<p>References are a ubiquitous, personality-focused preemployment assessment with relatively low validity. Nevertheless, few studies have explored interventions to improve the accuracy of references. Drawing on accountability theory, we manipulated a reference tool to either include or not include (a) a Reason Giving intervention that required participants to describe the reasons underlying the personality ratings they provided, and (b) a Monetary Incentive intervention which explained that participants would receive additional compensation for their references. We applied this tool in context of a simulated call center, to reflect the common situation where referees provide information about job applicants whom they do not know well. Two-hundred working supervisors and managers were randomly assigned to one of four possible rating conditions. Participants completed a personality rating form online while role-playing as referees for a set of prototypical call-center job applicants. The Reason Giving intervention had significant negative consequences for participants’ rating accuracy, but no other intervention effects were observed. This study suggests novel and potentially paradoxical pathways for understanding some of the observed inaccuracies in references. It also suggests that organizations who compensate referees may not get what they pay for in these contexts. Study limitations and future directions are discussed.</p>