The first experiment, conducted with college students based upon their witnessing of a disruptive classroom incident involving a stranger entering the classroom, examined whether multiple witnesses interacting with one another produce greater identification accuracy than lone witnesses recalling details of an incident. Findings showed that where people have seen an event with others and interacted afterwards, their testimony will show a high degree of homogeneity but will be highly unreliable in principle because it will be a composite of witness observation and reaction. The second experiment considered the effects on accuracy of the character of the interaction between the witness and the interviewer. Specifically, the effect on testimony of the status of the interviewer was explored. Results showed that overall testimony accuracy was higher with the low status questioner than with the high status questioner. It was also found that females were significantly worse than males with the high status questioner while being better than males with the low status questioner. The third experiment investigated the comparative effects of violent and nonviolent incidents on testimony accuracy of witnesses. Incidents and descriptions in the nonviolent filmed incident were recalled with greater accuracy. Principles regarding the accuracy of eyewitness testimony established by psychological research are listed and briefly discussed. The use of the event methodology in simulating real-life eyewitness conditions should convince criminal justice professionals of the reliability of research findings and their applicability to actual cases. References (about 40) are provided. For related documents, see NCJ70738-45 and 70347-48.