Mercedes La Voy has always been interested in studying memory, and has recently developed a specific interest in false memories. These interests led her to the lab of Dr. Steyvers and his work on recognition memory. She says that her research experience—including designing the experiment, counterbalancing the variables, and interpreting the results—have helped her to think critically and to solve problems. In Fall 2005, Mercedes moved on to the next step in her education, entering the Washington State University Clinical Psychology Ph.D. Program. When she’s away from her work, Mercedes enjoys going to the beach, cooking, and spending time with friends and family.

Eyewitness testimony relies on recognition memory of an event. This study examines the relationship between similarity of stimuli and recognition memory as it applies to eyewitness testimony. Participants viewed a study set followed by a test set, in which the similarity of alternatives was continuously manipulated. Items were presented in a lineup format, either side by side or serially. The results show a bias toward the central tendency of similarity, or the picture that is in the middle range of similarity. This causes accuracy to decline under certain conditions. This study has implications for police lineup situations, and suggests ways to address potential bias due to similarity relations among the set of alternatives.

Mark Steyvers School of Social Sciences

Mercedes La Voy’s work is a great example of how basic research in human memory can have real-world applications, such as in police lineups. Typical lineups involve individuals chosen to be somewhat similar to each other to force the witness to be confident about their potential identification. It is not clear however how similarity relations among the alternatives affect identification. Mercedes La Voy investigated this question by testing participants with lineups involving pictures of outdoor scenes, one of which was shown in an earlier study phase. Mercedes showed that participants often based their recognition-memory decisions not on their actual memory but on the similarity relations between alternatives—test items that looked most like other items in the lineup were often chosen as targets.