BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Talks.cam//talks.cam.ac.uk// X-WR-CALNAME:Talks.cam BEGIN:VEVENT SUMMARY:A Graphical Model Approach to Eyewitness Identification Data - Ama nda Luby (Carnegie Mellon University) DTSTART:20160929T083000Z DTEND:20160929T091500Z UID:TALK67703@talks.cam.ac.uk CONTACT:INI IT DESCRIPTION:Although eyewitness identification is generally regarded as re latively inaccurate among cognitive psychologists and other experts\, test imony from eyewitnesses continues to be prolific in the court system today . There is great interest among psychologists and the criminal justice sys tem to reform eyewitness identification procedures to make the outcomes as accurate as possible. There has been a recent push to adopt Receiver Oper ating Characteristic (ROC) curve methodology to analyze lineup procedures\ , but has not been universally accepted in the field. This work addresses some of the shortcomings of the ROC approach and proposes an analytical ap proach based on log-linear models as an alternative method to evaluate lin eup procedures. \;Unlike approaches that emphasize correct and incorre ct identifications and rejections\, our log-linear model approach can dist inguish among all possible outcomes and allows for a more complete underst anding of the variables at work during a lineup task. \;Due to the hig h-dimensional nature of the resulting model\, representing the results thr ough a dependence graph leads to a deeper understanding of conditional dep endencies and causal relationships between variables involved. We believe that graphical models have been under-utilized in the field\, and demonstr ate their utility for not only broader statistical insights\, but as an in tuitive way to communicate complex relationships between variables to prac titioners. We find that log-linear models can incorporate more information than previous approaches\, and provide flexibility needed for data of thi s nature LOCATION:Seminar Room 1\, Newton Institute END:VEVENT END:VCALENDAR