Social Science Research 51 (2015) 205–218
Contents lists available at ScienceDirect
Social Science Research journal homepage: www.elsevier.com/locate/ssresearch
Convenience on the menu? A typological conceptualization of family food expenditures and food-related time patterns Sarah Daniels ⇑, Ignace Glorieux, Joeri Minnen, T.P. van Tienoven, Djiwo Weenas Vrije Universiteit Brussel, Research Group TOR, Sociology Department, 1050 Brussels, Belgium
a r t i c l e
i n f o
Article history: Received 4 July 2013 Revised 23 September 2014 Accepted 29 September 2014 Available online 27 October 2014 Keywords: Household food consumption typology Convenience foods Meal preparation and consumption habits Time- and Household Budget data Latent Class Cluster Analysis
a b s t r a c t One of the most fundamental, but also controversial, food trends of the past years is convenience food. This article investigates the underexplored relationship between the heterogeneity in (convenience) food consumption (a feature of a food culture’s cuisine) and meal patterns (characteristics of a food culture’s structure). This study hopes to illustrate that convenience food can be interpreted both as a means to maintain a food culture’s structure and as a means to overturn it. Latent Class Cluster Analysis is performed using data from the HBS 2005 survey on families’ food expenditures to conceptualize convenience-orientation and to examine the relationships with families’ meal behaviors. Whereas outsourcing cooking is most prevalent among single-person households; two-or more-person households are most likely to buy unprocessed and natural foods and to spend most time cooking and eating in. A higher consumption of convenience food is also more likely to affect individuals’ kitchen than table habits. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction This paper focuses on one of the most important food questions and controversies that mark today’s foodways: to which extent convenience foods have become part of the food consumption patterns of families of different social groups, and to which extent they affect today’s meal preparation and table habits. So far, there is still little quantitative research about both families’ reliance on, or need for convenience, and their day-to-day kitchen and table habits. However, it is precisely the examination of this relationship, that will help to better understand the role of convenience food in a time where the seemingly simple question ‘‘what to eat’’ (Pollan, 2009) has become a fast-growing topic of concern, discussion and insecurity. Historically, food insecurity is not new, but due to many – especially post-war – food trends, innovations, regulatory changes, social and cultural developments and many more, it has become an even more complicated and omnipresent issue in today’s societies. The concerns about food and the ways that people think about food and use it in their daily lives, has never been so strong. One of the most significant changes that has brought this discussion about food to the table, was the way in which convenience food became a defining symbol of the commercialized and industrialized food supply and the ever-changing relationship between people, food and society (Scholliers, 2007, 2014). By using convenience foods, many of the activities, tasks and culinary skills essential to food preparation are taken over by services and industries intended to make cooking less time and energy consuming, both physically and mentally (Darian and Cohen, 1995; Scholderer and Grunert, 2004). Today’s availability of convenience food items goes far beyond the older existing canned products (already available since the early 19th century), varying from canned soups and sauces, cake mixes, ⇑ Corresponding author at: Pleinlaan 2, 1050 Brussels, Belgium. E-mail address: [email protected] (S. Daniels). http://dx.doi.org/10.1016/j.ssresearch.2014.09.010 0049-089X/Ó 2014 Elsevier Inc. All rights reserved.
206
S. Daniels et al. / Social Science Research 51 (2015) 205–218
deep-frozen pizzas, pre-cut vegetables to ready-to-(h)eat meals. Since a few decades, the consumption of convenience food has become more embedded in the way that most people cook, eat and live (Shove, 2003), that it is considered to be one of the largest and most fundamental food innovations in recent years (Brunner et al., 2010; Goody, 2013). The many discoveries, developments and innovations behind convenience food are some of the most crucial foundations on which today’s food industry is built (Capps et al., 1985). Today, convenience food has become such a self-evident part of daily life, that even non-users of convenience food hardly realize how their relationship to food became industrialized and commercialized, going from ground coffee to baked bread, breakfast cereals, margarine, sugar, cold cuts, yogurt, ice cream, and more. However, despite the fact that these commercial food products are generally ready-made (Burnett, 1989), they have become such a taken-for-granted part of individuals’ cooking routines that they are no longer labeled as convenience food, but as natural and ‘‘fresh’’ ingredients (Short, 2003). Despite of the fact that the basic principles of convenience food were already present for centuries (or at least in the supply of canned meat, vegetable and fish products), it was especially from the Second World War on that the food industry started to sell and depict convenience food as an indispensable kitchen ingredient, as a means to taste the comfort of modern life and to save work and time in the kitchen (Shapiro, 2004). The last decades have, however, not only shown many changes in the supply of and demand for convenience food, but also a clear growth in the accessibility and consumption of away-from-home food products as ‘‘the pinnacle’’ of the commercialization of food (e.g. fast food, takeout or restaurant meals) (Scholliers, 2007) and in the design (e.g. kitchens that increasingly became built around appliances instead of the opposite (Van Otterloo, 2000)) and utensils used in the kitchen (e.g. microwave ovens, freezers, steamers, espresso machines, bread making machines, . . .). What these trends seem to illustrate is that cooking not only seems to have become a household task more difficult to organize and schedule (regardless whether one finds it worth the time and effort or not), but also a less necessary task, no longer strictly dependent of the homemaker’s efforts (a task which is still seen as a woman’s responsibility). According to Alan Warde (1999) and Elizabeth Shove (2003), convenience food has increased in importance due to people’s enhanced needs for ease, efficiency, control, and to manage and spend their times (and lives) better. In just a few generations, we seem to have moved from a society in which people still worked for bread and butter, towards ‘a leisure society’ in which consumerism became a ‘way of life’ (Veblen, 1994 [1899]) and more and more people started to feel an ever greater time pressure than ever (Linder, 1970) in which ‘‘life has become a patternless checklist of having been there, done that’’ (Robinson and Godbey, 1997, 44). The many different activities that people nowadays need or want to combine, have become so complex, that people are increasingly looking for ways to save time, including into their home-kitchen. Given this, the increase in time-saving and convenient meal options may seem an obvious and successful evolution, while, in essence, it was, and still is, a somewhat complex, disapproved and contested trend (Algren et al., 2004; Shove, 2000; Warde, 1997). Despite that convenience food is one of the most fundamental food innovations of the last decades, the story of convenience has been far from an instant success and has become subject of plenty of discussion and controversy. Its consumption was, and still is, hampered, by the many social and emotional components attached to food and cooking, as ways of showing care for others, to communicate love, to make others feel welcome, and to symbolize social relationships. According to Fernández-Armesto (2002, p. 250) ‘‘the companionship of the campfire, cooking-pot and common table, which have helped to bond humans in collaborative living, are in risk of being shattered’’ due to the growing trend towards commercially processed food. There are strong claims that food habits would be individualized, fragmented, more irregular and hurried, that meals would be no longer act as social Zeitgebers, or organize and give meaning to time (Zerubavel, 1985), that the ‘‘habitual’’ mealtimes would break down and that food would have evolved to a state of gastro-anomy or normlessness (Fischler, 1979). Home-cooked meals are believed to be increasingly replaced by time-saving meal alternatives and snacks, to enjoy less priority, to be more likely to be eaten outdoors, alone, and on the run (Pollan, 2013). And much, if not all of these problems, are ascribed to the rise of this ‘‘convenience revolution’’ in individuals food spending patterns (Burnett, 1989, 310). Despite of the fact that both the production and consumption of convenience food are criticized for their possible destructive impacts on the environment, public transport and health (Kjaernes et al., 2007), the grazing-hypothesis is the main starting point of this article. This hypothesis assumes that people would increasingly eat in an incidental manner regardless of socio-cultural norms regarding food, time and context (Davis, 1995). Although, it has already been shown that contemporary meal habits (both in Belgium, as in other European countries or in the USA) are still bound to collective and customary rhythms, routines and social norms (Mestdag, 2007; Warde et al., 2007), the studies that have cooked this grazing-thesis, only used information about the ways in which individuals’ and families’ meal habits are organized in time and space, without taking into account any information about their food intakes, purchases and demand for convenience. However, it is only by examining these associations that consumers’ food consumption choices, and their kitchen and table habits can be fully understood (Marshall, 2005). To examine to which extent convenience food influences household’s cooking and eating habits, it is crucial to make the distinction between a food culture’s structure and a food culture’s cuisine. Whereas a cuisine is more likely to refer to the food products, recipes, cooking methods and kitchen appliances that people use to prepare their meals; a food culture’s structure is characterized by the way that people organize their meals into their daily life routines and habits (Rozin, 1982). The main difference between both food culture components, is that a structure is more steadfast and that a cuisine is, thus, more open to, for example, influences from foreign cuisines and more flexible to change towards a food melting pot (Askegaard and Madsen, 1998; Schrover et al., 2005). Perhaps this difference can
S. Daniels et al. / Social Science Research 51 (2015) 205–218
207
be best illustrated by the fact that people have more difficulties to adapt to foreign food culture’s structures (such as other mealtime hours), while at the same time they can look forward to the local cuisine and food specialties. The main goal of this article is to investigate (RQ1) to which extent convenience food has become part of Belgium’s food culture cuisine, (RQ2) to which extent it has become more embedded in the food habits of some groups in society than in others, and (RQ3) to which extent households’ reliance on convenience food in food provisioning is related to how they cook and eat (which can be understood as important indicators of a food culture’s structure). By doing this, we will be able to find out whether the reliance on convenience food has a similar impact on individuals’ kitchen habits, as on their eating habits. By spending attention to consumers’ purchasing behavior, food consumption patterns and home-kitchen and eating habits, and by combining a consumption-oriented, with a situation-oriented approach, we can examine to which extent convenience food can be both a means to destruct a food culture’s structure (because members of same households can easily eat on different mealtimes, for instance), as a means to maintain an existing structure (by quickly warming up a ready-to-eat pizza or a takeout meal to have dinner together during busy days). Although convenience food is often believed to be positively related to less structured, more hurried and individualized meal habits, others believe that convenience food can help to organize meals more effectively in today’s ‘‘busy society’’ (Shove, 2003; Warde, 1999). Whereas people, previously, had to eat when the food was ready, people, now, have more choices, more flexibility and are less tied to a ‘‘fixed’’ cuisine. This study adds new knowledge to the existing literature by using an internationally unique database (the 2005 Time- and Household-Budget Survey), which gives insights into both families’ food expenditures, their degree of reliance on convenience in food provisioning, and their day-to-day cooking and eating practices. This makes it possible not only to pay attention to the foods and meals that people consume, but also on how, when and in which contexts they prepare and eat their meals. The data apply to Belgium but the results of this article might reflect changes that may occur in other western countries, keeping in mind that each country has its own food culture cuisines and especially its own food culture structures. Despite the widespread belief in a global consumer cultural and ‘‘worldcuisine’’ (Goody, 1982), studies show considerable cross-national variations in culinary habits and convenience-orientation. While the trend of convenience is most apparent in America, people in Europe spend more time cooking from scratch, although the consumption of time-saving and pre-prepared meal alternatives also turns out be higher in the Northern European countries than in the South of Europe (Askegaard and Madsen, 1998; Warde et al., 2007). It is however not the purpose of this article to get into a detailed examination of Belgium’s food culture, since we belief that the (modern) idea of a ‘‘national food culture’’ is notoriously ambiguous, and that national food habits are hardly ever fully national (Belasco, 2002). The fact that ‘‘national’’ food cultures are commonly associated with stereotyped images (Fernández-Armesto, 2002), stresses the importance of using time-and household-budget data to examine everyday food and cooking practices. Although this paper will mainly address practices rather than discourses, the findings may also reveal many of the sociocultural meanings embedded in the processes of food shopping, preparation and consumption. This study hopes to offer new sociological insights into households’ food consumption, cooking and table habits as a mirror of contemporary culture, society and time-related problems and perceptions of time, using convenience food as central focus in this complex web of relationships. In order to make sure that our main research goal can be clearly investigated, it is necessary to address the different conceptual definitions and meanings of convenience food (Section 2.1). After this conceptualization, we formulate a number of hypotheses regarding the general consumption of convenience food in Belgium’s food culture (Section 2.2), the different lifestyle and household characteristics that may affect consumers’ demand for convenience in food provisioning (Section 2.3), and the meal patterns that tend to go along with higher levels of convenience in food choices (Section 2.4). The third section presents the data and the different analyses that will test these three main hypotheses. We conclude in the final section.
2. Hypotheses 2.1. A conceptualization of convenience food Convenience food is a multidimensional construct, includes a wide variation of commercially processed canned, jarred, deep-frozen, dried or fresh prepackaged products, has many levels of convenience, and is, as such, not easy to conceptualize. Part of the work, the knowledge, the culinary skills and the time needed to prepare food (going from the meal planning to the cleaning up), is transferred from the home-kitchen to the food industry, food manufacturers and distributors (Traub, 1979). Over the last decades, the ‘‘meaning’’ of convenience has changed from comfort and ease of use, to the saving of time and effort (the modern form of convenience), and the shifting and storing of time (the hypermodern form of convenience) (Shove, 2003). However, there still seems to be no consensus among researchers about the most important factor making convenience foods ‘convenient’. Many ambiguities are found in studies dealing with the conceptualization and measurement of convenience food. According to some, convenience foods are mainly convenient because they make meal planning and thinking, and not cooking, less demanding (Laudan, 2013). Others believe they in fact do make cooking less complex and time-intensive (Candel, 2001) or that they chiefly make the timing of meal preparation and eating more flexible and controllable (Warde, 1999). The extent to which convenience food actually saves time and effort is questionable and both subjective and context-dependent.
208
S. Daniels et al. / Social Science Research 51 (2015) 205–218
The general notion of convenience food is rather vague, arbitrary, and open to various interpretations, as Candel (2001) emphasizes. As a result, the notion of convenience is difficult to examine, although it touches on important issues, such as time, space and consumption (Warde et al., 1998). In this paper, we propose a typological analysis method to measure consumers’ food convenience-orientation, by examining the extent to which their monthly food budget is spent on fresh unprocessed or non-convenience foods on the one hand and on pre-prepared convenient food ingredients and ready-to-(h)eat meals on the other, taking into account the broad nature of convenience food and the fact that most individuals today, as Pollan (2013) states, occupy a place somewhere between those bright poles. Existing research often only focusses on the consumption of convenience foods, without examining how convenience food is situated within households’ overall food-spending habits. For this reason we will examine the broader role convenience plays in households’ food purchases. This will be done by constructing a typology of household domestic food consumption, accepting the idea that the degree of reliance on convenience foods varies from household to household. Although it has already been criticized that consumers are difficult to capture into a typology, and that reality is much more complex (Scholliers, 2007), the typological method that we use in this paper is, however, one of the most accurate and feasible research instruments to grasp the complexity of consumption and everyday life and to better understand the complex interplay between consumers, their conditions, food choices and meal habits (Halkier, 2001). 2.2. The consumption of convenience food Findings of the Belgian-Ready Meals Association (the only Belgian association covering and representing the interests of the entire Belgian ready-meals industry) suggest that nearly all Belgian households (91 per cent) consume convenience foods, irrespective of the amount and the types of convenience food products they buy (personal communication). According to FEVIA (the Food Industry Federation) the food industry is the most important industrial sector in Belgium, in which the production of convenience food belongs to the fastest-growing product categories (Verbeek et al., 2003). Although convenience food seems to have become an important part of today’s food and life habits, it is still considered as an ideal to prepare meals from scratch with unprocessed (or minimally processed) and fresh ingredients (Carrigan et al., 2006). This becomes clear from, for instance, the socio-cultural ideal of the home-cooked meal to reproduce and represent social relations, but also from counter-reactions such as the slow and organic food movements that emphasize localness, terroir-foods, food heritages, green lifestyles, and more, and from the many campaigns, pamphlets and booklets spreading messages about the art of ‘real cooking’. Therefore we expect that the majority of Belgian households do in fact consume convenience foods, albeit in combination with non-convenient and natural ingredients, which we expect, to still take the largest share in households’ food budget (Hypothesis 1: convenience food consumption). The next paragraph mentions the lifestyle and social background characteristics that are already shown to be important drivers for consumers’ demand for convenience. 2.3. Lifestyle and household characteristics and the consumption of convenience food The consumption of convenience food is mainly ‘‘disapproved’’ of when being used in cohabiting households with children: family contexts in which the socio-cultural ideal of the home- and fresh-made family meal is still socially valued and a rather common practice (Kemmer et al., 1998; Mestdag, 2007). It is also argued that, irrespective of the household situation, convenience foods would bring more feelings of ambivalence, guilt or dissatisfaction, when women, rather than men choose and consume pre-prepared food (Aarseth and Olsen, 2008; DeVault, 1994; Warde, 1997). Therefore, we expect the consumption of convenience food to be higher among single-person (or less ‘‘traditional’’ household situations) than multi-person households (Harris and Shiptsova, 2007). Additionally, we hypothesize the consumption of convenience food to be more customary among single men than single women. With this hypothesis we test whether the long dominant view of cooking as a woman’s responsibility and sign of her womanliness and motherhood (DeVault, 1994) still holds today (Hypothesis 2: the impact of gendered household traditions). Regarding the impact of having children, there are somewhat more contradictory findings in the literature. Some studies found that having children is negatively correlated with the consumption of convenience foods, mostly because of the social or emotional component of the meal as a symbol of a close family life (Prim, 2007; Verlegh and Candel, 1999) and the priority to feed children ‘‘good’’, fresh and nutritious and high-quality meals (Jabs et al., 2007). Contrariwise, others studies conclude the opposite, because convenience food would often be consumed by parents to have their children have a treat (Carrigan et al., 2006), as an easy meal alternative when they feel time-constrained (Bava et al., 2007), or as a simple solution for their teens who occasionally need to cook for themselves (Harris and Shiptsova, 2007). As a result, we will explore the relative impact that children have in households’ food consumption patterns (Hypothesis 3: parent/child feeding responsibilities). In pursuit of Verlegh and Candel (1999), we presume that single- or two-earner families are more likely to be limited in their time for food preparation, to outsource their cooking activities and to consume convenience foods than unemployed families (Hypothesis 4: paid work responsibilities). However, in the next hypothesis we address the question whether this relation will persist, controlled for to the strong age-sensitivity in the use of convenience foods. We presume that the consumption of convenience food is more common in some age-groups than in others and changes through the life course. Whereas younger age-groups would have a high likelihood of consuming processed and easy-to-prepare convenience foods or fast foods (Marquis, 2005), people in the ‘‘busy’’ age between 24 and 44 would also often try to
S. Daniels et al. / Social Science Research 51 (2015) 205–218
209
find the quickest way and use convenience foods as a strategy to get more control over their time and to plan and organize their meals more effectively (Jabs et al., 2007). Whereas people in this ‘‘busy’’ age-group would often experience an increase in distress and time pressure when trying to balance their multiple roles, work- and family demands (Elchardus, 1996), older age-categories or generations would have more time to cook or would hold more to their ‘‘habitual routine’’ to cook their meals from scratch (which can be understood as a generation effect). When they grew up they were less familiar with pre-prepared convenience food or at least not to the extent that it is today (Brunner et al., 2010). However, from the moment they are no longer able to cook their meals and no longer have someone to cook for or to share their meals with, the whole social meaning of cooking and eating can change and enhance the demand for less time-consuming cooking products and methods (Sidenvall et al., 2000). For care-dependent elderly, convenience food can be a way to prevent having poor dietary patterns and being dependent of the help from family or food preparation services. Despite that we expect a significant correlation between convenience foods and age (Hypothesis 5: age-related responsibilities), we are cautious in interpreting this association. Concerning the impact of individuals’ formal education, we expect that higher educated people consume more convenience food, because they are shown to have more meal scheduling problems and time constraints to cook from scratch compared to lower educated people, although this ‘‘busyness’’ might also be interpreted as a ‘‘social badge of honor’’ (Gershuny, 2000, 2005). On the other hand, higher educated individuals are also shown to be more consciousness on holding both a sustainable and healthy lifestyle (Mirowsky and Ross, 1998), and a rather aesthetic than a functional approach to food (Vandebroeck, 2013) as a marker of self-stylization (Bugge and Almås, 2006; Daniels et al., 2012; Warde, 1997, p. 41), which might suggest that lower educated individuals are more convenience-oriented than higher-educated individuals. In this study we will explore the credibility of both hypotheses (Hypothesis 6: educational background). When studying food consumption, it is not only important to focus on the relationship between consumers and their food choices, but also to focus on how and in which different ways they prepare and consume their meals, especially because food consumption is often taken out of context in studies. 2.4. Convenience food and meal patterns Many of the societal changes and trends that have occurred in the last decades, are likely to have had a significant impact on people’s food choices and meal patterns. This has raised questions about the extent to which food habits have changed. Many scientists debate whether today’s meal habits have individualized and can be characterized by a total absence or a disruption of the ‘‘conventional’’ structure in the daily organization and timing of meals. Convenience food is regarded to be one of the most significant food trends that would have contributed to this disruption and would be an important symbol of this changed relationship between people, food and time. By testing the role of convenience foods and its part in contemporary meal structures, we can examine the extent to which convenience food can be regarded as ‘‘a cause’’ of individualized and de-routinized meal patterns and/or as ‘‘a strategy’’ to maintain conventional meal patterns (Warde, 1999). We shall investigate the extent to which we find evidence for the grazing-thesis and the strong belief that a high demand for convenience food items plays a major role in grazing behavior. We will do this by testing whether convenience food is significantly associated with a higher likelihood to eat alone or in social isolation, to eat in front of TV rather than around the dining table, to eat much more quickly and less leisurely, to snack more often in between meals, to eat out of the home more regularly, and to eat at irregular and less conventional hours: indicators that in combination can be interpreted for what is theoretically understood as grazing (Hypothesis 7: the thesis of grazing). If this hypothesis is not supported, and consumers’ demand for convenience has hardly no impact on meal structures, then the assumption must be true that meals are still highly structured events, despite the many changes in households’ cuisine habits. If the results support this hypothesis, it would suggest that convenience food has become part of the food choices and meal behaviors of today’s consumers. If this is the case, we can hypothesize whether households’ need for convenience food items has a larger impact on their kitchen, than their table habits. Hence, we would expect to see that higher rates of convenience food consumption are associated with other aspects of convenience in food provisioning (Candel, 2001), such as the consumption of take-out, home delivery, fast food and restaurant meals (Hypothesis 8: convenience in food provisioning). We will control these hypothetical relations for week versus weekend days, gender, and household background differences. 3. Materials and methods 3.1. Data We use data from the Time- and Household-Budget Surveys, collected by Statistics Belgium in 2005. Both surveys cover time-use and expenditure information for the same households (3306), along with some socio-economic background information. This allows us to link households’ food purchases and expenditures on convenience foods (as one of the many expenditure-categories in their total registered household budgets) to the different ways members of the household spend their time and incorporate their cooking and eating activities into their daily lives (which is described in detail in the diaries for one random weekday and one weekend day). In general, it is assumed that self-completed diary surveys as these, are more
210
S. Daniels et al. / Social Science Research 51 (2015) 205–218
reliable and precise than self-administered questionnaire surveys in the study of daily human behavior (Robinson and Bostrom, 1994, 13) because ‘‘they record the nature and timing of activities as they sequentially occur in daily life.’’ Diary-research is a ‘‘sort of social microscope that allows to examine facets of daily life that are otherwise not [or hard] to examine (Gershuny and Robinson, 2013, 3; Robinson and Godbey, 1997, 5)’’. In order to take into account the intra-household non-response (because of the minimum age restriction of 12 and the voluntariness to participate in the research), we only use the time-use information of the household member that contributes the most to the family income, and, when applicable, of his or her spouse or domestic partner. Student households are excluded from the analyses. All data are weighted in terms of the different weekdays to have all days of the week equally represented, and on individual and households characteristics to be representative for the Belgian population. 3.2. Measurement and analytic strategies In order to examine the degree of convenience in households’ food consumption choices, we take a closer look at their food expenses and measure both the share of their total expenditures on food, and the share of their total at-home-food expenditures on convenience food items. The outcomes of these measurements are expressed in relative mean percentages and operationalized into 8 different food- and convenience-related variables (Table 3) we believe are useful for assessing households’ demand for convenience food and for categorizing people into different food consumption patterns. Although these different variables provide a consistent picture of families’ (convenience) at-home food expenditures as a share of their total (food) budget, they do, however, not give any information about the quality and health characteristics and the purchased quantities of the food products. To investigate the role convenience food plays in Belgium’s food culture cuisine and to conceptualize different categories of convenience-orientation, we construct a food typology by carrying out a Latent Class Cluster Analysis (abbrev. LCA) on these 8 variables. One of the main advantages of LCA over other ‘‘traditional’’ cluster methods, is that the technique can express households’ tendencies to belong to the different identified food categories (expressed as class membership probability scores), besides the modal assignment approach. It is also found that LCA has a lower risk of having biased results than other cluster techniques, because the method does not rely on model assumptions, such as normal distribution, linearity and homogeneity. For more details we refer to the technical report (Magidson and Vermunt, 2004). Because the creation of this typology is not theoretically driven, we select the most parsimonious cluster-model based upon the different model fit indicators (The Akaike Information Criterion, the Bayesian Information Criterion and the Consistent Akaike Information Criterion), by gradually extending the number of clusters and comparing the different values of the Bayesian estimates (Table 2). To test the meaningfulness of the best-fitting model, we perform a multiple classification analysis (abbrev. MCA) in a descriptive manner to validate the typology and to examine to which extent households from different backgrounds belong to the different identified food consumption patterns, using the continuous class membership probabilities as dependent variables. In the final part of this paper, we employ two different types of analyses to estimate the relations between families’ backgrounds, their demand for convenience food (i.e., the modal-assignment of the households to the food consumption category they are most likely to belong to) and their cooking and eating patterns. First, we use multilevel logistic regression analyses to examine how household characteristics and food consumption patterns are differently related to how individuals’ spend their time on cooking and eating (a description of these 7 dummycoded time–space related meal patterns can be found in Table 1). By using visual binning, these variables are transformed into rather more substantively interesting variables. The reason for working with multilevel regression analysis is due to the tree-level survey design of the time-use data, because the different cooking and eating activities (level-1) are nested within individuals (level-2), who are nested within households (level-3). If we would ignore this clustering, we could have a higher risk of having biased results and overestimated statistical powers and signification-levels. We also explore the cross-level interactions between gender and family structure, because at-home cooking is still highly gendered and a mainly female responsibility (especially amongst cohabiting couples) and women’s meal habits might thus differ from those of men. Adding these cross-level interactions changes the interpretations of the main coefficients of gender and family structure: they can no longer be interpreted independently from each other, but need to be interpreted in terms of reference values of the other terms in the interaction. Because of this cross-level interaction we only use information from heterosexual couples to examine to which extent cooking is still a woman’s responsibility in ‘‘traditional’’ families (DeVault, 1994). We believe that the long held cultural associations between gender and cooking have been somewhat diversified and re-interpreted from the long dominant definition of cooking as a an ordinary female chore, but the total sample of same-sex households is too small to examine the relationships between gender and cooking more in detail. We only focus on the fixed parts of the 7 multilevel models (expressed as ORs) and present the goodness-of-fit indicators. Second, we develop two logistic regression models in SPSS 20 to study the impact of the above mentioned household and food consumption variables as possible determinants of the away-from-home consumption of fast food, takeaway and restaurant meals (a description of these dummy-coded dependent variables can be found in Table 1). Because this information is provided on the household level we only need to carry out logistic regression analyses. To measure the extent to which households are outsourcing their meal preparations, we focus on their expenditures on convenience food products, fast food, takeout meals and restaurant visits, without paying attention to food preparation services or the (unpaid) support of family members or others, simply because this information is not provided by our data.
211
S. Daniels et al. / Social Science Research 51 (2015) 205–218 Table 1 Descriptives of the original and new dependent variables on the individual and household level. Original dependent variables Time-related variables (on daily basis and individual level) 1. Time spent cooking 2. Time spent snacking 3. The share of eating alone in the total time spent on eating 4. Time spent eating a meal (
Contents lists available at ScienceDirect
Social Science Research journal homepage: www.elsevier.com/locate/ssresearch
Convenience on the menu? A typological conceptualization of family food expenditures and food-related time patterns Sarah Daniels ⇑, Ignace Glorieux, Joeri Minnen, T.P. van Tienoven, Djiwo Weenas Vrije Universiteit Brussel, Research Group TOR, Sociology Department, 1050 Brussels, Belgium
a r t i c l e
i n f o
Article history: Received 4 July 2013 Revised 23 September 2014 Accepted 29 September 2014 Available online 27 October 2014 Keywords: Household food consumption typology Convenience foods Meal preparation and consumption habits Time- and Household Budget data Latent Class Cluster Analysis
a b s t r a c t One of the most fundamental, but also controversial, food trends of the past years is convenience food. This article investigates the underexplored relationship between the heterogeneity in (convenience) food consumption (a feature of a food culture’s cuisine) and meal patterns (characteristics of a food culture’s structure). This study hopes to illustrate that convenience food can be interpreted both as a means to maintain a food culture’s structure and as a means to overturn it. Latent Class Cluster Analysis is performed using data from the HBS 2005 survey on families’ food expenditures to conceptualize convenience-orientation and to examine the relationships with families’ meal behaviors. Whereas outsourcing cooking is most prevalent among single-person households; two-or more-person households are most likely to buy unprocessed and natural foods and to spend most time cooking and eating in. A higher consumption of convenience food is also more likely to affect individuals’ kitchen than table habits. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction This paper focuses on one of the most important food questions and controversies that mark today’s foodways: to which extent convenience foods have become part of the food consumption patterns of families of different social groups, and to which extent they affect today’s meal preparation and table habits. So far, there is still little quantitative research about both families’ reliance on, or need for convenience, and their day-to-day kitchen and table habits. However, it is precisely the examination of this relationship, that will help to better understand the role of convenience food in a time where the seemingly simple question ‘‘what to eat’’ (Pollan, 2009) has become a fast-growing topic of concern, discussion and insecurity. Historically, food insecurity is not new, but due to many – especially post-war – food trends, innovations, regulatory changes, social and cultural developments and many more, it has become an even more complicated and omnipresent issue in today’s societies. The concerns about food and the ways that people think about food and use it in their daily lives, has never been so strong. One of the most significant changes that has brought this discussion about food to the table, was the way in which convenience food became a defining symbol of the commercialized and industrialized food supply and the ever-changing relationship between people, food and society (Scholliers, 2007, 2014). By using convenience foods, many of the activities, tasks and culinary skills essential to food preparation are taken over by services and industries intended to make cooking less time and energy consuming, both physically and mentally (Darian and Cohen, 1995; Scholderer and Grunert, 2004). Today’s availability of convenience food items goes far beyond the older existing canned products (already available since the early 19th century), varying from canned soups and sauces, cake mixes, ⇑ Corresponding author at: Pleinlaan 2, 1050 Brussels, Belgium. E-mail address: [email protected] (S. Daniels). http://dx.doi.org/10.1016/j.ssresearch.2014.09.010 0049-089X/Ó 2014 Elsevier Inc. All rights reserved.
206
S. Daniels et al. / Social Science Research 51 (2015) 205–218
deep-frozen pizzas, pre-cut vegetables to ready-to-(h)eat meals. Since a few decades, the consumption of convenience food has become more embedded in the way that most people cook, eat and live (Shove, 2003), that it is considered to be one of the largest and most fundamental food innovations in recent years (Brunner et al., 2010; Goody, 2013). The many discoveries, developments and innovations behind convenience food are some of the most crucial foundations on which today’s food industry is built (Capps et al., 1985). Today, convenience food has become such a self-evident part of daily life, that even non-users of convenience food hardly realize how their relationship to food became industrialized and commercialized, going from ground coffee to baked bread, breakfast cereals, margarine, sugar, cold cuts, yogurt, ice cream, and more. However, despite the fact that these commercial food products are generally ready-made (Burnett, 1989), they have become such a taken-for-granted part of individuals’ cooking routines that they are no longer labeled as convenience food, but as natural and ‘‘fresh’’ ingredients (Short, 2003). Despite of the fact that the basic principles of convenience food were already present for centuries (or at least in the supply of canned meat, vegetable and fish products), it was especially from the Second World War on that the food industry started to sell and depict convenience food as an indispensable kitchen ingredient, as a means to taste the comfort of modern life and to save work and time in the kitchen (Shapiro, 2004). The last decades have, however, not only shown many changes in the supply of and demand for convenience food, but also a clear growth in the accessibility and consumption of away-from-home food products as ‘‘the pinnacle’’ of the commercialization of food (e.g. fast food, takeout or restaurant meals) (Scholliers, 2007) and in the design (e.g. kitchens that increasingly became built around appliances instead of the opposite (Van Otterloo, 2000)) and utensils used in the kitchen (e.g. microwave ovens, freezers, steamers, espresso machines, bread making machines, . . .). What these trends seem to illustrate is that cooking not only seems to have become a household task more difficult to organize and schedule (regardless whether one finds it worth the time and effort or not), but also a less necessary task, no longer strictly dependent of the homemaker’s efforts (a task which is still seen as a woman’s responsibility). According to Alan Warde (1999) and Elizabeth Shove (2003), convenience food has increased in importance due to people’s enhanced needs for ease, efficiency, control, and to manage and spend their times (and lives) better. In just a few generations, we seem to have moved from a society in which people still worked for bread and butter, towards ‘a leisure society’ in which consumerism became a ‘way of life’ (Veblen, 1994 [1899]) and more and more people started to feel an ever greater time pressure than ever (Linder, 1970) in which ‘‘life has become a patternless checklist of having been there, done that’’ (Robinson and Godbey, 1997, 44). The many different activities that people nowadays need or want to combine, have become so complex, that people are increasingly looking for ways to save time, including into their home-kitchen. Given this, the increase in time-saving and convenient meal options may seem an obvious and successful evolution, while, in essence, it was, and still is, a somewhat complex, disapproved and contested trend (Algren et al., 2004; Shove, 2000; Warde, 1997). Despite that convenience food is one of the most fundamental food innovations of the last decades, the story of convenience has been far from an instant success and has become subject of plenty of discussion and controversy. Its consumption was, and still is, hampered, by the many social and emotional components attached to food and cooking, as ways of showing care for others, to communicate love, to make others feel welcome, and to symbolize social relationships. According to Fernández-Armesto (2002, p. 250) ‘‘the companionship of the campfire, cooking-pot and common table, which have helped to bond humans in collaborative living, are in risk of being shattered’’ due to the growing trend towards commercially processed food. There are strong claims that food habits would be individualized, fragmented, more irregular and hurried, that meals would be no longer act as social Zeitgebers, or organize and give meaning to time (Zerubavel, 1985), that the ‘‘habitual’’ mealtimes would break down and that food would have evolved to a state of gastro-anomy or normlessness (Fischler, 1979). Home-cooked meals are believed to be increasingly replaced by time-saving meal alternatives and snacks, to enjoy less priority, to be more likely to be eaten outdoors, alone, and on the run (Pollan, 2013). And much, if not all of these problems, are ascribed to the rise of this ‘‘convenience revolution’’ in individuals food spending patterns (Burnett, 1989, 310). Despite of the fact that both the production and consumption of convenience food are criticized for their possible destructive impacts on the environment, public transport and health (Kjaernes et al., 2007), the grazing-hypothesis is the main starting point of this article. This hypothesis assumes that people would increasingly eat in an incidental manner regardless of socio-cultural norms regarding food, time and context (Davis, 1995). Although, it has already been shown that contemporary meal habits (both in Belgium, as in other European countries or in the USA) are still bound to collective and customary rhythms, routines and social norms (Mestdag, 2007; Warde et al., 2007), the studies that have cooked this grazing-thesis, only used information about the ways in which individuals’ and families’ meal habits are organized in time and space, without taking into account any information about their food intakes, purchases and demand for convenience. However, it is only by examining these associations that consumers’ food consumption choices, and their kitchen and table habits can be fully understood (Marshall, 2005). To examine to which extent convenience food influences household’s cooking and eating habits, it is crucial to make the distinction between a food culture’s structure and a food culture’s cuisine. Whereas a cuisine is more likely to refer to the food products, recipes, cooking methods and kitchen appliances that people use to prepare their meals; a food culture’s structure is characterized by the way that people organize their meals into their daily life routines and habits (Rozin, 1982). The main difference between both food culture components, is that a structure is more steadfast and that a cuisine is, thus, more open to, for example, influences from foreign cuisines and more flexible to change towards a food melting pot (Askegaard and Madsen, 1998; Schrover et al., 2005). Perhaps this difference can
S. Daniels et al. / Social Science Research 51 (2015) 205–218
207
be best illustrated by the fact that people have more difficulties to adapt to foreign food culture’s structures (such as other mealtime hours), while at the same time they can look forward to the local cuisine and food specialties. The main goal of this article is to investigate (RQ1) to which extent convenience food has become part of Belgium’s food culture cuisine, (RQ2) to which extent it has become more embedded in the food habits of some groups in society than in others, and (RQ3) to which extent households’ reliance on convenience food in food provisioning is related to how they cook and eat (which can be understood as important indicators of a food culture’s structure). By doing this, we will be able to find out whether the reliance on convenience food has a similar impact on individuals’ kitchen habits, as on their eating habits. By spending attention to consumers’ purchasing behavior, food consumption patterns and home-kitchen and eating habits, and by combining a consumption-oriented, with a situation-oriented approach, we can examine to which extent convenience food can be both a means to destruct a food culture’s structure (because members of same households can easily eat on different mealtimes, for instance), as a means to maintain an existing structure (by quickly warming up a ready-to-eat pizza or a takeout meal to have dinner together during busy days). Although convenience food is often believed to be positively related to less structured, more hurried and individualized meal habits, others believe that convenience food can help to organize meals more effectively in today’s ‘‘busy society’’ (Shove, 2003; Warde, 1999). Whereas people, previously, had to eat when the food was ready, people, now, have more choices, more flexibility and are less tied to a ‘‘fixed’’ cuisine. This study adds new knowledge to the existing literature by using an internationally unique database (the 2005 Time- and Household-Budget Survey), which gives insights into both families’ food expenditures, their degree of reliance on convenience in food provisioning, and their day-to-day cooking and eating practices. This makes it possible not only to pay attention to the foods and meals that people consume, but also on how, when and in which contexts they prepare and eat their meals. The data apply to Belgium but the results of this article might reflect changes that may occur in other western countries, keeping in mind that each country has its own food culture cuisines and especially its own food culture structures. Despite the widespread belief in a global consumer cultural and ‘‘worldcuisine’’ (Goody, 1982), studies show considerable cross-national variations in culinary habits and convenience-orientation. While the trend of convenience is most apparent in America, people in Europe spend more time cooking from scratch, although the consumption of time-saving and pre-prepared meal alternatives also turns out be higher in the Northern European countries than in the South of Europe (Askegaard and Madsen, 1998; Warde et al., 2007). It is however not the purpose of this article to get into a detailed examination of Belgium’s food culture, since we belief that the (modern) idea of a ‘‘national food culture’’ is notoriously ambiguous, and that national food habits are hardly ever fully national (Belasco, 2002). The fact that ‘‘national’’ food cultures are commonly associated with stereotyped images (Fernández-Armesto, 2002), stresses the importance of using time-and household-budget data to examine everyday food and cooking practices. Although this paper will mainly address practices rather than discourses, the findings may also reveal many of the sociocultural meanings embedded in the processes of food shopping, preparation and consumption. This study hopes to offer new sociological insights into households’ food consumption, cooking and table habits as a mirror of contemporary culture, society and time-related problems and perceptions of time, using convenience food as central focus in this complex web of relationships. In order to make sure that our main research goal can be clearly investigated, it is necessary to address the different conceptual definitions and meanings of convenience food (Section 2.1). After this conceptualization, we formulate a number of hypotheses regarding the general consumption of convenience food in Belgium’s food culture (Section 2.2), the different lifestyle and household characteristics that may affect consumers’ demand for convenience in food provisioning (Section 2.3), and the meal patterns that tend to go along with higher levels of convenience in food choices (Section 2.4). The third section presents the data and the different analyses that will test these three main hypotheses. We conclude in the final section.
2. Hypotheses 2.1. A conceptualization of convenience food Convenience food is a multidimensional construct, includes a wide variation of commercially processed canned, jarred, deep-frozen, dried or fresh prepackaged products, has many levels of convenience, and is, as such, not easy to conceptualize. Part of the work, the knowledge, the culinary skills and the time needed to prepare food (going from the meal planning to the cleaning up), is transferred from the home-kitchen to the food industry, food manufacturers and distributors (Traub, 1979). Over the last decades, the ‘‘meaning’’ of convenience has changed from comfort and ease of use, to the saving of time and effort (the modern form of convenience), and the shifting and storing of time (the hypermodern form of convenience) (Shove, 2003). However, there still seems to be no consensus among researchers about the most important factor making convenience foods ‘convenient’. Many ambiguities are found in studies dealing with the conceptualization and measurement of convenience food. According to some, convenience foods are mainly convenient because they make meal planning and thinking, and not cooking, less demanding (Laudan, 2013). Others believe they in fact do make cooking less complex and time-intensive (Candel, 2001) or that they chiefly make the timing of meal preparation and eating more flexible and controllable (Warde, 1999). The extent to which convenience food actually saves time and effort is questionable and both subjective and context-dependent.
208
S. Daniels et al. / Social Science Research 51 (2015) 205–218
The general notion of convenience food is rather vague, arbitrary, and open to various interpretations, as Candel (2001) emphasizes. As a result, the notion of convenience is difficult to examine, although it touches on important issues, such as time, space and consumption (Warde et al., 1998). In this paper, we propose a typological analysis method to measure consumers’ food convenience-orientation, by examining the extent to which their monthly food budget is spent on fresh unprocessed or non-convenience foods on the one hand and on pre-prepared convenient food ingredients and ready-to-(h)eat meals on the other, taking into account the broad nature of convenience food and the fact that most individuals today, as Pollan (2013) states, occupy a place somewhere between those bright poles. Existing research often only focusses on the consumption of convenience foods, without examining how convenience food is situated within households’ overall food-spending habits. For this reason we will examine the broader role convenience plays in households’ food purchases. This will be done by constructing a typology of household domestic food consumption, accepting the idea that the degree of reliance on convenience foods varies from household to household. Although it has already been criticized that consumers are difficult to capture into a typology, and that reality is much more complex (Scholliers, 2007), the typological method that we use in this paper is, however, one of the most accurate and feasible research instruments to grasp the complexity of consumption and everyday life and to better understand the complex interplay between consumers, their conditions, food choices and meal habits (Halkier, 2001). 2.2. The consumption of convenience food Findings of the Belgian-Ready Meals Association (the only Belgian association covering and representing the interests of the entire Belgian ready-meals industry) suggest that nearly all Belgian households (91 per cent) consume convenience foods, irrespective of the amount and the types of convenience food products they buy (personal communication). According to FEVIA (the Food Industry Federation) the food industry is the most important industrial sector in Belgium, in which the production of convenience food belongs to the fastest-growing product categories (Verbeek et al., 2003). Although convenience food seems to have become an important part of today’s food and life habits, it is still considered as an ideal to prepare meals from scratch with unprocessed (or minimally processed) and fresh ingredients (Carrigan et al., 2006). This becomes clear from, for instance, the socio-cultural ideal of the home-cooked meal to reproduce and represent social relations, but also from counter-reactions such as the slow and organic food movements that emphasize localness, terroir-foods, food heritages, green lifestyles, and more, and from the many campaigns, pamphlets and booklets spreading messages about the art of ‘real cooking’. Therefore we expect that the majority of Belgian households do in fact consume convenience foods, albeit in combination with non-convenient and natural ingredients, which we expect, to still take the largest share in households’ food budget (Hypothesis 1: convenience food consumption). The next paragraph mentions the lifestyle and social background characteristics that are already shown to be important drivers for consumers’ demand for convenience. 2.3. Lifestyle and household characteristics and the consumption of convenience food The consumption of convenience food is mainly ‘‘disapproved’’ of when being used in cohabiting households with children: family contexts in which the socio-cultural ideal of the home- and fresh-made family meal is still socially valued and a rather common practice (Kemmer et al., 1998; Mestdag, 2007). It is also argued that, irrespective of the household situation, convenience foods would bring more feelings of ambivalence, guilt or dissatisfaction, when women, rather than men choose and consume pre-prepared food (Aarseth and Olsen, 2008; DeVault, 1994; Warde, 1997). Therefore, we expect the consumption of convenience food to be higher among single-person (or less ‘‘traditional’’ household situations) than multi-person households (Harris and Shiptsova, 2007). Additionally, we hypothesize the consumption of convenience food to be more customary among single men than single women. With this hypothesis we test whether the long dominant view of cooking as a woman’s responsibility and sign of her womanliness and motherhood (DeVault, 1994) still holds today (Hypothesis 2: the impact of gendered household traditions). Regarding the impact of having children, there are somewhat more contradictory findings in the literature. Some studies found that having children is negatively correlated with the consumption of convenience foods, mostly because of the social or emotional component of the meal as a symbol of a close family life (Prim, 2007; Verlegh and Candel, 1999) and the priority to feed children ‘‘good’’, fresh and nutritious and high-quality meals (Jabs et al., 2007). Contrariwise, others studies conclude the opposite, because convenience food would often be consumed by parents to have their children have a treat (Carrigan et al., 2006), as an easy meal alternative when they feel time-constrained (Bava et al., 2007), or as a simple solution for their teens who occasionally need to cook for themselves (Harris and Shiptsova, 2007). As a result, we will explore the relative impact that children have in households’ food consumption patterns (Hypothesis 3: parent/child feeding responsibilities). In pursuit of Verlegh and Candel (1999), we presume that single- or two-earner families are more likely to be limited in their time for food preparation, to outsource their cooking activities and to consume convenience foods than unemployed families (Hypothesis 4: paid work responsibilities). However, in the next hypothesis we address the question whether this relation will persist, controlled for to the strong age-sensitivity in the use of convenience foods. We presume that the consumption of convenience food is more common in some age-groups than in others and changes through the life course. Whereas younger age-groups would have a high likelihood of consuming processed and easy-to-prepare convenience foods or fast foods (Marquis, 2005), people in the ‘‘busy’’ age between 24 and 44 would also often try to
S. Daniels et al. / Social Science Research 51 (2015) 205–218
209
find the quickest way and use convenience foods as a strategy to get more control over their time and to plan and organize their meals more effectively (Jabs et al., 2007). Whereas people in this ‘‘busy’’ age-group would often experience an increase in distress and time pressure when trying to balance their multiple roles, work- and family demands (Elchardus, 1996), older age-categories or generations would have more time to cook or would hold more to their ‘‘habitual routine’’ to cook their meals from scratch (which can be understood as a generation effect). When they grew up they were less familiar with pre-prepared convenience food or at least not to the extent that it is today (Brunner et al., 2010). However, from the moment they are no longer able to cook their meals and no longer have someone to cook for or to share their meals with, the whole social meaning of cooking and eating can change and enhance the demand for less time-consuming cooking products and methods (Sidenvall et al., 2000). For care-dependent elderly, convenience food can be a way to prevent having poor dietary patterns and being dependent of the help from family or food preparation services. Despite that we expect a significant correlation between convenience foods and age (Hypothesis 5: age-related responsibilities), we are cautious in interpreting this association. Concerning the impact of individuals’ formal education, we expect that higher educated people consume more convenience food, because they are shown to have more meal scheduling problems and time constraints to cook from scratch compared to lower educated people, although this ‘‘busyness’’ might also be interpreted as a ‘‘social badge of honor’’ (Gershuny, 2000, 2005). On the other hand, higher educated individuals are also shown to be more consciousness on holding both a sustainable and healthy lifestyle (Mirowsky and Ross, 1998), and a rather aesthetic than a functional approach to food (Vandebroeck, 2013) as a marker of self-stylization (Bugge and Almås, 2006; Daniels et al., 2012; Warde, 1997, p. 41), which might suggest that lower educated individuals are more convenience-oriented than higher-educated individuals. In this study we will explore the credibility of both hypotheses (Hypothesis 6: educational background). When studying food consumption, it is not only important to focus on the relationship between consumers and their food choices, but also to focus on how and in which different ways they prepare and consume their meals, especially because food consumption is often taken out of context in studies. 2.4. Convenience food and meal patterns Many of the societal changes and trends that have occurred in the last decades, are likely to have had a significant impact on people’s food choices and meal patterns. This has raised questions about the extent to which food habits have changed. Many scientists debate whether today’s meal habits have individualized and can be characterized by a total absence or a disruption of the ‘‘conventional’’ structure in the daily organization and timing of meals. Convenience food is regarded to be one of the most significant food trends that would have contributed to this disruption and would be an important symbol of this changed relationship between people, food and time. By testing the role of convenience foods and its part in contemporary meal structures, we can examine the extent to which convenience food can be regarded as ‘‘a cause’’ of individualized and de-routinized meal patterns and/or as ‘‘a strategy’’ to maintain conventional meal patterns (Warde, 1999). We shall investigate the extent to which we find evidence for the grazing-thesis and the strong belief that a high demand for convenience food items plays a major role in grazing behavior. We will do this by testing whether convenience food is significantly associated with a higher likelihood to eat alone or in social isolation, to eat in front of TV rather than around the dining table, to eat much more quickly and less leisurely, to snack more often in between meals, to eat out of the home more regularly, and to eat at irregular and less conventional hours: indicators that in combination can be interpreted for what is theoretically understood as grazing (Hypothesis 7: the thesis of grazing). If this hypothesis is not supported, and consumers’ demand for convenience has hardly no impact on meal structures, then the assumption must be true that meals are still highly structured events, despite the many changes in households’ cuisine habits. If the results support this hypothesis, it would suggest that convenience food has become part of the food choices and meal behaviors of today’s consumers. If this is the case, we can hypothesize whether households’ need for convenience food items has a larger impact on their kitchen, than their table habits. Hence, we would expect to see that higher rates of convenience food consumption are associated with other aspects of convenience in food provisioning (Candel, 2001), such as the consumption of take-out, home delivery, fast food and restaurant meals (Hypothesis 8: convenience in food provisioning). We will control these hypothetical relations for week versus weekend days, gender, and household background differences. 3. Materials and methods 3.1. Data We use data from the Time- and Household-Budget Surveys, collected by Statistics Belgium in 2005. Both surveys cover time-use and expenditure information for the same households (3306), along with some socio-economic background information. This allows us to link households’ food purchases and expenditures on convenience foods (as one of the many expenditure-categories in their total registered household budgets) to the different ways members of the household spend their time and incorporate their cooking and eating activities into their daily lives (which is described in detail in the diaries for one random weekday and one weekend day). In general, it is assumed that self-completed diary surveys as these, are more
210
S. Daniels et al. / Social Science Research 51 (2015) 205–218
reliable and precise than self-administered questionnaire surveys in the study of daily human behavior (Robinson and Bostrom, 1994, 13) because ‘‘they record the nature and timing of activities as they sequentially occur in daily life.’’ Diary-research is a ‘‘sort of social microscope that allows to examine facets of daily life that are otherwise not [or hard] to examine (Gershuny and Robinson, 2013, 3; Robinson and Godbey, 1997, 5)’’. In order to take into account the intra-household non-response (because of the minimum age restriction of 12 and the voluntariness to participate in the research), we only use the time-use information of the household member that contributes the most to the family income, and, when applicable, of his or her spouse or domestic partner. Student households are excluded from the analyses. All data are weighted in terms of the different weekdays to have all days of the week equally represented, and on individual and households characteristics to be representative for the Belgian population. 3.2. Measurement and analytic strategies In order to examine the degree of convenience in households’ food consumption choices, we take a closer look at their food expenses and measure both the share of their total expenditures on food, and the share of their total at-home-food expenditures on convenience food items. The outcomes of these measurements are expressed in relative mean percentages and operationalized into 8 different food- and convenience-related variables (Table 3) we believe are useful for assessing households’ demand for convenience food and for categorizing people into different food consumption patterns. Although these different variables provide a consistent picture of families’ (convenience) at-home food expenditures as a share of their total (food) budget, they do, however, not give any information about the quality and health characteristics and the purchased quantities of the food products. To investigate the role convenience food plays in Belgium’s food culture cuisine and to conceptualize different categories of convenience-orientation, we construct a food typology by carrying out a Latent Class Cluster Analysis (abbrev. LCA) on these 8 variables. One of the main advantages of LCA over other ‘‘traditional’’ cluster methods, is that the technique can express households’ tendencies to belong to the different identified food categories (expressed as class membership probability scores), besides the modal assignment approach. It is also found that LCA has a lower risk of having biased results than other cluster techniques, because the method does not rely on model assumptions, such as normal distribution, linearity and homogeneity. For more details we refer to the technical report (Magidson and Vermunt, 2004). Because the creation of this typology is not theoretically driven, we select the most parsimonious cluster-model based upon the different model fit indicators (The Akaike Information Criterion, the Bayesian Information Criterion and the Consistent Akaike Information Criterion), by gradually extending the number of clusters and comparing the different values of the Bayesian estimates (Table 2). To test the meaningfulness of the best-fitting model, we perform a multiple classification analysis (abbrev. MCA) in a descriptive manner to validate the typology and to examine to which extent households from different backgrounds belong to the different identified food consumption patterns, using the continuous class membership probabilities as dependent variables. In the final part of this paper, we employ two different types of analyses to estimate the relations between families’ backgrounds, their demand for convenience food (i.e., the modal-assignment of the households to the food consumption category they are most likely to belong to) and their cooking and eating patterns. First, we use multilevel logistic regression analyses to examine how household characteristics and food consumption patterns are differently related to how individuals’ spend their time on cooking and eating (a description of these 7 dummycoded time–space related meal patterns can be found in Table 1). By using visual binning, these variables are transformed into rather more substantively interesting variables. The reason for working with multilevel regression analysis is due to the tree-level survey design of the time-use data, because the different cooking and eating activities (level-1) are nested within individuals (level-2), who are nested within households (level-3). If we would ignore this clustering, we could have a higher risk of having biased results and overestimated statistical powers and signification-levels. We also explore the cross-level interactions between gender and family structure, because at-home cooking is still highly gendered and a mainly female responsibility (especially amongst cohabiting couples) and women’s meal habits might thus differ from those of men. Adding these cross-level interactions changes the interpretations of the main coefficients of gender and family structure: they can no longer be interpreted independently from each other, but need to be interpreted in terms of reference values of the other terms in the interaction. Because of this cross-level interaction we only use information from heterosexual couples to examine to which extent cooking is still a woman’s responsibility in ‘‘traditional’’ families (DeVault, 1994). We believe that the long held cultural associations between gender and cooking have been somewhat diversified and re-interpreted from the long dominant definition of cooking as a an ordinary female chore, but the total sample of same-sex households is too small to examine the relationships between gender and cooking more in detail. We only focus on the fixed parts of the 7 multilevel models (expressed as ORs) and present the goodness-of-fit indicators. Second, we develop two logistic regression models in SPSS 20 to study the impact of the above mentioned household and food consumption variables as possible determinants of the away-from-home consumption of fast food, takeaway and restaurant meals (a description of these dummy-coded dependent variables can be found in Table 1). Because this information is provided on the household level we only need to carry out logistic regression analyses. To measure the extent to which households are outsourcing their meal preparations, we focus on their expenditures on convenience food products, fast food, takeout meals and restaurant visits, without paying attention to food preparation services or the (unpaid) support of family members or others, simply because this information is not provided by our data.
211
S. Daniels et al. / Social Science Research 51 (2015) 205–218 Table 1 Descriptives of the original and new dependent variables on the individual and household level. Original dependent variables Time-related variables (on daily basis and individual level) 1. Time spent cooking 2. Time spent snacking 3. The share of eating alone in the total time spent on eating 4. Time spent eating a meal (
