Anger mediates the effect of ostracism on risk-taking

Abstract

The present research tested the effects of being ostracized on risk-taking behavior and how emotional responses to ostracism mediate these effects. In two experiments, undergraduates were either ostracized or included during an Internet ball-toss game (Cyberball). In Experiment 1 (N = 52), ostracism increased self-reported risk-taking inclinations, and the relation between ostracism and risk-taking was mediated by feelings of anger, but not control. In Experiment 2 (N = 72), ostracism increased risky driving during a driving video game. The effect of ostracism on risk-taking was again mediated by post-ostracism anger (but not control or ostracism-induced sadness). This research suggests that anger elicited by ostracism may increase risk-taking.

Keywords:

• ostracism
• risk-taking
• anger
• emotion

Acknowledgement

The authors would like to thank and acknowledge Tobias Greitemeyer for his advice and consultation on the project. This research was supported by a Linkage Project grant from the Australian Research Council (LP100200362) in collaboration with AAMI, Ltd. The funders had no role in the design, analysis, or interpretation of the results.

Notes

1. Terms such as social rejection, exclusion, and ostracism are often used interchangeably. However, each has specific situational antecedents and associated research paradigms. While the present study focuses on ostracism – the experience of being ignored and excluded by those around us – the fundamental theoretical premises are all compatible with research on social rejection and exclusion.

2. Note, however, that the psychological desire for control does not necessarily translate into controlled behavior.

3. We also tested two alternative two-step mediation models a) where anger influences control need-threat which in turn influences risk-taking and b) where control need-threat influences anger, which subsequently influences risk-taking. There was no evidence of a two-step ostracism-anger-control need threat – risk-taking mediation path (95%CI [−.08, .38]). There was support for the second model, however, in which the control need-threat (as a result of being ostracized) influences anger, which in turn affects risk-taking: (ab’ = .44, SE = .18, 95%CI: [.16, .92]).

4. Experiment 2 also contained a second between participants factor. While playing the driving game after the ostracism manipulation, participants listened to music with angry content (i.e. Go To Sleep and Still Don’t Give A F**k by Eminem) or two songs without angry content (i.e. Just The Two Of Us and Switch by Will Smith). These conditions tested the boundary effects of anger on risk-taking, by examining whether induced anger from a non-rejection related experience can increase risk-taking in a similar fashion, that is, the distinction between affect that is experienced as a result of an experience/stimulus and affect that is unrelated to the event, but could nonetheless affect subsequent behavior (Peters et al. 2006). After playing the driving game, participants reported on the extent to which they experienced ten different emotions elicited by the music. Relative to the control condition, participants in the angry music condition reported feeling significantly more angry, disgusted, resentful, sad, and stressed, and less aroused, excited, happy, and joyful (ps < .05). However, risk-taking was not influenced by music, music-induced anger, or the interaction between music and the ostracism manipulation (ps > .10), and the angry music manipulation had no significant effect on primary needs. t(70) = .382, p > .7. Therefore, risk-taking was not influenced by anger that was not induced by ostracism.

5. Judging the intentionality of some behaviors was apparent when viewing the videos – participants for example deliberately swerved or deliberately crashed their vehicle even though the behavior was avoidable and unnecessary. Additional cues to intentionality were whether participants did not try to stop the motion of the vehicle or correct their actions or try to turn away. Finally, coders, while blind to the experimental condition of the participant, could compare the participant’s baseline performance to the driving performance in judging each driving behavior.

6. Because the coding of the videos distinguished risky from explicitly aggressive driving, we conducted exploratory analyses on aggressive driving. There was weak evidence of an effect of post-Cyberball anger on aggressive driving, (b = .11, SE = .59, p = .06), as well as an indirect effect between ostracism and aggressive driving as mediated by anger (ab = 3.33, SE = 2.13, 95%CI [.04, 8.56]).

7. We also conducted two-step mediation analyses between ostracism and risk-taking, with anger and control in the mediation chain, respectively. There was no evidence of an indirect effect via this causation chain (95% CI −3.87, 4.82]). The two-step mediation chain with control need-threat and anger as mediators, respectively, did show evidence of a significant indirect effect between ostracism and risk-taking (ab’ = −1.18, SE = .80, 95%CI [−3.82, −.11]).

Additional information

Funding

This research was supported by a Linkage Project grant from the Australian Research Council (LP100200362) in collaboration with AAMI, Ltd. The funders had no role in the design, analysis, or interpretation of the results.