Self-affirmation and responses to health messages: a meta-analysis on intentions and behavior

Abstract

Objective: The present study aimed to quantify the magnitude of the effect of self-affirmation manipulations on health messages' influence on both intentions and behavior.

Methods: A systematic search was conducted for experimental studies comparing participants who self-affirmed prior to reading a threatening health message to those who did not self-affirm. Effect sizes for health intentions and behaviors were aggregated using a random-effects model.

Results: Data from 16 studies were included. The aggregate effect sizes for intentions and behavior were significant and small in magnitude (d+ = .26, 95% confidence interval [CI] = .04-.48; d+ = .27, 95% CI = .11-.43, respectively). A meta-regression analysis revealed that among studies assessing both outcomes, the size of an effect on intentions did not predict the size of an effect on behavior, β = .03, 95% CI = -.30-.36. Type of health behavior (damaging vs. promoting), timing of the health behavior (proximal vs. distal), type of self-affirmation manipulation (values vs. kindness), and the specificity of the health message (single vs. multiple health issues) did not moderate the effect of self-affirmation on intentions or behavior.

Conclusions: Self-affirmation influences health messages' effect on intentions and behavior; however, with the present study finding that intention effect sizes did not predict behavior effect sizes, and with past studies of heath behavior change finding that intentions do not always translate to behavior, little research supports a causal intention-behavior relation among self-affirmation studies. Future research is needed to address which specific health-related responses explain why self-affirmation elicits health behavior change.