Friends Don’t Let Friends Smoke: How Storytelling and Social Distance Influence Nonsmokers’ Responses to Antismoking Messages

ABSTRACT

This study examines nonsmokers’ responses to antismoking messages. Informed by construal-level theory (CLT), it investigates whether and how evidence type (narrative vs. non-narrative) and social distance might interact to influence nonsmokers’ attitudes toward others’ quitting smoking and intentions to persuade others to quit smoking. Results of a controlled experiment (N = 281) revealed an approximately significant two-way interaction pertaining to attitudes. Simple effects analyses revealed that narratives produced less-favorable attitudes toward others’ quitting smoking than nonnarratives when participants thought about their best friend. Yet, there was no difference in attitudes between narratives and nonnarratives when participants thought about socially distant others. The results also indicated that nonnarratives overpowered narratives to influence participants’ attitudes toward others’ quitting smoking. Moreover, social distance had a consistent impact on their risk beliefs, such that they perceive fewer health risks of their close friends than an average college student. Theoretical and practical implications of the results are discussed.

Notes

^1. The stimuli are available from the corresponding author upon request.

^2. We also conducted four separate CFAs for each key outcome variable. For the attitude variable, the model is a just-identified model that fits perfectly by default, RMSEA = .000; SRMR = .000; CFI = 1.000; and (0) = 0; for the intention variable, the model fit is good, RMSEA = .000; SRMR = .000; CFI = 1.000; and (1) = 0.001. The residuals of the third and fourth manifest variables were correlated; for the susceptibility variable, the model fit is acceptable, RMSEA = .161; SRMR = .021; CFI = .988; and (2) = 16.550; for the severity variable, the model fit is good, RMSEA = .033; SRMR = .007; CFI = 1.000; and (2) = 2.629.

^3. In order to control the interaction effect and properly interpret the main effects, we performed another four separate ANCOVAs by creating the interaction term as a variable and entering it in the model as a covariate. We found a significant main effect of evidence type on attitudes, F(1, 274) = 5.86, p = .016. Moreover, social distance had a significant main effect on perceived severity, F(1, 274) = 4.43, p = .036, and approximately significant effect on perceived susceptibility, F(1, 274) = 3.49, p = .063. No other main effect was found.