VuLogoPsychology Department

Health Psychology Home Page

Papers written by students providing scientific reviews of topics related to health and well being

  HomeWeight LossAlternative Therapy | Supplements | Eating Disorders | Fitness | About this Page |


Socioeconomic Status and Prevalence of Eating Disorders

Saanyol Se-ember Suswam

6 May 2010


            With the increased exposure of eating disorders (ED) as an ever more deadly sickness in our society a new concentration on the causes of these unsolved disorders has also greatly increased. In the investigations to find the root causes, biological, social, environmental, genetic, and psychological factors have been considered. Studies have revealed that perfectionist personalities along with unhealthy family views of body image and eating practices and several other characteristics are traits of those who are diagnosed with eating disorders. More specific qualities such as gender, race, age, and geographical location of patients have also been targeted in order to fully solve the mystery that is an eating disorder, namely anorexia and bulimia nervosa. In this study, however one other quality, the socioeconomic status (SES) of an individual suffering from an ED is going to be analyzed as a factor in the likelihood and development of eating disorders. The underlying question is: does socioeconomic status have any relationship in the prevalence of an ED in an individual?

Background of Eating Disorders

Although there are many disorders that could be classified as dealing with “eating” or “strange eating habits” for the purposes of this paper eating disorders will include anorexia nervosa, bulimia nervosa, and a brief mention of binge eating disorder later on as well. In this paper anorexia nervosa will be defined in an overall scope as “an eating disorder characterized by refusal to maintain a healthy body weight, and an obsessive fear of gaining weight due to a distorted self image.” Bulimia nervosa will be categorized as “an eating disorder characterized by recurrent binge eating, followed by compensatory behaviors. The most common form is defensive vomiting, sometimes called purging; fasting, the use of laxatives, enemas, diuretics, and over exercising are also common”, both as defined by the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision.


            For several years there has been a notion surrounding those diagnosed with eating disorders and the diseases themselves. This notion entails that eating disorders are somehow linked with high socioeconomic status.  Numerous studies conducted before 1970 proved that this stereotype was true; however, many other studies have been released since then using more critical approaches and have concluded that the new data does not support this previous claim. In fact, in supporting evidence studies have shown the opposite is true of bulimia- that the prevalence of bulimia raises in instances of lower socioeconomic status.  The “myth” seems to have been brought about by an overrepresentation of patients from higher classes, explained by the fact that these patients have more disposable income to be put toward the more expensive treatment and counseling of eating disorders (Yager, 1996). In this paper, our job is to find out why the conclusions of these early studies have been rendered invalid and how the current studies have proven to be reliable.

Review of Past Studies

            The relationship between socioeconomic status and eating disorders has been questioned in the past. However, many researchers say that while this discrepancy was rightfully questioned, there were many things incorrectly done in the past studies, that they feel had led to incorrect conclusions, specifically that, “studies conducted in the 1960s and 1970s…may have biased the result.” Another reason that researchers feel that more inquiry should be taken into the matter of this relationship is that “the methods emplyoed by from the studies are suspect, and in some cases, based largely on speculation as opposed to empirical evidence.” The speculation that Gibbons speaks of here refelcts the popular, but false notions about class. Specifically he is addressing  sayings like “the cup of the wealthy over floweth, while the poor hope for scraps” leaving little room in the mind of the common man to belive that a class in so much struggle for food would ever adopt such eating habits as what they believe anorexia and bulimia inclulde (Gibbons, 2001). A number of other factors could also be involved of the perpetuation of this incorrect belief including : “the powerful influence of clinical impression, sources of bias in referral patterns, problems in the methodology of existing research, and the failure to adequately separate bulimia nervosa from anorexia nervosa when discussing common predisposing factors”  (Gard, 1996). In this paper we will critically evaluate the methodology of the past studies and the soundness of the conclusions, in order to deem the results of these studies, that socioeconomic status was indeed correlated to likelihood of an eating disorder, valid or invalid. The only way that this can be done is by doing the same to the current studies: to evalutate their methodology and the soundness of those conclusions. By doing this we can finally gather enough relevant and meaningful information to answer the issue of socioeconomic status and eating disorders.


Past Studies


            According to Gard, “the following studies illustrate some of the mehtodological problems present in the early studies which demonstrate an increased prevalence of eating disorders in high socioeconomic groups” (Gard, 1996). In the first study mentioned, that took place in 1973,


            “Kendell et al. summaried demographic data on anorexia nervosa from retrospective reviews of the psychiatric case of three different areas: North-East Scotland, Camberwell in London, and Monroe County, New York State. The authors concluded that the findings dfsof their review provided substantial eveidence for a high socioeconomic status bias in anorexia nervosa. However, the number of cases identified in each area was small, and the one register reporting a significant relationship between socioeconomic status and  anorexia  nervosa was based on only 8 patients with the anorexia nervosa. The valididty of the authors’ conclusion was made more questionable by their assumption that ‘nearly all anorexics living in a register catchment area are sooner or later seen by a psychiatrist and reported to the register.’ No attempt was made to justify this relation at the time, and more recent studies indicate that it is ill founded” (Gard, 1996). 


In this study the results are clearly invalid because of two methodological flaws: firstly the sample size is not nearly large enough, and secondly the incorrect assumption that all anorexics living in that area would eventually be treated. This fallacy goes along with what we mentioned earlier that not everyone who has anorexia nervosa can afford treatment or is willing to be treated. In another study in 1976,


             “ Crisp et al. published the findiings of their study of the prevalence of anorexia nervosa in nine populations of shcoolgirls between 1972 and 1974. The authors conclude that severe anorexia in girls aged over 16 was found five and a half times more frequently in private schools compared to state schools and that this discrepancy almost certainly affected a socioeconmic status factor for the disorder. However, the methodological probmems of this sudy render the seminal conclusions of its authors highly questionable: Firstly, the authors report tha they surveyed the schoolgirls whereas in fact no survey took place. Instead, all cases were ‘case-spotted’ throught the authors collaboration with the teachers at the schools. One way in which bias could be introduced into this method of case detection is apparent from the authors' own observation that ‘it was immediately clear that all the independent schools were very familiar with the problem.’ It is possible that given the greater awareness of anorexia nervosa on the part of staff at private schools, that staff at these schools might attribute a psychological origin to thinness in a pupil whilst thinness in state schools might be attributed to deprivation by their less informed state school colleagues. The fact that four of the private schools included in the study were selected because there had been one or more referrals to the authors' clinic from those schools, whilst no referrals had come from either of the participating state schools, appears to confirm this suspected bias. Finally, seven of the nine participating schools were privately run whilst only two were state run. Thus, the number of pupils ‘surveyed’ in the private schools was four and a half time greater than those "surveyed" in the state schools” (Gard, 1996).


Likewise there are two other case series reports, one by Morgan and Russell (1975) and the other by Crisp et al. (1980) that although they are known as providing substantial evidence that there is in fact a high socioeconomic status bias in eating disorders, their validity has been shed many years later. In these particular studies, only the most severe cases are included, the subjects were all patients with long histories of illness and resistant to treatment, and at the time of their assessments, many were tertiary referrals. To further demonstrate the uncertainty of these two studies the authors questioned the validity of their own trials stating that, “it is uncertain, therefore whether higher social class is associated with increased casual factors or merely reflects differences in attitude to feeding disorders and a willingness to seek psychiatric help.” This quotation embodies the uncertainty of all of the early investigations surrounding the relation between eating disorders and high socioeconomic status. Furthermore in all of the studies we covered and in others many problems occurred rendering the conclusions invalid: the samples were too small and not population representative, they were already biased before trials began, and in some extreme cases the results were rendered inconclusive because the factor of socioeconomic status was not tested for and falsely published as the highlighted factor as is highlighted by Halmi (1974) (Gard, 996).


 Review of Recent Studies


            In more recent studies regarding eating disorders and socioeconomic the go more along the lines of reflecting notions like: “In the United States two epidemiological studies conducted in the mid-1980s, examining the prevalence of bulimia nervosa in high school students, found no significant socioeconomic status differences between bulimic and non-bulimic students”  (Gard, 1996). However, in order to prove, or disprove this new theory we need to examine the methodology of them and the soundness of their conclusions, just as we did for the early studies. If their methodology is just as botched as the early studies a inclusive conclusion will have to be drawn up.


Recent Studies


            In a study done by Lisa Rogers of the department of Psychiatry at the University of Minnesota Hospital and Clinic, Michael D. Resnick and Robert W. Blum of the Division of General Pediatrics and Adolescent Health at the University of Minnesota, and James E. Mitchell of the Department of Neuroscience at the University of North Dakota an attempt was made to examine the relationship between SES and eating disorder in a large community of adolescent girls, grades 7 to 12. The method of this study entailed: “Items were selected from a comprehensive self-report health survey completed by 17,571 adolescent girls. The relationship between SES and dieting behaviors and attitudes were examined using chi-square tests, ‘a statistical test commonly used to compare observed data with data we would expect to obtain according to a specific hypothesis’ (Fisher, 1963)  and analysis of variance (ANOVA) to control for body mass index.” In their results they concluded that “although there was some a significant positive relationship between SES and some of the unhealthy dieting behaviors, there was no relationship between self-report of clinically significant eating disordered behaviors (e.g., vomiting twice a week or more) and SES in this community sample” (Rogers, 1997).


            In another study done by Dr. Patricia Reagan and Dr. Joni Hersch conducted in 1995, they used an analysis of self reported data on 573 womien and 360 men ages 18-97 “from a cross-sectional, multistage area probability sample of individuals aged 18 yars and older residening in the Detrioit metropoliatan area.” In this study the results were, “for women the frequency of binge eating [was] negatively associated with age and family income, and positively associated with being married, depression, and time residing in polluted neighborhoods. For men, the frequency of binge eating is negatively associated with age. The frequency of binge eating was not affected by education, race, obesity, or current dieting.” Consequently Reagan and Hersch concluded that, “the frequency of binge eating is highest among adults younger than 40 years. It follows an income gradient for women, but not for men” (Reagan, 2005).  


            In an additional study done by Zaimin Wang and Nuala M. Byrne of the Centre for Health Research in the School of Human Movement Studies of Queensland University of Technology, and Justin A. Kenardy and Andrew P. Hills of the School of Psychology at eh University of Queensland in Brisbane Australia, the researchers aimed to examine the association between socioeconomic status, ethnicity, body dissatisfaction, and eating behaviors of 10- to 18-year old children and adolescents. The 768 participants were categorized by race as either Caucasian (74.75), Chinese or Vietnamese (18.2%), and Italian or Greek (7.0%). They were also differentiated based on SES: high (82%), middle (8.6%), and low (9.4%) according to parents’ occupations. The x2 was used to determine the interaction between variables. The testing arrived at these results: females and older participants were more likely to desire a body figure that was perceived as thinner than their perceived current figure. Additionally, the same groups were also more likely to be more preoccupied with eating problems. However, participants whose parents were professionals or in manager positions were more likely to desire a body figure that was thinner than their perceived body figure than those from white collar and blue-collar families (Wang, 2004).

            In a final study by Jennifer A. O’Dea and Peter Caputi, the aim was “to examine the effect of socioeconomic status, weight, and gender on the body image and weight control practices of children and adolescents, and to investigate whether health education about weight issues should target low socioeconomic groups. In this study the participants were a randomly selected group of school children who completed a questionnaire, and had their height and weight measured. The 1131 participants were aged 6-19 years from 12 schools in New South Wales. SES, age, gender, body weight, body image, skipping breakfast, physical self-esteem, attempts to lose or gain weight, and dietary control advice received from others were examined.” The results were this: “low SES children were more likely to be overweight, to skip breakfast, to perceive themselves as ‘too thin’ to be trying to gain weight and less likely to receive dietary or weight control advice. Physical self-esteem was lowest among overweight girls of middle/upper SES, despite the latter being more likely to be overweight. Being overweight does not appear to adversely affect the physical self-esteem of children of low SES, particularly boys.” The researchers also felt that, “Health educators should examine these issues with young people to help make health education and nutrition education most relevant and appropriate”

Analysis of Recent Studies


            In the first study, Rogers’ study, the researchers felt that there may be a significant relationship between SES and dieting or other behaviors associated with ED, however among the young women involved in their study who met psychiatric criteria to acutally be diagnosed with an eating disorder, SES does not apperar to be a significant factor. This study had a neccesary large sample group and tested for the specifc relationship of SES and ED both good insurances of a valid testing. However because the whole analysis was based on self-surveys of only adolescent girls there could be room for error in the reports of the girls. They could be incaccurately responding to the questions causing the data to be skewed. Secondly, the sample did not include any males. This is a great potential for invalidity and for very basic reasons, males of high SES may contribute to the small population of men who do indeed have ED. However it is understandable that since the male population of diagnosed anorexicts and bulimics is so minimal Rogers and her co-researchers decided to deem them negligible.

            In the second study, the sample pool was again large and even included a broader age span of participants and in a broader reigon. These three factors allowed for the study to be very indicative of several types of binge eating behaviors, a different type of ED which could be very a very insightful resource to the issue at hand and also has the diversifying affect in relation to simply anorexia nervosa and bulimia nervosa,  of different types of groups. This factor is good, however non essential data might have diverged attention from crucial data with the inclusion of a bit too many factors in the study. For example one of the conclusions included the data that the frecuency of binge eating is highest in adults under 40 years. Although true, this conclusion is already a widely know evaluation of people with eating disorders in general and seems a bit non-essential in this new data.

            In the third study, the sample was again large, much larger than the ones in the previous early studies conducted in the 60s and 70s. However the main focus of this study included education goals and other such means. This factor could have swayed the collection of sample data in favor of this goal and left out some crucial data. Furthermore, although the study’s results coincided with the popular belief of current studies that SES has no relationship, either positive or negative, with ED, there were so many other factors involved in the study, that this one focus on SES could have been clouded. Additionally, the sample was taken in a predominantly higher class society, leaving room for possible doubt of the diversity of the study. Finally one piece of data collected, the “advice received from others”, was very subjective and also not very represenative of areas outside of the regions where the data was collected and like the first study could have made the results skewed based on possible incorrect interpretations by the participants.

            Finally in the last study by O’Dea the results suggested that SES was correlated to ED, in the way that children of high SES were more likely to have body image problems and desire a body image that was thinner than their current percieved body image. However, these problems are only eating disorderd behaviors as defined by Rogers in the first study and do not acutally mean that the adolescents and children with these habits had acual eating disorders such as  anorexia nervosa and/or bulimai nervosa. Likewise the SES of the children was based on subjective judgments of the parents’ careers instead of tangible tax records or something like that.




            After investigating past studies regarding the realtionship between SES and eating disorders and then comparing the results and evaluating the methodologies used with current studies on the same issue the intial conclusion is that eating disorders and SES have no significant relationship. In the early studies the samples were too small and not inclusive enough to deem any of the results they found valid. Furthermore, in the more recent studies with larger sample sizes, more inclusive and exact methods of obtaining information the conclusions seem to be a bit murkey. While most researchers are reporting that no relationship is present, there is still a minority of reserchers reoprting that disorderd eating habits are prevelant in children and adolescents of higher SES. However, in studies such as Rogers’s the non-relation between SES and eating disorders seems so apparent. Thus the conclusion is outrightly that the two have no relation. Continually, it is hard not to see the realtionship between high SES and eating disorders, however, this notion can be quickly dissolved even without reading these new studies or speaking with a physician- it would be done by simply realizing that people of higher SES are unproportionately represented in treatment facilities because they have the means to be there. It is not that people of lower SES do not also suffer from ED, it is just likely that they cannot afford to be treated or any other such reasons that keep them away from treatment and counseling facilites. For futher illustration a simple example can be made: just because people with less money buy less medication than poeople with lots of money, does not mean that people with more money are sicker than people with less. They simply have more disposable income to spend on the medicines. Additionally, it is still understandable to think that a relationship is there in either direction, positive or negatie, because of the high prevelance of disordered behaviors in adolescents and the lower prevelance of body image dissatisfaction in adolescents and children of low SES, according to O’Dea and others. However, it is critical to remember that these instances do not prove eating disorders, thus the notion that eating disorders are linked to socioeconomic status is false.




Fisher, R. A. (1963). Statistical Tables: For Biological, Agricultural and Medical Research . New York: A Divison of Macmillan Publishing Co., Inc. .

Gard, M. C. (1996). THe Dismantlinng of a Myth: A Review of Eating Disorders and Socioeconomic Status. International Journal of Eating Disorders , 20 (1), 1-12.

Gibbons, P. (2001). The Relationship Between Eating Disorders and Socioeconomic Status: It's Not What You Think. Nutrition Noteworthy , 4 (1), 1-5.

Reagan, P. a. (2005). Influence of Race, Gender, and Socioeconomic Status on Binge Eating Frequency in a Population-Based Sample. International Journal of Eating Disorders , 38 (3), 252-256.

Rogers, L. R. (1997). The Realtionship Between Socioeconomic Status and Eating Disorderd Behaviors in a Community Sample of Adolescent Girls. Minneapolis, MN: John Wiley & Sons Inc. .

Wang, Z. B. (2004). Influences of ethinicity and socioeconomic status on the body dissatisfaction and eating disorders of Australian children and adolescents. Eating Behaviors , 6, 23-33.

Yager, J. (1996). Debunking Myths about Socioeconomic Status and Eating Disorders. Journal Watch Psychiartry , 1.




Psychology Department

The Health Psychology Home Page is produced and maintained by David Schlundt, PhD.


VuLogoVanderbilt Homepage

Return to the Health Psychology Home Page
Send E-mail comments or questions to Dr. Schlundt