J-P.P. has conceptualized the food barometer, adapted it to the Malaysian context and wrote the manuscript. W.S. has contributed to the conceptualization of the food barometer and to the writing of the manuscript. C.L. has contributed to the conceptualization of the food barometer, to its adaptation to the Malaysian context and to the writing of the manuscript. L.T. has contributed to the conceptualization of the food barometer, to its adaptation to the Malaysian context and to the writing of the manuscript. M.I.N. has contributed to the conceptualization of the food barometer and to the writing of the manuscript. E.M. has contributed to the conceptualization of the data processing plan and to the writing of the manuscript. M.A. has contributed to the adaptation of the food barometer to the Malaysian context. N.A.R. has contributed to the adaptation of the food barometer to the Malaysian context and S.A.B. has contributed to the conceptualization of the food barometer and to the writing of the manuscript. W.S. and M.I.N. have helped with the editorial revision of the manuscript, which was greatly appreciated.

The authors would like to acknowledge: from the Institute of Ethnic Studies (KITA- UKM): Prof. Mansor Mohamed Noor and Dr. Anis Yusal Yusoff for their participation to the adaptation to the Malaysian context; from Université de Toulouse: Simon Roser and Manon Collard, Daniela Chiang, Dr. Eric Olmedo, for their participation to the adaptation to the Malaysian context; from Taylor’s Toulouse University Center: Karen Ho, Mun Yee Lai, Saeed Pahlevan Sharif, Yu Wai Man, Wei Ting Ting and Sara Abourich for their participation to the adaptation to the Malaysian context; from Field Geek: Tan Kok Poo and Terry Chan for their participation to the adaptation to the Malaysian context, especially during the training of interviewers; and finally all the anonymous participants of the focus group and in-depth interviews.

2Its operational objectives are to produce useful data for different categories of stakeholders: first, public health actors (from epidemiologists to those engaged in health and nutritional education); second, economic actors (including agrofood chain stakeholders, restaurants and food service industries); and third, academics working in the different disciplines interested in food consumption and food cultures, from anthropology and food sociology to medical science, nutrition and dietetics.

3This article presents the methodological organisation of the MFB, focusing on two challenges dealing with ethnic and social factors interaction and collecting data at the level of practices (what people have done) and the norms and representations on food (what people think they have to do and what they believe they do).

The Malaysian Food Barometer

The research framework

4The aim of the barometer is to study the influence of socio-cultural factors, impacted by modernisation, on social and ethnic positions within Malaysian society; demographic and epidemiological transitions in the Malaysian population, emergence of a middle class and the stylisation of food consumption patterns. The research adopted a new approach in combining social position with ethnicity in the context of the Malaysian society. The study of food habits is based on an established model matching eating practices and social representations on meals (Poulain 2002b). The study describes and explores the “food decisions” of individuals through the contextualisation of their food lifestyles. The research framework assumes that eating decisions are a calculation developed in a space of freedom framed by ethno-cultural determinants and are built into social contexts with different kinds of social interactions. Thus this research differs from the reductionist point of view of the “rational choice theory” which has been a dominant approach so far in the nutritional and consumer sciences.

5Effects of urbanisation, middle class emergence, demographic transition, the associated diminution of the number of household members and the stylisation of food consumption, which are effects of modernization, are matched with ethnic groups to develop an Ethno-Socio Economic Status. Food habits are studied at the level of eating practices and social representations while eating decisions are understood as the expression of food lifestyle in certain contexts.

Respondents, sampling and questionnaire

6The MFB project studies the social, ethnic and cultural diversification of food habits in Malaysia using a macro-data collection approach. It is based on a national representative sample (n = 2000) of respondents aged 18 and above. The methodology for sampling uses a semi-randomized approach, based on the regions within Malaysia and their degree of urbanisation. A quota system based on age and ethnicity is also applied. In this way, the evolution of food consumption both at home and outside may be studied and the consequences in terms of market factors and public health can be identified.

7The questionnaire used for the interviews comprises six main parts: socio demographics and ethnic indicators, food norms, food intake in the last 24 hours, cooking practices, social representations concerning food and perception related to food consumption. The questionnaire comprises 66 items and more than 1400 variables, including Body Mass Index (BMI). 46 closed and multiple-choice standard questions used in sociology to describe the socio-demographics of a population, as well as questions that have been used in other barometer studies like the French Baromètre santé nutrition of INPES (Poulain et al. 2010). They have been adapted to the Malaysian context based on a preliminary qualitative survey, which included focus groups, semi-directed interviews and expert interviews.

Background to food studies in Malaysia

8Malaysian society is experiencing rapid modernization. A sizeable new middle class has emerged (Shamsul 1999, 2001; Embong 2002, 2007) since the New Economic Policy era of export-oriented industrialization (1971-1990). Moreover the traditional ways of life for the different ethnic communities are changing dramatically. “Within a few decades, this country has passed from under-nutrition to over-nutrition” (Shamsul 2012: 4). Significantly, within just a few decades, Malaysia has experienced the transition from mortality rates based on epidemic diseases, whose severity was reinforced by food scarcity, to a higher incidence of mortality through non-transmissible diseases for which obesity is a significant risk factor. “What a contrast between the student demonstration over the plight of peasants in Baling suffering from ‘hunger and poverty’ in 1974, and the establishment twenty years later of the Malaysian Association of the Study of Obesity (MASO) in 1994, concerned by the increase of obesity and other related medical problems” (ibid.: 5).

9For many years, nutritional surveys have been capturing these transformations. Tee’s annotated bibliography for the period 1900 to 1993 is instructive (Tee 1980, 1994). After 1993, research on nutrition in Malaysia intensified (Tee 1999; Ismail 2002), culminating in the Malaysian Adult Nutrition survey of 2003, which was based on a national representative sample (MANS, 2008). Nutritional studies on specific populations continue at a steady pace, including focii on women (Tilakavati et al. 2013); minority communities in Peninsula Malaysia (Khor and Mhod Shariff 2008; Gan et al. 2011) as well as in Sabah and Sarawak (Lee et al. 2012). A new Global Nutrition Survey has been recently launched by the Ministry of Health Malaysia. All these research projects focus mainly on the food intake of individuals, evaluated in terms of nutrient composition (macro and micro nutrients). Some ethnic dimensions are included in these studies, with the status of descriptive variables. Other pioneering nutrition studies include Ismail (2002), Rampal et al. (2007), MANS (2008) and Hafzan et al. (2012).

10In parallel, one can find a body of ethnographic literature on the food habits, beliefs and taboos relating to 1) Malaysian indigenous Orang Asli, originally mountainous jungle dwelling hunter-gatherers, mainly the Senoi and Negrito groups (Bolton 1972; Lim and Chee 1998), 2) specific times in the life-cycle such as pregnancy and confinement (Manderson 1981) or specific community of Malaysian society (Devasahayam 2003). In the general anthropological and historical literature on Malaysia we can also find some elements of discussion on Malay food culture both in monographs like those of Peter Wilson (1967) or in a synthesis as described by Anthony Milner (2011). In addition there are some cookbooks in existence, which discuss the historical and cultural aspects of Malaysian food (Brissenden 1996).

11To date a comprehensive survey focusing on the socio-cultural determinants of food habits at the national level has not been reported. This lack of information is particularly important in the context of the rapid social change, especially the trends towards modernization and urbanization of Malaysian society. Moreover nutritional studies are dominated by individualistic paradigms (including “rational choice theory”) that failed to provide the much-needed information on the cultural framework and the context of social interactions within which food decisions are made. The purpose of the Malaysian Food Barometer (MFB) is to fill this gap.

Modernization in a multicultural context

12In this article the term "modernization" is used in a purely descriptive way. It takes into account a series of socio-demographic factors such as urbanization or more precisely metropolization, increased purchasing power, level of education, number of individuals per household and stylization of consumption (consumption used to exhibit a social position in a logic of distinction). This is without prejudging the consequences of this modernization. The authors of the article are also interested in the factors of a possible convergence of social and cultural groups to a middle-class culture, to those that slow down convergence and support processes of differentiation between groups. In doing so, they are different from the classical theories of converging modernity (Mahbubani 2013).

13Several concurrent phenomena are involved in the process of modernization in Malaysia. Through a process of rapid urbanization and rural exodus, which accompanied the industrialization of the New Economic Policy Era, 1971-1990, (Aziz 2012) and the development of the services economy from 1970 (Hutton 2003) the urban population rose from 11% in 1951 to 51% in 1991, 62 % in 2000 and 72.7 % in 2012 (Jaafar 2004; BMCE Trade, 2013). Now in phase 3 of the demographic transition, the structure of the society is changing dramatically. The fertility rate has dropped from 3.29 children per woman in 2000 to 2.64 in 2012 (Leete 1996; Index Mundi 2013) and the size of the average household has reduced from 5.2 persons in 1980 to 4.3 in 2010 (Hirschman and Guest 1990; Shreeniwas 1997; Masayu et al. 2012; Hirschman 2011). The increase in purchasing power of newly waged and salaried employees is an additional factor that has combined with the reduction in family size to herald the emergence of a new middle class (Shamsul 1999; Embong 2007) and also a new working class (Smith 1999) with greater ability to participate in the economy of consumption.

14The related epidemiological transition shows a change in the causes of mortality from communicable diseases, epidemics, to non-communicable diseases such as diabetes, cardio-vascular diseases, cancers, degenerative diseases, and one of the main risk factors, obesity. Despite the success of modern medicine to combat the latter diseases, it is becoming increasingly clear that lifestyle changes, including food habits and physical activity, are vital components both for the prevention and management of these diseases. Malaysian life expectancy has increased by 3.2 years within the 12 years between 2000 and 2012 (Ismail 2002; Index Mundi 2013). Therefore, the current obesity epidemic in Malaysia and worldwide raises concerns for the potential negative effects of such transformations and has stimulated us to focus research on identifying those factors in food cultures and lifestyles which have led to the ever escalating prevalence of non-communicable diseases in Malaysia.

15All the aforementioned macro structural developments have profoundly affected the lifestyles and food habits of the various ethnic groups making up the Malaysian population. The stage of food modernity in Malaysia can be regarded as a consequence of the modernisation of Malaysian society and characterised by the transformation of food supply and of food consumption patterns as well as by new consumer expectations and aspirations in relation to food.

16In addition to these characteristics, which can be found in most developed countries, the Malaysian food consumption context has two major distinctive characteristics. Firstly, it is linked with Malaysia’s multi-ethnicity officially comprising three main ethnic groups (Malay, Chinese and Indian, plus some minority groups). Each main ethnic group has its own food culture with its typical dishes and ingredients, dietary taboos and restrictions, dining rituals, form and structure of meals, symbolic dimensions of food. This could appear as a simple question but the social organisation is much more complex. However, these “ethnic” categories are not totally homogenous in Malaysia. The “Indians” may belong to different religions; for instance, they may be Hindus, Muslims, Sikhs, Buddhists, Christians or members of New Religious Movements. Besides Bahasa Malaysia, the national language, and possibly English, they may speak different mother tongues, for instance, Tamil, Hindi, Urdu, Malayalam; they identify more or less strongly with a caste, come from different regions of India, or from countries neighbouring India, such as Pakistan or Sri Lanka; their families may have been in Malaysia for several generations or just arrived. The “Chinese” may be Buddhist, Taoist, Christian, Muslim converts; they may speak Hakka, Hokkien, Cantonese, Teochew, Mandarin. They may be Min people, Hakka, Cantonese, Wu. Furthermore, there are Malaysians in the official “Others” categories such as the non-Malay Bumiputra, the Dusun, Iban, Kadazan. Moreover, boundaries between these three main ethnic groups are not totally hermetic; they have a certain “porosity” resulting from interpersonal relationships across ethnic boundaries, through friendship, mixed marriage, overlapping religious affiliation and language competence and usage behind the primary “race” identity, religious conversion, “metissage” from historical institutionalized mixed marriages (for example, in the Baba-Nyonya community in historic times) or actual inter-ethnic breeding with or without conversion (Hirschman 1975, 1987; Clammer 1980; Tan 1982); not to mention the rise of individualism within which individual Malaysians develop personal preferences in choosing from a wide variety of dietary alternatives. In addition, there is also some “metissage” between the different food cultures. For example, the Nyonya cuisine from the Malacca region is a combination of Chinese and Malay food cultures, with some influence from the Portuguese. Some restaurants, such as those labelled Mamak, which were originally for Tamil Muslims, are now frequented by consumers of all ethnic groups and thus make a solid contribution to the development of a “Malaysian mixed” food culture. This means that some dishes and some food practices are commonly shared by or are compatible with more than one “ethnic” group.

Conceptualizing social position

17This poses some questions. How to study the transformations in the social structure in a multi-ethnic society such as Malaysia under the process of socio-economic modernisation? How does the Malaysian population reorganise during modernization in terms of social stratification and how do these transformations influence their traditional food “consumption cultures”? In the social sciences, two concurrent paradigms have been developed to describe the position of an individual within the social hierarchy. The Marxist tradition uses the concept of “social class”, assuming that society is divided into categories whose membership is defined in relation to possession of the means of production: those who own it and those who own nothing and have only their labour power to sell (Marx and Engel 1848). This fact of ownership or non-ownership groups of social actors form two basic social classes which have common interests, namely the capitalist class and the working class. The boundaries of each group are quite hermetic. Social classes in the Marxian sense stand in a relationship of conflict due to the differential of power and this is the key dynamic of the society. This concept is inseparable from the notion of class struggle.

18Alternative perspectives to the Marxist paradigm, rather than emphasizing a conflictual relationship, tend to emphasize the perspectives of “social position” and “social category”. But in doing so, they also use the term “social class”, because the term “class” belongs to the generic vocabulary of social sciences. The German sociologist Max Weber (1921) used the term “social class” also but in its larger meaning, similar to “category”. This latter sense of “class” views society as a continuum. Categories group individuals, but their borders are porous so that people can move from one category to another during the course of their life trajectories. Relations between categories are more competitive than confrontational. While the latter perspective is associated with Marxist theory, the former is in line with the class theories of Max Weber and Alexis de Tocqueville (1849). It is based on the idea of a system of social stratification but in a continuum with no boundaries between categories. Under the Weberian class theory, a social class occupies a position in the socio-economic system, while individuals attain social status due to the social characteristics they possess such as educational qualifications, beliefs and social skills.

19Both Marxist and Weberian theoretical perspectives support the idea that social classes or social categories can be described and studied objectively, from an outside point of view. The Marxist view takes into account the relationship to the means of production. The other conception, associated with Weber, considers the position in the social hierarchy as largely defined by economic and cultural resources. But both also include the possibility to study the subjective feeling of belonging of the social actors.

20The critique of the Marxist “social class” perspective was developed within the North American intellectual context by Robert Nisbet (1959). It argues that the relevance of social class has declined as a consequence of the modernization of society. The main arguments in this critique are 1) the de-industrialization and the concomitant rise of services, 2) economic development, 3) the rise of a consumption culture, and 4) the de-structuration of political life. Beside these major reasons, Nisbet suggests that increases in levels of education, female employment, the development of individual property holding and social diversity based on differences in religion, gender, culture, ethnicity or sexual orientation contribute also to the decline of the relevance of social class as a theory for understanding the nature of modern society. In addition, the combination of these phenomena do not only transform the objective living conditions of members of the society, but contribute to the deconstruction of their “class consciousness” (Nisbet 1959). The emergence of the concept of the middle class, negating the binary Marxist view of society, as being made up of owners and non-owners of production means, is part of the shift to view society as a Weberian social hierarchy of classes layered in a continuum.

21So this paradigm based on social position focuses on the transformation of value systems under the effect of modernization and post-modernization. Thorstein Veblen initially formulated this focus, switching the emphasis on consumption activities from “use value” to “symbolic value”, what he termed “conspicuous consumption” (Veblen 1912). After him, and in different ways, Robert Merton (1949), Jean Baudrillard (1969), Pierre Bourdieu (1984) to mention only a few, have reworked this fundamental question in social sciences. Supporting the idea that there is no “genuine” need, (that is challenging the notion of “use value”), Baudrillard considers that consumption, rather than production, is the driver of capitalist society. He distinguishes between four dimensions of value: functional, exchange, symbolic and sign. Moreover, with his distinction between economic, social, cultural and symbolic capital, Pierre Bourdieu has transformed the concept of social class. But it is important to note that, even in Bourdieu’s formulation, finally cultural, symbolic and social capital are considered as “translatable into the economic” (Bourdieu 1987).

22Ronald Inglehart and his research team analysed, with very large, recurrent quantitative surveys, the cultural, economic, and political changes in 43 societies (Inglehart 1997). His findings also support the idea that in the cycle of modernization, the consumption culture moves from a situation where economic determinants dominate cultural determinants to one where cultural determinants take the prevailing role. Re-using a concept of Simmel (1997) he called this stage the “stylization” of consumption.

23Other researchers, following Erik Olin Wright (2006), have proposed to update the Marxist social class perspective. Thus, the Marxist “social class” and the Weberian “socio-economic status” paradigms are still in competition and the controversy has re-emerged in recent years in Europe, especially in France. The first debate focuses on the “supposed decline” versus the “return” of social classes while the second one explores the “downgrading” of the middle class (Chauvel 2006). At the same time other socio-economists and demographers focusing on social mobility reject the vision of a reduction in the importance of the concept of middle class and support the idea of the emergence of a “new middle class” (Maurin and Goux 2012). Beyond these controversies, the question that interests us in this research is the cohabitation or the integration of social class determinants and cultural and ethnic determinants.

24Because the framing of food habits by cultural and ethnic factors is so strong, food is a very fruitful and dramatic empirical field to revisit the question of how social class and culture interact (Fischler 1990, 2011; Mennell et al. 1992; Warde 1997; Gronow 1997; Poulain 2002a; Tibère 2009).

Social class, socio-economic status and ethnicity

26It has been difficult to establish a dialogue about the connection between ethnicity and social class between North American sociologists, who more or less adopt a liberal frame of reference based on the individual, and European sociologists, mainly in the French and German traditions, who were more influenced by the Marxist theory (Bertheleu 2007). On the one hand, Marxists have been blinded by their desire to achieve a proletarian internationalism, and on the other, the Liberals are handicapped by "the belief in the inevitability of the melting pot" (Martiniello 1995: 22). In fact in trying to understand the opposition between a ‘classist’ theory, which views the economic aspect as dominant and the basis of social stratification, and an ‘ethnicist’ theory, in which this economic dimension is totally absent while social hierarchy is based on cultural determinants, we must ask ourselves what is the objectivity of these differences? In fact, it has been argued by Wieviorka (1993, 1998) that this opposition must be made more complex by taking into account the different conceptions of ethnicity and social positions in the class structure.

27The challenge of how to articulate the concepts of social position, or social class, and ethnicity has engaged many sociologists, anthropologists and political scientists. Although from the North American sociological tradition, Milton Gordon has developed the concept of “Ethclass” as an articulation of the Marxist concept of class with that of ethnicity. With this formulation, he tries to demonstrate the fact that ethnicity is more or less connected with social class (Gordon 1978). The paradigm of “Ethclass” assumes that social and ethnic categories are quite impervious. Adopting the same perspective, but working in the empirical field of post-colonial societies, the historian Pierre Briant (1998) has proposed the “Ethnoclass” concept in an attempt to describe the social re-partition of ethnic groups into certain social or economic functions as a result of the colonial experience. The concept of “Ethnoclass” has also been adopted by anthropologists and sociologists working in post-colonial plural or creole societies which still manifest compartmentalization, and studying the interactions between the different groups (Bernabé et al. 1989; Benoist 1998; Tibère 2006, 2009, 2013).

28The social stratification perspective has also developed a dialogue with scholars focusing on ethnicity, gender and social hierarchy. For instance, at the present time, Floya Anthias (2001) represents the tradition of dialogue between social stratification and ethnicity.

29In addition to the neo-Marxist and neo-Weberian perspectives, social movement theory has proposed new types of political allegiances emerging from a range of social forces, focusing around the environment and other specific campaigns often organised in terms of local concerns and identities (Touraine 1981).

30Malaysian society, which during the colonial era experienced ethnic partition according to economic functions (Indians in rubber plantations, Chinese in tin mines and the urban business sector, Malay peasantry in subsistence agriculture), is more or less a good example of the “Ethclass” concept. But the modernization of Malaysian economy, with the decline of importance of the agricultural sector in the economy and the development of service industries in the labour market could change the situation. The de-ethnicisation of occupation, a key element in the New Economic Policy of 1971-1990, achieved only limited success (Olmedo and Noor 2012). The very complex interlinking of cultural, linguistic, religious and political affiliations, which could be the key to understand Malaysian social cohesion (Shamsul 2010) is a relevant empirical field for studying this connection between social position and ethnicity.

31With regard to Malaysia we need to revisit the seminal work of Charles Hirschman (1975). At the beginning of his career this American sociologist studied the way to interrelate ethnicity and social stratification. In his first work Hirschman generated six models, matching ethnicity with social determinants such as “level of education”, “father’s occupation”, “urbanization versus rurality”, etc. His aim was to try and identify the impact of ethnicity and the other sociological variables over a range of dependent variables such as “occupation” (Hirschman, 1975). Twenty years later, he completed this work, using a very comprehensive analysis of Malaysian’s census classifications, from 1871 to 1980 (Hirschman 1987). This body of knowledge on ethnicity and social position plays a foundation role in this study.

33A second index based on ethnic and religion beliefs is being built. It includes a more subtle and complex ethnic dimension, not just the official national administrative designation of the three main racial categories in Malaysia, “Malay”: “Chinese” and “Indian”. It takes into account the fact that this variable of “ethnicity” in not monolithic but could be fragmented by “métissage”, mixed marriage, religious conversion, and the intensity of religious practice. The ethnic identity of individuals participating in the survey will be studied through different variables such as personal ethnic assignation, ethnic assignation of the partner for those living in couples, or of parents and grandparents, the religion practised and the intensity of the individual’s religious life (Poulain and Tibère, 2000). These indexes will be crossed to get an Eth-Socio-Economical position (EthSEP).

34In this context, the MFB is a tool to identify and study in depth the socio-cultural determinants of Malaysian food habits. It describes the “food social space” of Malaysian population. It also focuses simultaneously on the practices and on the representations of food cultures. The aim is to understand food lifestyles and the different food contexts of the various Malaysian ethnic groups and the “middle class” in order to elucidate processes in making food decisions. Additionally, being a follow-up survey, it makes possible the study of longitudinal transformations of food habits in Malaysia. Eventually these different sets of data and their analysis could be used to uncover the social infrastructure of eating decisions (patterns, scenarios and contexts) of Malaysians.

Data collection stakes and methodology

35In the early 1980s, sociologists and anthropologists developed an intense research focus on food consumption surveys (Douglas 1971, 1984; Fischler 1979, 1980, 1990; Garine 1979, 1994; Goody, 1982; USDA, 1985; Mennell, 1985; Poulain, 1985; Lambert, 1987; Mennell, Murcott and Van Otterloo 1992). The results of these works generated theoretical debates, which gave rise to methodological advances that are benefitting current investigations. These include awareness of the need to consider the relative status of variables and data collection techniques. Indeed, how can we be sure that the data collected corresponds to the actual practices of individuals? Quantitative research on eating practices faces an obstacle related to the use of declarative methods to uncover, or at least get data as close as possible to the actual behaviour of individuals. For example, when we ask individuals to describe the meals they ate the day before, and if they have not eaten "as usual" or if they ate differently from the normal pattern, that is what they think they should have done according to prevailing social norms, then they feel uncomfortable and are placed in a dilemma. What should they answer? What they did actually eat yesterday, or what do they usually eat or believe that they should have eaten?

36The methodological problem is that not all individuals solve this dilemma in the same way. Some of them, respecting the instructions of the interviewer, faithfully describe the food intake of the day before, while others, eager to report their usual way of eating, are tempted to change their statement from the actual to the usual, to reduce the cognitive dissonance they feel. All seek to translate what they think is the reality of their food practices. In the second case, the data collected can be said to be more related to their perception of "social norms" - which are a mix of social and nutritional requirements - than to their actual practices. Thus the data obtained have a fairly weak empirical value because they represent neither a complete picture of the real behaviour of individuals, nor of social representations (norms and values) relating to food in the social group being studied.

Daily food intake: between qualitative and quantitative approaches

37In an attempt to resolve this ambiguity, some studies (Poulain 1996; 2002b) have developed a collection method that facilitates distinction between practices and norms, using a questionnaire administered during a face-to-face interview. This is done by first inviting people to describe what they consider to be a “proper meal”, a “proper breakfast”, a “proper lunch”, etc. This is presented to them as taking place in an ideal setting, whereby nothing has disturbed the material organization of the preparation and consumption of these meals. This method is an extension of the work of Mary Douglas on “deciphering a meal” (1971). Through this process, the social norms are collected for the meals under consideration. In the second step, when the interviewee is “liberated” from the normative pressures by his or her statement, another series of questions is asked in order to help the individual to rebuild his or her actual food day. The interviewer begins by specifying that what now interests the research team is what really happened, what really has been eaten. She/he explains that, working at population level, it is not a problem if the meals eaten by the interviewee differ from what has been said in the first part of the questionnaire, when the respondent told what she/he thinks should be done, or what she/he usually does.

38The first type of data corresponds to social norms, that is to say, provides an aggregate of guidelines for food consumption that are rooted in cultural, social and family traditions. They result from the specific socialization of an individual. But these norms are also impacted by the prevailing discourse of public health, or by the pressure of prevailing models of desirable body shape. The second type of data always retains the status of declarative data, but is much closer to the actual practices of individuals. With such a method, the data collected increases in precision, and it becomes possible to distinguish norms and practices and their relationships with each other, particularly for the exploration of various forms of change.
The improvement in data collection methods is an important issue in this research. The ability to distinguish between norms and practices allows a deeper understanding of the transformation of eating habits. In France, four national surveys using the same methodology showed convergent results. The French informants displayed an ideational commitment to “normal” meals as a 3-course structure (first course, main course and dessert) which are rather traditional, but in fact, they simplify their structure in practice. It is understandable that until recently some studies which did not take the precautions described above gave results which appear to demonstrate a higher incidence of structural stability in the traditional French meal (starter, main dish, cheese, dessert) and in contrast with more simplified structures of meals consumed in a catering context (Poulain 1996, 2001). The distinction, in our survey, between norms and practices is a solution to improve the empirical quality of the data; but this has a cost because it greatly increases the length of the questionnaire and almost always requires face-to-face data collection.

39To rebuild the picture of food intake with more accuracy, the data collection method for the MFB project will adopt a middle position between a qualitative and a quantitative approach. The questionnaire is actually a guide to help the interviewee to remember the composition of his/her food intake, the structure of meals, the timing of food consumption, the conditions of food acquisition and the socio-technical contexts of its consumption, including the ethnicity of the persons with whom the food was shared. The form and sequence of the questions have been developed to reflect the Malaysian context, including the high frequency of meals and snacks purchased out of the home and the variety of places of consumption.

40Every food intake, either main meals or other kinds of intake, was described using the same questioning process (one sheet per food intake). The interviewer assisted the interviewees to identify the food intakes of the previous day, following the time flow of the day, from wake-up time until sleeping time. Moreover, the interviewee named him/herself the food intakes and describes it in his or her own words.

Conclusion

41The MFB research tool and the data it generated enriched many disciplinary contexts and theoretical perspectives in the Malaysian academic landscape: the Sociology and Anthropology of Food; Socio-Economy of Consumption, Medicine, especially Diabetes and Obesity Studies; Ethnic Studies; Modernization Theory; Class Theory. The MFB surveys enable us to study comprehensively the social and cultural determinants of human food consumption. Data collected, focusing on food intake, contexts of consumption and social representations related to food, adds to the body of knowledge already available and obtained from nutritional surveys. Through the analysis of the influence of modernization on social hierarchies and ethnic cultures, and ultimately on food eating patterns and food styles, the MFB could invite to revisit theories of convergence trying to show the influence and inertia of ethnic food cultures in societies such as Malaysia, undergoing rapid industrialization and social change in one generation, along with the rural-urban migration which this entails. The MFB is a new data source that could greatly facilitate the study of the evolution of food cultures and eating habits in Malaysia. It will also contribute to set up a wider dialogue, based on empirical data, between the social sciences and nutrition, public health, economics and political sciences. Finally it will facilitate the establishment of comparative analyses with various European countries in which parallel studies are being conducted using similar methodology.