Comparative dynamics of urban food styles

The current project began in February 2013 with a review of the literature in sociology, anthropology, nutritional sciences and psychology on "snacking" food practices, in connection with the theories of modernity in an attempt to characterize urban dietary changes. The first research works aimed to define and contextualize the issue and more specifically the difference between food and "snacking". This work allowed us to define the stance of this research that questions ethnocentrism and takes food practices as manifestations of a particular socio-cultural context. Therefore when approaching the food habits of precarious populations in Jakarta we have to bear in mind only the time to eat or the moment of consumption (often associated with the "snacking") but also the physical environment in which they operate. This work has helped to raise the state of the art on the issue and prepare the initial approaches to the problem.

A first fieldwork immersion took place for two months from October to December 2013. The first month of this trip was spent learning Indonesian language and the level of "Communicative beginner" has been certified by a local school. Immersion in Indonesian society continued with ethnographic work (participant observation and open interviews). Put under the theoretical perspective, these data allowed the development of guidelines for data collection in collaboration with local sociologists in 2014.

The qualitative fieldwork began in February with the construction of three instruments for data collection: non-participant observation grid, in-depth interviews guide and focus group script; aiming the description of practices and representations of food in connection with space uses and perceptions. Qualitative data collection took place from March to May in the city of Jakarta, in a precarious neighborhood that we chose in the first explorations in October-December 2013. The ethnographic observations were carried out for a month in street vendors settlements, in restaurants, in public spaces and in the households. The interviews were conducted during the month of May with 30 adults (men and women) living in the neighborhood, 5 street vendors (mobile and not mobile) and 5 stakeholders (head of district, director of health center, health volunteer, midwives). We conducted 4 focus groups with adolescents, women; men and grandmothers in order to perceive changes in the perception of food.

A first thematic analysis of the data was done on Atlas.ti qualitative software. This work allowed the construction of the questionnaire which revolves around the relationship between food-related functions of the home versus the street; the food model through a detailed 24-hour food recall (articulated around different social contexts of food contacts); and socio-economic variables of the household. Quantitative data collection took place during the month of October with 400 individuals representing the sub-district selected by random sampling. The data was entered on EpiData for quality control and then exported to SPSS. The cleaning and first descriptive statistics were done in December 2014.
2015 was dedicated to the processing and analysis of qualitative and quantitative data collected in Jakarta during 2014. These data was processed separately at first and articulated with the theoretical problem for hypothesis testing about (1) the categorization of food intakes in Indonesia (social dimensions of "real meals " versus "snacks") and their temporal distribution in the food journeys; (2) the function of the living space and food supply forms on eating behaviors and (3) the relationship between the autonomy of eaters and solidarity existing in precarious communities.

Several methods were used to collect qualitative data (in-depth interviews, focus groups and ethnographic observations) and each of them requires a different analytical approach. The software chosen for those analyses (Atlas. Ti) was prepared for analysis through the guidelines given by a first reading of the transcripts of the interviews. Specifically, this preparation requires the creation of a code-tree where the themes that are to be treated are defined and the relationships between them. The transcripts of the interviews and focus groups were analyzed through this reading grid. The study population consists of adults eaters, street vendors and stakeholders, these codes grids were therefore adapted to each population.

Quantitative data (400 questionnaires) was processed using SPSS software. Flat sorting allowed us to characterize the population with economic and social variables and to release a first descriptive overview of eating behavior, including the characterization of different types of food days. Following this work, each part of the questionnaire was worked with diverse statistical tools tailored to the objectives. Six dimensions of the 24h recall of food consumption of the day were analyzed: time, place of consumption, forms of supply, content, social context, and categorization in order to define and characterize each moment of consumption and the categorization process made by the consumers. Thanks to this exercise we were able to extract food journey profiles displayed on 24h timeline, highlighting the fact that in general the poor of Jakarta eat mostly food purchased outside the home, eaten preferably within the home but often in solitude. Questions relating to the representations of domestic space and "out-of-home" space were also treated to characterize not only the standards and practices associated with each space but also the ways in which different behaviors are associated with more “private” or “public” spheres and how they are socially constructed. Data on dietary practices and representations were then crossed with socio-economic indicators to understand the factors explaining such behaviors (ie gender, age, employment status of the head of the family and regularity of income).

The articulation of qualitative and quantitative results highlighted the foundations of food behavior from the standpoint of individual rationalities across the population. The main scientific objective of this research was to put in a socio-cultural perspective the categorizations “meals” versus “snacks” and “home” versus "out-of-home” consumption. We chose the context of the poor of Jakarta, the Indonesian capital, because the dietary pattern is characterized by frequent and lonely eating during the day, very little cooking at home and strong community ties. Thanks to a successful combination of qualitative and quantitative methods we managed to determine the dimensions that make sense for Jakartans eaters.
The study of food habits in Indonesia involves first a cultural approach on the place and function of meals and sharing within the food model, a function that varies from one society to another according to the different socio-cultural codes. Our goal was to free the analysis of western ethnocentrism in order to understand the dimensions at play in the meaning of a "real meal" (makanan) and a "snack" (jajanan). Qualitative data showed that the boundary is determined mainly by the presence of cooked white rice (not mixed) and attitude to eat and the trigger that started the meal: hunger, craving, money, obligation; rather than by the moment of the day and the social context (which are, for instance de the main factors that determine when and how people eat in France). These results were also validated by quantitative analyzes where we managed to create profiles of food days per dimension (alone / together intakes; at home / outside; cooked / bought). Each meal type has a specific set of norms (socially constructed) and therefore their temporal distribution in the food journey is determined by deferent factors. In Jakarta, social context is not the main reason to eat and eaters do not seek the synchronization of meals with other members of their household, they often eat alone (23% eaters did not shared any meal the day before); however, these lonely intakes are not related to snacks consumption (in opposition to what the literature review on the issue showed) on the contrary, it is the makanan plate (the main food) that is most consumed in solitude. So, as a main result of the study, we could prove that "commensality" and "real meal" are not consubstantial for Indonesian eaters and therefore the scope of analysis of the ways meals are distributed through the day (and the factors that organize such distribution) should include other dimensions and especially the socio-cultural categorization process.

We also studied the relationship between the living space and supply forms with dietary behavior. Indeed, the vast majority of food intakes (60%) is purchased and not cooked at home. The sale network of ready-to-eat products is very complex (shops, stalls and many restaurants; more or less mobile and present 24h / 24) and highly diversified (traditional dishes composed of white rice and supplements but also industrialized snacks, fruits, soups, drinks, etc.). Also residences have little space inside, are very close to each other and therefore very narrow streets host domestic practices of many types (kitchens installed outside homes, meals at door thresholds, etc.). Recourse to the street as a provider of baked goods depends on these conditions of proximity to the commercial world and on vicinity ties through economic rationalities: the street products require much less investment in ingredients and time and require less expertise than products made at home. Similarly, links between buyers and sellers (often neighbors or family members) help cope with irregularities in wages and build a community network of reciprocity: caring and sharing for subsistence thought trust chains, loans, gifts and sharing. This geographical proximity linked to the living conditions (poverty, vulnerability and uncertainty) promotes community ties and therefore causes an extension of the food management beyond the household. Thus we have shown that “out-of-home” eating behavior implies in this context that people buy ready-to-eat foods through individual choice processes but such process involves the community. Therefore the boundaries of "home" are dependent on the ties existing and "out-of-home" has very different functions depending on the social context.