The selection interview is a commonly used procedure for gathering applicant information upon which to base personnel decisions. Synder and Swann (1978) found that people tended to ask questions and generally seek information consistent with their initial beliefs about the person being interviewed. They suggested that this confirmatory questioning strategy might operate in settings such as the employment interview. Studies employing a fixed question selection response format where subjects selected interview questions from the same experimenter-provided list have failed to replicate, for selection interviews, Synder and Swann's finding. The present study employed a free question generation methodology to remove the potential confound caused by unaccounted for differences between experimenter-provided questions. Forty-eight college students generated pilot data for the construction of stimulus material for the study, 126 student interviewers participated in the experiment proper, and 90 students rated the freely generated questions for positivity/negativity. The results indicated that both male and female interviewers adopted confirmatory questioning strategies, asking significantly more negative questions of below average applicants than of above average applicants, regardless of applicant sex. Further results suggest that this same confirmatory bias exists for males and females when asking positive questions of same sex applicants. However, interviewers adopted disconfirmatory strategies when asking positive questions of opposite-sex applicants. (Author/NB)