The Nature of Crime by School Resource Officers

School resource officers (SROs) have become a permanent presence in many K-12 schools throughout the country. As a result, an emerging body of research has focused on SROs, particularly on how SROs are viewed by students, teachers, and the general public. This exploratory and descriptive research uses a different focus by examining the nature of crimes for which SROs were arrested in recent years with information gathered from online news sources. The current findings are encouraging insofar as they reveal that SROs are rarely arrested for criminal misconduct. When SROs were arrested, however, they are most often arrested for a sex-related offense involving a female adolescent. These sex-related incidents generally occurred away from school property or during nonschool hours and rarely involved the use of physical force. The implications of these findings for SRO programs are discussed.


Article
Teachers and administrators in many K-12 schools are no longer exclusively responsible for ensuring that students are behaving and attending class. They are often assisted by school resource officers (SROs) who are sworn law enforcement personnel deployed in schools. Indeed, it is estimated that as many as one half of public schools in the United States have at minimum a part-time police presence (Raymond, 2010), with additional estimates suggesting that as many as 17,000 SROs are employed nationwide (James & McCallion, 2013). The recent school shooting incident at Sandy Hook Elementary School in Connecticut has resulted in a further push for the placement of police officers in schools (Brenchley, 2013), particularly in elementary schools where a police presence has been rare (e.g., Na & Gottfredson, 2013). Research suggests that students and school personnel view SROs favorably (Watkins & Maume, 2012), but there are still prevailing questions in the literature about the precise role of SROs and their preparedness to work in an educational environment (Cray & Weiler, 2011).
The current study focuses on SROs by reviewing news articles reporting on the nature of criminal acts committed by school police officers. It is argued that police officers are more likely to be compliant when their purpose is well defined and they are sufficiently prepared to complete this function (Manning, 1977). There is some contentionlargely found in the debate surrounding the "school-toprison" pipeline-that SROs can suffer from role ambiguity and are insufficiently prepared to work with young persons in a school setting (e.g., Fowler, Lightsey, Monger, & Aseltine, 2010;Mukherjee, 2007;Petteruti, 2011). SROs therefore represent an important segment of law enforcement to study officer misconduct. Very few studies have systematically focused on the behavior of SROs despite the widespread use of these police personnel.

The Presence of Police Officers in Schools
Sworn police officers have long served a role in schools. Lambert and McGinty (2002) noted that prior to the 1950s police officers were periodically invited to schools to instruct students on traffic and bicycle safety. Police officers assumed a more permanent role in some schools beginning primarily in the 1950s and 1960s (Patterson, 2007). Schools in cities such as Flint (MI) and Miami (FL) established at that time what are now referred to as "School Resource Officer Programs" (Lambert & McGinty, 2002;Patterson, 2007;Streater, 2008). These programs place sworn police officers in schools on a full-or part-time basis with the primary goal of improving the safety of students and school personnel (Kupchik & Monahan, 2006).
The number of schools with a full-or part-time police presence has grown dramatically since the 1970s (Raymond, 2010;Watkins & Maume, 2012). During the 1995-1996 school year, for instance, there were 243 SROs employed in North Carolina public schools. Thirteen years later, the number of SROs assigned to work in North Carolina public schools had increased more than threefold to 849 (Center for the Prevention of School Violence, n.d.). A comparable pattern has emerged nationally. In 1999, 54% of all students aged 12 to 18 who completed the School Crime Supplement (SCS) to the National Crime Victimization Survey (NCVS) reported the presence of security guards and/or police officers in their school. By 2007, this number had increased to nearly 70% (Robers, Kemp, & Truman, 2013). Na and Gottfredson (2013) similarly found that more than 60% of urban and suburban high-school principals reported that at least one full-time SRO was stationed at their school during the 2007-2008 academic year. The most recent data from the Law Enforcement Management and Administrative Statistics (LEMAS) survey found that there were more than 17,000 SROs deployed in public schools nationwide in 2007 (James & McCallion, 2013).
There is some indication that the placement of police officers in schools has slowed since the early 2000s due to stagnate or shrinking funds. The Office of Community Policing Services (COPS) served as a federal funding source for SROs starting in 1999. Na and Gottfredson (2013) noted that "as of July 2005, COPS has awarded in excess of $753 million to more than 3,000 grantees to hire more than 6,500 SROs" (p. 621). May, Hart, and Ruddell (2011) indicated, however, that in recent years, the COPS Office made other funding areas a priority. Indeed, the COPS Office ended its Cops in Schools (CIP) program during the 2005 fiscal year and the use of federal funds to hire school security personnel through the Safe and Drug Free Schools and Communities Act ended during the 2009 fiscal year (James & McCallion, 2013). The elimination of these federal funding sources resulted in less grant or soft money being available for schools to hire SROs. May and colleagues (2011) also indicated that many schools and local police departments have had to reassess their own funding priorities because of tightening budgets related in part to the national recession of 2008 and 2009. Schools may have a difficult time justifying the expansion or preservation of "nonacademic" programs, such as an SRO program, when confronted with cost-saving decisions (May et al., 2011, p. 130;Rosiak, 2011). Data from the SCS to the NCVS offer some confirmation that the presence of SROs and other school security personnel has waned in recent years. The percentage of students who reported a police or security guard presence in the SCS has largely remained unchanged-at roughly 70%-from 2003 to 2011 (Robers et al., 2013).
The school shooting incident at Sandy Hook Elementary School in December 2012 placed a renewed focus on police officers in schools. Soon after this incident President Obama announced that federal money would be used to place police officers in schools (Brenchley, 2013). The COPS Office recently fulfilled this promise through its COPS Hiring Program by awarding 141 agencies nearly 45 million dollars to hire 356 SROs for the 2013 fiscal year (Community Oriented Policing Services, n.d.). Evidence also suggests that since Sandy Hook a number of states such as Indiana (Elliott, 2013), North Carolina (Burns, 2013), Ohio (Candisky, 2013), and Virginia (Meola, 2013) have appropriated or plan to appropriate funds for the hiring of additional SROs. This allocation of state and federal dollars suggest that once again there is an upward trend in the use of SROs, particularly in elementary schools where SROs were rarely deployed prior to the shooting at Sandy Hook Elementary School (e.g., Na & Gottfredson, 2013).

Views and Responsibilities of SROs
Few studies have systematically focused on SROs even though police officers have played a role in schools for several decades and are now deployed in many schools. The existing research offers few answers to such basic questions as how SROs are selected, the nature and extent of SRO training, and the strategic uses of SROs. An exception is a line of research that has addressed how SROs are generally viewed by students, teachers, or school administrators. Brown and Benedict (2005) found that the vast majority of primarily Hispanic students surveyed in Brownville, TX, were partial to having a police officer present at school. Schuiteman, Aikens, and Thornton (2001) reported a similar finding among a large sample of middle-and high-school students in Virginia (see also Kupchik & Bracy, 2010a). Other studies also indicate that teachers and school administrators generally favor a police presence in schools because the officers can assume a number of important roles or responsibilities including legal advisor, security coordinator for school-related activities, and a first responder for serious acts of school violence (e.g., Travis & Coon, 2005). Additional evidence also suggests that the public generally supports the placement of police officers in schools (Benigni, 2004). The level of public support for SROs is not surprising given that schools are often depicted as unsafe places in the media (Kupchik & Bracy, 2009), and schools are often characterized as "soft" targets in need of additional security measures (e.g., Hull, 2011;Trump, 2011). The visible presence of SROs may symbolically signal to parents and community members that student safety is an institutional priority (Watkins & Maume, 2012).
Most school and community stakeholders are supportive of a police presence in schools, yet concerns persist regarding the general fit and uses of SROs. Cray and Weiler (2011) broadly spoke to this point when they stated that school−police partnerships are challenging because they bring "together agencies with potentially competing orientations, practices, and assumptions regarding the behavior of K-12 students" (p. 168; see also Fowler et al., 2010;Portune, 1971). Much of the delinquency committed by students in schools would not be subject to police involvement if committed in the community because school incidents are largely of a minor nature (e.g., verbal threats; see Robers et al., 2013). Bittner (1990) noted that it is precisely for this reason that police officers have traditionally assigned little value to working with juveniles. Police officers earn institutional recognition by making the "big bust," not by resolving relatively trivial incidents involving minors (Bittner, 1990). Although there is strong reason to assume that events like Sandy Hook have altered many police officers' views on the utility and merits of working in schools with juveniles, there may be a prevailing belief in law enforcement that the "serious" police work occurs in the community with adults (e.g., Morimune, 2012;Commons, 2005).
Police officers placed in schools often spend time on tasks that are not traditional law enforcement functions such as patrol and investigations. Trump (2002) analyzed the survey responses of more than 600 SROs who attended the National Association of School Resource Officers (NASRO) Annual Conference in 2002 and found that roughly 60% of these SROs reported that they spent "most" of their time serving as an instructor/teacher or counselor/mentor. Only 41% of these officers indicated that most of their time was spent as a "law enforcement officer" (p. 45; see also . A 1996 survey of 217 SROs in North Carolina revealed that on average these officers reported spending one half of their time on law enforcement duties and the other half of their time on counseling and teaching (Barnes, 2008). Some SROs spend considerably more time on law enforcement functions (e.g., May, Cordner, & Fessel, 2004), but the prior research studies suggest that many SROs embrace the "triad" model (counselor, educator, and police officer) promoted by NASRO and the COPS Office (see Rosiak, 2011).
It is unclear though to what extent SROs are trained to be effective mentors, counselors, educators, or intervention specialists in schools. The work of Thurau, Buckley, Gann, and Wald (2013) indicates that entry-level police academy and in-service training that police officers typically receive is not sufficient for working with juveniles in the community and is certainly not sufficient for working with juveniles in an educational environment. Research suggests that because of this lack of training, many SROs are required or expected to complete some form of specialized training that focuses on the school setting (Clark, 2011;Cray & Weiler, 2011;. Specialized training is not completed by all SROs (e.g., Dycus, 2008, Fowler et al., 2010, and training opportunities are limited for some SROs because of logistical and/or budgetary reasons (e.g., May et al., 2011).

The Current Study
This exploratory study examines the nature of instances where an SRO was arrested for one or more criminal offenses and it was reported in the news media. There are at least three reasons that justify such a focus on SROs. First, despite the widespread use of SROs, there is little systematic research that focuses on the behavior of school security personnel (Brown, 2006;James & McCallion, 2013;Na & Gottfredson, 2013;Theriot, 2013). Second, Manning (1977) noted that police officers are more likely to engage in prosocial behavior when their role is well defined and they are sufficiently trained to execute this role. In the case of SROs, questions remain as to whether police officers are prepared to effectively work in a school setting because of possible role ambiguity, insufficient training, and the differing institutional norms and priorities of schools versus police agencies (e.g., Langberg, Fedders, & Kukorowski, 2011). Indeed, similar arguments have been raised when considering the role of SROs in supplying the "school-to-prison" pipeline (Fowler et al., 2010;Hirschfield, 2008;Kupchik, 2009). Because of such possible role, training, and institutional factors, SROs represent an important, and growing, segment of law enforcement to study officer misconduct. Third, research suggests that school personnel generally hold favorable views of SROs and consider them beneficial to the school environment (e.g., Travis & Coon, 2005). The general acceptance of SROs has evolved to a point where "there is no critical discussion among the public or policy-makers, and extremely little among academics, about whether police ought to be in public schools" (Kupchik & Bracy, 2010a, p. 80). Nonetheless, when officer misconduct occurs it places a school's SRO program in a negative light and may force some schools to reassess its investment in the program, particularly given that SRO funding-a nonacademic expenseis a budgetary challenge for many schools (e.g., May et al., 2011). In some schools, therefore, SRO retention may depend on averting any officer misconduct that undermines the typically favorable views of school-based police personnel. Systematic efforts to avert officer misconduct are aided by research that provides a better understanding of the nature of such incidents.
This research also examines the nature of crimes committed by police officers who are not deployed at schools. SROs often take on counseling and teaching responsibilities that require regular interaction with young persons in a nonthreatening environment. The adoption of such responsibilities clearly suggests that the day-to-day demands and challenges of SROs are disparate from those of communitybased police officers. Anecdotal evidence indicates that schools and police agencies are sensitive to these divergent responsibilities and try to hire a "teacher with a badge" or officers who are not overly preoccupied with "chasing bad guys" (e.g., Thurau & Wald, 2010). Given the unique dynamics of the job at school and the presumably distinct career orientation of SROs, there is reason to suspect the nature of misconduct by school and nonschool police officers is dissimilar.
The extent and nature of crime by police officers is still largely unknown (Anechiarico & Jacobs, 1996;Kane, 2007;Stinson, Liederbach, & Freiburger, 2010). Prior research has shown that police officers commit a variety of crimes onduty and off-duty. Crimes by police officers committed while on-duty include acceptance of bribes, theft, and drug shakedowns (Knapp Commission, 1972;Reiss, 1971;Stinson, Liederbach, Brewer, Schmalzried, et al., 2013). Some police officers commit crimes while off-duty and it is often difficult to determine whether such crimes are related to or stem from an officer's employment (Fyfe & Kane, 2006). All of this is complicated by the fact that police officers are generally exempt from law enforcement (Reiss, 1971). Police officers do not like to arrest other officers, and often minor transgressions are ignored whereas similarly situated nonpolice citizens would be arrested for their crimes (Black, 1976). The off-duty crimes for which officers are sometimes arrested include domestic violence and other assaults, driving under the influence (DUI), and various sex crimes (Stinson, Liederbach, Brewer, & Todak, 2013;Stinson, Liederbach, & Freiburger, 2012;).

Method
The current data were collected as part of an ongoing comprehensive study on police crime. The larger study was designed to locate cases reported in the news media in which sworn law enforcement officers had been arrested for one or more criminal offenses. The primary information source was the internet-based Google News™ search engine and its Google Alerts™ email update service. Google News is a computer-generated news site developed and operated by Google that aggregates news articles from more than 50,000 news sources (Bharat & Beckmann, 2010). Automated daily queries of the Google News search engine can be performed by user-defined search terms previously entered into the Google Alerts application.

Data Collection and Coding
Data were collected in real time on a daily basis from January 1, 2005, through December 31, 2011. Google Alerts searches were conducted using the same 48 search terms developed by Stinson (2009). The Google Alerts email update service sent a message each time one of the automated daily searches identified a news article in the Google News search engine that matched any of the designated search terms. The automated alerts contained a link to the URL for the news articles. The articles were located, examined for relevancy, printed, scanned, and indexed in a digital imaging database for subsequent coding and content analyses.
After the universe of news reports were identified, the case-related information was recorded using a multipage coding instrument. Data were coded on such information as arrested officer characteristics, the nature of criminal charges filed against an officer, and victim characteristics. Criminal offenses charged against officers were coded using a two-step process. Each offense charged was initially coded using the data collection guidelines of the National Incident-Based Reporting System (NIBRS) as the protocol for each criminal offense categories (U.S. Department of Justice, 2000, pp. 21-52). A total of 57 offenses are included in the NIBRS, including 46 incident-based criminal offenses in 1 of 22 basic crime categories, as well as 11 additional arrestbased minor crime categories (pp. 9-12). Additional non-NIBRS offense categories were added to the coding instrument following a pilot study because officers were often arrested for criminal offenses not included in the NIBRS. The non-NIBRS offense categories are as follows: indecent exposure, online solicitation of a child, criminal deprivation of civil rights, destroying or tampering with evidence, false reports/statements and perjury, hit and run, obstructing justice, official misconduct (including official oppression and violation of oath), and restraining/protection order violations.
The news articles were also coded using Stinson's (2009) typology of police crime. This typology differentiates police crime that is violence-related, sex-related, drug-related, alcohol-related, and/or profit-motivated. The use of this broad typology overcame initial issues we encountered with the data. For instance, the criminal charges filed against an officer in some cases failed to accurately reflect the nature of the criminal acts described in the narratives of the news articles analyzed, suggesting that preferential initial charging decisions were made as a professional courtesy to some of the officers who were arrested. On other occasions, officers were charged with generic "official misconduct" crimes in lieu of specific crimes that would prove an embarrassment for the employing law enforcement agency if the true nature of the offense were an element of the offense charged. This occurred in some cases involving SROs, where the news source disclosed that an SRO was arrested for initiating an inappropriate relationship with a student but more detailed information was unknown or not published. It is common for news media to withhold victim names and other personally identifiable information when reporting on sex crimes, especially when the victim is a child (Putnam & Finkelhor, 2006). These dynamics generated a need for a broad coding strategy that captured the nature of crimes committed by offenses. Use of the 65 specific arrest categories was also not possible due to the size of the current SRO sample (n = 52). The use of the specific offense categories resulted in many cells having few or no cases when frequency tables were generated.
Efforts were made to triangulate data sources in that multiple news articles and sources were available for many of the cases reported and, as available, copies of official court records were obtained specifically to determine eventual case outcomes. Coding of the offenses was completed by one of the authors. Procedures were undertaken to ensure the reliability of the data. One of the most widely accepted tests of reliability for content analyses is the percentage of agreement test, where the percentage of agreement among two or more coders is calculated (Riffe, Lacy, & Fico, 2005). A second coder was used to independently code a random sample of 5% of the total number of cases in our database (n = 106). The overall level of simple agreement (96.85%) between the two coders across the variables established a degree of reliability well above what is generally considered "acceptable" (Riffe et al., 2005, p. 147).

Results
All of the SROs in the current sample were male, and most had worked in law enforcement for more than 5 years prior to their arrest. The average length of service was slightly more than 10 years at the time of arrest and their average age was 39. Table 1 displays the distribution of offenses for which SROs were arrested. The second column in Table 1 indicates the number (n) of SROs who were arrested for any one of the broad category of offenses listed. The five offense categories listed in Table 1 are therefore not mutually exclusive. Two SROs in the sample, for instance, supplied female adolescents with alcohol and then had sexual contact with the girls. These two officers are found in the alcohol-and sex-related categories in Table 1. We found, however, that in most cases SROs were arrested for either a single offense or for multiple offenses that were similar in nature. One SRO in the sample, for example, was arrested for possession of child pornography and sexual abuse of middle-school students, both sexually related offenses.
It is also worth noting that all of the violence-related offenses in Table 1 were nonsexual in nature. Most SROs who committed sex-related crimes chose a teenage victim. A majority of these crimes were described in news sources as consensual even in instances where direct physical contact occurred. The only exceptions involved cases where it was not known or made clear by news sources whether an officer used physical force, perhaps as a courtesy to the young victim and her family. In these instances, it was impossible to determine whether a sex-related offense involved the use of physical force and thus violence-related.
The data presented in Table 1 show that by a considerable margin, SROs in the sample were most often arrested for sex-related offenses-nearly 62% of all offenses. When SROs did commit a sex crime, students were often the victim. Of the SROs who were arrested for sex-related crimes, 56% (n = 18) chose a victim(s) enrolled at the school at which they were employed. Six additional SROs committed sex crimes against a nonfamily member who was of school age, but it was unclear whether the victim was a student at the school where the SRO worked. We find that 75% (n = 24) of SROs arrested for sex-related crimes targeted nonfamily youths almost entirely of middle and high school age.
Many of the school-age victims were teenage girls. Specifically, 62% (n = 20) of the SROs arrested for sexrelated crimes were charged with offenses against teenage girls who were not their family member. One incident involved an SRO who kissed a student enrolled in a junior cadet program he supervised and then days later sent sexually explicit text messages to the girl's cell phone. Another SRO was charged with having sex with a female student. Prior to his arrest, school officials had expressed concerns about how this SRO was interacting with the female student. When direct physical contact occurred in the sex-related offenses, news sources often did not report on the precise time and location of the incident. The limited information we do have suggests that these incidents occurred mostly during nonschool hours and/or off school grounds.
Unlike the sex-related offenses, other offenses for which SROs were arrested typically did not involve students or minors. The alcohol-related offenses largely involved arrests for DUI, whereas the violence-related offenses were dominated by domestic violence incidents. Profit-motivated incidents dealt with crimes such as SROs stealing police-or school-issued equipment or falsifying the amount of overtime worked. These nonsexual crimes once again were generally not committed in the company of students or minors. Table 2 displays the general types of offenses for which police officers, excluding SROs, were arrested over the same time period. In all, 95% (n = 1,674) of these police officers were male, with an average age of 37 years, and had nearly 10 years of sworn law enforcement experience at the time of their arrest. The current results indicate that in terms of gender, age, and experience, arrested SROs were demographically similar to arrested nonschool police officers. Differences between the two groups of officers become apparent when the types of offenses for which they were arrested are  examined. Table 2 displays results from bivariate chi-square tests that assess whether significant arrest differences exist between SROs and other officers. These tests reveal that nonschool police officers were significantly more likely to be arrested for violence-related offenses, whereas SROs were significantly more likely to be arrested for sex-related offenses. The measure of association, Cramer's V, between officer type and arrest type indicates that distributional differences are most distinct for sex-related offenses. Among the nonschool officers in Table 2 who were arrested for a sex-related offense (n = 399), 45% targeted unrelated minors (compared with 75% in the case of SROs). Table 2 also reveals there are no significant differences between SROs and nonschool officers for alcohol-related, drug-related, and profit-motivated offenses. These bivariate statistical tests should be interpreted with caution, however, given the low expected contingency table cell counts for certain incidents including drug-related offenses involving SROs.

Discussion
Our methodology produced data on police crime that would be difficult, if not impossible, to obtain using other methods. There are no official data or comprehensive statistics available on the phenomenon of police crime (Kane, 2007;Stinson et al., 2010). This research compliments existing studies in the "newsmaking criminology" tradition whereby criminologists interpret news reports about crime (Barak, 1988(Barak, , 2007. Studies in newsmaking criminology most commonly involve the analysis of news content to gain knowledge about the nature of crime-related media coverage, but news content can also provide valuable information on the nature of the criminal behavior that underlies the media coverage (see, for example, Morris, 2010). The news reports we analyzed provide an unparalleled amount of information on a large number of crimes committed by police officers, including those by SROs. There are four primary limitations with the current data. First, the research is constrained by the content and quality of information provided on each case in the news articles analyzed. The amount of information on each case varied, and data for some of the variables of interest were missing in some cases. Second, the data are limited to cases involving an official arrest. Data on cases of police crime that failed to come to the attention of law enforcement or prosecutors, or failed to result in arrest were unavailable. Third, these data are the result of a filtering process that includes the exercise of discretion by media sources. Media sources exercise discretion in terms of the types of news stories covered and the nature of the content devoted to particular news stories (Carlson, 2007). This discretion raises concerns that media accounts of police crime may not be representative of actual officer misconduct. Some research suggests, however, that news coverage of certain types of officer misconduct is consistent with official police records of these events (e.g., Ready, White, & Fisher, 2008). Research also suggests that police agencies are not especially effective at controlling media accounts of officer misconduct (Chermak, McGarrell, & Gruenewald, 2006). Fourth, there are a relatively small number of arrest cases involving SROs (n = 52) in our unique sample. Among the limited number of studies that have focused on SROs, most were undertaken with small or nonprobability samples (e.g., Kupchik & Bracy, 2010b). Despite this weakness, these exploratory studies have been instrumental in advancing research on SROs. This descriptive study aims to do the same by focusing on the implications of the findings for SRO programs.
It is estimated that as many as one half of public schools in the United States have at minimum a part-time police presence (Raymond, 2010), and the number of police officers deployed in schools will certainly increase in coming years given that state and federal governments have made SROs a budgetary priority after the 2012 shooting at Sandy Hook Elementary School (Brenchley, 2013). The placement of SROs in many schools has not been accompanied by a corresponding investment in research that systematically focuses on school police officers. We contribute to this sparse line of research by examining news articles to identify the general types of offenses for which SROs are arrested. SROs are an important segment of law enforcement to study officer misconduct because of some level of role ambiguity, limited training, and the differing institutional norms and priorities of schools versus local law enforcement agencies.
Our results are encouraging insofar as they indicate that very few SROs are arrested for a criminal offense. More specifically, recent estimates suggest that there are more than 15,000 SROs nationwide (James & McCallion, 2013), yet we identified only 52 school police officers who were arrested for a criminal offense over a 7-year period. The troubling aspect of our findings is that when these few SROs were arrested, they were most often arrested for sex-related offenses that typically involved female students enrolled at the schools they patrolled or frequented. All of the studentinvolved cases in our study represent instances where an SRO engaged in inappropriate behavior by making sexually explicit remarks or initiating sexual contact with a student.
As mentioned, SROs are often encouraged to assume responsibilities not directly related to law enforcement including counseling and mentoring students (e.g., . The activities provide students with nonadversarial opportunities to interact with police officers (Jackson, 2002). Many of these encounters occur outside the classroom in small group or one-on-one situations where informal conversations about matters unrelated to school regularly occur. Indeed, SROs are often deployed in large urban public schools attended by students with unstable home lives (Lawrence, 2006). SROs in such schools likely encounter emotionally charged students whose words and actions are attributable to circumstances away from school. These encounters often result in the SRO taking on a nonpolice role comparable to that of a social worker or family member. As one SRO in a large urban school put it, his job "goes beyond policing" because he is often a "father figure," "uncle," or "everything" to some students (Snyder, 2011, p. 5).
Teachers are also often much more than educators. They often self-assess their own interactions with students to make sure they do not become too invested or friendly with some students (Aultman, Williams-Johnson, & Schutz, 2009). These questions can be difficult to answer when research consistently finds that students who express a stronger connection to their teachers perform better academically (Andrzejewksi & Davis, 2008). When SROs take on duties unrelated to law enforcement such as counseling students and teaching classes, they too are likely confronted by the same relational questions. It is perhaps not surprising then that we found that the types of crimes for which SROs are arrested are more consistent with the types of offenses committed by teachers (e.g., American Association of University Women [AAUW], 2001;Shakeshaft, 2004) rather than police officers (e.g., Stinson et al., 2012). The current findings generally revealed that the circumstances surrounding offenses for which nonschool police officers were arrested were often much different than the crimes committed by SROs. The community-based police officers were in particular less likely than SROs to be arrested for sex-related offenses, but they were more likely to be arrested for violence-related incidents such as a domestic dispute.
Our findings, along with research on teacher misconduct (e.g., Shakeshaft, 2004), highlight the importance of having school personnel establish and maintain relational boundaries with students. Indeed, education experts have strongly encouraged school administrators for decades to formally address student relational boundaries when training teachers in part because self-report research indicates that students are regular victims of teacher-perpetrated sexual misconduct (AAUW, 2001;Hutchings, 2009). The problem, however, is that such relational training is sparingly administered to teachers (Shakeshaft, 2004). The same likely holds true for SROs. Although most states do not require that SROs "receive any specialized training before becoming an SRO or working with youth" (Thurau & Wald, 2010;p. 998), research indicates that most SROs do eventually complete some specialized training-usually no more than 40 hr-such as participating in courses offered by NASRO and/or becoming certified to teach Drug Abuse Resistance Education (DARE) or other prevention-related curriculums (see, for example, Finn, Townsend, Shively, & Rich, 2005). A tertiary examination of the content of such specialized training opportunities suggests that relational boundaries with students are sparsely addressed if at all.
Our findings also suggest that such training should serve experienced police officers in addition to officers who are new to the job. Most of the SROs arrested for sexual incidents involving students had worked in schools for more than 2 years. Perhaps issues with student relational boundaries become more salient as officers spend more time in schools and have greater opportunities to informally interact with students. Prior research indicates that SROs assume more non-law-enforcement responsibilities the longer they remain at a school (e.g., . The news articles that covered the arrests of SROs for sexual misconduct often raised pointed questions about the supervision of these officers. Although many of these sexual incidents took place away from school, research with educators has shown that when sexual encounters occur between teachers and students, there is often an escalation of events where an emotional attachment becomes a physical one (Johnson, 2010). As that happens, there may be opportunities to identify questionable behaviors that occur at school-such as repeated closed-door or after-school meetings-and intervene (Shakeshaft, 2004). Although all school personnel should be cognizant of potential misbehavior and related warning signs, immediate supervisors bear a particular responsibility. In the case of SROs, their immediate supervisor is often located off-site at a local police department or tends to balance multiple responsibilities that severely limit the amount of time they spend at any one school. In fact, of the 19 SRO programs that Finn, Shively, and colleagues (2005) evaluated, they noted that most police supervisors "do not conscientiously supervise their SROs" (p. 47). In instances where an SRO's supervisor is located off-site or supervising multiple SROs deployed across different schools, it is all the more "essential that the SRO's police supervisor maintain regular contact with school administrators to assess the officer's progress throughout the school year accurately" (Clark, 2011, p. 98).
Many police departments have implemented comprehensive personnel assessment systems to collect data on all sworn officers within their agency (Walker, Alpert, & Kenney, 2000). These early intervention systems could be used to identify officers at risk of engaging in criminal behaviors and sexual misconduct (International Association of Chiefs of Police, 2011). Police and school administrators are in a position to adopt such promising personnel systems for SROs, but the effectiveness of these systems is certainly reliant on the consistent oversight and supervision of SROs.
The current findings indicate that SROs are rarely arrested for criminal misconduct. These infrequent incidents typically involved sexual misconduct directed at a student. The nature of these incidents suggests there may be benefit to having SROs receive some basic training on relational boundaries in the school environment, something that has been strongly advocated for teachers given the sexual nature of offenses they commit (Shakeshaft, 2004). In addition, preventing infrequent acts of misconduct by SROs is likely difficult given that most police officers presumably have substantial autonomy in schools, but the use of early intervention systems that ensure active and consistent supervision of SROs may facilitate the identification of problematic behavior before it escalates to a serious nature. The effective training and monitoring of SROs will be a source of continued discussion and debate as state and federal governments make greater investments in the placement of police officers in schools.

Authors' Note
The opinions, findings, and conclusions or recommendations expressed in this article are those of the authors and do not necessarily reflect those of the Department of Justice.

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) disclosed receipt of the following financial support for the research and/or authorship of this article: This project was supported by Award No. 2011-IJ-CX-0024, awarded by the National Institute of Justice, Office of Justice Programs, United States Department of Justice