Test whether research dissemination format affects undergraduate criminal justice students’ knowledge acquisition and perceptions of body-worn cameras (BWCs), and whether effects vary by evidence-based practice awareness or academic reading experience. A randomized online survey experiment assigned 1,044 students to control, non-referenced, citation-based, or executive-summary formats presenting the same mixed BWC findings. Binary logistic regression, OLS regression, average marginal effects, and second-difference tests were estimated with multiple imputation. All dissemination formats increased knowledge acquisition relative to the control: non-referenced OR = 4.03, citation-based OR = 4.40, executive-summary OR = 3.84. Effects were larger among students with prior academic reading experience. Citation-based (b = − 0.152) and executive-summary (b = − 0.149) formats predicted less favorable BWC perceptions; moderation was not observed for BWC perceptions. Research dissemination can improve understanding of mixed evidence, but effects depend on prior research exposure and do not necessarily increase policy support.
Evidence-based policing (EBP) is widely recognized as essential in modern democracies, yet it often remains an aspiration rather than a daily operational reality. Although research on policing interventions has grown significantly, institutionalizing these findings into sustainable practice remains a formidable challenge. In her 2025 Joan McCord Lecture, Lum argues that fieldwork is the indispensable link between scientific inquiry and the operationalization of evidence-based policing. Through collaborative partnerships, researchers and practitioners generate shared knowledge—a “special sauce”—necessary to move beyond ad hoc implementation toward institutionalizing evidence-based practices. This practical wisdom enables both groups to engage the five elements of evidence-based policing (Lum Koper, 2017): generating evidence, translating findings, building institutional receptivity, operationalizing strategies, and institutionalizing practices as “business as usual.” Without mutually respected fieldwork, criminology will struggle to make its work genuinely meaningful to the criminal justice field.
Despite numerous evaluations assessing public perceptions of specialized police equipment, few studies have examined how citizens perceive officers’ use of everyday tools, such as mobile devices, in practice. We address this gap by assessing citizens’ perceptions of different note-taking modalities during a routine police encounter. Using a 3 × 2 between-subjects experimental vignette design, we test the effects of officers’ note-taking modality (mobile device, notepad, or no device) and the degree of explanation provided during the encounter (no explanation or a brief explanation of intended actions) on citizens’ perceptions of the police and their willingness to engage with law enforcement in the future. Using a mobile device for note-taking was perceived similarly to the no-device condition, regardless of whether a brief explanation was provided. In contrast, using a notepad was perceived more favorably and generated a greater likelihood of future engagement with police than the no-device condition. Police agencies may benefit from informing officers about the potential harmful effects of mobile phone use during face-to-face encounters. Traditional pen-and-paper methods may produce more favorable attitudes toward police, though further research is needed to verify whether these findings are observed in real-world settings.
The current study measures whether AI-generated transcripts of body-worn camera footage accurately capture who is speaking and what is said, compared to human-edited transcripts. We analyzed AI and human-edited transcripts from 176 body-worn camera videos (73 incidents, approximately 23 h of footage) from a large U.S. police department. We used multiple text-similarity metrics to compare speaker extraction, speaker matching, and overall document similarity. AI-generated transcripts captured much of the same lexical content as human-edited transcripts but performed noticeably worse on identifying and attributing speakers, especially in complex, multi-speaker encounters. AI transcription appears suitable for large-scale content search and descriptive text analysis but should not replace human transcription for attribution-sensitive uses. The findings underscore the value of hybrid human-AI workflows and the need for clear policies distinguishing low-stakes from high-stakes uses of AI-generated transcripts in policing.
This study reports the one-year impact of a social media campaign aimed at deterring drunk driving in El Paso, Texas. Every Friday evening, a police department and community group uploaded booking photos of people arrested for driving while intoxicated (DWI) to social media accounts. Alcohol-related vehicle crashes were evaluated to assess the campaign’s deterrent effect. Bayesian structural time-series analysis was conducted with 55 months of vehicle crash data. To isolate the campaign’s effect, several control-series were incorporated, such as public transportation usage and alcohol sales, as well as crash trends in other Texas cities. Overall analyses suggest the campaign was associated with a modest reduction in alcohol-related crashes. Subgroup analyses by driver age revealed strong evidence the campaign was associated with reductions among drivers ages 31 years and older. Drivers between the ages 21 to 30 showed no evidence of a reduction. The findings underscore both the potential and legal limits of social media campaigns as tools for crime prevention.
Experimentally test whether retention bonus conditions impact law enforcement officers’ willingness to extend their employment. Officers from a stratified-random sample of sixteen agencies responded to a vignette that randomized the amount of a one-time bonus and obligation to repay if the officer leaves within 3-years. OLS models estimate bonus conditions on willingness to remain, including moderating effects of pre-experiment retention intentions. Most were likely to stay for an additional period if they received the bonus (87.4
Assess potential spillover effects Operation Cookie Monster (OCM), a law-enforcement takedown of an online market selling stolen account credentials, had on users and administrators of an untargeted hacker forum, a different platform type where similar goods are shared. Treating OCM as a natural experiment, we used data from two sources: observations of forum advertisements and an extension of a prior field experiment. We analyzed changes in posting frequency by user rank, credential types, and administrator content removal before OCM to six months after. Activity declined among experienced non-paying users and most credential types, while increasing among paying users and posts advertising social media and cloud accounts. Administrator moderation intensified after OCM, becoming more selective and retroactive, before gradually declining. OCM produced temporary diffusion of benefits alongside evidence of crime displacement. Enforcement actions can affect untargeted related platforms, but offender adaptation limits longer-term impact.
To evaluate the effectiveness of therapeutic communities (TCs) in terms of reducing post-prison recidivism for individuals with substance use disorder. Propensity score matching was used to create statistically balanced groups of individuals released from (a) minimum- and medium-security TCs in Illinois and (b) non-TC prisons. Logistic regression models then estimated the effect of TC treatment on several types of post-release arrest and reincarceration. Treatment at both the minimum- and medium-security TCs led to reductions in general arrests and reincarceration at one and three years post-release. The medium-security TC also had significant effects on each arrest subtype measured (i.e., violent, drug, and domestic violence). For all outcomes, the effects of treatment at a TC diminished over time. The results align with prior literature showing that TCs can reduce recidivism. However, more research is needed on the moderating and mediating mechanisms that account for these effects.
Police chiefs are among the most visible and consequential leaders in local government. Despite their importance, little empirical research examines which leadership credentials the public values most when evaluating candidates for the chief of police position. This study addresses that gap by examining how the public weighs competing credentials when evaluating police chief applicants, with particular attention to evidence-based policing and the implementation of professional standards. The study employs a mixed-methods design, combining a discrete-choice experiment with a conjoint design and thematic analysis of qualitative data. Emphasis on professional standards was the most influential attribute in respondents’ selection decisions. Population size served and commitment to evidence-based policing tied for second, followed by years of experience. The applicant’s current position was the least influential attribute. Respondents preferred applicants with external credentials in professional standards and evidence-based policing over those who helped implement those practices but lacked such credentials. The public places significant value on external credentials in professional standardsand evidence-based policing when evaluating police chief candidates.
Examine whether participation in a cognitive-behavioral Batterer Intervention Program (BIP) is associated with reductions in criminal thinking among individuals convicted of intimate partner violence (IPV), and whether such reductions persist after program completion. Using a prospective longitudinal design, we tracked changes in criminal thinking across repeated measurements spanning a two-year period among 624 court-mandated men and women from five Midwestern counties (1,717 total observations). Criminal thinking was assessed using a modified version of the Texas Christian University Criminal Thinking Scales (TCU-CTS). Growth curve modeling was used to estimate within-person change during and after program participation. Criminal thinking declined significantly during program participation but stabilized after program completion, with no evidence of reversal. Cognitive change achieved during BIP participation may persist beyond program termination, underscoring the dynamic and modifiable nature of criminal thinking. Findings are based on a non-randomized, single-state sample, limiting causal inference and generalizability. Replication with randomized samples across diverse settings is needed to strengthen causal claims and inform policy on BIP effectiveness.
This study examines how point-of-view (POV) in short-form crime media shapes punitive attitudes and whether this relationship depends on viewers’ emotional responses during exposure. Data come from a randomized laboratory experiment (N = 60) in which participants viewed crime videos varying in POV (first-, second-, and third-person). Expressed fear was measured in real time using automated facial expression analysis, and punitive attitudes were assessed as an ordinal measure of preferred punishment. Ordinal logistic regression models estimated the independent and interactive effects of POV and fear. Second-person POV increased support for more severe punishment. Expressed fear moderated this relationship: under first- and third-person conditions, higher fear was associated with greater punitiveness, whereas under second-person POV, higher fear was associated with reduced support for the most extreme sanction. Emotional responses shape punitive attitudes in a context-dependent manner conditioned by media presentation. Further implications for research and policy are discussed.
This study examined how a mock defendant’s exposure to genetic and environmental risk factors shapes public sentencing judgements using an experimental survey design. 1,668 survey participants were randomly assigned to one of eight vignette conditions. Participants received one of four possible background scenarios describing the defendant’s risk exposure (no-risk, genetic risk, environmental risk, or both risks) and were informed about the defendant’s most recent offense (burglary or armed robbery). Participants were more likely to recommend incarceration for defendants described as having been exposed to genetic risk factors, but less likely to do so for those exposed to environmental risk factors, compared to those exposed to no risks. A moderation analysis found no conditioning effect of crime type. Results suggest that genetic risk may be viewed by the public as an aggravating factor, while environmental risk may be seen as a mitigating factor. Implications are discussed.
This study examined the effectiveness of a novel voice message intervention in reducing shoplifting in supermarket stores. Six different messages were introduced to 13 experimental stores, with three stores as controls. The messages were delivered via a human-shaped device resembling a police officer, equipped with a motion sensor and a speaker. The study spanned two two-month periods, with different messages in each period. The loss rates of seven targeted items were recorded at the end of each period and compared with those of the same month of the previous year. Effects varied across product categories, with nonuniform reduction across all items. The message appealing to moral norms and conscience was the most effective in preventing shoplifting. These findings show that moral appeals made via voice messages can reduce shoplifting in certain contexts. This approach offers a low-cost and easily implementable alternative to conventional store layout modifications.
This study tests whether headline themes from news coverage of police use-of-force influence general public approval of legally reasonable police use-of-force. In a between-subjects survey experiment (N = 2,495), a nationally stratified sample of U.S. adults was randomly assigned to view four real newspaper headlines representing one of four thematic conditions or a neutral control. The headline themes were derived from a structural topic model of three decades of news coverage. Respondents then completed three items measuring approval of police use-of-force. Treatment effects were estimated using Bayesian model averaging across four ordinal regression models. Protest Reform and Race–themed headlines reduced approval by about 7
This study provides the first context-specific meta-analytic synthesis of job satisfaction among South Korean police officers in a collectivist policing system and examines the patterns of associations between job satisfaction and its correlates. A total of 185 empirical studies conducted in Korea were systematically reviewed, yielding 644 independent effect sizes. Random-effects meta-analytic models were applied to estimate average associations between job satisfaction and its correlates. In the Korean police, job satisfaction shows strong positive associations with variables related to intrinsic motivation and work conditions and rewards, including job motivation (r = .579), job itself (r = .534), empowerment (r = .500), and promotion (r = .587). Among organizational factors, relational and support-related organizational variables showed relatively large associations with job satisfaction, including social capital (r = .468), communication (r = .466), and perceived organizational support (r = .448). By contrast, variables that are negatively associated with job satisfaction, such as job stress (r = − .215), exhibit relatively smaller effect sizes, and demographic variables report very small effect sizes overall (e.g., age r = .005, education r = − .026). Job satisfaction is also strongly associated with key work attitudes and behavioral indicators including organizational commitment (r = .587), organizational citizenship behavior (r = .576), and turnover intention (r = − .482). The findings suggest that job satisfaction in Korean policing is systematically associated with individual motivational factors, work conditions and rewards, and relational organizational conditions. This study contributes to criminological research by providing a contextualized empirical synthesis of the correlates of job satisfaction in policing.
Objective: To assess how threat imminence shapes public judgments of deadly force used by police and by civilians, and whether the public believes the imminence requirement currently applies to police. Methods: We fielded a nationally representative US survey experiment (N = 1,183) that manipulated whether a suspect posed an imminent or non-imminent threat and whether a civilian or police officer faced the decision to use deadly force. Respondents evaluated support for shooting and reported beliefs about relevant legal rules. Results: Imminence significantly affected support for deadly force. Although respondents were more supportive of police than civilians using deadly force across conditions, they generally disapproved of shooting a non-imminent threat. Nearly 95% of respondents incorrectly believed that Supreme Court precedent restricts police deadly force to imminent threats. Conclusions: The public supports the imminence requirement for police and largely believes it already exists in law. This "imminence divide" between public expectations and legal standards may undermine police legitimacy when officers act lawfully but contrary to public judgment.
Injectable naltrexone (XR-NTX) is an FDA-approved opioid agonist treatment (OAT) for the prevention of relapse to opioid abuse. While there is a growing consensus about its efficacy in community settings, less is known about the effect on justice-involved people, particularly in the time immediately after incarceration. The unique needs of this population warrant additional inquiry, especially regarding the prevention of mortality, as they are at an increased risk of death during the reentry period. The current study follows a wait-list randomized control design and was conducted within a large state-level correctional system located in the Eastern United States. Eligible individuals could receive no-cost XR-NTX before their release from incarceration or six months after release. Analyses followed an intention-to-treat (ITT) framework. A policy of providing injectable naltrexone did not significantly increase or decrease mortality among people leaving prison, though lower mortality rates and longer time to failure were observed among those people starting OAT treatment while incarcerated. The evidence on XR-NTX remains mixed concerning justice-related outcomes (e.g., recidivism), as well as mortality among people leaving prison. The elevated risk- and needs-profile of this population, considered in conjunction with the existing evidence base on MAT, suggests that this remains an important avenue for future research.
This study examines whether police activities—including self-initiated stops, frisks, searches, arrests, and summonses—differ between uniformed and plainclothes officers in New York City and whether the effect of wearing a uniform on these activities varies by assignment type (detective, bureau, SCU/NST [Street Crime Unit/Neighborhood Safety Team] period). Using Stop-Question-Frisk (SQF) data from 2017 to 2024 (N = 86,028), the effects are estimated with three average treatment effect estimators: Inverse Probability Weighted Regression Adjustment (IPWRA), Nearest Neighbor Matching (NNM), and Propensity Score Matching (PSM). Multivariate logistic regression analyses are also conducted, including interaction terms between police uniform and assignment type. The IPWRA results indicate that, compared to plainclothes officers, uniformed officers are significantly less likely to initiate stops but significantly more likely to make arrests. Multivariate logistic regression with interaction terms indicates that uniform alone significantly decreases the odds of initiating stops while increasing the odds of conducting searches. In contrast, its effects on frisks and summonses depend on assignment to specific bureaus. Police uniforms affect police activities, suggesting that officers in uniform behave differently from those not in uniform.
This study examined a hybrid treatment approach that integrates a traditional batterer intervention program (BIP) with a restorative justice (RJ) component, Circles of Peace (CP), to address domestic violence (DV) crimes in cases of intimate partner violence (IPV). In Utah, 274 offenders (both male and female), court-mandated to treatment for misdemeanor DV offenses in IPV cases, were randomly assigned to treatment. A parallel randomized controlled trial compared a standard 16-week BIP (BIP-only; n = 138) with a 12-week BIP followed by four weeks of CP (BIP-plus-CP; n = 136). Poisson regression models were used to examine group differences in new offenses and crime severity scores. Offenders assigned to BIP-plus-CP demonstrated statistically significant reductions in both the number of new offenses and the severity of those offenses. Combining BIP and RJ components showed greater effectiveness in reducing recidivism and the severity of offenses among DV offenders.
