Randomized controlled trial of social interaction police training

Classified as probably not a substantial claim. May combine two claims.

Theory or research evidence is rarely integrated into police training, and when it is integrated, it is rarely rigorously evaluated.

“Yet, as Nagin and Telep (2017) recently noted, rarely is theory or research evidence integrated into police training, and even when it is, it is rarely subjected to rigorous scientific evaluation.”

• cites Nagin and Telep (2017) sourcePartly supported

  The Introduction and ‘Altering Procedurally Just Treatment’ section describe limited evidence on procedural justice training and few evaluations of its effects on officers’ field behavior (page numbers unavailable). They do not establish that theory or research evidence is rarely integrated into police training generally; the paper also describes several training programs that incorporated procedural justice principles. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

In RCTs, complex regression models with many covariates may inflate standard errors and bias estimates.

“In RCTs, complex regression models with a series of covariates are not only overly complicated but may, in fact, inflate standard errors and bias estimates (Berk et al., 2014;Freedman, 2008).”

• cites Berk et al., 2014Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Freedman, 2008 sourcePartly supported

  The Introduction reports that regression adjustment can bias estimates and that conventionally calculated standard errors can be severely biased; the Abstract says adjustment may worsen precision. However, the paper’s comparison uses one covariate and does not establish that having many covariates or model complexity causes these problems. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Research on training programs that focus on officers' social interaction skills has increased slightly in recent years.

“Research on training programs that focus on officers' social interaction skills has increased slightly in recent years (Hansson & Markström, 2014;Krameddine et al., 2013).”

• cites Hansson & Markström, 2014 sourcePartly supported

  The Background (page unavailable) discusses several studies of police training intended to improve interactions with people with mental illness, including communication skills and role-play. It does not say that research on these programmes has increased in recent years or quantify any increase; instead, it says the field has not gained much attention. (LLM, conf=0.78)

  Experimental — accuracy not yet measured
• cites Krameddine et al., 2013 sourcePartly supported

  The Introduction discusses prior research on police training in communication and de-escalation, and the Behavioral measures section identifies those skills as goals of this study’s training program. Neither passage says that research on such programs has increased in recent years, even slightly. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Nagin and Telep identified six programs in a review of the application of procedural justice theory to policing.

“Nagin and Telep (2017) identified six such programs in a review of the application of procedural justice theory to policing.”

• cites Nagin and Telep (2017) sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Vignettes have been used in various policing contexts to assess hypothetical officer behavior.

“Vignettes also have been previously used in a variety of policing contexts to assess hypothetical officer behavior (McLean, 2019;Nix, Picket, & Mitchell, 2019;Nix, Pickett, Wolfe, & Campbell, 2017;Phillips, 2009).”

• cites McLean, 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Nix, Picket, & Mitchell, 2019 sourcePartly supported

  The Experimental procedures section (page not provided) describes hypothetical vignettes covering several types of police-civilian encounters. But the Dependent variables section says the questions measured officers’ cognitive and emotional reactions, not their hypothetical behavior; the Conclusion distinguishes those reactions from behavioral responses. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Nix, Pickett, Wolfe, & Campbell, 2017 sourceNot checked — no full text of the source

May combine two claims.

Most studies reviewed by Nagin and Telep showed promising outcomes, but the training programs varied on several dimensions.

“Most of the studies reviewed by Nagin and Telep showed promising outcomes, but the training programs varied on a number of dimensions.”

• cites Nagin and Telep (2017) sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Difference-in-difference is a common strategy in economics and has been applied to criminological issues.

“This is a common strategy used in economics that also has been successfully applied to criminological issues (see Branas et al., 2011;Smith & Petrocelli, 2018).”

• cites Branas et al., 2011 sourcePartly supported

  The Statistical analyses section uses difference-in-differences models for crime and other safety outcomes, supporting its application to a criminological issue. The References section cites a work on natural and quasi-experiments in economics, but the paper does not establish that difference-in-differences is a common strategy in economics. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites Smith & Petrocelli, 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The randomized controlled trial found that the social interaction training improved officers' attitudes, with treatment-group officers giving higher priority to procedurally fair communication in a hypothetical officer-citizen encounter.

“Survey data and a series of difference-in-difference tests found that participating in the training program improved attitudes with treatment group officers placing higher priorities on procedurally-fair communication during a hypothetical officer-citizen encounter.”

• Also stated in: § Policy Implications · hedged

  “The results indicate that the training program was effective at increasing the priority officers place on procedurally fair communication but was not effective at reducing the number of reported use of force incidents for officers undergoing the training program.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Study officers accounted for 119 reportable use of force incidents in Fayetteville.

“Officers involved in the study -that is, officers who were randomly assigned to either treatment or control in February 2016 -accounted for 119 reportable use of force incidents.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The Tucson PD database contains 36 months of data from January 2016 to December 2018.

“The Tucson PD database contains 36 months of data (January 2016 to December 2018).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In the pre-test survey, 166 officers indicated they had been randomly assigned to receive T3 training and 228 officers indicated they were assigned to the control group.

“During the pre-test survey, 166 officers indicated that they had been randomly assigned to receive T3 training and 228 officers indicated that they were assigned to the control group.”

May combine two claims.

For maintaining self-control, the control group decreased from pre-test to post-test while the treatment group remained constant.

“With respect to maintaining self-control, the control group again decreased, but the treatment group remained constant, rather than increasing.”

May combine two claims.

Other effects in Tucson were small and non-significant.

“The other effects are again small and non-significant.”

May combine two claims.

The behavioral analysis did not support the survey analysis, with no significant differences in reportable use of force incidents in Fayetteville or Tucson attributable to T3 training.

“However, the behavioral analysis did not support the survey analysis with no significant differences detected in the number of reportable use of force incidents in Fayetteville or Tucson attributed to the T3 training program.”

May combine two claims.

For maintaining self-control, the low-dose treatment group had a substantial and statistically significant increase, while high-dose and control groups were similar.

“The high-dose treatment and control groups have very similar lines, while the low-dose treatment group experienced a substantial (and statistically significant) increase.”

May combine two claims.

Tucson PD saw reductions in reported uses of force in both treatment and control groups, but the treatment group's reduction was not statistically significant.

“Tucson PD saw reductions in the number of reported uses of force in both the treatment and control groups though the treatment group's reduction was not statistically significant.”

May combine two claims.

In the full sample, low-dose treatment had significant effects on procedural justice priorities (β = .12, p < .05) and maintaining self-control (β = .11, p < .05), and high-dose treatment had a significant effect on physical control priorities (β = -.13, p < .05).

“In the full sample model there is a significant treatment effect for procedural justice priorities (β = .12,p < .05)and maintaining self-control (β = .11,p < .05)when receiving the lower dosage of treatment, and there is a significant treatment effect for physical control priorities (β = -.13,p < .05)when receiving the higher dosage of treatment.”

May combine two claims.

In Fayetteville, low-dose treatment had a significant effect on procedural justice priorities (β = .21, p < .05) and high-dose treatment had a significant effect on physical control priorities (β = -.22, p < .05).

“In the Fayetteville sample, there are significant treatment effects for procedural justice priorities (β = .21,p < .05)when receiving the lower dosage and physical control priorities (β = -.22,p < .05)when receiving the higher dosage of treatment.”

May combine two claims.

Tucson's Midtown and West divisions showed no effect, while South and East divisions showed effects similar to Fayetteville.

“Finally, Tucson's Midtown and West divisions clearly experienced no effect, while Tucson's South and East divisions showed similar effects to those observed in Fayetteville.”

May combine two claims.

The effect of treatment on procedural justice is clearer than the other two effects, with five of seven district points on the hypothesized side of zero.

“First, the effect of treatment on procedural justice is clearer than the other two effects as five of the seven points are on the hypothesized side of the line at 0.0, which represents no effect.”

May combine two claims.

The low-dose treatment group showed a substantial and statistically significant increase in procedural justice priorities, while control and high-dose groups remained relatively stable.

“There is a clear demonstration that the low-dose treatment group saw a substantial (and statistically significant) increase in their procedural justice priorities scores while the control and high-dose treatment groups remained relatively stable.”

May combine two claims.

In Fayetteville, reportable use of force incidents were rare (n=119), and neither model showed differences between treatment (low- or high-dose) and control groups or from before to after training.

“Reportable use of force incidents are rare in Fayetteville (n=119 for the entire study period) and neither model suggests that there are any differences between the treatment (either low-dose -Figure 5 or high-dose -Figure 6) and control groups or from before to after the training program began.”

May combine two claims.

There was a significant treatment effect on procedural justice priorities in the full sample (β = .12, p < .05) and in the Fayetteville sample (β = .24, p < .01).

“There is a significant treatment effect on procedural justice priorities in both the full sample (β = .12,p < .05)and in the Fayetteville sample (β = .24,p < .01).”

May combine two claims.

In the district-level sensitivity analysis, only two significant effects were found: procedural justice priorities in Fayetteville's Central District (β = .33, p < .05) and physical control priorities in Fayetteville's Campbellton District (β = -.38, p < .05).

“Due to further reductions in sample size, it is not surprising that only two significant effects -the treatment effect on procedural justice priorities in Fayetteville's Central District (β = .33,p < .05)and the treatment effect on physical control priorities in Fayetteville's Campbellton District (β = -.38,p < .05)-were found.”

May combine two claims.

When broken down by dosage, low-dose treatment significantly improved procedural justice priorities and increased emphasis on maintaining self-control, while high-dose treatment significantly de-prioritized physical control.

“Second, when broken down by dosage, each outcome experienced a significant treatment effect: low-dose treatment resulted in improved procedural justice priorities and an increased emphasis on maintaining self-control, and high-dose treatment resulted in a de-prioritization of physical control.”

May combine two claims.

Tucson's East Division had point estimates for procedural justice priorities and maintaining self-control similar to Fayetteville, though neither was statistically significant.

“Specifically, Tucson's East Division had point estimates similar to Fayetteville for procedural justice priorities and maintaining selfcontrol though neither of these effects were statistically significant.”

May combine two claims.

The treatment group's decrease in physical control priorities was slightly greater but not significant.

“The treatment group decreased at a slightly greater rate, but again, this change is not significant in the difference-in-difference analysis.”

May combine two claims.

Tucson's South Division had point estimates for procedural justice priorities and physical control priorities similar to Fayetteville's districts, though neither was statistically significant.

“Similarly, Tucson's South Division had relatively similar point estimates for procedural justice priorities and physical control priorities compared to Fayetteville's districts, though again neither were statistically significant.”

May combine two claims.

For physical control priorities, low-dose and control groups had different overall levels but relatively stable trends.

“While there is a difference in the overall levels between the low-dose treatment group and the control group, their trends are relatively stable.”

May combine two claims.

In the full sample, procedural justice priorities decreased from pre-test to post-test for the control group but increased for the treatment group.

“Specifically, while procedural justice priorities decreased from the pre-test to the post-test for the control group, it increased for the treatment group.”

Classified as probably not a substantial claim.

Encounters between individuals are governed by exchange rituals that create order, including a mutual expectation that each person will show respect and regard for the other.

“Drawing on Goffman's (1956Goffman's ( , 1961) classic work on social interaction, Sykes and Clark (1975) asserted that encounters between individuals are governed by a set of exchange rituals that create order in interactions, including the mutual expectation that each individual will show respect and regard for the other.”

• cites Sykes and Clark (1975) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Officers expect a higher degree of deference than they feel they must reciprocate.

“That is, officers expect a higher degree of deference shown to them that they do not feel they have to reciprocate.”

• cites Sykes and Clark (1975) sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Police-citizen interactions are dynamic exchanges that can escalate into use of force or de-escalate into peaceful resolution.

“An interactionist approach to police use of force, however, conceptualizes officer-citizen interactions as dynamic exchanges that unfold with varying potential to escalate into a use of force event or de-escalate into a peaceful resolution (e.g., Alpert & Dunham, 2004;Sykes & Clark, 1975).”

• cites Alpert & Dunham, 2004 sourceNot checked — no full text of the source
• cites Sykes & Clark, 1975 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The rules for showing deference in police-citizen encounters are asymmetrical because officers have formal authority to enforce the law and maintain order.

“However, Sykes and Clark highlight that the rules for showing respect and regard, or deference, within police-citizen encounters is asymmetrical in nature given the officer's formal authority to enforce the law and maintain order.”

• cites Sykes and Clark (1975) sourceNot checked — no full text of the source

May combine two claims.

The study's findings provide evidence that social interaction training may improve police officer attitudes but perhaps not their behaviors.

“Our findings provide evidence that such training may improve police officer attitudes, but perhaps not behaviors.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The physical control priorities scale had two items, with alpha 0.60 and r=0.42.

“Finally, physical control priorities contained two items that related to the physical restraint of the suspect (α=0.60,r=0.42).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The control group contained more officers than originally assigned because officers who graduated from the academy or were on long-term leave during the assignment stage were defaulted to the control group for the survey analysis.

“Interestingly, more officers were in the control group than were originally assigned.This occurred due to the time gap between random assignment and the administration of surveys.”

• Also stated in: § Procedure and Sample · plain

  “Any officers who graduated from the academy or who were on long-term leave (e.g., military or maternity leave) during the assignment stage defaulted to the control group for the survey portion of the analysis.”

May combine two claims.

Control group officers were dismissed from daily roll call before T3 training was delivered to treatment officers, so there was no contamination in treatment delivery.

“Officers in the control group were dismissed from the daily roll call prior to the delivery of T3 training to the treatment group officers, so there was no contamination in treatment delivery.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Survey administration emphasized anonymity and told officers that honest feedback was necessary for a faithful evaluation of the training program.

“We emphasized the anonymous nature of the survey and that honest feedback was necessary for a faithful evaluation of the training program.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The procedural justice priorities scale had eight items and a Cronbach's alpha of 0.88.

“Procedural justice priorities contained eight items that related to communicating and building rapport with the subject (α=0.88).”

May combine two claims.

In the survey data, 114 officers (69% of the pre-test treatment group) remained in the treatment group at post-test, indicating meaningful attrition.

“A meaningful amount of attrition was observed in the survey data throughout the study with 114 officers (representing 69% of the pre-test survey treatment group) indicating they were in the treatment group at the time of the post-test survey.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Officers seek to maintain the authority afforded by asymmetric deference because it lets them control a situation and reach outcomes that serve their interests.

“Officers seek to maintain the authority afforded through this asymmetry because it allows them to control a situation-including the ability to question, command, and even physically coerce if necessaryand arrive at outcomes that serve their interest (Bittner, 1967;Muir, 1977;Van Maanen, 1978).”

• cites Bittner, 1967 sourceNot checked — no full text of the source
• cites Muir, 1977Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Van Maanen, 1978Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The T3 training is a practical application of Alpert and Dunham's asymmetric model.

“As a result, the training is a good example of a practical application of Alpert and Dunham's (2004) asymmetric model.”

• cites Alpert and Dunham (2004) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Officers were told that the survey was voluntary, results would be reported in aggregate, and only researchers would have access to the raw data.

“Officers were told that the survey was voluntary, all results would be reported in the aggregate, and that no one other than the researchers would have access to the raw data.”

May combine two claims.

At post-test, the only significant difference between treatment and control respondents was that treatment group members were slightly older and more educated.

“The only significant difference between the treatment and control group respondents at the post-test was in the age and education categories with the treatment group indicating that they were slightly older and more educated than the control group.”

May combine two claims.

Officers graduating or returning from long-term leave were not assigned unique identifiers and thus were excluded from the use of force analysis.

“Furthermore, officers graduating from the academy or returning from long-term leave would not have been assigned a unique identifier, and therefore, would not be included in the use of force analysis as a member of the treatment or control group.”

May combine two claims.

When citizens do not meet the asymmetrical deference expectation, officers try to reassert authority, generally by taking verbal control of the interaction.

“When citizens do not abide by this asymmetrical deference expectation, officers will attempt to reassert their authority, which Sykes and Brent (1980) found was generally accomplished by taking verbal control of an interaction.”

• cites Sykes and Brent (1980)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

At post-test, treatment and control respondents were similar on gender, race/ethnicity, experience, rank, and military service.

“Respondents in each group were similar on gender, race/ethnicity, experience, rank, and military service.”

May combine two claims.

In this study, correlations between age/education and outcomes were non-significant with point estimates below .1, so the age and education differences from attrition are unlikely to affect treatment effect estimates.

“For our study, these correlations were all non-significant with point estimates less than .1,suggesting that even though there were age and education differences due to attrition, it is unlikely these differences would impact the treatment effect estimates.”

May combine two claims.

The White et al. (2019) study is less likely to be affected by contamination but lacks the repetitive practice principle central to T3.

“Thus, the White et al. (2019) study is less likely to be affected by contamination but does not contain the repetitive practice principle that is a key to the T3 program.”

• cites White et al. (2019)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Reviewed studies varied in outcomes measured, with some using attitudinal measures and others using behavioral measures of officers' social interaction skills.

“Second, studies varied in the outcomes measured.Several studies included attitudinal measures as outcomes (Rosenbaum & Lawrence, 2013;Schaefer & Hughes;Skogan et al., 2015), while others relied on behavioral measures of officers' social interaction skills (Owens et al., 2016;Wheller et al., 2013).”

• cites Rosenbaum & Lawrence, 2013Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Schaefer & HughesNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Skogan et al., 2015 sourceNot checked — no full text of the source

May combine two claims.

Among the reviewed studies, four delivered one-time communication skills training to in-service officers with no future retraining, while two integrated communication skills training into the curriculum for new recruits.

“Specifically, four studies featured one-time delivery of communication skills training to in-service officers with no future re-training (Owens et al., 2016;Schaefer & Hughes, 2016;Skogan et al., 2015;Wheller et al., 2013), while the other two programs involved integrating communication skills training into the curriculum for new recruits (Robertson et al., 2014;Rosenbaum & Lawrence, 2013).”

• cites Owens et al., 2016Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Schaefer & Hughes, 2016Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Skogan et al., 2015 sourceNot checked — no full text of the source

May combine two claims.

An evaluation of a procedural justice training program found no significant differences in attitudes toward procedural justice or legitimacy, but videotape review suggested a treatment effect on officer performance, though the sample was too small for strong conclusions.

“Similarly, Rosenbaum and Lawrence (2013) evaluated a procedural justice training program and did not find significant differences in attitudes towards procedural justice or legitimacy.However, a review of videotapes of simulated officer-citizen interactions suggested a treatment effect on officer performance, though the sample of videotaped encounters was too small for strong conclusions.”

• cites Rosenbaum and Lawrence (2013)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Officers trained to improve interactions with sexual assault victims did not report changes in attitudes regarding sexual assault victims such as rape myth acceptance, but did show improvement in ratings of their interview skills during a simulated interview.

“Lonsway and colleagues (2001) noted that officers trained to improve interactions with victims of sexual assault did not report any changes in attitudes regarding sexual assault victims such as rape myth acceptance.However, officers involved in the training did see an improvement in ratings of their interview skills during a simulated interview with a sexual assault victim.”

• cites Lonsway et al. (2001) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

For the pre- and post-test survey data, the research team attended roll calls for every shift in each patrol district over a two-week period to ensure all patrol officers had an opportunity to be surveyed.

“For the pre-and post-test survey data, members of the research team worked with agency personnel at both departments to attend roll calls for every shift in each patrol district over a twoweek period to ensure that all patrol officers had an opportunity to be surveyed.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Blocked goals in police-citizen interactions can create an action-reaction chain that escalates into force and continues until one party acquiesces.

“Alpert and Dunham assert that this blockage creates an action-reaction chain that can escalate into force and continue until one party acquiesces.”

• cites Alpert and Dunham (2004) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Incidents involving both treatment and control officers indicate potential contamination, but contamination is virtually impossible to avoid in a long-term training program.

“This type of incident indicates a potential for contamination; however, in a study of a long-term training program, it is virtually impossible to avoid contamination in use of force incidents.”

May combine two claims.

At pre-test, treatment and control groups did not differ significantly on gender, age, years of service, or race.

“Still, a comparison of officers in the treatment and control groups at the pre-test revealed no significant differences in terms of gender, age, years of service, or race.”

May combine two claims.

Officers may not report attitudinal changes but may change their behaviors, or vice versa, making measurement of both attitudes and behaviors desirable in assessing training programs.

“Thus, it is possible that officers may not report attitudinal changes but will change their behaviors (or vice versa) in officer-citizen interactions making the measurement of both attitudes and behaviors desirable in assessing training programs.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The maintaining self-control scale had seven items and a Cronbach's alpha of 0.77.

“Maintaining self-control contained seven items that related to the officer remaining calm and thinking through his/her options (α=0.77).”

May combine two claims.

Treatment or control assignment was random, but dosage level was not randomly assigned to districts.

“Thus, the dosage-level was not randomly assigned to districts, but the assignment of treatment or control was random.”

Classified as probably not a substantial claim.

These suggestions rely on training officers to improve their social interaction skills.

“Notably, each of these suggestions relies on training officers to improve their social interaction skills.”

• AssumptionRests on an assumption: Traditional police training rarely teaches that listening, empathizing, and maintaining respect may be valuable social interaction skills that can help turn encounters away from force.

Probably not a substantial claim: may not pass S3 (load-bearing).

The pre-test survey covered 74% of officers randomly assigned to the treatment condition.

“Accordingly, we were able to survey 74% of officers randomly assigned to the treatment condition.”

May combine two claims.

Because the study was a program-level RCT, it can identify effects of the whole program but not of its individual components.

“due to the research design (a program-level RCT), we know only the effects of the larger program but not the effects of its individual parts.”

Classified as probably not a substantial claim. May combine two claims.

Differences in training success could be due to training delivery, departmental culture, workgroup environment, or other unmeasured factors.

“These findings could have been the result of differences in training delivery, departmental culture, the workgroup environment, or some other unmeasured factor.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The divergent attitudinal effects may reflect differences in the focus of training during the second half of the program.

“These divergent effects could reflect differences in the focus of the training for the second half of the program.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Prior work suggests de-escalation typically includes prioritizing communication, especially procedurally fair communication, and de-emphasizing physical control.

“However, prior work has suggested that it typically includes the prioritization of communication -especially procedurally-fair communication -and a de-emphasis on physical control (Todak, 2017;Todak & James, 2018).”

• cites Todak, 2017Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Todak & James, 2018 sourceNot checked — no full text of the source

May combine two claims.

The effect size for maintaining self-control in Fayetteville is consistent with the full model but is not significant due to reduced power from the smaller sample.

“Additionally, note that the effect size for maintaining self-control in the Fayetteville sample is consistent with the effect size in the full model, but is not significant due to the reduced power of the smaller sample size.”

Classified as probably not a substantial claim. May combine two claims.

Newey-West standard errors were used to handle autocorrelation, and the Cumby-Huizinga test ensured correct lag estimation.

“Newey-West standard errors were used to handle autocorrelation due to the timeseries nature of the data, and the Cumby-Huizinga test was used to ensure that the correct lag was estimated for each model (Cumby & Huizinga, 1992;Linden, 2015).”

• cites Cumby & Huizinga, 1992 sourceNot checked — no full text of the source
• cites Linden, 2015 sourceNot checked — no full text of the source

May combine two claims.

Additional officers experienced career changes during the observation period, causing more attrition.

“We also had a number of officers experience similar career changes during the observation period, which resulted in a few more cases of attrition.”

May combine two claims.

The study found no evidence that attrited officers differed from retained officers except that retained officers were slightly younger and more likely to have an Associate's degree than a Bachelor's degree.

“we have no evidence that the officers were different from those that remained in the treatment and control group aside from the latter being slightly younger and more likely to have an Associate's degree rather than a Bachelor's degree.”

May combine two claims.

Contamination of treatment and control officers occurred in 10% of use of force incidents in Fayetteville and 18% in Tucson.

“Contamination of treatment and control group officers occurred in 10% of use of force incidents in Fayetteville PD and 18% of incidents in Tucson PD.”

May combine two claims.

The possibility that training effects erode over time is a troubling indicator, because the first three months' content was consistent across both groups.

“though this in itself may be a troubling indicator that the effects of training erode over time as the content of the first three months was consistent across both groups.”

• AssumptionRests on an assumption: The content of the first three months of training was consistent across both treatment groups.

May combine two claims.

In Tucson, none of the effect sizes exceeded |0.04|, so lack of power is unlikely to explain the absence of effects.

“In the Tucson sample, however, none of the effect sizes is greater than |0.04|, so a lack of power is unlikely to be the cause of the absence of effects among officers in that agency.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Police training is not a one-size-fits-all solution, and T3 is no exception.

“In short, police training is not a one size fits all solution, and the T3 program is no exception.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Contamination was the second limitation of the study.

“Our second limitation is the effect of contamination on the results.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Literature on police subculture suggests officer attitudes and behavior can vary by immediate small work group.

“This is not entirely surprising given the established literature on police subculture that suggests officer attitudes and behavior can vary by their immediate, small work group (e.g., shift or squad; see Klinger, 1999;Ingram, Paoline, & Terrill, 2018;Ingram, Terrill, & Paoline, 2018).”

• cites Klinger, 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Ingram, Paoline, & Terrill, 2018 sourceNot checked — no full text of the source
• cites Ingram, Terrill, & Paoline, 2018 sourceNot checked — no full text of the source

May combine two claims.

The non-significant effects for maintaining self-control and physical control priorities in Fayetteville are in the expected direction and may reflect a power issue due to the small sample (108 officers).

“Consider that the non-significant effects for maintaining self-control and physical control priorities in the Fayetteville sample is in the expected direction and may simply reflect a power issue given the smaller effect size compared to the other outcomes (there are only 108 officers in the entire Fayetteville post-test data).”

Classified as probably not a substantial claim. May combine two claims.

The simultaneous reduction in Tucson may be due to contamination or may be spurious because of unobserved agency changes at the time training started.

“This simultaneous reduction may be due to the effects of contamination or may be spurious due to changes that we were unaware of occurring in the agency at the time the training program started.”

• AssumptionRests on an assumption: T3 training encourages officers to prioritize delaying physical contact.

May combine two claims.

Although previous models suggested no treatment effects in Tucson, the district-level analysis suggests some Tucson patrol divisions experienced effects.

“The previous models suggested that Tucson did not experience any treatment effects.However, this analysis suggests that some of Tucson's patrol divisions experienced effects.”

May combine two claims.

The data do not indicate who was first on scene, so the study cannot determine whether contaminated incidents resulted from T3 officers delaying or control officers delaying.

“Unfortunately, the data we have do not indicate who was first on scene to see if contaminated incidents were the result of T3 officers on scene delaying encounters or control group officers delaying encounters.”

May combine two claims.

Randomization at the shift level was not realistic because there were limited shifts and use of force opportunities vary by shift.

“There were limited numbers of shifts in the two police departments and the opportunity for use of force incidents necessarily vary according to these shifts (e.g., use of force incidents are less likely at 8 am on a Monday morning than at 9 pm on a Friday night), so randomization at the shift level was not realistic.”

• AssumptionRests on an assumption: Use of force incidents are less likely at 8 am on Monday than at 9 pm on Friday.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Moving from theory to a realistic training program implementable across agencies is extremely difficult but achievable with appropriate resources.

“moving from the realm of theory to a realistic training program that can be implemented in agencies across the country is extremely difficult, but certainly achievable with the appropriate resources.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The research team observed training once a month throughout the program.

“Members of the research team observed training once a month for the duration of the program.”

May combine two claims.

Only patrol officers were included in the study, so officers in other roles were excluded from random assignment and analysis.

“Only patrol officers were included in the study, so officers assigned to other roles would not be included in the random assignment and are also not included in the analysis.”

May combine two claims.

An incident involving use of force does not by itself indicate failure to de-escalate, but on average the treatment group should show fewer incidents if training improved de-escalation.

“Thus, an incident resulting in a use of force is not indicative of a failure in that incident or by that officer, but on average, we would expect to see fewer use of force incidents across the entire treatment group if they were more successful at de-escalation as a result of the training.”

• AssumptionThe authors state this as an assumption: Over the three years of use of force data, treatment and control groups had the same average number of opportunities to de-escalate.

May combine two claims.

Cronbach's alpha is related to the number of items in a scale, so the low alpha for physical control priorities is not unexpected.

“Cronbach's alpha is related to the number of items in a scale(Carmines & Zeller, 1979), so the low alpha value for physical control priorities is not unexpected.”

• cites Carmines & Zeller, 1979 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Anonymity would decrease the amount of socially desirable answers received, though social desirability may still be an issue.

“This is not to suggest that social desirability may not still be an issue, but rather that anonymity would decrease the amount of socially desirable answers that we received.”

• AssumptionRests on an assumption: Anonymity decreases the amount of socially desirable answers received.

Probably not a substantial claim: may not pass S3 (load-bearing).

A study by White et al. (2019) uses block randomization to randomize entire squads rather than individual officers to avoid contamination.

“a study currently being conducted byWhite et al. (2019) uses block randomization to randomize entire squads rather than individual officers to avoid contamination.”

• cites White et al. (2019)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Integration of training into recruit programs may involve more repetitive training than one-time delivery, but it is limited because communication skills training is not repeated after recruit training.

“While the integration of training into the recruit program may have involved more repetitive training than the onetime delivery programs, they are still notably limited in that once recruit training is completed, the communication skills training is no longer repeated.”