Citation

Kim DY, Suk HW, Cho S, Hwang H, Lee JH. Pers. Individ. Dif. 2014; 60(Suppl): S66.

Copyright

(Copyright © 2014, Elsevier Publishing)

DOI

10.1016/j.paid.2013.07.289

PMID

unavailable

Abstract

This study aimed to investigate predicting factors for detecting guilty individuals who committed a mock crime using a functional data analysis, to reduce an individual variation of psychophysiological data and to find weighted integration of function. Fifty-four participants (30 male, 24 female) chose either guilty condition (stealing a wallet) or innocent condition (non-crime). Skin conductance level (SCL), temperature, and pulse were measured during conducting a concealed information test. A functional logistic regression indicated that SCL and pulse were predictive variables for detecting guilty. Specifically, participants were more likely to be guilty if they showed a higher degree in SCL responding to both crime-relevant and irrelevant questions and a higher increase of pulse after the crime-relevant questions. These results suggest that SCL and pulse may predict whether a suspect is guilty or innocent and that a functional data analysis could provide more accurate prediction by excluding individual variation.