Two hundred and ninety-one new adult male inmates convicted of murder, robbery, arson, rape, forcible indecency, and kidnapping completed questionnaires about their pretrial interviews. Logistic regression analyses revealed that marital status, feelings of guilt, and perceptions of the strength of the evidence significantly affected the likelihood of confession by those who had already decided to confess prior to interview. By contrast, prior arrest history and interview style affected the likelihood of confession by those who had not previously decided to confess. Suspects who had no previous arrests and had undergone either Undifferentiated-high or Relationship-focused interviews were more likely to make full confessions. In Undifferentiated-high interviews, police officers employed all of the interviewing techniques examined: they listened closely to the suspects' accounts, attempted to build good relationships, and discussed the crimes, while also presenting evidence and confronting the suspects. In Relationship-focused interviews, police officers listened attentively to the offenders' accounts, tried to build good relationships, and discussed the crimes directly. Prisoners who experienced Relationship-focused interviews felt satisfied with their confessions and admitted guilt at trial.