Abstract
In Study 1, participants completed five extant shame and guilt proneness inventories based on different theoretical conceptions of the difference between shame and guilt. Factor analyses revealed that despite very different theoretical distinctions, the shame proneness subscales loaded on one factor, and the guilt proneness subscales loaded on one factor. In Study 2, we altered scale items so that hypothetical transgressions were committed in either public or private, and likelihood response options were either typical of a “shame-prone response” (negative self-evaluation; avoidance behavior) or a “guilt-prone response” (negative behavior evaluation; approach behavior). Our findings indicate that shame and guilt proneness can be measured both by responses to transgressions (e.g., negative self-evaluation and avoidance responses vs. negative behavior evaluation and approach responses) and the situational context in which the transgression occurs (e.g., public vs. private). We provide recommendations regarding optimal measurement of shame and guilt proneness.
Study 1 is based on the first author's dissertation. Scott T. Wolf is now at the Department of Psychological Sciences, University of Missouri.
The authors thank Melanie Beck, Kate Conover, and David Watkins for their help with data collection.
Notes
1 Separate factor analyses for male and female participants yielded similar results.
2 Similar results were found using factor scores.
3 In addition to the results presented in Study 2 (which used the actor's perspective), we collected data from an additional 93 participants using an observer's perspective (e.g., “Sam cheats on an exam but does not get caught. What is the likelihood Sam would avoid the instructor?”). These results were very similar to those using the actor's perspective. The main difference between the actor and observer results was that the measures using the actor's perspective (as presented in Study 2) were more highly correlated with the theoretically-relevant measures. This difference suggests that an optimal measure of individual differences in guilt and shame proneness should have participants indicate how they themselves would respond following a transgression as opposed to indicating how another person, or the average person, would respond. More information about the observer data can be obtained by contacting the authors.
4 For ease of presentation, response metrics were altered slightly from the original measures, but the items themselves were retained unchanged. The response options for scales with similar metrics were equated so items from multiple scales could be randomly interspersed with each other.
We caution against making too much of the Lewis-type Shame–Lewis-typeGuilt × Public–Private interaction given the significant Self-evaluation–Avoidance × Public–Private interaction and the significant Behavior evaluation–Approach ×Public–Private interaction. As noted above, the significant main effects for the Lewis-type shame and guilt contrasts along with the significant public–private moderation of the Lewis-type shame and guilt contrasts indicate that combining the constituent parts of Lewis-type shame and Lewis-type guilt is problematic.