Healthy eaters beat unhealthy eaters in prototype evaluation among men, but abstinence may pose a risk for social standing

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=rhpb20 Download by: [Tampere University] Date: 19 September 2016, At: 23:58 Health Psychology and Behavioral Medicine An Open Access Journal ISSN: (Print) 2164-2850 (Online) Journal homepage: http://www.tandfonline.com/loi/rhpb20 Healthy eaters beat unhealthy eaters in prototype evaluation among men, but abstinence may pose a risk for social standing Marja Kinnunen, Nelli Hankonen, Ari Haukkala, Britta Renner, Piia Jallinoja, Clarissa M. L. Bingham & Pilvikki Absetz To cite this article: Marja Kinnunen, Nelli Hankonen, Ari Haukkala, Britta Renner, Piia Jallinoja, Clarissa M. L. Bingham & Pilvikki Absetz (2015) Healthy eaters beat unhealthy eaters in prototype evaluation among men, but abstinence may pose a risk for social standing, Health Psychology and Behavioral Medicine, 3:1, 323-336, DOI: 10.1080/21642850.2015.1095097 To link to this article: http://dx.doi.org/10.1080/21642850.2015.1095097 © 2015 The Author(s). Published by Taylor & Francis. Published online: 19 Oct 2015. Submit your article to this journal Article views: 280 View related articles View Crossmark data Healthy eaters beat unhealthy eaters in prototype evaluation among men, but abstinence may pose a risk for social standing Marja Kinnunen a *, Nelli Hankonen b,c , Ari Haukkala c , Britta Renner d , Piia Jallinoja a,e , Clarissa M. L. Bingham a,f and Pilvikki Absetz a,g a Welfare and Health Promotion Unit, National Institute for Health and Welfare, Helsinki, Finland; b Department of Social Psychology, School of Social Sciences and Humanities, University of Tampere, Tampere, Finland; c Department of Social Research/Social Psychology, University of Helsinki, Helsinki, Finland; d Department of Psychology, University of Konstanz, Konstanz, Germany; e Consumer Society Research Center (KTK), Department of Political and Economic Studies, University of Helsinki, Helsinki, Finland; f Cancer Society of Finland, Helsinki, Finland; g School of Health Sciences, University of Tampere, Helsinki, Finland (Received 28 January 2015; accepted 14 September 2015) Although previous studies have shown that unhealthy eating is associated with a positive image of a typical unhealthy eating peer (prototype), their focus on prototypes is typically narrow/ limited. The present study addresses this gap in the literature and investigates two aspects in peer evaluations: (1) healthy vs. unhealthy and (2) abstainer vs. chooser. Moreover, their mean differences, interrelationships and associations with eating behavior will be examined. Methods: Men in military service (N= 1824, Age M = 19.8) reported their eating behaviors and evaluated either (a) Vegetable chooser and Vegetable abstainer prototypes (N= 920) or b) Fat chooser and Fat abstainer prototypes (N= 904) on a scale containing 17 antonyms. Exploratory and confirmatory factor analyses suggested a three-factor structure: Selfregulation, Social standing in peer group and Appearance (Comparative Fit Indexes 0.82– 0.87; Tucker–Lewis index 0.78–0.84; Root Mean Square Error of Approximation 0.07– 0.08). Results: Healthy eaters (i.e. Vegetable chooser and Fat abstainer) were evaluated higher on Self-regulation and Appearance than respective unhealthy eaters (i.e. Vegetable abstainer and Fat chooser) (ps < .001). However, Fat abstainer was rated lower than Fat chooser on Social standing in peer group (p< .001). Neither Fat chooser nor Fat abstainer prototype was associated with fatty food consumption. Fruit and vegetable consumption was associated with higher Self-regulation ratings for Vegetable chooser (β= .10, p< .01) and lower Self-regulation ratings (β=−.13, p< .001) and Appearance ratings (β=−.08, p< .05) for Vegetable abstainer. Conclusions: Among young men, healthy eating peers are evaluated as more self-regulative and better looking than unhealthy eating peers. Rating healthy eaters more positively is related to higher fruit and vegetable consumption. Prototypes play a role in young men’s eating behavior, although the role differs for vegetables and fatty foods. Interventions to increase fruit and vegetable consumption should consider addressing the vegetable eater prototype. Keywords: prototypes; social image; eating behavior; food choice; peers © 2015 The Author(s). Published by Taylor & Francis. *Corresponding author. Email: marja.kinnunen@thl.fi This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Health Psychology and Behavioral Medicine, 2015 Vol. 3, No. 1, 323–336, http://dx.doi.org/10.1080/21642850.2015.1095097 1. Introduction The consumption of fruits and vegetables among young men is infrequent and far from the national recommendation’s minimum of 500 g per day (Bingham et al., 2010; National Nutrition Council, 2005). Eating behavior is driven by biological needs such as satisfying hunger. Nonetheless, food choices are also dependent on cultural and social norms (Stead, McDermott, MacKintosh, & Adamson, 2011) and also provide a way to express group membership like ethnicity or social class, and thus, need to be examined from wider perspectives. Understanding the antecedents of healthy and unhealthy food consumption would enable evidence-based intervention design. An influential theoretical framework depicting social influences on health-related behavior is the prototype-willingness model, emphasizing the role of prototypes in the decision-making process. Prototypes are defined as social images of a typical peer who engages in certain behavior (e.g. “a typical smoker of your age”) (Gibbons & Gerrard, 1995; Hahn & Renner, 1998). According to the model, positive or negative valence of the prototype predicts the willingness to engage in the respective behavior. Two recent meta-analyses have shown support for the prototype-willingness model (Todd, Kothe, Mullan, & Monds, 2014; van Lettow, de Vries, Burdorf, & van Empelen, 2014). One of the strengths of the prototype-willingness model is the investigation of the social impact of peers upon behavior through social images (rather than direct self-reported influence). Van Lettow et al.’s(2014) meta-analysis shows that prototypes have effect not only on intention and willingness, but on behavior as well. For example, young smokers tend to evaluate a typical peer smoker more positively than a non-smoker (Piko, Bak, & Gibbons, 2007). Also, a more positive perception of a typical peer drinker predicts higher willingness to drink in social situations (Zimmermann & Sieverding, 2011) as well as actual alcohol consumption (Gerrard et al., 2002). Studies exploring social images of healthy and unhealthy eaters show that healthy eaters are generally perceived more positively compared to unhealthy eaters (Barker, Tandy, & Stookey, 1999; Gerrits, de Ridder, de Wit, & Kuijer, 2009; Gerrits et al., 2010; Oakes & Slotterback, 2005). When asked to describe typical healthy eaters, people generally give positive descriptions, but also tend to associate some negative characteristics with healthy eaters. While healthy eaters are described with virtuous self-regulative characteristics like “active”(Gerrits et al., 2009), “dutiful”and “self-controlled”(Barker et al., 1999), they are also perceived as “self-centered” (Fries & Croyle, 1993). Unhealthy eaters in turn get more negative descriptions related to their appearance, for example, “chubby”, but are also perceived positively with characteristics, for example, “sociable”,“entertaining”and “masculine”(Fries & Croyle, 1993; Gerrits et al., 2009; Oakes & Slotterback, 2005). Gender differences in the perceptions of the prototype content have not always been found, (Gerrtits et al., 2009), but male adolescents have been found to hold more positive unhealthy eater prototypes and more negative healthy eater prototypes (Steinhilber, Fuchs, & Dohnke, 2013). The associations of eater prototypes and eating seem to be complex, although several studies have shown a positive relationship between positive prototypes of unhealthy eater and related eating behaviors (Barker et al., 1999; Gerrits et al., 2009,2010). For example, gender seems to moderate the effect of eater prototypes on eating behaviors. Prototype perception partially explains gender differences in eating among adolescents, especially among older age adolescents (Steinhilber et al., 2013). Also, favorable images of unhealthy eaters associate with higher fatty food consumption among girls, but with higher soft drink consumption among boys (Gerrits et al., 2009). Dohnke, Steinhilber, and Fuchs (2015) tested the full prototype-willingness model on eating behavior and found that among adolescents, only the unhealthy eater prototype was associated 324 M. Kinnunen et al. with eating. The unhealthy eater prototypes predicted subsequent unhealthy eating, and not only through intention and willingness as the model predicts, but also directly. This finding is actually in line with van Lettow et al.’s(2014)finding that health-risk prototypes generally exhibit stronger associations with outcome variables than health-protective prototypes (van Lettow et al., 2014). In terms of eating, similarly favorable images of healthy eaters have not always been found to relate to one’s own food choices (Gerrits et al., 2009), probably because healthy eaters tend to get positive ratings not only from other healthy eaters but from unhealthy eaters as well (Barker et al., 1999). So far, the question in literature has been whether and how prototypes influence the healthiness of eating behavior. However, Rivis, Sheeran, and Armitage (2006) noted that prototypes could actually refer either to a typical person engaging in a behavior (actors) or to a typical person abstaining from performing a certain behavior (abstainers). Hence, prototypes can be inhibitory (i.e. not wanting to be like a typical peer who avoids healthy foods) or facilitating (i.e. wanting to gain the favorable characteristics of the healthy food choosing peer). Combining the different types of prototypes, four prototype dimensions can be distinguished: (1) risky behavior actor prototype representing a typical peer who engages in risky behaviors (e.g. eats unhealthy foods), (2) a risky behavior abstainer prototype representing a typical peer who abstains from risky behavior (e.g. does not eat unhealthy foods), (3) a healthy behavior actor prototype representing a typical peer who actively engages in healthy behavior (e.g. eats healthy foods) and finally (4) a healthy behavior abstainer prototype representing a typical peer who abstains from healthy behaviors (e.g. does not eat healthy foods). Interestingly, the (few) positive descriptions obtained to describe unhealthy eaters often reflect social and relaxing aspects, like “fun-loving”(Barker et al., 1999)or“sociable”(Gerrits et al., 2009). Also, the (few) negative descriptions obtained for healthy eaters tend to reflect abstinence from something, for example, “serious”or “highly strung”(Barker et al., 1999). It is possible that as respected and eligible healthy eating may be, abstaining from unhealthy foods is not a desirable strategy for pursuing health in social situations. One of the core ideas of the prototype-willingness model is that health behaviors take place in social context, in which intention to do something either does not exist at all or is so weak that it can be overridden by willingness to do something entirely different (Gibbons & Gerrard, 1995). Most recent findings suggest that in eating context, the actual behavior can be predicted directly by prototypes (Dohnke et al., 2015). Therefore, the distinction outlined by Rivis et al. (2006) raises interesting questions; does the abstaining or active nature of (healthy or unhealthy) eating behavior make a difference in terms of prototype favorability? Healthy eating differs from other health behaviors in a marked way: For smoking, there is only one way of choosing a healthy option, that is, by abstaining from smoking. Similarly, for physical activity, the only healthy way is to actively engage in being physically active. However, one can eat healthily both by choosing healthy foods and by abstaining from unhealthy ones. It is unclear how this influences prototypes. Previous studies have examined associations between unhealthy or healthy eater prototypes and actual unhealthy or healthy eating behavior. To our knowledge, none has investigated all the four above-mentioned different prototypes, and their associations with respective eating behavior. The present study will address this gap in the literature. The aims of the study are (1) to explore healthy behavior actor or abstainer prototypes (“Vegetable chooser”and “Vegetable abstainer”) and risky behavior abstainer or actor prototype (“Fat abstainer”and “Fat chooser”), (2) to identify whether these four prototypes have similar factor structures and (3) whether these factors are interrelated. The final aim (4) is to explore the associations of these obtained factors with eating behavior. Health Psychology and Behavioral Medicine 325 Based on previous research among adolescents (Gerrits et al., 2009,2010), we expect that the prototype evaluations will depict the following types of prototype factors among young adult men too: one that represents self-regulatory characteristics, another that contains items that relate to sociable characteristics and a third that relates to appearance. Considering the previous research on evaluations of healthy and unhealthy eaters (e.g. Barker et al., 1999; Gerrits et al., 2009), we present the following hypotheses: (1) the healthy eater prototypes (Vegetable chooser and Fat abstainer) are rated more positively by their appearance and self-regulative characteristics than the respective unhealthy eater prototypes (Vegetable abstainer and Fat chooser). (2) Assuming that the sociable characteristic factor is found, choosers are rated over abstainers in the sociable characteristics factor within both healthy and unhealthy eating behavior. (3) As most healthy or unhealthy eater prototypes found in earlier literature seem to reflect self-regulative traits vividly (Gerrits et al., 2009), we expect self-regulative prototypes to play a crucial role in predicting respondents’own eating behavior. And (4) finally, assuming that abstaining from food is not socially acceptable, and knowing from previous research that unhealthy prototypes have more predictive power on behavior (Dohnke et al., 2015; Gerrits et al., 2009), we assume that Vegetable abstainer prototypes will have the strongest association with respective eating behavior. 2. Methods The present study is a part of the DefenceNutri Study (Bingham et al., 2012; Hankonen, Kinnunen, Absetz, & Jallinoja, 2014) that took place in two garrisons: Kainuu Brigade in North-Eastern and Armoured Brigade in Southern Finland. In Finland, every man is liable for military service. During their service time, conscripts live in garrisons. At garrisons, men are provided three meals plus an optional evening snack. The provided meals are planned to meet the national nutritional recommendations as well as the demands the conscripts face during the service (e.g. if having a physically rough camping, the provided meals contain more energy). Fatty foods, such as pizzas, kebabs or burgers and so on are eaten as snacks rather than as planned meals. However, after the first weeks, if not on encampment, they often spend weekends and some other free days at home. Nearly 80% of the age cohort completes the service. About 95% of those who serve enter service at the year when they turn 19 or 20 years (Defence Command Public Information Division, 2012). Data were collected in 2008 and 2009 from four cohorts entering their service in January and July. Two service units from both garrisons were included to participate, each unit containing from 80 to 150 men. Participation was voluntary, and informed signed consent was obtained from those willing to participate. Participation took place during the service time. The study protocol was approved by the ethics committee of the Hospital District of Helsinki and Uusimaa. 2.1. Participants The four cohorts contained 2.125 male conscripts of whom 1.824 (85%, Age: M = 19.8 years, SD = 0.81, range 19–28) filled in questionnaire in Finnish language during the first week of their military service. Of all participants, 98.6% were 22 or younger. Excluding the 25 older participants from the analyses did not change the results for the correlations or regression analyses. Thus, they were included in all analyses. Dropout was due to interruption of service, being ill during the measurements, military transfers or refusal to participate in the study. The participants reported their highest education achieved as follows: secondary school 12%, vocational training 42%, upper secondary school 44% and 3% had higher or some other education. Majority of the participants (76%) still lived with their parents, and only 22% lived alone or with a spouse. 326 M. Kinnunen et al. 2.2. Measures 2.2.1. Prototypes and prototype factors Previous eater prototype studies have measured “typical healthy eater/unhealthy eater of your age”or “a typical person of your age who eats the following foods …” We decided to choose foods that (a) are very essential part of any healthy diet (i.e. fruit and vegetables) and (b) foods that are unhealthy and more likely to be consumed with peers than alone and (c) available in garrison area (fatty food). Both food groups have been targeted by health education for decades, and hence, their associations with health are relatively well known among population too (Jallinoja, Niva, Helakorpi, & Kahma, 2014). Moreover, the consumption of fruits and vegetables among conscripts is too low as regards recommendation (Bingham et al., 2009). Participants were randomly assigned into two groups; either on Fat group where they evaluated a typical “Fat chooser”and a typical “Fat abstainer”(N= 904) or on Vegetable group where they were asked to evaluate a typical “Vegetable chooser”and a typical “Vegetable abstainer” (N= 920).The question: “We will ask you to evaluate a typical conscript, who behaves in a way described below”was followed by the definitions of an abstainer and an actor: Abstainer: “A typical conscript who tries to avoid fatty foods and snacks”(Fat condition)/ “vegetables and fruits”(Vegetable condition) and Actor: “A typical conscript who strives for choosing fatty foods and snacks”(Fat condition)/“fruits and vegetables”(Vegetable condition) “at every meal”. Participants evaluated each four types on scales of 17 antonyms. The adjectives were selected based on earlier studies (Barker et al., 1999; Fries & Croyle, 1993) and on a qualitative pilot study among male conscripts. The adjective pairs were asked to be answered on a 5-point scale with the antonym adjectives at the extremes. Participants were asked to “choose an option from the scale that most accurately represents your opinion”. The antonyms were the following: easygoing–uptight, popular–unpopular, unreliable–reliable, convincing–unconvincing, physically fit–physically unfit, childish–grown-up, fat–skinny, self-indulgent–austere, masculine–feminine, careless–meticulous, responsible–irresponsible, insecure–confident, muscular–lanky, fashionable– unfashionable, dumb–intelligent, attractive–unattractiveandgoodcompany–dull. Appropriate reversions were done for higher scores to reflect more positive evaluations. Exploratory factor analyses with Maximum likelihood extraction and Varimax rotation were conducted to investigate the factor structure of the four prototypes (Vegetable chooser, Vegetable abstainer, Fat chooser and Fat abstainer). The obtained factor solution was investigated with confirmatory factor analysis. The first analyses yielded a three-factor solution for Fat chooser and Vegetable avoider and four-factor solutions for Fat avoider and Vegetable chooser. For the three-factor solution the criteria were eigenvalues ≥1 and scree plots confirmed the results. For the four-factor solution, eigenvalues were ≥1, but the scree plots were less clear. Two items were dropped (convincing and austere) due to cross-loadings and field researchers’feedback from the comments from conscripts concerning the meaning of the items. In subsequent analyses, the three-factor solution was supported by eigenvalues and scree plots. The three-factor explanatory factor solutions explained from 46.3% to 52% of variance. The factor structure was examined with confirmatory factor analysis, with Full Information Maximum Likelihood for estimating missing data. As an indication of acceptable model fit, Comparative Fit Indexes ranged from 0.82 to 0.87, Tucker–Lewis index from 0.78 to 0.84 and the Root Mean Square Error of Approximation from 0.07 to 0.08. For all four prototypes, three similar factors emerged. The factors were Self-regulation (items: reliable, grown-up, meticulous, responsible and intelligent), Social standing in peer group (items: easy-going, popular, masculine and good company) and Appearance (items: muscular, fashionable, attractive, skinny and physically fit). Prototype items were subsequently summed into Health Psychology and Behavioral Medicine 327 composite scores for the analyses. The Cronbach’s alphas were satisfactory at 0.60 level. Alphas for other Fat abstainer sum scores: Self-regulation α= 0.68 and Social standing in peer group α= 0.64. For Fat chooser sum scores: Self-regulation α= 0.73, Social standing in peer group α= 0.66 and Appearance α= 0.68. For Vegetable abstainer sum scores: Self-regulation α= 0.72, Social standing in peer group α= 0.60 and Appearance α= 0.65. For Vegetable chooser sum scores: Self-regulation α= 0.74, Social standing in peer group α= 0.73 and Appearance α= 0.66.). As fat and vegetable prototypes were measured in different groups, the statistical comparisons were made within the group (i.e. fat prototypes in one and vegetable prototypes in another). 2.2.2. Eating behavior Eating was measured by a 36-item food frequency questionnaire asking on how many days during the past week in civilian life the respondent had consumed certain foods (item range 0–7). The questionnaire was based on several corresponding questionnaires among Finns (e.g. Paalanen et al., 2006) and adjusted for conscripts on the bases of a previous food diary study among Finnish conscripts (Bingham et al., 2009). For the purposes of the present study, two eating indexes were created: Fruit and Vegetable Index to describe the consumption of fruits or berries and fresh vegetable (the mean of the two items), and Fatty Food Index to measure fatty food consumption (the mean of five items: French fries, potato chips, pizza and kebab, hamburgers and hot dogs, meat pies and savory pastries). The Fatty Food Index was based on earlier studies of eating behavior among men in military service (Bingham et al., 2011; Jallinoja et al., 2011). The Fruit and Vegetable Index varied between 0.00 and 7.00 (M = 2.93, SD = 1.88) and the Fatty Food Index between 0.00 and 6.20, (M = 1.01, SD = 0.78). 2.3. Statistical analysis IBM SPSS 22 was used to obtain descriptive statistics, for explanatory factor analysis with maximum likelihood extraction and varimax rotation. For confirmatory factor analyses, Mplus version 6.1 was used. The first two hypotheses were tested with t-tests. A Pearson correlation coefficient was used to explore if the factors are related, as well as for the third hypothesis. For the fourth and fifth hypotheses, to predict fat or vegetable eating by the respective abstainer or chooser prototype factors, linear regression analysis was also used. 3. Results Our first hypothesis anticipated typical healthy eater prototypes (Vegetable chooser and Fat abstainer) to be rated more positively than respective unhealthy eater prototypes (Vegetable abstainer and Fat chooser). The findings supported the hypothesis for two of the three factors: Self-regulation and Appearance. The means and confidence intervals of the different prototype factors are reported in Figure 1. According to two-tailed t-tests, a typical Vegetable chooser was rated higher than Vegetable abstainer on Self-regulation (t(909) = 22.94, p<.001) and on Appearance (t(912) = 21.02, p<.001). Fat abstainer was rated higher than Fat chooser on Selfregulation (t(892) = 25.35, p<.001) and on Appearance (t(892) = 25.88, p<.001). The ratings of Social standing were more strongly influenced by the abstaining vs. choosing nature rather than the healthiness of behavior. This supports our second hypothesis, which anticipated more positive ratings for choosers rather than abstainers; both Vegetable chooser and Fat chooser were rated higher on Social standing than the respective abstainers (Vegetable (t(910) = 2.91, p= .004; Fat (t(892) = −8.78, p<.001). Confirming the first hypothesis, single-item comparisons 328 M. Kinnunen et al. revealed that healthy eater prototypes were perceived more positively along 16 items than respective unhealthy eater prototypes (all ps≤.001). In respect of masculinity, the association was reverse: both Vegetable chooser and Fat abstainer were seen as less masculine than Vegetable abstainer (p= .001) and Fat chooser (p< .001). The profile for single items can be seen from Figure 2. The correlation matrix for the prototype factors of all four prototypes is shown in Table 1. Vegetable chooser prototype factors had significant negative correlations with all Vegetable abstainer prototype factors with two exceptions. Fat chooser prototype factors had significant negative correlations with all Fat abstainer prototype factors with two exceptions. Between chooser vs. abstainer prototypes, the correlations were strongest for Self-regulation. Moreover, within each prototype, the factor correlations were positive with the strongest correlations between Self-regulation and Appearance (see Table 1). Figure 1. Means and confidence intervals for the Self-regulation, Social standing in peer group and Appearance prototype factors. Figure 2. Profile diagram for the single items. Health Psychology and Behavioral Medicine 329 Table 1. Means, standard deviations and correlations between eater prototype factors. Vegetable chooser Vegetable abstainer Mª SD Self-regulation Social standing Appearance Self-regulation Social standing Vegetable chooser Self-regulation 3.44 0.53 1 Social standing 3.17 0.50 .35** 1 Appearance 3.34 0.49 .53** .46** 1 Vegetable abstainer Self-regulation 2.75 0.57 −.39** −.11* −.25** 1 Social standing 3.10 0.47 −.06 −.20** −.05 .23** 1 Appearance 2.74 0.55 −.38** −.20** −.39** .52** .32** Fat abstainer Fat chooser Self-regulation Social standing Appearance Self-regulation Social standing Fat abstainer Self-regulation 3.47 0.54 1 Social standing 3.07 0.48 .26** 1 Appearance 3.35 0.53 .42** .34** 1 Fat chooser Self-regulation 2.75 0.49 −.35** −.04 −.22** 1 Social standing 3.26 0.50 .15* −.11** .07 .17** 1 Appearance 2.59 0.52 −.37** −.12* −.40** .55** .12** Notes: a Min. = 1, Max. = 5. *p< .05 **p< .001. 330 M. Kinnunen et al.