Legal systems use divergent and contradictory definitions of credibility, illustrated by recent examples from the apex courts of the United States, Canada, and Norway. Combining concepts from legal analysis and legal psychology presents a way forward. First, it is noted that psychology has largely studied intuitive initial judgments of credibility, whereas legal evaluations of credibility are generally deliberative and considered. Combining both approaches with the theoretical distinction in legal psychology between the credibility of a person and the reliability of their testimony, a preliminary model is proposed that describes the ways in which a reliability assessment can diverge from a credibility assessment. Because reliability also consists of an assessment, it is therefore subject to the weaknesses of human reasoning and can be mistaken, or, in the model's terms, not in accordance with the factual truth. The model links thinking about credibility and reliability with the relevant literature from legal psychology.
Many local jurisdictions use pretrial risk assessments to guide release and supervision decisions. Risk-informed approaches rely on factors identified through administrative records, limiting comprehensive understanding of defendants' risk, needs, and protective factors. Few studies have examined how defendants perceive factors that might contribute to pretrial success. We conducted semistructured interviews with 73 defendants across 10 counties to explore how defendants perceived their risks, needs, and protective factors during pretrial supervision. We asked participants to think about times during pretrial when things were or were not going well, especially vis-& agrave;-vis their ability to meet pretrial conditions and pretrial failure (i.e., failure to appear in court or a new criminal offense). We identified themes using abductive coding, which involved deductive broad coding and inductive narrow coding. Code frequency guided identification of prominent themes. Participants' legal status and pretrial requirements reflected their preexisting risks and needs but also served to exacerbate risks and needs. Participants identified behavioral health challenges as a prominent need, which included subclinical symptoms like anxiety about the uncertainty of the pretrial period. Prominent protective factors included social support from family, friends, and pretrial staff. Findings revealed a more dynamic interplay between risks, needs, and protective factors than currently captured in pretrial risk assessment tools. Future prospective research designs may expand conceptualization of risks and needs and explore the utility of protective factors in risk management. Whether protective factors could serve as intervention targets during pretrial supervision remains to be seen but represents an important avenue for future research.
Although risk assessment tools in South Korea are formally administered to inform security-related dispositions rather than to serve as sentencing factors, debate persists over whether their visibility to judges during sentencing deliberations may nevertheless induce negative bias against defendants, potentially leading to harsher sentences. Yet empirical evidence based on actual court rulings remains scarce, particularly outside Anglo-American jurisdictions. This study examined whether the implementation and the scores of two widely used tools in South Korea-the Korean Offender Risk Assessment System-General and the Korean version of the Psychopathy Checklist-Revised-affect sentencing severity when statutory sentencing factors are controlled. A total of 735 first-instance judgments for murder (including attempted murder and parricide) between 2021 and 2023 were analyzed. Risk assessment data for both the Korean Offender Risk Assessment System-General and the Psychopathy Checklist-Revised were available in 116 cases. Sentence length served as the dependent variable, with control variables encompassing guideline-based sentencing ranges, crime classification, special and general sentencing factors, victim-related factors, defendant-related factors, and aggression factors. Regression models tested whether the mere implementation of an assessment functioned as a biasing signal to judges (Hypothesis 1) and whether higher assessment scores independently predicted harsher sentences (Hypothesis 2). Results showed that neither the implementation of assessments nor the their scores significantly influenced sentencing severity once statutory factors were accounted for. These findings suggest that risk assessments, although visible to judges during sentencing, do not systematically amplify punitive outcomes in South Korean murder cases, providing real-world evidence relevant to ongoing policy debates on the role of risk assessments in judicial decision-making.
Research on youth testimony in child sexual abuse (CSA) cases has rarely examined age and gender differences in depth. The current study investigated how attorneys' questions and victims' responses vary by victim age (children vs. adolescents) and gender (boys vs. girls). The sample included 70 CSA trial transcripts (35 male, 35 female victims; M-age = 13.64, SD = 2.33; range = 8-17 years) from Arizona and California. Boys and girls were matched on case characteristics (age at testimony, relationship to perpetrator, number of perpetrators, abuse frequency and severity). Transcripts were coded for question type, response productivity, and response reluctance. Prosecutors asked more open-ended questions (consistent with best practice recommendations) more often, whereas defense attorneys asked more option-posing and leading questions. Age differences emerged, with attorneys adjusting their questioning tactics based on whether the witness was a child (8-12) or an adolescent (13-17). For example, defense attorneys asked proportionally more leading questions of adolescents compared with children. Although gender did not have an impact on question type, gender differences in responses emerged. Boys were more productive and less reluctant to prosecutors than defense attorneys. Overall, findings suggest that attorneys consider age more than gender when formulating questions, but boys and girls may still vary in how they respond to such questions. These variations, in turn, may affect how victims are perceived and ultimately the outcome of a case, highlighting the need for ongoing research and training on the nuanced and complex ways that characteristics in victims shape how CSA criminal cases unfold.
The majority decisions in two recent Supreme Court of Canada cases (R. v. Barton, 2019; R. v. Chouhan, 2021) recommended that jury instructions be used as a method to counteract potential racial bias in the courtroom. This study sought to test the effectiveness of these recommended instructions, including general antibias instructions provided by the National Judicial Institute and the more specialized race-based instructions recommended by the Court, compared with control instructions that did not mention race, in trials involving racialized accused and victims. Community-member participants (N = 1,091) recruited from Prolific Academic watched an audiovisual trial presentation involving a White/Black/Indigenous accused and victim in which the judge provided one of three types of instructions (control/general race-based/general and specialized race-based), then made verdict decisions. Results reflect recent experimental jury research trends toward nonsignificant differences as a function of accused race and/or increased harshness toward White accused. However, we observed two counterintuitive effects of the modified instructions. The general antibias instructions invoked increased leniency toward an accused with an Indigenous victim, and the general antibias plus specialized race-based instructions invoked harshness in conditions with White accused and racialized victims. These results suggest that more work is needed to investigate these instructions before their implementation and that the Court may wish to explore other avenues for racial bias reduction.
Indigenous men with criminal records experience compounded exclusion in the labor market due to intersecting stigmas of race and criminal history, a phenomenon explained by the Double Jeopardy Hypothesis. This article explores these dynamics through three experiments examining hiring biases against Indigenous applicants compared with their Black and White counterparts. Results reveal that Indigenous men with criminal records are disproportionately underselected for employment opportunities. Techniques like name anonymization, revised language describing criminal records, and Ban-the-Box policies were found to be ineffective for Indigenous applicants, contrasting with their success for other groups. These findings highlight how compounded disadvantages uniquely exclude Indigenous men with criminal records from equitable labor market access, reinforcing systemic barriers. This research contributes to the limited empirical literature on Indigenous peoples with criminal records and employment and underscores the need for tailored equity-focused interventions informed by the Double Jeopardy Hypothesis.
Globally, relationship separation with subsequent litigation of a parenting dispute represents a high-risk period for family violence (FV). The Federal Circuit and Family Court of Australia piloted Lighthouse, a new, innovative triage process beginning with confidential screening of FV risks to guide triage and subsequent case management during the family court process. The current study adopted an exploratory design to examine whether there were observable differences in court events and safety markers for parenting matters involved in the Lighthouse pilot (n = 335) and those operating as usual (nonpilot; n = 639). Matters were matched on equivalent allegations of risk indicated by litigants on mandatory court forms at court entry. We used court-linked records across a minimum of 2 years postfiling to examine differences in (a) timing and type of court events (e.g., lag to and use of hearing types), (b) safety-oriented court actions (e.g., court-appointed children's lawyer), and (c) mode and timing of matter finalization. Our findings show that the implementation of Lighthouse, including new confidential standardized risk screening at court entry, was associated with more rapid court responses, characterized by short lag to and greater use of interim and chambers hearings. Increased safety-oriented court actions were also evident in Lighthouse pilot registries, along with a pronounced focus on children's well-being. The change aligns with the intention of recent family law policy reforms in Australia to prioritize the safety of families involved in family law proceedings. Findings, although preliminary, may inform policy, practice, and funding settings across family law and FV sectors.
This exploratory study examined five potential problems with attorneys' questions and 5- to 9-year-old children's responses to before/after questions in 156 child sexual abuse trials. Two problems have received some attention in prior research: the extent to which children were asked to backward sequence through before questions, and order of mention, which considers whether the order in which two events were mentioned in the question matched the order in which the events purportedly occurred. We found that 43% of before/after questions included before, and 55% of questions asking about two events violated order of mention. We also examined three issues involving potential ambiguity that have been overlooked: grain size ambiguity, in which the temporal interval to which before/after refer is unspecified; reference time ambiguity, in which before/after could refer to a prior time or the time when the question is asked; and focal ambiguity, in which the before/after question could be asking about sequence or occurrence. Eighty-four percent of questions exhibited grain size ambiguity, 18% of questions exhibited reference time ambiguity and 40% of questions exhibited focal ambiguity. Overall, 78% of before/after questions were option-posing and 84% of children's responses to those questions were unelaborated, which makes disambiguation of responses difficult. Even when questions were unambiguous, we found suggestive evidence that words used to specify grain size or reference time created other difficulties for children. The results highlight the need for additional research examining difficulties in before/after questions, and for educating attorneys and others who interview children about subtle difficulties.
Beyond a reasonable doubt (BRD) is the standard of proof required for conviction in criminal trials in Anglo-American legal systems. We examined how lay people (N = 220) define BRD in qualitative terms, and quantitatively measured their perceptions of existing definitions of this standard of proof as well as the grounds they believe create "a reasonable doubt." The qualitative data suggest that although individuals used relevant synonyms for BRD (i.e., degree of doubt present, certainty or probability required, amount and strength of proof provided, and alternative explanations beyond the defendant being guilty), the concept is vague in individuals' minds, has variable meanings across people, and its interpretation can increase or reduce the threshold for conviction. There was no significant difference between the sure and firmly convinced definitions of BRD (used in England and Wales, and the U.S. Federal courts, respectively) in terms of self-reported understanding, confidence in application, and judged likelihood of wrongful conviction or wrongful acquittal. However, for each of these two definitions, greater self-reported understanding was significantly positively associated with greater confidence in applying the standard of proof, and both self-reported understanding and confidence were significantly negatively associated with judged likelihood of wrongful conviction and wrongful acquittal. Finally, a weak prosecution case and strong defense case were each rated as creating significantly greater grounds for a reasonable doubt than either victim and offender behavior or the punitive impact of a guilty verdict. The present findings have implications for jury instructions.
Juvenile diversion programs aim to redirect youth from formal legal processing and toward community-based interventions. Despite the benefits of diversion, there is little research characterizing the U.S. state, district, territory, and freely associated state policy that governs diversion programming. We extracted statutes, administrative codes, and regulations pertaining to juvenile diversion across all 50 U.S. states, the District of Columbia, and eight territories and freely associated states. We analyzed these using a blended inductive-deductive qualitative coding approach. We describe themes and examine their presence and absence across policies. Themes spanned several domains, ranging from general diversion policy considerations and goals (e.g., mandating diversion programming), requirements for screening and referral to diversion programming (e.g., defining eligibility criteria), requirements for diversion programming and implementation (e.g., operationalizing diversion strategies and interventions), and requirements for program oversight and evaluation. Policies varied considerably. Findings highlight opportunities for future research to understand how local jurisdictions interpret and implement policies, align research and policymaking to support rigorous evaluation of existing policies, and consider how existing evidence can inform the development and implementation of policies.
Research indicates that showing evaluators videorecorded lineup procedures in cases involving administrator suggestion fails to decrease belief in suggestively obtained identifications. We theorized that this is because viewing a suggestive lineup procedure has dual effects on evaluations of eyewitness identifications: a suggestiveness-correction effect that decreases belief in suggestively obtained identifications and a counteracting prejudicial effect that increases belief in suggestively obtained identifications. In two preregistered experiments, undergraduate students (N = 785, Experiment 1) and jury-eligible community members (N = 782, Experiment 2) viewed a video of a mistaken eyewitness's testimony or viewed the testimony video followed by a video of the lineup procedure. In half of the lineup procedure videos, the administrator made suggestive remarks, and in the other half, the administrator behaved neutrally. Evaluators assessed the eyewitness's accuracy and the suggestiveness of the lineup procedure. Mediation analyses revealed the predicted effects: Although viewing a videorecorded lineup procedure enabled suggestiveness correction (i.e., decreased belief in suggestively obtained identifications via increased perceptions of procedural suggestiveness), this suggestiveness correction was counteracted by a prejudicial effect (i.e., increased belief in suggestively obtained identifications). These dual effects occurred regardless of whether evaluators were told that the lineup was conducted single- or double-blind. Our findings help explain why prior research has failed to find corrective effects of showing evaluators videorecorded lineup procedures in cases involving administrator suggestion and suggest that alternative remedies are needed.
Legislators often propose bills in honor of a victim following a tragic event involving public outrage (e.g., AMBER Alerts, Megan's Law). Such bills and subsequent laws that involve a victim's name and story are known as eponymous legislation. Legal commentators have raised concerns that this legislation garners unwarranted emotional support from the public. We conducted three experiments to empirically test whether eponymous legislation received greater public support than legislation without a victim's name and story. In Study 1 and Study 2, participants read a proposed bill that either included a victim's name and narrative (eponymous condition) or not (control) and rated their approval of the bill, cognitive and affective responses, and reasoning. In both of these studies, participants were significantly more likely to support the eponymous bill than the control bill. Study 2 further suggested the effect was mediated by sympathy. In Study 3, we manipulated the victim's name and narrative separately and found that narrative, not name, drove increased support. Narrative also increased sympathy in Study 3, but sympathy did not mediate the effect of narrative on bill support. Findings support legal commentators' concerns about eponymous legislation. Legislators need to be cautious when proposing such bills to ensure that they are not ineffective or harmful despite receiving increased public approval.
Although sleep disruption carries implications for the evidentiary value of legal statements and confessions, there is only limited evidence and little systematic attention to this topic. In response, we reviewed existing research regarding the impact of sleep-related fatigue on psychological functioning in the context of investigative and legal processes, alongside its policy implications. First, we outline the sleep-wake mechanisms and the consequences of their disruption for psychological functioning. Second, we summarize the prevalence and severity of sleep disruption among populations interacting with the criminal justice system. Third, we review relevant U.S. law on the admissibility of statements under sleep deprivation, with a particular focus on confessions. Fourth, we tie the discussed processes and relevant evidence together to posit three dose-dependent consequences of sleep-related fatigue for the validity of statements provided by subjects involved in legal processes-(a) unreliable recollections, (b) increased duress, and (c) poor reasoning. Finally, we integrate these findings by considering how sleep disruption impacts the evidentiary basis of statements, while providing the first evidence-based guidelines on what level of sleep-related fatigue may be influential for statements in evidence. We conclude by suggesting practical steps to document its influence and advance the understanding of the impacts that sleep disruption has on legal processes.
When sexual offenses are tried in criminal court, it is common for multiple complainants (i.e., alleged victims who are witnesses) to testify against the same defendant in a single trial, known as "trial joinder." When complainants are joined at trial, often the law requires that they be evaluated independently, but based on a psychological literature of context effects, multiple complainants testifying in the same trial violates this assumption of independence. Across two studies, this research examined judicial outcomes of sexual offenses when trials were joined. In Study 1, we explored the real-world relationship between the number of complainants at trial and verdict using a dataset of 4,237 Canadian criminal cases of child sexual abuse. Descriptively, the data showed a nonlinear pattern whereby conviction rate increased, peaked, and then decreased as the number of complainants at trial increased. Using an experimental paradigm in Study 2, we replicated the same nonlinear pattern predicting mock jurors' perceptions of guilt; however, the effect was not observed when predicting complainant credibility. Consistent with the joinder effect literature, the current research suggests that trial joinder has a biasing effect on perceptions of guilt in sexual assault cases. Extending the literature, the current research also suggests the pattern of guilt bias may be more complex than initially thought. The research is aligned, however, in that trial joinder risks judicial fairness by altering perceptions of guilt due to irrelevant factors. Implications for justice policy are discussed.
The repeated-suspect effect is well documented with adult eyewitnesses (Godfrey & Clark, 2010; Lawson & Dysart, 2014) but has not been explored with child eyewitnesses. Given this, we examined whether children are vulnerable to this effect (Study 1). We also tested whether the multiple independent lineups (MIL) technique could produce an effective reflector variable (i.e., the number of lineups the suspect is identified in) in helping assess the reliability of the face identification following a prior exposure to a suspect's face (Study 2). In both studies, child participants (aged 6-11 years) witnessed a live event and then received a showup or not. The next day, they completed a lineup task (either a single face lineup or the MIL task). Results suggest that children show evidence of the repeated-suspect effect (Study 1), and the MIL technique could produce an effective reflector variable in situations involving an eyewitness seeing a suspect in a prior instance (Study 2). Across both studies, no significant differences for age emerged, suggesting that school-age children are vulnerable to the repeated-suspect effect and that the MIL technique may function similarly for this age group of children. Overall, we echo the call that researchers have made to police to only show an eyewitness a suspect on one occasion.
Children interviewed in alleged maltreatment cases sometimes provide avoidant (e.g., "I don't remember") or denial (e.g., "Nothing happened.") responses. In the current study, we conducted a sequential dialogue analysis to investigate (a) how interviewers at the Dutch child protection (Safe Home) reacted to such utterances and (b) how these reactions in turn affected children's subsequent responses. We examined 57 denial and 281 avoidance statements across 68 Safe Home interviews with alleged child maltreatment victims (38 according to the National Institute for Child Health and Human Development protocol and 30 without a protocol). Interviewers oftentimes rejected avoidant (63%, 178/281) and denial statements (49%, 28/57), irrespective of National Institute for Child Health and Human Development protocol training. Yet, the likelihood of children to respond productively (i.e., by providing new details) was seven times higher when their initial reluctances were met with acceptance by the interviewer rather than rejection. Based on these results, we provide recommendations for child interviewing training.
Understanding the level of confidence an eyewitness intends to communicate is important for criminal justice. However, eyewitnesses vary considerably in what they mean-even when using the same phrase-and evaluators vary in their interpretations too. Building on that work, we show how the criminal justice system could leverage where agreement does exist. We produced a tool (a lexicon) for helping eyewitnesses and legal actors to understand how people on average interpret confidence. The tool could be used to guide the communication of eyewitness confidence. Three replications demonstrate that the structure of the produced lexicon is relatively, though not perfectly, consistent across samples and English-speaking cultures. Study 5 shows that eyewitnesses' numeric translations of phrases from the lexicon when they spontaneously produce them result in similar distributions of meaning as those we obtained from non-eyewitness interpreters in Studies 1-4. Furthermore, the lexicon elicited a similar confidence-accuracy relationship as asking for confidence on a 0%-100% scale or in eyewitnesses' own words. While one might argue that this suggests reliance on the eyewitness's own words is sufficient, we argue that research-based evaluations of own-words confidence-here and in prior research-likely overestimate the confidence-accuracy relationship because researchers systematically interpret own-words confidence. In practice, individuals make unaided (and variable) interpretations of verbal confidence judgments. Therefore, our results suggest a lexicon could produce a superior confidence-accuracy relationship in practice.
