Two teams of psychics or 'sensitives,' four in one group and eight in the other, participated in the study. In addition, the two control groups consisted of 11 college student volunteers and 12 homicide detectives. Four actual murder cases, two solved and two unsolved, were selected for study. Case information was available in all four cases for victim data and in two cases for suspect data. One piece of physical evidence from each crime was sealed in an envelope. The two nonpsychic groups were instructed to take each piece of evidence and attempt to intuit or guess characteristics of both the victim and the suspect in each crime. The psychics worked in two teams and were asked to perform the same task. Most of the psychics generated lengthy discourses with dramatic and confident statements. In contrast, the detective group produced very terse and highly qualified statements. The student groups appeared to feel slightly more at ease with the task than did the detectives. Their statements tended to be lists of information without apparently sensory-derived descriptions. The data generated by each subject were classified into 20 response categories and an additional miscellaneous category. Since the psychic group produced approximately 10 times as much information as either of the two comparison groups, it is more likely by chance alone that their data would produce more useful information. Despite this statistical advantage, the psychics were unable to produce information that was significantly better than the comparison groups. The data provided no support for the theory that psychics could produce investigatively useful information. In addition, the data failed to show that the psychics could produce any information relating to the cases beyond a chance level of expectancy. Independent verification procedures are recommended for cases in which psychics are used. Three tables and three references are provided.