About 70% Of Participants Know That the Canonical Deception Paradigms Measure Dishonesty

Four paradigms dominate the study of behavioral ethics: the coin task, die task, matrix task, and sender task. Over 60,000 people have participated in experiments with these paradigms, and results from these studies inform theory and practice. In this article, I challenge the construct validity of these tasks. In Study 1, I show that most participants know that these tasks are designed to measure their dishonesty. This awareness leads most participants to believe that the experimenter expects and wants them to lie and leads many to conceptualize lying in these experiments as an amoral activity. In Studies 2, I show that participant sophistication significantly correlates with participants’ overreporting. In Study 3, I conduct three construct validation tests by running participants through three foundational experiments and show that sophistication affects behavior in these experiments. Rather than measuring deception, these canonical paradigms reflect how participants perceive their relationship with the experimenter.