The 90-minute interviews were conducted in 1989 and 1990. Results revealed that 53 percent of the sample drank alcohol on the day of the homicide, while 32 percent of the sample said they were drunk at the time of the homicide. One-third of the sample smoked marijuana in the 24-hour period before the homicide, while 18 percent of the sample were high on marijuana when the homicide occurred. Twenty-two percent used cocaine in the prior 24-hour period, and 14 percent were high on the drug. Most participants who used a drug were not hardcore users of that drug. About one in five of the participants could be considered polydrug abusers. Thirty percent of the sample believed that the homicide was related to their drug use. Alcohol was the drug most likely to be implicated in these homicides. Findings suggested that drug users become violent when they become high and that self-reports are more useful than existing criminal justice records for understanding the relationship between offender drug use and participation in homicide. Tables and 32 references (Author abstract modified)