The Stigmatizing Effect of Visual Media Portrayals of Obese Persons on Public Attitudes: Does Race or Gender Matter?

Abstract

Overweight and obese persons are frequently stigmatized in news media. The present study is the first to systematically compare public reactions to positive and negative images of obese persons accompanying news reports on obesity (while manipulating gender and race of the target)and their effects on generalized attitudes and social distance toward obese persons. The authors conducted 3 randomized experimental studies using online surveys to assess public perceptions of positive versus stereotypical images of obese adults (who varied by gender and race) accompanying a neutral news report about obesity. The sample included 1,251 adults, who were recruited through a national survey panel during May of 2010. Participants who viewed negative, stereotypical images of obese targets increased social distance, antifat attitudes, and ratings of laziness and dislike toward obese persons, whereas positive, nonstereotypical images induced more positive attitudes. These findings remained consistent when accounting for sociodemographic variables. African American female obese targets portrayed in images evoked higher ratings of dislike and social distance compared with Caucasian targets, but ratings were similar for male and female targets. This study provides evidence that images of obese person accompanying written media influence public attitudes toward obese people, and may reinforce weight stigmatization if images contain stereotypical portrayals of obese persons. Implications for efforts to report about obesity in the news media are discussed.

Notes

¹This image gallery can be accessed at http://www.yaleruddcenter.org/press/image_gallery_intro.aspx =.

Note. Pearson χ² tests; braces indicate across which categories a test was performed. Original responses were measured with a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

*p < .05.

Note. Pearson χ² tests; braces indicate across which categories a test was performed. Original responses were measured with a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

*p < .05.

Note. Dependent variables: z scores (estimation sample: M = 0, SD = 1). Coefficients are from linear regression models. Marginal means are adjusted means as predicted by the regression models.

*p < .05. **p < .01. ***p < .001.