The long-term objective of this project is to examine whether a well-studied cognitive bias (i.e., anchoring effect) may effect the validity of studies examining the efficacy of norm-based interventions. Prior research suggests that personalized normative feedback (PNF) interventions are successful as reducing self-reported alcohol use, but there is very little support suggesting that they successfully reduce alcohol-related problems. Given the very strong relationship between alcohol use and alcohol-related problems, this finding is peculiar. As stated in NIAAA's mission statement, NIAAA seeks to provide "leadership in the national effort to reduce alcohol-related problems." The proposed studies examine whether the presentation of normative information to individuals in the treatment group of a PNF intervention biases estimates of self-reported alcohol use, which could explain why reported alcohol use decreases while consequences are unaffected. Study 1 examines several hypotheses. Using the study design to rule out behavioral change (i.e., cross-sectional design), Hypothesis 1 states that providing normative information will produce an anchoring effect such that self- reported alcohol use will be affected in the absence of behavioral change. To further examine the boundary conditions of the anchoring affect, the idea that the anchoring effect occurs without conscious awareness is tested (nonconscious anchoring effect hypothesis). Hypothesis 2a predicts that the anchoring effect will occur even when participants have been forewarned of the anchoring effect, and Hypothesis 2b predicts that the anchoring effect will be present even after controlling for social desirability bias. Further, Study 1 examines whether the greater the anchoring effect is stronger when the anchor is more applicable (anchor applicability hypothesis). Hypothesis 3a predicts that the anchoring effect will be stronger when presented with gender- specific normative feedback as compared to gender-neutral feedback, and Hypothesis 3b predicts that the anchoring effect will be stronger for individuals who more closely identify with the referent group. To extend the findings from Study 1, Study 2 uses a longitudinally design (i.e., 1-month follow-up) to examine two specific hypotheses. Hypothesis 4 predicts that the anchoring effect predictions will better account for reported change in alcohol use than predictions derived from Social Norms Theory. Hypothesis 5 predicts that perceived drinking norms will mediate the effect that normative information has on alcohol use estimates, even when these changes in estimated alcohol use cannot be attributed to actual behavior change. Using ecological momentary assessment (EMA) methods, Study 3 examines one unique hypothesis. Hypothesis 6 predicts that the anchoring effect will be present when using past week retrospective alcohol use measures, but not when using aggregated daily diary assessments. The proposed studies have broad implications to the development and testing of alcohol interventions.
This project will examine a cognitive bias that may undermine the validity of self-reported alcohol consumption after being provided with normative information. Specifically, the Anchoring Effect may dissimulate the intervention efficacy for norm-based interventions. If the hypotheses are supported, then the standard assessment of alcohol consumption used to evaluate personalize normative feedback interventions will be called into question and improved methods will be suggested to better assess intervention efficacy.