Download data and study materials from OSF
Carnegie Mellon University
The Ohio State University
The Ohio State University
Sample size: 1340
Field period: 06/13/2018-05/29/2019
An American’s yearly chance of being killed by a terrorist attack sits at about 1 in 3.5 million. Yet over 40% of the American public consistently believes that they or their family members are likely to be the victim of a terror attack. Can these inflated estimates of the risks of terrorism be brought closer to reality? With trillions of dollars spent on the War on Terror since 9/11, this question is not just theoretically but practically important. In order to investigate, we field a nationally representative survey experiment containing a brief vignette with corrective information about the actual risks of terrorism vs. other dangers facing Americans. Additionally, we vary whether there is a political elite endorsement accompanying the information, with either a Democratic politician, Republican politician, or senior military officer driving home the message.
We explore whether corrective information about the actual risk of terrorism facing Americans will mitigate misperceptions. Our core hypothesis is that the efficacy of this “corrective message” will hinge on endorsements of this new information by political elites. In particular, we test the impact of an endorsement from elites that (1) align with the respondent's partisan affiliation (co-partisanship) (2) can claim partisan ownership of the terrorism issue in recent years (Republicans), and (3) have credibility and trust across partisan lines (military officer).
Our study employs a fully randomized design with four treatment groups and one control group. The control group only receives a vignette that captures the perception of terrorism as threatening. The treatment groups each read the control vignette and then are provided a corrective vignette with information on the risk posed by terrorism relative to other everyday threats. The first treatment group’s vignette only includes these risk statistics. The remaining three treatment group vignettes also include an endorsement of the risk statistics by one of three fictitious elites -- a Democratic congressman, a Republican congressman, or a military general. We also field the same experimental design simultaneously on MTurk in order to provide additional evidence and investigate the ability of MTurk to recover the results of experiments like ours as compared to high-quality, representative platforms. Moreover, the parallel MTurk experiment includes a two-week follow to examine the persistence of any observed effects.
Post-treatment questions were asked about the extent to which respondents saw terrorism (along with other safety hazards) as threatening both personally and nationally, their degree of worry about becoming a terror victim, their expectations about future major attacks, and their perceptions about the importance of counterterrorism efforts.
Overall, we found that presenting people with factual information about the relative risk of terror attacks substantially altered their perceptions and beliefs. After seeing the treatment, respondents were much less likely to view terrorism as an important threat to U.S. national security and to prioritize combatting it as a key U.S. foreign policy objective. The effects we observe are substantial in substantive terms and relatively persistent in our two-week MTurk follow-up wave. Additionally, while there is some variation due to elite endorsements, the main story is informational -- rather than partisan or institutional -- in nature. These results show that the American public’s inflated fear of terrorism is far more malleable and correctible than previously thought. In this sense, countering terrorism may largely require providing context and perspective.
Silverman, Daniel, Daniel Kent, and Christopher Gelpi. 2021. "Putting Terror in Its Place: An Experiment on Mitigating Fears of Terrorism among the American Public." Journal of Conflict Resolution, doi:10.1177/00220027211036935.
Papers presented at the 2018 APSA annual meeting and the 2018 ISA annual meeting.