Article Critique

Assessing Similarities and Differences in Self-Control between Police Officers and Offenders

Ryan C. Meldrum1 & Christopher M. Donner2 & Shawna Cleary3 & Andy Hochstetler4 & Matt DeLisi4

Received: 2 August 2019 /Accepted: 21 October 2019 / Published online: 2 December 2019 # Southern Criminal Justice Association 2019

Abstract Research provides consistent evidence that non-offenders have greater self-control than offenders. While such differences exist across a range of samples, the ability of measures of self-control to discriminate between different groups merits additional attention. We advance research on this topic by comparing the self-control of police officers to offenders. Results indicate police officers score higher than offenders do on global self-control. Results also indicate that, when analyzing differences across the six dimensions of self-control conceptualized by Gottfredson and Hirschi (1990), police officers consistently score lower in impulsivity, self-centeredness, and anger than offenders. At the same time, police officers have a greater preference for physical activities than offenders do, and the risk-seeking and simple tasks dimensions are inconsistently associated with being a police officer relative to an offender across the different models estimated. Discussion centers on the implications of these findings for theory and for the screening of potential police recruits.

Keywords Self-control . Police officers . Prisoners . Grasmick et al. (1993) Scale

American Journal of Criminal Justice (2020) 45:167–189 https://doi.org/10.1007/s12103-019-09505-4

* Ryan C. Meldrum [email protected]

Christopher M. Donner [email protected]

Shawna Cleary [email protected]

Andy Hochstetler [email protected]

Matt DeLisi [email protected]

Extended author information available on the last page of the article

Introduction

Self-control is a core individual-level construct that has profound implications for behavior transcending multiple contexts across the life course (Gottfredson & Hirschi, 2019; Hay & Meldrum, 2015; Moffitt, Poulton, & Caspi, 2013; Pratt, 2016). Toward the right tail of the self-control distribution, reflecting individuals with higher self-control, there are numerous behavioral benefits. Persons with greater self-control are, on average, better students, have greater work performance, have higher incomes and accumulate more wealth, and experi- ence generally low psychopathology evidenced by fewer psychiatric symptoms, less use of alcohol, and abstention from drugs and risky behaviors. Those with greater self-control also enjoy more cohesive, agreeable relationships, have higher self-esteem and self-efficacy, and experience heightened wellbeing and happiness (e.g., Baumeister & Alquist, 2009; DeLisi, 2013; Krueger, Caspi, Moffitt, White, & Stouthamer-Loeber, 1996; Moffitt et al., 2011; Tangney, Baumeister, & Boone, 2004). To illustrate, in a recent study using decades of data from a prospective birth cohort, Caspi et al. (2016) found that persons characterized by high self-control left little to no adverse societal footprint in terms of their involvement in social problems, social burden, and crime.

Toward the left tail of the self-control distribution, reflecting individuals with lower self-control, there are numerous behavioral liabilities. Gottfredson and Hirschi’s (1990) theoretical construct nicely instantiates low self-control with its presentation of a person who is impulsive, risk seeking, self-centered, easily angered, prefers simple tasks, and action-oriented. In sharp contrast to their peers with higher self-control, those with low self-control impose a disproportionate and substantial societal burden in terms of their involvement in unhealthy behaviors and attendant medical costs, accidents, substance use, and dysfunctional behaviors (Gottfredson & Hirschi, 2019; Caspi et al., 2016; DeLisi, 2011; Hay & Meldrum, 2015; Moffitt et al., 2011). Moreover, low self-control is associated with the full spectrum of criminal, externalizing, and antisocial behaviors evidenced by multiple meta-analytic reviews (de Ridder, Lensvelt-Mulders, Finkenauer, Stok, & Baumeister, 2012; Pratt & Cullen, 2000; Vazsonyi, Mikuška, & Kelley, 2017). As Vazsonyi et al. (2017, p. 59) recently stated, “self-control theory has established itself as one of the most influential pieces of theoretical scholarship during the past century, as it continues to stand up to a plethora of rigorous empirical tests.”

Against this backdrop of the established importance of self-control and evidence supporting the core argument of Gottfredson and Hirschi’s (1990) general theory of crime, the current study contributes to the self-control literature by comparing self- control levels of offenders to non-offenders (e.g., Turner & Piquero, 2002). Though this topic has received considerable attention in the literature, to date no studies have evaluated such differences when juxtaposing the self-control levels of police officers and offenders, and we believe such a study is worthy of empirical investigation. As we will discuss, there are several reasons to suspect that police officers would, on average, have substantively higher levels of global self-control than offenders, though there is also reason to suspect exceptions may exist for certain dimensions of self-control emphasized by Gottfredson and Hirschi (1990). Consequently, this study will contrib- ute to the existing literature on the generality and dimensionality of self-control, while also providing important implications for police policy and practice.

In the following sections, we first provide a brief overview of self-control theory and its arguments concerning differences in self-control between offenders and non-

168 American Journal of Criminal Justice (2020) 45:167–189

offenders. Next, we draw attention to the policing literature, noting the traits that police agencies desire among officers and the manner in which these traits overlap with Gottfredson and Hirschi’s (1990) conceptualization of self-control. In the process, we also review research investigating how self-control relates to officer behavior. After outlining the goals of the current study and stating our hypotheses, we present an empirical analysis that compares the self-control of offenders against police officers.

Theory and Prior Research

Self-Control Theory

In an effort to provide a general theory of crime, Gottfredson and Hirschi (1990) proposed low self-control is “… the individual-level cause of crime” (p. 232, original emphasis). Their theory assumes that people make rational decisions and that crime does not require any special motivation; it is simply an expression of one’s natural predisposition to pursue pleasure and avoid pain (Gottfredson & Hirschi, 1990). The authors further contend that those who lack self-control are more likely to pursue the immediate pleasure of criminal behavior when presented with an opportunity to do so.

In conceptualizing self-control, Gottfredson and Hirschi (1990) define it as “the differen- tial tendency of people to avoid criminal acts whatever the circumstances in which they find themselves” (p. 87). Individuals with low self-control tend to engage in crime and behaviors analogous to crime because they lack the capacity to consider the long-term consequences of their behavior (see also Gottfredson & Hirschi, 2019). They go on to posit that crime and its analogous acts are immediately gratifying, simple, and exciting, and they presume that people involved in these types of behaviors will exhibit similar characteristics. Specifically, they argue that individuals lacking self-control (1) have a here-and-now orientation, so that they seek immediate gratification; (2) prefer easy and simple endeavors and tend to dislike activities that require diligence, tenacity, and persistence; (3) engage in risky and exciting, rather than cautious and cognitive, behaviors; (4) are quick-tempered; (5) are attracted to endeavors that entail little skill or planning; and (6) are unkind, insensitive, and self-centered. Gottfredson and Hirschi (1990) further assert that, “There is considerable tendency for these traits to come together in the same people, …it seems reasonable to consider them as comprising a stable construct useful in the explanation of crime” (pp. 90–91).

Gottfredson and Hirschi’s (1990) theoretical premise advances the hypothesis that offenders should have lower self-control relative to non-offenders (pp. 130–131). Prior research has consistently supported this assertion, and these self-control differences are across a range of different samples (e.g., Beaver, DeLisi, Mears, & Stewart, 2009; Carroll et al., 2006; Turner & Piquero, 2002; Winfree Jr, Taylor, He, & Esbensen, 2006).1 For example, Turner and Piquero (2002) compared self-control levels of 393 offenders and 120

1 In previous research, the determination of differentiating ‘offenders’ from ‘non-offenders’ has been based, for the most part, on the participant’s own self-reported involvement in crime and delinquency. For example, Turner and Piquero (2002), using NLSY data of adolescents, categorized ‘offenders’ as those who self-reported engaging in at least one of 14 delinquency items within the preceding 3 years. Similarly, Winfree Jr et al. (2006), using adolescent self-report data from a national evaluation of the GREAT program, classified ‘offenders’ as those who self-reported engaging in at least one of 17 delinquency items within the preceding year.

American Journal of Criminal Justice (2020) 45:167–189 169

non-offenders over seven waves of data collection. Across the first four waves, using a behavioral measure of self-control, they found significant mean-differences in self-control in three of the four waves. In each of these, non-offenders had statistically lower means, which indicated higher self-control. Across the last three waves, using an attitudinal measure of self-control, they found significant mean-differences in self-control in all three waves. Again, non-offenders had statistically lower means, which indicated higher self-control.

In a similar manner, Winfree Jr et al. (2006) examined self-control differences among a sample of 2921 offenders and 1650 non-offenders. To measure self-control, they utilized the four impulsivity and four risk seeking items from the Grasmick et al. (1993) scale to create an impulsivity scale, a risk seeking scale, and an eight-item global self-control scale. Their results demonstrated significant mean-differences across all three self-control measures, with non-offenders consistently yielding higher self-con- trol. Moreover, results from a multivariate regression model indicated that being in the offender group was significantly related to higher impulsivity, higher risk seeking, and lower global self-control. While such findings are illuminating, the generality of self- control can be further demonstrated by comparing offenders not simply to a general population sample of non-offenders but to a sample of individuals that should be (but not always are) high in self-control: police officers.

Policing and Self-Control

Police officers interact with the public on a daily basis, and, as law enforcers and peacekeepers, they have an obligation to “serve and protect.” Whether they are attempting to diffuse a domestic violence situation, conducting a traffic stop, rendering first aid at an accident scene, assisting a disabled motorist, or maintaining order at a civil protest, they are entrusted by society to behave with steadfast professionalism and integrity. The nature of the profession, including regular encounters with rude, defiant, and sometimes violent individuals, does not make this commitment easy. Further, with the job comes a tremendous amount of authority and discretion (e.g., Bittner, 1970; Brooks, 1993; Skogan & Frydl, 2004; Reiss, 1971; Walker, 1993). Officers have the legally prescribed power to deprive a citizen of his/her freedom of movement, and they can use legally appropriate physical force to do so. Within this context, officers have to ‘wear many hats,’ and the job frequently places them in stressful situations where quick decisions need to be made. Moreover, they, particularly patrol officers, often perform job duties outside of direct supervision.

Given the uniqueness of the policing profession, it is easy to understand why there are certain personality traits/characteristics that police officers are expected to possess—and that agencies try to identify in their applicants through the hiring processes. In line with Gottfredson and Hirschi’s (1990) conceptualization of self- control, both scholarly and professional sources emphasize that officers should be thoughtful and deliberate (rather than impulsive), courteous and caring (rather than self-centered), and slow to anger (rather than having a volatile temper) (e.g., Capps, 2014; Morison, 2017; Ohio Law Enforcement Foundation, 2001).

Police departments across the United States are in general agreement that self- control—and/or its underlying elements—is a desirable characteristic of police officers. For example, Larry Capps, a former assistant chief of the Missouri City (TX) Police Department, identified having a controlled temper as a key trait that police officers

170 American Journal of Criminal Justice (2020) 45:167–189

should possess (Capps, 2014). He suggests that a controlled temper involves self- control (or self-discipline), and that it requires an abundance of competence, confi- dence, and emotional maturity. This is particularly important when officers encounter citizens who have lost their tempers, as trying to resolve a volatile situation becomes exponentially more problematic if officers respond by losing their own temper.

Other examples also illustrate the centrality of different elements of self-control in the policing profession. According to California’s Commission on Peace Officer Standards and Training (2014), there are certain behavioral traits that departments should evaluate when selecting/hiring applicants for law enforcement positions. Among these are impulse/anger control, even temper, stress tolerance and recovery, thoroughness, attention to detail, situational/problem analysis, and decision-making/ judgment. Similarly, the Ohio Law Enforcement Foundation (2001) identified self- control and discipline as key characteristics that departments should consider during their hiring process. In their own recruiting efforts, the Bainbridge Island (WA) Police Department (Bainbridge Island Police Department, 2012) recognized being analytical, having a calming demeanor, having compassion and empathy, being detail-oriented, being emotionally resilient, having frustration tolerance, being non-impulsive, being patient, and having self-control as key characteristics that are sought in their applicants.

More recently, a forum of approximately 50 law enforcement practitioners from around the country convened to discuss challenges and strategies for twenty-first century law enforcement hiring practices. Recruiters selected the practitioners for this forum, in large part, because their agencies had implemented innovative hiring pro- grams that have shown promise in their communities and that may be useful models for other jurisdictions (Morison, 2017). The forum identified seven key traits of the “21st

century police officer.” Among these were empathy, self-control, and problem-solving skills. Moreover, community residents advocate for these same qualities. According to research from Whetstone, Reed, and Turner (2006), community members expect a high degree of competency from police officers, and their findings revealed that community members expect officers to possess—among other qualities—self-discipline, patience, and attention to detail.

In essence, self-control and several of its underlying dimensions articulated by Gottfredson and Hirschi (1990) are key traits that police administrators (and the community) look for in their recruits/officers. To corroborate this assertion, policing scholars have identified aspects of self-control in several studies as predictors of “successful” officers. For example, research from Hargrave and Berner (1984) found that police supervisors in California generally agreed effective officers were, among other things, emotionally controlled. Similarly, Hogue, Black, and Sigler’s (1994) research of Alabama police officers identified several preferred characteristics, such as emotional stability, patience, and being slow to anger. Using the NEO Personality Inventory (NEO-PI), Detrick and Chibnall (2006) found that the best entry-level officers (as rated by Field Training Officers) were low in neuroticism and high in conscientiousness, the latter being a concept that correlates highly with self-control (e.g., De Vries & Van Gelder, 2013; Jones, 2017). Looking deeper into the subscales, however, revealed more nuanced results. The best officers were low in the angry- hostility subscale, but were on par with their “average officer” counterparts on the impulsiveness subscale (both subscales under neuroticism). The best officers also rated higher on the self-discipline subscale, but were on par with their counterparts on the

American Journal of Criminal Justice (2020) 45:167–189 171

deliberation subscale (both subscales under conscientiousness). Interestingly, the best officers rated higher in extraversion and had higher scores on the excitement seeking subscale. Overall, the sample of training officers concluded that the best officers were emotionally controlled, slow to anger, highly conscientious, and disciplined.

Related research also links low self-control and related constructs to negative police behavior. For example, Hiatt and Hargrave (1988) demonstrated officers that departments disciplined for misconduct scored significantly higher on the Minnesota Multiphasic Per- sonality Inventory (MMPI) hypomania scale, indicating that these officers had higher levels of disinhibition and lack of restraint. Additionally, Hargrave and Hiatt (1989) found that problem officers had significantly lower scores on the self-control subscale of the California Personality Inventory (CPI). Likewise, Girodo (1991) found that high extraversion, high neuroticism, and disinhibition were significant NEO-PI predictors of on-the-job misconduct among a sample of federal undercover drug agents. Sarchione, Cuttler, Muchinsky, and Nelson-Gray’s (1998) research further identified that officers who had been formally disciplined for misconduct scored significantly lower on three subscales of the CPI (respon- sibility, socialization, and self-control). While not directly assessing the effects of self-control on police misconduct, Pogarsky and Piquero (2004) used the impulsivity items from the Grasmick et al. (1993) scale to assess whether impulsivity mediated the relationship between deterrence and police misconduct, finding that impulsivity had a direct effect on misconduct. Recent findings also reveal that low self-control predicts officers’ citizen complaints (behav- ioral self-control measure; Donner & Jennings, 2014) and officers’self-reported engagement in misconduct (Grasmick et al., 1993 measure; Donner, Fridell, & Jennings, 2016).

The Current Study

Past research comparing self-control levels of offenders to non-offenders finds non- offenders possess greater self-control. Likewise, the policing literature consistently identifies self-control—and several of its elements—as traits that police officers should embody. Given the unique position that police officers occupy and the legally pre- scribed authority, discretion, and tools (e.g., firearms) that accompany the profession, high self-control appears to be a natural prerequisite. Taken together, these observations lead to the conclusion that police officers, on average, should possess significantly higher levels of self-control than offenders.

To our knowledge, no study has made this direct comparison. While it might seem obvious to expect that police officers would have more self-control than offenders would, we view this gap in the literature as something worthy of empirical investigation for several reasons. First, comparing the self-control of police officers to offenders offers a unique test of the ability of measures of self-control to discriminate between individuals who should, according to Gottfredson and Hirschi (1990, pp. 130-131), occupy opposing ends of the self-control distribution. Second, if minimal differences in self-control between police officers and offenders are observed, this would potentially raise important concerns about existing screening procedures used in the recruitment processes of potential officers. Third, being able to glean further insight into the self- control of police officers is of great importance, particularly at a time of increased public scrutiny of officer behavior and concerns over officer misbehavior.

Accordingly, the current study compares the self-control levels of a sample of police officers to offenders by combining multiple existing datasets, each of which includes the

172 American Journal of Criminal Justice (2020) 45:167–189

Grasmick et al. (1993) self-control scale. Based on theory and prior research (e.g., Gottfredson & Hirschi, 1990; Turner & Piquero, 2002; Winfree Jr et al., 2006), the primary hypothesis tested is that police officers will score, on average, substantively higher on global self-control relative to offenders. In addition, our review of the policing literature consistently identifies that police officers should be low in impulsivity, slow to anger, considerate (i.e., low in self-centeredness), and able to navigate a complex and stressful job (i.e., low in preference for simple tasks). Yet, in considering the other two dimensions of self-control (risk seeking, physically oriented) emphasized by Gottfredson and Hirschi (1990), the police recruitment literature seemingly places less emphasis on these two aspects. This may partially reflect the fact that the nature of police work involves an acceptance of risk (e.g., Herbert, 1998; Maskaly & Donner, 2015; Skolnick & Fyfe, 1993; Van Maanen, 1975) and an expectation of physicality (e.g., Anderson, Plecas, & Segger, 2001; Bissett, Bissett, & Snell, 2012; Hunter, Bamman, Wetzstein, & Hilyer, 1999; Shephard & Bonneau, 2003). Police officers must sometimes run towards danger: they pursue fleeing suspects, rescue citizens from burning cars and buildings, and use hand-to-hand combat to disarm suspects and intervene in fights. Further, officers must be able to react instantly to whatever crisis is at hand—this requires a certain level of physical fitness. In fact, evaluations of police recruits include physical fitness, and officers must increase physical fitness through training and, while in police academies, they learn strategies for dealing with risks inherent to police work (e.g., Bureau of Justice Statistics, 2016).

Given these realities, it is possible—and perhaps even likely—that differences in levels of risk seeking and being physically oriented between police officers and offenders could be minimal, even as significant differences are observed for global self-control and its other four dimensions. Thus, our secondary hypothesis is that, when comparing self-control levels of police officers to offenders at the dimension-level, we expect offenders will score higher than police officers will in impulsivity, simple tasks, self-centeredness, and anger, but that there will be minimal or perhaps no differences in scores between officers and offenders for the risk-seeking and physical-oriented dimensions.

Method

Participants and Procedure

To examine similarities and differences in the self-control levels of police officers and offenders, we combined four different datasets. Two of these datasets provide information on offenders, while the other two provide information on police officers. Below, we briefly describe these four different data sources.2 Readers interested in

2 Authors of the current study played a principal role in the design and collection of the data for each of the four data sources. With regard to the selection of these four specific data sources, they were included in the current study because they each contained the Grasmick et al. (1993) self-control scale. To our knowledge no other data sources outside of the two we utilize in the current study exist that include data on the self-control levels of police officers for each of the six dimensions included in the Grasmick et al. (1993) scale. Similarly, very few datasets on prisoners exist that include the Grasmick et al. (1993) scale other than the two data sources used in the current study (e.g., Mitchell & MacKenzie, 2006). Existing relationships among the authors of the current study facilitated the utilization of the two offender datasets and two police officer datasets.

American Journal of Criminal Justice (2020) 45:167–189 173

more detailed information concerning the methodologies employed to produce each of the four datasets are referred to existing studies cited in the below descriptions.

To create our sample of offenders, we first made use of survey data collected in 2001 from male prison parolees located at four work-release facilities located in a Midwest- ern state.3 All of the participants had been released from a state prison within the prior six months and were serving conditional parole sentences. To collect the survey data, brochures were first distributed at all four work-release facilities letting potential participants know researchers were administering questionnaires in small groups. It was made clear to all individuals that participation was voluntary, confidential, and that they had the right to refuse to answer any of the questions on the survey. Of the 480 parolees who were invited, 208 participated, yielding a participation rate of 43%. Research staff were present when surveys were administered in small groups (from September through December 2001) in order to answer questions and provide clarifi- cation about items on the survey. Compensation in the amount of $30 was provided to participants. Of the parolees who participated in the original study, 29% were incar- cerated for violent crimes (murder, rape, assault, robbery), 22% were incarcerated for drug crimes (possession and selling), and the remaining 49% were incarcerated for a variety of other offense types (burglary, motor vehicle theft, fraud, etc.). For additional information about this data source, see DeLisi, Hochstetler, & Murphy (2003).

Next, we utilized survey data collected in 2000–2001 from 295 male prison inmates located at two prison facilities (one medium security and the other a facility that housed both medium and maximum-security inmates) in Oklahoma. Three separate random samples were drawn at the time of the original study: (1) inmates convicted of sex offenses participating in a sex offender treatment program, (2) inmates convicted of a sex offense not participating in a sex offender treatment program, and (3) inmates with no record of having committed a sex offense. After random selection, potential participants were informed by memoranda they were chosen to participate in a study about the social, economic, and criminal history backgrounds of inmates.4 All individ- uals were provided a cover letter attached to a survey questionnaire outlining informed consent, and it was made clear that participation was voluntary and that no compen- sation was being provided for participation. The overall participation rate across the three inmate groups was 40%. Of the 295 participants, 68% were incarcerated for a sex- related offense (top three by frequency: rape, lewd molestation, sodomy) and the remaining 32% were incarcerated for crimes other than sex offenses (top three by frequency: 1st degree murder, armed robbery, felony drug possession). For additional information about this data source, see Cleary (2014). After combining the information for two offender data sources, verifying the presence of common indicators of self- control and demographic characteristics (described below), and removing cases with missing data, complete data on each of the items used in the current study was available for 457 of the 503 male prison inmates and parolees.

To create our sample of police officers, we first made use of survey data collected via an online platform—Qualtrics—in 2012 from a geographically diverse, multi-

3 Following the IRB protocols of the original study, the name of the state is blinded. 4 All memoranda were generated by the individual prisons, which in addition took on the responsibility for scheduling data collection within each prison (the principal researcher and assistants were present for all data collection). Questionnaires were self-administered in the visitation rooms of the two prison facilities.

174 American Journal of Criminal Justice (2020) 45:167–189

agency sample of 101 first-line police supervisors in the United States who were partici- pating in the National Police Research Platform. The three organizations consisted of one large police department in the West, one large police department in the Midwest, and one statewide training academy in the South that trains police employees from multiple depart- ments within the same state. Following IRB approval and securing agency cooperation, initial exposure to the larger research project was provided by research team members during the first week of new supervisory training. Subsequently, email solicitations were sent to the subset of supervisor subjects who had at least 0.5 years of experience in the role of first-line supervisor. At the time of survey solicitation, the respondents had been participating with the Platform project between 0.5 and 3.5 years. Of the 475 individuals who were contacted, 101 police supervisors fully completed the survey instrument, which represents a participation rate of 21%.5 This data source serves as the basis for several published studies (Donner, Fridell, & Jennings, 2016).

Next, we used survey data collected via an online platform—Opinio—in December of 2018 and January of 2019 from non-supervisory police officers (e.g., patrol, detectives) from a medium-sized police department located in a Midwestern state. After obtaining IRB approval for the larger study and securing cooperation from the police department’s administration, a member of the research team recruited participants during seven roll call briefings over the period of four days. Emails were then sent to 249 non-supervisory officers with an invitation to voluntarily complete an online survey. A $10 donation was offered to a police officer memorial fund for every completed survey. Of the 249 officers invited to participate, 113 completed the online survey, yielding a participation rate of 45%.

Before proceeding to the descriptions of the measures for the current investigation, we should specify that because the sample of prison inmates and parolees consisted entirely of males, the decision was made to limit the analysis of police officers to males as well. After the removal of female police officers from the data, verification of the presence of common indicators of self-control and demographic characteristics for the combined sample of male police officers (which had to match the indicators for the sample of offenders), and removal of cases with missing data, complete information on each of the survey items used in the current investigation was available for 174 (81%) of the 214 police officers. Overall, we analyzed data on a sample of 631 offenders and officers.

Measures

Self-Control Each of the four data sources contained the self-control items developed by Grasmick et al. (1993). The Grasmick scale, which taps into the six dimensions of low self-control as outlined by Gottfredson and Hirschi (1990), is a widely used measure of low self-control within the criminological literature (e.g., DeLisi, Hochstetler, & Murphy, 2003; Pratt & Cullen, 2000; Vazsonyi et al., 2017). Of the 24 items originally appearing as part of this scale, 22 were common across each of the four data sources. Of the two items that were not included across each data source, one pertained to the impulsivity dimension (Item #1: “I often act on the spur of the moment without stopping to think”), and the second pertained to the self-centeredness dimen- sion (Item #3: “If something I do upsets people, it’s their problem, not mine”). In

5 Low-to-moderate response rates are not uncommon in policing research, especially given the online methodology and the sensitive nature of some of the survey items (e.g., Bishopp & Boots, 2014; Gould, 2000).

American Journal of Criminal Justice (2020) 45:167–189 175

addition, while each of the items measuring self-control among police officers was based on a four-category response set ranging from “strongly disagree” (= 1) to “strongly agree” (= 4), only one of the two data sources for the sample of offenders was based on this four-category response set. The items from the other data source for offenders included a fifth “neutral” option in the middle of the response scale, making the maximum value equal to 5 (“strongly agree”).

To account for the difference in response options across data sources, we first standard- ized responses to each of the 22 individual items. We then reverse-coded each item so that higher values indicate greater self-control. Following this, we created a 22-item measure of global self-control by averaging together all of the items (α = 0.89). In addition, to test our secondary hypothesis, we created four-item averages for four of the six dimensions of self- control (simple-tasks [α = 0.82], risk-seeking [α = 0.79], physical activities [α = 0.72], anger [α = 0.85]) and three-item averages for the two dimensions of self-control where an item was dropped because it was not common across all four data sources (impulsivity [α = 0.73], self-centeredness [α = 0.76]). Each of the seven multi-item measures (i.e., global self- control and the six subscales) were then standardized so that each measure had a mean of zero and a standard deviation of one. Descriptive statistics for the global measure of self- control, the six measures representing each of the dimensions of self-control, and each of the other measures described below are reported in Table 1.

Police Officers and Offenders Twenty-eight percent of the sample is comprised of police officers, while 72% is comprised of offenders. For the analysis, we created a dichotomized variable labeled police officer to distinguish officers from offenders in the data; police officers were assigned a value of 1 and offenders were assigned a value of 0. Thus, the key distinction we focus on in our analyses (outlined below) is whether this dichotomy is associated with differences, both substantively and statistically speaking, in global self-control and in each of the six dimensions.

Demographics For the portion of the analyses that involve multivariate modeling, we include three covariates capturing the age, race, and education level of each participant; recall sex is a constant in the sample as all participants are male. Age is measured in whole years; the youngest participant in the sample is 18 and the oldest participant is 76. The mean age for the police officers in the sample (μ = 39.3) is similar to the mean age for the offenders (μ = 38.2) based on a t-value of 1.29 (p = 0.20). Race is measured dichotomously with the variable labeled White (= 1; non-White = 0). More than three-quarters of the police officers in the sample (82.2%) are White, while slightly less than two-thirds of the offenders (64.6%) are White (t = 4.35, p < 001). Education is also measured dichotomously with the variable labeled as More Than High School (= 1; high-school degree/GED or less = 0); participants who reported taking at least some college-level classes were coded as 1. As would be expected, nearly all of the police officers in the sample report at least some education beyond a high-school degree (94.3%), while only a little more than one-third of offenders report any education beyond high-school (36.5%); the difference between the two samples based on a t-test is statistically significant at p < .001.6

6 Both race and education were originally measured as categorical variables across each of the four datasets. Yet, the coding scheme differed between the datasets. Thus, in order to create uniform measures for race and education when combining the four datasets, the dichotomous measurement approach was employed.

176 American Journal of Criminal Justice (2020) 45:167–189

Analytic Plan

The analyses unfolded in two stages. In the first stage, we conducted a series of t-tests to examine mean differences in levels of self-control between police officers and offenders.7 Specifically, t-tests were conducted for the global measure of self-control as well as the six individual dimensions of self-control. The results of these t-tests are accompanied by histograms, which overlay the global self-control (and its individual dimensions) distribution of scores for the police officers on top of the distribution of scores for the offenders. Together, the t-tests and the overlaid histograms provide an initial test of our hypotheses and offer insight into similarities and differences between police officers and offenders with regard to self-control.

In the second stage, we estimated a series of OLS regression models to examine the extent to which being a police officer, relative to an offender, is associated with greater self-control when accounting for age, race, and education. In these models, the self- control measures are modeled as the dependent variables, the dichotomous variable police officer is modeled at the independent variable, and age, race, and education level

7 Prior to this, we examined whether mean differences in global self-control existed between (1) the two separate samples of police officers and (2) the two separate samples of offenders. A t-test for mean differences in global self-control between the two samples of police officers indicated the sample of police supervisors scored slightly lower (μ = 0.27) than the non-supervisory sample of officers (μ = 0.54) based on a t-value of −2.80 (p < .01). Likewise, a t-test for mean differences in global self-control between the two samples of offenders indicated the sample of offenders from Oklahoma scored lower in self-control (μ = −0.38) than the other sample of offenders (μ = 0.13) based on a t-value of −5.21 (p < .001). Given the exploratory nature of this study, we elected to pool together the two officer samples and the two offender samples for the analysis. Because both the officer and offenders are drawn from larger populations of each, we have little reason to believe that any one of the four samples included in our analyses represents an extreme outlier. The fact that the visual distribution of scores for global self-control (presented in the results section) provides little evidence of a bimodal distribution for both the officer and offender sample reinforces this belief.

Table 1 Descriptive Statistics (N = 631)

Variables % Mean SD Min. Max. Skew Kurtosis

Global Self-Controla – 0.00 1.00 −3.58 3.22 −0.17 3.03 Impulsivitya – 0.00 1.00 −2.85 2.44 −0.36 2.62 Simple Tasksa – 0.00 1.00 −3.07 2.56 −0.42 2.95 Risk-Seekinga – 0.00 1.00 −2.70 2.97 −0.01 2.83 Physical Activitiesa – 0.00 1.00 −2.27 3.59 0.09 3.10 Self-Centerednessa – 0.00 1.00 −3.13 2.45 −0.37 2.67 Angera – 0.00 1.00 −2.51 2.42 −0.30 2.66 Police Officerb 27.6% – – 0 1

Age – 38.48 10.01 18 76

Whitec 69.4% – – 0 1

More Than High Schoold 52.5% – – 0 1

SD = standard deviation; a higher scores indicate greater self-control; b reference is offender; c reference is non-White; d reference is high school diploma/GED or less

American Journal of Criminal Justice (2020) 45:167–189 177

are modeled as covariates. To be clear, these OLS models were not estimated in order to claim that being a police officer, relative to an offender, causes someone to be higher or lower in self-control. Rather, these models identify the strength of the association between officer/offender group membership and self-control when holding constant age, race, and education level. We focus on the standardized effect of the police officer variable in each model to identify the relative magnitude of the differences in self- control between police officers and offenders and to further assess our hypothesis that larger differences will be observed for certain dimensions of self-control (i.e., impul- sivity, simple tasks, self-centeredness, anger) than other dimensions (i.e., risk-seeking, physical activities). Following the presentation of these models, we present the results of a supplementary analysis.

Results

Bivariate Analyses

Figure 1 displays the results of the t-tests and the overlaid histograms for police officer and offender self-control (recall higher values indicate greater self-control).8 The top portion of Fig. 1 displays the results for the global measure of self-control. What is clear is that the distribution of scores for police officer self-control generally falls on the right side of the range of values, whereas the distribution of scores for offender self- control is more centered along the range of values. This visual difference between the distributions of scores is reinforced by the difference in the mean values between the two groups: police officers have a mean value of 0.40, offenders have a mean value of −0.15, and this difference is statistically significant based on a t-value of 6.42 (p < .001; Cohen’s d = 0.57). Thus, we find preliminary evidence in support of our first hypothesis that police officers score, on average, higher on global self-control relative to offenders.

The two additional rows of histograms in Fig. 1 provide the results as they pertain to the six dimensions of (low) self-control reflected in the Grasmick et al. (1993) scale. What is evident from these six histograms, and the accompanying t-test information, is that comparing global self-control between police officers and offenders masks the fact that, for particular dimensions of self-control, average differences between the two groups are much larger than for other dimensions. In partial support of our secondary hypothesis, there are small to moderate differences between police officers and of- fenders for the dimensions of impulsivity (Cohen’s d = 0.58), simple tasks (Cohen’s d = 0.37), self-centeredness (Cohen’s d = 0.60), and anger (Cohen’s d = 0.74). For each of these four dimensions, the t-values are large and statistically significant, and the absolute difference in means between the two groups is a value of at least 0.35.

When focusing on the two dimensions of self-control we hypothesized to be less likely to differentiate police officers from offenders, we find little support. First, for the

8 As a further consideration, it should be pointed out that the visual overlay of the histograms only represents the relative distribution of scores for the two samples (i.e., offenders and officers). It does not take into account the fact that the distribution of scores for the offenders is based on a larger sample size (N = 457) than that of the officers (N = 174). This should be kept in mind when considering what the distribution of scores would look like for the combined sample of offenders and officers that is examined in the subsequent multivariate regression models.

178 American Journal of Criminal Justice (2020) 45:167–189

risk-seeking dimension, the mean value for police officers is 0.26, which can be compared to the mean value for offenders of −0.10. The t-value of 4.01 (p < .001) and the Cohen’s d value of 0.36 make evident that police officers are less prone to seek out risks than offenders. Second, for the physical activities dimension, the mean value for officers is −0.17, which is lower than the mean value for offenders of 0.06 (t- value = −2.62, p < .01, Cohen’s d = 0.23), indicating that the police officers have a slightly stronger preference for physical activities relative to the offenders.

Multivariate Analyses

We next turned our attention to the OLS regression models. Table 2 displays the results predicting the global measure of self-control (Model 1) and each of the six dimensions of self-control (Model 2 through Model 7). Beginning with Model 1, police officers score 0.38 points higher (p < .001) on global self-control relative to offenders when holding constant age, race, and education level. Stated in terms of standardized effects, police officers score 0.17 standard deviations higher on global self-control relative to offenders. Thus, Model 1 provides additional support for our first hypothesis. In addition, Model 1 indicates that individuals with anything more than a high-school education score higher on global self-control (β = 0.16, p < .001).

Model 2 through Model 7 provide estimates when each of the six separate dimen- sions of self-control is modeled as a dependent variable in place of the global measure of self-control. Overall, the pattern of results is in many ways similar to the pattern that emerged from the t-tests and histograms. First, the largest differences in self-control

Fig. 1 Overlaid histograms: Offender and officer self-control (N = 631)

American Journal of Criminal Justice (2020) 45:167–189 179

Ta b le 2

O L S re g re ss io ns

of gl ob al se lf -c o nt ro l an d in di vi du al d im

en si o ns

(N = 6 31 )

M od el 1 : G lo ba l

S C

M od el 2 :

Im pu ls iv it y

M o de l 3:

S im

pl e

T as ks

M o de l 4:

R is k-

S ee k in g

M od el 5:

P hy si ca l

A ct iv it ie s

M o de l 6:

S el f-

C en te re dn es s

M od el 7 :A

n ge r

V ar ia bl e

b S E

b S E

b S E

b S E

b S E

b S E

b S E

β β

β β

β β

β

P ol ic e O ff ic er

a .3 8* **

.1 0

.4 4* **

.1 0

.0 7

.1 0

.3 6* *

.1 0

-. 40 * **

.1 0

.5 6* * *

.1 0

.5 9 ** *

.1 0

.1 7

.2 0

.0 3

.1 6

-. 18

.2 5

.2 6

A ge

.0 05

.0 0 4

-. 00 6

.0 04

-. 00 5

.0 03

.0 13 * *

.0 04

-. 00 0

.0 04

.0 07

.0 04

.0 0 9*

.0 04

.0 5

-. 06

-. 05

.1 3

-. 00

.0 7

.0 9

W h it eb

-. 11

.0 8

-. 21 *

.0 8

.0 8

.0 8

-. 14

.0 9

-. 08

.0 9

-. 05

.0 9

-. 0 8

.0 8

-. 0 5

-. 10

.0 4

-. 07

-. 04

-. 02

-. 0 4

M or e th an

H .S .c

.3 3* **

.0 9

.2 8* *

.0 9

.5 0* * *

.0 9

.0 1

.0 9

.3 2 **

.0 9

.0 4

.0 9

.1 9 *

.0 9

.1 6

.1 4

.2 5

.0 0

.1 6

.0 2

.1 0

C o ns ta nt

-. 3 8*

.1 6

.0 9

.1 6

-. 14

.1 6

-. 49 **

.1 6

.0 0

.1 6

-. 39 *

.0 3

-. 5 6* * *

.1 6

A dj us te d R -S qu ar e

.0 8

.0 8

.0 7

.0 4

.0 2

.0 7

.1 1

N ot es : a re fe re n ce

is o ff en de r;

b re fe re nc e is no n- W hi te ; c re fe re nc e is hi g h sc ho ol di p lo m a/ G E D or

le ss ; F or

ea ch

va ri ab le ,r ow

1 pr es en ts th e un st an da rd iz ed

co ef fi ci en t( b)

an d st an d ar d

er ro r (S E ), w hi le ro w

2 pr es en ts th e st an da rd iz ed

co ef fi ci en t (β ); F or

al l m od el s, hi gh er

va lu es

on th e de pe nd en t va ri ab le in di ca te gr ea te r se lf- co nt ro l. * p<

.0 5;

** p<

.0 1 ; * ** p<

.0 0 1

(t w o -t ai le d)

180 American Journal of Criminal Justice (2020) 45:167–189

between police officers and offenders are observed for (low) anger (β = 0.26, Model 7), (low) self-centeredness (β = 0.25, Model 6), (low) impulsivity (β = 0.20, Model 2), and (low) risk seeking (β = 0.16, Model 4), where police officers score higher than offenders do on each of those four dimensions. Second, Model 5 indicates police officers have a lower preference for mental over physical activities than offenders (β = −0.18, Model 7). Unlike the results of the bivariate analyses, however, Model 3 indicates there is no statistical difference between police officers and offenders in the preference for simple tasks9 (β = 0.03, p = 0.51).10

Supplementary Analyses

As a means to further consider exactly which dimensions of self-control are more strongly associated with being a police officer relative to an offender, we estimated a logistic regression model in which each of the six self-control subscales and the demographic variables were included as predictors of being a police officer (relative to being an offender).11 The results of this model are presented in Table 3. As shown, when considering each of the six dimensions of self-control simultaneously, being lower in impulsivity (OR = 1.79), lower in self-centeredness (OR = 1.65), and lower in anger (OR = 1.58) is positively associated with the likelihood of being a police officer (relative to an offender). Conversely, being lower in simple tasks and lower in a preference for physical activities is negatively associated with the likelihood of being a police officer (simple tasks OR = 0.60; physical activities OR = 0.54). There is no statistically significant association between risk seeking and the likelihood of being a police officer. Lastly, being White (OR = 2.72) and having more than a high school education (OR = 34.03) is positively associated with the likelihood of being a police officer as opposed to an offender. Overall, the results of this supplementary model are consistent with the results of the OLS models for four of the six self-control subscales (impulsivity, physical activities, self-centeredness, and anger) but inconsistent with regard to the simple tasks and risk-seeking subscales.

Discussion

In this study, we compared the self-control of a sample of current and former prisoners to a sample of police officers in order to advance research on the ability of measures of

9 Education appears to be mediating the effect, as when education is removed from the model, the standard- ized effect of being a police officer (β = 0.16) is statistically significant (p < .001). 10 At the suggestion of an anonymous reviewer, we also examined interactive effects between race and status (Officer vs. Offender) for each of the seven models presented in Table 2. After applying a bonferroni correction for multiple testing (.05/7), only in the model for not preferring physical activities (i.e. preferring mental activities) did the interaction between race and status achieve statistical significance. Specifically, the interaction indicated that while police officers in general are more likely to prefer physical over mental activities than offenders, this association is particularly evident among non-white participants. We also examined potential interactive effects between education and status (Officer vs. Offender) for each of the models, finding no evidence of a moderating effect. 11 We first estimated the model using OLS regression to obtain variance inflation factors (VIFs) and tolerance statistics for each of the predictors. There was no evidence of problematic multicollinearity among the predictors; all VIFs were below 2.0 (max VIF = 1.87) and all tolerance statistics were above 0.40 (min = 0.53).

American Journal of Criminal Justice (2020) 45:167–189 181

self-control to differentiate offenders from non-offenders. In support of our primary hypothesis, police officers scored higher on a global measure of self-control than offenders. This finding, which was consistent throughout both bivariate and multivar- iate analyses, would come as no surprise to Gottfredson and Hirschi (1990; 2019), as well as authors of policing literature who maintain that police officers should have high levels of self-control (Hargrave & Berner, 1984; Hogue et al., 1994; Detrick & Chibnall, 2006). Yet, it is also worth pointing out that we observed a considerable degree of overlap in the distribution of scores for police officers and offenders on global self-control; there were many instances where police officers score substantively lower in self-control than offenders (and where offenders score substantively higher in self- control than police officers). Thus, it is likely that a number of other factors differentiate police officers from offenders in addition to self-control.

Our secondary hypothesis was that police officers and offenders would differ on four of the dimensions of self-control (impulsivity, simple tasks, self-centeredness, and anger), but exhibit few or perhaps no differences on the physical activity and risk- seeking dimensions. Through bivariate analyses, we observed differences between police officers and offenders for the impulsivity, simple tasks, self-centeredness, and anger dimensions, as hypothesized. Yet, contrary to expectations, officers scored lower in risk seeking and higher in preferences for physical activities than offenders. Within the OLS models, substantively similar patterns were observed for five of the six dimensions of self-control, with the notable exception being that officers were no more or less likely to prefer simple tasks relative to offenders.

At the same time, a slightly different pattern of results emerged from the supple- mentary logistic regression model, which offered somewhat greater support for our secondary hypothesis. Specifically, being low in impulsivity, low in self-centeredness, and low in anger was associated with the likelihood of being an officer as opposed to an offender, while risk seeking did not differentiate police officers from offenders. While these results are consistent with what we hypothesized, we did not anticipate that

Table 3 Logistic regression of being a police officer relative to an offender on self-control subscales and demographics (N = 631)

Variables b SE OR

Age −.01 .01 .99 White 1.00** .29 2.72

More Than H. S. 3.53*** .38 34.03

Impulsivity .58** .17 1.79

Simple Tasks −.51** .16 .60 Risk Seeking .07 .15 1.08

Physical Activities −.62*** .13 .54 Self-Centeredness .50** .17 1.65

Anger .46** .17 1.58

Nagelkerke R2 0.54

Higher values on each of the self-control subscale scores indicate greater self-control; *p < .05, **p < .01, ***p < .001 (two-tailed)

182 American Journal of Criminal Justice (2020) 45:167–189

having a lower preference for simple tasks or a lower preference for physical activities would be negatively associated with the likelihood of being a police officer, which is what emerged from the logistic regression model. Overall, across the different bivariate and multivariate analyses, support for our secondary hypothesis was found with regard to the impulsivity, self-centeredness, and anger dimensions, but less so with regard to the simple tasks, risk-seeking, and physical activities dimensions.

Theoretical and Policy Implications

Gottfredson and Hirschi (1990) note that offenders tend to lack restraint; on the other hand, police officers must show restraint in the face of verbal and/or physical attacks by suspects, victims and citizens alike. Resisting the impulse to strike or retort back surely indicates high self-control, perhaps illustrating why police officers in our study scored particularly low on tendency toward anger, impulsivity and self-centeredness. As previously discussed, policing is an inherently risky and physical occupation; further, although the work of police officers is often portrayed as simple—they go out and catch the “bad guys”—it involves much more. Policing involves report writing, testifying in court, and interacting with citizens individually and in groups; these are not simple tasks. Our findings that police officers scored substantively higher than offenders on a global measure of self-control, (low) impulsivity, (low) self-centeredness, and (low) anger supports Gottfredson and Hirschi’s (1990) overall concept of self-control. More- over, the findings are consistent with past research explicitly comparing the self-control of offenders to non-offenders (Turner & Piquero, 2002; Winfree Jr et al., 2006).

The results from this study also have the potential to inform police policy and practice. Given the nature of the policing profession (e.g., authority, discretion, limited supervision), it is imperative that departments employ officers with desirable characteristics, and the professional and scholarly literature reviewed earlier advocates for police administrators to identify applicants with traits consistent with high self-control. Accordingly, many agencies attempt to do so through a battery of testing hurdles during the hiring process. Much of this process, however, involves “selection by elimination” (see e.g., Metchik, 1999). Applicants who are determined to be unsuitable (i.e. those low in self-control and other undesirable characteristics) are “screened out” of the hiring process.

Historically, the hiring process to “screen out” undesirable candidates has included scenario-based questions in written tests and oral panel interviews, background inves- tigations, and psychological assessments (e.g., Arrigo & Claussen, 2003; Cochrane, Tett, & Vandercreek, 2003; Kane & White, 2012; Palmiotto, 2001). Though these hurdles are commonplace across agencies in the United States, it may be more useful for police administrators to “select in” desirable candidates (see e.g., Sanders, 2003; White, 2008). This is because screening out “undesirable” candidates may not auto- matically result in an applicant pool of “desirable” candidates; it may, in fact, result in a candidate pool of both “desirable” and “neutral” candidates. Given that prior literature has identified high self-control as a desirable characteristic (e.g., Hogue et al., 1994; Detrick & Chibnall, 2006), police administrators and practitioners could, in part, begin to design their hiring process around identifying applicants who score particularly high on the construct. If administrators truly wish to hire “desirable” candidates, they would be wise to more carefully—and intentionally—“select in” those candidates with desir- able qualities.

American Journal of Criminal Justice (2020) 45:167–189 183

The difference in approaches may seem subtle, but “screen out” strategies often involve ever-changing lists of disqualifying criteria recognized after the fact, while “select in” strategies rarely need altering and are viewed as stable indicators of decision-making predispositions. Here, police administrators, background investiga- tors, and oral board interviewers might seek to identify applicants who possess both attitudinal (e.g., on the Grasmick et al., 1993 scale) and other behavioral indicators of high self-control as opposed to simply hiring applicants who have no identifiable evidence of disqualifying characteristics. Further, given that low self-control is a known predictor of police misconduct (e.g., Donner, Maskaly, & Thompson, 2018; Pogarsky & Piquero, 2004) and intentions to use force more quickly (Staller et al., 2019), it seems even more prudent that agencies hire individuals who are high in self-control.

Limitations and Directions for Future Research

Though this study provides a unique test of Gottfredson and Hirschi’s (1990) claims concerning differences in self-control between offenders and non-offenders, it is not without limitations. A first limitation concerns the sample. Specifically, the police officers and offenders included in this study were not randomly drawn, and the sample analyzed was the result of an amalgamation of data collected over a period covering roughly 18 years (offenders contributing data in the years 2000–2001 and officers contributing data as recent as the start of 2019); participation rates were also below 50% across each of the four data sources. Thus, while the sample analyzed in the current study is particularly unique, the pattern of findings might have been different had this study been based on a contemporaneous random sample of offenders and police officers that are more representative of the respective populations. As no such data currently exists, we relied on what was available. Future efforts should be directed at making comparisons that are more generalizable between the self-control levels of offenders and officers to assess the validity of the patterns of findings revealed herein. Related to this issue is that fact that, analytically speaking, we treated offenders and police officers as two homogenous samples. There may be importance differences among offenders (e.g., white-collar vs. sex offenders) and officers (supervisors vs. foot patrol) with regard to self-control that should be considered in future research.12

Second, we measured self-control with attitudinal items from the Grasmick et al. (1993) self-control scale. Though this strategy has been widely used and validated in previous research (see e.g., Pratt & Cullen, 2000; Vazsonyi et al., 2017), future researchers could utilize measurements that are more in line with Hirschi and Gottfredson’s (1993) preference for behavioral measures or more theoretically consis- tent with Hirschi’s (2004) reconceptualization of self-control. Third, the use of a prisoner sample to represent offenders has an inherent selection bias, as these individ- uals were caught, convicted, and imprisoned. The use of such a sample fails to account for the many offenders who are not caught or who are caught and sentenced to punishments less severe than prison. This limitation, of course, applies to any study

12 We would like to thank an anonymous reviewer for raising this issue. The exploratory nature of our study, combined with the relatively small sample size, lead us to examine average differences between all offender types and all police officer types. However, as we comment, future research based on large random samples of offenders and officers could compare the self-control levels of different types of offenders to different types of officers.

184 American Journal of Criminal Justice (2020) 45:167–189

that makes use of a prison sample, of which there have been many in the criminal justice literature (e.g., Ireland, 2011; Mitchell & MacKenzie, 2006; Reisig & Mesko, 2009; Smith, 2015). That said, future researchers should consider samples that com- prise a wider spectrum of the “offender” pool.

Fourth, due to data limitations, the present study does not take into account the importance of organizational factors and police culture on officer behavior. This influence may be exercised directly (through policies and supervision) or indirectly (through values and culture). According to Skogan and Frydl (2004), “Police behavior is affected by broad forces, including features of the organizations that hire, train, and supervise police, as well as the environment in which they work” (p. 155). In fact, prior studies of organizational explanations of police behavior speak to the influence of recruitment and selection (e.g., Sechrest & Burns, 1992), police leadership (e.g., Goldstein, 1975), organizational response to police deviance (e.g., Sherman, 1978), and police culture and socialization (e.g., Herbert, 1998). Future research should attempt to replicate our results, while accounting for the importance of organization and culture.

Fifth, the framework of our study—comparing the self-control levels of offenders to non-offenders (i.e. police officers)—makes the implicit assumption that the officers in our sample are not themselves “offenders.” Though police departments make every effort to hire “non-offenders” (or, subsequently fire officers who engage in post-hiring misconduct), it is possible that our sample of officers contained individuals who have committed undetected law violations, which could affect the results.13 It is also important to note that the police code of silence, as part of the larger police culture, may contribute to the dark figure of police misconduct in which police officers violate criminal laws but are not reported or caught (e.g., Ivkovic, 2005). Thus, future researchers studying this topic should therefore attempt to identify—and only include—officers who have no discernable criminal/misconduct history. Lastly, while we compared the self-control levels of a sample of offenders with a sample of police officers, an additional issue worthy of future investigation would be to assess where the average level of self-control of non-offenders who are not members of the policing profession falls relative to offenders and police officers.

Conclusion

While prior research consistently provides evidence that non-offenders have greater self-control than offenders, no prior study has made a direct comparison of offenders to police officers. We viewed this gap in the literature as a topic worthy of empirical examination and believe that this study contributes to both the self-control and policing literatures. Through a combination of unique officer and offender datasets, our findings demonstrate that police officers, in fact, do score significantly higher than offenders on a global measure of self-control. Additionally, when analyzing differences between police officers and offenders across the six dimensions of self-control, we consistently found that officers are lower in impulsivity, self-centeredness, and anger, but that they

13 This possibility is perhaps supported by the observation that a handful of police officers in the sample scored below the offender mean on global self-control (see Fig. 1).

American Journal of Criminal Justice (2020) 45:167–189 185

are slightly higher with regard to preferring physical to mental activities. Overall, these findings offer support for the generality of self-control theory. Moreover, they yield important practical implications for police administrators who have a significant interest in hiring desirable candidates into the policing profession.

Acknowledgements The authors would like to express their appreciation to the anonymous reviewers for their comments and suggestions on an earlier draft of this manuscript.

References

Anderson, G. S., Plecas, D., & Segger, T. (2001). Police officer physical ability testing–re-validating a selection criterion. Policing: An International Journal of Police Strategies & Management, 24(1), 8–31.

Arrigo, B. A., & Claussen, N. (2003). Police corruption and psychological testing: A strategy for pre- employment screening. International Journal of Offender Therapy & Comparative Criminology, 47(3), 272–290.

Bainbridge Island Police Department. (2012). Key traits and characteristics sought in police officers. Bainbridge Island, WA: City of Bainbridge Island.

Baumeister, R. F., & Alquist, J. L. (2009). Is there a downside to good self-control? Self and Identity, 8, 115–130. Beaver, K. M., DeLisi, M., Mears, D. P., & Stewart, E. (2009). Low self-control and contact with the criminal

justice system in a nationally representative sample of males. Justice Quarterly, 26(4), 695–715. Bishopp, S. A., & Boots, D. P. (2014). General strain theory, exposure to violence, and suicide ideation among

police officers: A gendered approach. Journal of Criminal Justice, 42, 538–548. Bissett, D., Bissett, J., & Snell, C. (2012). Physical agility tests and fitness standards: Perceptions of law

enforcement officers. Police Practice and Research, 13(3), 208–223. Bittner, E. (1970). The functions of the police in modern society: A review of background factors, current

practices, and possible role models. Bethesda, MD: National Institute of Mental Health. Brooks, L. (1993). Police discretionary behavior. In R. G. Dunham & G. P. Alpert (Eds.), Critical issues in

policing (pp. 140–164). Long Grove, IL: Waveland Press. Bureau of Justice Statistics. (2016). State and local law enforcement training academies, 2013. Washington,

DC: Department of Justice. California Commission on Peace Officer Standards and Training. (2014). Behavioral traits evaluated in the

selection process. Sacramento, CA: State of California. Capps, L. E. (2014). Characteristics of an ideal police officer. FBI law enforcement. Retrieved from https://leb.

fbi.gov/articles/perspective/perspective-characteristics-of-an-ideal-police-officer. Carroll, A., Hemingway, F., Bower, J., Ashman, A., Houghton, S., & Durkin, K. (2006). Impulsivity in

juvenile delinquency: Differences among early-onset, late-onset, and non-offenders. Journal of Youth and Adolescence, 35(4), 517–527.

Caspi, A., Houts, R. M., Belsky, D. W., Harrington, H., Hogan, S., Ramrakha, S., et al. (2016). Childhood forecasting of a small segment of the population with large economic burden. Nature Human Behaviour, 1(1), 1–10.

Cleary, S. (2004). Sex Offenders and Self-Control Explaining Sexual Violence. New York: LFB Scholarly Publishing LLC.

Cochrane, R. E., Tett, R. P., & Vandercreek, L. (2003). Psychological testing and the selection of police officers. Criminal Justice and Behavior, 30(5), 511–537.

de Ridder, D. T., Lensvelt-Mulders, G., Finkenauer, C., Stok, F. M., & Baumeister, R. F. (2012). Taking stock of self-control: A meta-analysis of how trait self-control relates to a wide range of behaviors. Personality and Social Psychology Review, 16(1), 76–99.

De Vries, R. E., & Van Gelder, J. L. (2013). Tales of two self-control scales: Relations with five-factor and HEXACO traits. Personality and Individual Differences, 54(6), 756–760.

DeLisi, M. (2011). Self-control theory: The Tyrannosaurus rex of criminology is poised to devour criminal justice. Journal of Criminal Justice, 2(39), 103–105.

DeLisi, M. (2013). Pandora’s box: The consequences of low self-control into adulthood. In C. L. Gibson & M. D. Krohn (Eds.), Handbook of life-course criminology: Emerging trends and directions for future research (pp. 261–273). New York, NY: Springer.

186 American Journal of Criminal Justice (2020) 45:167–189

DeLisi, M., Hochstetler, A., & Murphy, D. S. (2003). Self-control behind bars: A validation study of the Grasmick et al. scale. Justice Quarterly, 20(2), 241–263.

Detrick, P., & Chibnall, J. T. (2006). NEO PI-R personality characteristics of high-performing entry-level police officers. Psychological Services, 3(4), 274.

Donner, C. M., Fridell, L. A., & Jennings, W. G. (2016). The relationship between self-control and police misconduct: A multi-agency study of first-line police supervisors. Criminal Justice and Behavior, 43(7), 841–862.

Donner, C. M., & Jennings, W. G. (2014). Low self-control and police deviance: Applying Gottfredson and Hirschi’s general theory to officer misconduct. Police Quarterly, 17(3), 203–225.

Donner, C. M., Maskaly, J., & Thompson, K. N. (2018). Self-control and the police code of silence: Examining the unwillingness to report fellow officers' misbehavior among a multi-agency sample of police recruits. Journal of Criminal Justice, 56, 11–19.

Girodo, M. (1991). Drug corruption in undercover agents: Measuring the risk. Behavioral Sciences & the Law, 9(3), 361–370.

Goldstein, H. (1975). Police corruption: A perspective on its nature and control. Washington, DC: The Police Foundation.

Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford University Press. Gottfredson, M. R., & Hirschi, T. (2019). Modern control theory and the limits of criminal justice. New York,

NY: Oxford University Press. Gould, L. A. (2000). A longitudinal approach to the study of the police personality: Race/gender differences.

Journal of Police & Criminal Psychology, 15, 41–51. Grasmick, H. G., Tittle, C. R., Bursik, R. J., & Arneklev, B. J. (1993). Testing the core empirical implications of

Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime & Delinquency, 30(1), 5–29. Hargrave, G. E., & Berner, J. G. (1984). POST psychological screening manual. Sacramento, CA: State of

California, Department of Justice. Hargrave, G. E., & Hiatt, D. (1989). Use of the California psychological inventory in law enforcement officer

selection. Journal of Personality Assessment, 53(2), 267–277. Hay, C., & Meldrum, R. (2015). Self-control and crime over the life course. Thousand Oaks, CA: Sage. Herbert, S. (1998). Police subculture reconsidered. Criminology, 36(2), 343–370. Hiatt, D., & Hargrave, G. E. (1988). MMPI profiles of problem peace officers. Journal of Personality

Assessment, 52(4), 722–731. Hirschi, T. (2004). Self-control and crime. In R. Baumeister & K. Vohs (Eds.), Handbook of self- regulation:

Research, theory, and applications (pp. 537–552). New York: Guilford Press. Hirschi, T., & Gottfredson, M. (1993). Commentary: Testing the general theory of crime. Journal of Research

in Crime and Delinquency, 30(1), 47–54. Hogue, M. C., Black, T., & Sigler, R. T. (1994). The differential use of screening techniques in the recruitment

of police officers. American Journal of Police, 13(2), 113–124. Hunter, G. R., Bamman, M. M., Wetzstein, C. J., & Hilyer, J. C. (1999). Validation of fitness/ physical abilities

tests for evaluating the ability to do job-related tasks. Strength and Conditioning Journal, 21(2), 33–39. Ireland, J. L. (2011). The importance of coping, threat appraisal, and beliefs in understanding and responding to fear of

victimization: Applications to a male prisoner sample. Law and Human Behavior, 35(4), 306–315. Ivkovic, S. K. (2005). Fallen blue knights: Controlling police corruption. New York: Oxford University Press. Jones, S. (2017). Does choice of measure matter? Assessing the similarities and differences among self-control

scales. Journal of Criminal Justice, 50, 78–85. Kane, R. J., & White, M. D. (2012). Jammed up: Bad cops, police misconduct, and the New York City police

department. New York: New York University Press. Krueger, R. F., Caspi, A., Moffitt, T. E., White, J., & Stouthamer-Loeber, M. (1996). Delay of gratification,

psychopathology, and personality: Is low self-control specific to externalizing problems? Journal of Personality, 64(1), 107–129.

Maskaly, J., & Donner, C. M. (2015). A theoretical integration of social learning theory with terror management theory: Towards an explanation of police shootings of unarmed suspects. American Journal of Criminal Justice, 40(2), 205–224.

Metchik, E. (1999). An analysis of the" screening out" model of police officer selection. Police Quarterly, 2(1), 79–95.

Mitchell, O., & MacKenzie, D. L. (2006). The stability and resiliency of self-control in a sample of incarcerated offenders. Crime & Delinquency, 52(3), 432–449.

Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., Harrington, H., Houts, R., Poulton, R., Roberts, B. W., Ross, S., Sears, M. R., Thomson, W. M., & Caspi, A. (2011). A gradient of childhood

American Journal of Criminal Justice (2020) 45:167–189 187

self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences of the United States of America, 108(7), 2693–2698.

Moffitt, T. E., Poulton, R., & Caspi, A. (2013). Lifelong impact of early self-control. American Scientist, 101(5), 352–359.

Morison, K. P. (2017). Hiring for the 21st century law enforcement officer: Challenges, opportunities, and strategies for success. Washington, DC: Office of Community Oriented Policing Services.

Ohio Law Enforcement Foundation. (2001). The complete guide to hiring law enforcement officers. Dublin, OH: Law Enforcement Foundation.

Palmiotto, M. J. (2001). Can police recruiting control police misconduct? In M. J. Palmiotto (Ed.), Police misconduct: A reader for the 21st century (pp. 344–354). Upper Saddle River, NJ: Prentice Hall.

Pogarsky, G., & Piquero, A. R. (2004). Studying the reach of deterrence: Can deterrence theory help explain police misconduct? Journal of Criminal Justice, 32(4), 371–386.

Pratt, T. C. (2016). A self-control/life-course theory of criminal behavior. European Journal of Criminology, 13(1), 129–146.

Pratt, T. C., & Cullen, F. T. (2000). The empirical status of Gottfredson and Hirschi's general theory of crime: A meta-analysis. Criminology, 38(3), 931–964.

Reisig, M. D., & Mesko, G. (2009). Procedural justice, legitimacy, and prisoner misconduct. Psychology, Crime & Law, 15(1), 41–59.

Reiss, A. J. (1971). The police and the public. New Haven, CT: Yale University Press. Sanders, B. A. (2003). Maybe there’s no such thing as a “good cop”: Organizational challenges in selecting

quality officers. Policing: An International Journal of Police Strategies & Management, 26(2), 313–328. Sarchione, C. D., Cuttler, M. J., Muchinsky, P. M., & Nelson-Gray, R. O. (1998). Prediction of dysfunctional

job behaviors among law enforcement officers. Journal of Applied Psychology, 83(6), 904–912. Sechrest, D. K., & Burns, P. (1992). Police corruption: The Miami case. Criminal Justice and Behavior, 19(3),

294–313. Shephard, R. J., & Bonneau, J. (2003). Assuring gender equity in recruitment standards for police officers.

Canadian Journal of Applied Physiology, 27, 263–295. Sherman, L. W. (1978). Scandal and reform: Controlling police corruption. Berkeley, CA: University of

California Press. Skogan, W., & Frydl, K. (2004). Fairness and effectiveness in policing: The evidence. Washington, D.C.:

National Academic Press. Skolnick, J. H., & Fyfe, J. J. (1993). Above the law: Police and the excessive use of force. New York: Free Press. Smith, H. P. (2015). The meaning of the cut: A phenomenological inquiry into prisoner self- injury. Justice

Quarterly, 32(3), 500–531. Staller, M. S., Müller, M., Christiansen, P., Zaiser, B., Körner, S., & Cole, J. C. (2019). Ego depletion and the

use of force: Investigating the effects of ego depletion on police officers’ intention to use force. Aggressive Behavior, 45(2), 161–168.

Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. Journal of Personality, 72, 271–324.

Turner, M. G., & Piquero, A. R. (2002). The stability of self-control. Journal of Criminal Justice, 30(6), 457–471. Van Maanen, J. (1975). Police socialization: A longitudinal examination of job attitudes in an urban police

department. Administrative Science Quarterly, 20(2), 207–228. Vazsonyi, A. T., Mikuška, J., & Kelley, E. L. (2017). It's time: A meta-analysis on the self- control-deviance

link. Journal of Criminal Justice, 48, 48–63. Walker, S. (1993). Taming the system: The control of discretion in criminal justice, 1950–1990. New York,

NY: Oxford University Press. White, M. D. (2008). Identifying good cops early: Predicting recruit performance in the academy. Police

Quarterly, 11(1), 27–49. Whetstone, T. S., Reed, J. C., & Turner, P. C. (2006). Recruiting: A comparative study of the recruiting

practices of state police agencies. International Journal of Police Science and Management, 8(1), 52–66. Winfree Jr., L. T., Taylor, T. J., He, N., & Esbensen, F. A. (2006). Self-control and variability over time:

Multivariate results using a 5-year, multisite panel of youths. Crime & Delinquency, 52(2), 253–286.

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

188 American Journal of Criminal Justice (2020) 45:167–189

Ryan C. Meldrum is an associate professor in the Department of Criminology and Criminal Justice at Florida International University. His researchinterests include testing and refining theories of delinquency and crime, prosecutorial discretion, and adolescent development. His recent research has appeared in journals such as Justice Quarterly, Criminal Justice & Behavior, Intelligence, Sleep Health, and Developmental Psychology.

Christopher M. Donner is an Assistant Professor in the Department of Criminal Justice & Criminology at Loyola University Chicago, and he received his Ph.D. in Criminology from the University of South Florida. His current research focuses on police issues, with a particular emphasis on police integrity and misconduct. His recent publications have appeared in a variety of outlets, including the Journal of Criminal Justice, Policing and Society, and Deviant Behavior.

Shawna Cleary is a Professor in the School of Criminal Justice at the University of Central Oklahoma, where she is also the School’s Graduate Advisor and Director of Field Studies and Internships. She has served as a member of the Attorney General of Oklahoma’s Domestic Violence and Sexual Assault Advisory Council since 2006. Dr. Cleary is the author of Sex Offenders and Self-Control: Explaining Sexual Violence.

Andy Hochstetler is Professor in the Department of Sociology at Iowa State University. His recent research has appeared in outlets such as the American Journal of Public Health, Justice Quarterly, and Rural Sociology among others.

Matt DeLisi is College of Liberal Arts and Sciences Dean’s Professor, Coordinator of Criminal Justice Studies, Professor in the Department of Sociology,and Faculty Affiliate of the Center for the Study of Violence at Iowa State University. Dr. DeLisi is the only scientist in the world who is Fellow of both the Academy of Criminal Justice Sciences and the Association for Psychological Science.

Affiliations

Ryan C. Meldrum1 & Christopher M. Donner2 & Shawna Cleary3 & Andy Hochstetler4 & Matt DeLisi4

1 Department of Criminology and Criminal Justice, Florida International University, Miami, FL 33199, USA

2 Department of Criminal Justice and Criminology, Loyola University Chicago, Chicago, IL 60660, USA

3 School of Criminal Justice, University of Central Oklahoma, Edmond, OK 73034, USA

4 Department of Sociology, Iowa State University, Ames, IA 50011, USA

American Journal of Criminal Justice (2020) 45:167–189 189

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.