Modelling self reported confessions and cooperation with police interrogators

This study modelled self-reported confessions and cooperation with police interrogators. Incarcerated men (N = 100) were interviewed about their most recent police interrogation. A logistic regression analysis was performed to predict confession decision using nine predictors: Humanitarian Style, Legal Advice, Interrogation Length, Perception of Evidence, Age, Previous Conviction, Number of Convictions, Offence Seriousness, and Attitude Toward Police. A model containing Perception of Evidence, Humanitarian Style, Previous Convictions, Number of Convictions, and Legal Advice predicted confession decision 79% of the time (versus 60% for a base model). A multiple regression analysis, using the same predictors, revealed that Humanitarian Style, Previous Convictions, and Number of Convictions accounted for 29% of the variance in self-reported cooperation. The implications of the findings for interrogations practices are discussed.