Seminarios Académicos

Abstract: Despite the often-contentious use of deception in behavioral science experiments, empirical insights are rare. We take advantage of a naturally occurring study using false information in stimulus materials–one of the most common forms of deception–to measure causal effects on participant behaviors and attitudes. By augmenting this deception study with a non-deceptive arm, we assess concerns of deception ‘poisoning’ the participant pool, often held in experimental economics. Specifically, we examined how recent exposure to deception affected behavior in prosocial and honesty tasks, attrition rates, attention and attitudes in a subsequent study. Our large sample (nDeceptionStudy = 1,067, nSpilloverStudy = 650) is drawn from an interdisciplinary laboratory silent on deception usage, thereby avoiding common selection effects. We observed that behavior in honesty and prosocial tasks is largely unaffected by prior deception exposure, against commonly held expectations of adverse changes. This was the case even when prior deception was made salient via a randomly assigned ‘strong’ debriefing. In addition, participant attrition, attention was unaffected by prior deception. Elevated suspicion from deception was found only in relation to making charity payments, contrary to expectations that deception exposure in general increases suspicion. Trust in researchers was adversely affected by past deception, and completion rates were lower among those treated with the ‘strong’ debriefing–a treatment unrepresentative of current practices. Both participants and researchers drawn from the campus where the laboratory was located (n = 28) commonly overestimated spillover effects from deception. While these findings may reduce concerns of systematic distortions in behavior from false information deception, they also present a dilemma: adequately informing participants of deception could cause more harm, and effective redress remains unknown.