The 16 subjects who served as the 'witnesses' were undergraduates from the introductory psychology course at the University of California, Los Angeles. These students were not informed in advance that their memory for a staged incident would be tested. Instead, they were asked to volunteer for participation in an experiment on 'improving your memory.' The experimental procedure was to have confederate research assistants enter a classroom session and perform a staged incident which was recorded on videotape. The student observers then served as 'witnesses' to be interviewed about the incident. Subjects in the control condition were asked to respond to a series of nonleading questions about the persons, objects, and events of the incident with as much detail as they could recall. In the cognitive interview condition, the subjects completed the same questionnaire, but only after they had learned several memory retrieval mnemonics to be used in answering the questions. The retrieval mnemonics emanated from two overriding perspectives in cognitive theory: the encoding specificity principle (Tulving and Thomson 1973) and the multicomponent view of a memory trace (Bower 1967; Wickens 1970). The encoding specificity principle holds that the effectiveness of a retrieval probe is determined by its similarity to the encoding (acquisition) operations. Thus, a retrieval environment that effectively reinstates the original environment should enhance memory. The multicomponent view of memory maintains that the memory trace is not a unitary, holistic representation of the original event, but rather it is a complex array of many features. Findings showed that the cognitive interview produced significantly more correct information without an accompanying increase in the amount of incorrect information. Tabular data and 19 references are provided.