Applying Deception Detection to True and False Confessions: A Novel Approach to Conducting Experiments in Legal Psychology

Background & Context

False confessions are a significant contributor to wrongful convictions, with DNA exoneration cases revealing that a substantial proportion of wrongfully convicted defendants had confessed to crimes they did not commit. This troubling phenomenon not only inflicts suffering on the wrongfully convicted and their families, but also imposes economic costs on communities and undermines the legitimacy of the legal system.

While prior research had identified structural interrogation practices and individual characteristics that increase the risk of false confessions, knowledge on how to identify true versus false confessions remained minimal and undeveloped. The challenge of distinguishing genuine from fabricated confessions represents a critical gap in the intersection of deception detection research and legal psychology.

Deception detection research had not been fully applied to confessions, and paralinguistic cues such as pauses and pitch changes had never been empirically tested. This study sought to address this gap by examining whether training in paralinguistic cues could enhance the ability to differentiate true from false confessions, while simultaneously pioneering a novel online experimental methodology.

Research Design & Methodology

The research employed a two-stage experimental design. In the first phase, true and false confessions to guilt-inducing or shame-evoking events were obtained through an adaptation of the established Kassin, Meissner, & Norwick (2005) method, which involved collecting actual confessions from participants about genuine and fabricated transgressions.

In the second phase, participants (n = 63) were randomly assigned to receive either training in paralinguistic cues to deception or a placebo training condition. The paralinguistic training focused on vocal characteristics such as pauses, pitch variations, and other speech patterns theoretically associated with deception.

Subsequently, participants rated the perceived veracity of six confessions and indicated their confidence level in each judgment. A secondary methodological goal was to assess the viability of using an online experimental survey as an alternative to traditional laboratory-based experiments in legal psychology.

Results & Key Findings

Consistent with prior deception detection research, the results indicated that accuracy rates were around chance levels and that confidence was not related to accuracy. This fundamental finding replicated decades of research showing that people—even trained individuals—struggle to detect deception reliably.

The training manipulation revealed a troubling pattern:

• The accuracy rate of the paralinguistic training group did not significantly differ from the control group, possibly due to random effects in treatment assignment
• The paralinguistic training group indicated significantly higher confidence than the control group
• Training increased observers' confidence without improving their actual ability to distinguish true from false confessions

The core finding—training in paralinguistic cues increased confidence but not accuracy—highlights a dangerous outcome: practitioners who receive deception detection training may become more certain in their judgments while remaining no more accurate than untrained individuals, potentially leading to more assertive but equally erroneous decisions in legal contexts.

Discussion & Significance

This study makes an important contribution to the false confession literature by demonstrating that paralinguistic cue training does not enhance discrimination between genuine and fabricated confessions. The confidence-accuracy dissociation observed here mirrors findings from broader deception detection research, reinforcing the conclusion that subjective confidence is a poor indicator of judgment accuracy in detecting deception.

The findings have profound implications for interrogation training programs. Many law enforcement training protocols emphasize behavioral and paralinguistic indicators of deception, yet this research suggests such training may create overconfident investigators who are no better—and potentially more dangerous—than untrained observers. The increased confidence without corresponding accuracy gains can lead to stronger convictions about suspect guilt, potentially intensifying interrogation pressure on innocent individuals.

The study also advances methodological innovation by testing online survey methods for legal psychology experiments. Assessing the use of an online experimental survey as an alternative to traditional laboratory experiments addresses practical constraints in conducting confessions research and offers a scalable approach for future investigations.

Limitations & Considerations

The study acknowledges several limitations. Random effects in treatment assignment may have influenced the paralinguistic training group results, suggesting that methodological variations in how participants were assigned to conditions could have affected outcomes. The relatively small sample size (n = 63 evaluators) may have limited statistical power to detect training effects if they existed.

The confessions examined involved relatively minor transgressions rather than serious criminal acts, which may limit generalizability to actual forensic contexts where stakes are considerably higher. Additionally, the use of an online methodology, while innovative, introduces questions about experimental control and participant engagement compared to traditional laboratory settings. The specific paralinguistic training content and duration may not have been optimal, leaving open the possibility that different training approaches might yield different results.

Practical Applications

For agents in the legal system—law enforcement, legal decision-makers, and psychologists—awareness of cues to guilt or innocence in confessions, along with risk factors for false confessions, is essential to prevent false convictions and preserve the legitimacy of the legal system. This research provides empirical evidence that current approaches to training evaluators to detect deceptive confessions are ineffective and potentially counterproductive.

The practical takeaway for interrogators, prosecutors, and judges is clear: subjective assessments of confession veracity—even after training—should not be trusted. The findings support policy reforms emphasizing objective corroboration of confessions through independent evidence, mandatory recording of entire interrogations, and caution about relying on paralinguistic or behavioral "cues" to truthfulness. Training programs that promise enhanced deception detection abilities should be critically evaluated, as they may produce false confidence rather than genuine expertise.