Weight bias as a function of person variables and contact experiences

ABSTRACT

We tested a mediation model of weight bias that considers person attributes and contact experiences with overweight individuals. In Study 1, we administered a survey to assess Openness, Agreeableness, Attributional Complexity, contact experiences with overweight individuals, and weight bias. Mediation analyses found that Agreeableness predicted less weight bias through contact experiences. In Study 2, we asked participants to interact with a peer whose weight and attributions regarding the weight were experimentally manipulated. We then measured acceptance of the peer. Agreeableness was found to indirectly predict more acceptance of an overweight peer through Empathy and contact experiences. These results show that contact theory is applicable to the domain of weight bias, and support person-situation approaches to prejudice.

KEYWORDS:

• Agreeableness
• contact theory
• openness
• prejudice
• weight bias

Funding

This research was supported by an IPFW summer research grant awarded to the first author.

Notes

1. Consistent with contact theory, this scale emphasizes the importance of quality contact experiences over the frequency of contact experiences. Two items in the scale correspond to frequency of contact while the other 13 items correspond to quality of contact experiences. If separated, the frequency and quality subscales are highly intercorrelated, r = .42, p < .001). Deleting either or both of the frequency items weakens the internal reliability of the scale (from .850 to .841, .845 or .839) and does not significantly change any of the results.

2. For exploratory purposes, we examined the interaction of openness and agreeableness on weight bias. The interaction was not significant (B = .530, SE = .616, p = .391).

3. In total, 10 participants were excluded because they failed one or more of these manipulation checks (three participants failed more than one). To check our gender manipulation, we asked participants, “Was your interaction partner a male or a female?” and gave them three response options: (1) male, (2) female, or (3) not sure. One male participant misidentified his partner as female, and two male participants were “not sure” about their partner’s gender. One female participant was “not sure.” To check our weight manipulation, we asked participants, “How would you describe your interaction partner’s weight?” They were give seven response options: (1) very underweight, (2) moderately underweight, (3) slightly underweight, (4) neither underweight nor overweight, (5) slightly overweight, (6), moderately overweight, and (7) very overweight. Two participants in the overweight condition misidentified their partners as underweight (marking 1, 2, or 3 on the scale), and one participant in the normal weight condition misidentified her partner as being overweight (marking 4, 5, or 6 on the scale). To check our attribution manipulation, we asked participants, “To what extent was your interaction partner’s weight due to personal behaviors and choices versus due to forces outside of his or her control?” They were give seven response options: (1) much more due to personal behaviors, (2) somewhat more due to personal behaviors, (3) slightly more due to personal behaviors, (4) equally due to personal behaviors and uncontrollable forces, (5) slightly more due to uncontrollable forces, (6), somewhat more due to uncontrollable forces, (7) much more due to uncontrollable forces. Four participants in the “uncontrollable” condition incorrectly responded to this item (with a 1, 2, or 3), and two participants in the “controllable” condition incorrectly responded (with a 5, 6, or 7). After removing the 10 participants who failed one or more of the manipulation checks, the mean response on the “describe your interaction partner’s weight” item was M = 3.84 (SD = .54) for participants in the normal weight condition and M = 5.44 (SD = 1.18), for those in the overweight condition, F (1, 190) = 154.924, p < .001. The mean response to the “personal behaviors vs outside forces” item was M = 2.59 (SD = 1.47) for those in the internal attribution condition, M = 5.31 (SD = 1.73) for those in the external attribution condition, and M = 3.40 (SD = 1.31) for those in the neutral (no attribution mentioned) condition, F (2, 189) = 52.34, p < .001. Tukey post-hoc tests determined that all comparisons of mean differences were significant (all ps < .01). These data indicate that our manipulations of weight and attributions for the weight were both successful.

4. The stimuli in the Appendix were developed using pilot tests, and were supported with results similar to those reported in Note 3.

5. After the research session was over, but before the full debriefing, the participants were asked to indicate what they thought the study was about, if they had any questions regarding the study, and if they found any aspect of the study to be unclear or odd. None of the participants indicated any significant or accurate suspicions regarding the deceptions involved. We attribute this in large part to extensive pilot-testing of our procedures.

Additional information

Funding

This research was supported by an IPFW summer research grant awarded to the first author.

Notes on contributors

Jay W. Jackson

Jay W. Jackson is Professor of Psychology and Director of the Group Processes and Intergroup Relations Lab, Department of Psychology, Indiana University – Purdue University, Fort Wayne (IPFW). Audrey James is affiliated with the Group Processes and Intergroup Relations Lab at IPFW. Joan Rose Poulsen is Associate Professor of Psychology and Director of the Undergraduate Program in Psychology, Division of Science, at Indiana University – Purdue University, Columbus. Jennifer Dumford is affiliated with the Group Processes and Intergroup Relations Lab at IPFW.