Can the goal of honesty be primed?

In a simple study involving 64 participants, Rasinski, Visser, Zagatsky, and Rickett (2005) reported that requiring people to make semantic judgments involving four words related to honesty (embedded among other words) increased the likelihood that they would later admit to having engaged in problematic alcohol-related behaviors (e.g., drinking to the point of blackout). If valid, this honesty-priming effect would offer a powerful intervention to improve the validity of self-report data in many different contexts. To determine whether the effect is repeatable, we first attempted two replications using the same materials, tasks, and measures used by Rasinski et al. Experiment 1 repeated the study with a sample of 150 students. No priming effects were observed here, nor in a follow-up study using adults recruited on the web (Experiment 2). Experiment 3 used the same priming manipulation together with a more refined measure of response candor (derived from Paulhus, 1991). Again, the honesty-related primes had no detectable effects.