Interview Sequences and the Formation of Subjective Assessments
This paper studies how the interview assessment of a candidate depends on the other candidates seen by the same evaluator. We leverage novel administrative data on about 29,000 one-to-one interviews conducted within the admission process of a large merit-based study grant program. Identification relies on the quasi-random assignment of candidates to evaluators and time slots. We find that a candidate's assessment decreases when her evaluator receives a better overall candidate draw. The previous candidate has a substantial additional influence, which produces a strong negative autocorrelation in admission outcomes. The observed pattern suggests that evaluators exhibit a contrast effect due to the associative recall of prior candidates.