly have met in cafés and pubs on an everyday basis (Rotenberg 1981). The decline in the number of eating events can be observed in many industrialized countries. An interesting question is what happens when societies shift from industrial to post-industrial practices. To what extent do people still share common eating times and can we identify typical national peaks in the daily eating frequency? In modern life, the daily organization of eating also depends on other institutional and practical arrangements in society, such as the size and composition of households, including the increasing number of single person households, the arrangement of public catering in the workplace and in educational institutions, and the accessibility of commercial venues for eating out. It has been suggested that the home may no longer be the primary location for meals, as they are increasingly taken in public places such as restaurants, cafés, fast-food outlets, and so forth, or in canteens (Dinkins 1992; Dumangan and Hackett 1995; Julier 2013; Mogelonsky 1998). If this is true, we would expect to find a decline in eating at home in the Nordic countries, and a rise in eating out.

4

Everyday Eating

Meals in social life: Disappearance of family meals? Eating is an important operator of social life and a primary social function (Fischler 2011). Meals in private households have been described as the medium by which families are created and recreated on a daily basis (DeVault 1991; Jackson 2009; Julier 2013), and family meals are seen as important arenas for the socializing of children into central cultural norms and values (Fischler 2011). Family meals are often associated with a range of moral virtues. For some time, writers in the mass media and scientific literature have been asking whether traditional, regular dining patterns and meal formats have given way to a more irregular and destructured style of eating in which the collective rules organizing temporal, social and spatial dimensions of eating are disappearing (Murcott 1995; Murcott 1997; see also Mestdag 2005 and Southerton, Diaz-Méndez and Warde 2012). The most popular form of this thesis points to the abandonment of family meals and has been advocated in one form or another for at least half a century. Regular meals eaten at home together with all, or most, family members have given way, it is said, to snacking and the consumption of fast food. Commercialization, globalization, and even ‘Americanization’, as well as the changing social status of women, are often mentioned as the main developments lying behind this process of erosion. Market researchers use the term ‘grazing’ (Caplan 1997) to describe this allegedly new, hurried way of eating. A hamburger and a Coke bought from a burger joint and enjoyed hurriedly while driving a car is perhaps the most stereotypical expression of grazing. In a similar vein, JeanPierre Poulain (2002; see also Poulain 2008) has referred to ‘vagabond feeding’ or ‘nibbling’. In Sweden, the term ‘frukostisering’ (meaning breakfastization) has been coined to refer to this trend (Ekström 1990). An increase in the number of meals people eat, and a decrease in meals taken with family members, would indicate changes towards erosion of traditional social eating in households.

Manners of eating: De-structuring or de-traditionalizing? In addition to, and in parallel with practical arrangements around family and working life, eating is shaped by strong forms of normative regulation. In all societies and cultural groups, models for proper eating exist that dictate not only what foods should and should not be eaten, where, when and by whom specific foods should be eaten, and how foods should be ordered in terms of

Introduction

5

their combination and sequence (Douglas 1975; Fischler 1988a; Murcott 1982), but also the very conduct and social context of eating. While in the medieval and early modern periods eating practices and manners varied greatly between social classes, such variations appear to be diminishing in modern times. This could in principle mean either that eating has become individualized or that it has become more homogenous. The phrase ‘civilizing process’ has been coined to describe historical changes taking place between the medieval period and the end of the nineteenth century, during that time, among other things, table etiquette became more formal and detailed, and then spread from the upper classes in France to elites across Europe, and gradually also to the lower social classes (Elias 2000). In a reverse development, identified as informalization occurring in the course of the twentieth century (Wouters 2007), table manners have become less formal and more liberal. This is sometimes seen as a result of diminishing class distinctions, and of a ‘trickle-up’ process linked to large-scale social mobility in which the ideals and norms of ascendant groups mix with those of the previously superior group. Still, even in the 1960s Bourdieu identified distinctive differences between the conduct of festive meals in working class and bourgeois families: an abundance and loose sequencing order in working class meals, witnessing informality and liberty; by contrast, the strict rhythm, formality and stylization of bourgeois meals, implying restraints, self-control and an emphasis on form and manner (Bourdieu 1984: 196). Following Wouters’ (2007) theory of informalization, however, it might be suggested that the rules of correct eating conduct, or table manners, are not as strict as they once were and allow for more individual variation and heterogeneity in behaviour. It has also been suggested that the meal is losing its significance as an event in its own right, since eating events increasingly take place simultaneously with other activities (such as walking down a street, working at a desk, or watching television) (Bugge and Døving 2000; Senauer et al. 1991; Andersen 1997). It is a common sociological observation, often repeated in the diagnoses of our times, that European cultures have gradually become de-traditionalized and individualized (Giddens 1990; Heelas et  al. 1996), and that social norms of behaviour, or etiquette, have become less formalized. Changes in eating practices could thus be a part of a more general trend toward greater relaxation of social rules and manners. If the traditional rules, norms and standards applying to good food and the conduct of meals are dissolving, the sociological literature has raised the question of what comes in their place (Warde 1997). Claude Fischler’s concept of gastro-anomie (Fischler 1988b; Fischler 2011) captures an extreme version of

6

Everyday Eating

the idea that eating patterns have become irregular and individualized and that cultural eating norms are disappearing. Fischler did, admittedly, set the bar high: his ideal society involved well-organized classical French cuisine, with its extensive three-course lunches and dinners prepared and eaten at home in accordance with traditional table manners. In many theories of modernization, the evolution of more relaxed norms is rather welcomed, as it extends the freedom of the individual and expands the cultural space of human action (Wouters 2007). However, with respect to food and meals, this evolution is often associated with the fear of loss of social cohesion and order (Andersen 1997; Fischler 2011; Mennel et al. 1992). In the following study we investigate possible trends towards de-traditionalization and informalization by analysing specific traits of eating events: duration, seating arrangements, parallel activities to eating.

Meals: Disruption of structure? Typically, studies of meals are based on the notion that eating has a certain structure and rules. In anthropological and sociological research, the meal has been identified as a structured social affair that typically distinguishes family and friends from other contacts (Mäkelä 2009). Meals are characterized by copiousness, ceremoniousness, and the structural and sensory qualities embedded in binary oppositions of savoury and sweet, hot and cold, and liquid and dry (Mäkelä 2009; Douglas 1975). More specifically, research on meals has focused on the structure, or grammar, of meals and other eating events. One central conceptualization has been that of the proper meal with definite and necessary components – a notion originating in British studies that is frequently applied in Nordic research (e.g. Douglas and Nicod 1974; Murcott 1982; Ekström 1990; Bugge and Døving 2000). From the perspective of meals as structured social affairs, the alleged individualization of eating points to the evaporation of structuring rules. There have been concerns not only that meals and their preparation are losing their significance in everyday life, but also that meals are losing their structure. With the increasing role of convenience foods, such as ready meals and semi-prepared meals, cold meals are replacing hot meals and people are spending less time eating meals than before. This is challenging the idea that a proper meal consisted of clear elements: centre, staple, vegetables, trimmings and bread (Mäkelä 2001). On the other hand, meals may well become simpler and yet still be regarded as ‘proper’ by those who consume them (e.g. Marshall and Anderson 2002).

Introduction

7

A typical shortcoming in existing research is that commensality, in terms of the temporal, social, and spatial setting of the meal, is usually studied separately from what is eaten (Domaneschi 2012). Yet the very combination of these different aspects is crucial to a proper understanding of the changes in our eating habits. It is evident that across countries and cultures there are differences in, for instance, what types of foods are eaten at mealtimes, whether the food is hot or cold, the number of ingredients and components eaten, the sequences of different dishes, the meal venue, and those with whom meals are typically eaten (Warde 2005; Warde et al. 2007). The structure and content of meals must therefore be part of the analysis of how modern eating habits change.

Researching de-traditionalization and disruption of eating patterns It is, after all, an empirical question whether today’s daily eating habits resemble the more traditional, home-based, family-centred ideal, or are increasingly characterized by individualization, grazing, and gastro-anomy (cf. Jackson et al., 2009). Yet, with some exceptions (Poulain 2002; Mestdag 2005; Mandemakers and Roeters 2015), empirical studies to date looking into the matter are sparse. One reason for this may be that it is difficult to examine the issue in sufficient detail using typical questionnaire studies. This is because terms such as ‘grazing’ and ‘vagabond feeding’, and the meanings they convey, point to simultaneous changes in several factors, including shared temporal rhythms (grazing and vagabond feeding), spatial coordination (grazing and vagabond feeding), social meals (individualization; eating alone), and the ability of cultures to impose norms on cuisine (gastro-anomy). To investigate all, or even most, of these aspects of eating rather comprehensive data sets are needed in which microbehavioural aspects of everyday life are tracked in detail. Methods such as timeuse studies and 24-hour recall studies of the sort common in nutrition research would be appropriate for this purpose. Using empirical data, time-use studies have revealed some general trends (e.g. a decline in the time devoted to cooking), but they have also shown that trends in the time allocated to eating in the home and outside the home vary from country to country (Warde et  al. 2007; Cheng, Olsen, Southerton and Warde 2007; Mestdag 2005; Mestdag and Vandeweyer 2005; Mestdag and Glorieux 2009; Statistics Sweden 2003; Bonke 2002). However, time-use analyses address only a few aspects of eating. For a more comprehensive picture of the daily eating

8

Everyday Eating

patterns in populations, an analysis targeting a wider range of specific aspects of the organization of eating is necessary. Our study is a contribution to the project of providing such an analysis. We focus on documenting and describing as rich as possible the present state of eating practices and their change over time paying special attention to their multidimensionality. We are interested in findings that are (more or less) common to all the four countries. We interpret such findings as possible signs or indicators of general social and cultural changes such as individualization and de-traditionalization, or globalization and commercialization. Our aim is not to explain causally why these trends have developed. There are many factors, structural as well as cultural, common to all four countries or specific to only one or some of them, involved in such processes. Therefore, it would be almost impossible to isolate the decisive factor or factors in each particular case. By analysing which socio-economic groups are mainly responsible for, or leading, the transformation, we can reflect upon whether these changes are here to stay, or whether they are more likely to be only transitory stages connected to a specific generational life phase that will disappear with ageing. Our data are cross-sectional, and it goes without saying that a more definite idea about this would demand following the development of the eating practices of a generation or age group repeatedly and for a longer time.

Our study design To investigate the patterns of everyday eating and how they change we needed rather comprehensive data, which could track micro-behavioural aspects of daily life and data from more than one time point. We conducted two surveys in four Nordic countries (Denmark, Finland, Norway and Sweden) in the two years 1997 and 2012 using a questionnaire that focused on details about one day of eating – the day before the interview. Both surveys included representative samples of the four populations, in all, 4,823 in 1997 and 8,248 in 2012. In 1997 interviews were conducted as telephoned interviews, in 2012 they were webbased. (For a detailed discussion of the data collection methods, the quality, comparability and reliability of the two data sets, see the appendix in this book and the project website http://ifro.ku.dk/everydayeating_data.) We framed the general dimensions of eating using the concept of an eating system. This is a concept combining the physical, organizational and sociocultural aspects of eating (Mäkelä et al. 1999; see also Kjærnes 2001). The eating

Introduction

9

system consists of three elements: the eating pattern, the meal format and the social organization of eating. Here, the eating pattern refers to the rhythm and number of eating events as well as to the alternation of hot and cold eating events. Meal format refers to the composition of the main course (centre, staple, trimmings, gravy, dressings) and the sequence of the meal in terms of starter, main course and dessert. The social organization of eating relates to the venue of the meal, commensal partners and specific organization of the event, and who cooked the meal (see Mäkelä et al. 1999; Kjærnes 2001). In all these dimensions, eating events can be more or less complex or developed. At one end of the scale would be a snack consisting of a single food item, or perhaps a bite, eaten alone in a non-formal setting – say, driving the car. At the other end of the scale would be a three-course dinner eaten at a dinner table with the whole family and guests. The conceptual framework provided by the notion of an eating system is useful when we are trying to understand eating as a whole in which various dimensions operate together and are dependent on each other. The concept of an eating system comes close to the concept of practice. Practice has gained currency in recent research on eating and is defined as a combination of material objects, know-how and meanings, together with their spatial and temporal organization (Southerton et al. 2012; Warde 2016). As Domaneschi (2012) has noted, practices are also inevitably connected with social structures, experiences and conventions. In our study we focused on the concrete timing and organization and the structure of the eating events. It goes without saying that this focus leaves out socially relevant aspects of eating, such as norms about eating, or the specific products consumed and their origins, labels, trademarks, or brands. Such data would open up for analyses of e.g. the relation between ideals and practices or the role of fashions, fads and marketing in our food practices, but this must be left to future research. Our basic concept of the eating system is more than comprehensive enough for the purpose and resources of one study, and with this concept, we hope to have provided a systematic and well-structured account of the most basic aspects of daily eating in the four Nordic countries.

Our questionnaires The survey questionnaires were designed to reflect the basic theoretical concept of an eating system. The main part of the questionnaires focused on respondents’ eating events the day before the interview (see Figure 1.1). It resembled, in other words, the method of 24-hour-recall regularly used in diet studies within

10

Everyday Eating

Figure 1.1 The construction of meal formats.

nutrition science. Questions about eating events were posed in chronological order, to a maximum of ten events in 1997 and 13 events in 2012. This difference in the two questionnaires was purely technical, since no one had eaten so many times. The questions focused on the time of eating and the structure of meals (whether they were hot or cold, the number and types of courses), and on the foods and dishes eaten. They also addressed the social context and conduct of the meal, including who cooked the meal, where the eating took place, the company present, the duration, the seating arrangements for eating events, and finally whether other activities, like reading or watching the TV, took place while the food was being eaten. In the 2012 study, respondents were also asked to categorize each eating event as a type of meal (breakfast, lunch, dinner, lateevening meal, in-between, other). Questions about specific eating events on the previous day were followed by enquiries about the respondents’ sociodemographic background, and their attitudes to food and eating-related issues such as healthy eating habits, experiences of food deprivation, and sustainable food consumption. In both years, interviews were conducted within a fixed time-frame at the end of April with no national holidays and approximately equal numbers of interviews for each day of the week were ensured. The comparison of systematic data collected from four countries over two periods made it possible to distinguish between culturally or nationally idiosyncratic phenomena, and to identify cross-national trends and phenomena. This, we believe, is a strength of the present study. It can be argued that the comparison of more divergent countries – for instance, Southern and Northern European countries – with different historical paths and different kinds of welfare regime would have been interesting and fruitful. On the other hand,

Introduction

11

the similarity of the countries investigated here does offer benefits, since by standardizing many relevant cultural and social factors it reduces the number of potential explanations for observed changes in food culture. The comparison of these four countries offers, in its turn, a more composite picture of eating in modern welfare societies than would have been possible had we focused on only one country.

Four Nordic populations Historically, politically and demographically the four Nordic countries resemble one another in significant respects, partly as a consequence of their shared history. Both Norway and Finland were (at different times) parts of the Swedish empire: Finland, for several centuries, and Norway in the nineteenth century, before which it belonged to the Kingdom of Denmark for nearly five hundred years. The populations of Denmark, Finland and Norway are almost the same size (around five million), but Sweden’s population is roughly twice this size. The four countries are all predominantly Lutheran, and are ethnically and culturally relatively homogenous. With the exception of Sweden, their ethnic and language minorities are, or have been until recently, quite small. All four are relatively wealthy, welfare societies of the social democratic type (on the Nordic model, see Esping-Andersen 1990). They differ in many respects from other European states but global and European trends can be identified in them too. In many respects, the four countries are under the same socio-economic forces that shape the rest of Europe and the world at large. In terms of generalizability of the results of the study it should thus be kept in mind that our findings report changes in daily life in one version of modern societies. Nordic populations have a high share of elderly people. This share did not change much between 1997 and 2012; in 2012 the share of older people (65–80 years) varied across the Nordic region from 15 to 18 per cent. Many people live alone in the Nordic countries. The share of single person households varied from Norway’s 23 per cent to Sweden’s 31 per cent. Between 33 per cent (Sweden) and 43 per cent (Denmark) of households included children. Another important characteristic of Nordic populations is the relatively high level of education. This increased between 1997 and 2012, and the proportion of the population with the highest – university – level of education varied from 24 per cent (Denmark) to 32 per cent (Sweden). Despite these national differences, the four countries resemble each other quite extensively in all four dimensions. Changes or developments between 1997 and

12

Everyday Eating

2012 were also relatively small. For instance, over the period the share of households with children declined only slightly, except in Finland where the change was greater. (For more details, see Tables  2.8 and 3.0 on our website http://ifro.ku.dk/everydayeating_data). Across the Nordic region, national rates of adult participation in the workforce varied somewhat in both periods covered by our study (72–80 per cent). However, the rates did not change significantly between 1997 and 2012. Norway had the highest rates, while Finland showed the lowest rates in both years. A distinctive feature of the Nordic countries is the high rate of female employment, ranging from Finland’s 70 per cent to Sweden’s 77 per cent in 1997, and from Denmark’s 72 per cent to Norway’s 77 per cent in 2012. Although men had higher rates of employment in both years and in all countries, women have had an active role in working life for at least a generation. However, part-time employment was much more common among women than among men. The proportion of employees in part-time work increased between 1997 and 2012 in all four countries, but only modestly. It was lower in Finland than in the three other countries. In both years, there were fewer people unemployed in Norway than there were in the other three countries, where the rates were rather similar and did not change significantly between the first year of research and the second (Nordic Statistical Yearbook 2014). With their high levels of female participation in the workforce, a large proportion of older people, high levels of education, and many single person households, the Nordic populations share many of the traits often defined as key features of Western developed societies. The four Nordic countries are therefore well positioned to illuminate the ways in which eating takes place in modern everyday life.

The contents of the book In the following chapters, we take up many aspects of the social organization of eating in daily life. We begin by describing changes in the foods, i.e. the contents of dishes and meals eaten in the four countries, and then move on to the social organization of eating. This includes the timing, the daily rhythm, the place, the social context and conduct of eating. Throughout, we pay special attention to interconnections and changes in these aspects over time. We go deeper into the context of eating by analysing the prevalence of family meals in the four countries in both years, and patterns of eating out in restaurants and cafés. We then analyse

Introduction

13

the complexity of dishes and meals, and use this as an indicator of their orderliness, in order to study whether social or festive eating occasions are more complex than solitary and ordinary ones. We also ask who, if anyone, does the cooking in order to explore possible changes in the gendered organization of domestic food work. In each case, we pose the question whether it is possible to identify social groups, determined by age, gender, education, the place of residence or social status, which are mainly responsible for observed changes in eating habits in one way or another and/or are distinctive from others. The three final chapters explore themes that take us beyond changes in the social organization of eating and focus on issues that are on public agendas about food consumption: health, environmental sustainability and poverty and deprivation. We ask, first, if there is a connection between a healthy diet and the social orderliness of eating. This is an intriguing question with obvious relevance to health policy, but studies rarely address it systematically. Another topic with direct political relevance is the question to what extent people in Denmark, Finland, Norway and Sweden have adopted eating habits that can be considered environmentally benign. Perhaps more importantly, it can be asked to what extent sustainable eating activities can be explained by motivational factors and attitudes alone, and to what extent they, in fact, are an integral part of a broader set of ‘proper’ attitudes to food and eating encompassing a general interest in food, cooking and healthy eating. Poverty and lack of proper food, which many assume to have become marginal phenomena in rich European countries, have recently attracted growing public attention. We investigate the extent to which people have experienced food shortages and the dispersion of such shortages among socio-economic groups. At the end of the book, in the appendix, we discuss details about the construction of our study samples and the quality of our data. We present the response rates of the two surveys and the representativeness of the study samples when compared to the four national populations. We also discuss potential non-response bias and the substantial quality of our data by comparing with other relevant studies. In the main, we base the analyses we present in the book on simple statistical description and summaries. However, at times, complex analyses from statistical techniques such as sequence analysis, logistic regression, correlational coefficients, and predicted probabilities are used. We describe these analytical techniques in detail in the chapters’ method section for the benefit of readers unfamiliar with statistical methods. Readers that are familiar with these techniques may skip these text passages and instead focus on the remaining parts of the chapter.

14

Everyday Eating

The aim of the book is to draw a composite picture of modern eating in the four Nordic countries, and to describe how eating habits have changed over a fifteen-year period. The strength of our study is that it allows comparing over time and geography the overall changes on a population level. Such studies are rare. The focus is on routines, and on ordinary and regular aspects of our daily food practices. By throwing light on when, where, what, with whom and how we eat, we hope to provide an empirical account of modern eating that can put the more spectacular ideas often presented in the media and public food discourse into perspective and frame them in a social context.

2

The Food We Eat Lotte Holm, Drude Skov Lauridsen, Jukka Gronow, Nina Kahma, Unni Kjærnes, Thomas Bøker Lund, Johanna Mäkelä, and Mari Niva

In much food research, the social context and meaning of food is actually the focal point: often the food itself plays a minor role, whereas analyses centre on social interaction, social structures, social discourses, or on economic and technological developments. The food that is put on the table is of course also a material substance that we incorporate while eating our daily meals. Therefore, the food we eat in daily life is an important starting point for sociological food research. In this chapter we ask: Has the food we eat changed in the period from 1997 to 2012? As far as our data allow, we pursue the sociological discussions described above. The main part of the questionnaire focused on respondents’ eating events the day before the interview. Questions about each eating event were posed in chronological order, to a maximum of ten events in 1997 and 13 events in 2012. The questions focused on number, timing, social context and structure of eating events. When asking about what was eaten we distinguished between hot and cold food. Cold eating events were registered on a list of food options, whereas hot eating events were recorded in a more complex manner, registering number of courses, and centre, staples, vegetables and trimmings in the main course. In both types of events drinks were registered on a list of drink options. In open questions following hot eating events, respondents could report names of the dishes they ate. The questions and the reasoning behind are described in more detail elsewhere (Mäkelä et  al. 1999; Kjærnes et  al. 2001). For the questions regarding what was eaten at various events, the change of survey method may be of significance. In 1997 the lists of food items were not read aloud by the telephone interviewers, so that respondents reported on the basis of their own recollection of which foods they had eaten. In 2012, the computer-assisted (CAWI) methodology implied that respondents were presented lists of potential 15

16

Everyday Eating

food items. They were thus prompted to recollect items that would perhaps otherwise be forgotten. We have reconstructed meals on the basis of timing and the character of each eating event. Breakfast is defined as the first eating event in the day, lunch as eating events taking place between 11.00 am and 2.00 pm, and dinner as events taking place in the following country-specific time spans: Denmark 5.00–8.00 pm, Finland 4.00–6.00 pm, Norway 4.00–7.00 pm, Sweden 5.00–7.00 pm. These are the time spans defined as ‘lunch’ and ‘dinner’ by more than 50 per cent of respondents in the 2012 study. The resulting time-spans concord with the 1997 data where the timing of evening meals differed considerably between the countries whereas lunch times were more similar (Gronow and Jääskeläinen 2001). In this analysis we focus on describing which are the most typical foods, dishes and meals eaten in the four countries. By so doing we will then – as far as our somewhat limited data allows – pursue the sociological discussions we are interested in. Instead of trying to address trends such as gastro-anomy and informalization directly, we will approach them indirectly by looking at the complexity and simplicity of eating events, which is, we contend, inextricably linked with the alleged trends. Globalization and de-traditionalization are addressed by assessing whether differences between the Nordic countries are increasing or decreasing in the period. To this effect we assess whether traditional national dishes are still on the dinner table in 2012 and whether distinct national lunch traditions (cold or hot meals) are still discernible in 1997 and 2012. The effects of the health trend that is discernible from dietary surveys (Pedersen et al., 2010; Helldán et al. 2013) are analysed by studying whether core foods from health discourses, fruit and vegetables, are introduced to types of eating events where they are traditionally not included. We have used SPSS (version 21) for the analysis. Analysis of significant association between variables is carried out by Chi-square tests, since all variables employed are either binary or multiple nominal.

Breakfast Bread is by far the most common food item in both 1997 and 2012 in the Nordic countries. However, in Denmark and Sweden it seems to be on the decline. At the same time breakfast cereal is increasing in three of the countries. This change points to a more modern, simple and quickly prepared breakfast. The exception

Table 2.1 The most frequently chosen food and drink items for breakfast in 1997 and 2012 (% of all breakfasts) Denmark 1997

Finland 2012

%

1997 %

Norway 2012

%

1997 %

Sweden 2012

%

%

Food

Bread Cereal Cake Yogurt

64 15 12 11

Bread Cereal Yogurt Fruit

58 25 20 14

Bread Porridge Yogurt Cereal

63 16 10 7

Bread Yogurt Porridge Othera

64 24 17 16

Bread Cereal Egg Fruit

Drink

Hot drinks Milk Juice Water

71 18 9 9

Hot drinks Water Milk Juice

61 33 20 17

Hot drinks Juice Milk Water

72 14 13 11

Hot drinks Water Juice Milk

67 36 19 16

Hot drinks 49 Hot drinks Milk 48 Water Water 10 Milk Juice 9 Juice

N

1187

2045

1196

2015

1177

78 7 5 4

1997

Bread Egg Cereal Fruit

2064

2012 %

%

70 16 11 11

Bread Cereal Yoghurt Porridge

73 19 17 12

Bread Yoghurt Cereal Egg

65 31 20 17

54 36 28 29

Hot drinks Milk Juice Water

63 26 15 10

Hot drinks Water Juice Milk

64 29 21 15

1244

N is the first eating event defined as breakfast given year and country on all weekdays and weekends. a ‘Other’ is the last choice from the following list: Bread, cereal, savoury pastry, cold cuts, vegetables, porridge, yoghurt, egg, fruit, cake, snacks and sweets, other.

2053

17

18

Everyday Eating

is Finland, where instead the more traditional porridge has kept its place, and its consumption has in fact increased a bit. In Denmark and Norway, porridge is not an important item in either year, but in Sweden, porridge was among the four most important items in 1997, but its consumption declined, and in 2012, egg had taken its place as the fourth most popular item (Table  2.1). Another important change that also points to a changing character of breakfasts is the rather dramatic increase in yoghurt consumption in all countries. Fruit has increased with between 5 and 12 percentage points in all countries, but only in Denmark and Norway is fruit among the four most frequent breakfast food items in 2012. Egg consumption has also increased in all countries, with between 4 and 12 percentage points, mostly in Norway and Sweden, and only in these two countries are eggs among the most important items. The decline in bread consumption suggests that also items used as toppings for conventional open sandwiches, such as cheese or meat cuts, may have reduced. The choice of beverages has undergone some change too (Table 2.1). In both years, beverages are first of all coffee and tea, juice and milk. However, both water and juice appear to have become more frequent in all countries and in Norway and Sweden, milk drinking declines markedly and significantly (with 20 and 11 percentage points, respectively) whereas it is rather stable in Denmark and Finland. In all countries the average number of items for breakfast appears to have increased, but still, having just one item is the most typical in all countries (Table 2.2). However, between 20 and 30 per cent now eat two items for breakfast.

Table 2.2 Number of food components at breakfasts in 1997 and 2012 (% of all breakfasts) Denmark Components

1997

2012

1 2 3 4 or more

80 16.7 2.9 0.4

69 22.3 6.9 1.8

Total 100 N of all breakfast 1121 Sig

***

100 2045

Finland 1997 69.3 24.5 4.9 1.2 100 1056 ***

Norway

2012

1997

2012

52.9 30.4 11.9 4.8

88.9 9 2 0

71.2 21.2 5.7 1.9

100 2015

100 1085 ***

100 2064

Sweden 1997 63.1 25.9 9.2 1.9 100 1186

2012 50.9 28.1 14.8 6.2 100 2053

***

N is the first eating event defined as breakfast given year and country on all weekdays and weekends (excluding events that only consisted of drinks). Test: chi2. Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant.

The Food We Eat

19

In both years Norway and Denmark have more simple breakfasts, and Sweden and Finland more complex breakfasts, but the differences between countries have diminished markedly, especially within the two groups of countries. The rising significance of fruit for breakfast suggests influence from health discourses, as does the decline in cake for breakfast in Denmark.

Lunch The study of eating patterns in 1997 revealed highly diverse national lunch patterns, the main distinction being between hot, cooked lunches dominating in Finland and Sweden, and cold lunches with open sandwiches dominating in Denmark and Norway. Within the countries these features of lunches were more homogeneous indicating that we were observing a basic feature of Nordic food cultures, but also national differences in how lunch was organized. So how does this look in 2012? About two-thirds of the respondents had had something to eat between 11.00 am and 2.00 pm, most of them having eaten once during that time interval. Slightly more people had eaten in this time slot in Denmark and Sweden (72– 75 per cent), compared to Norway and Finland (62–67 per cent). The changes from 1997 to 2012 were insignificant. Figure 2.1 shows that most meals eaten were a substantial meal (hot or cold), while few meals were only a cake, snacks or fruit. Large and consistent national differences are found again in 2012 when it comes to whether lunch consisted of cold food, like sandwiches, or a hot, cooked meal. These differences do not seem to have diminished from 1997 to 2012. There is a small tendency towards more hot lunches, most significant in Sweden, none in Finland. Somewhat more meals were just a snack (not in Sweden). These two small changes explain why cold lunches have decreased in Denmark, Norway and Sweden. But, still, around 70 per cent of the lunches in Denmark and Norway are cold meals, while a similar proportion of lunches in Finland and Sweden are hot. Nordic cold lunches most often consist of open sandwiches with a topping of cheese, cold cuts or pâté. But in 2012 salads and other vegetables have increased dramatically in all countries and also fruit, to a somewhat lesser degree. Salads and vegetables increased with 9–14 percentage points and are now eaten at 20 and 12 per cent of cold lunches in Denmark and Norway respectively, and at 18 and 23 per cent of these (more unusual) lunches in Finland and Sweden. Fruit has increased with 5–10 percentage points, except in Sweden. It is now part of 16 and 18 per cent of cold lunches in Denmark and Norway, and 16 and 17 per cent in Finland and Sweden.

20

Everyday Eating

Figure 2.1 Types of meals in 1997 and 2012 between 11am and 2pm (pct). N in 1997/2012 and significance level from chi2 (Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant): Denmark: = 910/1622 (Sig.:***), Finland 903/1369 (Sig.:Ns), Norway 766/1367 (Sig.:*), Sweden 934/1552 (Sig.:***).

The hot lunches have a format similar to the Nordic dinner (staple plus centre and vegetables, see also below) and that format seems to be shared across all four countries. Around two thirds consist of just a main course, but quite a few also have a first course or a dessert. The number of courses has been significantly reduced from 1997 to 2012 in Denmark and Norway, likewise in Sweden a small reduction can be found, while in Finland no change can be found. Figure  2.2 shows that in all countries around half of the hot lunches included meat. Lunches with fish were most common in Norway, least so in Denmark, but fish for lunch decreased in all countries but Finland. Lunches with a vegetable centre, however, had become more common in all countries, being most popular in Denmark and Finland (21 and 19 per cent, respectively). The category ‘other’ includes, for example, pizza, omelet and soups and constituted around 18 per cent of the hot lunches in 1997 in Sweden, 13 per cent in Finland and decreased somewhat between the two years in these countries. In Denmark we found a small increase and in Norway a marked increase in these types of meals. Lunch is eaten somewhat later in the day during weekends and this has not changed. While the composition was quite distinct from weekday lunches in 1997, with significantly more meat centres, this was less apparent in 2012 (data not shown). Meat had to some extent been replaced by fish (increasing slightly,

The Food We Eat

21

Figure 2.2 Main ingredients of the centre at hot lunch in 1997 and 2012 (pct). N in 1997/2012 and significance level from chi2 (Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant): Denmark: = 173/287 (Sig.:***), Finland 630/949 (Sig.:Ns), Norway 122/254 (Sig.:**), Sweden 579/1075 (Sig.:**).

contrary to the tendency, on weekdays),1 but more often by mixed dishes without a clear main ingredient. We do not know what this shift signifies, whether it indicates a tendency of informalization, with a less strict lunch format during weekends, a ‘modernization’, with more dishes that do not adhere to the traditional format, or, whether the mixed dishes are in fact traditional casseroles (see below in the section on Dinner). Around 95 per cent had a drink with their lunch, the most common being water, which also increased markedly and significantly from 14–29 per cent in 1997 to 42–62 per cent in 2012 (Table 2.3). Having milk with lunch varies between the countries, but had been reduced from 22–37 to 10–28 per cent. Tea and coffee are also popular, most so in Norway (included in half of the lunches, around 25–30 per cent in the other countries), increasing only in Finland. Soft drinks as well as alcoholic drinks are rare in 2012 (5–8 and 1–6 per cent, respectively). Likewise, juice has been a rare type of beverage for lunch but it is becoming more popular in Norway in 2012 (11 per cent). Taken together, the national lunch patterns seem highly persistent, with hot lunch perhaps having strengthened its position slightly. Still, cold lunches dominate in Denmark and Norway, hot in Sweden and Finland. The meal format 1

This should not be interpreted as a tendency of eating less meat, since this is not what we see in aggregate statistics on meat consumption.

22 Table 2.3 Beverages consumed for hot lunch and cold lunch in 1997 and 2012 from 11 am–2 pm (pct) Denmark

Finland

Norway

Type of beverage

1997

2012

sig

1997

2012

sig

Milk Water Soft drinks Juice Alcohol Hot drinks Other No drink N

22.3 21.1 10.2 3.6 11.6 24.2 10 1.4 937

10.1 58.8 8 3.2 6.3 24.1 3 5.2 1622

*** *** * Ns *** Ns *** ***

36.5 27.3 1.8 5.5 2 22.2 7.6 4 947

27.5 51.5 4.7 8.3 1.7 28 2.9 3.4 1369

*** *** *** ** Ns ** *** Ns

1997 21.6 14.1 7.4 5.2 0.5 53.8 6.9 3.5 772

Sweden

2012

sig

1997

2012

sig

11.1 41.7 5.5 11.2 0.7 52.5 2 5.9 1367

*** *** Ns *** Ns Ns *** *

19.6 28.5 5.5 5.8 8.1 32 6.7 5.7 970

9 62 6.4 3.9 3.7 26.9 2.1 4.3 1552

*** *** Ns * *** ** *** Ns

N is the drinks at eating events defined as meals hot or cold in the timeframe, given year and country on all weekdays and weekends. Eating events used: 1997:2–10, 2012:2–13. More than one choice was possible, so % does not add up. Test: chi2. Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant.

The Food We Eat

23

is also stable, but a few more had a vegetable centre. There has been a noticeable shift towards water as the dominant drink with lunch. These findings do not support assumptions of ‘breakfastization’ tendencies often purported in Sweden, with people more often having lunch with items typical of breakfast.

Dinner As Table 2.4 shows, a large majority of hot dinners consisted of only one course in all countries both in 1997 and 2012. The share of one-course dinners increased over time in Denmark and Sweden, remained constant in Norway and declined in Finland. An opposite development can be seen in the share of two-course dinners – in which desserts were much more typical than starters (data not shown) – which declined in Denmark and Sweden and increased in Finland. The share of three-course dinners remained very low. It thus seems that Nordic dinners are still characterized by ‘platefuls’ (Murcott 1982) of main course, and desserts and especially starters are relatively rare. This pattern is more prevalent on weekdays and less so at weekends, but the difference is not very large (data not shown). Complexity of the dinner can also be measured by looking at how many meal components – centre, staple, bread, vegetables and trimmings (Mäkelä 2001) – were included. Our data suggest that the complexity of the platefuls has increased, from 1997 to 2012, as the mean number of components has increased in all countries (minimum one, maximum five). In Denmark it increased from 2.6 to 3.3, in Finland from 2.4 to 3.0, in Norway from 2.6 to 3.2 and in Sweden from 2.8 to 3.3. Especially trimmings appear to have become more common. If we take the number of dishes at the dinner or the complexity of the dishes as a rude indicator of the degree of formalization vs informalization of the Nordic dinners our results point in two directions: on the one hand, the number of courses has remained more or less constant, on the other hand, the number of the components in individual meals has increased. Our data suggest that people in the Nordic countries like to eat first and foremost meat dishes at dinner (Figure 2.3). In both years, more than half of the respondents in all countries identified meat as the main ingredient of their dinner. The share varied between 55 per cent (Norway) and 73 per cent (Denmark) in 1997, and between 55 per cent (Finland) and 63 per cent (Denmark and Sweden) in 2012. However, meat dishes seem to have become somewhat rarer and vegetable dishes somewhat more popular both in Denmark and Finland, but

24 Table 2.4 Number of courses at hot dinner in 1997 and 2012 (pct) Denmark (5–8 pm ) Number of courses 1 2 3 Total N Sig. Mean St.d.

1997 73.3 24.1 2.5 100 833 1.2917 0.50729

2012 78.2 19.6 2.3 100 1452 * 1.2410 0.47807

Finland (4–6 pm ) 1997 72.4 25 2.6 100 392 1.3010 0.51195

Norway (4–7 pm ) 2012

64.6 33 2.4 100 666 * 1.3784 0.53262

1997 69.8 28.7 1.5 100 599 1.3172 0.49704

Sweden (5–7 pm ) 2012

72.9 25.5 1.6 100 1115 NS 1.2870 0.48696

1997 71.9 26.3 1.8 100 452 1.2987 0.49539

2012 76.6 19.7 3.7 100 898 ** 1.2706 0.52074

N is the hot eating events in the time frame, given year and country on all weekdays and weekends. Hot eating events used: 1997:2–10, 2012:2–13. Test: chi2. Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant.

The Food We Eat

25

Figure 2.3 Main ingredients of the centre at hot dinners in 1997 and 2012 (pct). N in 1997/2012 and significance level from chi2 (Sig. defines the p-level of significance: p < 0,001 = ***, p < 0,01 = **, p < 0,5 = *, p ≥ 0,5 = Not significant): Denmark: = 833/1452 (Sig.:***), Finland 309/666 (Sig.:*), Norway 599/1115 (Sig.:***), Sweden 452/898 (Sig.:***).

not in Norway and Sweden. Fish was in both years most popular in Norway, where its use as a main ingredient, however, has declined – while in the other countries changes in fish dishes are minor. Vegetables as a main ingredient has increased in all countries, while the ‘other’ category shows minor changes apart from decreasing in Sweden. More can be learned about dinners when looking at open-ended qualitative responses about the name of the dinner dishes. Our data does not allow this analysis for 1997, and the following analysis is therefore based on responses only from 2012 about all self-defined dinners irrespective of the timing of the meal. In general, the qualitative descriptions of dinners were very short. Hardly any fancy (à la French cuisine) names appear in the list. A typical answer was simply ‘steak’ or a bit more elaborated ‘steak with potatoes’. On rare occasions, the respondent had taken the time to describe the dish in detail: ‘Mørbrad biff m/ fløtegratinerte poteter, bernaice og brokoli’ (NO). The dish depicted was almost always a plateful (Murcott 1982). Interestingly, in all countries there were references to buffet but this did not refer to the traditional Scandinavian smorgasbord but to offerings such as ‘SushiThai buffe’ (SE) in ethnic restaurants. Generally, ‘buffet’ may indicate ‘a bit of this and a bit of that’. However, there were recurrent dishes that belong to Nordic/Scandinavian food traditions. One example is the almost iconic comfort dish Pyttipanna/ Biksemad, which is a bubble and squeak-type dish known and reportedly eaten in all four countries. Another classic in all four countries was meatballs, which

26

Everyday Eating

come in many forms and variations. In Norway and Denmark there were also a lot of fishballs. In all countries people mentioned dishes that are probably more typical to the respective culinary culture than to all Nordic countries. For example, in Norway there was ‘komle med pølse og salt kjøtt, kålrabistappe, gulrot, bacon og smør’, which means a large potato dumpling with sausage and salted meat, kohlrabi mash, carrot, bacon and butter. In Finland, liver casserole, one of the first ready-to-heat convenience foods on the market, was mentioned. The Danish ‘forloren hare’ (mock hare meatloaf) was reported several times. In Sweden, there was ‘ärtsoppa, senap med pannkakor och sylt’, i.e. the pea soup with mustard traditionally eaten on Thursdays and followed by pancakes and jam as dessert. In all countries various dishes made of minced meat were very popular. People had eaten hamburgers with or without the bun, but minced meat appeared also in soups and stews. In Finland, one of the most popular and characteristic casseroles (‘makaronilaatikko’) is made of macaroni, minced meat, milk and eggs. In Norway, minced meat is frequently used in convenience type dishes like spaghetti with meat sauce, taco and pizza. In addition to minced meat, people reported eating a lot of poultry, i.e. broiler chicken. Today, broiler chicken is one of the most often used meat types in the Nordic countries, and our data show it was used in a variety of dishes from Chicken Korma to grilled whole chicken. Often, chicken seemed to be an ingredient in dishes inspired by culinary cultures other than Nordic. An interesting detail in both the naming of the meat and consuming it was the fact that no one in Denmark or Finland reported eating ‘entrecote’, a dish that several respondents in both Norway and Sweden had eaten. Otherwise, beef and pork were eaten a lot. Fish in various forms and species was also a favourite, especially in Norway and Denmark where many had dishes based on minced fish. In Sweden and Finland fishballs and cakes were rarer but fish was consumed in soups or cooked in the oven. The picture presented above based on the open-ended responses, is backed up by national top ten lists of popular dishes. In Denmark, the latest data show that chicken, meat sauce and meatballs are the most popular dishes in Denmark (Fagt and Jensen 2012). In Finland the top four favourite dishes in 2013 included sauce made of minced meat, fried fish, steak or schnitzel, and meatballs (Suomen Gallup Elintarviketieto OY 2013: 67). Dessert patterns – in terms of what is eaten for dessert – seem relatively stable. In all countries berry or fruit desserts were the most typical dessert types in both years. In 1997, between 8 per cent (Finland and Norway) and 14 per cent (Sweden), and in 2012 between 6 per cent (Denmark and Sweden) and 11 per cent (Norway)

Table 2.5 Types of beverages at hot dinner in 1997 and 2012 (pct) Denmark (5–8 pm)

Finland (4–6 pm)

Norway (4–7 pm)

Sweden (5–7 pm)

Type of beverage

1997

2012

sig

1997

2012

sig

1997

2012

sig

1997

2012

sig

Milk Water Soft drinks Juice Alcohol Hot drinks Other No drink N

15.8 27.1 9.6 2.6 31.6 1.9 12.1 6.7 833

9.5 58.4 13 3 26.9 5.3 3.6 2.1 1452

*** *** * Ns * *** *** ***

52.3 27 2.3 6.1 3.3 2.3 7.1 4.3 392

38.4 45.8 6.8 7.1 7.2 12.2 0.9 2 666

*** *** ** Ns ** *** *** *

7 44.2 15.2 3 6.3 7.2 18.9 2.7 599

3.7 66.3 13 3.5 14.9 8.1 4.8 2.4 1115

** *** Ns Ns *** Ns *** Ns

27.9 28.3 7.7 3.3 25 8.6 6 3.5 452

17.1 50.9 8 4.6 22.2 9.5 3 1.6 898

*** *** Ns Ns Ns Ns ** *

N is the drinks at eating events defined as meals hot in the timeframe, given year and country on all weekdays and weekends. Eating events used: 1997:2–10, 2012:2–13. More than one choice was possible, so N and % doesnot add up. Test: chi2.. Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant.

27

28

Everyday Eating

of hot dinners included a berry or fruit dessert. Ice cream or other frozen desserts, cake or sweet pies competed for the second and third places in all countries, and in Norway also pudding. All in all, the differences between the years and the popularity of the most common dessert types were rather small. For beverages enjoyed during dinner there were significant differences between the four countries and also quite large changes over time. Table  2.5 shows, first, that drinking water at dinner increased considerably, around 20 percentage points or more in all countries, and drinking milk declined particularly in the milk-drinking countries Finland and Sweden. Drinking soft drinks increased somewhat in Finland and Denmark and remained stable in Norway and Sweden. Having juice with dinner remained at a relatively low level. Drinking alcohol at dinner showed a variable pattern across countries. It was quite low in all countries, highest in Denmark, but here it declined. We found an increase in Norway and Finland, but still drinking alcohol with dinner was, in 2012, more popular in Denmark and Sweden than in Finland and Norway. Hot drinks during or immediately after dinner were relatively rare, but their share increased in all countries, and in Finland they were, in 2012, more common than alcohol.

In-betweens Between the main meals, people in Nordic countries eat snacks or in-betweens. We have constructed variables for such in-betweens as eating events that only included either fruit, cake or snacks and sweets. Table 2.6 presents the proportion of cold eating events (excluding breakfasts) which only includes these items. The table shows that in all countries the frequency of cake has declined markedly and significantly, whereas snacks and sweets have increased. Fruit increases markedly in Denmark, less so in Norway, not at all in Finland and decreases in Sweden. These changes suggest a somewhat mixed result with respect to health, but they do suggest a cultural change in the role of cake eating. In all countries cake eating in the evening goes down, whereas snack eating goes up (data not shown).

Discussion Our analysis shows that the food we eat in the Nordic countries in the period from 1997 to 2012 is characterized by stability but also some change. Core

Table 2.6 The frequency of fruit, cake or snacks and sweets at ‘in-between’ in 1997 and 2012 (pct) Denmark Snack in detail Fruit Cake Snack and Sweets Combinations N of in-betweens

Finland

Norway

1997

2012

sig

1997

2012

sig

1997

20.5 57.1 19.8 2.6 767

39.2 18.1 36.9 5.8 1176

*** *** *** **

32.4 54.6 10.4 2.5 709

27.6 38.5 29.5 4.3 691

Ns *** *** Ns

30.2 47.6 18.8 3.3 483

2012 40.3 29 23.9 6.8 837

Sweden sig

1997

2012

sig

*** *** * **

30.4 50 16.3 3.3 852

38.6 29.1 26.2 6.1 818

*** *** *** **

Combinations mean the combination of fruit, cake, snacks and sweets. N is all meals that were only snack given year and country on all weekdays and weekends. Test: chi2. Sig. defines the p-level of significance: p < 0.001 = ***, p < 0.01 = **, p < 0.5 = *, p ≥ 0.5 = Not significant.

29

30

Everyday Eating

elements in traditional Nordic national food cultures persist such as the existence of two marked lunch cultures based on hot and cold meals respectively, the dominant position of meat in dinner dishes, and the rather simple meal formats of both hot lunches and dinners, as ‘platefuls’. But many changes can be identified too, such as the marked tendency that water is becoming a most popular everyday drink for all meal types, the radical decline in cake served as an in-between, the apparent increase in vegetarian hot lunches and dinners, the introduction of fruit and vegetables at meals where they are not to be found by tradition (breakfasts, cold lunches) and the rise in cereals and yoghurts for breakfast. As for the trends of change suggested in the sociological literature, our results are mixed. The many Nordic traditional dishes eaten for dinner as well as the sharp distinction between hot and cold lunch countries suggest that national distinct traits persist. On the other hand, the marked changes we identify are in most cases shared by all countries suggesting a certain harmonization between the food cultures. Thus, we do see some homogenizing tendencies but the underlying distinct national patterns do not seem to dissolve. There are mixed indications of potential informalization in the sense of changes in simplicity and complexity of meals. In most but not all countries we see that hot dinners and lunches are reducing to one-course meals and some changes in the items chosen for breakfast suggest that foods without any preparation such as yoghurt and cereal are prioritized. On the other hand, our data suggest that the number of items included in meals is increasing. This, however, needs to be confirmed by other studies, as our shift in study design results in some uncertainty on this point. The trend towards including foods that are categorized as healthy in more types of meals is evident in our data about both breakfast and lunch, and to some extent dinners, too. Fruit is increasingly included in both breakfasts and lunches, and salads and vegetables in cold lunches. Vegetables have also become a more typical main ingredient in both lunches and dinners. The decline in cake for inbetween meals is open for more composed interpretations: on the one hand, a sign of more healthy in-betweens such as fruit has increased for these meals – but, on the other hand, snacks has gone up, so another explanation could be that the traditional cake culture is in decline and being substituted by simpler and easier, ready-bought snacks.

The Food We Eat

31

Conclusion There is much stability with respect to the food we eat in Nordic countries. Distinct national patterns with respect to types of meals and dishes eaten persist in spite of evidence of marked changes going on which are quite similar in all four countries. Still, we do not see a general harmonization of Nordic food cultures. The evidence regarding informalization in terms of complexity of meals points in two directions and needs to be analysed further by other studies. The importance of health discourses is reflected in the inclusion of healthyfoods meals that are new compared to tradition.

32

3

The Daily Rhythm of Eating Thomas Bøker Lund and Jukka Gronow

In this chapter we focus on the temporal dimension of eating and ask whether culturally shared timing of eating rhythms have disappeared or, less dramatically, are declining. The temporal dimension of eating is included in this study because we believe that a shared temporal rhythm is an underlying base that all the other aspects of meal culture (such as spatial coordination, social meals, cuisine, and dietary quality) rest on. The daily schedule of activities in general (Southerton 2012), and eating, specifically (Southerton, Diaz-Méndez and Warde 2012) – that is, how many times, and when, typical eating takes place, and how those events are ordered – is arguably a basic characteristic of any cultural pattern. For instance, some kind of shared and/or regularly repeated daily order and rhythm of eating is a prerequisite of the sociability of eating. Although having lunch at the same time as others, like family members or colleagues, does not in itself guarantee that it was enjoyed in the company of those others, without shared mealtimes eating together would not be possible at all. Collective conventions about cuisine are an extra layer of food culture in which the shared timings of meals and the consumption of meals together with others are required. The assumption that when eating becomes de-socialized and irregular, the healthiness of the associated diet is threatened (a hypothesis referred to in Caplan (1997), which also receives some support in nutrition studies (e.g. Hammons and Fiese 2011) and presupposes that the temporal coordination of eating has declined. A further reason for focusing on temporal rhythm is that, from another theoretical perspective, the individual timing of eating and institutional timings can be viewed as functionally interwoven. Eating times are arguably both conditioned by, and change with, other institutional arrangements and, in return, exercise their own influence on them (Southerton, Diaz-Méndez and Warde 2012). Also, the few large-scale historical studies that exist suggest that changes 33

34

Everyday Eating

in the production mode of societies and the resulting re-institutionalization of work routines, also determine eating routines profoundly. Although we concentrate on the temporal rhythms of eating, we do not neglect further dimensions of destructuration, grazing, and gastro-anomy. Thus, where temporal disorder is identified we also ask whether spatial and social dimensions, as well as other eating-related norms, decline correspondingly.

Our research questions and data considerations Questionnaire data collected in the Nordic countries (Denmark, Finland, Norway and Sweden) in 1997 and 2012 are used to examine whether culturally shared eating rhythms have disappeared or are decreasing. We set up two basic research questions: Firstly, we examine whether culturally shared eating rhythms can be discerned in 1997 and 2012 in all four Nordic countries. As part of this, we also ask whether any noteworthy changes have taken place in the period 1997–2012. Secondly, where it is possible to discern such shared rhythms, we investigate whether there is a sub-population that deviates from this collective rhythm. We also analyse whether this sub-population has increased between 1997 and 2012, and whether other socio-structural factors explain such a sub-population, including socio-demographic factors and employment status. Finally, we seek to establish the extent to which this sub-population deviates with respect to other dimensions of food culture (spatial, social and cultural norms regarding manners/healthy food). Fifteen years is, admittedly, not a very long period over which to record cultural and social changes in what are arguably rather basic features of everyday life (mundane eating habits). It would have been preferable to compare our two time-points with, say, the 1950s or 1960s, since those decades are arguably representative of the collective and stable eating culture of which observers think when characterizing the decline of shared eating culture. Indeed, hardly any signs of destructuration were observed from 1988 to 1999 in a Belgian study (Mestdag 2005), while some change was observed when comparing changes in meal commensality from 1966 to 1999 in Belgium (Mestdag and Glorieux 2009). Nevertheless, we believe that a comparison of 1997 and 2012 offers a unique opportunity to evaluate the main aim of this paper, and to extend some of the findings from Belgium to the Nordic countries with updated data covering a very recent year: 2012.

The Daily Rhythm of Eating

35

To recapitulate: this chapter aims to assess whether collective rhythms of eating have disappeared or are disappearing. We begin with a presentation of the data and study setup, after which we describe the research method. Results from the population-wide analysis and the sub-population analysis are presented. Finally, in the discussion our results are put into a specific perspective; we discuss whether there is a link between the macro-institutional and micro-institutional timing of eating, or whether the destructuration hypothesis, predicting the end of collective timing, is more to the point. Since sub-populations with different temporal rhythms of eating are compared on spatial and social dimensions of eating, as well as on norms of proper conduct (manners/health), the study also offers an empirical assessment of the grazing and gastro-anomy assumption that all dimensions of eating culture are disintegrating conjointly.

The research method In addressing the first research question, we analyse the population-wide distribution of events during the day in each country and in both years. All eating events, as reported by the respondents, are included in this analysis (and in the following analyses). Two diagnostics are used. Firstly, the temporal pattern of eating rhythms during the whole day is displayed graphically by reporting the proportion of the sample eating in any one-hour time slot on weekdays (Monday–Friday). Secondly, the average number of meals eaten during the day is reported, along with population proportions eating one meal, two meals, three meals, four meals, etc., separating weekdays and weekends, so as to examine whether the three-meal pattern identified in earlier historical studies (Rotenberg 1981) still exists. In addressing the second question, we start by conducting a sequence analysis (Abbott 1990) and subsequent cluster analysis. The combination of these two methods makes it possible to identify sub-populations with differential patterns in the kind of temporal data employed here. The sequence analysis uses SQAdos, which is an add-on program for the statistical package Stata (BrzinskyFay, Kohler and Luniak 2006). Sequences with 24 time points were inserted with a dichotomous event space (eating or non-eating) as element input. Optimal matching analysis was conducted in order to determine pairwise sequence similarities between the respondents. After this, the derived distance matrix from optimal matching was used as input in a cluster analysis, using Ward’s linkage method. To ensure the data had been sufficiently explored, and to disclose

36

Everyday Eating

possible cross-country commonalities in the ways in which daily eating rhythms are differentially patterned, we initially inspected cluster analysis results from three optimal matching procedures described by Lesnard (2010). We report the results we obtained using the Levenshtein II distance measure (indel costs 1, substitution costs 2), as this measure identified a distinctive cluster that differed from the remaining clusters. At the same time, the specific features of the cluster were clearly identifiable in all four countries. With respect to the temporal dimension of daily eating, the patterns of the sub-populations that emerged from sequence analysis and subsequent cluster analysis are interpreted, firstly, by displaying graphically the proportion within the derived clusters that eat in each one-hour time slot during the day. Secondly, the extent to which the clusters time their eating with collective time is assessed by reporting the proportion eating their first meal during country-specific breakfast time intervals and in country-specific lunch intervals, and the proportion eating their second or later meal in country-specific lunch intervals and dinner intervals. Thirdly, the proportion within clusters that in 2012 reported eating breakfast, lunch, dinner, and an intermediate meal (snack) is displayed so as to determine the extent to which the three-meal pattern (breakfast–lunch– dinner) is being followed.1 We compare the sub-populations in respect of other dimensions of food culture that also are assumed to decline along with temporal destructuration of eating. Here we report the percentage of meals that were eaten alone (commensality), at home (spatial destructuration), and in less than ten minutes (grazing). The notion that cultural norms do not offer guidance on proper eating and table manners is examined by displaying the percentage of the meals taken at home that were eaten at a dinner table, and while watching TV. The assumption that food culture lacks the general ability to guide proper food habits is examined by comparing the dietary quality of food eaten in the sub-populations. For this purpose, a validated measure of the dietary quality of people’s eating habits, known as Dietary Quality Score (Toft et al. 2007), is used. The eight food frequency questions needed to identify dietary quality were put to the respondents, after which they were assigned to an unhealthy, average, or healthy dietary group, again following Toft et al. (2006). In order to examine whether the sub-population that deviates from the collective eating rhythm increased between 1997 and 2012, and whether socio-

1

We included all self-reported ‘in-between’ and ‘evening meal’ events as snacks. The ‘evening meal’ category was presented to the respondents using country-specific labels.

The Daily Rhythm of Eating

37

structural factors can predict the respective sub-populations, we conducted a logistic regression in which a (binary) variable indicating eating rhythm was inserted as dependent variable. In this analysis we used eating rhythm observations from weekdays only in order to strip out weekend patterns, which are generally less structured. We entered the year (1997/2012) together with potentially relevant socio-demographic variables as explanatory factors: gender, age, area of residence, household type, household income, education, occupational status, and social class (we used the ESEC (European SocioEconomic Classification) measure of social class (Rose and Harrison 2007)). To assess whether the eating rhythm clusters are associated with macro-institutional settings and institutional changes (Rotenberg 1981; Southerton, Diaz-Méndez and Warde 2012), the clusters are characterized according to prevalence at the weekend (Saturday and Sunday) and on weekdays (Monday–Friday), and according to prevalence in different employment groups. In the latter case, we distinguish between students, pensioners, the unemployed, and the employed. Among those employed, we also distinguish those that did and did not work during the day the eating rhythm was recorded.

National peak times and regularities of eating In both 1997 and 2012 we identified clear national patterns with daily rhythms. There were peak hours during which a great number of people had eaten something, as well as quieter hours when eating was not at all common (see Figure  3.1). In general, these regularities were rather similar in all of the countries. They corresponded quite well with the three-meal pattern identified in earlier historical studies (Rotenberg 1981) and recent time-use studies from other countries (Mestdag 2005; Southerton, Diaz-Mendez and Warde 2012). The first peak, which can be reasonably assumed in most cases to be breakfast, occurred at 7.00–8.00 am in all four countries, even though smaller peaks could be identified both one hour earlier (6.00–7.00 am) and one hour later (8.00–9.00 am), and again at 9.00–10.00 am. The second peak occurs around lunchtime, and is even more distinctive than the earlier, morning peaks of eating. Typically, the lunch hour in Finland and Norway is 11.00 am–12.00 pm; approximately 40 per cent of respondents were found to have eaten during this time. In Denmark approximately 50 per cent, and in Sweden over 40 per cent, of respondents had eaten during the typical lunch hour of these two countries, which is 12.00–1.00 pm.

Percentage eating

Percentage eating

Norway

50% 1997 (N = 736) 60%

2012 (N = 1447) 50%

30%

20%

10%

2012 (N = 1445)

40%

30%

20%

10%

0% Percentage eating

40%

4– 5 5– am 6 6– a m 7 7– am 8 8– am 9 9– am 1 10 0 a – m 11 11 –1 am 0– 2 a 1 m 1– pm 2 2– pm 3 3– pm 4 4– pm 5 5– pm 6 6– pm 7 7– pm 8 8– pm 9 9– pm 1 10 0 p – m 11 11 –1 pm 0–2 p 1 m 1– am 2 2– am 3 3– am 4 am

5 5– am 6 6– am 7 7– am 8 8– am 9 9– am 1 10 0 a – m 11 11 –1 am 0– 2 a 1 m 1– pm 2 2– pm 3 3– pm 4 4– pm 5 5– pm 6 6– pm 7 7– pm 8 8– pm 9 9– pm 1 10 0 p – m 11 11 –1 pm 0–2 p 1 m 1– am 2 2– am 3 3– am 4 am

4–

50%

Percentage eating

60%

Denmark

0%

1997 (N = 957)

4– 5 5– am 6 6– a m 7 7– am 8 8 – am 9 9– am 1 10 0 a – m 11 11 –1 am 0– 2 a 1 m 1– pm 2 2– pm 3 3– pm 4 4– pm 5 5– pm 6 6– pm 7 7– pm 8 8– pm 9 9– pm 1 10 0 p – m 11 11 –1 pm 0–2 p 1 m 1– am 2 2– am 3 3– am 4 am

5 5– am 6 6– am 7 7– am 8 8– am 9 9– am 1 10 0 a – m 11 11 –1 am 0– 2 a 1 m 1– pm 2 2– pm 3 3– pm 4 4– pm 5 5– pm 6 6– pm 7 7– pm 8 8– pm 9 9– pm 1 10 0 p – m 11 11 –1 pm 0–2 p 1 m 1– am 2 2– am 3 3– am 4 am

4–

38 60%

Time of the day 60%

Time of the day

Figure 3.1 Population-wide eating rhythms in the Nordic countries on weekdays (Mon–Fri) in 1997 and 2012. Finland

50%

Sweden

Time of the day

1997 (N = 977)

40%

2012 (N = 1432)

30%

20%

10% 0%

Time of the day 1997 (N = 906)

40%

2012 (N = 1459)

30%

20%

10% 0%

The Daily Rhythm of Eating

39

The picture of later eating frequencies – those in the afternoon and evening – is less clearly defined. Although peaks can be identified, eating seems to be more spread out during these later hours, and national patterns differ more. However, in Denmark there is a very sharp peak at 6.00–7.00 pm, like the similar one at lunchtime, when almost 50 per cent of respondents in 1997, and over 40 per cent in 2012, were found to have eaten something. In no other Nordic country are these later peaks so well-defined. A possible explanation for the lower peaks from mid-afternoon onward is that work-related obligations, which probably shape daily eating schedules substantially (cf. Rotenberg 1981; Southerton, DiazMéndez and Warde 2012), are less limiting during the later hours of the day.

Stability or change 1997–2012 On most accounts the daily rhythms of eating in 1997 and 2012 were rather similar. Closer inspection of these two years, however, reveals some systematic changes. Most notably, the early morning peaks between 7.00–8.00 am have become less accentuated in all the four countries. Second, the lunch hour peaks between 12.00 pm and 1.00 pm have flattened out somewhat, both in Denmark and in Sweden, while the lunch hour peak in Norway, between 11.00 am and 12.00 pm, has risen a little. The other changes observed were not as systematic. We observed smaller increases and decreases in the time slots following each other, and these changes, differing from one country to another, were not welldefined. Although there appears to be a systematic flattening out of peaks during the mornings and at lunchtime, the change is by and large a minor one. On the whole, the clear peaks discernible in Figure 3.1 indicate the existence of culturally shared eating rhythms in both 1997 and 2012 at this population-wide level of analysis. The same conclusions can be drawn if we look at country averages for the whole day. Thus, changes in the average number of eating events per day are quite small and almost negligible for weekdays in Denmark, Finland and Sweden (Table 3.1). The only relatively substantial deviation over time emerges in Norway, where the average number of meals on weekdays increases from 3.74 in 1997 to 3.91 in 2012. During weekends the change is also negligible in Finland and Sweden, while in Denmark the average number of meals increased somewhat from 3.55 to 3.81. In Norway it rose even more, from 3.26 in 1997 to 3.60 in 2012. Despite these changes in average number of meals in Denmark and Norway, the most common pattern in all of the four countries, in both 1997 and 2012, was to eat three or four times a

40 Table 3.1 Number of eating events in the Nordic countries on weekdays and weekendsa in 1997 and 2012 – in percentb and means Denmark Weekday 1997/2012 0–1 events 2 events 3 events 4 events 5 events 6–10 events Means St. Dev. Total (N) a b

1 7 29 36 21 8

1 7 31 28 21 12

3.94 4.04 0.98 1.28 736 1449

Finland

Weekend 1997/2012 1 11 38 34 14 2

2 7 36 30 17 8

3.55 3.81 0.98 1.28 257 602

Weekday 1997/2012 1 8 26 36 21 7

2 9 26 33 18 11

3.93 3.96 1.15 1.25 977 1440

Only Sunday in Finland and Norway in 1997. Percentages do not sum to 100 in all instances due to rounding error.

Norway

Weekend 1997/2012 2 13 30 34 15 6

3 10 35 31 15 7

3.63 3.71 1.19 1.16 211 576

Weekday 1997/2012 1 8 31 41 16 3

1 7 28 39 17 8

3.74 3.91 0.98 1.13 958 1476

Sweden

Weekend 1997/2012 1 17 46 31 5 1

1 10 40 32 14 3

3.26 3.60 0.82 1.00 221 592

Weekday 1997/2012 0 6 31 35 19 8

1 7 35 32 17 8

3.93 3.85 1.06 1.16 903 1462

Weekend 1997/2012 1 12 35 33 16 4

2 9 43 28 14 5

3.63 3.60 1.04 1.06 168 598

The Daily Rhythm of Eating

41

day. For instance, at weekends 41 per cent of Norwegian respondents reported eating four times a day in 1997, and 39 per cent in 2012. The corresponding figures for eating three times were 31 and 28 per cent, respectively. We can reasonably assume that this indicates in most cases a three-meal pattern (breakfast, lunch, dinner, with or without a snack break (cf. Rotenberg 1981)), suggesting that this daily pattern is still a collective anchor point in all countries. Coupling this with the peaks identified in Figure  3.1, indicating collectively shared temporalities, we conclude that national eating rhythms were still very much alive in 2012.

Within-country differences in daily eating rhythms In order to identify possible differential patterns in daily eating rhythms within the countries we conducted a sequence analysis and subsequent cluster analysis (described in more detail in the section ‘The research method’) and requested a solution with five clusters. Following this analysis, it was possible to detect, in each country, a cluster that was both different from the other clusters within each country and quite similar in all four countries (see cluster 1 in Figure  3.2). We will focus on cluster 1, partly because it has country-invariant characteristics, and partly because it can be interpreted as a rather interesting ‘unsynchronized’ pattern (as outlined below) that departs considerably from the culturally shared eating rhythms set out in the last section. To save space, Figure 3.2 displays through graphs (to the left) and numerical information (to the right) all five clusters emerging from the analysis for just one country: Norway. A number of features are characteristic of cluster 1 alone. The eating day starts considerably later, since by 10.00 am only 24 per cent of individuals in cluster 1 have eaten their first meal of the day, while either a majority or virtually everyone has done so in clusters 2–5. In cluster 1, 52 per cent eat their first meal of the day during typical lunch hours (between 11.00 am and 2.00 pm). In clusters 2–5, practically no one eats a first meal in that period and a large proportion eats a second or later meal. In cluster 1, eating is also less frequently undertaken during typical dinner hours (between 4.00 pm and 7.00 pm), although the difference between cluster 1 and the remaining clusters is less marked compared to breakfast and lunch. Overall, the timing of eating in cluster 1 certainly deviates from the discernible collective peaks that were identified in the first part of the analysis. Until now, we have assumed that these peaks reflect breakfast, lunch and dinner. In principle, however, the unsynchronized eaters may subscribe to the three-meal

42 Figure 3.2 Graphical display of five daily eating rhythms in Norway derived from sequence analysis, proportion eating at different time-slots of the day, eating main meals, and mean number of meals eaten.

43 Figure 3.2 (Continued) .

44 Figure 3.2 (Continued) .

Table 3.2 Indicators of eating rhythm among unsynchronized and synchronized eaters Denmark

Finland

Unsynchronized Synchronized

Norway

Unsynchronized Synchronized

Sweden

Unsynchronized Synchronized

Unsynchronized

Synchronized

7%

87%

18%

95%

24%

88%

20%

80%

1st meal in lunchintervalsc,b

23%

3%

40%

2%

52%

1%

30%

1%

2nd or later meal in lunch intervalsc,b

37%

87%

26%

74%

22%

74%

29%

86%

2nd or later meal in country-specific dinner intervalsd,c

73%

90%

61%

68%

61%

73%

41%

61%

Mean number of mealsb

3.10

4.36

3.09

4.18

2.93

4.00

2.91

4.20

Ate breakfaste

75%

97%

64%

95%

64%

97%

76%

98%

1st meal in countryspecific breakfast intervalsa,b

e

Ate lunch

72%

93%

56%

80%

58%

83%

53%

86%

Ate dinnere

91%

95%

74%

72%

88%

91%

74%

72%

Ate snacke,f

37%

66%

36%

56%

29%

50%

34%

59%

a

45

Breakfast intervals: Denmark: Eating before 9 am, Finland: before 10 am, Norway: before 10 am, Sweden before 9 am. b Pooled data from 1997 and 2012 (Denmark N = 3247; Finland N = 3244; Norway N = 3256; Sweden N = 3309). c Lunch intervals: Denmark: 11 am–2 pm, Finland: 11 am–2 pm, Norway: 11 am– 2 pm, Sweden: 11 am–2 pm. A time slot was defined as lunch interval if at least 11% of the country-specific sample in 2012 reported having eaten lunch. d Dinner interval: Denmark: 5 pm–8 pm, Finland: 4 pm–7 pm, Norway: 4 pm–7 pm, Sweden: 5 pm–7 pm. A time slot was defined as dinner interval if at least 13% of the country-specific sample in 2012 reported having eaten dinner. e Data from 2012 only (Denmark N = 2060; Finland N = 2044; Norway N = 2079; Sweden N = 2065). f All instances of self-reported ‘in-between’ and ‘evening meals’ were included as snacks.

46

Everyday Eating

pattern norm (or four-meal pattern, if snacks are included) just as much as synchronized eaters but do so with a very different temporal rhythm. However, the figures on the right-hand side of Figure 3.2 reveal that cluster 1 tends to deviate from this: here a considerably lower proportion of cluster 1 reports eating breakfast (64 per cent) than we find in clusters 2–5 (95–99 per cent), and similarly for lunch (58 per cent as against 76–92 per cent, respectively). There is no difference in the proportion eating dinner, while snacks are eaten less often (cluster 1: 29 per cent; clusters 2–5: 34–53 per cent). Interestingly, on average, cluster 1 also contains considerably fewer eating events than the other clusters (cluster 1: 2.93 meals per day; clusters 2–5: 3.72–4.19). Our finding of a temporal deviant cluster does not support the popular idea that the lack of temporal norms, or deviation from them, leads to constant nibbling. The distinctive features of cluster 1 suggest that it is ‘out of tune’ or ‘unsynchronized’ with the society in which it is located. This is not to say that this eating rhythm is socially uncoordinated with, say, peers and family. The suggestion is merely that the social anchor points here are not derived to the same extent from the harmonization of eating rhythms with such macro-societal institutions as normal working hours. Table  3.2 displays numerical results similar to those in Figure  3.2 for all countries. Cluster 1, which is labelled ‘unsynchronized’, is isolated and presented alone, whilst clusters 2–5, which are labelled ‘synchronized’, are collapsed and presented en bloc. It can be seen that the features described above in connection with Norway characterize the unsynchronized eating rhythm in a strikingly similar way across all countries. Those who are unsynchronized eat fewer times during the day, and do not synchronize their eating events with collective hours. In addition, they adhere less to the three-meal pattern, reporting the taking of breakfast, lunch and a snack less often.

Is de-synchronization associated with other dimensions of destructuration? Table 3.3 features additional indicators with relevance to some of the assertions about the destructuration of meal culture that were described in the introductory section of this chapter. It can be seen that the picture underlying vagabond feeding and grazing – one of people eating in isolation from each other at random locations, on an ad hoc basis and in a hurry – does not seem to be more typical of unsynchronized eaters. Thus, across countries, there is hardly any

Table 3.3 Propensity to eat meals alone, at home, in a hurry, in places other than at table, while watching television, and Dietary Quality Score Denmark

Finland

Unsynchronized Synchronized a

Norway

Unsynchronized Synchronized

Sweden

Unsynchronized Synchronized

Unsynchronized

Synchronized

37.6%

37.5%

52.6%

41.7%

42.1%

40.2%

42.7%

35.7%

Ate meal at home b

73.3%

72.1%

78.4%

75.2%

72.7%

70.8%

75.7%

70.7%

Ate in less than 10 min.c

19.7%

23.6%

29.1%

28.5%

23.6%

26.4%

17.4%

20.3%

Ate meal sitting at tabled

53.3%

64.4%

60.2%

75.2%

46.1%

56.9%

59.6%

68.7%

Watched televisione

37.1%

27.7%

32.3%

24.0%

38.3%

29.9%

34.1%

26.0%

Unhealthy

29.2%

16.1%

37.0%

19.4%

27.7%

15.1%

38.0%

20.7%

Average

61.6%

66.0%

59.0%

74.1%

64.4%

71.2%

58.5%

73.3%

Healthy

9.3%

18.0%

4.0%

6.5%

7.9%

13.7%

3.5%

6.0%

Ate meal alone

Dietary Quality Scoref

a

Computed as percentage of all meals eaten with information about social companionship (1997 and 2012 data pooled: Denmark N = 12422; Finland N = 11021; Norway N = 11964; Sweden N = 12380). b Computed as percentage of all meals eaten with information about social companionship (1997 and 2012 data pooled: Denmark N = 12338; Finland N = 11023; Norway N = 11968; Sweden N = 12380). c Computed as percentage of all meals eaten with information about meal duration (1997 and 2012 data pooled: Denmark N = 12976; Finland N = 13079; Norway N = 12148; Sweden N = 12854). d Either kitchen table or table in dining room. Computed as percentage of all meals eaten at home with information about place of eating (1997 and 2012 data pooled: Denmark N = 8921; Finland N = 8344; Norway N = 8507; Sweden N = 8735). e Computed as percentage of all meals eaten at home with information about activities during eating (1997 and 2012 data pooled: Denmark N = 8923; Finland N = 8345; Norway N = 8510; Sweden N = 8760). f 2012 data (Denmark N = 2007; Finland N = 1969; Norway N = 2034; Sweden N = 2029).

47

48

Everyday Eating

difference between those who are synchronized and those unsynchronized in the frequency with which meals are eaten alone, or eaten at home, or in the duration of the relevant meals.

Unsynchronized eating – is there an increase over time, and can the pattern be explained by socio-structural factors? The unsynchronized cluster displays a systematic increase between 1997 and 2012. This can be seen from the totals in Table 3.4, where in Denmark, Norway and Sweden the share of the unsynchronized rhythm rises by approximately five percentage points. The increase in Finland is much greater, with more than a doubling of the share of the unsynchronized rhythm during the study period (approximately 12 to 25 per cent). Results from the multivariate logistic regression analysis (Table 3.5) suggest that the increase is indeed a real change over time, since the year variable remains

Table 3.4 Prevalence of unsynchronized eating rhythm on weekdays and at weekends, employment status categories, and in total Denmark

Finland

Norway

Sweden

17.1% 32.6%

9.3% 21.1%

15.7% 26.8%

16.9% 29.9%

22.7% 41.2%

24.0% 26.5%

19.8% 30.0%

20.8% 36.3%

21.4% 40.7% 42.6% 28.0% 23.9% 18.9%

22.6% 30.4% 43.6% 38.1% 15.9% 34.0%

20.5% 33.5% 49.2% 31.7% 24.0% 31.4%

23.6% 34.0% 45.1% 40.8% 19.7% 26.0%

23% 28.2%

11.5% 24.8%

17.8% 22.8%

20.4% 25.3%

1997 Weekdays Weekends 2012 Weekdays Weekends Employment statusa Working day Day off Unemployed Student Pensioner Other Total 1997 2012 a

Data from 2012 only (Denmark N = 2060; Finland N = 2044; Norway N = 2079; Sweden N = 2065).

Table 3.5 Logistic regression results. What explains unsynchronized eating? (Weekdays only)

49

Year (ref: 1997) 2012 Education (ref.: elementary school) Medium (high school or vocational) High (short to long-term education) HH income (ref.:1 quintile) 2. quintile 3. quintile 4. quintile 5. quintile Would/could not answer Social class (ref: manual class) Intermediate Salariat No information Employment status (ref: in work) Unemployed Student Pensioner Other

Denmark (N = 2184)

Finland (N = 2431)

Norway (N = 2436)

Sweden (N = 2306)

B

S.E

B

B

B

0.380**

0.135

1.141**

S.E. 0.145

0.379**

S.E. 0.131

0.370**

S.E. 0.132

0.284 −0.018

0.156 0.164

−0.057 0.057

0.172 0.201

−0.048 −0.388*

0.170 0.181

0.098 −0.193

0.158 0.166

−0.398 −0.100 −0.171 −0.349 −0.218

0.196 0.191 0.215 0.238 0.186

−0.513* −0.370 −0.430* −0.273 −0.122

0.204 0.214 0.190 0.236 0.183

−0.452* −0.323 −0.335 −0.321 −0.374

0.181 0.200 0.213 0.226 0.197

0.140 −0.089 −0.604** −0.263 −0.132

0.178 0.184 0.205 0.221 0.197

−0.182 −0.341* 0.320

0.148 0.168 0.212

−0.247 −0.240 −0.214

0.155 0.170 0.211

0.139 0.318 0.151

0.159 0.170 0.231

−0.085 −0.259 0.192

0.150 0.164 0.218

0.369 −0.505* 0.339 −0.285

0.238 0.231 0.218 0.306

0.200 0.211 0.243 0.256

0.500 0.000 0.605* 0.432*

0.347 0.241 0.241 0.205

0.519* −0.131 0.522* 0.082

0.218 0.227 0.240 0.261

0.776** 0.062 0.120 0.133

50

Table 3.5 (continued)

Gender (ref.:female) Male Residential area (ref. below 100,000) 100,000 or more inhbs. Household type (ref: single) 2 adults HH with children Age (ref: 60–80 y.) 15–29 y. 30–44 y. 45–59 y. Constant −2 log likelihood (Nagelkerke’s R2) (p < 0.05; = *, p < 0.01 = **)

Denmark (N = 2184)

Finland (N = 2431)

Norway (N = 2436)

Sweden (N = 2306)

B

S.E

B

B

B

0.321*

0.112

−0.329**

0.119

−0.126

0.118

−0.111

0.120

−0.260*

0.118

−0.013 −0.684**

0.147 0.175

−0.233 −0.292

0.157 0.165

−0.368* −0.543**

0.156 0.163

−0.331* −0.502**

0.149 0.165

0.933** 1.031** 0.529* −1.770** 2114.87 (0.085)

0.250 0.227 0.219 0.308

1.264** 0.969** 0.481* −2.716** 2083.28 (0.142)

0.275 0.256 0.244 0.333

1.639* 1.378* 1.036* −2.496** 2192.85 (0.079)

0.267 0.244 0.236 0.342

1.131** 0.741** 0.431 −1.840** 2129.98 (0.087)

0.268 0.251 0.243 0.322

0.549**

S.E. 0.118

0.575**

S.E. 0.113

0.380**

S.E. 0.113

The Daily Rhythm of Eating

51

statistically significant when the effect of socio-demographic factors is taken into account. The regression results also reveal that socio-structural factors related to stratification do not in general explain the extent to which eating is, or is not, synchronized. Thus, neither social class, household income, nor education shows statistical significance systematically across the four countries. In Denmark and Sweden residence in a larger city (over 100,000 inhabitants) decreases the probability of an unsynchronized eating rhythm. With respect to employment status, in Finland and Sweden the unemployed are significantly more likely to have an unsynchronized eating rhythm than those in work. In Norway, pensioners, and the residual ‘other’ category encompassing, for example, people on maternity leave, are significantly more likely to be unsynchronized than those employed, while in Denmark students are significantly less likely than others to be unsynchronized. In general, women are slightly more synchronized in their eating patterns than men in all countries. However, the results from the regression analyses show that socio-structural factors related to life course (age and type of household) predicts the synchronization of eating rhythm consistently (with the exception that household differences are not statistically significant in Finland (a level of significance of 0.05)). Not surprisingly, people living alone have a higher incidence of unsynchronized rhythm than those in households with two or more adults, and in particular households with children. A likely reason for this is that, for those living alone, it is simply easier to ‘de-synchronize’ because they have no need to coordinate their meals with the divergent work and leisure activity schedules of other family members. Turning to age, the clear trend is that older people eat more frequently in accordance with synchronized time than the young. In this respect, the adoption of societal eating rhythms can be seen as a gradual process facilitated throughout life by the habit of timing of one’s daily activities to meet with the commitments of work, family and social life.

Timing of eating with working life The interpretative claim made above – that the adoption of a collective rhythm is facilitated by gradually increasing commitment to work and societal institutions in general – is corroborated further by the fact that eating rhythm is associated with the social timing of working life. Thus, in both 1997 and 2012, and in all countries, the unsynchronized rhythm is more prevalent at the

52

Everyday Eating

weekend, where coordination with work is unimportant, than it is during the working week (see Table 3.4). With the exception of Finland in 2012, this greater prevalence at weekends is statistically significant. From Table 3.4, it can also be seen clearly that individuals who work during the day exhibit a substantially lower prevalence of unsynchronized eating than others in the active work force and students. There is some variation across the countries with respect to the prevalence of unsynchronized eating among students, and among individuals in the workforce having the day off, relative to people who are working. However, most significantly, in all countries the probability of an unsynchronized eating rhythm among the unemployed is approximately double (and even more in Norway) than it is among those working.

Discussion Addressing the assertion, made in popular discourse and by sociological writers, that our food habits are becoming destructured, this study set out to determine whether culturally shared temporal eating rhythms are discernible in the Nordic countries in 1997 and 2012, and whether significant changes have taken place over the period defined by these time points. This was examined in two ways. Firstly, we looked at the population-wide timing of daily eating. Here we noted some interesting changes in daily eating patterns in all of the Nordic countries. This included the fact that eating in the mornings was starting later, and that there was an emergence of less distinctive, or accentuated, peak hours of eating. Some of the changes were common to all or most of the countries, some observable only in one of them. By and large, however, these changes were by no means dramatic, and the basic temporal pattern of daily eating, including the national differences, remained largely intact in the years 1997 and 2012. Equally, in all countries, and in both years, it was clear that the most prevalent number of meals eaten during the day was three and four meals. This corresponds well with the three-meal pattern, with or without a further snacking episode – the predominant schedule of daily eating in industrialized societies (Rotenberg 1981). Clearly, the idea that grazing and vagabond eating are coming to dominate eating practices in contemporary Nordic society is far from accurate. Moreover, our results corroborate findings made in Belgium, where temporal destructuration was identifiable in neither 1988 nor 1999, no changes were detected over time, and a three-meal pattern was the norm (Mestdag 2005).

The Daily Rhythm of Eating

53

This conclusion does not rule out the possibility of within-country differentiations in which smaller social groups have deviant temporal eating rhythms. Indeed, in the second part of the analysis an unsynchronized eating pattern was identified. A late start to the eating day, the displaced timing of eating, and a smaller number of eating events, characterizes this pattern. In it, main meals like breakfast and lunch are skipped to a much greater extent. Furthermore, we found that this unsynchronized sub-population has grown between 1997 and 2012. Since the division into a synchronized and unsynchronized group is a country-invariant finding, it seems that it is universal and not a culturally idiosyncratic phenomenon. The interesting question is: what exactly are we dealing with here? Can the division be regarded as a general, time-invariant phenomenon, emerging from structural differentiation in society (e.g. social stratification or life course factors)? Or does the growth of the unsynchronized pattern over the study period signal the appearance of a new generational practice? The latter suggests that the idea of a general temporal destructuration of food culture has something to it. Our findings are ambiguous. They do not directly support one of these explanations over the other. Although the unsynchronized pattern did indeed grow in prevalence from 1997 to 2012, the increase was relatively modest – except in Finland, which, we note, also entered the study period at a considerably lower level than Denmark, Norway and Sweden. At the same time, a clear structural determinant of unsynchronized eating was identified: in addition to its association with gender, it is much more frequent in single-occupier households and among young people. Since the age effect was very clear in all countries and in both years, our data suggest that de-synchronization typifies a specifically situated life phase – essentially, the phase young singles find themselves in – and will pass when restrictions associated with family life and work exert their influence. This conclusion finds further support in the fact that there were clear relationships between institutional timing and individual timing. Thus, unsynchronized eating was less prevalent among people who worked during the day, while the unemployed and people recording their eating times on weekend days had a greater tendency to be unsynchronized. Despite this ambiguous finding, we believe it is justifiable to extrapolate from our results and claim that collectively shared eating rhythms are likely to persist. Specifically, the fact that life phases and the obligations of working life seem to

54

Everyday Eating

explain the synchronized eating rhythm strengthens the assertion (Rotenberg 1981; Southerton, Diaz-Méndez and Warde 2012) that macro-institutional and micro-instructional patterns are deeply interwoven. This also suggests, as a corollary, that the loosening of daily eating rhythms we did identify is connected with developments in societal institutions and practical arrangements for eating. On this view, the decrease in synchronization can be regarded as a consequence of more extensive changes in the social timing of work that allows – and even demands – more flexibility and individuality in people’s everyday lives. However, the suggestion that a ‘24-hour’ labour market has emerged is not backed up by historical studies. In Belgium the proportion of people working standard hours (roughly 8.00 am to 5.00 pm) in fact grew between the 1960s and the 1990s (Glorieux et  al. 2009). And in Denmark, the proportion of people working in such standard hours did not change over the period 2001–2009 (Bonke 2012). Therefore, setting aside the unlikely possibility that social timing and coordination will vanish from dominant institutions, such as the workplace, collective shared eating rhythms will remain relatively intact. We also examined the idea of a simultaneous rupture of eating culture on several dimensions (temporal, social, spatial, norms). We found that unsynchronized eaters do tend to watch TV more often while eating – something often regarded as a sign of the loosening of social norms. Unsynchronized eaters also exhibit a greater prevalence of unhealthy dietary habits (supposed consequences of gastro-anomy (Murcott 1997)). This last finding, in particular, requires further investigation in studies seeking to explain the relation, if any, between eating temporality and norms surrounding people’s eating practices. On the other hand, unsynchronized eaters did not eat more often in isolation or at random locations, or consume food more hurriedly. Even more decisively, we saw that the unsynchronized eater consumes fewer meals during the day, which flies in the face of the grazing hypothesis. Therefore, our results suggest that the idea of a simultaneous decline in meal commensality, spatial and temporal coordination, and cultural norms guiding cuisine, is doubtful. It was a limitation of this study that different data collection methods were used in 1997 (telephone interviewing) and in 2012 (internet survey). As discussed earlier, this could, potentially, impair comparability, since among other things the differential effect from memory error could produce different results at the two time points. However, the very similar findings, made in 1997 and 2012, regarding food rhythms at the population-wide and sub-population

The Daily Rhythm of Eating

55

level suggest that, in the present study, the use of different data collection methods did not seriously affect the results reported. We would therefore maintain, with due caution of course, that the results from the 1997 and 2012 field methods are comparable. This is not to deny that the results we have presented and discussed would ideally be reproduced using data with a similar collection mode.

56

4

The Social Context and Conduct of Eating Lotte Holm, Drude Lauridsen, Thomas Bøker Lund, Jukka Gronow, Mari Niva and Johanna Mäkelä

Eating is a routinized activity embedded in the flow of activities, events and relations making up everyday life. It is influenced by the broader social organization of time and space in society and its institutional arrangements, and it is regulated by norms and conventions that vary with social and cultural context. Thus, the study of eating patterns is an entry point to the wider study of the social organization of daily life.

Theoretical background There is a common sociological assumption, often repeated in analyses of our times, that our cultures are becoming gradually de-traditionalized and individualized (Giddens 1990; Heelas, Lash and Morris 1996), and that social norms of behaviour – or ‘etiquette’ – are less formal than they once were (Wouters 2007). In connection with food and eating, the concept of gastro-anomy (Fischler 1988b; Fischler 2011) summarizes the idea that eating patterns have become irregular and individualized, and that cultural norms are dissolving. In discussions of modernization this development is often welcomed on the grounds that it fosters the freedom of the individual and enlarges the cultural space of human action (Wouters 2007). However, where food and meals are concerned, the development is often associated with fear of loss of social cohesion (Andersen 1997; Mennel, Murcott and van Otterloo 1992; Fischler 2011). Eating is an important operator of social life and a primary social function (Fischler 2011). In private households, meals have been described as the medium by which families are created and recreated on a daily basis (DeVault 1991; Jackson 2009; Julier 2013), and family meals are seen as important arenas in which children are socialized into central cultural norms and values (Ochs and 57

58

Everyday Eating

Shohet 2006; Fischler 2011). In the social world beyond private households, meals are events, which organize social groups through inclusion and exclusion and make social equality or hierarchy manifest (Douglas and Gross 1981). It is often hypothesized that, in post-industrial societies, eating has become more individualized and flexible, and that eating patterns have become irregular and destructured (Falk 1994; Mintz 1996; Whit 1995). It is inferred that more eating takes place alone, a development that has been linked to concerns about the disruption of family meals and family life in general (Murcott 1997; Jackson 2009). But how we eat, where, when and with whom, is influenced by the social organization of work and family life, and the physical locality of private homes and workplaces, as well as the rhythm and organization of the workday. The daily organization of eating also depends on institutional and practical arrangements in society, such as the size and composition of households, public catering at work places and educational institutions and the accessibility of commercial venues for eating out. It has been suggested that the home may no longer be the primary location for meals, as eating is increasingly taking place in public places and commercial premises such as restaurants, cafés, fast-food outlets, and so forth, and in public welfare institutions such as canteens (Mandemakers and Roeters 2015; Cheng et al. 2007; Dinkins 1992; Dumangan and Hackett 1995; Mogelonsky 1998; Julier 2013). It follows that we might expect a decline in eating at home in the Nordic countries, and a rise in eating out. By tradition, eating is subject to strong normative regulation; in all societies and communities models of proper eating exist, not only with regard to how foods should be ordered, in terms of combination and sequence (Douglas 1975b; Murcott 1982; Fischler 1988a), but also regarding the very conduct of eating. According to the theory of informalization, manners have become less formal and more lenient in the course of the twentieth century (Wouters 2007). Following this theory, we could presume that rules of good eating conduct, or table manners, are not as strict as they once were and allow for more individual variation and heterogeneous behaviour. It has been suggested that the meal is losing its significance as an event in its own right, but increasingly takes place simultaneously with other activities (Gemming et al. 2015; Bugge and Døving 2000; Senauer, Asp and Kinsey 1991; Andersen 1997). Empirical data, which provide systematic evidence for changes over time in the localization and conduct of eating, are rare. Time-use data that trends with regard to time spent eating in the home and outside the home are not uniform across countries (Warde, Cheng, Olsen and Southerton 2007; Cheng et al. 2007; Mestdag

The Social Context and Conduct of Eating

59

and Glorieux 2009; Mestdag and Vandeweyer 2005; Mestdag 2005; Statistics Sweden 2003; Bonke 2002). Evidently, time-use analyses address very few aspects of eating. For a more composite picture of the daily eating practices of populations a more comprehensive analysis of various aspects of the organization of eating is required. Our study is a contribution to this kind of analysis.

Aim of the analysis We aim to analyse in some detail changes in the conduct of eating in the Nordic countries between the late 1990s and the early 2010s. This was a period in which discussions of a possible destructuring of meal patterns intensified. It is an empirical question whether the discussed changes did actually occur during this relatively short period. With the debates described above in mind, we ask whether, over the 15-year period, signs can be identified that suggest that eating has become de-localized from private households, individualized in terms of more eating alone occurring, and informalized, with the arrival of more informal and lenient codes of conduct around eating.

Analysis and data We identified features that were distinctive of meals in 2012 as compared with 1997, and we asked what socio-demographic background factors, if any, could be called upon to explain the changes. We also asked whether and how the changes we identified had affected each of the daily meals (breakfasts, lunches, dinners and in-betweens) by assessing how typical the practices we focused on were for these meal events in 2012. We addressed our overall research interest by asking specifically about what changes, if any, could be identified in the following domains: localization of eating; social company in which eating takes place; and the conduct of meals, by which we meant seating arrangements, activities occurring alongside eating and the duration of eating events. The main part of our questionnaire had enquiries about the respondents’ eating events the day before the interview. The questions focused on the time of eating and the nature of the meals; also, on the social context and conduct of the meal, including where the eating took place, in what social company, the duration and the seating arrangements for the eating event; and finally, whether other

60

Everyday Eating

activities went on while eating was taking place (see chapter 1). In the 2012 study, respondents were also asked to categorize each eating event as one of: breakfast, lunch, dinner, late evening meal, in-between, or other. In the 1997 survey, questions about social context and the conduct of eating events were asked for all hot-food events, but only for those cold-food events in which the respondents reported having had more than one food or drink item. We therefore excluded events from the analysis in which only one (cold) food or only one drink was taken. In 1997 there was a technical problem with the filters in the Finnish questionnaire, and as a result the contextual questions were posed only for eating events in which hot food was eaten. Thus, in most analyses of the 1997 data Finland had to be excluded; in these analyses the sample including only the other three countries was 3,596 in 1997 and 6,155 in 2012. Further, in the regression analyses respondents with missing socio-demographic data were removed. This reduced the sample sizes to 3,143 in 1997 and 5,376 in 2012. The central measures we employed to examine trends in de-localization, individualization, and informalization were based on five questions, all of which were put after the time and content of each eating event had been registered. These questions focused on the following themes: place of the eating event (de-localization), social context (individualization), duration of the event (informalization), activities while eating (informalization), and seating arrangements (informalization). Details of the response options offered to respondents can be seen in Table 4.1. The reported number of eating events differed between individuals (See Chapter 3; Lund and Gronow 2014), and this needed to be taken into account in order to obtain a comparative metric across respondents (in e.g. propensity to eat solitary meals, and propensity to watch TV while eating). We therefore chose to work with relative proportions instead of absolute numbers. To do this, we computed the number of times each respondent reported the activity in question. These counts were then translated into proportional measures by dividing them with the person’s total number of meals eaten during the day. In the analysis, in an initial effort to identify the largest changes, we first outlined descriptive statistics (in average percentages) for all five measures in 1997 and 2012. We did so in a relatively explorative manner, and all of the original response options offered to the respondents were retained. Following this, the measures for which changes across time were detected were examined in more detail. More specifically, we set out to examine which of the socio-demographic factors, if any, could explain the identified changes in eating alone, eating short

Table 4.1 Average percentages of social context and commensality of meals for respondents 1997 and 2012 in Denmark, Norway and Sweden Denmark 1997

Norway 2012

1997

Sweden 2012

1997

2012

Mean

Std.D

Mean

Std.D

Mean

Std.D

Mean

Std.D

Mean Std.D

Mean

Std.D

Location Home Work/school Someone else’s Cafe/restaurant Other n

75 12 6 2 3 1183

(29) (21) (17) (10) (12)

71 16 6 3 4 2042

(31) (24) (17) (12) (12)

74 15 4 2 4 1170

(28) (21) (14) (10) (14)

72 15 5 3 5 2061

(29) (22) (14) (11) (15)

73 14 6 5 2 1243

(29) (22) (17) (13) (8)

72 14 5 6 4 2052

(29) (22) (14) (14) (13)

Company Alone Family Friends Colleagues Others n

33 50 8 10 1 1183

(36) (39) (21) (20) (8)

38 43 7 12 3 2042

(34) (36) (18) (21) (11)

36 43 7 12 2 1170

(35) (37) (18) (20) (10)

41 42 6 11 2 2061

(32) (35) (16) (20) (11)

37 41 9 13 1 1243

(35) (38) (20) (22) (8)

36 46 7 13 2 2052

(34) (38) (18) (21) (11)

Duration 40 min n

11 29 29 10 21 1183

(20) (29) (27) (18) (27)

22 37 25 9 7 2042

(25) (29) (25) (17) (16)

14 41 28 8 9 1170

(22) (32) (28) (17) (19)

26 39 22 8 5 2061

(27) (29) (24) (16) (13)

15 37 26 8 14 1243

(23) (28) (27) (16) (22)

15 (22) 41 (30) 27 (27) 10 (19) 7 (15) 2052 (continued)

a

61

62

Table 4.1 (continued) Denmark 1997 Mean Activitiesa,b,c TV Radio Reading No activities Computer, tablet etc. n Seating arrangements Kitchen/dining table Coffee/sofa tables Elsewhere Did not sit n

21 26 12 44

Norway 2012

Std.D

Mean

Std.D

(30) (33) (23) (38)

36 14 12 33 13 1924

(36) (27) (23) (36) (26)

58 30 10 3 1924

(39) (36) (23) (11)

1131

1997 Mean 26 21 10 46

Sweden 2012

Std.D

Mean

Std.D

(31) (31) (23) (38)

35 12 14 34 14 1957

(35) (25) (25) (35) (27)

51 40 5 4 1957

(38) (37) (17) (14)

1126

1997 Mean Std.D 22 22 18 39

2012 Mean

Std.D

(30) (32) (26) (37)

31 10 18 36 13 1961

(36) (24) (28) (36) (27)

(34) (29) (14) (6)

63 28 7 2 1961

(38) (36) (20) (9)

1194

a,c

72 23 3 3 1131

(34) (32) (12) (12)

65 31 2 2 1126

(36) (35) (11) (12)

72 19 4 1 1194

Note: Where respondents answered ‘don’t know’ they are registered as missing. a Multiple response option. b Differences with regard to categories in 1997 and 2012: In 2012 respondents were also asked whether they used computers/smart devices. c Variables were only measured for meals taken in the home. Respondents with no meals at home were omitted from this part of the analysis.

The Social Context and Conduct of Eating

63

meals, eating somewhere other than at a kitchen/dining table, and eating while watching TV. These four measures were recoded from proportions (as described in the previous paragraph) to dichotomous variables (see Table 4.2 for details) and then used as dependent variables in a series of logistic regression analyses. In addition to the year variable, the following variables were inserted into the regression analysis as (main effects) independent variables: gender, age, education (low, medium, high), social class (working class, intermediate, salariat (Rose and Harrison 2007)), household type (cohabitant, living alone), occupational status (working/student, retired, unemployed or homemaker), and whether the eating events reported were from a weekday (Monday–Friday) or a weekend day (Saturday/Sunday). We then entered interactions effects between year and all background variables. The interaction effects were entered one by one into the regression models. For each of these, we examined whether the omnibus likelihood ratio test of the model with interaction effect was better than the main effects model (at the 0.05 level of significance). If this was the case, we then examined whether the main effect from the year variable was altered (in the direction of the effect, from a significant to an insignificant effect, or vice versa). We report likelihood ratio tests of the interaction model where both of these conditions are met. All of the analyses were conducted and outlined separately for each of the three countries included. Since, in 2012, new response options to the question about parallel activities were added (using a computer and using tablet, smartphone, or similar device), we also decided to examine the sociodemographic predictors of these activities. We modelled this analysis using the main effects variables that are described above (except for the year variable). Plausibly, the social implications and significance of eating alone, or while using a computer, tablet or smartphone, or eating very fast may depend on the meal event in question (e.g. dinner or mid-morning snack). In order to examine how these key practices were related to different meal events, we turned to logistic generalized estimating equations (GEEs). These account for the correlated nature of the data (the data was correlated because each respondent is represented with several observations in the data, i.e. all eating episodes during the day). Five GEEs were conducted. The dependent variables were: (1) using a pc, tablet, or smartphone, (2) watching TV, (3) not eating at kitchen/dinner table, (4) eating a short meal ( = 0.05 = not significant. aN for meals alone and duration: only respondents that recorded answers on variables used in the following regressions. bN for TV and computer etc.: respondents who had at least one meal at home – and valid answers on variables used in regressions.

67

68

Everyday Eating

Table 4.3 reports the four regression analyses. The analyses suggested that the increases in short meals (Denmark and Norway), eating while watching television, and eating somewhere other than at the kitchen/dinner table, was a general change in the populations’ practices, and not exclusive to certain subgroups, since the year variables remained significant when the effect of sociodemographic factors was taken into account. This was also the case regarding eating alone in Denmark. However, patterns were more complex with respect to this in Norway and Sweden. Here, the slightly higher propensity to eat alone in Norway in 2012 (see Table  4.1) was explained by an increase among the employed/students and the unemployed/homemakers (cf. the interaction effect between employment status*year). In Sweden, there was a clear interaction effect between age*year, where the three youngest age brackets in particular increased their probability of eating alone from 1997 to 2012. Despite this, we still observe a slight general decrease in the probability of eating alone in 2012, in Sweden. This is explained by a lower proportion of single households in the 2012 Swedish data compared to the 1997 data (see online supplementary material for sociodemographic details). The regression analyses also showed that the ways in which socio-demographic factors were related to the five meal characteristics differed somewhat. Not surprisingly, eating alone was first of all related to living in a one-person household. In Denmark and Norway, eating alone was less likely among people that were employed or in education. The significant employment*year interaction effect, though, showed that in Norway this difference attenuated from 1997 to 2012. In Norway, it was more likely for people to eat alone during weekdays, than during weekends. Eating short meals was more common among women than men and among the young than the old. They were also more common during the weekdays than during the weekends. In all countries eating somewhere other than at a dinner table was more likely among respondents living alone. In Sweden it was less likely among those with older age, and most likely among the young. Watching TV while eating was in all countries related to living alone and to age, as the oldest age group was less likely and the youngest age group the most likely to do so. In Norway it was associated with being unemployed or being a homemaker; in Sweden it was less likely among employed/students, and in Denmark it was more likely among those with lower educational attainment. In all countries, eating while using computers, smartphones or tablets was more prevalent among the young, those living alone, and more likely among people with higher educational attainment, than those with secondary education.

Table 4.3 Meal characteristics and socio-demographic variables. Logistic regression analyses for Denmark, Norway and Sweden

2012 (1997 (ref)) Woman (Man (ref)) 15–29 years 30–44 years 45–59 years >60 years Employed/Student Retired Unemployed/ homemaker Living alone (cohabitant(ref)) Tertiary (ref) Below Secondary Secondary Salariat (ref) Working class Intermediate class Constant N Psuedo-R2 Nagelklerke

Eating alone > 2/3 of meals

Short meals (any meals < 10 min)

Watching TV at any meal

Not at kitchen/dinner table > 2/3 of meals

PC, tablet, etc. 2012 only

Denmark Norway Sweden

Denmark Norway Sweden

Denmark Norway Sweden

Denmark Norway Sweden

Denmark Norway Sweden

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

OR

1.582** 0.955 — 0.963 1.074 1.204 — 2.243*** 1.629*

1.314* 0.928 — 1.305 1.279 1.123 — 1.759** 2.383***

1.362* 1.108 — 0.923 1.244 1.516 — 1.445 2.466***

3.532*** 1.511*** — 0.717* 0.523*** 0.399*** — 0.579*** 0.859

2.801*** 1.562*** — 1.160 0.835 0.421*** — 1.096 1.021

1.029 1.194* — 0.921 0.738* 0.483*** — 0.878 1.190

2.053*** 1.293** — 0.857 0.775 0.739 — 1.078 1.501*

1.967*** 1.143 — 0.907 0.784 0.901 — 0.902 1.855**

1.412*** 1.041 — 0.786 0.705** 0.493*** — 1.338 1.216

2.286*** 1.155 — 0.587*** 0.437*** 0.276***

2.710*** 0.912 — 0.580*** 0.397*** 0.198***

2.569*** 0.998 — 0.502*** 0.420*** 0.170***

1.086 1.367

1.293 1.019

0.868 0.958

— 1.077 — 0.616** 0.286*** 0.175*** — 1.378 1.599*

— 1.000 — 0.513*** 0.359*** 0.167*** — 1.078 1.070

— 1.164 — 0.382*** 0.256*** 0.116*** — 0.951 1.618*

13.694*** 8.737*** 11.045***

1.179

1.115

1.308**

1.565*** 1.349**

1.394***

3.794***

3.216*** 3.116***

1.715***

1.907*** 1.809***

— 1.214 0.956 — 1.395 1.167 0.033 2798 0.361

— 1.099 0.944 — 0.971 1.169 0.459 2798 0.153

— 1.079 1.101 — 0.875 1.037 0.456 2825 0.112

— 1.034 0.950 — 0.896 0.986 0.743 2896 0.036

— 1.438** 1.299* — 0.931 1.028 0.613 2656 0.053

— 1.032 1.053 — 1.004 0.924 0.942 2769 0.025

— 1.190* 1.451 — 1.012 0.983 0.171 2656 0.096

— 1.355* 1.382 — 0.866 1.225 0.241 2701 0.087

0.849 0.718* — 1.131 1.010 0.674 1710 0.108

0.739 0.597** — 0.989 1.001 0.951 1704 0.109

— 1.245 1.120 — 1.285 1.092 0.055 2825 0.235

— 1.034 0.637** — 1.510* 1.206 0.041 2896 0.308

69

Exponentiated coefficients. *p < 0.05, **p < 0.01, ***p < 0.001.

— 1.242 1.155 — 1.191 1.212 0.762 2701 0.037

— 0.879 0.984 — 1.383 1.242 0.167* 2769 0.098

1.041 0.655** — 1.149 1.122 0.787 1692 0.140

70

Everyday Eating

The social context and conduct of specific meal types, 2012 The question is, then, whether or not the changes we have identified pertain to eating events traditionally characterized by some formality and ceremony, such as main meals like dinner, or whether they relate mostly to in-betweens, which by tradition are more informal eating events (Douglas and Gross 1981). In Figure 4.1 the propensity to eat alone, to eat short meals, to eat somewhere other than at the kitchen/dinner table, to watch TV, to use a pc/tablet/smartphone at different meal events is displayed for each of the four countries. These propensities are based on predicted probabilities calculated from the logistic GEE, where meal event significantly predicted all activities in all countries. Figure 4.1 shows that in all countries eating alone was most typical of breakfast meals and in-betweens. It was rarer for lunches, and even rarer still for dinners (the latter was not the case in Finland, however). The very short meals of less than ten minutes’ duration were highest for in-between eating events in all countries except Sweden, where ‘other’ events were equally often very short. However, between 20 per cent (Sweden) and 40 per cent (Norway) of breakfasts were also very short, and in Denmark and Norway between 15 and 20 per cent of lunches were very short (probably due to the cold lunches in these countries), while this was only rarely the case for dinner in any of the countries. Eating at places other than at the kitchen or dinner table was most typical of the late evening meal, of in-betweens, and of ‘other’. This was the case in all countries, but the prevalence of this kind of eating varied between countries; it was most prevalent in Denmark and Norway but less so in Finland and Sweden. Eating while watching TV followed a similar pattern. It was most typical of late evening meals and it was more prevalent in Denmark and Norway than it was in Sweden and Finland. Eating events characterized as ‘other’ also very frequently took place while watching TV. In Denmark around 40 per cent of dinners took place while watching TV. This was less frequent in the other countries, ranging from 25 to 30 per cent. It is worth noticing that watching TV while eating breakfast was less frequent than for dinner in all of the countries; between 15–20 per cent of breakfasts appeared to take place in front of the TV. The frequency of using a pc, tablet or smartphone while eating was quite similar at breakfast, late evening meal, in-betweens and ‘other’, but was rarer at lunch, and in Denmark and Norway it was even less common at dinner. In Sweden the use of such devices was equally rare during dinner and lunch (5 per cent of these meals), while in Finland it was rarer during lunch than it was during dinner. The overall picture presented by Figure 4.1 was therefore that social context and conduct during eating varied greatly over different meal events. While

Figure 4.1 All figures are calculated predicted probabilities from Logistic GEE results. In addition to the main explanatory variable of interest here (meal event type), age and living a. b alone and their interaction (age*living alone) were included as control variables; aDenmark N = 5761; Norway = 5544; Sweden = 5547; Finland N = 5871. bDenmark N = 8207; Norway = 7949; Sweden = 7822; Finland N = 7935.

71

72

Everyday Eating

breakfast appeared to be characterized relatively often by eating alone, fast eating and eating while using a computer or the like, this was not the case to the same degree for the other main meals.

Discussion Our analysis showed that changes in the social organization of eating over the period 1997–2012 were by no means dramatic. Still, some signs pointed towards processes of individualization and informalization in eating. Our data showed that the place where Nordic peoples eat had not changed over the 15-year period we investigated. The home and the workplace were the main arenas for meals, and going out to eat in restaurants or cafés was still a relatively rare activity. Thus, the changes we could identify in the period of study related to social company, and even more markedly to the conduct of eating; in 2012 there were some small changes regarding propensity to eat alone. In Denmark, more eating alone took place, and in Norway there was an increase, which took place first of all among employed/students. The opposite trend, i.e. lower probability of eating alone in 2012, was observed in Sweden. More eating took place in haste, while watching television, and using computers or smartphones was quite frequent in 2012. Correspondingly, in 2012 people ate less often than they did in 1997 at dinner tables, and more often at coffee and sofa tables and ‘other places’. The regression analyses showed that some of the studied activities were explained to a large extent by socio-demographic factors, especially people’s age and whether or not they lived alone. However, even when we took such factors into account, there was still a statistically significant increase between 1997 and 2012 in many of the indicators of informalization and the indicator of individualization (in Denmark), suggesting that we were probably dealing with general cultural changes. The changes we found did not seem to apply evenly to the different meal events of the day. Dinner, by tradition the most formal and ceremonial meal, was less often eaten alone than all other meal types, and less often taken quickly, and while using a computer, tablet or smartphone. However, it was quite often eaten while watching television. There were some differences between countries in the way the conduct of eating varied across meal types. In Finland and Sweden, for instance, a lower percentage of respondents reported having fewer short lunches and lunches with computer technology than in Denmark and in Norway, a difference that is likely to reflect the different lunch

The Social Context and Conduct of Eating

73

traditions in the two pairs of countries; hot lunches prevail in Sweden and Finland, cold lunches in Denmark and Norway (Holm et al. 2015). While eating quick meals and eating alone suggest that meals may have lost some of their social significance, it is important to remember that these practices did not apply evenly across all meal events. Especially worth mentioning is that eating quickly and alone was much less frequent for dinner than it was for other meal events. It could therefore be the case that the importance of the dinner, which is often seen as the main meal in families (Julier 2013), was increasing relative to other meals. On the other hand, watching television while eating dinner was quite frequent, and this may be interpreted as showing that there was less socializing during dinner in 2012. It is interesting that the use of computers, tablets and mobile phones was not very typical for dinners. How is this to be aligned with the idea that dinner remains the pre-eminent social meal event? Perhaps watching television is capable of being a social activity that people partake in together, while the use of computer technology is more stubbornly solitary. The fact that it was television, and not the computer or tablet screen, which was the more typical dinner activity may be taken to suggest that, in the new developments around dinner, we are seeing, not less socializing, but socializing of a new sort. It was a strength of our data that they are records of micro-behavioural aspects of eating on specific days rather than more generalized accounts of habitual practices. Presumably more exact information is received when asking in this way about specific events rather than general habits. The questionnaire inspired by the 24-hour recall method worked well in achieving such exact data on specific events (Illner et al. 2012). However, individuals’ daily eating varies from day to day, and it is a weakness that our method only covers 24 hours of eating. Repeated 24-hour recalls would have provided some measure of individual variation in daily practices, but for resource reasons this was not possible. Further, we did not ask about whether respondents used computers, tablets or smartphones while eating in 1997, and our data therefore cannot document a change in this activity. However, there is good reason to presume that eating while using such devices is a new phenomenon, as in 1997 tablets and smartphones did not exist, and only very few households (in Denmark, for example, less than 10 per cent (Statistics Denmark 2016)) had access to internet. It was a strength of our data that they reported details about the social conduct and context of eating at different points in time. Such data are rare. There are, therefore, not many sources with which we can compare our results. Time-use

74

Everyday Eating

studies provide one option, as they are repeated and, like our data, based on careful registration of activities during specific days (not, as is often the case, on respondents’ reports on general and usual habits). In the Nordic countries, analyses of time-use data show that people spent more time eating in 2000 than they did in the late 1980s and early 1990s, but more detailed analysis of the context and character of eating events is rare (Statistics Sweden 2003; Bonke 2002; Varjonen and Peltoniemi 2012). One Finnish study shows, however, that the increased duration is explained largely by coffees and snacks, not by ‘proper’ meals (Varjonen and Peltoniemi 2012). This may be taken to resonate with our finding that short eating episodes were more frequent. Where other aspects of eating are concerned, the comparison with time-use studies from other countries reveals mixed results. While our data suggested that eating at home did not change, it declined according to analysis of time-use data from the early 1970s to the end of the 1990s in the USA, the UK, Norway and the Netherlands (Warde et al. 2007). This is an earlier period than that addressed in our study, but a Dutch study from a period closer to ours (1975–2005) has recently shown a steady decline in home-eating in the Netherlands (Mandemakers and Roeters 2015). While eating out in restaurants, cafés and the like remained quite infrequent in the Nordic countries in our data, it increased quite dramatically in France and the UK, but not in Norway and the Netherlands, between the early 1970s and late 1990s (Warde et al. 2007) (see also Yates and Warde 2015a for a more recent analysis of survey data from United Kingdom). The study from the later period in the Netherlands showed a very clear increase in eating out in restaurants or takeout outlets (Mandemakers and Roeters 2015). While it has been suggested that the duration of eating episodes at home did not change in the UK (Warde et al. 2007), there were more short meals in the Nordic countries according to our data. On the other hand, our finding that eating alone in Denmark and eating while watching television increased in the Nordic countries is similar to results based on Belgian time-use data (Mestdag and Glorieux 2009; Mestdag 2005). Can we conclude, then, that Nordic eating patterns were indeed characterized by de-localization, individualization and informalization? Hardly any signs of delocalization could be identified. Private households or institutional arrangements in workplaces and educational centres were still the most important sites for eating. Thus, unlike findings from other countries (Cheng et al. 2007; Mandemakers and Roeters 2015), our data showed that commercialization, in the form of more eating out in restaurants and cafés, was not a major trend in the Nordic countries. However, commercialization may take other forms, such as increased

The Social Context and Conduct of Eating

75

use of ready-made meals or other convenience services. Further studies are needed in order to decide whether, and to what extent, this could be the dominant form of commercialization in the Nordic countries, or whether food provisioning in these countries remains, to a greater extent than it does in other countries, privately organized. With respect to individualization, we could detect some signs of diminishing sociability in eating, especially in Denmark. Quite naturally, living alone as well as being unemployed, or a homemaker, or retired, resulted in more solitary meals, which clearly underlines the fact that, in general, frequent eating in company depends on being routinely exposed to the social company of others, either in the household, or at work, or in an educational setting. However, while neither this nor any other socio-demographic factor accounted for the changes over time in Denmark, the identified change in Norway was explained by changes going on among people in employment or studying. Where informalization is concerned, our data suggested that considerable slackening of eating etiquette had taken place. In 2012 we saw more fast eating, more eating on the sofa, and more eating with parallel activities, such as TV, but also that eating while using other screens, such as TVs, computers, tablets and smartphones was frequent. Eating quickly was more typical among younger people, but the quite remarkable shortening of eating time that took place (especially in Denmark and Norway) appeared to be a general tendency. At the same time, the data showed that meals of lengthy duration had become less frequent. Together these trends could then, on the one hand, reasonably be interpreted as a sign of the diminishing social and cultural importance of daily eating events in the lives of the Nordic people. On the other hand, when we bear in mind that in 2012 a considerable proportion of meals took place while using computers, tablets or smartphones, we see that there is another interpretation, namely that new forms of sociability involving the use of these items had emerged. The survey data used in the present analysis clearly did not provide a full picture of the social organization of eating. While data on frequency of simple activities at different points in time may show that practices have changed, more research of the character of eating practices is needed, including research employing qualitative methodologies, to decipher the meaning, implications and consequences of new patterns of eating. This would provide a basis on which more could be learned about the implications of the changes identified in this study for ordinary daily life in the four Nordic countries. While potentially negative health implications of those practices of eating that were on the rise

76

Everyday Eating

have been suggested (See Chapter 10; Holm, Lund and Niva 2015), more scrutiny is needed to ascertain whether, for example, watching TV while eating is in conflict with socializing and with the transfer of cultural values to future generations, and whether the move from dinner tables to sofa tables is in fact accompanied by informalization in table manners. While such changes clearly mark a move away from traditional and dominant rules of etiquette, we need to know more about the codes of conduct and forms of social interaction that are modified in the process.

5

Family Meals on the Decline? Lotte Holm, Thomas Bøker Lund, Drude Lauridsen, and Jukka Gronow

During several decades the fate of family meals has been a recurrent theme in public debate and in research. While in the late 1990s a decline in family meals was often anticipated, empirical research has shown that the fear that families have stopped eating together, is not warranted, at least in a European context (Fjellstrom 2009). While it is debated to what extent the universality of the family meal was ever a historical reality (Cinotto 2006; Ekström 1990; Murcott 1997), there is less doubt that the family meal is an idea that continues to be of monumental significance, sometimes described as a romanticized myth created to moralize about modern eating patterns (Sobal et al. 2002) or modern family life (Murcott 1997). In research, the significance of family meals is a theme within social science, in evolutionary and developmental studies and in public health. Family meals are linked to discussions of social cohesion, and social behaviour; to child socialization and children’s achievement of language skills and knowledge; and to public health in the broadest sense. Even though the practice of assembling family members on a regular basis for a shared meal has not vanished, the frequency of this phenomenon and the way it is practised may have undergone change, which potentially could imply that the importance ascribed to family meals has changed too. In this paper we analyse current changes in the frequency, organization and conduct of family meals, which have happened over a 15-year period in the four Nordic countries, Denmark, Finland, Norway and Sweden.

The significance of family meals In social research, family meals are linked to theories of commensality. In his discussion of the role of commensality for social life, Fischler (2011) argues that 77

78

Everyday Eating

eating is a primary social function because in food procurement cooperation is essential and this involves reciprocity and redistribution (Fischler 2011). Following Simmel (2001), he underlines that commensality is an essential dimension of the common meal which finds its most salient expression in its daily social occurrence. Commensality creates bonds between those sharing their food, and meals regulate social life and individual behaviour. All cultures have rules and customs regulating arrangements of commensal eating events, distribution and sharing, as well as table manners (whether eating takes place around a table or not). One obvious effect of commensality is socializing individuals into specific rules involving cooperation. It follows that shared meals are important places for socialization, and from an evolutionary perspective it has been suggested that food sharing might be one very important of several acts of socialization, that contribute to the development of pro-social behaviour. Throughout human history the acquisition, distribution and consumption of food were closely tied to cooperation and morality within and across families. Inspired by this, a recent empirical study confirmed a hypothesis to the effect that people who have often engaged in shared meals during childhood had a more pro-social personality as compared to others who did so less frequently during childhood (De Backer et al. 2014). In sociological research the significance of sharing meals for family cohesion is highlighted. Thus, family meals are seen as the medium through which the family unit is produced and reproduced on a daily basis (DeVault 1991). According to interview reports from several countries, family meals are in many households viewed as a priority in spite of practical obstacles in daily life (Brannen, O ’Connell and Mooney 2013; Bugge 2005; Ekström 1990; Iversen and Holm 1999; Jansson 1993; Green et  al. 2009). A study from the United States concluded that this was because, compared with other family time, the conventions of the family dinner table, the norms for behaviour, and the structured consumption sequence are viewed as facilitating and enforcing intimacy and collective identity. The weekday evening meal represents a symbolic coming together and a nostalgic rendering of family time as a refuge from the outside world (Gutierrez et  al. 2008). Similarly, on the basis of cross-national data, it has been argued that frequent family meals may serve as a predictable and stable aspect of family life that helps to promote security within the family (Davidson and Gauthier 2010). Family meals are sites for socialization of children. Ochs and Shohet (2006) discuss how this socialization takes place through apprenticeship and language socialization. Through active observation and direct participation in meals and

Family Meals on the Decline?

79

through mealtime communication, children are socialized to commensality through which they learn their own position relative to others; they are socialized into culturally divergent symbolic, moral, and emotional meanings associated with food and eating, and they learn about ways of acting, thinking and feeling in the world (Ochs and Shohet 2006). Thus, for children’s experience with language and its social use, mealtime conversations are viewed to be important. While the amount and style of communication varies across social classes and ethnic groups, in both middle- and working-class families, mealtime talk incorporates discussions and explanations of current events, world knowledge, and even abstract general principles (Snow and Beals 2006). Conversation during family meals is considered important for children’s performance in school. Analysis of mealtime conversations in American families showed, for example, that there were positive relationships between narrative and explanatory talk at mealtimes during preschool years, and five-year-old children’s scores on literacy-related measures which are heavily drawn on in school (Beals 2001). Social interaction and conversation are frequently highlighted as the important factor which ensures the significance of family meals for family cohesion and child socialization. However, it has been maintained that the food itself is an important feature. For family meals to be arenas for moral socialization they need to be events where foods are shared, i.e. participants need to eat from the same pot. When several people share the food, issues of fairness and morality become important; each individual needs to make sure that enough is left for the others to be satisfied, and this is central for how family meals are arenas for learning pro-social behaviour. This aspect is thought to be more decisive than whether or not people eat at the same time; even if meal participants eat at different points in time, they still need to exert the self-control necessary to ensure that the food is shared in a fair manner (De Backer et  al. 2014). In a similar vein, others argue that the benefits of having a family meal can be undermined if the family consumes fast food (Martin-Biggers et al. 2014). An example of this appears in an empirical study from the United States where mothers complain that they are unable to establish proper family meals, because of the availability of individually packed ready-made meals, which family members prefer to the home-cooked meal (Ochs and Beck 2013). In public health literature empirical studies highlight the positive effects of eating family meals, especially for children and adolescents. Thus, positive associations are reported between frequent family meals and overall diet quality (Fulkerson et al. 2014; Woodruff and Hamming, 2008) and lower risk of obesity (Martin-Biggers et al. 2014; Hammons and Fiese 2011; Valdés et al. 2013). While

80

Everyday Eating

most authors maintain that the evidence for the link between family meals and obesity is still inconsistent and weak (Valdés et  al. 2013; Casazza et  al. 2015; Martin-Biggers et al. 2014), family meals have become an important component of health promotion for children (Fruh et  al. 2011), and regular family meals are recommended as part of childhood obesity prevention in the United States (Rao 2008). Further, frequency of family meals is found to be inversely associated with teen substance abuse (tobacco, alcohol and drugs), and disordered eating practices, and positively associated with good language skills, academic performance and family cohesion and connectedness (Martin-Biggers et  al. 2014). In public health literature the mechanisms involved in the positive impacts of family meals are often not discussed. But one factor is seen to impede the positive impacts of family meals, and is frequently discussed. Watching TV while having family meals is associated with poor diet quality (Feldman et  al. 2007; Blass et  al. 2006), poor health for children (Fiates et  al. 2008) and higher prevalence of overweight and obesity for both children and adolescents (Liang et al. 2009; Fuller-Tyszkiewicz et al. 2012). This suggests then, that also in public health it is the social interaction and conversation, which are crucial for the positive effects of family meals, but which are hampered by TV viewing. In all, in research, family meals are assigned a significant role in health; in children’s literary, language experience, and school performance; in the cohesion of families; in the socialization of younger generations not only into cultural norms and values related to eating, but also into pro-social behaviour and different ways of acting, feeling and thinking in the world. It is worth underlining, that as the practice of having family meals is quite widespread, it is not possible to decide whether the virtues and benefits assigned to the family meal, are in fact linked to ordinary family life in general, and not to the family meal as such. Still, the question of whether or not the practice of having family meals changes, is of interest; whether or not it is seen as the central factor for child socialization, health, and social cohesion, or merely an indication of ordinary family life.

Family meals in decline? Whether or not family meals are in decline has for decades been a topic on public agendas in many countries, but qualitative studies confirm that in current daily life, the idea of sharing meals on a daily basis with family members is still important, also in Nordic countries (Jansson 1988; Ekström 1990; Bugge and

Family Meals on the Decline?

81

Døving 2000; Iversen and Holm 1999). At the same time, several authors have questioned whether family meals were ever a regularly practised and stable event, as often anticipated in public debate. Thus, Ekström underlined that ‘the shared meal was not, historically speaking, shared by everyone, as not everybody belonged to the sharing group’. Thus, women, servants and children each had a different social status and ‘everybody did not eat from the same pot’ (Ekström 1990:75). Murcott proposes a similar argument in her discussion of how family forms and household structures have varied considerably between social classes, leaving considerable variation with respect to, e.g. whether women or children were part of family meals (Murcott 1997), and in a small case study of households in early twentieth-century West Yorkshire, Jackson and colleagues found great variations in whether and how families ate together (Jackson et al. 2009). Murcott suggests that the ideal of the family which is expressed in concern for the decline of the family meal, is probably an ‘ideal-typical model of the middleclass and (respectable) working-class family’ (Murcott 1997). Cinotto shows, in a historical account of family meals in the United States, that current ideals of family meals refer to a brief history of middle-class practices in the 1950s. She suggests that media representations of family life from this period are the reason why these specific practices have come to epitomize traditional family meals (Cinotto 2006). Parallel to this, Jackson argues that the family meal is a venerated social institution; a myth we live by; more an ideal than a daily practice. Thus, the persistent idea that family meals are on the decline is a case of moral panic, a proliferation of poorly informed public debate with disregard for solid evidence (Jackson et al. 2009). Time-use studies provide data that are suited to analysing how the prevalence of family meals develops over time in more recent history. But, to our knowledge, there are only a few studies addressing this question, and results are equivocal since studies are conducted in different national settings, and at different time periods. Mestdag and Vandeweyer found that in Belgium over a 30-year period from the mid-1960s to 1999, time spent eating with family went down by almost half – from 51 to 27 minutes per day (Mestdag and Vandeweyer 2005). Cheng and colleagues found that in the United Kingdom time spent eating in the home declined between 1975 and 2000, but as the duration of eating episodes was stable, the authors interpreted that eating took place in the company of others, and concluded that the temporal pattern of eating and drinking at home remained remarkably stable between 1975 and 2000. However, no data about the company of eating episodes were presented in this study (Cheng et al. 2007). On the basis of these two studies it is not possible to draw any conclusions

82

Everyday Eating

about current developments in family eating, and there is thus a lack of more recent data and comprehensive analyses.

Social organization of family meals In addition to the question of whether family meal-eating is a current and prevalent phenomenon is the question whether the form of such meals has changed. As described in chapter 4 in this book it has been suggested that codes of conduct and rules of etiquette have undergone a process of informalization, according to which, norms and rules have become more lenient and less formal, and our Nordic surveys confirmed that trends supporting this idea could be identified in our data material. But would that also be true for family meals? Evidently, new practices with little or no conversation between family members at meals would have great impact on the processes of socialization supposed to take place at such meals. There are strong traditional images of what a family meal should look like: All family members should be present, they should sit around a table, eating the food that is prepared by the mother in the family. They should engage in friendly conversation about events in their daily lives and matters of the world, and they should enjoy the emotional bonds between family members which are created and confirmed during this daily encounter. Already Bourdieu, in his analysis of festive meals in the French working class and bourgeoisie of the 1960s showed that working-class meals were often structured differently in terms of codes of conduct. The timing of eating would not necessarily be synchronized between those sharing the meal, and young children would perhaps eat their dessert in front of the television (Bourdieu 1984). While this analysis does not relate to ordinary everyday meals, newer studies indicate that current family meals often display similar lenient codes of conduct. Thus, Brannen and colleagues found in a qualitative study of 40 dualearner British families, that shared family meals would often take place with the adults sitting on the sofa, the children on the floor and everybody enjoying TV programmes (Brannen et  al. 2013). Also, in American middle-class families, eating in separate rooms, at different times, and eating different foods were characteristics of family meals in some families. A plethora of individually packed convenience foods to be found in American supermarkets, and stocked in large quantities in the households, supported a practice of accommodating individual tastes of both adults and children and often undermined maternal

Family Meals on the Decline?

83

ambitions to establish meals where foods were shared (Ochs and Beck 2013). In both these studies, as in most research about family meals, it is a stable phenomenon that the responsibility for arranging family meals is assigned to women (Wood 1995). Providing loving care to family members through food and family meals has been described as central in feminine identity (Fürst 1997). However, overall, the gendering of cooking has changed in recent years (See Chapter 8; Holm, Ekström, et al. 2015) and the question is whether responsibility for family meals is now increasingly shared between women and men. While research confirms that ideas about the organization of family meals are quite persistent (Jackson 2009), there are qualitative studies that indicate that realities in modern life are quite different (Ochs and Beck 2013). However, to our knowledge, no large-scale studies about this exist and data about how the organization and conduct of family meals differ between population groups are lacking.

Aim of analysis In the following we analyse family meals in the four Nordic countries in 1997 and 2012. We focus on how typical such meals were, how they were distributed across populations and aspects of the organization and conduct of such meals. Families eat different kinds of meals together, and in research, definitions of family meals vary with respect to which kind of meals are eaten and how many family members need to be present (Martin-Biggers et al. 2014). In sociological studies, the term ‘family meal’ often refers to cooked meals eaten in the evening, when family members have returned home from their daily activities outside the household. In studies, this meal is often viewed as the main meal of the day in terms of time spent preparing and eating it, and the degree of ceremony and value and the importance assigned to the meal in terms of gathering family members (DeVault 1991; Fjellstrom 2009; Julier 2013). In the present study we adopt a rather strict definition of family meals, closely related to the ideals of main meals in families highlighted in social research. We are interested in family meals as arenas for child socialization, family cohesion and social interaction and we have therefore chosen to focus on events that are shared by all household members. Thus, we operationalize family meals as cooked meals (‘hot meals’ is the term used in our study) which are eaten at home in multi-person households with all family members present and eating at the same time. We analyse the frequency of having family meals, in 1997 and 2012,

84

Everyday Eating

and how it distributes across socio-demographic groups. We also investigate whether or how the conduct of family meals has changed over the 15 years between our two studies. We analyse whether key aspects of meals, which in earlier chapters 3 and 4 were shown to change between 1997 and 2012, also changed for family meals. Thus, we ask whether the gendering of cooking family meals has changed (See Chapter 8; Holm, Ekström, et al. 2015), and whether the duration, sitting arrangements and conduct of family meals have changed (See Chapter 4). Further, we focus on factors that have been raised as important concerning the presumed impact of eating together on family cohesion and the socialization of children, namely whether families shared the same food, or whether they watched TV while having family meals (De Backer et  al. 2014; Martin-Biggers et al. 2014). The data analyses employed (which include chi2 tests, Spearman’s correlation coefficients, logistic regression, and predicted probabilities from logistic regression) will be described in the relevant passages of the result section.

Have family meals become more or less frequent from 1997 to 2012? First, we ask whether Nordic peoples eat fewer family meals in 2012 than they did in 1997. Figure  5.1 shows frequencies of eating family meals separated between weekend days and weekdays. Figure 5.1 shows that, in general, there are only slight changes in the frequency of family meals. In Denmark, Norway and Finland the changes in prevalence of family meals are not statistically significant. The only country in which a statistically significant change has taken place is in Sweden. Here, the family meal has increased in prevalence. In weekdays, 56.5 per cent had a family meal in 1997, while 61.5 per cent had this in 2012, which is a borderline statistical significant increase (Chi2 3.795(1); p = 0.051; N = 1620). In weekends in Sweden, 64.9 per cent had a family meal in 1997, while 72.7 per cent had this in 2012 (Chi2 4.205(1); p = 0.040; N = 663). We conducted logistic regression analyses for each country separately in order to analyse socio-demographic variations in eating family meals. The analyses showed that in Sweden and Finland those in employment had family meals less often than students, retired or unemployed people. This was not the case in Denmark and Norway. In all countries, however, age and household composition were significantly related to having family meals. Next, we therefore

Family Meals on the Decline?

85

N 1997, 2012: Denmark: 758,1417; Norway 857,1506; Sweden 781,1511; Finland 812,1452

Figure 5.1 Prevalence of family meals in the Nordic countries in 1997 and 2012 – divided into weekdays and weekends.

take a closer look at how different combinations of these two variables relate to family meals. In Figure  5.2, the probability of having a family meal across different theoretically defined life-phase groups is presented. Results are based on predicted probabilities from logistic regression conducted separately for each country and each year within countries. Judged from the patterns observed in Figure 5.2, for each year, the frequency of having family meals is influenced by a combination of family composition and age, which again reflects phases of life. In Denmark, Sweden and Finland we find the least family meals in younger households without children. When small children (aged 0–13 years) enter the picture, family meals are more frequent, and then they decline again in households with teenage children (14 years or older), only to increase somewhat in the older households without children, and even more so in the eldest households. In Norway in 1997, families with teenage children have even fewer family meals than young households without children. The figure also shows that in all countries, differences between life-phase groups are somewhat smaller in 2012 than in 1997. In 2012, the frequencies of family meals are quite similar in the oldest age groups, and national patterns differ more among the younger age groups. In 2012, Finland has the lowest frequency of family meals in all life phases except the oldest. Except for the three younger age groups in Finland, more than half of the populations in all four countries had a family meal yesterday.

86

Everyday Eating

Figure 5.2 Family meals and life-phase groups 1997 and 2012. Prevalence (in percent) of family meal according to age and household composition (with or without either small or teenage children) – (Sample sizes (1997/2012): Denmark N = 658/N = 1327, Norway N = 770/N = 1437; Sweden N = 712/N = 1412; Finland N = 766/N = 1346).

We now take a closer look at more details of family meals. First, we turn to the question as to whether or not families actually shared family meals, i.e. shared the same food, or, whether they ate different foods, accommodating individual tastes and preferences. We then focus on the arrangement and conduct of eating family meals, and on who cooked the meal. The importance of sharing the same food at family meals has been highlighted both from a theoretical point of view (De Backer et al. 2014) and from empirical

Family Meals on the Decline?

87

studies of the difficulties with arranging family meals in American households where a plethora of individually packed ready-made meals are available (Ochs and Beck 2013). Following an interest in this, we asked in the 1997 survey whether, at family meals, ‘anyone was served specific food’ and in 2012, ‘Did everybody eat the same food?’ The two questions are not identical, but both address the issue of sharing food versus accommodating individual preferences by serving individualized meals. In 1997, only in 12–13 per cent of the cases were individual preferences accommodated, while in 87–88 per cent of the cases all family members ate the same food. In 2012 this had increased to 91–94 per cent in the four countries (change was significant, at 60 Education Basic (ref) Vocational High school 1–3 years after >3 years

Denmark (5.1)

Finland (4.9)

Norway (4.9)

Sweden (5.3)

5.9 (+0.4)

6.6 (+0.7)

5.5 (+0.4)

5.8 (−0.5)

6.2 (+1.0) 6.0 (+0.6) 6.1 (+0.9)

6.7 (+0.91)

5.5 (+0.4) 5.6 (+0.6)

5.9 (+0.6) 6.0 (+1.6)

7.1 (+1.6) 5.7 6.7 (+1.0)

5.3 5.9 (+0.6)

7.0 (+1.3)

6.0 (+0.7)

5.4 5.8 (+0.3) 4.0 5.6 (+1.7) 6.1 (+2.0) 7.4 (+3.3)

4.5 6.1 (+1.6) 7.1 (+2.6) 7.9 (+3.4)

4.5 5.0 (+0,5) 5.5 (+1.1) 6.6 (+2.1)

4.3 5.5 (+1.2) 6.0 (+1.8) 6.9 (+2.7)

5.2 5.5 (+0.3) 5.4

5.8

5.1

5.2

6.0 (+0.6)

6.9 (+1.1)

5.7 (+0.6)

6.3 (+1.0)

5.4

5.4

5.9 (+0.5)

5.8 (+0.4)

the complexity of the hot dinners the most, and the presence of a partner to a lesser extent. Only the presence of children did not have an effect on hot dinner complexity. In Finland the presence of relatives or colleagues increased the complexity of hot dinners. The presence of children or friends did not prove significant in explaining the complexity. Also in Norway and Sweden, the

The Complexity of Meals

105

presence of a partner, relatives or friends increased the complexity of hot dinners. Thus, the presence of children, colleagues or others did not have a significant effect on the complexity in either Norway or Sweden. Meal venue was not a significant predictor for hot dinner complexity in Denmark or Norway. In Finland and Sweden, the hot dinners eaten at someone else’s home or at a café or restaurant were more complex than dinners eaten at home. Hot dinners at weekends and on weekdays were equally complex in Denmark, Finland, and Norway. Only in Sweden were the weekend dinners slightly more complex than those on weekdays. The duration of the dinner had a positive effect on hot dinner complexity in all of the countries, the more complex dinners taking more time. Individual background (gender, age group and education) was also linked to hot dinner complexity. In Norway, women’s hot dinners were slightly more complex than men’s, whereas in other countries there was no gender difference. Moreover, in all the countries, the over-60-year-olds had slightly more complex dinners than the youngest group (age 15–29). The differences related to education were small and few; only in Denmark and Sweden the group with the highest education had slightly more complex hot dinners than the group with basic schooling. While especially meal sociability and duration were linked to hot dinner complexity, none of the independent variables conveyed significant differences in relation to cold dinner complexity. The primary reason for this is that there is not much variation in the complexity of cold dinners between or within the different countries (no data shown).

Discussion In the Western countries popular concerns about changing food habits have to do with sociability, the decline of the family meal, the erosion of national culinary cultures, and the increased popularity of convenience and fast foods (Rozin et al. 2006; Warde et  al. 2007). This potential destructuration of eating practices (Mäkelä 2001) may mean that meals are becoming simpler and shorter, and that in general the meanings of eating in everyday life are changing. In this study, we approached these questions empirically by looking at the complexity of meals in Denmark, Finland, Norway and Sweden, through a comparative analysis of meal complexity in these countries, examining to what extent the social context of the meals as well as social backgrounds have an effect on meal complexity.

106

Everyday Eating

The previous Nordic research from 1997, which was used as a basis for the 2012 study, indicated that eating patterns in the Nordic countries were home based: Most of the eating took place at home either alone or with the family. Eating with colleagues was common on weekdays, but eating with friends was rather rare (Kjærnes 2001; Holm 2001). The findings contradicted other research that has suggested that because Nordic people are active outside home in leisure time, it could be expected that much of the eating would also take place in the public sphere. Our results indicate that still, in 2012, meals at other people’s homes and eating out in cafés or restaurants constituted only a minor proportion of everyday eating. For lunches, commensality had a bearing on the complexity of the hot meal only in Finland. This suggests that lunch has a different kind of status in Finland, resembling the trends that are discernible for dinner in all the other countries. In Finland and Sweden both hot lunches and hot dinners eaten in public, at restaurants, or cafés were more complex than private home lunches and dinners, whereas in Denmark and Norway they were not. For lunches, this may be partly explained by the fact that in Denmark and Norway hot lunches were relatively rare, and with dinners, it may be speculated that in these countries after having eaten a simple cold lunch, people prefer more elaborate hot dinners irrespective of whether they have dinner at home or elsewhere. In general, both the hot lunches and dinners that took more time were more complex in all the four countries. The difference in complexity between weekday and weekend meals was substantial only in hot dinners in Sweden. Despite the fact that at weekends more time and effort is spent on cooking, this is not reflected in complexity as measured by the number of various elements in the meal. However, it may well be that the ingredients, seasoning and cooking methods differ between weekends and weekdays. The elaborateness of Swedish dinners as compared to the other three countries is supported by findings in an OECD study, which shows that in Sweden as many as four out of five of the adult population report cooking daily. Cooking is nearly as popular in Norway and Denmark, although less time is devoted to the activity. Also, in Finland more than 70 per cent of the respondents engage in cooking (OECD 2011). These qualitative differences call for further research on food preparation. The complexity of hot dinners had a strong relation to commensality, but in cold dinners differences were not apparent. Hot dinners within the immediate family differed from each other in that the presence of a partner increased meal complexity, whereas the presence of children had no effect on the complexity compared to dinners eaten alone. The presence of companions from outside the immediate family was unusual in most of the countries. The national meal conventions are

The Complexity of Meals

107

practical and somewhat uniform: Commensal partners and meal venues are determined by family structure and employment status (see also Kjærnes et  al. 2009). Sobal and Nelson (2003) came to the conclusion that usual commensal partners, such as colleagues and family members, are seen as related to routines in meal eating, whereas unusual partners for meals may be rare but socially significant and related to special meal occasions. Our results imply that the presence of members of the immediate family does not increase meal complexity. Having relatives over for dinner had a positive impact on hot dinner complexity across the countries. The presence of friends, colleagues or others, which is exceptional, is reflected in meal complexity (see also Grignon 2001; Holm 2001). Relatively few consistent socio-demographic patterns were found to explain the differences in complexity. In all the four countries, the respondents in the oldest age group had more complex hot dinners than the youngest age group. In Finland and Sweden, the oldest age group also had more complex hot lunches. The educational differences were small, although in Denmark and Sweden the group with the highest level of education had more complex hot dinners. For hot lunches some educational differences were found only in Sweden. Regarding indicators of social differentiation, it could be argued that the increased welfare of practically all social groups has made it possible, economically speaking, to procure a large number of food items for the household irrespective of education and income disparities. This, in turn, could be why there is little social stratification. Thus, regarding age, both the results from this analysis and recent results from Lund and Gronow (See Chapter 3) suggest that age and life-course effects appear to have an important influence on eating practices. It seems that social and situational factors relating to where and with whom the meal is eaten mainly account for the differences in meal complexity. The country comparison carried out in this article is based on the assumption that people within each country have more in common with each other than with people in neighbouring countries (Kuipers 2012), although ideas of convergence between different countries have been displayed in recent research (see Warde et  al. 2007). From a comparative perspective, there were few differences between the four Nordic countries. The similarity may result from similar ingredients being used for cooking across the borders (see Amilien 2012). Thus, the complexity measure is not the best tool to depict these differences, as actual dishes and national peculiarities cannot be captured with a simple count measure. Further research is needed in order to study these specifics, and the relationship between lunch and dinner complexity.

108

7

Eating Out, Having Guests Thomas Bøker Lund

Several chapters of this book reveal clear collective patterns in national food cultures (in timing, when meals are shared or individual, hot or cold lunches), implying that the act of eating is heavily influenced by the temporal and cultural organization of societies. On the other hand, it is well known from other literature that there are quite large socio-economic differences with regard to, for instance, obesity and dietary health (Darmon and Drewnowski, 2008; Drewnowski and Specter 2004). In this chapter, we look at a number of eating related practices where there is reason to assume that there are socio-economic differences. For each practice, we look at the socio-demographic gradient, and whether the gradient has grown, decreased or remained stable from 1997 to 2012. The first kind of practice we will focus on is the extraordinary eating event, like eating in a restaurant or entertaining guests at home. This is a contrast to mundane, everyday eating activities. We will examine socio-economic differences with respect to the act of serving food to guests and frequency of going out to eat, respectively. It is not unlikely that socio-economic differences could be involved in both activities. Eating out could certainly follow an income gradient, insofar as purchase power inhibits or facilitates restaurant visits. Restaurant activity may also be more frequent among higher educated and upperoccupational classes, indirectly because of higher purchase power, and directly because restaurants, at least historically, have facilitated conversations and high status (culinary) experiences, which arguably is a good match for upper-class lifestyles, compared e.g. to going to a football match. Indeed, research from the United Kingdom shows clear differences in restaurant frequency (Warde et al. 1999). Similarly, what is served for guests at home could also be influenced by income differences where serving a hot meal may be difficult to handle economically for low-income households. Occupational factors and education also influence lifestyles, including what appears proper and natural to do at 109

110

Everyday Eating

social events (e.g. Bourdieu 1984). The kind of experiences people are expected to provide to guests could therefore also vary as a function of educational assets and social class. We also look at two attitudinal factors: liking foreign foods, and healthy eating motivation. Both attitudes are a part of people’s overall lifestyles. Lifestyles have been proposed to be associated with social position where, according to e.g. Bourdieu, the upper strata of society will develop attitudes that are detached from mere economic necessity, and thus favour exotic and light food over satiating or cheap food (Bourdieu 1984). This makes it likely that a propensity to appreciate foreign food and motivation to eat healthily is higher in the upper classes. Despite the expectation of a socio-economic gradient in the two case areas that will be studied, there are also arguments to the effect that the gradient is decreasing or has disappeared altogether. Thus, the Nordic societies have become more affluent, and today there is also a wider range of high-end and low-end outlets compared to, say, ten years ago (Horesta 2014), making it much less economically burdensome to eat out. Similarly, the availability of foreign food products in supermarkets has grown, and outlets serving international cuisines have become more diverse and widespread. Everyday confrontation with such foods and dishes could imply that the reservation and distrust by nationally and traditionally oriented cultural segments has declined. Appreciating foreign foods can therefore have become quite common. Finally, according to some accounts, health motivation is an example of a societal innovation and trend, which initially will attract the upper strata of society, but over time, will trickle down to lower societal segments (Mackenbach 2008; Victoria 2000). The stark focus on public health information that exists in the Nordic societies (Vallgårda 2011) could therefore suggest that socio-economic difference in healthy eating motivation is declining following such a trickle-down process.

A short description of the method To portray social differentiation and how it changes over time in the case areas described above, we make use of four measures. In Table 7.1, frequencies and means for each year of these measures are presented. Note that the question ‘Have you had guests during the last week?’ resulted in further questions to the respondent prompting on what was served. It is these follow-up questions that are used in the analysis. To test whether there are socio-economic differences

Eating Out, Having Guests

111

and whether the influence of these has changed over time we conducted four regression analyses for each country where the variables in Table  7.1 were inserted as dependent variables. Social class, education and income (our socioeconomic factors of interest) were inserted as main effect variables together with respondent age, population density in the area of residence and the year (1997 or 2012). Also, interaction effects between year and the three socio-economic factors were included to test whether there were changes in the effect from socioeconomy on the practices studied. During analysis it transpired that change from 1997 to 2012 was significantly affected by age. Indeed, this sometimes explained what at face value appeared to be socio-economic changes. To avoid false conclusions about socio-economic differences and change, the age*year interaction was therefore included in the analysis. In all analyses (reported in Figures  7.2 to 7.5), the predicted probabilities (from logit regressions) or predicted scores (from OLS regressions) are presented, so that the effect from age (and any other confounding variables) is taken into consideration when summarizing the influence from socio-economic factors. Since age turned out to be decisive for some of the studied cases, age results are also displayed. In some of the sections, we also bring in additional analysis from the 2012 data to explore variations between, and patterns within, countries in more detail.

Results Eating at restaurants and cafés – socio-economic differences in all countries and a decrease in the socio-economic gradient in Norway, Sweden, and Finland According to our data, eating out at least once per month has increased significantly in Sweden and Finland. It has decreased a little in Denmark; while there were no changes in Norway (see Table 7.1). In all countries and both years, socio-economic factors are associated with propensity to eat out at least once per month (this is so for all three of the measures: income, education and social class). Specifically, higher social position increases the propensity to eat out (see Figure 7.1). In 1997, the gradient was less marked in Denmark, somewhat higher in Norway, and highest in Sweden and Finland. For instance, in Denmark 49 per cent of those with the highest income and 34 per cent of those with the lowest income ate out at least once per month. The similar figures were 61 versus 41 per cent in Norway, 73 versus 48 per cent

112 Table 7.1 Descriptive statistics in 1997 and 2012 and tests for difference between 1997 and 2012 Denmark 1997 1. Frequency of eating out (avg.) p value (t-test)

2012

Norway 1997

2012

Sweden 1997

Finland

2012

1997

2012

3.50 0.631

3.48

3.72 0.671

3.75

3.89 0.000

4.33

3.55 0.000

4.07

2. Had guests during the last week (%) p value (chi2)

62.3% 0.000

43.0%

71.1% 0.000

48.6%

63.3% 0.000

42.9%

51.1% 0.002

45.4%

Served a hot meal to guests (%) p value (chi2)

28.9% 0.071

26.0%

18.1% 0.010

21.9%

27.4% 0.068

24.6%

16.8% 0.298

15.5%

Served cake and coffee to guests (%) p value (chi2)

27.0% 0.000

16.1%

35.1% 0.000

21.1%

31.2% 0.000

19.6%

30.2% 0.751

29.6%

3. I like foreign food (%) p value (chi2)

73.5% 0.005

77.8%

71.4% 0.000

78.7%

70.4% 0.000

81.3%

44.8% 0.000

59.1%

1.99 0.02

2.10

2.15 0.03

2.02

1.89 0.251

1.94

2.43 0.000

2.09

4. Healthy eating motivation (avg.) P value (t-test)

A Composite variable (range 0 to 5), based on responses to four statements: (1) ‘To what degree did you think about the health or risks associated with the foods you ate yesterday’; (2) ‘I am ready to hold back in my everyday diet if this gives me better health in the long run’; (3) ‘I am more concerned about the taste of food than how healthy it is’; 4) ‘It’s important to stick to the advice of experts on healthy eating’. The original response options were likert scales. However, the 1997 and 2012 data differed in the extent to which the extreme (‘very’) or modest (‘partly’) response options were chosen. Therefore, responses to all four items were dichotomized (to indicate agreement) and subsequently summed.

Eating Out, Having Guests

113

Figure 7.1 Probability of eating out at least one time per month per occupational class, educational groups, income groups, and age in years – in 1997 and 2012. Note: results are based on predicted probabilities from a logit regression.

in Sweden and 61 versus 36 per cent in Finland. By 2012, the difference remained more or less the same in Denmark: 48 per cent in the highest income quintile and 37 per cent in the lowest quintile. The income gradient decreased in the remaining countries: 59 versus 44 per cent in Norway, 76 versus 60 per cent in Sweden and 70 versus 51 per cent in Finland. A similar picture over time is observed with respect to educational differences. Thus, the gradient was lowest

114

Everyday Eating

in Denmark in 1997 and remained at the same level in 2012. In the other countries, the gradient decreased from 1997 to 2012. There were also clear differences across occupational classes. Over time, the most significant change is observed in Finland, where the working class increased the propensity to eat out at least once per month; and with a much higher pace than the salariat. Despite the declining importance of socio-economic differences, where Norway, Sweden and Finland, by 2012, have reached the Danish level, there are still clear socio-economic differences.

Serving hot meals to guests – little socio-economic difference except in Sweden, large difference in kind of food served to guests (in 2012) Having guests come over has declined considerably in Denmark, Norway and Sweden, while it remained stable in Finland (Table  7.1). Having guests was particularly low in Finland in 1997, so by 2012 there is in fact country convergence, where around 45 per cent of the households in all Nordic countries report to have had guests come by their house during the last week. Serving coffee and cake for guests declined in the three countries where guest activity declined (this remained stable in Finland). Propensity to serve a hot meal for guests, on the other hand, only changed modestly from 1997 to 2012. Interestingly, despite a decline in socializing at people’s homes, serving hot meals is still a significant social event when it happens. In Denmark, there is hardly any socioeconomic gradient in serving hot meals for guests (see Figure 7.2). In 1997 there was a modest higher propensity for the higher educated to serve hot meals, but by 2012 this had disappeared. In Norway, the same picture of a common cultural practice among Norwegians emerged. While the lowest income quintile served somewhat fewer hot meal to guests in 1997, by 2012 this difference had disappeared. In Finland, serving hot meals for guests was not associated with socio-economy in any of the study years. In Sweden, however, there are some differences. In 1997, higher social class was positively associated with serving hot meals. In 2012, education but especially income differences were influential. Thus, in the lowest quintile, 20 per cent reported serving a hot meal for guests during the last week, while it was 34 per cent for the highest quintile. Except for Sweden, serving hot meals does not appear socio-economically patterned. Again, it should be mentioned that additional analysis of the 2012 data nuances the picture. In 2012 respondents were asked what they would

Eating Out, Having Guests

115

Figure 7.2 Probability of serving hot food to guests during last week per occupational class, educational groups, income groups, and age in years – in 1997 and 2012. Note: results are based on predicted probabilities from a logit regression.

‘usually serve for family and friends’ when inviting them to dinner at home. Propensity to serve ‘something new and/or a bit exotic’ was significantly associated with educational level and income in all countries. To illustrate, in Norway, 40 per cent of lower educated from the lower income quintile would usually serve this, while the propensity was 70 per cent among highly educated in the highest income quintile.

116

Everyday Eating

Figure 7.3 Probability of serving cake and coffee to guests during last week per occupational class, educational groups, income groups, and age in years – in 1997 and 2012. Note: results are based on predicted probabilities from a logit regression.

Serving coffee and cakes to guests – practically no socio-economic differences Similar to serving hot meals for guest, there are very few socio-economic differences with respect to serving coffee and cakes for guests (Figure 7.3). There was a modest negative association between income and this propensity in Denmark in 1997. However, this had disappeared by 2012. In Norway, there

Eating Out, Having Guests

117

was a negative effect from social class in 2012 (less common to serve coffee and guests in the salariat); but again, a very modest difference. In Finland and Sweden, there were no socio-demographic differences associated with serving coffee and cake in either 1997 or 2012. Specifically for Norway and Sweden, changes in the association between life course and serving coffee and cake for guests appear to be in play, albeit in quite different ways in the two countries. In Norway, serving coffee and cakes increased with age in 1997, while this pattern had disappeared by 2012. Conversely, in Finland serving coffee and cakes was negatively associated with age in 1997, while this pattern had almost disappeared by 2012.

Liking foreign foods – largest socio-economic difference in Finland In all countries, appreciation of foreign foods increased from 1997 to 2012. The increase was relatively modest in Norway, Sweden and Denmark, while it was somewhat higher in Finland (from 44 to 59 per cent), even though this still means that foreign foods are clearly less appreciated in the Finnish population. In all countries, we find socio-economic differences in propensity to like foreign foods (Figure 7.4). However, the differential is not uniform across countries. In Denmark and Norway there are modest socio-economic differences. In Denmark education plays a modest role in explaining liking foreign foods. This association remained more or less intact from 1997 to 2012. In Norway, both education and income is associated with liking foreign foods. This is the case in both years, and in both years the association is quite modest. In Sweden there is a somewhat stronger social gradient especially from education and to some extent income, and this differential was more or less similar in 1997 and 2012. In Finland, we find the most substantial differences between socio-economic groups. All socio-economic factors included in the analysis, but especially education and income, influences propensity to like foreign food. The socio-economic gradient decreased a little from 1997 and 2012 but was still evident in 2012. Another key point in the increase from 1997 to 2012 in propensity to like foreign foods is the finding that this kind of food has become much more accepted in the older generations. Where appreciation of foreign foods decreased significantly with age in 1997, this is much less so in 2012 in Denmark, Norway and Sweden. In contrast, in Finland, appreciation of foods decreased to the same extent in tandem with age in 1997 and 2012.

118

Everyday Eating

Figure 7.4 Probability of ‘liking foreign food’ per occupational class, educational groups, income groups, and age in years – in 1997 and 2012. Note: results are based on predicted probabilities from a logit regression.

Healthy eating motivation – remarkably small socio-economic variation in both years Socio-economic factors play a quite modest role in the propensity to be motivated to eat healthily (Figure 7.5). This is the case in all countries, and in both years. The largest explanatory effect is seen in Denmark where, in both years, education influences motivation.

Eating Out, Having Guests

119

Figure 7.5 Health motivation per occupational class, educational groups, income groups, and age in years – in 1997 and 2012. Note: results are based on predicted scores from an OLS regression. The dependent variable had five values (0 to 4), and is a composite variable based on responses to four statements: (1) ‘To what degree did you think about the health or risks associated with the foods you ate yesterday’; (2) ‘I am ready to hold back in my everyday diet if this gives me better health in the long run’; (3) ‘I am more concerned about the taste of food than how healthy it is’; 4) ‘It’s important to stick to the advice of experts on healthy eating’. The original response options were likert scales. However, the 1997 and 2012 data were different in extent to which the extreme (‘very’) or modest (‘partly’) response options were chosen. Therefore, responses to all four items were dichotomized (to indicate agreement).

120

Everyday Eating

However, even here the explained variation is modest (class, education and income explains R2 0.023 in 1997, and R2 0,010 in 2012). The variation is remarkably low. Indeed, the modest association between education and motivation to eat healthily is also identified in an analysis focusing only on the 2012 data which is presented in chapter 10. The fact that education, social class and income influence healthy eating motivation to a similar low degree in 1997, supports the overall conclusion that health orientations with respect to food and eating does not follow a social gradient in the Nordic countries. This conclusion should be seen in few of the fact that higher education is associated with better dietary quality (see chapter 10). However, apparently this association is not explained by higher health motivation.

Summary: A declining importance from socio-economic factors? Aiming to examine potential importance of socio-economic status on food and eating related practices in the Nordic countries, and identify changes over time, we looked at two case areas; the extraordinary eating event and attitudes to food. Even though socio-economic gradients were identified in both areas, and in both years, the overall impression is that socio-economic differences do not play a particular decisive role in the studied countries. This is perhaps best illustrated by the finding that healthy eating motivation is, at best, only modestly associated with socio-economic differences in both years. Even in the area where socio-economy meant the most, eating out at restaurants and cafés, the socio-economic differences is not particular high. Furthermore, the socio-economic gradient in eating out declined from 1997 to 2012 in all countries except for Denmark. The findings suggest the conclusion that it is primarily in Finland and Sweden that socio-economic differences still remain in 2012. However, this gradient reveals itself in different sectors of food and eating in the two countries; in Finland there is a considerable gradient in the propensity to like foreign foods (see also Purhonen et al 2014: 200–204), while the gradient in Sweden is found in the propensity to serve hot foods to guests. All in all, the results could cautiously suggest that social status differences in eating and food are relatively modest, and appear to be on the decrease. It should be mentioned, though, that the comparison of changes over time was somewhat limited. Particularly when comparing extraordinary eating events we could only

Eating Out, Having Guests

121

look at socio-economic changes in frequency of eating out. A range of qualitative differences in kinds of foods and dishes that are eaten and appreciated could have been relevant in this analysis. The best measure available in the data for appreciation of kinds of food was the propensity to like foreign foods. In fact, for this measure the gradient remained at a stable, however modestly low, level in both years in three of the Nordic countries (not in Finland). Our results indicate that the Nordic social-democratic welfare state (Esping-Andersen 1990) on the whole has done a quite good job in facilitating the development of commonly shared food attitudes and habits regarding extraordinary eating events. It should be mentioned, though, that even though the number of people who have experienced food insecurity are quite low in the four Nordic countries there are recent indications of somewhat higher food insecurity in Finland (see chapter 9; see also Sarlio-Lähteenkorva and Lahelma 2001) and Denmark (Nielsen et al. 2015). This type of insecurity is determined by lack of purchase power (from income restraints). While the attitudes and practices in households facing food insecurity from the outset may not be any different from others, food poverty/insecurity can have a detrimental effect on the involved households’ life circumstance, obesity, and dietary choices.

122

8

Cooking and Gender Lotte Holm, Marianne Pipping Ekström, Sara Hach, and Thomas Bøker Lund

Meals are cornerstones in social life and central parts of caregiving and the building of relations in households. Food preparation and cooking are routine tasks in ordinary daily household work, and the preparation of hot meals has been highlighted as an especially significant example of such work. Historically, this work has been assigned to women and several studies link cooking to female gender roles and identity (Charles and Kerr 1988; Murcott 1982). However, as women increasingly engage in paid work outside households the gendering of cooking is likely to change. Many studies show that in most Western countries the gap between women’s and men’s time spent on general housework has narrowed (Sullivan 2000; Bianchi, Sayer, Milkie and Robinson 2012). While some researchers argue that this is a consequence of structural changes with rising educational and income level of women, other researchers emphasize the resilience of gender inequality, maintaining that household work is not gender-neutral (Coltrane 2000; Hochschild and Machung 1989 cited after Sayer 2005). Despite the narrower gap it is still the case that most household tasks are done by women. Some studies even suggest that the development towards more gender equal engagement in core household tasks has slowed down or even come to an end (Schneider 2011; Bianchi et al. 2012). In order to understand the continued gender gap in time spent on household work it may be important to distinguish between different kinds of household tasks. According to Schneider (2011) some tasks are important fields for ‘doing gender’ as they are imbued with cultural meaning and serve as resource for the construction of femininity for women and as a threat to masculinity for men. Other tasks are more mundane and do not, to the same extent, contribute to the construction of gender identity. Cooking is such a task that is closely linked to what historically has been seen as female identity. Food cooked in private 123

124

Everyday Eating

households is marked by the logic of the gift, which involves an ethic of care – an orientation ‘where one is relating and responding to another’s needs’ (Fürst 1997: 444) and where production is geared towards the structuring of social and emotional relations. The main evening meal is especially important in this respect as it is often the meal of the day that brings household members together. It has been shown how women’s preparation of cooked dinners for their families is seen as a normal and important female duty that confirms gender roles and expectations in households and which many women find pleasurable (Murcott 1983). This could imply that cooking dinner is a specific task, which women tend to keep to themselves. On the other hand, it could also be hypothesized that this task is attractive to men, especially because of the significance of cooked dinners in households (Kemmer 2000). Some time-use studies show that men’s involvement in household work has increased also in tasks related to food. Still, as is the case for other household tasks (Coltrane 2000), most food work is done by women (Harnack, Story, Martinson, Neumark-Sztainer and Stang 1998; Sayer 2005; Sullivan 2000; Schneider 2011; Chesters 2012). Is cooking of dinner especially resistant to change due to the importance of this activity in the construction of female identity? The question can therefore be posed whether and to what extent the convergence between men’s and women’s time spent on household work also takes place with respect to cooking dinner? Is cooking dinner becoming a shared responsibility for men and women, or is it still an activity that is assigned first of all to women? The aim of the present analysis was to contribute to the clarification of to what extent the gendering of cooking dinner is changing. Based on two surveys of eating patterns conducted in four Nordic countries in 1997 and 2012 we analyse changes in the gendering of cooking dinners and ask: who is cooking dinner and how has this changed over the last 15 years? Further, we want to add to the existing literature by examining the socio-demographic variations in how gender division of cooking has changed.

Cooking as gendered work Historically, men and women have had distinct roles in food production and distribution (Counihan 2012). Almost universally, women are in charge of reproduction: cooking, feeding, and teaching table manners. Counihan suggests

Cooking and Gender

125

that the obligations of cooking and feeding others ‘have been ambiguous sources of oppression, violence, drudgery, power, and creativity for women’ (Counihan 2012: 104), and have given them a channel for creating important social ties. In some qualitative empirical research about cooking in households, cooking is highlighted as dull and repetitive work that women do in the service of others (Kerr and Charles 1986); in other studies cooking is highlighted as an important activity through which women mark themselves as ‘adequate’ women and thus central for women’s female identity (DeVault 1991: 18; Murcott 1983). Much of the research on gender relations and food preparation in households was conducted in historical contexts in which women predominantly worked as housewives and were economically dependent on their male partners as breadwinners. More recent qualitative studies of the gendered division of domestic labour show changes in expectations around gendered household roles and responsibilities (Magnusson 2006; Singleton and Maher 2004; Bugge 2005; Roman and Peterson 2011; Murcott 2000). This probably reflects historical developments, and particularly the fact that women have experienced rising educational attainment and increasing levels of participation in the workforce (Kemmer 2000). However, while qualitative studies are indicative of changes occurring in discourse about gender roles in households, the disclosure of structural changes in the gendered division of domestic labour requires observations of different points in time. Time-use studies have provided such analyses of changes in gender division of domestic labour (Sayer 2005; Sullivan 2000; Schneider 2011). Such studies show that in all Nordic countries, just as in Great Britain (Sullivan 2000), the United States (Sayer 2005) and Australia (Chesters 2012), women’s general domestic work has decreased and men’s has increased (Bonke 2012; Suomen Virallinen Tilasto (SVT) 2009; Ekström and Hjälmeskog 2006; Statistics Sweden 2012; Vaage 2012). Fewer studies have focused specifically on cooking. A study by Sullivan (2000) suggests that in Britain men’s cooking has increased, as the time they use for cleaning and cooking has nearly doubled between 1975 and 1997. However, it is not clear to what extent this increase applies for both cleaning and cooking. For the United States, Sayers (2005) showed a marked decline in the gap between men and women’s time spent on cooking from 1965 to 1989. Bianchi (2012) discusses how this development was strongest in the period 1970–90, which saw the greatest labour force increase for American women. By the end of the twentieth century, this trend levelled off (see also Schneider 2011). This leads some to suggest that ‘the gender revolution’ is over (Cotter, Hermsen, and

126

Everyday Eating

Vanneman 2011 cited from Bianchi 2012). Time-use studies in Nordic countries rarely focus specifically on cooking. An exception is the Swedish time-use studies, which show that from 1990 to 2010 women’s time spent on cooking declined by 40 minutes, while men’s increased by 28 minutes. However, women still spend about two hours more on cooking than do men (Ekström and Hjälmeskog 2006; Statistics Sweden 2012). Despite (and because of) the continued unequal division of cooking it is relevant to track and understand whether possible changes towards equality take place only in some population groups or in the wider population. Several timeuse studies analyse the time men spend on household work related to age, employment status of partners, gender ideology, life-course issues (Coltrane 2000; Sullivan 2000; Chesters 2012) and income (Schneider 2011), but results vary somewhat between countries and studies, and no studies analyse sociodemographic variations in cooking in the Nordic countries. Specifically, for household cooking, age and social class indices have been suggested as important socio-demographic determinants of change (Murcott 2000). With respect to age-related changes towards a more equal gender division in housework, in some accounts the development is particularly associated with men from cohorts born between 1965 and 1979 (often referred to as the ‘Generation X’) when women’s workforce participation increased and the modern feminist movement gained significance (Gorman-Murray 2008). It is suggested that men born after 1965 are likely to be socialized in dual earner families and thus are more likely to share the breadwinner role with female partners (Chesters 2012). It follows, that this generation can be expected to have been first movers, setting the trend of integrating new masculine routines of sharing household tasks for younger generations to follow and at a faster pace than older generations. Another explanation, discernible in the literature, links the emergence of the new masculinity to class relations, and especially the middle class. Adoption of a new masculinity can be seen as an example of what Giddens terms ‘life politics’ (Giddens 1991) – that is, the politics of lifestyle and choice which form part of the processes of self-actualization in post-traditional contexts. The creation of morally justifiable forms of life and of an ethics of how individuals should live their lives, are examples of the new post-materialist politics associated with new social movements in the middle class (Eder 1985). Post-materialistic values fit well with the struggle of middle class for status and acknowledgement, as they set the middle class apart from the materialism of the lower classes and allow them to challenge the authoritative and universalistic character of the norms and values of bourgeois high culture (Eder 1985). New masculinity is one such post-

Cooking and Gender

127

materialist value. It may be adopted not only as an anti-authoritarian political value, but as a social value of opposition which can be practically realized in daily life. As such it ought to be more common among middle-class men. A third explanation would be that new household gender division of labour is a consequence of structural changes that, functionally speaking, operate similarly in all socio-demographic segments of society. Women’s workforce participation includes women from all social strata and the ideology of gender equality is rather all-encompassing. The Nordic welfare states have implemented social security systems that contain an ideology of gendered equality and this also finds expression in public institutions such as schools and in citizens’ rights. (Melby, Ravn and Wetterberg 2009; Borchorst 2009). Since all citizens disregarding their socio-cultural status, or age, or sex, encounter gender equality discourses it could be argued that new gender roles and identities are slowly bound to become more common in all segments of society. There are a few empirical studies from the Nordic countries showing that, in spite of widespread ideals of gender equality, there is substantial variation in the practical division of labour in middle-class households (Magnusson 2006; Roman and Peterson 2011; see also Murcott 2000 for a parallel discussion based on studies from the United Kingdom). This calls for more specific analyses of the gendering of household tasks, such as cooking.

Aims of the analysis The aim of this article is to investigate changes in the gendering of cooking in Nordic households over the last fifteen years. We focus on dinners, which research show to be seen as important for the social life in households (DeVault 1991; Murcott 1982; Bugge 2005; Iversen and Holm 1999; Ekström 1990; Mäkelä 1996). We use the term ‘dinner’ as referring to a main hot meal eaten in the late afternoon or early evening. First, we ask whether, and to what extent, a more equal gender division of labour has emerged. Do men take a greater part in cooking dinner in 2012 than they did in 1997? Do women cook less frequently? Following this, we analyse socio-demographic variations in men’s and women’s reported cooking, and whether changes in gendered activities from 1997 to 2012 can be linked to specific socio-demographic background factors such as age or social class, or, whether changes are more general and allencompassing. We end by discussing implications and potential explanations for our results.

128

Everyday Eating

Data and methods – design of study Subsample In the present analysis we use a subsample of respondents in which we consider gender-based division of labour to be particularly relevant. We have excluded individuals from single-person households as well as those in single-parent households. We focus the analysis on hot meals eaten at home, between 4.00 and 10.00 pm, in the company of others. The rationale for this criterion is threefold: these are the meals that typically involve the most cooking; they are meals where people other than the respondents may very likely have prepared the meal, and in the survey from 1997, data on cooking are available only for hot meals. In the remaining article we refer to these meals either as ‘dinners’ or as ‘dinners eaten at home in social company’. The subsample thus includes respondents who live in households with at least one other adult (above 20 years) and who had eaten a hot meal in their homes and in social company on the day before the interview. The subsample totaled 4022 respondents: 1268 in 1997 and 2754 in 2012.

Variables in the analysis Central to the analysis is the question whether the man or the woman did the cooking. In the questionnaire, following the registration of each relevant eating event, respondents were asked: Who did the cooking? The following multiple reply option was given: ‘me’, ‘my partner/spouse’, ‘children’, ‘my mother’, ‘my father’, ‘others’, ‘don’t know’. This variable was recoded into fewer categories: ‘man’, ‘woman’, ‘man and woman together’, ‘others’; and in regression analyses even fewer categories were created: ‘man (or woman) participating in cooking the meal’, or not. Respondents were not asked about the sex of their partner and the category ‘my partner/spouse’ was categorized as belonging to the opposite sex relative to the respondent. In order to investigate whether the gendering of cooking relates to sociodemographic characteristics, the following variables were included in the analysis: education, social class, age, household type (1 = two adults, 2 = two adults with children), occupational status (1 = working/student, 2 = unemployed, 3 = pensioner, 4 = other), level of urbanization (1 = area with below 100,000 residents, 2 = area with more than 100,000 residents). Education was classified into three categories: low (primary school), medium (vocational and secondary education) and high (more than 1 years of additional

Cooking and Gender

129

education after secondary education). The household income variable was recoded into quintiles relative to country-specific income distributions within the sample. The social class variable contained three categories, echoing the three-class scheme recommended by Rose and colleagues (Rose and Harrison 2007): working class (lower service/sales/clerical occupations, lower technical occupations, and unskilled work), intermediate (clerical occupations and lower grade technicians/ supervisory occupations) and salariat (employers, professionals, higher grade managerial and technical occupations). Class schemas today (e.g. the EGP scheme (Erikson and Goldthorpe 1992)) typically differentiate social class more finely than is possible with just three categories. However, we decided to introduce three classes in order to ensure conceptual equivalence of our data over time.1

Analysis and measures First, the unadjusted proportion of men, women, man and woman together, or others, that cooked dinner in 1997 and 2012, respectively, was displayed. For this analysis, the complete sample encompassing responses both from men and women was included (N = 4022). Following that, socio-demographic differences in gendered cooking practices, and changes over time, were analysed in a multivariate setting using logistic regression models. This analysis was conducted separately for men (N = 2028) and women (N = 1994), as we only have sociodemographic information available for the interviewee. The dependent variable indicated whether the man or woman, respectively, took part in cooking or not (0 = no part, 1 = took part in cooking). In this model, we also examined whether possible changes from 1997 to 2012 in propensity to cook dinner can be explained by socio-demographic factors by means of inserting interaction effects between year and the following variables: year*age was tested with respect to exploring whether changes over time are related to the ‘generation x’ hypothesis, year*social class, year*income and year*education with respect to the assumption that trends towards gender equality in cooking primarily is found in the middle class, and finally, year*income and year*employment status with respect to the idea that de-feminization of household tasks is caused by changes in workrelated obligations. A ‘general to specific’ analysis strategy was employed, removing insignificant variables at the 0.10 level one by one, and starting with 1

For the 2012 data we constructed the ESeC schema following the procedures outlined by Rose and Harrison (2007). We used a collapsed three-class version of the 1997 class variable, which corresponds, by and large, with the three-class version of the European socio-economic classification system employed in the 2012 data.

130

Everyday Eating

insignificant interaction effects, in a backward stepwise strategy, using Waldtests. P-values