The Roles of Police Officers in Schools

Abstract

Deploying police officers, known as school resource officers (SROs), in schools is a popular school crime prevention strategy. This study tested whether specific SRO roles, rather than the presence or absence of SROs, influenced school crime and reporting of crimes to law enforcement differently. Specifically, schools with officers serving a law enforcement only role as well as those with officers who also teach and/or mentor (“mixed SROs”), were compared with schools without officers. The study used a longitudinal sample (N = 480) from the School Survey on Crime and Safety for the years 2004, 2006, and 2008. Results suggest that the level of crime recording and reporting generally increased with SRO presence. Further, schools with law enforcement only SROs recorded more crimes than non-SRO schools, and contrary to hypotheses, schools with mixed SROs reported more crimes to law enforcement. Future research should expand on the typology of SROs used in this study.

Keywords

school resource officers school crime policing police roles

Concern over the safety of students, teachers, and administrators in U.S. schools continues to grow. This is due in part to the widespread media coverage of mass school shootings, such as the events at Columbine High School in 1999 and Sandy Hook Elementary School in 2012. Deploying police officers, known as school resource officers (SROs), in schools has been a common response. SROs mainly serve a law enforcement purpose but may also perform additional roles such as teacher or mentor (Brown, 2006; Finn & McDevitt, 2005; Finn, Shively, McDevitt, Lassiter, & Rich, 2005; James & McCallion, 2013; Thomas, Towvim, Rosiak, & Anderson, 2013).

Despite public perceptions that school crime is on the rise, serious, violent crime is in fact relatively rare in U.S. schools (Morgan, Musu-Gillette, Robers, & Zhang, 2015). For instance, the number of violent deaths of students, specifically homicides, decreased from 34 in 1997–1998 to 15 in 2011–2012 (Robers, Kemp, & Truman, 2013). Further, the percentage of students aged 12–18 who reported weapon carrying on school grounds declined from 12% in 1993 to 5% in 2013 (Robers et al., 2013). Other forms of crime, such as theft and vandalism, are much more common in schools (Morgan et al., 2015).

Schools implement a variety of strategies aimed at reducing crime and increasing safety (Gottfredson & Gottfredson, 2002; Morgan et al., 2015). These strategies include a wide range of practices such as violence prevention curricula and counseling strategies but also include exclusionary practices such as zero tolerance policies which remove high-risk youth from schools. Several security practices are used in both private and public schools such as metal detectors and security cameras (Morgan et al., 2015). One of the most popular security procedures is deploying SROs to work in schools.

Funding from federal grants has facilitated a large increase in the number of SROs placed in schools (Beger, 2002; Brown, 2006; Cook, Gottfredson, & Na, 2010; James & McCallion, 2013). The percentage of principals reporting that SROs were stationed in their schools increased from 22% during the 1996–1997 school year (Kaufman et al., 1999) to 43% during the 2009–2010 school year (Robers et al., 2013).

Little is known about the consequences of placing SROs in schools: the practice might reduce school violence but might also result in processing more youth through the juvenile or criminal justice systems (Addington, 2009; Brown, 2006; Crews, Crews, & Burton, 2013; Jennings, Khey, Maskaly, & Donner, 2011), which would likely have detrimental consequences for youth. Deploying SROs in schools might lead to more youth being placed in the justice systems if they arrest juveniles for offenses that might otherwise be handled through traditional disciplinary channels in schools such as detention or in school suspension. Existing research on the effects of SRO placement has found both positive effects (e.g., improved attitudes toward police and reductions in crime) and negative effects (e.g., increases in the recording and reporting of crimes and negative views of SROs). A possible reason for the disparate findings across studies is that SRO roles have not been examined. Officers who serve multiple roles might contribute more to school safety than those who play a law enforcement only role. Specifically, these officers might offer services that can reduce students’ propensities to engage in crime in addition to their traditional security functions.

This study builds upon previous research by testing the extent to which the effect of placing SROs in schools differs according to SRO roles. Previous research uses mostly cross-sectional data to examine merely the presence of SROs in schools. The current study contributes to the literature by employing a longitudinal design to test differences in the recording and reporting of crimes by SRO role approach.

Literature Review

Prior Research on the Effectiveness of SROs

Research findings on the effectiveness of SROs have been inconsistent. Some studies indicate that SROs have a significant impact on crime (Jennings et al., 2011; Johnson, 1999; Theriot, 2009, 2013), whereas others suggest that they do not reduce crime (Brady, Balmer, & Phenix, 2007; Jackson, 2002; Na & Gottfredson, 2011). Specifically, some studies find reductions in crimes such as fighting, possession of weapons and drugs, and burglary (Johnson, 1999) and generally positive perceptions of SROs (May, Fessel, & Means, 2004). However, other studies find negative effects of SRO placement such as increased recording and reporting of crimes similar to the crime types examined by Johnson (Brady et al., 2007; Na & Gottfredson, 2011). In addition, some studies find that students often have negative views of SROs and feel that they decrease school connectedness (Theriot, 2013).

The inconsistencies of these findings may be due to the general low level of scientific rigor of these studies (Petrosino, Guckenburg, & Fronius, 2012). Several existing studies used a post-test only design (Johnson, 1999; May et al., 2004; Theriot, 2013). Others used a pre–post design but lacked a non-SRO comparison group (Brady et al., 2007; Jennings et al., 2011). Others included a non-SRO school comparison group, but major nonequivalencies between the SRO and comparison schools were not adequately controlled (Jackson, 2002; Theriot, 2009).

Na and Gottfredson (2011) conducted one of the more rigorous evaluations of SROs to date. Using a longitudinal sample including 480 schools, they examined whether the increased use of SROs had an effect on school crime and found that schools that increased SROs reported more crime. Although the results from this study are more conclusive than those from some of the previous work on SROs, neither this study nor any of the others examined the effect of SRO roles on crime outcomes. The following section describes the roles of SROs and possible mechanisms through which the roles may affect crime outcomes.

The Roles of SROs

The National Association of SROs describes the roles of SROs as the “SRO triad,” which includes the roles of law enforcer, counselor/mentor, and teacher (Beger, 2002; Brown, 2006; Jackson, 2002; James & McCallion, 2013; Thomas et al., 2013). The absence of specific standards defining the roles and responsibilities of SROs makes it difficult to fully describe what SROs do. However, schools can be classified as having law enforcement only SRO roles or mixed SRO roles. In the former, SROs serve only the law enforcement function. In the latter, they serve the law enforcement role in addition to teaching and/or mentoring. The law enforcement role includes functions most commonly associated with the policing job such as patrolling school grounds, conducting investigations, performing sweeps for drugs and weapons, making arrests, and checking student IDs (Beger, 2002; Johnson, 1999).

Officers with mixed roles (hereafter referred to as “mixed officers”) may serve multiple roles. They might provide classroom instruction on various topics such as law, investigations, conflict resolution, and violence prevention (Thomas et al., 2013). The counseling function allows the officers to serve as a resource for students. Students may seek out SROs for advice, and in turn, SROs may identify at-risk students who may need intervention. SROs can serve as mentors through individual counseling sessions, coaching sports teams, or simply interacting informally with students (Travis III & Coon, 2005).

Only a few studies have focused on SRO roles (Finn et al., 2005; Finn & McDevitt, 2005; Travis III & Coon, 2005). Finn, Shively, McDevitt, Lassiter, and Rich (2005) examined 19 sites and found that the amount of time dedicated to each specific role depended on various factors such as the amount of school crime, the desires of the school administration, and the level of experience of the SRO. Finn and McDevitt (2005) found that the amount of time spent on each role depended on the size of the site and whether it was new or established. These case studies indicated that there is no set guideline for the proportion of time officers should dedicate to each of the roles.

Travis III and Coon (2005) surveyed school principals about the level and frequency with which officers perform different activities. The results indicated that, overall, officers spent most of their time in the law enforcer role. However, a substantial percentage of schools have officers performing additional roles as well. For example, several schools had officers running Police Athletic League teams, providing guidance to students, and teaching different classes such as DARE and conflict resolution. Additional factors such as school size, region, crime, and school level were all related to SRO roles. While this study was one of the first to examine SRO role approaches, it did not address the relationship between the roles of SROs and crime in schools.

In summary, these studies demonstrate that there is much variation across schools in SRO activities, which could influence SRO effectiveness. The following section discusses the mechanisms through which different SRO roles might influence school crime differently.

Mechanisms linking SRO approach to school crime

There are several mechanisms through which the law enforcement only approach may affect school crime. This approach could decrease crime through the deterrent mechanisms involved in security and surveillance (Jackson, 2002; Johnson, 1999; Na & Gottfredson, 2011). However, this reduction in actual crime may be offset by an increase in the level of detected crimes, as SROs increase patrols and surveillance on school grounds.

Scholars have suggested that SROs performing a law enforcement role may also increase school crime. First, they may increase student perceptions that the school environment is unsafe and may cause juveniles to resort to violence as a means of self-protection (Addington, 2009; Beger, 2002). Second, these officers may be more likely to use aggressive tactics, which may generate negative attitudes, undermine their perceived legitimacy, and reduce their effectiveness (Tyler & Fagan, 2008).

Additionally, when schools incorporate SROs, more responsibility for handling discipline problems is likely to shift from administrators and teachers to SROs (Theriot, 2009). SROs may react more harshly to offenses than do administrators, and this may be especially likely for law enforcement only SROs. This shift may result in youth being removed from school through disciplinary actions such as suspension or arrest. Removal from school may reduce the probability of promotion to the next grade, setting students back academically, weakening their social bonds to school, and increasing their risk of dropping out (Brown, 2006; Cook et al., 2010; Theriot, 2009). SROs serving the law enforcement only role may be more likely to crack down on crime, increasing the level of crime recorded and potentially the amount of crime they report to law enforcement as well.

Mixed approach

The mixed SRO role (officers who teach and/or mentor in addition to law enforcement) may decrease crime by changing perceptions of school safety and police, enhancing informal interactions between students and SROs, and providing access to additional resources. Youth in mixed role schools may begin to trust SROs more, which may improve attitudes about school and decrease crime. Teaching law-related concepts may give the SRO a chance to explain to juveniles how and why decisions are made. This may increase perceived legitimacy and reduce crime (Paternoster, Brame, Bachman, & Sherman, 1997).

Mixed approach officers might also decrease crime indirectly through referrals to outside agencies. These SROs may have more connections with agencies to which youth may be referred or may be able to provide a wider variety of services to youth than they otherwise would have received. Crimes reported to law enforcement may also decrease through these mechanisms because these officers may form stronger bonds with students through their informal interactions and may be more inclined to provide potentially helpful services rather than punishing them with harsher disciplinary actions.

However, mixed approach SROs may also increase detected crimes, thus offsetting decreases in actual school crime. As students come to view the SROs as legitimate and develop trust in the officer, they may be more likely to report a crime to the officers that they otherwise would not have reported. This could lead to more crime being recorded even if actual crime in schools did not increase.

SRO Role Effects by Crime Type

SRO role effects may vary by crime type. Na and Gottfredson (2011) found that as schools increase their use of SROs, they record more crimes involving weapon and drugs. This may be explained by the facilitation of zero tolerance for these crimes. Gregory and Cornell (2009) define zero tolerance as “a highly structured disciplinary policy that permits little flexibility in outcome by imposing severe sanctions for even minor violations of a school rule” (p. 107). In addition, schools with SROs report a higher percentage of nonserious violent crimes to law enforcement. Effects on other forms of crime were not found (Na & Gottfredson, 2011). It is also possible that specific SRO roles might influence crime recording and reporting differently for different crimes.

Kalven and Zeisel’s (1966) liberation hypothesis explains why there may be differential effects on the recording and reporting of different crimes by SRO role. The liberation hypothesis suggests that discretion is enhanced when the offense in question is a lower level crime. This hypothesis can be applied to SRO decision-making regarding school crime. SROs’ discretion is likely restricted for serious violent crimes because the severity of the offense demands a clear and consistent response. Discretion may also be limited for crimes that are often targeted by zero tolerance policies. On the contrary, there may be more discretion for less serious offenses. Further, discretion may vary by SRO role. For example, mixed approach SROs may use their discretion to provide juveniles with counseling for getting into a fight since it is considered to be a lower level offense, whereas law enforcement only SROs may not have the training or resources to provide an alternative option to use their discretion in this way.

This implies that SRO role effects should be assessed separately by crime type. Nonserious violent and property crimes can be considered as lower level offenses and more discretion may be used for these crimes, but they should be examined separately as Na and Gottfredson (2011) found that SRO presence was associated with a reporting effect for nonserious violent crimes, but not for property crimes. Weapon and drug crimes can be grouped with serious violent crimes in terms of the amount of discretion in recording and reporting these crimes because they are often the target of zero tolerance policies (Na & Gottfredson, 2011). Although the severity of serious violent crimes may differ from the severity of some weapon and drug crimes, the focus on these offenses for zero tolerance purposes may limit the discretion of the SROs regardless of the role approach used.

The Present Study

The purpose of this study is to examine whether SRO role approaches influence the recording and reporting of school crimes to law enforcement differently. The analyses compared outcomes for schools with and without SROs, and the magnitude of the SRO effect was assessed separately for schools that used law enforcement only and mixed approaches. Negative binomial regressions were used to assess effects on the recording of crimes in schools. Logistic regressions were used to measure effects on the reporting of crimes to law enforcement as this outcome was dichotomized. Outcomes were assessed separately for serious violent/weapon/drug crimes, nonserious violent crimes, and property crimes. The following hypotheses were tested:

Hypothesis 1: Schools with SROs will be associated with a higher rate of recorded crimes for all three crime types than schools without SROs, given that SROs may serve as a formal mechanism to detect and focus on crime. Additionally, SRO presence will be associated with a higher likelihood of reporting crimes to law enforcement.

Hypothesis 2: Schools with law enforcement only SROs will record and report more nonserious violent and property crimes (relative to non-SRO schools) than schools with mixed SROs, as mixed SROs may be perceived more positively and may refer youth to additional services rather than record and report crimes to law enforcement.

Hypothesis 3: There will be no difference in the effects of SROs on the recording and reporting of serious violent/weapon/drug crimes by SRO role, as officers are likely to have limited discretion in handling these offenses regardless of role.

Method

Data

The data for this study originate from the School Survey on Crime and Safety (SSOCS). The SSOCS is funded by the U.S. Department of Education. In this series, U.S. schools are periodically randomly sampled and school administrators are asked to report on crime incidents and school safety strategies in their schools (Ruddy, Neiman, Hryczaniuk, Thomas, & Parmer, 2010). The sampling frame consists of all the schools in the National Center for Education Statistics Common Core Data Public Elementary/Secondary School Universe data file, which includes a list of all national public schools (Ruddy et al., 2010). The schools are stratified by instructional level, type of locale, and enrollment size (Ruddy et al., 2010). Additionally, middle and high schools are oversampled as crime is a more frequent event in these schools.

For this study, data were compiled from the SSOCS for three consecutive school years (2003–2004, 2005–2006, and 2007–2008; Na & Gottfredson, 2011). Each of these three data sets was designed cross-sectionally. However, Na and Gottfredson (2011) merged these three school years of data to construct a longitudinal sample. Due to random sampling, several schools appeared in more than one of the data sets. The merging of the three data sets allowed for 480 schools to be assessed longitudinally. These 480 schools appeared in both 2004 and 2006, 2006 and 2008, or 2004 and 2008. Therefore, the effects of the various roles of SROs in schools were assessed at Time 2 while controlling for the outcomes measured at Time 1.¹ For example, if a school appeared in 2004 and 2006, then 2004 was considered Time 1 and 2006 was considered Time 2. There were 10 schools that appeared in all 3 years, and these schools were included only in the 2006 and 2008 pair to assess the most recent changes. Time 1 measures were controlled in analyses of Time 2 outcomes.

Measures

Dependent variables

The two outcomes assessed were crimes recorded and crimes reported to law enforcement. School administrators were asked to report the number of incidents that the school recorded and reported to law enforcement for several offenses. For example, administrators were asked to report on the number of attacks with weapons. Crimes recorded and crimes reported to law enforcement were operationalized by combining several of these items into three crime-type categories. These categories included nonserious violent crimes, property crimes, and serious violent/weapon/drug crimes. Nonserious violent crimes consisted of physical attacks, fights, or threats of physical attacks without a weapon; and property crimes included theft and vandalism. Serious violent crimes included rape, sexual battery other than rape, robbery with or without a weapon, physical attack or fight with a weapon, and threat of physical attack with a weapon. Finally, weapon and drug crimes consisted of possession of a firearm or explosive device; possession of a knife or sharp object; and distribution, possession, or use of illegal drugs or alcohol. Crimes recorded were measured as a count of the number of incidents recorded in the schools, and crimes reported were measured as a percentage of these crimes the school reported to the local law enforcement agency. Table 1 displays the descriptive statistics for the dependent variables, in addition to the individual crimes that were used to comprise these categories.² Due to the mostly bimodal distributions of crimes reported to law enforcement (e.g., schools tend to report all or very few of the crimes to law enforcement), these variables were dichotomized at the median.³

Table 1.

Descriptive Statistics for Dependent Variables (Time 2).

Variable	Mean	SD	Min	Max	n ^a,b
Crimes recorded
Serious violent/weapon/drug	11.59	15.02	0	103	480
Rape	.02	.16	0	2	480
Sexual battery	.15	.77	0	10	480
Robbery with weapon	.03	.28	0	4	480
Robbery without weapon	.74	3.72	0	50	480
Attacks with weapon	.24	2.27	0	40	480
Threats with weapon	.70	5.19	0	100	480
Firearms	.23	1.14	0	22	480
Knife or sharp object	2.30	3.32	0	25	480
Drugs	5.12	7.42	0	56	480
Alcohol	2.05	3.83	0	35	480
Nonserious violent crime	51.58	60.24	0	500	480
Attacks without weapon	18.40	22.98	0	150	480
Total no. of attacks/fights	22.23	27.57	0	250	480
Threats of attack—no W	10.94	22.27	0	200	480
Property crime	15.64	23.12	0	220	480
Thefts/larceny	10.11	15.32	0	150	480
Vandalism	5.53	13.51	0	200	480
Percentage of crimes reported
Serious violent/weapon/drug	79.17	32.90	0	100	400
Rape	100	0	100	100	10
Sexual battery	99.02	5.72	66.67	100	30
Robbery with weapon	83.33	40.82	0	100	10
Robbery without weapon	81.08	36.94	0	100	60
Attacks with weapon	83.40	34.76	0	100	30
Threats with weapon	75.88	40.54	0	100	60
Firearms	85.16	35.28	0	100	60
Knife or sharp object	76.40	40.43	0	100	310
Drugs	87.86	28.60	0	100	330
Alcohol	76.02	40.77	0	100	240
Nonserious violent crime	41.16	38.78	0	100	440
Attacks without weapon	43.99	41.92	0	100	440
Threats without weapon	42.59	43.00	0	100	330
Property crime	55.84	39.77	0	100	420
Thefts/larceny	57.87	41.61	0	100	380
Vandalism	58.50	44.71	0	100	340

^aThe number of observations in Table 1 for crimes reported to law enforcement varies because some schools did not experience those crimes and therefore could not report them to law enforcement. ^bUnweighted sample sizes are rounded to the nearest 10 to comply with Institution of Education Sciences requirements for restricted use data.

Independent variables

The primary predictors in this study were the roles of SROs in schools. This study examined schools which used the law enforcement only approach or the mixed approach and used schools in which no SROs were stationed as the control condition. In the SSOCS, school administrators were asked whether SROs were stationed in their schools as well as the various functions the SROs served for that given school year (Tonsager, Neiman, Hryczaniuk, & Guan, 2010).

Two SSOCS questions were used to construct a measure of the law enforcement role. Principals were asked whether the SRO engaged in security enforcement and patrol as well as maintained school discipline (Tonsager et al., 2010). If the principal reported that the SRO engaged in either of those two activities, then the schools were considered to have the law enforcement function. The mixed approach was captured through questions which asked principals whether SROs mentor students and teach a law-related education course/train students (e.g., drug-related education, criminal law, or crime prevention course; Tonsager et al., 2010). If the principal indicated that the SRO served the law enforcement role, but neither of the additional functions, then it was coded as 1, indicating that it used the law enforcement only approach (0 = otherwise). A second dummy variable for mixed approach schools was created. Schools were coded 1 on this variable if principals reported that SROs were used for teaching and/or mentoring also.⁴ When both of these dummies were entered into a regression equation, the omitted category (i.e., the schools that are coded 0 for both dummies) represented the control condition, no SROs.

The SRO role approaches of the schools at Time 2 were used as the independent variables. The roles at Time 2 were used rather than Time 1 because if the roles were only assessed at Time 1, then the independent variable would not capture the role approach the school was using during the schools year in which the dependent variable was measured. In this research, measuring the independent variables concurrently with the dependent variables is justifiable because SROs are often deployed to schools at the beginning of the academic year, while the SSOCS survey is administered at the end of the school year. Table 2 displays the summary statistics for these roles.

Table 2.

Descriptive Statistics for Primary Independent Variables (Time 2).

Variable	Mean	SD	Max	n ^a
No school resource officers	.229	.421	1	480
Law enforcement only	.149	.357	1	480
Mixed	.621	.486	1	480

^aUnweighted sample sizes are rounded to the nearest 10 to comply with Institution of Education Sciences requirements for restricted use data.

Control variables

Several variables were used as controls to rule out spurious relationships between SRO roles and the study outcomes. First, the Time 1 measure of the dependent variable in each equation was controlled.⁵ Several factors are associated with school crime including community context, location, and neighborhood crime rates (Gottfredson & Gottfredson, 1985; Gottfredson, Gotffredson, Payne, & Gottfredson, 2005). Community context control variables included school location and crime in the area in which the school is located. The types of location (urban fringe, town, rural, and city) were coded as dummy variables. The demographics of the students were measured by the percentage of male students as well as a measure of socioeconomic status. Originally, the percentage of students eligible for free lunch and percentage of students belonging to a racial or minority group were going to be included separately as controls. This study encountered the same issue identified by Na and Gottfredson (2011) in that these variables were too highly correlated to be included as two separate variables. Therefore, the same modeling decision was made to create a low socioeconomic status index by averaging these two variables (Na & Gottfredson, 2011, p. 628). Various additional school characteristics including average daily attendance percentage, student/teacher ratio, total enrollment, and school level (coded as dummy variables) were also controlled.

Additional variables that may be associated with SRO roles and the outcomes were also controlled. Variables measuring whether schools had prevention curricula and/or student counseling were included. Finally, non-SRO security measures were included to control for potential confounding effects. These variables included policies requiring students to pass through metal detectors, random metal detector checks, random dog sniffs, the use of clear book bags/banning book bags, security cameras, and requiring students to wear photo ID. All control variables were measured at Time 1. Table 3 displays these descriptive statistics.

Table 3.

Descriptive Statistics for Control Variables.

Variable	Mean	SD	Min	Max	n ^a
Crimes recorded (T1)
Serious violent/weapon/drug	11.40	16.20	0	175	480
Nonserious violent crime	62.48	78.21	0	656	480
Property crime	11.81	13.92	0	111	480
Crimes reported (T1)
Serious violent/weapon/drug	83.02	30.50	0	100	400
Nonserious violent crime	44.84	39.42	0	100	450
Property crime	61.30	40.42	0	100	430
Had prevention curriculum	0.827	0.378	0	1	470
Had student counseling	0.941	0.236	0	1	470
Percent male	49.78	7.11	0	90	470
SES index	35.72	26.89	1.45	99.95	460
Percent average attendance	93.11	7.402	7	100	470
(ln) Total enrollment	6.86	0.633	4.41	8.34	470
Crime in school location
Low level of crime	0.752	0.433	0	1	470
Moderate level of crime	0.183	0.387	0	1	470
High level of crime	0.065	0.247	0	1	470
Student/teacher ratio
Low student/teacher ratio	0.309	0.463	0	1	470
Medium student/teacher ratio	0.436	0.496	0	1	470
High student/teacher ratio	0.255	0.436	0	1	470
School location
City	0.272	0.445	0	1	470
Urban fringe	0.417	0.494	0	1	470
Town	0.107	0.310	0	1	470
Rural	0.204	0.404	0	1	470
School level
Elementary	0.061	0.240	0	1	470
Middle	0.371	0.484	0	1	470
Secondary	0.568	0.496	0	1	470
Pass through metal detector	0.019	0.136	0	1	470
Random metal detector checks	0.095	0.293	0	1	470
Random dog sniffs	0.486	0.500	0	1	470
Clear/banned book bags	0.091	0.287	0	1	470
Security cameras	0.575	0.495	0	1	470
Require student IDs	0.135	0.341	0	1	470

^aUnweighted sample sizes are rounded to the nearest 10 to comply with Institution of Education Sciences requirements for restricted use data.

Method

Schools that were represented in at least two of the merged school years were used (2004 and 2006 OR 2006 and 2008 OR 2004 and 2008). The first step was to identify preexisting differences among schools using different SRO role approaches, so that these differences could be controlled in subsequent analyses. A difference in means test was conducted for the continuous variables, comparing each role approach to the non SRO condition and comparing the two role approaches to each other. To test for differences by SRO role for the categorical variables, logistic regressions were used. The variables from those analyses that differed significantly by role were then included as controls in subsequent analyses.⁶

The different distributions of the dependent variables required different statistical models. A comparison of means and standard deviations indicated that the number of crimes recorded is right skewed. Therefore, negative binomial regressions were used to assess the relationships between police roles and crimes recorded.⁷ The outcome variables for crimes reported to law enforcement were analyzed using logistic regressions, as these are binary outcomes.

First, the effects of SRO presence without regard to SRO role were assessed using a dummy variable contrasting any police presence to no SRO presence. Then, three regressions used negative binomial (one for each of the crime-type categories examined) to assess the relationship between SRO roles and crimes recorded. Three regressions then used logit to assess the relationship between roles and crimes reported to law enforcement. Following Osgood (2000), the coefficients in the negative binomial regression were exponentiated. Coefficients greater than 1 indicated that schools which used those approaches had higher crime rates than schools that did not use SROs, and coefficients less than 1 indicated the opposite. For the logit regressions, the odds ratios (ORs) were compared to test the hypotheses.

As previously discussed, schools with the law enforcement only approach were expected to experience higher levels of recorded and reported nonserious violent and property crimes (relative to the control schools) than the mixed approach schools. Further, it was expected that the coefficients for the SRO roles would be the same for the more serious crimes. To compare effects for the law enforcement schools and the mixed approach schools, a t-test was run to test the null hypothesis that the coefficient on law enforcement is equal to the coefficient on mixed.

Results

The analyses of preexisting differences by SRO role revealed several differences that would have to be controlled when assessing the effects of SRO role on the outcomes. These analyses showed that SROs are placed in schools with higher levels of Time 1 crimes recorded and reported, larger schools, and secondary schools. Schools with SROs are also more likely to use other security devices as well such as random dog sniffs and security cameras. Schools whose SROs use the law enforcement only approach also differ from those that use a mixed approach. Mixed SROs are more likely to be in secondary schools, in addition to schools that have prevention curricula and lower student-to-teacher ratios. Further, law enforcement only SROs are more likely to be in schools with medium student-to-teacher ratios. These preexisting differences are controlled in subsequent analyses.

Preliminary analyses examined the relationships between SRO presence, regardless of SRO role, and the outcomes. Schools with SROs recorded a 27% higher rate of property crimes and a 57% higher rate of serious/weapon/drug crimes than non-SRO schools, after controlling for preexisting differences between the schools that might have produced these outcomes. Schools with SROs did not record more nonserious violent crimes.

As previously mentioned, crimes reported to law enforcement were dichotomized at different points for different crime types, but they always contrast lower reporting to higher reporting. Schools with SROs were 2.80 times more likely to be in the high reporting group for nonserious violent crimes and 1.73 times more likely to be in the high reporting group for property crimes.

Tables 4 and 5 show results from regressions assessing the relationships between SRO role and each outcome, controlling on preexisting differences. Schools with law enforcement only SROs recorded more crime than schools without SROs. Specifically, schools with law enforcement only SROs had a 41% higher rate of recorded nonserious violent crimes, 48% higher rate of property crimes, and 75% higher rate of serious/weapon/drug crimes than schools without SROs. Mixed schools recorded more serious/weapon/drug crimes than non-SRO schools. Relative to mixed schools, law enforcement only SROs recorded about 23% more serious/weapon/drug crimes and this difference was statistically significant.

Table 4.

Negative Binomial Regressions of Crimes Recorded on School Resource Officer Roles at Time 2.^a

	Nonserious Violence		Property		Serious/Weapon/Drug
Variable	Exp (bx)	SE	Exp (bx)	SE	Exp (bx)	SE
Law enforcement only (T2)	1.414*	.242	1.484*	.278	1.754***^c	.285
Mixed (T2)	1.140	.127	1.223	.150	1.528***	.164
Crime (T1)	1.004***	.001	1.005	.003	1.022***	.004
(ln) Total enrollment^b	1.889***	.187	1.626***	.153	1.954***	.189
Percent male	0.992	.006	1.006	.007	0.998	.006
Prevention curriculum	1.155	.146	1.086	.152	0.970	.122
Random dog sniffs	1.072	.108	0.932	.105	1.210*	.117
Require student ID	1.333*	.187	1.227	.193	1.262	.167
Security cameras	1.089	.108	1.044	.116	1.100	.105
Low student/teacher ratio	1.085	.141	0.803	.114	0.724**	.091
Medium student/teacher ratio	1.103	.133	0.925	.123	0.780*	.091
School level	0.842	.089	1.454***	.149	1.351**	.128
Constant	0.658	.460	0.113**	.086	0.026***	.019

Note. N = 470. T1 = Time 1; T2 = Time 2. ^aUnweighted sample sizes are rounded to the nearest 10 to comply with Institution of Education Sciences requirements for restricted use data. ^bFollowing Osgood (2000) and Na and Gottfredson (2011), the coefficients for enrollment need to be transformed because they are elasticities. Therefore, one was subtracted from the original coefficients, and the test statistics were calculated from (b − 1)/SE rather than b/SE. ^cThis value is significantly different from the mixed approach value.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

Table 5.

Logistic Regressions of Crimes Reported on School Resource Officer Roles at Time 2.^a

	Nonserious Violence		Property		Serious/Weapon/Drug
Variable	OR	SE	OR	SE	OR	SE
Law enforcement only (T2)	1.820^b	.701	0.672^b	.268	1.023	.418
Mixed (T2)	2.404***	.621	2.050**	.512	1.292	.361
Crime (T1)	1.010	.005	1.011	.009	1.002	.010
(ln) Total enrollment	1.492	.333	1.740*	.408	1.761*	.453
Percent male	1.007	.015	1.010	.015	1.009	.016
Prevention curriculum	0.603	.173	1.025	.292	1.052	.325
Random dog sniffs	1.583*	.348	0.992	.217	1.282	.310
Require student ID	1.548	.495	1.460	.447	0.914	.310
Security cameras	1.166	.264	1.217	.272	0.864	.215
Low student/teacher ratio	1.242	.374	1.270	.379	1.358	.440
Medium student/teacher ratio	1.245	.340	1.187	.319	1.409	.418
School level	2.204***	.513	1.260	.277	1.180	.285
Constant	0.002***	.004	0.003***	.005	0.014*	.028

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

In terms of the control variables, larger schools were associated with higher rates of all crime types. Schools with random dog sniffs were associated with a 21% higher rate of serious/weapon/drug crimes, and schools requiring student IDs had a 33% higher rate of nonserious violent crimes. Further, the higher the level of school, the more property and serious/weapon/drug crimes were recorded.

In assessing the relationships between SRO roles and crimes reported to law enforcement, law enforcement only schools did not report significantly more crimes than non-SRO schools. Mixed schools were more likely to be in the high reporting group for nonserious violent and property crimes than schools without SROs (OR: 2.40 and 2.05, respectively). In comparing law enforcement only schools to mixed schools, the results demonstrated that mixed schools were more likely to report nonserious violent and property crimes to law enforcement.⁸

Discussion

The present study adds to the existing literature on SROs by assessing the effects of SRO role approaches on the recording and reporting of crimes. The findings partially supported the hypothesis that SRO presence increases crimes recorded and reported. SRO presence was associated with more recording of property and serious/weapon/drug crimes and more reporting of nonserious violent and property crimes. These results are partially consistent with the earlier findings of Na and Gottfredson (2011) which showed that an increased use of police is associated with recording more weapon and drug crimes and reporting more nonserious violent crimes.

It was also hypothesized that law enforcement only schools would record and report more nonserious violent crimes and property crimes than mixed schools relative to non-SRO schools, and this hypothesis was supported. In terms of reporting these crimes to law enforcement, the coefficients for the mixed schools were larger than those for the law enforcement only schools, and the law enforcement only coefficient for property crimes was in the direction opposite to that hypothesized. This was contrary to the hypothesis as mixed approach schools reported more nonserious violent and property crimes than law enforcement only schools.

Lastly, it was hypothesized that there would be no difference in the effects of SROs on the recording and reporting of serious violent/weapon/drug crimes by SRO role. However, this hypothesis was not fully supported. The effect of SROs on reporting these crimes to law enforcement was the same regardless of SRO role. However, there was a significant difference in the recording of these crimes by SRO roles as law enforcement only schools had a 23% higher rate of recorded serious violent/weapon/drug crimes.

The results can be explained by several of the mechanisms discussed earlier. Schools with law enforcement only officers may have recorded more crime than schools without SROs through any of the mechanisms previously mentioned such as detection, reducing legitimacy, or creating an environment that youth perceive to be unsafe (Jackson, 2002; Na & Gottfredson, 2011). However, this research cannot test which mechanism is operating. Additionally, although it was expected that there would be no difference in the recording of serious violent/weapon/drug crimes regardless of SRO role, there are several reasons why schools with SROs focusing only on law enforcement may have recorded more of these crimes than mixed schools. For example, mixed schools may have lower rates of serious violent/weapon/drug crimes due to using potential connections with external agencies to provide services to youth. Mixed SROs may have been more likely to refer these youth to a drug treatment service than record the offense.

Differential effects for mixed schools compared to law enforcement only schools were found for nonserious violent and property crimes reported to law enforcement. However, this finding was not anticipated because crimes reported were higher for mixed schools than law enforcement only schools. In terms of serious violent/weapon/drug crimes, there were no differences in reporting these crimes to law enforcement by SRO role. This finding suggests that schools risk a “net widening” effect when SROs are placed in schools. This is especially true for lower level crimes and when SROs serve mixed roles. The mechanism behind this finding is not clear, and future research is needed to investigate this mechanism. A potential explanation may be that school administrators are more comfortable reporting crimes to law enforcement via their SROs when they see police providing services such as teaching and mentoring.

Limitations

This study did not have the ability to test the possible mediating mechanisms that might have explained the results such as a detection mechanism or access to additional resources. Measures of potential mediators were not available in the data. Future research should test for the above-mentioned mediators to determine the mechanisms through which the roles affect the outcomes of interest.

Second, the measurement of the dependent variables may have been influenced by having SROs stationed in the schools (Na & Gottfredson, 2011). Schools with and without SROs may differ in the accuracy of their principals’ crime reports. That is, non-SRO schools may be more likely to underreport incidents. Schools with SROs may appear to have more crime, but principals in those schools may be under more pressure to accurately report incidents. Therefore, subsequent research should also measure the dependent variables in ways that could not be influenced by SRO presence, such as student self-report surveys (Na & Gottfredson, 2011).

Third, unmeasured selection effects might confound the study results. Although this study attempted to control for preexisting differences between schools, there may have been unobserved characteristics that were related to both the SRO role and the study outcome. Future research would ideally randomly assign schools to different SRO role approaches.

Fourth, the data only allowed for a limited evaluation of the roles of the SROs. Travis III and Coon (2005) state that SROs perform a wide variety of activities and that the frequencies of these activities vary from daily to a few times a week/month/year. This study used a relatively crude typology of SRO role approaches based on whether principals reported that the SROs served broad roles. In addition, the data did not have available information on characteristics or demographics of the SROs. Future research should expand on this typology to address whether there are differences between the roles according to the level and frequency of activities as well as looking at characteristics of the SROs such as demographics, level of training, and experience.

Finally, the sample used in this study is not nationally representative. Although this sample did originate from the nationally representative SSOCS, this sample only included schools that appeared more than once in the three cross-sectional samples. Therefore, this sample overrepresented schools that had characteristics associated with crime such as secondary schools, large schools, and schools not located in rural areas (Na & Gottfredson, 2011).

Implications for Research and Practice

Subsequent studies should test mediators, use alternative measures of the dependent variables such as self-reports, control for preexisting differences across schools that remained unmeasured in this study, and examine the level and frequency of SRO activities. Additionally, subsequent research should assess the effects of the role approaches over multiple time points.

Policy recommendations should be made only after subsequent research confirms these findings. In general, SROs are recording and reporting more of certain crimes. There are several possible mechanisms that may be operating for schools with law enforcement only SROs to record more crimes, although these mechanisms cannot be identified in this study. Further, it is likely that the mentoring and teaching functions were added to the traditional law enforcement role to reduce crime by allowing officers to become more embedded in the students’ environment, establishing trust and providing more resources for them. However, the results indicated that mixed SROs were more likely to report less serious offenses to law enforcement. This might result in the unintended effect of youth being referred to the juvenile or criminal justice systems for offenses that otherwise may have been handled in traditional ways. Although this study stops short of establishing causal links between SRO role approaches and crimes recorded and reported, it represents a first step in examining this relationship and opens the door to future opportunities to address this policy-relevant question.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Addington

L. A.

(2009). Cops and cameras: Public school security as a policy response to Columbine. American Behavioral Scientist, 52, 1426–1446.

Beger

R. R.

(2002). Expansion of police power in public schools and the vanishing rights of students. Social Justice, 29, 119–130.

Brady

K. P.

Balmer

Phenix

(2007).School-police partnership effectiveness in urban schools: An analysis of New York City’s impact schools initiative. Education and Urban Society, 39, 455–478.

Brown

(2006). Understanding and assessing school police officers: A conceptual and methodological comment. Journal of Criminal Justice, 34, 591–604.

Cook

P. J.

Gottfredson

D. C.

(2010). School crime control and prevention. In Tonry

(Ed.), Crime and justice: A review of research (pp. 313–440). Chicago, IL: The University of Chicago Press.

Crews

G. A.

Crews

A. D.

Burton

C. E.

(2013). The only thing that stops a guy with a bad policy is a guy with a good policy: An examination of the NRA’s “National School Shield: proposal. American Journal of Criminal Justice, 38, 183–199.

Finn

McDevitt

(2005). National assessment of school resource officer programs (Final Project Report). Cambridge, MA. Retrieved from http://www.ncjrs.gov/pdffiles1/nij/grants/209273.pdf

Finn

Shively

McDevitt

Lassiter

Rich

(2005). Comparison of program activities and lessons learned among 19 school resource officer (SRO) programs. Retrieved from http://www.ncjrs.gov/pdffiles1/nij/grants/209272.pdf

Gottfredson

G. D.

Gottfredson

D. C.

(1985). Victimization in schools. New York, NY: Plenum Press.

10.

Gottfredson

D. C.

Gottfredson

G. D.

(2002). Quality of school-based prevention programs: Results from a national survey. Journal of Research in Crime and Delinquency, 39, 3–35.

11.

Gottfredson

G. D.

Gotffredson

D. C.

Payne

A. A.

Gottfredson

N. A.

(2005).School climate predictors of school disorder: Results from a national study of delinquency prevention in schools. Journal of Research in Crime and Delinquency, 42, 412–444.

12.

Gregory

Cornell

(2009). “Tolerating” adolescent needs: Moving zero tolerance policies in high school. Theory into Practice, 48, 106–113.

13.

Jackson

(2002). Police-school resource officers’ and students’ perception of the police and offending. Policing: An International Journal of Police Strategies & Management, 25, 631–650.

14.

James

McCallion

(2013). School resource officers: Law enforcement officers in schools. Congressional Service Research, 1–35.

15.

Jennings

W. G.

Khey

D. N.

Maskaly

Donner

C. M.

(2011). Evaluating the relationship between law enforcement and school security measures and violent crime in schools. Journal of Police Crisis Negotiations, 11, 109–124.

16.

Johnson

I. M.

(1999). School violence: The effectiveness of a school resource officer program in a southern city. Journal of Criminal Justice, 27, 173–192.

17.

Kalven

Jr. Zeisel

(1966). The American Jury. Boston, MA: Little Brown.

18.

Kaufman

Chen

Choy

S. P.

Ruddy

S. A.

Miller

A. K.

Chandler

K. A.,

… Klaus

(1999). Indicators of School Crime and Safety, 1999.U.S. Departments of Education and Justice (NCES 1999– 057/NCJ-178906). Washington, DC.

19.

May

D. C.

Fessel

S. D.

Means

(2004). Predictors of principals’ perceptions of school resource officer effectiveness in Kentucky. American Journal of Criminal Justice, 29, 75–93.

20.

Morgan

R. E.

Musu-Gillette

Robers

Zhang

(2015). Indicators of School Crime and Safety: 2014. Washington, DC: National Center for Education Statistics, U.S. Department of Education, and Bureau of Justice Statistics, Office of Justice Programs, U.S. Department of Justice.

21.

Gottfredson

D. C.

(2011). Police officers in schools: Effects on school crime and the processing of offending behaviors. Justice Quarterly, 30, 619–650.

22.

Osgood

D. W.

(2000). Poisson-based regression analysis of aggregate crime rates. Journal of Quantitative Criminology, 16, 21–43.

23.

Paternoster

Brame

Bachman

Sherman

L. W.

(1997). Do fair procedures matter? The effect of procedural justice on spouse assault. Law & Society Review, 31, 163–204.

24.

Petrosino

Guckenburg

Fronius

(2012). Policing schools’ strategies: A review of the evaluation evidence. Journal of MultiDisciplinary Evaluation, 8, 80–101.

25.

Robers

Kemp

Truman

(2013). Indicators of School Crime and Safety:2012 (NCES 2013-036/ NCJ 241446). Washington, DC: National Center for Education Statistics, U.S. Department of Education, and Bureau of Justice Statistics, Office of Justice Programs, U.S. Department of Justice.

26.

Ruddy

S. A.

Neiman

Hryczaniuk

C. A.

Thomas

T. L.

Parmer

R. J.

(2010). 2007–08 School Survey on Crime and Safety (SSOCS) Survey Documentation for Public-Use Data File Users (NCES 2010-307). Washington, DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.

27.

Theriot

M. T.

(2009). School resource officers and the criminalization of student behavior. Journal of Criminal Justice, 37, 280–287.

28.

Theriot

M. T.

(2013). The impact of school resource officer interaction on students’ feelings about school and school police. Crime & Delinquency, 20, 1–24.

29.

Thomas

Towvim

Rosiak

Anderson

(2013). School Resource Officers: Steps to effective school-based law enforcement. National Center for Mental Health Promotion and Youth Violence Prevention (Vol. 7).

30.

Tonsager

Neiman

Hryczaniuk

Guan

(2010). School Survey on Crime and Safety: 2007–08, Public-Use Data File Codebook (NCES 2010-334). Washington, DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.

31.

Travis

III, L. F.

Coon

J. K.

(2005). The role of law enforcement in public school safety: A national survey. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/211676.pdf

32.

Tyler

T. R.

Fagan

(2008). Legitimacy and cooperation: Why do people help the police fight crime in their communities? Ohio State Journal of Criminal Law, 6, 231–275.