In this paper with Edgar Kausel, we conducted three lab studies to examine mechanisms and interventions of interviewer overconfidence. We found that interviewers judged their own interview questions as more valid when the same questions were rated by someone else. We also found that putting in more effort increased the perceived validity of the questions but not the question quality.
Our findings suggest that letting interviewers write their own questions is not a good idea. They often write poor questions, but also overestimate their validity. Instead, it is best to provide interviewers with high-quality, pre-written questions that they can choose. Read the full paper here: