Interviewers made pre-interview and postinterview judgements of inmates' likelihood of completing parole successfully and rated inmates' honesty during the interview. The study was conducted in 1982 and used four different interviewers, three males and one female. A measure of reliability was obtained using an observer who sat in on 10 interviews; 45 interviews were tape recorded. Data on disciplinary infractions in prison and parole violations serious enough to result in revocation were collected 32 months after the interviews. The results showed that information obtained during the interview significantly affected the interviewers' postinterview judgements. Perceived honesty was affected significantly by the severity of the inmates' crime and by how much information the inmates volunteered, how polite they were, and how much inmates fidgeted during the interview. Subsequent analysis of the 47 inmates released via parole indicated that interviewers were less accurate in predicting parole failures after than before the interview. Tables and nine references. (Author abstract modified)