Jeff Kukucka

Do Confessions Taint Perceptions of Handwriting Evidence? An Empirical Test of the Forensic Confirmation Bias

Citing classic psychological research and a smattering of recent studies, Kassin, Dror, and Kukucka (2013) proposed the operation of a forensic confirmation bias, whereby preexisting expectations guide the evaluation of forensic evidence in a self-verifying manner. In a series of studies, we tested the hypothesis that knowing that a defendant had confessed would taint peoples evaluations of handwriting evidence relative to those not so informed. In Study 1, participants who read a case summary in which the defendant had previously confessed were more likely to erroneously conclude that handwriting samples from the defendant and perpetrator were authored by the same person, and were more likely to judge the defendant guilty, compared with those in a no-confession control group. Study 2 replicated and extended these findings using a within-subjects design in which participants rated the same samples both before and after reading a case summary. These findings underscore recent critiques of the forensic sciences as subject to bias, and suggest the value of insulating forensic examiners from contextual information.

Does video recording alter the behavior of police during interrogation? A mock crime-and-investigation study.

A field study conducted in a midsized city police department examined whether video recording alters the process of interrogation. Sixty-one investigators inspected a staged crime scene and interrogated a male mock suspect in sessions that were surreptitiously recorded. By random assignment, half the suspects had committed the mock crime; the other half were innocent. Half the police participants were informed that the sessions were being recorded; half were not. Coding of the interrogations revealed the use of several common tactics designed to get suspects to confess. Importantly, police in the camera-informed condition were less likely than those in the -uninformed condition to use minimization tactics and marginally less likely to use maximization tactics. They were also perceived by suspects-who were all uninformed of the camera manipulation-as trying less hard to elicit a confession. Unanticipated results indicated that camera-informed police were better able to discriminate between guilty and innocent suspects in their judgments and behavior. The results as a whole indicate that video recording can affect the process of interrogation-notably, by inhibiting the use of certain tactics. It remains to be seen whether these findings generalize to longer and more consequential sessions and whether the camera-induced differences found are to be judged as favorable or unfavorable.

Cognitive bias in forensic mental health assessment: Evaluator beliefs about its nature and scope.

Decision-making of mental health professionals is influenced by irrelevant information (e.g., Murrie, Boccaccini, Guarnera, & Rufino, 2013). However, the extent to which mental health evaluators acknowledge the existence of bias, recognize it, and understand the need to guard against it, is unknown. To formally assess beliefs about the scope and nature of cognitive bias, we surveyed 1,099 mental health professionals who conduct forensic evaluations for the courts or other tribunals (and compared these results with a companion survey of 403 forensic examiners, reported in Kukucka, Kassin, Zapf, & Dror, 2017). Most evaluators expressed concern over cognitive bias but held an incorrect view that mere willpower can reduce bias. Evidence was also found for a bias blind spot (Pronin, Lin, & Ross, 2002), with more evaluators acknowledging bias in their peers’ judgments than in their own. Evaluators who had received training about bias were more likely to acknowledge cognitive bias as a cause for concern, whereas evaluators with more experience were less likely to acknowledge cognitive bias as a cause for concern in forensic evaluation as well as in their own judgments. Training efforts should highlight the bias blind spot and the fallibility of introspection or conscious effort as a means of reducing bias. In addition, policies and procedural guidance should be developed in regard to best cognitive practices in forensic evaluations.

The Journey or the Destination? Disentangling Process and Outcome in Forensic Identification

Jeff Kukucka Towson University, Towson, Maryland Recently, this journal published an archival study by Langenburg, Bochet, and Ford (2014) that explored the impact of two sources of contextual bias—namely, exposure to case information and interaction with investigators—on the identification decisions of certified latent print examiners. This paper is commendable as an effort to examine the effects of bias on real-world casework. However, it is also limited by several methodological oversights and ultimately perpetuates a common misconception about contextual bias. Below, I briefly summarize this study before elaborating my concerns with its conclusions. The study included a sample of 885 criminal cases in which law enforcement had solicited the help of examiners from an accredited latent fingerprint laboratory. First, the authors categorized each case as either high, moderate, or low in (a) the degree of interaction between the examiner and investigators, and (b) the amount of case information available to the examiner. Then, they compared identification decisions from 466 cases that were deemed “low” in both dimensions against 18 cases that were deemed “high” in both, and found that the rates of identification were nearly identical (22% and 21%, respectively). From this, the authors concluded that the aforementioned contextual factors had no appreciable biasing effect on examiners’ identification decisions. This conclusion is misleading insofar as it is derived solely from the raw number of identification decisions, and not from any index of their validity, origin, or strength. As others have noted (e.g., Dror 2009), we must be careful to distinguish between the decision-making process and the decision outcome. Even if bias fails to change the latter, it may nonetheless affect the former. For example, if the forensic evidence suggests a certain conclusion and contextual factors encourage the same conclusion, the examiner’s decision will not change, but his or her confidence in this decision may increase as a result. Unfortunately, the data from this study tell us nothing about the process by which examiners arrived at their decisions, and thus it remains unknown whether and how context impacted them. Even if we ignore the importance of process as the authors have done, their conclusion with respect to outcomes is also unsound. Their inference that context had no effect on identification decisions is based on a comparison of cases with low versus high degrees of bias. This assumes that bias affects decisions in only one direction, which is a dubious assumption. Instead, contextual factors may bias examiners toward identification in some cases, and bias them toward Address correspondence to Jeff Kukucka, Towson University, Towson, MD 21252. E-mail: jkukucka@towson.edu

Police reports of mock suspect interrogations: A test of accuracy and perception.

A 2-phased experiment assessed the accuracy and completeness of police reports on mock interrogations and their effects on people’s perceptions. In Phase 1, 16 experienced officers investigated a mock crime scene, interrogated 2 innocent suspects—1 described by the experimenter as more suspicious than the other—and filed an incident report. All 32 sessions were covertly recorded; the recordings were later used to assess the reports. In Phase 2, 96 lay participants were presented with a brief summary of the case and then either read 1 police report, read 1 verbatim interrogation transcript, or listened to an audiotape of a session. Results showed that (a) Police and suspects diverged in their perceptions of the interrogations; (b) Police committed frequent errors of omission in their reports, understating their use of confrontation, maximization, leniency, and false evidence; and (c) Phase 2 participants who read a police report, compared to those who read a verbatim transcript, perceived the process as less pressure-filled and were more likely to misjudge suspects as guilty. These findings are limited by the brevity and low-stakes nature of the task and by the fact that no significant effects were obtained for our suspicion manipulation, suggesting a need for more research. Limitations notwithstanding, this study adds to a growing empirical literature indicating the need for a requirement that all suspect interrogations be electronically recorded. To provide a more objective and accurate account of what transpired, this study also suggests the benefit of producing verbatim transcripts.