Exploring Optimistic Bias and the Integrative Model of Behavioral Prediction in the Context of a Campus Influenza Outbreak

Abstract

This study aims to determine the utility of constructs from the Integrative Model of Behavioral Prediction in predicting behaviors associated with flu prevention in a college campus. Building on previous studies that document substantial optimistic bias among college students, this study explores possible roles for optimistic bias in the Integrative Model of Behavioral Prediction. The authors conducted an online survey among a stratified random sample (N = 429) of undergraduate students in a university that experienced a large H1N1 influenza outbreak in the fall of 2009. The authors offer 3 major findings: (a) the 3 primary components in the Integrative Model of Behavioral Prediction predicted the intention to perform hand washing and sanitizing, with self-efficacy being the strongest predictor; (b) unrealistic optimists had significantly lower intentions to perform hand hygiene practices compared with realists or unrealistic pessimists; and (c) comparative optimism was a significant moderator of the relation between self-efficacy and behavioral intention, controlling for perceived risk. Theoretical and practical implications of these findings are discussed.

Acknowledgments

The authors are grateful to Jennifer Austin, Sharon Dittman, and Heather Stone with Gannett Health Services for their valuable input on relevant flu-prevention behaviors, communication strategies, and survey instrument development. They are also grateful to students enrolled in Planning Communication Campaigns for their efforts in identifying salient flu-prevention beliefs among the student population. The authors also thank each of the study participants for taking the time to complete the survey.

Notes

Note. BI = behavioral intention (4-point scale); OE = outcome expectancy beliefs (5-point scale); NB = normative beliefs (5-point scale); SE = self-efficacy beliefs (5-point scale); CO = comparative optimism (difference between risk perception of self and others, 7-point scale).

**p < .01.

Note. Because the latent mean analysis does not provide absolute mean scores (Cohen, 1988), relative mean scores were computed by setting one group's mean as 0; Cohen's d =latent mean difference divided by the same standard error shared in both groups; on the basis of Cohen's (1988) guideline, effect size less than .2 is small and above .8 is large. Independent samples t tests were performed to compare observed mean differences.

*Mean difference from unrealistic optimists statistically significant at p < .05.

**Mean difference from unrealistic optimists statistically significant at p < .01.

***Mean difference from unrealistic optimists statistically significant at p < .001.

Note. BI = behavioral intention (4-point scale); OE = outcome expectancy beliefs (5-point scale); NB = normative beliefs (5-point scale); SE = self-efficacy beliefs (5-point scale).

*p < .05. **p < .01. ***p < .001.