Forty-three convicted sex offenders read each of four different offence vignettes that involved a man forcing a female victim into sex and the offender's subsequent police interview. The experimental manipulation involved giving participants each of four different scenarios concerning how the police interviewed the offender. These were interviews characterized by humanity, dominance, displaying an understanding of sex offenders’ cognitive distortions, or a neutral, control interview. Participants were required to rate the interviews on a variety of dimensions, such as the offender's likelihood of confessing, and the fairness of the interview. Where participants were told the man had been interviewed with humanity and compassion, they rated the offender as more likely to confess and rated the interview as fairer than the other conditions. In contrast, participants rated the offender interviewed with a dominant approach as less likely to confess, and for this procedure to be less fair than the other conditions. Displaying an understanding of sex offenders’ cognitive distortions appeared to have had no influence on perceived likelihood of confessions but was perceived to make the crime appear less serious.