Examining the effectiveness of physical threats in road safety advertising: The role of the third-person effect, gender, and age

Threatening advertisements have been widely used in the social marketing of road safety. However, despite their popularity and over five decades of research into the fear-persuasion relationship, an unequivocal answer regarding their effectiveness remains unachieved. More contemporary "fear appeal" research has explored the extent other variables moderate this relationship. In this study, the third-person effect was examined to explore its association with the extent male and female drivers reported intentions to adopt the recommendations of two road safety advertisements depicting high physical threats. Drivers (N = 152) first provided responses on pre-exposure future driving intentions, subsequently viewed two advertisements, one anti-speeding and one anti-drink driving, followed by measurement of their perceptions and post-manipulation intentions. The latter measure, post-manipulation intentions, was taken as the level of message acceptance for each advertisement. Results indicated a significant gender difference with females reporting reverse third-person effects (i.e., the messages would have more influence on themselves than others) and males reporting classic third-person effects (i.e., the messages would have more influence on others than themselves). Consistent with such third-person effects, females reported greater intention not to speed and not to drink and drive after being exposed to the advertisements than males. To determine the extent that third-person differential perceptions contributed to variance explained in post-manipulation intentions, hierarchical regressions were conducted. These regressions revealed that third-person scores significantly contributed to the variance explained in post-manipulation intentions, beyond the contribution of other factors including demographic characteristics, pre-exposure intentions and past behaviour. The theoretical and applied implications of the results are discussed.