- 1 « Social sciences applied to food ».
1This study took the form of a qualitative survey and was based on research conducted in Hanoi from May 2005 to September 2005 within the Markets and Agriculture Linkages for Cities in Asia (Malica) research consortium. This consortium includes the CIRAD (Centre of International Cooperation in Agronomic Research for the Development, France) and several Vietnamese organizations such as the Institute of Sociology (IOS) of the Vietnamese Academy for Social Sciences (VASS). The study was conducted within the framework of a master’s degree titled “Sciences Sociales Appliquées à l’Alimentation”1, directed by Jean-Pierre Poulain (University of Toulouse-Le Mirail, France). Funding was provided by the CIRAD and work was performed in Hanoi under the Malica research consortium. The results of the present study have been presented at the Asia-Euro Tourism, Culture and Gastronomy Conference that took place in November 2006 at Taylor’s College, Kuala-Lumpur (Malaysia). The author would like to thank Muriel Figuié (CIRAD) for her help in supervising the research.
- 2 Henceforth Avian Influenza will be noted as AI.
2A new highly pathogenic Avian Influenza (AI2) virus, H5N1, was reported in Vietnam at the end of 2003. This virus is a subtype of the Influenza A virus, with the H5 and N1 subtypes. Influenza viruses are defined by the two types of protein spikes that protrude from their lipid envelopes (Hemagglutinin and Neuraminidase). This virus causes high mortality in animal breeding and threatens human health. The risk that it might recombine with the human influenza virus and trigger a world-wide influenza pandemic has been repeatedly stressed by various UN agencies (the WHO - World Health Organization - and FAO - Food and Agriculture Organization - and OIE - World Organization for Animal Health). At the time of the survey, in 2005, the risk was greatest in Vietnam because of the number of human deaths there (42, according to WHO), although Indonesia has largely overtaken Vietnam since. Vietnam is also a poor country (ranking 117th in terms of GDP/capita) and it is interesting to analyse the food practices of Hanoi inhabitants considering the important socio-economic improvements that have taken place there during the past two decades, especially in urban areas. Food insecurity is becoming rare in Hanoi and consumers’ diets are more balanced and diversified. Meat consumption increased from 18.9 kg/capita/year in 1995 to 28.6 kg/capita/year in 2002. Chicken represents 8% of this amount (2.4 kg/capita/year in 1995 and 5.4 kg/capita/year in 2002, according to FAO, Food Balance Sheets). Food poverty has considerably decreased, from 24.8 % in 1998 to 9.9 % in 2002 (3.9% in urban areas) (GSO, 2004). Most Vietnamese homes keep poultry, especially chicken and ducks, though mainly in rural areas. Since few Hanoi consumers are self-sufficient in poultry and eggs, they would seem to be solely at risk of infection with the H5N1 virus when handling poultry at purchasing and preparation stages. Hanoi consumers are major protagonists in this phenomenon. While the notion of consumers is strictly speaking global and masks differences between people and the influences of social contexts on individuals, we will use it here in order to focus mainly on Hanoi inhabitants who are responsible for household cooking. We could not use the term eater because analysing purchasing practices was also particularly relevant to understand how they coped with AI risk. Several studies have been held on AI but very few involved consumers. Therefore, there are many good reasons to take their risk perceptions and avoidance tactics into consideration.
3First, since consumers are primarily exposed to infection when preparing a poultry-based meal (whether purchasing, slaughtering or preparing the meat), AI is seen as a new food-related risk by consumers in Hanoi, even though no case of transmission by absorption has been validated by UN agencies (WHO and FAO) or OIE. Indeed, consumer food patterns (Beardsworth and Keil, 1997; Poulain, 2001, 2002-1, 2002-2, 2006) must be examined and identified. In fact, Vietnamese consumers prefer to buy live poultry (as a guarantee of freshness) then slaughter it at home. Such practices, of course, could represent a risk factor for the spreading of the virus. We must add that poultry, and more particularly chicken, is not just food; it has a symbolic function in Vietnamese culture, as in China (Simoons, 1991). Women, who tend to be responsible for housekeeping, have to know how to choose a good live chicken at their local open-air market. At festive meals, particularly at the Lunar New Year feast, slaughtering a chicken at home is common practice. Likewise to mark the anniversary of a relative’s death. The entire bird is considered edible, including liver and blood (Fournier, 2005). All of these practices, which experts regard as potential risk factors, make Hanoi consumers major protagonists in the present health crisis. Moreover, not only did they follow some official recommendations but they also adopted their own prevention measures. As regards to the AI risk, the food patterns of Hanoi consumers could have been seen as dangerous. As a reminder, food pattern corresponds to a body of rules that are imprinted on the eater in his primary socialization (Poulain, 2002-2). Nevertheless, a kind of “breathing space” exists in this body of rules: innovations may update the food pattern, turning it into a dynamic social concept. The food patterns in Hanoi, for instance, may have been modified by the prevention measures adopted in response to the AI risk.
4Food patterns constitute an important point to understand how Hanoi consumers coped with AI risk. Their perceptions of this new food-related risk are equally important. Our theoretical framework uses the term risk as defined in a report of The Royal Society (1992). Indeed, that risk is a function both of the probability of a fatal effect on health and of the seriousness of this effect, due to one or several dangers. Moreover, social sciences debate use of the term perception as commonly used in the risk field. Is risk perception the appropriate term when talking about how a risk is taken into consideration? It presumes in fact that there is an objective risk. On the contrary, using the term risk representation is useful in order to take into consideration both “imaginary” risks and “real” ones. This expression also focuses on the cognitive construction of risk (Peretti-Watel, 2000). But, given that the AI risk in Vietnam is an objective one, we will use the term risk perception when analysing the ways in which Hanoi consumers have grasped the AI risk. Social sciences studies on risk perception invite us to consider several actors. First, there is a difference between risk as defined by experts and risk as perceived by laymen (Douglas and Wildavsky, 1984; Slovic, 1987; Fischler, 1998; Marris, 1999; Peretti-Watel, 2000; Poulain, 2002-1; Joly, 2003). Experts tend to take a quantitative view of risk, whereas laymen see risk in qualitative terms (Marris, 1999). The laymen’s definition of risk includes social and psychological factors in addition to the qualitative dimension (individual sensation of control as regards risk, Slovic, 1987; Marris, 1999). Furthermore, the layman’s definition of a risk is influenced by individual values, notably their conception of science or perception of their own body (Peretti-Watel, 2000; Poulain, 2002-1). We will begin by considering Hanoi consumers as laymen so as to distinguish them from the experts who compiled official recommendations on AI. Hanoi consumers do not constitute a homogeneous group; however, their perceptions of risk depend on how well the risk was communicated about in the first place and on social variables such as sex, age, income, education, etc. (Kraus N.N and Slovic P, 1988; Peretti-Watel, 1999).We therefore examine the correlation between perceptions and practices related to AI and these social variables. Few sociological studies consider the individual as a plural actor (Corbeau, 1997; Lahire, 1998) whose perceptions, practices and decision-making processes are likely to be influenced by social contexts. The second part of this study will therefore consider Hanoi consumers as a heterogeneous group so as to analyse the effect of social factors on their coping strategies in the face of a new food-related risk.
5Hanoi consumers, as we said earlier, are major actors in the AI crisis. They have the potential to influence the demand and supply of poultry products and their purchasing and food practices could serve to fuel the epidemic.
6The first part of this study will therefore examine the avoidance tactics used by Hanoi consumers to protect themselves from the AI risk. We will then analyse the impact of this new food-related risk on local food patterns. We will conclude by exploring the extent to which Hanoi consumers have become accustomed to this risk, perhaps even regarding it as an acceptable risk, as a result of these avoidance tactics.
7Data collection: background information on the AI phenomenon in Vietnam was gathered from 11 experts. These consisted of five poultry traders, the head of the board of management of an open-air market, one representative from an animal food company, one FAO AI expert, one VSF-CICDA veterinarian and two representatives of Vietnamese consumer associations (VINASTAS and the Club of Consumers as Women). Qualitative surveys were based on 22 in-depth interviews and 5 focus groups. Individual interviews were held with consumers in charge of cooking and shopping for their households: 22 women aged between 22 and 60, found with the help of students from the Institute of Sociology of Hanoi. Focus groups were then organized so as to analyse how social variables or consumption practices could have influenced risk-coping strategies. Focus groups helped to put respondents at ease and identify hotly felt opinions on burning topics. The 5 focus groups consisted of 6 or 7 people and were organized with the help of the Club of Consumers as Women of Hanoi. The first group was composed of people who no longer consumed avian products; the second of people who never stopped eating poultry; the third of “rich” consumers; the fourth of “poor” consumers; and the fifth of young and old consumers. These focus groups aimed at contrasting different points of view on the following topics: comparison of AI with other meat-related risks; knowledge and trust related to information on AI; and individual risk-coping strategies. Individual interviews and focus groups discussions were recorded, transcribed and translated into French before analysis.
8Before tackling the methodology for the present study, we need to take a brief look at the AI crisis background in Vietnam as reported by the WHO. The outbreak started with reports of several animal and human cases in late 2003. Vietnam officially declared AI to OIE on January 8th 2004 as the number of infected provinces continued to rise. February 2004 then brought a lull that lasted until May 2004 when the second episode began with less intensity. The virus seemed to be under control by autumn 2004 but the following winter brought the third and most important episode in Vietnam (in terms of number of animal and human deaths and infected areas). This lasted from December 2004 to April 2005. A fourth phase of outbreaks occurred from October-November 2005 until January 2006. Our qualitative survey took place in summer 2005, between the third and fourth phases. The episodic nature of the outbreaks is important since consumer points of view could well have varied from one episode to another.
9In this section we will look at the risk-coping strategies used by Hanoi consumers, ranging from the most general avoidance tactics to the most individual ones. This topic covers five areas, all of which reveal differences between consumers and risk inequalities.
10AI was reported at the end of 2003 in Vietnam. On January 8th 2004, Vietnam became the first country to declare officially the epizootic to the OIE. In a survey conducted by Malica/CIRAD in Hanoi in July 2004, “all the respondents claimed that they already knew about AI, mainly from television, and that they had done so at least since February 2004” (Figuié and Fournier, 2008: 445). We have to add that in Hanoi the media are controlled by the government; therefore recommendations delivered by the media were perceived as coming from the government.
11In order to see if there was a gap between official recommendations and the real practices of Hanoi consumers concerning AI, we will take a look at the process of poultry consumption. Poultry consumption can be divided into three phases: purchasing, preparation and consumption. Some of the official recommendations covering these stages have been adopted – such as thoroughly cooking poultry prior to consumption. Other recommendations have not been followed, possibly because they were regarded as unacceptable. On December 19th 2004, the newspaper The Workers (N°50) recommended buying poultry in supermarkets. However only rich consumers can afford to purchase food in supermarkets; in Hanoi, supermarkets are not accessible to everyone because of high prices and distance (Figuié and Nguyen Duc, 2006). Other recommendations were not adopted because consumers regarded them as inefficient, preferring to employ homespun prevention measures based on “common” food practices and “magical thinking” (Rozin, 1976, 1994; Fischler, 1990). “We are what we eat” is a deeply-ingrained concept in our relation with food and impacts significantly our food decisions (Rozin, 1994: 24), forming part of our “rational thinking”. Our qualitative survey showed that laymen’s knowledge was principally influenced by such “magical thinking”. Concerning the preparation stage, on April 11th 2004 The Workers (N°42) recommended wearing gloves and masks when handling poultry, even when already slaughtered. But interviewees felt that these recommendations could not be adopted: “Where is the pleasure in cooking with gloves and mask?” (Interview N° 21, June 30th 2005). The recommendation was ignored and common food practices accompanied by “magical thinking” took over. Consumers continued to wash poultry in salted water in the traditional manner, believing that salt was strong enough to clean the meat and kill any virus. According to experts, this is a totally ineffective way to eradicate the H5N1 virus.
12Some Hanoi consumers, especially the oldest and poorest ones, thought it was possible to recognize H5N1-infected poultry, especially if the bird was still alive. When buying poultry on open-air markets, they would handle live animals, assessing both the quality and safety of the product, feeling its weight, feathers, etc. Holding a live chicken, a woman we met on an open-air market in Hanoi explained: “You see, its comb must be totally red, its cloaca must be dry and its belly must be tender” (Interview N° 14, June 22nd 2005). With AI, this kind of practice was counterproductive: according to experts, such close proximity was a very dangerous practice. Thus, the oldest and poorest consumers lost one of their reference points when buying poultry. Choosing a bird and trusting in their own expertise became a problem.
13In Hanoi, poultry is available from open-air markets, supermarkets and known farm producers (80% of the poultry in Vietnam is produced by smallholder producers - Agrifood Consulting International, 2007). New arrivals in Hanoi regularly return home to the countryside, often coming back to Hanoi with poultry, so maintaining a sort of urban-rural network. “I am from Ha Tay province, and when I go back there to see my family, my parents always give me a chicken to bring to Hanoi.” (Interview N°11, June 19th, 2005). People think that farm chicken (gà ta) is stronger than industrial chicken (gà công nghiêp), which are said to be bred in overcrowded conditions and fed with the remains of infected poultry. Influenced by “magical thinking”, Hanoi consumers do not eat the blood and liver of industrial poultries, even though these parts of the bird have a symbolic value in Vietnamese food. People prefer organs coming from farm chickens, popular symbols of nature, freedom and strength. For Hanoi consumers, purchasing a farm chicken from an acquaintance in the countryside is a guarantee of safety. Likewise, dealing with a known producer represents security. For Hanoi consumers this is a “traditional” purchasing practice as opposed to a new reference. What we observed, however, was that supermarkets had in fact become the real new reference for both quality and safety of poultry products. Our interviewees admitted that, ideally, they would prefer to buy poultry at the supermarket for reasons of safety and quality even though Hanoi supermarkets continued to sell industrially farmed chickens. The following diagram illustrates how ‘ideal’ practices compare with ‘real’ practices.
Figure 1. Real and ideal poultry purchasing practices in Hanoi (based on Poulain, 2001: 68)
14Figure 1 analyzes the perception of the supermarket as it is a relatively new place to purchase foodstuff in Vietnam. We tried to compare supermarkets with places where people traditionally purchased their food. The most common place to purchase poultry in Hanoi being the open-air market, we compared it with the supermarket. Based on this approach, we noticed that even though all our interviewees regarded the supermarket as the ideal place to purchase poultry in Hanoi, a social differentiation appeared when it came to real practices. The richer and/or younger the consumers, the more they actually bought poultry in supermarkets. As there was no real discrepancy between their ideal and real purchasing practices, their behaviour was quite consistent – unlike the poorest and oldest consumers who usually shop on open-air markets, despite their preference for supermarkets in the light of the current health crisis. As a reminder, supermarkets in Hanoi are not accessible to everyone because of high prices and distance (Figuié and Nguyen Duc, 2006).
15As mentioned earlier, Hanoi consumers usually shop every day, mainly on local open-air markets where they have close relationships with retailers who, at the time of the health crisis, were regarded as reliable sources of supply. In practice, the open-air market remained the principal place for poultry purchase in Hanoi even though supermarkets and rural acquaintances were the reference for both quality and safety. Consumers set great store by the relationships they enjoyed with their usual retailers. Since 2005, poultry sold in open-air markets in Hanoi is subjected to much stricter government inspections. Birds approved for sale must be certified as safe by a veterinary stamp (compulsory since the start of the third AI outbreak). However Hanoi consumers did not trust this veterinary stamp; for a majority, where veterinary controls could be fixed, trust established with a retailer could not. “I always go to the same retailer’s because I have good relations with her. I can’t recognize a good poultry and that’s why I let my usual retailer do it because I totally trust her.” (Interview N° 3, June 10th, 2005). It was felt that tradespeople who failed to inform truthfully their customers about the origin of products would lose their customers and their job. Hence, the trust established with them constituted a reassuring practice in order to avoid the AI risk. But individual interviews and focus groups discussions show that there is a social differentiation in purchasing practices in general and in the relations with poultry traders in particular. Indeed, some consumers, mostly the younger and richer ones, totally trust their usual retailer. As they feel unqualified to recognize good poultry in terms of quality and safety, they regard their usual retailers as the safest option. As mentioned earlier, however, the older and/or poorer consumers believe they can tell whether poultry is infected with the H5N1 virus. Thus, their experience seems to be more important than trust established with the retailer.
16Religion and more particularly Buddhism seem to have influenced AI risk perceptions and the risk-avoidance tactics used by consumers. Buddhism is the main religion in Vietnam, accounting for 80% of the Vietnamese population in 1999 (Dovert and Treglodé, 2004). Buddhists tend to regard death as largely unimportant since it does not constitute an end to life as such. They also believe that human destines are pre-decided. Such beliefs, of course, imply another way of life and vision of the future, inevitably altering risk perception. “I was first frightened. But I went on eating poultry because I am a Buddhist and if I have to die of avian flu, then it must have been decided.” (Interview N°6, June 16th, 2005). Other people explained that they did not need to build or adopt reassuring practices concerning purchasing and preparation of poultry. In practice therefore, religious beliefs and more particularly Buddhism seem to represent more of a danger regarding the health crisis in Hanoi than an avoidance practice. Consciously or not, however, believers feel protected from the AI risk and their religion forms part of their risk-coping strategies.
17According to a quantitative survey carried out by Malica/CIRAD in July 2004, only 18% of Hanoi consumers stated that they continued to eat poultry throughout the epizootic (even when poultry sales were forbidden in Hanoi) (Figuié and Fournier, 2008). Other people stopped for different lengths of time; in particular, 74% of consumers stopped eating poultry before the ban on sales in February 2004, 90% out of fear of getting AI (Figuié and Fournier, op. cit.). Given that almost the entire population stopped eating poultry, we wondered what people ate instead. The answer depended on social background (Fournier, 2005): richer people turned mainly to beef and seafood (such as prawns) while poorer folk opted for pork, soya and cheaper seafood such as shrimps or whitebait. Abstinence therefore became another means of avoiding risk.
18Hanoi consumers coped with the AI risk using a variety of avoidance tactics. In exploring how their food patterns evolved, we wanted to see what effect this had on consumer compliance with official recommendations - whether modified food patterns actually contributed to the risk or evolved as part of a set of avoidance tactics.
19The risk of AI filtered into the food sphere of Hanoi consumers, making them perceive it as a food-related risk. For that reason, we needed to analyse the change in food patterns in Hanoi in order to see whether consumers had adapted to the risk and basically accepted it. Food is affected by social variables because people eat in different ways. This requires an identification of food patterns i.e. the consumption practices, including purchasing and preparation, as well as the mechanisms used to add or reduce the perceived value of foodstuffs (Beardsworth and Keil, 1997; Poulain, 2001, 2002-1, 2002-2, 2006). A food pattern is composed of overlapping categories which structure the relation eaters have with their food. Food decision-making and reasoning are built into these categories, which are implicitly used by members of a given society (Poulain, 2006: 248). When a food pattern works, many decisions become implicit: the individual no longer has to think about where to buy or how to prepare certain foods. The question is: what was the impact of the AI risk on such reasoning in terms of influencing consumer food patterns in Hanoi?
20We have seen, on the one hand, that food patterns in Hanoi represented a potential risk factor for the spread of AI. Though many people stopped eating poultry since the beginning of the crisis, many continued to consume poultry for important events. “I haven’t eaten poultry since the beginning of the crisis except for the Têt feast because this is the tradition.” (Interview N°12, June 20th, 2005).
21On the other hand, we also noticed that new food habits grew out of official recommendations and avoidance tactics adopted by consumers. Consumers who previously slaughtered poultry at home now increasingly get their retailer to do it for them. Given that the risk of AI infection is greatest when handling live poultry, this new practice represents a major prevention measure. While recommendations on food safety may impact on family privacy (Delavigne, 2001), so do recent developments in the social and economic fabric of Vietnamese society. For instance, since increasing numbers of women now go out to work, they have less time to cook and more money to buy slaughtered poultry (slightly more expensive than a live poultry). Another new food habit relates to the widely recommended practice of cooking poultry at temperatures known to destroy the virus: 70°C. Poultry meat is now consumed thoroughly cooked.
22To sum up, we can see that food patterns have represented a potential risk factor for the spread of the AI virus while also discouraging consumers from adopting new regulations. The same findings are recorded in a study on compliance with diet in a medical context (Fournier et Poulain, 2008: 100). On the other hand, AI led to new food habits such as buying pre-slaughtered poultry and eating poultry meat thoroughly cooked. Hanoi food patterns have been modified as a result of official recommendations and risk-avoidance strategies adopted by consumers themselves. The fact that consumers differed in their response to the AI risk shows that the change in food patterns was subject to social differentiation.
23Risk-avoidance practices adopted by Hanoi consumers led us to wonder whether they were less frightened at the end of the survey (2005). In order to understand whether their perception of the risk might have changed, we need to consider some of the concepts developed by the sociology of risk. Peretti-Watel (1999) points out that we perceive risks as more dangerous for others than for ourselves. This is known as the “optimism bias”: people tend to consider the risk for themselves or their relatives (personal anxiety) as lower than for people in general (general anxiety). This phenomenon is related to what Bandura (1997) calls “high self efficiency”. In the qualitative survey, we asked respondents to classify AI risk in terms of frequency, seriousness and fear, comparing it to other meat-related risks faced by Hanoi consumers (Figuié, Bricas, Nguyen Than and Nguyen Duc, 2004). Frequency and seriousness were taken to indicate general anxiety while fear reflected personal anxiety. Compared with the other risks in the portfolio, AI was consistently regarded as “not at all frequent” and as “the least-frequent meat-related risk”. On the other hand, it was seen as a “very serious” risk, possibly the most serious of all the risks under consideration. Fischler (1998) suggests that individuals tend to underestimate everyday risks but overestimate less frequent ones. Focus groups participants explained that their fear stemmed from the invisibility and newness of the virus and the sheer scale of media coverage. Here too, there was evidence of social differentiation: the richer and younger consumers seemed the least frightened by AI, suggesting that they did not consider themselves exposed to risk (Peretti-Watel, 2000). The AI risk features certain characteristics listed in the “psychometric model” (Fischhoff et al., 1978; Slovic, 1987) that makes a risk less acceptable: it is new, potentially frightening and partly seen as artificial. The government’s intention had been to use the media, particularly television, to reassure consumers about the high standards of battery farming. As it turned out, the media campaign had the opposite effect: Hanoi consumers became increasingly worried and even frightened by what they saw on TV. People had no idea that the poultry they bought on a regularly basis came from battery farms. As in France at the time of the BSE crisis (Poulain, 2002-1), the Vietnamese media introduced Hanoi consumers to some unpalatable realities about everyday food production. The effect was obvious in our focus groups: suddenly workers in battery farms seemed to be implicated in the spreading of the H5N1 virus. This led to a shift in the perception of the nature of the AI risk, from “natural” at the beginning of the crisis to “artificial” by the time of the last outbreak.
24We have seen, on the other hand, that individual protection, through risk-avoidance practices, is considered possible. For Peretti-Watel (2000), the layman’s perception of risk hinges on the notion of control: no matter how illusory, it still makes the risk more acceptable. It nevertheless raises questions related to the respondents’ potential “illusion of control” (Bandura, 1997). Consumers have been widely known to over-estimate their level of information and ability to control their own food safety, at least as regards home-cooked meals (Redmond and Griffith, 2003; Peretti-Watel, 2000). We have to point out however that Hanoi consumers seemed to have become accustomed to the AI risk as a result of their own avoidance practices (such as buying pre-slaughtered poultries), old quality guarantees (such as buying from regular retailers) and trust in official recommendations (such as cooking poultry at 70°C). As shown by a Malica/CIRAD quantitative survey conducted in July 2004, the overwhelming majority of consumers stopped eating poultry when the crisis reached its peak (January-February 2004) (Figuié and Fournier, 2008). Six months later, only 6% of consumers no longer ate poultry (Figuié and Fournier, op. cit.), and 8% in May 2006 (Figuié, Nguyen Minh and Tran Thi, 2006). These figures tallied with the opinions of consumers and retailers we interviewed in 2005, who reckoned that most Hanoi consumers had gone back to eating poultry. The consumption of poultry products in Hanoi today is roughly 50% down on pre-AI levels (Figuié et al., op. cit.), though most people seem to eat poultry again, albeit less frequently. The fact that consumers appear to have become accustomed to the AI risk may be seen to relate neither to time nor habits but rather to risk-avoidance practices adopted or invented by consumers themselves. It was these practices that made the risk more acceptable.
25This study looks at the sociology of food and sociology of risk, analyzing the strategies adopted by Hanoi consumers to protect themselves from the risk of AI. Because Hanoi consumers regarded AI as a food-related risk, they adopted numerous risk-avoidance practices. Despite their fear of AI, Hanoi consumers became accustomed to the idea of risk through their own avoidance practices, old quality guarantees and confidence in some official recommendations. By the end of the survey, AI seemed to be less frightening, especially for rich and/or young people. Almost the entire population appeared to have gone back to eating poultry, if less frequently. Food patterns in Hanoi were originally perceived as potentially dangerous and likely to contribute to spreading the disease. On the other hand, AI led to new food habits. As a reminder, a food pattern is characterized by a dual movement: it corresponds first to a body of rules that were imprinted upon eaters in their primary socialization (Poulain, 2002-2); some food practices are partly determined by social variables. But a kind of “breathing space” exists within this body of rules: innovations may update the food pattern, turning it into a dynamic concept. Here, social contexts and other experiences can influence food decision-making and the way individuals relate to food in the longer term. Hanoi consumers’ responses to this new food-related risk can be seen to tally with this dual movement of food patterns.
26It is also important to explain the limits of this study. For instance, difficulty to analyze risk perceptions in another culture raised a number of problems related to the methodology, language being the most important one. We chose to conduct our qualitative surveys through interpreters (Vietnamese/French). In order to understand how Hanoi consumers faced the AI risk, we had first to tackle their perception of risk – which of course meant communicating with them in their native language. Explaining their perception of risk is difficult enough, but translation adds another bias. We also need to consider the political context of Vietnam: when discussing government reactions to AI, for example, consumers tended to give answers they believed to be socially acceptable.
27To conclude, the Vietnamese government should consider local food practices and risk perceptions before broadcasting recommendations that may not be suitable or feasible. This study suggests looking at the workings of the Vietnamese government, focusing on the political processes involved in the AI crisis.