گوشت و مردانگی در میان جوانان چینی، بزرگسالان ترکی و هلندی در هلند

Abstract The achievement of sustainability and health objectives in Western countries requires a transition to a less meat-based diet. This article investigates whether the alleged link between meat consumption and particular framings of masculinity, which emphasize that ‘real men’ eat meat, may stand in the way of achieving these objectives. From a theoretical perspective, it was assumed that the meat–masculinity link is not invariant but dependent on the cultural context, including ethnicity. In order to examine the link in different contexts, we analyzed whether meat-related gender differences varied across ethnic groups, using samples of young second generation Chinese Dutch, Turkish Dutch and native Dutch adults (aged 18–35) in the Netherlands. The Turkish group was the most traditional; it showed the largest gender differences and the strongest meat–masculinity link. In contrast, the native group showed the smallest gender differences and the weakest meat–masculinity link. The findings suggest that the combination of traditional framings of masculinity and the Western type of food environment where meat is abundant and cheap is bound to seriously hamper a transition to a less meat-based diet. In contrast, less traditional framings of masculinity seem to contribute to more healthy food preferences with respect to meat. It was concluded that cultural factors related to gender and ethnic diversity can play harmful and beneficial roles for achieving sustainability and health objectives.

Introduction Achieving the objectives of sustainability, food security and public health in Western countries requires a transition to a less meat-based diet, as has been put forward by many scientists (see Aiking, 2014, Friel et al, 2009 and Westhoek et al, 2014). One of the potential barriers to this transition is the alleged link between meat consumption and particular framings of masculinity, which emphasize that ‘real men’ eat meat (e.g., Fiddes, 1991, Meier, Christen, 2012, Roos et al, 2001, Rothgerber, 2013, Rozin et al, 2012, Ruby, Heine, 2011 and Sobal, 2005). This link is indicative of the close association between food consumption and gender frames (i.e. cultural understandings of differences between women and men). As gender frames are salient and relevant in relation to all aspects of food-getting (O'Doherty Jensen & Holm, 1999), they may significantly direct the consumption of gendered foods in a gender-congruent direction (Ridgeway, 2009). Theoretically, therefore, the link between meat and masculinity will not be invariant but dependent on the cultural context. Although there is little research on this context dependency, there are indications of recent changes in framings of masculinity in combination with more healthy food preferences with respect to meat and vegetables (Sellaeg & Chapman, 2008). In Western countries, however, gender frames are also changing due to the increasing inflow of new ethnic groups (Van de Vijver, 2007). Immigrants are a growing part of their populations, in particular in urban centers, and ethnicity is one of the main factors that play a role in food choices (Gilbert & Khokhar, 2008). The various combinations of gender, ethnic background, and types of acculturation raise important new concerns and questions on the role of gender differences in a potential transition to a less meat-based diet. To explore the nature of the meat–masculinity link in these contexts, the present paper examines whether meat-related gender differences vary across ethnic groups, using samples of young second generation Chinese Dutch, Turkish Dutch and native Dutch adults (aged 18–35) in the Netherlands. Up to now, ethnic diversity has been given little attention in discussions regarding the promotion of more sustainable and healthy food consumption, except for the topic of obesity, which is also related to gender differences (De Wilde, Verkerk, & Middelkoop, 2014). Our choice of Chinese Dutch (hereafter called Chinese) and Turkish Dutch (hereafter called Turkish) migrants was based on pragmatic and theoretical considerations. As to the first one, Turkish migrants are currently the largest minority group in the Netherlands and Chinese migrants are expected to become the largest one in the coming decades (Garssen & Van Duin, 2009). The theoretical justification is that the Eastern cultural background of these migrants is significantly different from the majority culture in the Netherlands in ways that may improve our understanding of the context dependency of the meat–masculinity link and its potential implications for sustainability and health objectives. Differences between the Far East, the Middle East and the West involve general value differences (Schwartz, 1999 and Schwartz, 2006) as well as differences in food culture (Nam, Jo, & Lee, 2010). Among the migrants these differences have been affected, to a certain extent, by acculturation experiences as a result of living in the Netherlands (Gilbert & Khokhar, 2008), which may work out differently for the sexes as girls are more likely to be bicultural than boys (Berry, Phinney, Sam, & Vedder, 2006). These theoretical issues will be addressed briefly below, as far as they are relevant for the links between meat and masculinity. The masculinity of meat is thought to be an echo of the time of the hunter-gatherers and the participation of men in hunting large game and subsequent meat-sharing activities, which has gained them a reputation of being tough and daring (Rozin et al., 2012). This scene from the past nicely illustrates that the links between gender and meat can be considered at different levels of society. This lines up with recent insights (Ridgeway, 2009) showing that gender frames are multilevel structures, which involve mutually reinforcing processes at the level of societal institutions (e.g. men's jobs in the meat industry), social interactions (e.g. the man's role as meat carver at the dinner table) and individual identities (the way a man likes his meat). Due to various causes, including the growing share of industrialized meat production in Western countries, however, the status of these practices may have changed (see de Boer, Hoek, & Elzerman, 2006). The actual practice of slaughtering has been hidden more and more behind the scenes of social life and the serving of large parts of the animal to be carved at the table has largely gone out of use. Furthermore, although masculine identities are linked with particular body practices (Connell & Messerschmidt, 2005), these practices are changing too, with a tendency toward masculine performances (e.g. martial arts) that are fast and controlled and not slowed down by fat (Spencer, 2014). As a result, the traditional links between meat and masculinity may have become weaker, except on special occasions, such as ‘outdoor cooking’ events (Dummitt, 1998). One of the drivers of this development may be that people in the West have come to develop values that strongly emphasize human equality (Siedentop, 2014), including gender equality and a more shared commitment to domestic participation (Aarseth, 2009). This process contrasts with the experiences of migrants from Eastern countries, where, in terms of Schwartz's cultural value orientations, hierarchical relationships and conservative values are more important (Schwartz, 1999 and Schwartz, 2006). A recent study among migrants and natives in the Netherlands shows that Turkish men and women hold more traditional gender-role beliefs and report less sharing of household tasks than non-Eastern migrants and natives (Van de Vijver, 2007). A comparable study of Chinese migrants is lacking, but a study of Chinese migrants in the USA suggests a different pattern of cultural adaptation in which Asian masculinity has changed over time to include the view that masculinity can contain elements of both masculinity and femininity (Chua & Fujino, 1999). Although this pattern was found in a highly educated sample, it agrees with other observations that East Asians are able to adapt flexibly to multiple demands as they tend to tolerate contradiction, to accept and anticipate change, and to prefer a ‘middle way’ (e.g. the concept of ‘Zhong Yong’; see Spencer-Rodgers, Williams, & Peng, 2012). Regarding the differences in gender frames, therefore, it may be expected that the Turkish are the most traditional, followed by the Chinese and the native Dutch. These cultural differences interact with the abundance of industrially produced meat, which is a typical Western phenomenon (Grigg, 1999 and Swatland, 2010). In the Far East, meats were traditionally used as flavorings or condiments (Nam et al., 2010). Due to their fast economic growth, however, the level of meat production and consumption in Eastern countries has grown rapidly, which is leading to what has been characterized as an unhealthy Western type of diet, often based on traditional recipes with major additions and changes (Popkin, Du, 2003 and Zhai et al, 2014). After immigration to another country, the majority of ethnic groups appear to modify their eating habits by combining parts of their traditional diet with some of the less healthy elements of the Western diet (Gilbert & Khokhar, 2008). As meat is abundantly available, it is accessible to a broad category of consumers, including ethnic groups who were used to a low-meat diet. In the course of this process, various differences between the sexes may also change, as shown by differences between women and men in the prevalence of obesity. In non-Western countries, the prevalence of obesity is often greater in women than in men (Garawi, Devries, Thorogood, & Uauy, 2014), but this pattern seems to have reversed in recent years, resulting in a higher prevalence in boys and men (De Wilde et al, 2014, Neslisah, Emine, 2011, Song et al, 2013 and Yang, 2007). This reversal cannot be explained in a simple way, but the meat–masculinity link might play a role in this process, as meat consumption or factors directly related to meat consumption are positively associated with weight gain (Gilsing et al, 2012 and Tucker et al, 2014). Although much has been written on the masculinity of meat, Fekete et al. (2012) note that there is surprisingly little literature that looks at gender differences in meat consumption systematically (i.e. instead of using gender as control or moderating variable), taking due account of the importance of age and living situation (e.g. marital status). The general pattern is that men eat more meat than women. According to a nation-wide German survey, in fact, about 50% more meat and meat products (MRI, 2008, p. 44) and, according to a Dutch survey, about 33% more animal protein (van Rossum, Fransen, Verkaik-Kloosterman, Buurma-Rethans, & Ocké, 2011, p. 54). A French study among healthy men and women, aged 20–30 and 65–75 years, found that men consumed more protein, including meat products, than women and that this difference was larger among the younger generation than among the elder one (Rousset, Patureau Mirand, Brandolini, Martin, & Boirie, 2005). A German study with middle-aged and older subjects also found that gender differences were more pronounced in those aged 45–59 compared to those aged 60–75, probably as a result of the influence of women on older men's nutrition (Fekete et al., 2012). As Sobal (2005) notes, spouses tend to correspond in the types of food they consume and this means that the gender difference in meat eating frequency may be small as compared to the gender difference in portion size, with men preferring the larger meat portions (Schösler, de Boer, & Boersema, 2012). These studies confirm that gender differences are not invariant but they do not provide much information on the meat–masculinity link. This is understandable because gender frames are multilevel structures (at the levels of societal institutions, social interactions and individual identities) that cannot be measured directly. For that very reason it is important to compare gender differences in different contexts. The present study makes this comparison by examining statistical interactions between the effects of ethnic group and gender on a number of meat-related variables that have proven their strategic relevance in earlier work on sustainability and health. The variables are preferred meat portion size, number of meat eating days per week, familiarity with, and use of, meat replacers (Schösler et al., 2012), familiarity with the benefits of a meatless day and willingness to adopt it (de Boer, Schösler, & Aiking, 2014), main reasons for frequently eating meat and for not frequently eating meat (Schösler, de Boer, & Boersema, 2014). Although the present study is not focused on acculturation and cultural identity, we will also take into account to which ethnic groups the participants considered themselves to belong.

Results Table 1 provides descriptive information about the three samples, distinguished by gender. The samples differed significantly in age and level of education (p < .001). Compared to the natives, the Turks were slightly older and had a lower level of education; the Chinese were slightly younger and had a higher education level. About 90% of the migrants were born in the Netherlands. The Turkish sample included a very small percentage of Kurds (3%) and the Chinese sample included a small percentage of Hong Kongese (12%). All the Turks (or Kurds) felt to be Turkish (or Kurdish) and about 30% of them also felt Dutch (i.e. bicultural). About 90% of the Chinese sample felt to be Chinese (Hong Kongese) and about 40% of them also felt Dutch. About 40 to 50% of the three samples lived with their parents and about 40 to 60% with a partner. Living situation was closely related with age (correlations not shown), which may explain why the Turks lived more often with a partner than the others (p < .001). None of the samples displayed significant (p > .05) differences between men and women in age, level of education, living situation or the percentage that felt Dutch. The participants' involvement in food shopping and cooking did not differ between the samples (p > .05), but it did differ between men and women. Turkish and Chinese men were less involved in food shopping than Turkish and Chinese women (p < .01). In each of the samples, men were less active in cooking than women (p < .001). Among the men, the Turks were significantly (p < .001) less active in cooking (5%) than the Chinese and the natives (about 25%). Table 1. Characteristics of the men and women in each of the samples. Turks Chinese Natives Men (n = 171) Women (n = 179) Men (n = 164) Women (n = 186) Men (n = 172) Women (n = 185) Age 18–24 year 36% 26% 42% 45% 44% 44% 25–29 year 25% 35% 35% 36% 27% 24% 30–35 year 39% (100%) 39% (100%) 23% (100%) 19% (100%) 29% (100%) 32% (100%) Education Low 22% 22% 12% 12% 5% 6% Middle 55% 58% 50% 47% 64% 62% High 23% (100%) 20% (100%) 38% (100%) 41% (100%) 30% (99%) 32% (100%) Country of birth % The Netherlands 91% 89% 90% 91% 100% 100% Ethnic identification % Feels Turkish/Kurdish 100% 100% 1% % Feels Chinese/Hong Kongese/Asian 92% 85% 1% 1% % Feels Dutch 29% 32% 41% 44% 99% 98% Living situation % Lives with parents (or parents in law) 46% 40% 47% 50% 42% 42% % Lives with partner (spouse) 54% 61% 39% 38% 42% 45% Food activities % Involved in food shopping 37% 51% 34% 48% 36% 40% % Active in cooking (including joint cooking) 5% 59% 26% 47% 24% 51% Table options Table 2 presents the results of the regression analyses to assess possible interactions between the effects of ethnic group and gender on the meat-related variables and BMI category. In these analyses age and level of education were included as controls. The results show that interactions involving ethnic group and gender were significant for several of the dependent variables. Dependent variable in the first row of Table 2 is preferred meat portion size (overall average 133 g). The preferred meat size increased slightly with age (B = 5.8), was larger among men (B = .21), in particular among migrant men (B = .17); it was somewhat smaller among the bicultural migrants (B = −18.9) but higher among the bicultural Turks (B = 60.0). For ease of interpretation of the regression coefficients, Table 3 presents the average preferred meat portion size split out by ethnic group and gender; the first row of Table 3 shows that almost all the bicultural Turkish men chose the largest portion. In contrast, the smallest average was found among the bicultural Chinese women. Table 2. Regression of the meat-related variables and BMI category on age, education and the dummy variables that represent ethnic group and gender. Meat portion size Meat eating days Familiar with replacers Use of replacers Familiar with meatless day Willing to reduce BMI category  B =  B =  B =  B =  B =  B =  B = Constant 99.2 (8.3) 3.49 (.38) .39 (.10) 2.37 (.30) −.00 (.10) 2.32 (.30) 1.76 (.10) Age divided by 10 5.8* (2.4) .07 (.11) .05 (.03) .10 (.09) .02 (.03) −.16 (.09) .13*** (.03) Level of education (low–high) −.5 (1.9) −.08 (.09) .01 (.02) −.01 (.07) .15*** (.02) .22** (.08) −.00 (.02) Ethnic group (migrant = 1, else 0) 9.1 (4.6) 1.48*** (.21) −.08 (.05) −.48** (.18) −.12* (.05) −.48** (.18) −.14* (.06) Turkish (=1, else 0) −9.0 (4.9) .05 (.23) .04 (.06) −1.04*** (.19) .03 (.06) .59** (.20) .08 (.07) Gender (man = 1, else 0) 21.0*** (4.3) .57** (.19) −.03 (.05) −.52** (.16) .01 (.05) −.16 (.15) .21*** (.05) Migrant man (=1, else 0) 17.3** (6.1) −.13 (.28) .03 (.07) .23 (.23) .06 (.07) −.04 (.23) .09 (.08) Turkish man (=1, else 0) −3.3 (6.0) −.05 (.28) −.15* (.07) .25 (.22) −.06 (.07) −.92*** (.22) .35*** (.08) Bicultural migrant (=1, else 0) −18.9*** (4.3) −.09 (.20) .10* (.05) .59*** (.16) .06 (.05) .16 (.17) −.04 (.06) Bicultural Turk (=1, else 0) 60.0*** (6.3) −.62* (.29) .46*** (.08) −.65** (.23) .42*** (.07) −.25 (.23) −.09 (.08) R square .246 .128 .112 .276 .126 .190 .265  N = 1001 1056 1056 570 1056 404 853 Note: Table entries are unstandardized regression coefficients and standard errors. All dependent variables were coded from low to high. The dummy variable Bicultural refers to those who felt to be Turkish and/or Dutch or Chinese and/or Dutch. * p < .05; ** p < .01; *** p < .001. Table options Table 3. Meat consumption, meat reduction and BMI, distinguished by ethnic group and gender. Turks Chinese Natives Mono-cultural Bicultural Mono-cultural Bicultural Men (n = 121) Women (n = 122) Men (n = 50) Women (n = 57) Men (n = 97) Women (n = 105) Men (n = 67) Women (n = 81) Men (n = 172) Women (n = 185) Meat consumption Mean preferred meat portion size and SD (g) 145 (45) 118 (44) 197 (12) 148 (26) 160 (37) 122 (37) 143 (34) 103 (35) 134 (42) 113 (40) Mean number of meat eating days per week and SD 5.4 (1.9) 5.1 (1.3) 4.8 (1.1) 4.2 (0.7) 5.4 (1.9) 5.0 (2.1) 5.3 (2.0) 4.9 (2.1) 4.1 (2.1) 3.5 (2.0) Meat reduction % Familiar with meat replacers 31% 54% 100% 98% 44% 46% 57% 55% 51% 54% % Uses meat replacers regularly (if familiar) 0% 2% 0% 0% 14% 15% 26% 33% 22% 29% % Familiar with meatless day impact 24% 28% 86% 70% 36% 26% 39% 36% 40% 40% % Potentially willing to reduce (including “yes maybe”) 11% 24% 8% 79% 53% 54% 37% 38% 42% 38% BMI % Don't know/Don't want to say 1% 30% 0% 0% 29% 35% 24% 21% 12% 25% % Overweight (BMI > 25) of those with data available 63% 13% 74% 0% 26% 4% 24% 3% 28% 14% Table options The number of meat eating days per week (overall average 4.6) is the second dependent variable (depicted in the second column of Table 2 and the second row of Table 3). The frequency of meat eating was higher among the migrants (B = 1.48) and among men (B = .57), but slightly lower among the bicultural Turks (B = −.62). It should be noted that about 93% of the participants reported to have one warm meal per day, except for the Chinese, of whom 22% reported to have more than one. Among the migrants taken together, having more than one warm meal per day was weakly correlated with a higher number of meat eating days per week (Spearman's rho = .19, p < .001). Familiarity with meat replacers is the third dependent variable in Table 2 and Table 3. About 50% was familiar with meat replacers and this was slightly lower among Turkish men (B = −.15), slightly higher among the bicultural migrants (B = .10), especially the bicultural Turks (B = .46). The use of meat replacers is the fourth dependent variable in Table 2 and Table 3. Use frequencies were lower among the migrants (B = −.48), in particular the Turks (B = −1.04), also lower among men (B = −.52), higher among the bicultural migrants (B = .59), except for the bicultural Turks (B = −.65). As Table 3 reveals, almost none of the Turks reported to use meat replacers, but relative frequent users of the replacers were the bicultural Chinese and native woman. The fifth dependent variable in Table 2 and Table 3 is familiarity with the impact of a meatless day. About 38% said to be familiar with this impact, which increased with level of education (B = .15), and was slightly lower among the migrants (B = −.12), but higher among the bicultural Turks (B = .42). This was the only meat-related variable in Table 2 that is not correlated with gender. Willingness to reduce (the sixth dependent variable) increased with level of education (B = .22), was lower among the migrants (B = −.48), and lowest among the Turkish men (all Turkish B = .59, Turkish men B = −.92). As shown by Table 3, about 10% of the mono-cultural and bicultural Turkish men said they would be willing to reduce meat consumption – the highest percentage (79%) was found among bicultural Turkish women. About 19% of the participants were unable or unwilling to report their height and their body weight. Table 3 reveals that this percentage was higher among the women, particularly the mono-cultural Turkish and Chinese. BMI category increased with age (B = .13), was lower among the migrants (B = −.14), but higher among the men (B = .21), in particular the Turkish men (B = .35). As shown by Table 3, the lowest percentage with overweight (BMI > 25) was found among the bicultural Turkish women and the highest among mono-cultural and bicultural Turkish men. The Turkish men had almost no missing data on this variable; in this group (n = 170), BMI category had a positive correlation with preferred meat portion size (r = .24, p < .01) but not with the number of meat eating days (r = .09, p > .05). Each of the reasons for frequently and not frequently eating meat (hereafter called promoting and limiting reasons) was used as a dependent variable in the regression analysis to test for interactions between gender and ethnic group. The results (not shown here) revealed 6 significant (p < .01) interactions in regressions of promoting reasons and 1 in regressions of limiting reasons on the dummy variables. Table 4 shows that most gender differences were found in the Turkish sample. For instance, Turkish men chose more often “being a meat lover” (68% versus 32% among women) and less often “it fits well with what I normally eat” (58% versus 73% among women). These and other gender differences were smaller among the Chinese and especially the natives. Overall, the participants mentioned more promoting than limiting reasons. Both Turkish men and women often chose health as a limiting reason (54% and 55%); Chinese men and women often mentioned ‘my finances’ (41% and 34%). Table 4 also reveals that the natives chose more often than the migrants health as a promoting and a limiting reason, and that they referred more often to ethical limiting reasons (i.e. the environment or animal welfare), without much difference between the genders. Table 4. Gender differences in reasons for frequently and not frequently eating meat in each of the samples. Turks Chinese Natives Men (n = 171) Women (n = 179) Men (n = 164) Women (n = 186) Men (n = 172) Women (n = 185) Frequently eating meat Being a meat lover 68% 32%*** 59% 37%*** 58% 48%* It fits well with what I normally eat 58% 73%** 30% 40% 33% 27% Being used to it 44% 58%* 44% 36% 35% 35% To get satiety 26% 5%*** 24% 28% 30% 29% It's healthier to eat meat frequently 9% 10% 13% 19% 40% 38% Others in the household want to eat meat 11% 32%*** 10% 18%* 16% 20%* Not liking anything else 12% 2%*** 9% 14% 17% 18% The speciality of the occasion 10% 14% 2% 3% 4% 7% It is a sign of being wealthy 2% 2% 2% 4% 6% 13%* Not frequently eating meat It's healthier not frequently eating meat 54% 55% 18% 25% 29% 35% I like to vary 27% 38%* 21% 20% 36% 36% Not liking meat that much 36% 44% 10% 16% 25% 35%* Because of my finances 16% 10% 41% 34% 34% 28% Being used to it 32% 26% 16% 22% 17% 20% It's better for the environment 2% 2% 29% 23% 37% 38% Others in the household don't want to eat meat 23% 34%* 2% 7% 20% 18% Because animal welfare is important 4% 3% 10% 22%** 17% 23% Because of my religion 1% 1% 6% 4% 8% 11% Note: Statistically significant gender differences in a sample are marked by asterisks. * p < .05; ** p < .01; *** p < .001.