False Confession Identification Accuracy in Groups of College Students Kassondra Bertulis1, Dr. Kento Yasuhara1, Dr. Charles Honts2 1University of New Haven and 2Boise State University Introduction Method Results Discussion Erroneous convictions are a serious u Participants Next, Chi Square analyses were run to compare This further illustrates that training, weather problem for the criminal justice system. This formal (Kassin & Fong, 1999), or informal Ø The participants of this study were college students who had taken classes in which false is demonstrated by the Innocence Project, Confessions were discussed and students who had exposer such as in a class setting does not students from the University of New Haven. increase accuracy when identifying false which has exonerated 312 people through the Students ranged in academic major. not in there hit rates (correctly identifying a true confessions. use of DNA evidence since it’s inception u Apparatus/Materials confession as true) and false alarm rates Future research should examine different groups (“Understanding the Cause,” n.d.). False (incorrectly identifying a false confession as true). Ø The survey engine, Qualtrics, was used to involved in the Criminal Justice systems ability statements and full false confessions account to identify false confessions such as defense design and distribute the survey to the The alpha level for this experiment was set at p=. for 25% of those exonerated, pointing to a attorneys, prosecuting attorneys, and judges. students. The survey could be taken on any 05 major flaw in the criminal justice system: that Additionally, different training methods and computer at any time. The survey contained Hit Rate (True Identified as True) topics should be compared to see if there is in people are confessing to crimes that they did eight short video confessions to crimes and fact a reliable way to increase accuracy in false not commit. Further, these innocent people Percentage 0% 25% 50% 75% 100% Total confession identification. questions accompanied each video. are being convicted and incarcerated. Such Taken Class u Procedures 0 5 5 13 5 28 issues seem counterintuitive, as one would Acknowledgments Ø The students received an email that was think those involved in the criminal justice Not Taken 1 10 24 20 10 65 generated through the Qualtrics survey Class system, such as police officers, prosecutors, engine. In the email there was an Total I would like to thank Dr. Kento Yasuhara for and defense attorneys, would be experts at 1 15 29 33 15 93 individualized link for the survey. When the advising my project and being extremely detecting false confessions. To test this The Pearson Chi Square significance for Hit student clicked on the link a new window supportive during this learning process. I would assumption, Kassin, Meissner, and Norwick Rate is p=.378 opened displaying the consent form. Once also like to acknowledge Carol Withers and the (2005) examined both true and false they indicated their acceptance of the consent False Alarm (False Identified as True) SURF team for making this project possible. confession identification accuracy between form the first video was displayed. A series of Percentage 0% 25% 50% 75% 100% Total Additionally I would like to thank Dr. Charles college students and police officers. The eight videos were displayed to the Honts for providing the confession videos for results indicated that college students were Taken 4 9 7 6 2 28 participants one at a time. With each video Class this research. generally more accurate than police officers, was a question asking them if they thought Not Taken 5 23 17 13 7 65 however, only slightly better than chance. Class the confession in the video was true or false. References Kassin and Fong (1999) examined whether Total They were also asked to rank their confidence 9 32 24 19 9 93 training college students in how to detect of their judgment on a seven point Likert verbal and non-verbal cues of deception made The Pearson Chi Square significance for False scale. Once all eight videos and their Kassin, S. M., & Fong, C. T. (1999). “I’m them more accurate in identifying a false Alarm is p=.872 questions were completed the students were Innocent!”: Effects of training on judgments denials that were made during interrogations. thanked for their participation and given the of truth and deception in the interrogation It was found that the participants who took No significant differences were found between choice to click on a new link to enter their room. Law and Human Behavior, 23, 499– part in the training were less accurate then the students who had taken a class and those who demographic information for a gift card 516. untrained group. had not in either hit rates or false alarms. drawing. This page was completely separate Kassin, S. M., Meissner, C. A., & Norwick, R. J. The current study examined the from their responses. Discussion (2005). 'I'd Know a False Confession if I Saw difference in accuracy rates for identifying Results One': A Comparative Study of College true and false confessions between students Students and Police Investigators. Law and There is no evidence to suggest that previously who have taken classes where false Human Behavior, 29(2), 211-227. doi: being exposed to knowledge about false confessions were discussed and those who 10.1007/s10979-005-2416-9 Overall, the students were more accurate when confessions in class made students any more have not. Understanding the Cause: False confessions. identifying true confessions rather then false accurate when identifying confessions. This Retrieved from: http:// confessions with an average accuracy rate of finding is consistent with the findings of the www.innocenceproject.org/ understand/ 61.7% overall for true confessions and 54.1% previous literature. overall for false confessions. False-Confessions.php .