Police confidence in lie detection: an assessment of crime types, Miranda and interview techniques

Abstract

PurposeExperimental research on lie detection has indicated that accuracy rates hover around chance but that police are significantly better in detecting deception in “high” stakes rather than “low” stakes situations. This paper has three objectives: to compare confidence levels in lie detection for property crime and violent crime detectives; to compare differences in confidence levels for custodial and noncustodial interviews; and to evaluate the relationship between interrogation techniques and lie detection confidence.Design/methodology/approachThe study uses self‐report data from a sample of Texas police detectives.FindingsThe results of this study show that property crime detectives are significantly more confident in their lie detection ability than are violent crime detectives. The results also highlight the fact that police detectives are significantly less confident in their lie detection abilities when the suspect has been provided his or her Miranda warnings.Originality/valueThe study highlights the disparity in findings derived from self‐reported data and experimental studies on veracity judgments and the need to account for contextual factors that ultimately impact the ecological validity of this research.