Scientific Uncertainty as a Moderator of the Relationship between Descriptive Norm and Intentions to Engage in Cancer Risk–Reducing Behaviors

Abstract

This study examined motivational factors underlying six behaviors with varying levels of scientific uncertainty with regard to their effectiveness in reducing cancer risk. Making use of considerable within-subjects variation, the authors examined the moderating role of the degree of scientific uncertainty about the effectiveness of cancer risk–reducing behaviors in shaping relationships between constructs in the Integrative Model of Behavioral Prediction (Fishbein & Yzer, 2003). Using cross-sectional data (n = 601), the descriptive norm-intention relationship was stronger for scientifically uncertain behaviors such as avoiding BPA plastics and using a hands-free mobile phone headset than for established behaviors (e.g., avoiding smoking, fruit and vegetable intake, exercise, and applying sunscreen). This pattern was partially explained by the mediating role of injunctive norms between descriptive norm and intentions, as predicted by the extended Theory of Normative Social Behavior (Rimal, 2008). For behaviors more clearly established as an effective means to reduce the risk of cancer, self-efficacy was significantly more predictive of intentions to perform such behaviors. The authors discuss practical implications of these findings and theoretical insights into better understanding the role of normative components in the adaptation of risk-reduction behaviors.

Notes

¹Measures relevant to coffee drinking were also assessed but excluded in this article because of the unique nature of coffee drinking. Although there is uncertainty regarding the relationship of coffee drinking and cancer risk, it differs from other uncertain behaviors in that it was associated with increased cancer risk as early as the 1980s, and has been associated with both increasing and reducing the risk of various cancers over time. No studies suggest that avoiding BPA plastics or using hands-free mobile headsets increase cancer risk. We thus excluded coffee drinking to separate it from the other behaviors.

²A tolerance of less than .20 and/or a variance inflation factor of 10 and above indicates a multicollinearity problem.

³According to Muller and colleagues (2005), the reduction in the magnitude of interaction coefficients is taken as evidence that a mediator (partially) mediates the effect of the Treatment × Moderator interaction on outcomes. Only when the moderation of the residual treatment effect is nonsignificant, it is considered “full” mediated moderation.