ABSTRACT
Public opinion researchers often find changing attitudes about pressing public health issues to be a difficult task and even when attitudes do change, behaviors often do not. However, salient real-world events have the ability to bring public health crises to the fore in unique ways. To assess the impact of localized public health events on individuals’ self-reported behavior, this paper examines Floridians’ intentions to take preventative measures against the Zika virus before and after the first locally transmit- ted case of Zika emerged in Florida. We find that local and national media coverage of Zika increased significantly following its first transmission in the U.S. Critically, we also find that Floridians surveyed after this increase in media coverage were more likely to pay attention to Zika-related news, and self-report intentions to take protective action against the virus. These results suggest that behavioral intentions can shift as health threats become more proximate.
Supplementary material
Supplemental data for this article can be accessed at https://doi.org/10.1080/10410236.2019.1662556.
Notes
1. The backfire effect has not always been found in subsequent studies (Haglin, Citation2017; Wood & Porter Citation2019).Similarly, social media has been found to serve as a corrective to false information in the Zika virus context (Bode & Vraga, Citation2018). And beyond these more general psychological forces, public health campaigns can be more more or less successful depending on the topic they address (Snyder & Hamilton, Citation2002; Snyder et al., Citation2004).
2. We estimate all of this in Stata 15 using the rdrobust and rdplot packages from Calonico, Cattaneo, Farrell, and Titiunik (Citation2017).
3. Hereafter, the phrase “pre-event” refers to the time period before the CDC reported the first locally contracted case of Zika in the United States and the phrase “post-event” refers to the time period after the CDC reported the first locally contracted case of Zika in the United States.
4. As this measure was created based on our intuition and the work by Chan et al. (Citation2018), other analysts might arrive at a different approach to scoring these items. We conducted analyses entering each indicator as a separate covariate for our key models of self-protective action. The results are highly similar and do not change the substantive inferences regarding the pre- vs. post-event demarcation, and thus are not reported further in the main text in the interest of parsimony.