People who approve of one's physical activity vs. people who are influential: A randomized cross-over experiment to compare normative referent elicitation procedures.

Objectives: To compare injunctive norm scores assessed based on Ajzen's (2002) original normative referent elicitation method (e.g., norms from those who approve/disapprove of physical activity) and a revised method (e.g., norms from those who are personally influential).

Methods: Undergraduate students (N = 1008, N= 383) were initially randomly assigned to either the condition that used Ajzen's original method or the condition that used the revised method, and completed an online survey. A month later, participants completed the survey in the reverse condition. Data collection occurred October 2021-July 2022. Data were analyzed to examine the differences between the two conditions regarding the mean injunctive norm scores and their indirect effects on physical activity via intentions (i.e., moderated mediation).

Results: Mixed ANOVA results indicated a statistically significant effect of condition (p < .001): the injunctive norm score assessed based on the revised method was lower compared to the original method. After accounting for baseline physical activity, a statistically significant moderated mediation was found with intention strength as the mediator (95% bootstrap CIs [.02, .20]): there was a statistically significant indirect effect of injunctive norms on physical activity via intention strength under the revised condition, but this effect was statistically nonsignificant under the original condition. There was no statistically significant moderated mediation with decisional intention as the mediator (95% bootstrap CIs [-.04, .13]).

Conclusions: Null findings regarding the relationship between injunctive norms and physical activity intentions typically found in the literature may be partially explained by the way in which norms were measured.