Understanding self‐affirmation effects: The moderating role of self‐esteem

OBJECTIVES: Whether people benefit from self-affirmation in overcoming resistance to health messages may depend on their level of self-esteem. However, extant theories offer rival hypotheses about the role of self-esteem. We pitted these hypotheses against each other, building on the argument that how self-esteem shapes self-affirmation effects may depend on how and when people self-affirm. DESIGN: A 3 (Self-affirmation type: value essay, attribute scale, control task) by 2 (Timing: pre-message, post-message) plus 1 (message-only) experimental, between-subjects design. METHODS: Participants (N = 422) completed a value essay, an attribute scale, a control task, or no task, either before or after reading a message about health risks of excessive drinking. Participants then completed defensiveness and message acceptance measures. RESULTS: Self-esteem moderated self-affirmation effects across different types and timing of self-affirmation. While participants with relatively high self-esteem consistently benefited from self-affirmation, participants with relatively low self-esteem experienced adverse effects from self-affirmation. Self-affirmed participants with high self-esteem showed lower psychological discomfort, lower personal risk discounting, and higher message acceptance than unaffirmed participants, while self-affirmed participants with low self-esteem showed higher psychological discomfort, higher personal risk discounting, and lower message acceptance than unaffirmed participants. CONCLUSIONS: These findings support the proposition that self-esteem serves as affirmational resources, which implies that self-affirmation interventions to reduce defensiveness to health messages may be beneficial for people with high self-esteem but less so for people with lower self-esteem. Similar to much other self-esteem research, self-esteem was negatively skewed, which warrants caution in generalization of the findings across all levels of self-esteem.