To test the assumption that individuals who share a personal relationship are more likely to corroborate one another's false alibi than are strangers, 81 undergraduate students were provided the opportunity to either corroborate or refute a confederate's alibi for a suspected theft. In a 'friendship' condition, feelings of affiliation between the participant and the confederate were experimentally induced by increasing the perceived similarity between the pair, and by having the pair interact during a collaborative task. Later during the experimental session the confederate became a suspect for a mock crime and provided a false alibi that she was with the participant during the entire session. Contrary to what we hypothesized, participants in the 'stranger' condition were as likely to corroborate the false alibi as those who underwent friendship-enhancing activities. When the confederate acted in a highly suspicious manner, however, she was much less likely to have her false alibi corroborated by participant than when the confederate's behaviour was less suspicious. The results put into question our assumptions of what makes a credible witness and emphasizes the need for further empirical research on the behaviour of alibi corroboration.