This paper discusses the reliability of college faculty peer evaluations. Three evaluation methods are compared: observation; impressionistic; and documentation review. The quality of information provided by each method is examined based on five areas: the purpose of the data being collected; what is being measured; who is responsible for doing the peer evaluation; when the evaluation is conducted; and how the evaluation is conducted. It is noted that evaluation problems still exist, such as judges who lack sufficient knowledge in the candidate's field, too few judges, close collegial relationship, adversarial relationships, lack of anonymity of the judges, and the overall problems inherent in a system that allows decisions to be made without a reasonable rationale. It is suggested that the peer documentation review technique is a critical aspect of peer evaluations, that further research is needed in this area, and that higher education institutions need to supplement student rating information with peer evaluations in annual reviews. Contains 20 references. (GLR)