بازنگری توسل به ترس: یک ساختار تجدید استعلام از مدل انگیزه حفاظت

Replicating Johnston and Warkentin (2010), we demonstrate that social influence and self-efficacy are the main drivers of compliance with fear appeals. Contrary to the original study, we find that the acknowledgment of a severe threat encourages subjects to seize on the proposed recommendation, bolstering perceptions of efficacy. With this sole exception, the original results are fully replicated in a different research context employing a different population.

Fear appeals are communications presenting the threat of impending danger to motivate compliance with a proposed recommendation (Keller & Lehmann, 2008). The Protection Motivation Model (PMM: Maddux & Rogers, 1983), widely employed in marketing (e.g., Pechmann, Zhao, Goldberg, & Reibling, 2003), identifies the drivers of compliance in the audience's assessment of a) threat severity, b) threat susceptibility, c) efficacy of the coping response, and d) self-efficacy in adopting the proposed coping response. Previous meta-analyses, however, show weak direct effects of threat severity and threat susceptibility on behavioral intentions (e.g., Milne, Sheeran, & Orbell, 2000). The current paper presents a conceptual replication of Johnston and Warkentin's (2010) structural model of the PMM, in which the effect of threat severity and threat susceptibility on behavioral intentions is mediated by coping response efficacy and self-efficacy. In addition, we extend the original model by including the relationship between coping response efficacy and self-efficacy. The provision of a clear solution to a problem is likely to increase self-efficacy, as the audience is lead to believe that they can cope successfully with the threat by adopting the recommended course of action. The replication of the original model notably underlines a) a negative effect of threat severity on both measures of efficacy and b) poor goodness-of-fit indices, not reported in the original study due to the component-based approach undertaken (i.e., partial least squares). An alternative model demonstrates higher reliability.