The role of emotional and cultural intelligences on police leaders’ perceptions of reform

Abstract

Police leaders’ support for policing reform is an important consideration because this population serves in some ways as gatekeepers of their respective organizations, setting organizational priorities and operationalizing change. Using survey data from 731 US police leaders who attended the FBI National Academy program in 2023, we evaluate the salience of emotional intelligence and cultural intelligence as key leadership contexts. Specifically, this study examines whether emotional and cultural intelligence relates to participants’ support for civilian oversight boards, and the belief that de-escalation and implicit bias training have a positive impact on policing. Controlling for other factors, the results indicate that emotional and cultural intelligence are significantly related to reform attitudes. Notably, we control for participants’ perceptions that systemic racism poses a problem in criminal justice and find that views of systemic racism also remain significant with most outcomes. Implications for police leadership training and policing reform are discussed.

INTRODUCTION

Arthur Niederhoffer’s (1967) seminal work Behind the Shield laid the groundwork for policing scholars to begin to understand police cynicism. Niederhoffer essentially presents a strain-based approach, arguing ‘that police officers experiencing the conflicting norms and values during the transition from the traditional to the emerging professional model of policing were faced with anomic conditions, which generated frustration and subsequent cynicism’ (Hickman et al. 2004: 1). The current era is not dissimilar to the timeframe in which Niederhoffer explored cynicism in that serious racial divisions marked both eras. Following the widespread civil unrest and serious outcries for police reform in 2020, it is not surprising that scholars are once again concerned about how police officers will respond to public demands for reform. Such reform-oriented attitudes of police leaders should be of particular concern because this group is responsible for setting their department’s training agendas, establishing organizational policies, priorities, and goals, as well as shaping the organizational culture toward reforms. Furthermore, this group is responsible for engaging in dialogue with citizens and other stakeholders about potential reforms. Leaders who are more open and adaptive to changes and who can navigate the increasingly complex environment of policing may be more inclined to support reform in formal and informal ways.

Society expects modern police leaders to be able to handle complex social relationships both internally, within their organizational spaces, and externally, with the community. Managing these complex environments requires leaders to demonstrate good emotional control and be culturally competent. The broader organizational literature provides two concepts that help address these demands: emotional intelligence (EI) and cultural intelligence (CQ) (Mayer and Salovey 1997; Earley 2002). Founded within the theory of multiple intelligences (Snyderman and Rothman 1987), EI and CQ are two forms of social intelligence (Gardner 1993; Mayer and Salovey 1997; Earley 2002), and they are considered essential to good leadership and the ability to make effective decisions in culturally diverse situations (Goleman 1998; Livermore 2015; Solomon and Steyn 2017).

Emotionally intelligent leaders tend to demonstrate greater empathy, have better social skills and self-awareness, and better self-regulate their emotions (Goleman 1998). Similarly, cultural intelligence is another aspect of social intelligence that extends emotional intelligence to focus on how someone responds to intercultural settings (Earley 2002). In the current study, we explore how American police leaders’ EI and CQ relate to their support for three common police reform-oriented efforts in the USA: support for civilian oversight, support for de-escalation training, and support for implicit bias training. We suggest police leaders with better EI and CQ may be more inclined to support these common policing reforms. We hypothesize that these two factors will positively relate to police leaders’ support for reforms, and we test these relationships in multivariate contexts that control for other potentially confounding factors.

LITERATURE REVIEW

Civilian oversight boards

The Final Report of the President’s Task Force on 21st Century Policing, Recommendation 2.8, addresses the call for law enforcement agencies to adopt at least some form of civilian oversight to help strengthen trust with their community, and the Task Force notes the need for additional research (Task Force 2015: 26). Civilian oversight boards likely first emerged in the late 1960s (Clark and Friedman 2020). These groups act in part as a community barometer, and they may help inject a level of cultural competence or awareness to police decision-making (Aplin 2022). The most recent Law Enforcement Management and Administrative Statistics (LEMAS) census of US police and sheriff’s agencies (2020) shows that only 16 per cent of police and 10 per cent of sheriff’s departments have some form of civilian oversight board, but notably, there have been increases since the 2016 census, which showed 11.3 per cent and 6.4 per cent, respectively (Brooks 2023; Goodison and Brooks 2023). As one might suspect, large agencies are more likely to have a civilian oversight board when compared to those departments serving rural populations. The LEMAS reports from 2020 show that 24.1 per cent of the largest sheriff’s departments (500 or more full-time sworn) and 65.9 per cent of police agencies serving populations of 500,000 or more had civilian oversight.

Research concerning civilian oversight boards is lacking, at least partly because of how these boards are structured within agencies from one community to the next (Engel et al. 2020). However, what we do know is that oversight boards are frequently referenced in relationship to addressing racially biased policing, improving community trust, improving accountability, and ultimately in discussions of more democratic modes of policing (Task Force 2015; Nix et al. 2018; Adams et al. 2022; Karpiak et al. 2022; Keyes and Keyes 2023; White and Schafer 2024). Racially biased policing has been one of the most prevalent themes in discussions of civilian oversight (Keyes and Keyes 2023). This is not a uniquely American issue, as other countries have struggled to incorporate and structure civilian oversight (e.g. Savage 2013; Dixon 2020). Regardless, at least in American contexts, as social pressures have mounted against police, those officers who feel a perceptual ‘war on cops’ (increased public scrutiny and dangerousness) may feel express less support for civilian oversight boards (Nix et al. 2018).

Among police leaders, elected sheriffs have seemed less supportive than police chiefs (Adams et al. 2022), which may relate to the fact that sheriffs are elected and tend to police more rural, less diverse, and more politically conservative areas than police chiefs (Falcone and Wells 1995). However, those leaders working in agencies with civilian oversight have expressed more support, and leaders’ perceptions of the civilian oversight board’s success are influenced by larger agency peers (Adams et al. 2022). White and Schafer (2024) have also demonstrated that police leaders’ perceptions of whether systemic racism is a problem in criminal justice are a strong predictor of their support for civilian oversight boards.

De-escalation and implicit bias training

De-escalation and implicit bias training are common elements of policing reform efforts (Task Force 2015). Leaders who support de-escalation training suggest it will ‘help resolve police-citizen encounters with less frequent and severe uses of force and, thereby, also increase officer safety’ (Engel et al. 2020: 153). Evidence has been somewhat mixed (White et al. 2023), but there is at least some support for the fact that de-escalation training can accomplish at least some of these goals, and citizens perceptions of encounters may improve as the result of training (Engel et al. 2022; White et al. 2023; Huff et al. 2024). Important to the current study, Engel et al. (2022: 217) show that police supervisors’ receptivity to de-escalation training (i.e. they saw it as valuable to their work and liked the training) was the only significant predictor in their models of the supervisors’ confidence and self-reported behavior in supervising and reinforcing de-escalation tactics with their subordinates.

Like de-escalation training, implicit bias training has been widely adopted in law enforcement despite a lack of evidence on its effectiveness (Engel et al. 2020; Worden et al. 2020, 2024; Glaser 2024). The basic belief is that implicit biases, which may operate below the level of conscious thought, may shape officers’ behavior (Fridell and Lim 2016; Worden et al. 2020; Lai and Lisnek 2023). Although understanding the effects of these types of training is complex, successes hinge in some ways on the behaviors and attitudes of supervisors, and police leaders may see value beyond individual officer-level impacts to include improved community trust (Engel et al. 2020; Worden et al. 2020).

Wolfe et al. (2022), in an evaluation of an officer roll-call level training that focused on procedural justice, rapport building, empathy, and self-control in police–citizen interactions, note that a key element in the developmental success of these types of training programs is the officers’ motivation and receptivity to the training itself. They find that these aspects are related to officers’ beliefs that the training is over an area that can be trained and is likely to have meaningful impacts—something they contextualize as their internal locus of control. Relatedly, White and Schafer (2024) demonstrate that police leaders’ support for implicit bias training was significantly related to an increased acknowledgment of systemic racism in criminal justice, and they report that support for implicit bias training was negatively related to the belief in the ‘war on cops’. These findings suggest that factors affecting law enforcement supervisors’ beliefs in the efficacy of de-escalation and implicit bias training are important in successfully advancing these reform-oriented efforts. Furthermore, leaders’ openness to these trainings, and specifically their perceptions that the trainings may make a meaningful impact, is what is under investigation in the current study, not the actual effectiveness of either training.

Emotional and cultural intelligences

EI is the ‘ability to recognize the meanings of emotions and their relationship and to use them as a basis in reasoning and problem solving’ (Mayer et al. 2001: 234). It is usually contextualized as including some aspects of empathy, social skills, self-awareness, and regulation (e.g. Goleman 1998). Widely discussed in the popular leadership literature, meta-analysis demonstrates that EI training can improve a person’s emotional intelligence (Hodzic et al. 2018). There are differing conceptualizations and approaches to measuring EI, but some approaches are more practical than others in survey research (Wong and Law 2002), and meta-analysis demonstrates that regardless of how it is measured, it is predictive of work-related outcomes (O’Boyle et al. 2011). The current study applies a self-reported psychometric scale developed by Wong and Law (2002), which is based on a four-branch, abilities-based model of EI (Mayer and Salovey 1997) that includes self-emotional appraisal, others’ emotional appraisal, use of emotions, and regulation of emotions (Mayer and Salovey 1997: 270–1). The Wong and Law (2002) scale has been among the most influential in the study of EI and leadership (Saha et al. 2023).

Some acknowledge that EI is critical to job performance in occupations high in emotional labor—including policing (Daus and Ashkanasy 2005). However, a recent systematic review of the policing literature revealed only about 20 studies have assessed EI, many of which are from outside the USA (Magny and Todak 2021), and there were notable methodological concerns (see White and Schafer 2023). Magny and Todak’s (2021) review is consistent with studies of other occupations in finding that EI was positively related to many positive outcomes for the individuals (e.g. job satisfaction) and the organization (e.g. work engagement). A study by White and Schafer (2023) is one of the few to have addressed EI among police leaders. They report that EI is positively associated with better workgroup fit among command staff-level officers.

Like EI, CQ is a multidimensional concept of intelligence defined by ‘a person’s capacity to adapt effectively to new cultural contexts’ (Earley 2002: 274) or their ability to function and manage culturally diverse settings effectively (Ang et al. 2007). CQ shares many parallels to EI, but while the two are statistically related, they are analytically distinct concepts (Earley and Mosakowski 2004; Thomas et al. 2015; Schlaegel et al. 2021). As Earley and Ang (2003) point out, CQ is distinct in that someone might possess seemingly high EI (e.g. greater empathy) within their own culture but struggle to adapt when confronted by cultural differences. CQ, therefore, ‘helps us explain the variation we observe in cross-cultural effectiveness among individuals who are otherwise intelligent, emotionally mature, and seem to have good social skills’ (Thomas et al. 2015: 1111).

Police reform efforts often emphasize training in cultural diversity and responsiveness toward not only racial and ethnic diversity but also other marginalized groups. This is true both historically, in examples of many reform-oriented commissions (see Walker 1998) and in more contemporary examples of calls for reform (e.g. Task Force 2015). The long-standing and broad-based attention toward improving officers’ cultural competency to reduce biased-based policing has done little to gauge officers’ CQ by applying this framework as a metric of improved competency. However, systematic reviews of the CQ literature demonstrate that cross-cultural training can improve CQ in many cases (Solomon and Steyn 2017; Phanphairoj and Piromsombat 2019).

Current study

The current study explores the reform-oriented attitudes of US police leaders relative to their EI and CQ. Using a convenience sample of leaders from across the USA, we use a series of ordinary least squares (OLS) regressions to specifically test the relationships between EI, CQ, and support for civilian oversight boards, and support for both implicit bias and de-escalation training. Controlling for other relevant factors, we hypothesize that EI and CQ will be positively related to these outcomes.

METHODS

Sample and data

Data for this study come from a sample of police leaders attending the Federal Bureau of Investigation’s National Academy (NA). The NA is an advanced 10-week, in-residence leadership training course that has been offered to state and local police supervisors by the FBI at their training academy in Quantico (VA). The FBI hosts four classes annually, with 200–250 participants per session. Selection is competitively based on applications, but each state in the USA is generally represented, with some seats in each session reserved for international law enforcement partners. Participants tend to be mid- to late-career mid-level or senior executives, although the NA does include a few frontline supervisors. Following IRB approvals, the research team administered a paper survey to participants in the program’s first week at all four sessions in 2023.

The survey received an overall response rate of 98.7 per cent. However, because this study involves reform efforts that concern primarily American policing contexts, we limited our analysis to only responses from US-based leaders (N = 758). A few cases were removed during data cleaning due to a high proportion of missing data, leaving us with a final sample of 731 (n = 731). The sample is reasonably reflective of American law enforcement leadership at these levels. For example, the sample was 88 per cent male and 75.9 per cent White (8.1 per cent Black, and 7.7 per cent Hispanic/Latino), compared to the most recent LEMAS census data from the Bureau of Justice Statistics, which reported about 11 per cent of frontline supervisors, 8.7 per cent of intermediate supervisors, and 3.6 per cent of chiefs among local agencies were female (Goodison 2022). LEMAS data also show about 75 per cent of frontline, 79 per cent of intermediate supervisors, and 87 per cent of chiefs are White, while 9.9 per cent of frontline supervisors, 9.6 per cent of intermediate supervisors, and 5.5 per cent of chiefs are Black (10 per cent, 7.2 per cent, and 3.8 per cent, respectively, are Hispanic) (Goodison 2022).

Dependent variables

Three dependent variables were used in the analyses. First, Support for Civilian Oversight was measured by first situating it in the following context: ‘Civilian oversight boards can include those that are structured in ways that investigate officer conduct, monitor agency/officer actions, or as an auditing/ombudsman style’. Respondents were then asked to what extent do they agree with the following statements: ‘Civilian oversight boards can offer police departments an added level of transparency that helps ensure public trust and confidence; Civilian oversight boards can serve as a needed accountability mechanism for police departments; Civilian oversight boards help improve integrity of police departments by ensuring officers at all levels adhere to the organization’s policies and procedures’ (emphasis in original) (Cronbach’s alpha = 0.909). These items were measured on a 5-point scale of agreement from Strongly Disagree (1) to Strongly Agree (5), summed into an additive scale where higher scores represent greater support for civilian oversight.

The other two dependent variables include Support for De-Escalation Training and Support for Implicit Bias Training. These were each measured by asking respondents to indicate their agreement with three statements: ‘De-escalation (Implicit bias) training will… reduce the frequency of negative police-citizen interactions; reduce the frequency of use of force incidents; and, reduce the severity of use of force incidents’. These were measured on the same 5-point scale of agreement (Cronbach’s alphas = 0.875 for De-escalation training and 0.937 for Implicit bias training). Both were summed into respective scales where higher scores equate to more supportive attitudes for each. All three dependent variables approximated normal distributions and met the assumptions of OLS regression analysis.

Independent variables

We include two independent variables: EI and CQ. To measure Emotional Intelligence, we utilized the Wong and Law (2002) Emotional Intelligence Scale, which uses 16 questions to test four underlying dimensions of EI: self-emotional appraisal, other’s emotional appraisal, use of emotion, and regulation of emotion. Each subdimension is measured with four indicator questions, all set to a 7-point agreement scale from Strongly Disagree (1) to Strongly Agree (7), where higher scores represent greater levels of agreement. The Cronbach’s alphas for each subdimension were sufficient (>0.70), and the Cronbach’s alpha for the full scale was 0.868. As Wong and Law (2002) used, we summed the scores for each subdimension and then averaged respondents’ scores across the four dimensions. Higher scores represent higher EI.

We measured Cultural Intelligence using a 10-question, previously validated scale adopted here from Thomas et al. (2015). The scale measures CQ by assessing three underlying subdimensions: knowledge, skill, and metacognition. Knowledge is measured with two questions, skill with five questions, and metacognition with three questions. All responses are set to a 5-point agreement scale asking respondents how well the statements describe them, from Not at All (1) to Extremely Well (5). As expected, each subdimension exhibited good internal consistency (Cronbach’s alphas = 0.80, 0.80, and 0.77, with an overall scale α = 0.886). As with the EI scale, we averaged the respondents’ scores, where higher scores represent higher levels of CQ.

Control variables

We include several individual- and agency-level factors as control variables. We include the respondents’ gender as Male. We asked respondents about their racial/ethnic identity and recorded them as White, Black, Hispanic, and Mutliracial/Other Races.

We include respondents’ Tenure in law enforcement in continuous years. Because para-military ranks can have different responsibilities at different levels, we asked respondents if, at their rank, they were considered a member of their agency’s Command Staff. We also ask respondents to report their political orientation on a scale from 1 (very liberal) to 10 (very conservative)—presented here as Political Conservativism.

We asked respondents about their attitudes toward Systemic Racism in the criminal justice system, which has been shown to affect our outcome variables (White and Schafer 2024). This was measured by asking respondents’ agreement to five questions used previously by White and Schafer (2024). These questions include, for example, ‘I believe systemic racism is a serious issue facing our criminal justice system’. These items were measured on a 5-point scale of agreement from Strongly Disagree to Strongly Agree, and summed together, so higher scores represent greater belief in systemic racism in criminal justice (Cronbach’s alpha = 0.895).

Finally, we asked respondents about their agency. We asked if their agency has a civilian oversight board (Has COB), and we asked about agency size, which we categorized as Large Agency (500 or more full-time sworn). Respondents were asked to indicate their agencies’ jurisdiction type (e.g. local, county, state, special jurisdiction, etc.), which we have recoded into two dummy variables. The first represents those who indicated they worked in an agency with an Elected Agency Head, and the second indicates they worked for a Local Agency (municipal or county), which serves to separate them from state police, special jurisdictions, and federal officers. The descriptive statistics are presented in Table 1. Table 2 presents the bivariate correlations for the main study variables.

RESULTS

Before discussing the multivariate analyses, a few issues that should be addressed emerged in the bivariate results. First, police leaders in our sample supported the three reform efforts studied. The mean scores for De-Escalation Training (M = 11.85, SD = 2.13) and Implicit Bias Training (M = 10.60, SD = 2.74) were above their scale midpoints of 9. The only outcome leaders seemed less supportive of overall was Civilian Oversight, where the mean scores (M = 8.77, SD = 3.02) were below the scale midpoint of 9.

A series of t-tests and ANOVAs with Games–Howell post hoc analyses were run to examine some of the bivariate results. Racial differences existed in support for civilian oversight and support for implicit bias training. Differences in Support for Civilian Oversight were significant (F = 13.619, P < .001), with Black leaders reporting much higher support (M = 11.13, SD = 3.32), when compared to Hispanic (M = 8.75, SD = 3.32), Multiracial/Other Races (M = 8.72, SD = 3.35), and White leaders, who had the lowest level of support (M = 8.53, SD = 2.81). White, Hispanic, and All Other/Multiracial groups did not significantly differ. Overall, the differences between Black and other groups demonstrated a medium effect size (η² = 0.053). There were no significant differences among groups on Support for De-escalation Training (F = 2.105, P = .098). However, regarding Support for Implicit Bias Training, Black and Hispanic leaders did not significantly differ from one another, and White and All Other/Multiracial did not differ. Black (M = 11.88, SD = 2.58) and Hispanic leaders (M = 11.55, SD = 2.87) reporting significantly stronger support than White (M = 10.39, SD = 0.68) and Multiracial/Other Races (M = 10.44, SD = 10.89). The effect size was smaller than civilian oversight (η² = 0.028). As a result of the racialized differences, we created two dummy variables for use in the regression models. We created a variable to measure Black Alone, and Black and Hispanic.

The bivariate results also demonstrated three other individual-level control variables were significantly related to at least one of the outcomes. Females showed more support for Civilian Oversight (M = 9.71, SD = 3.18) compared to males (M = 8.69, SD = 2.98). An independent samples t-test shows that these differences are significant (P = .005), with a modest effect size |d| = −0.341. Similarly, females reported significantly more support for implicit bias training (M = 11.21, SD = 2.30), compared to males (M = 10.54, SD = 2.77), but the effect was small (d = −0.244). Relative to the scaled variables, the correlation coefficients show respondents’ political orientation was significantly related to support for Civilian Oversight (r = −0.228, P < .01), Support for De-escalation training (r = −0.099, P < .01), and Support for Implicit Bias Training (r = −0.121, P < .01). More politically conservative leaders were less supportive of the three outcomes across the board. Likewise, perceptions of Systemic Racism as an issue in criminal justice were significantly related to support for Civilian Oversight (r = 0.398, P < .01), De-escalation training (r = 0.179, P < .01), and Implicit Bias Training (r = 0.333, P < .01). Thus, those leaders who more firmly belief that systemic racism is an issue facing the criminal justice system were more supportive of the reform efforts. Finally, education, tenure, and being a member of command staff were not significant at the bivariate level.

Agency-level controls were largely non-significant at the bivariate level, with some exceptions. The self-reported presence of a civilian oversight board in respondents’ agencies was significantly related to support for civilian oversight. Those in agencies with civilian oversight (M = 9.94, SD = 2.89) expressed significantly more support than those in agencies without oversight (M = 8.26, SD = 2.93) (P < .001, |d| = 0.575). Those in large agencies tended to be significantly less supportive of de-escalation training, as did those in agencies with elected agency heads, but the overall effect sizes of these differences were small (|d| = −0.204 and −0.223).

We present the multivariate results for each outcome variable in separate tables. In each table, we include three models where we first include all the controls, then EI, and finally, CQ. In Table 3 we provide the results for Support for Civilian Oversight. The control variables alone explained about 23 per cent (R² = 0.229), with views of Systemic Racism being the strongest predictor (β = 0.319, P < .001), followed by working in an agency with a civilian oversight board (β = 0.256, P < .001). Working in a large agency remained negatively associated with support (β = −0.122, P = .001), and though it was non-significant at the bivariate level, working in a local agency emerged as negatively related (β = −0.075, P = .038). Adding EI to the model demonstrated that EI was a significant predictor of support for civilian oversight (β = 0.098, P = .004), but it only increased the explained variance by about 1 per cent (R² = 0.239). Adding CQ in the final model showed it was not significantly related to support for civilian oversight (β = 0.020, P = .595). Although bivariate results demonstrated that Black leaders were significantly more supportive of civilian oversight, those differences were non-significant across the multivariate models. Likewise, gender, tenure, education, being in the command staff, political conservativism, and working in an agency with an elected head were all non-significant.

Table 4 presents the findings relative to Support for Implicit Bias Training. Among the control variables, leaders’ views of systemic racism remained a strong significant predictor (β = 0.305, P < .001), and working in a large agency was also significant (β = −0.088, P = .02), explaining 13 per cent of the variance (R² = 0.130). Adding EI in the second model shows EI was again, a significant predictor of Support for Implicit Bias Training (β = 0.198, P < .001). Tenure emerged as significant in the second model (β = 0.076, P = .033), and the overall explained variance increased by about 3.7 per cent (R² = 0.167). When we added CQ to the third model, it was significant (β = 0.084, P = .031), but it added less than 1 per cent of additional explained variance (R² = 0.172). As might be expected, adding CQ to the model reduced the strength of EI (β = 0.165, P < .001), suggesting partial mediation, but EI remained the stronger of the two predictors. Although not shown in Table 4, we ran another model with CQ alone, and the explained variance was lower than with EI alone (R² = 0.150). Black and Hispanic/Latino leaders were significantly more supportive in the base model (β = 0.086, P = .024), but these effects were reduced to non-significance once EI and CQ were included. Views of systemic racism and working in a large agency remained significantly related across all three models.

In Table 5, we present the findings for Support for De-escalation Training. In the first model, with control variables alone, views of systemic racism remained a significant positive predictor (β = 0.192, P < .001), while working in a large agency (β = −0.098, P = .013) and working for an elected agency head (β = −0.080, P = .035) were both negatively related to support for de-escalation training. The inclusion of EI in the second model shows it was a significant predictor of support (β = 0.201, P < .001), and it increased the explained variance by about 4 per cent (R² = 0.099). Adding CQ in the third model showed it to be a stronger predictor of support than EI (β = 0.169, P < .001), and it served to increase the explained variance by another 2.2 per cent (R² = 0.121).

DISCUSSION

The results demonstrate modest support for the fact that police leaders who are more emotionally and culturally intelligent are more supportive of the studied reform-oriented efforts. We hypothesized that more emotionally and culturally intelligent leaders would report stronger support for civilian oversight boards, implicit bias, and de-escalation training. These hypotheses held true, except for support for civilian oversight, where only EI was significant. Despite being significant, however, the amount of explained variance attributed to EI and CQ across the three outcomes was quite small, ranging from 1 per cent to just under 5 per cent. All the same, the results add to a small but important body of knowledge on EI and CQ in policing, and they serve to help us better understand what factors might impact leaders’ perceptions of reform efforts.

EI has particularly been included in law enforcement training for several years, but studies of EI among law enforcement samples, and especially among police leaders, are lacking (White and Schafer 2023). The evidence presented in this study suggests that incorporating both EI and CQ in law enforcement training could make a small, but incremental impact on police leaders’ support for reform efforts. Previous research suggests that both EI and CQ can be improved with training (Solomon and Steyn 2017; Hodzic et al. 2018; Phanphairoj and Piromsombat 2019). To the extent that law enforcement training can improve leaders’ EI and CQ, such training may better equip them to handle policing and police leadership in an increasingly diverse environment that demands more democratic modes of policing. Regardless of the actual effectiveness of these reform efforts, which is a wholly different research question, the fact remains that police leaders are the ones interacting with citizen groups demanding change. In that context, leaders’ lack of support for popular reforms might signal their agency’s unwillingness to embrace change at all. Police leaders’ lack of support might also be implicitly and explicitly transmitted to their personnel, potentially serving to help institutionalize organizational resistance to at least some methods intended to help reform aspects of US policing. Future research should attempt to replicate our work with police leaders, frontline officers, and other samples to see if the results are generalizable across populations and ranks.

In addition to our main study findings, there are several other points that warrant discussion from some of the individual-level control variables. First, as divisive as police reform is on a national political level, it is not surprising that more politically conservative leaders showed significantly less support for the three reform efforts (Horowitz et al. 2020), but these results were reduced to non-significance once placed in the multivariate context. Second, building on the work presented by White and Schafer (2024), we show that the more leaders believe systemic racism is a problem in criminal justice, the more supportive they are of reform efforts. Like the reforms themselves, clear political divisions exist regarding how America sees the issue of systemic racism (Hurst 2023). Notably, political conservativism and views of systemic racism are correlated but not overly so (r = −0.401, P < .001).

Diversity-related training often includes some discussion of concepts relating to systemic racism. Our results suggest that not only EI and CQ training, but trainings that help police leaders understand the issues and impacts of systemic problems may help encourage them to understand the needs for police reform. Leaders’ perceptions of systemic racism within the justice system were key in shaping their views of reform. To the extent such perceptions might be malleable through training and other interventions, this may suggest a pathway for building greater support for, or acceptance of, efforts to reform policing. Whether perceptions of systemic racism within the justice system hold importance among frontline personnel remains an unresolved area of inquiry. These findings are important because research supports that officers’ motivation to train is associated with receptivity toward training, and their motivation to train is associated with their belief ‘that the training focuses on a topic that can be trained and such training can have meaningful impacts’ (Wolfe et al. 2022: 28). The current study demonstrates that EI and CQ, as well as a belief that systemic racism is a problem for criminal justice, are positively related to such beliefs about the efficacy of these types of training. While true at the individual level, Wolfe et al. (2022: 29) also note that for training such as these to be successful, ‘agencies need to work on creating a “training culture” where officers believe that training is beneficial’. In our study, we consider police leaders’ perceptions, and this group is the one who can shape their organizations’ training culture. Therefore, the factors shaping their support of these types of training have implications not only for the leaders and leadership training but also for trainings at lower levels within their organizations, particularly among frontline officers. Those leaders who acknowledge that systemic racism is a more pressing issue for criminal justice are more likely to support de-escalation and implicit bias training.

Finally, there were racial differences in leaders’ support for the three reforms. Particularly, Black leaders were significantly more supportive of civilian oversight boards, and Black and Hispanic/Latino leaders being significantly more supportive of implicit bias training. However, these differences were largely non-significant in the multivariate models. Gender differences that were observed did not remain significant in the multivariate models. Education level and being a member of command staff were not significant, and tenure produced some mixed findings, with some evidence that more tenured individuals are more supportive of implicit bias training.

Turning to agency-level factors, our study supports previous findings that suggest those working in agencies with civilian oversight boards are more supportive (Adams et al. 2022). In our results, working in an agency with a civilian oversight board was positively correlated with support for their use. Even though larger agencies are more likely to have civilian oversight boards (Adams et al. 2022), respondents in our samples from larger agencies were statistically less supportive. Those working in local agencies were also less supportive than others, which in our sample include federal-level officers and state police. All the same, officers anecdotally trust their colleagues, and the water cooler-type chatter among police leaders should not be ignored as a potential mechanism for spreading reform-oriented efforts. The inclusion of some positive opinions from leaders working in agencies with civilian oversight boards at leadership conferences, training events, and within trade publications may serve to demonstrate what the research shows—that those leaders who work in an agency with a civilian oversight board support their use more than those who have not.

Our results also showed working in large agencies of 500 or more full-time officers was negatively related supportive of civilian oversight, de-escalation, and implicit bias training. Our sample includes mainly mid-level supervisors (Lieutenants, Captains, Majors, Assistant Chiefs/Chief Deputies) with an average of more than 21 years in policing. As such, many of these leaders had likely been exposed to patterns of civil unrest over their years of service and may have policed protest events as a patrol officer and/or a patrol-level supervisor. Although tenure was only (positively) related to support for implicit bias training, the accumulated experiences of large department leaders may alternatively add to their more cynical views of common reform-oriented efforts. The current data did not allow us an opportunity to explore this aspect, and this is an area we suggest that future qualitative research might consider.

We assumed that those working in sheriff’s departments with an elected agency head may have less supportive attitudes, but this was only true regarding support for de-escalation training. Contrary to previous research, there was no association between elected agency heads and support for civilian oversight boards (Adams et al. 2022). This finding should not preclude others from continuing to examine these differences in future research.

Finally, it is important to recognize that our study has limitations. Though our data came from a broad cross-section of American police leaders, it is a convenience sample drawn from a selective advanced 10-week leadership class. The program’s design often prevents smaller agencies from being able to send their officers; mid-sized and larger agencies tend to be over-represented within the NA. Although the sample is reasonably consistent with prior research from the NA and with LEMAS data concerning the demographics of police supervisors at these levels, the generalizability of results should be considered. The amount of variance explained in our models demonstrates that many factors that could explain outcomes were not addressed in our survey, so future research should continue to explore different explanations for popular police reforms. Future studies should also continue to explore the antecedents and outcomes of EI and CQ in policing contexts.

CONCLUSION

These study findings highlight the utility of EI and CQ for understanding the attitudes, perceptions, and beliefs of police leaders, particularly related to views of topics central to the current dialog around police reform. Leaders rated higher on EI and CQ tended to be more supportive of training and procedures, which are frequently identified as key elements to reform US policing, net of other relevant variables. Additionally, leaders with higher EI and CQ were more likely to acknowledge the existence of systemic racism problems within US criminal justice. That acknowledgment itself is important in shaping support for reform efforts. Because EI and CQ are malleable and research suggests they can be developed, and because the attitudes and views of police leaders are necessary (although perhaps not sufficient) to influence the attitudes and views of line personnel, training and educational programs to develop EI and CQ among police leaders might represent a key vector for facilitating internal police reform efforts.

The utility of EI and CQ for researchers and police reformers is likely not a panacea. However, these concepts are important additions to the array of concepts that might help us better understand the attitudes, beliefs, and behaviors of police leaders and supervisors. Emerging research supports these constructs have theoretical and statistical relevance, although more inquiry is needed to develop the parameters of their salience further. To the extent that police leadership training can incorporate and develop police leaders’ EI and CQ, such efforts serve to help make leaders more adaptive and receptive to the evolving demands for reform.

Conflict of interest

None declared.

Funding

None declared.

REFERENCES