Body Image, Peer Effects and Food Disorders (PDF)

Body Image, Peer Effects and Food Disorders: Evidence from a Sample of European Women Joan Costa-Fonta,b,d and Mireia Jofre-Bonetb,c aEuropean Institute and bLSE Health, LSE, London UK cDepartment of Economics, City University, London, UK dCESifo, Munich, Germany Working Paper No. 15/2009 First published in November 2009 by: LSE Health The London School of Economics and Political Science Houghton Street London WC2A 2AE All rights reserved. No part of this paper may be reprinted or reproduced or utilised in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieve system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this publication is available from the British Library ISBN [978-0-85328-008-8] Corresponding author Dr Joan Costa-Font LSE Cowdray House London WC2A 2AE. E-mail: [email protected] The authors express their gratitude to participants at the London Health Economics Group seminar at LSE, FEDEA-Health workshop in Barcelona, and the internal seminar at the Economics Department of City University for helpful comments. Also, support from CESifo Munich Germany is gratefully acknowledged.  Abstract Excessive preoccupation with self-image has been pinpointed as a factor contributing to the proliferation of food disorders, especially among young women. To provide an economic basis for this argument this paper models how ‘self-image’ and ‘other people’s appearance’ influence health-related behaviour. Self-image (identity) is claimed to be biased towards anorexic women by social norms and peer pressure, increasing the probability of women experiencing a food disorder. This paper empirically tests this claim using data from a representative, cross-sectional European survey for 2004. A two-step empirical strategy was used. First, the probability was estimated of a woman ‘being extremely thin’ and at the same time ‘seeing herself as too fat’. The findings revealed robust evidence suggesting that (different definitions of) peer effects average out, and that a larger peer body-mass decreases the likelihood of being anorexic. Second, the two processes were estimated separately, using a recursive system, which suggested that self-image was associated with body weight when unobservable variables explaining both processes were controlled for. (These processes were found to be positively and significantly correlated). As expected, several definitions of peers’ body mass were found to decrease the likelihood of women being thin or extremely thin, when common unobservable variables were controlled for. Key words: self-image, identity, body image, eating disorders, anorexia JEL: I12, Q18. 2  Content Abstract......................................................................................................................................2 1. Introduction........................................................................................................................ 4 6 9 12 21 30 31 2. Eating disorders and body image: a summary................................................................... 3. An economic decision model for eating disorders............................................................. 4. Data and Methods ............................................................................................................ 5. Results and discussion ..................................................................................................... 6. Conclusion ....................................................................................................................... 7. References........................................................................................................................ 3  1. Introduction It is becoming increasingly clear that standards of physical appearance are important and powerful motivators of human behaviour. However, the content and formation of these ideal- body standards have yet to be explored in economics literature. Case studies of eating disorders constitute a prime example how changes in social attitudes towards physical appearance explain irregular health behaviour such as anorexia and bulimia nervosa among women. These two phenomena are difficult to distinguish from each other because they have similar characteristics, namely distorted body image accompanied by an eating obsession, hence they are referred to here as ‘eating disorders’. Eating disorders can have damaging, and even devastating and life-threatening effects (American Psychiatric Association [APA] 2000). People who weigh at least 15% less than the normal weight for their height may not have enough body fat to keep their organs and other body parts healthy (APA, 2000). Disordered eating behaviour is a condition that can have long-term physical and social consequences (Hill, 1993). Indeed, about 6% of those who suffer from anorexia nervosa die from it (Birmingham et al, 2005). As these disorders tend to be longstanding, the prevalence rates for bulimia nervosa among young adult women have risen and are now at 1%-3% (Hudson et al, 2007). Given that the relatively young are more at risk, it becomes especially important to understand how food disorders are engendered. The reasons for the increasing trend towards food disorders are yet to be fully understood. Goldfarb et al. (2009) proposes a model explaining anorexic disorders (low calorie intakes, purging behaviours) that is based on taste variations and on rational choices to be underweight. However, Goldfarb’s model does not attempt to include or explain the formation of self-image, which determines individual tradeoffs between desired weight and health behaviour. In social science literature the formation of social identity is seen as a key factor and it is thought that food disorders are probably the result of some ‘socially transmitted’ standard of ‘ideal’ body image affecting food intake and exercise.1 Traditional social psychology literature regards social image as being continually under construction and essential in determining physical, psychological and social equilibrium (Schilder, 1958). When applied to food disorders, this could explain some extreme forms of weight aversion. 1 Fairburn and Cooper (1984) report on an experiment proving that women have a clear aspiration to be thinner, and that this aspiration is more marked in women with bulimia nervosa. 4  This is consistent with evidence suggesting that network phenomena appear to be relevant to the biological and behavioural trait of obesity (Christakis and Fowler, 2007). However, this is contested in Cohen-Cole and Fletcher (2008). Trogdon et al. (2008) using a sample of adolescents found that mean peer-weight was correlated with individual weight, suggesting that early health behaviour is determined by social influences. However, the specific mechanisms behind peer-pressure are unknown and require careful examination; the fact that members of the peer group have a similar self-identity is a necessary, but not sufficient, condition for the presence of social-multiplier effects. This is precisely because they share common observable and unobservable characteristics and exogenous influences. Economic policy-making in the area of health prevention requires a better understanding of the effects of social identity and self-image on health, and the development of empirical strategies to measure these effects. Recent contributions to economics literature enable baseline modelling. Akerlof and Kranton (2000) wrote the seminal paper in this area and included an application to gender attitudes. Bodenhorn and Ruebeck (2003) created models for the influence of identity on ethnic preferences. However, there is not much in the literature on the role of social identity as a determinant of health. Blanchflower et al 2008 used Eurobarometer data for 29 countries to show that overweight perceptions and dieting were influenced by individual relative body mass index (BMI). Lakdawalla and Philipson (2002) referred to an ‘ideal weight’, and Etile (2007) examined the role of social norms on obesity and concluded that social norms have an effect on ideal body-weight (for women). Gardner (1996)2 discussed the role of body-image in behavioural reactions in cases where individuals perceived a large gap between their desired image and the one they actually had3 suggesting that this gap gave rise to permanently distorted self-perceptions of the body. Altogether, the power exerted by media stereotypes of beauty and the social norms that individuals are immersed in – especially the association between thinness, aesthetic ideals and success (Hill, 1993) – is widely accepted. Further, it has been suggested that the consequent fear of rejection based on physical appearance is behind the increase in the number of persons suffering from eating disorders. Hence, eating disorders are ‘socially formed’ rather than a personal pathology (Bordo, 2003). 2 According to Gardner (1996), body image includes two components: a perceptual component, which includes estimations of size and appearance, and an attitudinal component, which includes feelings and attitudes towards one’s own body. 3 Slade (1988) defined body image as ‘the picture we have in our mind of the size, outline and shape of our body and the feelings we have about these characteristics and parts that make them up.’ 5  Hutchinson (1982) points out that ‘body image’ refers not only to the description of the body but to the place ‘where body, mind and culture meet’. Accordingly, different cultural backgrounds are likely to exert idiosyncratic influences on the prevalence of food disorders, and these need to be controlled for. The aim of this paper is to build an economic model of eating disorders, especially anorexia, that relates social and environmental factors to ‘self-image’ and objective weight. Some of the implications of this model are taken to the data and the effect of underlying determinants is estimated. A representative European data-set on women is used since according to the APA (2000) women account for 90% of all anorexia nervosa. This paper focuses particularly on the effect of ‘peer weight’ (which is likely to influence self-image [social identity]) on the likelihood of anorexia, and the influence of self-image on individual weight. In a joint- modelling exercise, the paper then estimates the determinants of the probability of a woman being extremely thin and, at the same time, seeing herself as too fat. It then takes the two processes apart and estimates a recursive probit model of being extremely thin and perceiving one self as being too fat, finding that the unobserved factors explaining both processes are correlated. This paper supports the hypothesis that social pressure through peer-shape is determinant in explaining anorexia nervosa and distorted self-perception of one’s own body. To the author’s knowledge, there is no previous study examining anorexia that uses an economic decision-model perspective combining self-image – or self-identity4 – formation and individual health production. The structure of the paper is as follows: Section 2 provides some background on the issue of self-image and healthy eating among women. Section 3 proposes an economic model for eating disorders. Section 4 sets out the empirical strategy used, describes the data-set and estimates a reduced-form equation derived from the model. Section 5 presents the estimation results and Section 6 contains a discussion and conclusions. 2. Eating disorders and body image: a summary Different factors have been suggested as possible determinants of anorexia nervosa. Some of these are related to ‘nature’, i.e. gender, genes and predisposition. Other factors are related to 4 Note that self-image and (physical) self-identity are used interchangeably throughout the text. 6  ‘nurture’ i.e. parental values and socio-cultural influences. However, in the main, these determinants only seem to make individuals more (or less) susceptible to having their food and exercise intake shaped by the strong socio-environmental pressures that define what an ideal body looks like. Gender and anorexia. Girls who achieve sexual maturity ahead of their peers, with the associated development of breasts, hips, and other physical signs of womanhood, are at increased risk of becoming eating-disordered (Bordo, 2003). These girls often wrongly interpret their new curves as signs of fatness and feel uncomfortable because they no longer look like their peers, who still have childish bodies. A young woman in this group may ‘tackle’ her body, partly because she wants to take control and ‘fix’ her insecurity and partly because she is under the influence of a culture that equates success and happiness with thinness. For this group of young women, dieting, bingeing, purging, exercising, and other strange forms of behaviour are not random, but the result of a conscious decision process. Genes, family and anorexia. There is some evidence indicating that eating disorders may run in families. Parents influence their off-springs’ values and priorities, including those towards food. Additionally, it has been suggested that there may be a genetic component to traits such as obsessive behaviour, which include eating disorders. According to recent research (Archives of General Psychiatry 2006; 63:305-312) genetic factors account for more than half (56%) of the risk of developing anorexia nervosa and work on the genetics of bulimia and binge-eating is under way. There are suggestions that women who develop anorexia nervosa have excess activity in the brain's dopamine receptors, which regulate pleasure. This may explain why they feel driven to lose weight but receive no pleasure from shedding pounds (Guido Frank, et al 2005). Some people with eating disorders report having felt smothered in overprotective families. Others have felt abandoned, misunderstood and alone. Parents who overvalue physical appearance can unwittingly contribute to an eating disorder, as can parents who make critical comments, even in jest, about their children's bodies. Furthermore, families that include a person with an eating disorder tend to be rigid and ineffective at resolving conflicts. In some such cases mothers are emotionally cool while fathers are physically and/or emotionally absent. At the same time, there are high expectations of achievement and success. Children in this type of family learn not to disclose doubts, fears, anxieties, and imperfections. Instead 7  they try to solve their problems by manipulating weight and food, in an attempt to achieve the appearance of success, even if they do not feel successful (Bordo, 1993). People who are vulnerable to eating disorders are, in most cases, going through relationship problems, loneliness in particular. Even those who appear to have normal relationships reveal great fear of the criticism and rejection that would occur if their perceived flaws and shortcomings should become known (Bachar et al, 2001). Socio-environmental factors: the media. Many people believe media stereotyping helps explain why about 90% of people with eating disorders are women and only 10% are men (Thompson and Heinberg, 2002). In westernised countries, characterized by competitive striving for success, women often experience unrealistic cultural demands for thinness. According to Health magazine (April 2002), in the United States (US) 32% of female TV- network characters are underweight, while only 5% of the female audience is underweight. Similarly, only 3% of female TV-network characters are obese, while 25% of US women fall into that category. The differences between media images of happy, successful men and women are interesting. While women appear young, beautiful and thin, men are young or old, but strong and powerful in all the areas that matter – physically, in business, and socially. Thin is not desirable in men; power, strength and firmness are. Despite TV being a dominant media type, some studies have found magazine-reading to be a more consistent predictor than television-viewing (Harrison and Cantor, 2006). Studies of undergraduate women have associated reading fashion magazines with having higher preference for lower weight, having lower confidence on their own body image, feeling frustrated for this reason etc (Turner et al., 1997). The ‘ideal’ body image portrayed by the media influences social interaction and this may in turn make it more dominant. This circularity only makes the power of social interactions in shaping people’s self-identity more extreme. To sum up, females of similar age, education and background are likely to have been exposed to similar media and social environments and, accordingly, to have similar ideal self- identities. To measure the strength of such socially transmitted influences on individual behaviour it seems appropriate to use the concept of peer or social-multiplier effect, as 8  applied in Glaeser et al. (1996) and in Sacerdote (2000). This concept arises not only when women have similar behaviour or representations (self-identity) due to sharing a common environment, but also when they belong to certain unobservable social groups (see Manski, 1993). 3. An economic decision model for eating disorders Current empirical evidence makes modelling eating disorders difficult, as one of the assumptions of consumer-choice theory is the principle of non-satiation. According to extant literature, food seems to need to be modelled as an economic ‘good’ up to a certain caloric intake – which is idiosyncratic due to socially influenced self-perception – and as an economic ‘bad’ thereafter. In order to model anorexia, the self-identity model of Akerlof and Kranton (2000) was found to be particularly useful and was adapted to the subject of interest. It was assumed that individuals choose food and exercise-related ‘actions’ in order to maximize an implicit utility function that depends not only on their net caloric intake (food consumption minus what is consumed by exercise), but also on their self-image (or self-identity) and health. Besides these individual factors, the utility function of individuals is conditioned by their peers’ net caloric intake - and also their appearance and their characteristics - and by socio-cultural environmental factors. Thus, the utility function can be modelled as: )1( ) , ; , , , , ( j j j j j j j j j Z z H SI c a a U U   where aj is j’s net caloric intake; a-j is the appearance of the j’s group of reference; cj reflects j’s other actions – not related to caloric intake; SIj is j’s self-image; Hj is j’s health-production function; zj are j’s characteristics; and Zj the environmental factors in which j is immersed. It is assumed that utility depends on the rather abridged concept of ‘net caloric intake’ because food and exercise are a source of satisfaction beyond the body weight they achieve. Similarly to Akerlof and Kranton (2000), self-image SIj depends not only on j’s net caloric intake, aj, but also on others’ body-weight-related actions or appearance, a-j; and is conditioned by j’s individual characteristics and environmental factors, zj and Zj; and by j’s 9  status’, sj - as a person with higher status may have a better self-image than an identical one with lower status.5 Thus, the equation for self-image6 is written as: ) 2 ( ) , , ; , ( j j j j j j j Z z s a a I SI   Finally, a health-production function Hj is added. This depends on j’s net caloric intake, aj; all j’s other actions, cj; j’s status’, sj; and any other individual and environmental factors, zj and Zj . The health-production equation is written as follows: )3 ( ) , ; , ( , j j j j j j j Z z s c a H H  Standard utility maximization subject to a budget constraint under the usual regularity assumptions would lead to an associated first-order condition as follows:  ) 4 ( 0                a h j j j j si j j j j u j j j j P a H H U a SI SI U a U da dU        where � is the usual income-multiplier and Pa the monetary price of net caloric intake or the combination of food price and exercise monetary cost including the opportunity cost of the time invested in it. Equation (4) can be rearranged as follows:    ) 4 ( 0 b P a SI SI U a H H U a U da dU a j j j j j j j j j j j j                                                  5 Here status can be interpreted as loosely reflecting not only social status but also physical appearance, and other status-determining attributes. 6 Akerlof and Kranton (2000) also include j’s ideal identity and the prescribed norms associated to j’s status but, to avoid unnecessary modelling complications only peer image is used, given the individual’s status, characteristics and socio-cultural environmental factors, to capture what is socially normal for j in his/her environment. 10  Even if very simplistically, equation (4b) reflects the fact that the net-caloric-intake-related choices (food and exercise) of a person with eating disorders will in principle take into account the positive effect that net-caloric-intake has on individual utility and on health, but also the effect that it has on utility and health through its impact on self-image. A person without any eating disorder and in a normal range of net-caloric-intake would be expected to receive a positive marginal utility from net-caloric-intake, from health and also from an improved self-image. Also, it is assumed that a normal net-caloric-intake has a positive marginal impact on health, since nutrition is necessary for survival. Thus, the first two summands in equation (4b) are expected to be positive. In contrast, in general, one can expect a possibly negative marginal impact of net caloric intake on self-image after a certain level of net-caloric-intake, which would make the sign of the second term in equation (4b) negative. The net-caloric-intake chosen to optimise overall utility will vary depending on the relative magnitude of the positive and negative signs in equation (4b) above, bearing in mind that both anorexic and non-anorexic women will eventually confront the economic principle of non-satiation. The difference lies in satiation among anorexic women taking place at lower levels of consumption7. In other words, the ‘bliss point’ of food consumption for anorexics is lower, because the negative effect of eating on self-image is greater for them. Given the empirical evidence, a person with anorexia will have an extraordinarily large negative term associated with the effect of net-caloric-intake on self-image. In this special case the utility of net-caloric-intake would achieve a maximum at a much lower level than for a non-anorexic person (see Figure 1). Note that the sign of the self-image term is idiosyncratic insofar as it depends on the impact on each individual of the societal ideal- body-shape that is in fashion. 7 Running for 30 minutes is fun, running for 10 hours is not. A house heated to 22 degrees is pleasant, one heated to 42 degrees not so. Equally, the marginal utility of eating and drinking is obviously negative after some point. 11  Figure 1. Optimal equilibrium with and without anorexia Thus, an anorexic individual chooses a net-caloric-intake aj that is under the healthy/optimal net-caloric-intake âj associated with his/her characteristics had that individual not been anorexic. This minimum-necessary net-caloric-intake threshold can be thought of as the one that would keep individual j on a body mass index (BMI) considered ‘healthy’. Utility Net caloric intake âj aj From equation (4b), it is easy to infer an implicit reduced form of net-caloric-intake that depends on individual status, individual characteristics and the social environment, which includes the appearance/net-caloric-intake of others. In particular, under standard normality and linearity assumptions, the likelihood of being anorexic, e.g. the probability that the net- caloric-intake of an individual j is below his/her minimal healthy level âj can be expressed as: P(a j  ˆ a j)  (s j,z j,Z j,a j) (5) The next section describes how equation (5) is taken to the data. 4. Data and Methods Data Two types of variables are used: individual-level variables and socio-environmental variables. The former are taken directly from the answers to the Eurobarometer 59.0 questionnaire, study number 3903. Eurobarometer 59.0 is one of the Eurobarometer Surveys that have been conducted each spring and autumn since autumn 1973, adding countries as the European Union has expanded. The usual sample in standard Eurobarometer Surveys is 1,000 12  people per country, with the exception of Luxembourg (600) and the United Kingdom (1,000 in Great Britain and 300 in Northern Ireland). Also, since Eurobarometer 34, 2,000 people have been sampled in Germany (1,000 in East Germany and 1,000 in West Germany) in order to monitor the integration of the five new Länder into unified Germany and the European Union. In each of the 15 member states, the survey is carried out by national institutes associated with the European Opinion Research Group.8 A special issue, Eurobarometer 59.0, was carried out in all European Union countries between 15th January and 19th February 2003 on behalf of the European Opinion Research Group. The questions from this Special Eurobarometer centred around attitudes towards life-long learning, health issues, dietary habits and alcohol consumption, safety issues, partnership, household tasks, childcare and family planning. It focussed particularly on the incidence of chronic illness, on long-term treatment, on dental health and, in more depth, on health maintenance (by discussing doctor's visits and various screening tests), on women's health and medical tests relating specifically to women's health, and on general and children’s safety.9 Given that the mechanisms that give rise to anorexia and bulimia particularly affect women (Hill, 1993) this paper focuses on women’s behaviour and thus only evidence on women was selected. This gave a sample of 8,740 valid observations on women above 15 years of age. The paper scrutinises a set of individual variables ranging from socio-demographic characteristics to biometric measures and behavioural attitudes. This set of variables includes: (self-reported) weight, height, own-body perception, healthiness of eating habits, age, gender, being married, educational level, professional category, political attitudes, and residence in an urban or rural area. Furthermore, to reflect the freedom and quality of the answers, in some of the specifications the number of people present during the interview and the level of cooperation is included. Women are categorised as anorexic if they are extremely thin but at the same time perceive themselves as being ‘just fine’ or ‘too fat’.10 For that purpose, an indicator variable called 8 From Standard Eurobarometer 59 / Spring 2003 - European Opinion Research Group EEIG: http://ec.europa.eu/public_opinion/archives/eb/eb59/eb59_rapport_final_en.pdf 9 Special Barometer: Health, Food and Alcohol and Safety. Special Eurobarometer 186 / Wave 59.0 - European Opinion Research Group EEIG: http://ec.europa.eu/public_opinion/archives/ebs/ebs_186_en.pdf 10 DSM-IV, the American Psychiatric Association’s manual classifying mental illness, categorises as anorexic a woman who satisfies the following four criteria: 1. She refuses to maintain body weight above a minimal threshold adjusted for age and height. 2. She has an intense fear of gaining weight. 3. She suffers from an undue 13  ‘anorexia’ was created, which took a value of 1 if a woman had a BMI of less than 17.511 and at the same time saw herself as being ‘fine’ or ‘too fat’. A second anorexia indicator variable labelled as ‘severe anorexia’ was created if the individual also declared herself to be eating adequately. Finally, to identify seeing oneself in the ‘right weight range or above it’, a variable called ‘normal or too fat’ was created, which took a value of 1 if the individual declared she saw herself as normal or too fat, and of 0 otherwise. A variable was also created that measured health consciousness through the number of declared gynaecological check-ups received during the previous six months (0 to 6). Figure 2 reports the prevalence of the three variables examined in different age groups. Extreme thinness in terms of very low BMI was highest during early youth (age group 15-24) and progressively decreased until 55-64 years of age, increasing slowly again in the late years of life. Anorexia, as defined here, had a prevalence of 3% for women aged between 15 and 24, just slightly higher than severe anorexia. Both conditions followed a decreasing pattern till the age of 35, after which they remained relatively constant at about 1%. The paper found that the prevalence of anorexia is just below 4% for younger age groups and just below 2% among women aged 25-34. Therefore, women below 34 are expected to present a different pattern from women over 34. influence of body shape on self-esteem or denial of the seriousness of current low body weight. 4. She suffers from amenorrhoea. The Eurobarometer data do not provide information on criteria number 4. However, the definition of anorexia used here covers criteria 1 by including those who are extremely thin, and criteria 2 and 3, by including those who, besides being extremely thin, have a distorted perception of their own body appearance. 11 Since height and weight are self-reported, the correction suggested by Connor Gober et al. (2008) was introduced. 14  Figure 2: Prevalence of extreme thinness and anorexia among different age groups 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 15-24 25-34 35-44 45-54 55-64 65+ Extreme Thinness Anorexia Severe Anorexia The first panel in Table 1 provides some overall statistics for the data. The average age of the women in the sample was 45. Of these, 57% were married, 37.5% were heads of household and 27% lived in a small town or rural area. Roughly 26% had completed primary education, 41% had finished compulsory secondary education, 24% had studied up to 18 years of age, and 9.4% held a university degree. The average value of the variable ‘being health conscious’ for the full sample was 1.25, indicating the average number of gynaecological screenings received over the previous 6 months. The second panel in Table 1 shows the descriptive statistics for young women, who had a higher incidence of food disorders. This group included the women between 15 and 34 years of age, with an average age of 25 years. Only 47% were married, 28% were heads of household and 30% lived in a small town or rural area. Eight per cent had completed primary education, 41% compulsory secondary education, 23% had studied until the age of 18, and 27% held a university degree. For this younger group, the average number of gynaecological check-ups during the last half year was 1. 15