Honesty after a labor relationship

At the end of a controlled experiment where research assistants were hired for coding news from online newspapers, the experimenter-employer asked a number of them to roll a die and report the result in order to be paid in cash an amount linear on the reported number from 1 to 6 that could go from 1.6 to 9.4 USD. Another (control) group of similar students, recruited in a similar manner, were also invited to perform the same die-roll task, but they had no prior labor relationship with the experimenter-employer. Our treatment group showed in average higher levels of honesty as their distribution of reported numbers was less skewed to the right, that is, the long-term labor relationship group was more likely to report numbers that are closer to the uniform (honest) distribution than our control, and than other reported numbers in this kind of experiments. We conjecture that the previous experimenter-subject relationship of the treatment group induced higher levels of honesty among the participants. One of the possible reasons is that the labor relationship created for the group of †treatment†students included a series of shocks that involved the possibility of involuntary unemployment, bringing incentives for the students to signal honesty as a trait that could be valued in the labor market. This paper contributes to the growing literature on understanding the motives for honesty and cheating.