April 4-6, 2024 • Hyatt Regency • Lexington, KY
Innovations in Health Communication
Innovations in Health Communication
Abstract: Descriptive Norms ≠ Injunctive Norms? A Meta-Analysis
Introduction. Social norms are weaved into every corner of health communication, from binge drinking and eating disorders among adolescents to the act of covering one’s face when coughing. Despite a voluminous body of work (Shulman et al., 2017) and no shortage of theories or models (e.g., Lapinski & Rimal, 2005), the empirical clarity of social norms remains relatively woolly (Yanovitzky & Rimal, 2006). Still, general consensus suggests that perceived social norms comprise two different types of evaluations—descriptive and injunctive norms.
Purpose. Conceptually speaking, descriptive and injunctive norms are distinct; for instance, individuals may estimate a high prevalence of unsafe sex (i.e., descriptive norm) even if this behavior is strongly denounced by authority figures or important referents (i.e., injunctive norms). After all, the motivating force behind injunctive norms is social approval, whereas the motivating force behind descriptive norms is the belief that if most others are doing it, it is probably a wise thing to do (Walter et al., 2017). Things get far murkier, however, when examining the interplay between descriptive and injunctive norms, as studies often find considerable overlap between the two concepts and their relationship to key antecedents and health-related outcomes (e.g., Frohe et al., 2018). Thus, there is a current need to disentangle inconsistencies, elucidate underlying mechanisms, and outline boundary conditions.
Method. The present meta-analysis zeroed-in on the relationship between descriptive and injunctive norms across four health contexts: tobacco use (k = 12, N = 11,305), vaccine adherence (k = 9, N = 8,253), substance use (k = 21, N = 17,052), and binge drinking (k = 23, N = 14,250). Primary studies were obtained from eight electronic databases (e.g., All Academic) using the Boolean string “descriptive AND injunctive” (990 studies after removal of duplicates). For inclusion in the analysis, studies had to (a) include self-report measures of descriptive and injunctive norms; (b) measure norms pertaining to the contextual variables of interest; and (c) report relevant statistics for calculating effect sizes. Informed by theories of social norms, the meta-analysis also examined four types of possible moderators: individual level factors (e.g., age), group-level factors (e.g., group identity), outcome expectations (benefits to oneself vs. to others), and behavior factors (e.g., perceived publicness). Correlations from primary studies, pooled-effect sizes, homogeneity statistics, a sensitivity analysis, moderation analyses, meta-regression analyses, and publication bias tests were calculated using Comprehensive Meta-Analysis (version 4) software (Borenstein et al., 2022).
Results. Descriptive and injunctive norms are moderately and positively associated (r = .19, 95% CI [.09, .29] p < .001), supporting the claim that the two are distinct, albeit related, judgments. Regarding health-related outcomes, results show substantial parity in associations between descriptive/injunctive norms with topic-relevant attitudes and behaviors (r = ~.17, 95% CI [.07, .26] p < .001). Moreover, notable differences emerge when considering the boundary conditions imposed by theory-driven moderators.
Conclusion. Results point to an interesting interplay between descriptive and injunctive norms as being associated with distinct boundary conditions but– by and large– generating similar main effects. We offer theoretical implications and directions for future research.
Purpose. Conceptually speaking, descriptive and injunctive norms are distinct; for instance, individuals may estimate a high prevalence of unsafe sex (i.e., descriptive norm) even if this behavior is strongly denounced by authority figures or important referents (i.e., injunctive norms). After all, the motivating force behind injunctive norms is social approval, whereas the motivating force behind descriptive norms is the belief that if most others are doing it, it is probably a wise thing to do (Walter et al., 2017). Things get far murkier, however, when examining the interplay between descriptive and injunctive norms, as studies often find considerable overlap between the two concepts and their relationship to key antecedents and health-related outcomes (e.g., Frohe et al., 2018). Thus, there is a current need to disentangle inconsistencies, elucidate underlying mechanisms, and outline boundary conditions.
Method. The present meta-analysis zeroed-in on the relationship between descriptive and injunctive norms across four health contexts: tobacco use (k = 12, N = 11,305), vaccine adherence (k = 9, N = 8,253), substance use (k = 21, N = 17,052), and binge drinking (k = 23, N = 14,250). Primary studies were obtained from eight electronic databases (e.g., All Academic) using the Boolean string “descriptive AND injunctive” (990 studies after removal of duplicates). For inclusion in the analysis, studies had to (a) include self-report measures of descriptive and injunctive norms; (b) measure norms pertaining to the contextual variables of interest; and (c) report relevant statistics for calculating effect sizes. Informed by theories of social norms, the meta-analysis also examined four types of possible moderators: individual level factors (e.g., age), group-level factors (e.g., group identity), outcome expectations (benefits to oneself vs. to others), and behavior factors (e.g., perceived publicness). Correlations from primary studies, pooled-effect sizes, homogeneity statistics, a sensitivity analysis, moderation analyses, meta-regression analyses, and publication bias tests were calculated using Comprehensive Meta-Analysis (version 4) software (Borenstein et al., 2022).
Results. Descriptive and injunctive norms are moderately and positively associated (r = .19, 95% CI [.09, .29] p < .001), supporting the claim that the two are distinct, albeit related, judgments. Regarding health-related outcomes, results show substantial parity in associations between descriptive/injunctive norms with topic-relevant attitudes and behaviors (r = ~.17, 95% CI [.07, .26] p < .001). Moreover, notable differences emerge when considering the boundary conditions imposed by theory-driven moderators.
Conclusion. Results point to an interesting interplay between descriptive and injunctive norms as being associated with distinct boundary conditions but– by and large– generating similar main effects. We offer theoretical implications and directions for future research.