Eyewitness testimony has as long history in the court system, and is very persuasive to juries. Jurors are hard pressed to ignore a witness' assertion of a perpetrator's identity. However, the juror's perception of eyewitness testimony is problematic as it has been documented as inaccurate and unreliable in numerous experiments. With the advent of DNA testing and efforts such as The Innocence Project, it has become apparent that faulty eyewitness accounts are central to many wrongful convictions. The intent of this thesis was to explore how law enforcement can facilitate more accurate eyewitness accounts via their interview process. Research suggests that a key problem in the current interviewing system is "post-event information," or outside information introduced by leading questions, exposure to police conversations or other witnesses' accounts. This information can contaminate a witness's memories of events and lead them to report things they did not see. The current experiment explores the effects of 1) warning and educating witnesses about suggestibility and 2) interviewing with leading or open-ended questions. Accuracy scores were then compared for each condition. The hope was to gain insight into methods for improving accurate recall of events and reducing memory contamination from "post-event information."

Contrary to interviewing guidelines, a considerable portion of witness interviews are not recorded. Investigators’ memory, their interview notes, and any subsequent interview reports therefore become important pieces of evidence; the accuracy of interviewers’ memory or such reports is therefore of crucial importance when interviewers testify in court regarding witness interviews. A detailed recollection of the actual exchange during such interviews and how information was elicited from the witness will allow for a better assessment of statement veracity in court. Two studies were designed to examine interviewers’ memory for a prior witness interview. Study One varied interviewer note-taking and type of subsequent interview report written by interviewers by including a sample of undergraduates and implementing a two-week delay between interview and recall. Study Two varied levels of interviewing experience in addition to report type and note-taking by comparing experienced police interviewers to a student sample. Participants interviewed a mock witness about a crime, while taking notes or not, and wrote an interview report two weeks later (Study One) or immediately after (Study Two). Interview reports were written either in a summarized format, which asked interviewers for a summary of everything that occurred during the interview, or verbatim format, which asked interviewers to record in transcript format the questions they asked and the witness’s responses. Interviews were videotaped and transcribed. Transcriptions were compared to interview reports to score for accuracy and omission of interview content. Results from both studies indicate that much interview information is lost between interview and report especially after a two-week delay. The majority of information reported by interviewers is accurate, although even interviewers who recalled information immediately after still reported a troubling amount of inaccurate information. Note-taking was found to increase accuracy and completeness of interviewer reports especially after a two week delay. Report type only influenced recall of interviewer questions. Experienced police interviewers were not any better at recalling a prior witness interview than student interviewers. Results emphasize the need to record witness interviews to allow for more accurate and complete interview reconstruction by interviewers, even if interview notes are available.