Abstract
Do norms compelling self-censorship of negative communication work? An attributional analysis suggests that awareness of self-censorship norms causes people to be suspicious of other people's positive communications about groups, thus causing the norms to backfire. Three studies tested this informational contamination hypothesis. Participants read stories in which they imagined that some friends' conversations painted a particular fraternity in a good light. Results from all three studies revealed that when participants were exposed to a cue encouraging self-censorship—the presence of a member of the talked-about fraternity—this self-censorship norm backfired, instead leading them to talk disparagingly about the fraternity in a different context. Mediation analyses implicate an informational contamination explanation for the backfiring effect more so than a reactance-based explanation.
ACKNOWLEDGMENTS
We thank Dov Cohen, Peter Darke, Darrin Lehman, Del Paulhus, and (especially) Mark Schaller for their invaluable help in theory and/or manuscript development.
Notes
Note. N = 199. Higher scores on attribution measure = more suspicion of informational contamination; higher scores on disparaging communication = more intent to communicate negatively about the fraternity.
Note. N = 300. Higher scores on attribution measure = more suspicion of informational contamination; higher scores on disparaging communication = more intent to communicate negatively about the fraternity.
1An additional study using a more powerful self-censorship cue supported this conclusion (Conway & Gornick, Citation2009). Instead of manipulating self-censorship norms by including a member of a talked-about fraternity, this additional study manipulated such norms by including a professor who was widely known to advocate group rights. This less subtle manipulation produced a significant backfiring effect when participants communicated to their other “friend” in a different context (consistent with Studies 1–3); however, when communicating in the presence of the professor, this backfiring effect was attenuated, and the moderation pattern was significant. Thus, the more obvious and recognizable the cue, the more likely the “immediate presence” moderating variable will exert its influence.