The question–behavior effect: Genuine effect or spurious phenomenon? A systematic review of randomized controlled trials with meta-analyses.

Abstract

Simply answering questions about a specific behavior may change that behavior. This is known as the mere-measurement or question-behavior effect (QBE). Our objective was to synthesize the evidence for the QBE on health-related behaviors. Included studies were randomized controlled trials that tested the effect of questionnaires or interviews about health-related behaviors and/or related cognitions compared with a no-measurement control condition or another form of measurement. Subgroup analyses were conducted to identify potential moderators. 41 studies were included assessing a range of health behaviors. Meta-analyses showed a small overall QBE effect (SMD = 0.09; 95% CI [0.04, 0.13]; k = 33). Studies showed moderate heterogeneity, variable risk of bias, and evidence of publication bias. No dose-response relationships were found from studies comparing more with less intensive measurement conditions. There were no significant differences in QBE by behavior, but QBEs for dental flossing, physical activity, and screening attendance were significantly different from 0. Findings were not altered by whether behavior or cognitions were measured, attitudes were or were not measured, studies used questionnaires or interviews, or outcomes were objective or self-reported. There is some evidence for the QBE on health-related behavior. However, risk of bias within studies and evidence of publication bias indicate that the observed small effect size may be overestimated, especially given that some studies included intervention techniques in addition to providing questionnaires. Preregistered high-quality trials with clear specification of intervention content are needed to confirm if and when measurement leads to behavior change.