Damned if you do, damned if you don’t: Effects of empathy and responsibility in Muslim leaders’ mediated responses to extremist attacks

ABSTRACT

Muslim American leaders are often called upon to publicly respond to violent attacks carried out by Muslim extremists. Yet it is unclear what types of responses are most likely to satisfy non-Muslim Americans and ultimately improve attitudes toward Muslim Americans as a group. In three experiments, we examined how expressions of empathy and responsibility within Muslim leaders’ mediated responses to extremism affect response satisfaction. We did so immediately after real terrorism incidents carried out internationally (Study 1) and domestically (Study 2), as well as within a controlled fictional incident (Study 3). Across studies, expressing (vs. not expressing) empathy decreased the perception that the issued statement was motivated by external pressure, which was associated with (a) increased response satisfaction and trust in Muslim Americans and (b) decreased perceptions that Muslim Americans were collectively responsible for the incident. In contrast, accepting (vs. denying) responsibility increased the perception that the response was issued out of a sense of collective guilt, which, in Study 3, led to (a) decreased response satisfaction and trust in Muslims and (b) increased Muslim collective responsibility. These findings illustrate the perilous task facing group leaders who use the media to publicly respond to actions of extremist ingroup members.

Supplementary material

Supplemental data for this article can be accessed here.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. It is important to note that empathy and sympathy are terms that have been used by some scholars to describe the same construct. Although our goal is to empirically examine the relationship between expressions of empathy and perceptions of sympathetic motivations, our intent is not to imply that empathy and sympathy are easily distinguishable conceptually.

2. Power analyses (α = .05, f² = .2) found achieved power levels (1- β) of .99–.95 across all three reported studies.

3. An exploratory factor analysis of the motivation items supported a two-factor solution. Items pertaining to genuine sympathy loaded on the first factor (loadings > .9) and items pertaining to external pressure loaded on the second factor (loadings > .7). The collective guilt item did not load on to any factors and was thus analyzed independently.

4. Across all studies, participants in each condition did not significantly differ in age, gender, political identification, or identification as an American (all ps < .1). See Supplemental Appendix A for additional details.

5. Given the high intercorrelations, we first performed a MANOVA, testing the effects of empathy and responsibility on all outcome variables. There was a significant effect of both empathy and responsibility Fs(4, 466) = 2.49; 6.40, ps = .04; .0001, Wilks’ Λ = 0.98; 0.95, indicating univariate follow-up analyses were appropriate. The same analysis was done in subsequent studies justifying the use of univariate follow-up analyses in all three studies. See Supplemental Appendix A for additional details.

6. No significant interactive effects were found in the ANOVA models so interaction terms were not included in tests of mediation for the purposes of model simplification.

7. All mediation analyses were reconducted in a path analysis framework, which yielded similar results. Mediation analyses are reported here for simplicity of presentation.

8. Indirect effects are considered statistically significant if their 95% CI do not cross zero.

9. See Table A1 in the Supplemental Appendix A for full demographic characteristics.

10. Although there are no direct effects of experimental manipulation on response satisfaction and trust in Muslims, Hayes (2013) argued that tests of mediation are still warranted in the absence of direct effects. The same analytic procedures for mediation analyses from Study 1 were used in Study 2.

11. Analytic procedures for mediation analyses from previous studies were used.

12. We note that attempts were made across all three studies to include different forms of control conditions. In a pure control condition that omitted both responsibility and empathy, the actual quoted text from the Muslim leader was omitted entirely. Unfortunately, this made measures related to the Muslim leader’s response uninterpretable. For example, asking participants if the response was motivated by external pressure was meaningless, because they had not read any actual response. Ultimately, we decided to exclude this condition prior to our analyses. We tried to implement a no responsibility condition in Studies 2 and 3, by removing any language relating to responsibility. Unfortunately, in doing so, we created conditions in which the quote from the Muslim leader was substantially shorter than in other conditions. In addition, in Study 2 there were sampling errors in this condition. Given these shortcomings, we also excluded these conditions prior to our analyses. As we have noted, future research should more properly employ control conditions.