The identification-guilt paradox revisited – Evidence for a curvilinear relationship between group-level self-investment and group-based guilt

ABSTRACT

Empirical evidence on the relationship between strength of in-group identification and collective guilt is mixed. Some studies point to anegative relationship, other work has found apositive or no link. We assume that the relationship between identification (group-level self-investment and self-definition) and collective guilt is inverted U-shaped (quadratic). Furthermore, we expect that self-investment in the group should be more central for the experience of collective guilt than self-definition. The results of three studies (N_total = 304) support our assumptions. They provide evidence for aquadratic relationship between self-investment and collective guilt about an in-group’s misconduct across different contexts, indicating that respondents with medium levels (vs. high/low levels) of self-investment experienced more guilt. No quadratic effects were found for self-definition.

KEYWORDS:

• Social identity
• group identification
• collective guilt
• non-linear relationship
• self-investment

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Reminders of group transgressions may also evoke other collective emotions, such as collective anger. The literature that we discuss in our theoretical introduction concentrates on the relationship between identification and guilt. Consequently, our focus in the present work is on collective guilt, even though alternative emotions such as collective anger and shame have been discussed in the literature as well (Giguère, Lalonde, & Taylor, 2014; Hansen & Sassenberg, 2006).

2. We do not preclude complementary explanations of this paradox. For example, Roccas et al. (2006) presented data on how in-group attachment and in-group glorification can affect collective guilt in opposing directions, resulting in suppressor effects. However, we suggest that exploring non-linear relationships might still be a worthwhile endeavor, both from a theoretical perspective as well as from a methodological point of view (more complex relationships than linear models).

3. The questionnaires of Studies 1a-1c also included items of no direct relevance to the present paper (collective anger, defensive beliefs about the ingroup’s antisocial behavior, interdependent self-construal, personal self-worth). To streamline the paper, we do not report further on these measures.

4. Two outliers with studentized residuals exceeding |2| were excluded for this analysis. Keeping the outliers in the analysis changed the significance of the quadratic term to p = .072.

5. Two outliers with studentized residuals exceeding |2| were excluded for this analysis. Keeping the outliers in the analysis changed the significance of the quadratic term to p = .134.

6. One outlier with studentized residuals exceeding |2| was excluded for this analysis. Keeping the outlier in the analysis changed the significance of the quadratic term to p = .066.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.