Content uploaded by Matthew T. Theriot
Author content
All content in this area was uploaded by Matthew T. Theriot on Mar 18, 2018
Content may be subject to copyright.
School resource officers and the criminalization of student behavior
Matthew T. Theriot ⁎
College of Social Work, University of Tennessee, 302 Henson Hall, 1618 Cumberland Avenue, Knoxville, TN 37996-3333, United States
abstractarticle info
As school resource officer (SRO) programs continue to be widely implemented, there is concern that an
increasing police presence at schools will “criminalize”student behavior by moving problematic students to
the juvenile justice system rather than disciplining them at school. If true, this has serious implications for
students and schools; yet research on this topic is limited and the discourse is often based on speculation or
anecdotal evidence. To address this issue, this study evaluated the impact of SROs on school-based arrest
rates by comparing arrests at thirteen schools with an SRO to fifteen schools without an SRO in the same
district. Poisson and negative binomial regression models showed that having an SRO did not predict more
total arrests, but did predict more arrests for disorderly conduct. Conversely, having an SRO decreased the
arrest rate for assault and weapons charges. Implications of these findings for understanding SROs and their
role in criminalizing student behavior are discussed.
© 2009 Elsevier Ltd. All rights reserved.
Introduction
Following a handful of high-profile incidents of lethal school v iolence
in the 1990s, growing attention has been given to the protection of
students and faculty at school. Though contrary to statistics showing
that school crime nationally was declining, relatively rare, and usually
nonviolent (Dohrn, 2002; Jackson, 2002; Miller, Gibson, Ventura, &
Schreck, 2005), school shootings like those in Littleton, Colorado, and
Jonesboro, Arkansas, fed growing public fear of juvenile and school
crime. This ledto the rapid implementation and expansion of numerous
school security measures, ranging from the use of high-tech security
devices like metal detectors and surveillance cameras to student-driven
peer mentoring programs, school resource officer programs, and
punitive zero-tolerance policies for disciplinary infractions (Eisenbraun,
2007).
Empiricalevaluations of these various security strategies are limited,
have varying levels of methodological rigor (D. C. Gottfredson, 2001),
and often report conflicting findings (Brown, 2005). For example, while
research done by Green (1999) and Johnson (1999) reported that metal
detectors and school resource officers, respectively, enhanced school
security, Schreck, Miller, andGibson (2003) found them to be ineffective
while Mayer and Leone (1999) found that they actually led to more
school disorder. Moreover, while development of a positive school
environment is considered critical to violence prevention (Eisenbraun,
2007; D. C. Gottfredson, 2001), common security measures like strip
searches and use of undercover agents actually lower students' self-
esteem and cause emotional distress (Hyman & Perone, 1998).
According to Beger (2003), such strict measures foster an “adversarial
relationship”between students and school personnel and interrupt
student learning (p. 340). Conflicting findings like these make it difficult
to determine what works to prevent school violence while showing
clearly that more research is needed (Brown, 2005; Eisenbraun, 2007).
Criminalizing student behavior
Moreover, several criminologists and legal scholars have expressed
concerns that some strategies designed to make schools safer—
particularly the growing number of school resource officers (SROs)—
might actually criminalize student behavior and lead to a substantial
increase in the number of school-based arrests. SROs are sworn law
enforcement officers assigned full-time to patrol schools. As they
become more common on school campuses, it is argued, discipline
problems traditionally handled by school principals and teachers now
are more likely to be handled by a school police officer (Hirschfield,
2008). Thus, as a scuffle between students becomes assault or disrupting
class becomes disorderly conduct, it is expected that the number of
youths referred from public schools for delinquent and criminal
prosecution will climb, especially for behaviors that pose no legitimate
threat to school safety (Beger, 2003; Brown, 2006; Dohrn, 2001, 2002;
Hirschfield, 2008; Lawrence, 2007). According to Dohrn (2002),
American schools have been transformed into “prisonlike”facilities,
replete with locked doors, metal detectors, camera surveillance, and
greater police presence (p. 283).
More information on this matter is urgently needed given the
implications of criminalization for students, schools, juvenile and
criminal justice systems, and communities. Students removed from
school miss educational opportunities. These students also face
humiliation and stigma from classmates and teachers after being led
from school in handcuffs. Being stigmatized and labeled as an offender
Journal of Criminal Justice 37 (2009) 280–287
⁎Tel.: +1 865 974 8109; fax: +1 865 974 3351.
E-mail address: mtheriot@utk.edu.
0047-2352/$ –see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcrimjus.2009.04.008
Contents lists available at ScienceDirect
Journal of Criminal Justice
also might result in greater scrutiny, surveillance, and questioning
from school staff and security. This type of regular suspicion and
harassment could lead some youth to drop out of school (Scheffer,
1987) and could even contribute to a rise in community and school
crime rates. Furthermore, having a criminal record might negatively
impact access to jobs and institutions of higher education (Dohrn,
2001).
Currently, however, data are limited and confidentiality rules
protecting juvenile court records make it difficult to calculate the
number of arrests made by SROs (Center on Juvenile and Criminal
Justice, 2000). Much of the discourse about criminalization is based on
speculation, anecdotal evidence, or descriptive statistics. The present
study therefore contributed to the literature by quantifying and
evaluating the impact of school resource officers on school arrest
rates. By comparing schools with an SRO to schools without an SRO in
the same district, this study sought to identify differences in the
number of arrests and types of charges. Such comparisons are critical
for understanding the effect of SROs on school arrests while also
considering their possible role in criminalizing behavior.
School resource officer programs
While a few school resource officer (SRO) programs have existed
since the mid-1900s, the number has swelled since the late 1990s.
Today, these officers represent a significant and popular trend in
school violence prevention. Following the fatal shooting of a school
principal by a middle school student, for example, Tennessee
Governor Phil Bredesen announced that he would “look into making
the SRO job a part of the framework for every public school”(Kovac,
2006, p. B7). It is not surprising then that, according to the National
Association of School Resource Officers (NASRO, n.d.), a member
service organization boasting about 10,000 members, school-based
policing is the fastest growing area of law enforcement. While it is
difficult to know the exact number of school resource officers, it is
estimated that there might be more than 20,000 law enforcement
officers patrolling schools in the United States (Brown, 2006).
School resource officers in the United States (also known as school
police officers or school liaison officers) typically are employed by a local
law enforcement agency and assigned to work in a school or schools.
They perform traditional law enforcement functions like patrolling
school buildings and grounds, investigating criminal complaints,
handling students who violate school rules or laws, and trying to
minimize disruptions during the school day and at after-school
activities (Lawrence, 2007). SROs also are charged with educating
students and school staff about crime and violence prevention, acting
as mentors to students, and helping to improve the school environ-
ment (Rich & Finn, 2001). Officers usually are armed and often in
uniform. While some schools utilize area law enforcement officers on a
part-time or irregular basis, true SROs frequently have received
extensive training in school-based policing and are a consistent fixture
at the school. For these reasons, Rich and Finn urge clear differentiation
between official SROs and other “sworn officers who focus exclusively
on law enforcement activities in schools”(p. 4).
School resource officers and criminalization
To date, most published research on school resource officers or
school-based policing focused on the implementation of such
programs at schools (e.g., Briers, 2003) or on describing officers'
duties while at school (e.g., Finn, Shively, McDevitt, Lassiter, & Rich,
2005; Rich & Finn, 2001). There also was literature discussing the
development of collaborative partnerships between school and law
enforcement personnel (e.g., May, Fessel, & Means, 2004; Patterson,
2007) as well as students' attitudes about school police officers
(Hopkins, 1994; Hopkins, Hewstone, & Hantzi, 1992; Jackson, 2002).
Though such research, commentaries, and process evaluations are
essential for understanding school resource officers, they rarely
discussed the notion of criminalization or provided data about arrests
made at school.
Nevertheless, in support of the criminalization hypothesis, there
were numerous published reports documenting incidences in which
students were arrested for seemingly minor offenses. For example,
Rimer (2004) described how a fourteen-year-old student was arrested
and detained for violating a school's dress code. The Center on Juvenile
and Criminal Justice (2000) similarly described how a fourteen-year-
old disabled student in Florida was arrested and charged with felony
robbery after stealing $2 from a classmate. The student was held for
several weeks in an adult detention center before charges were
dropped. In another example, a twelve-year-old student in Louisiana
was arrested and charged with making terroristic threats and detained
for two weeks after telling classmates in the school's lunch line that he
would “get them”if they ate all of the potatoes.
Authors also noted the rising number of school-based arrests in
some districts as validation of the idea that SROs contribute to
criminalizing behavior. Rimer (2004) reported that the number of
school-based arrests in one Ohio county increased from 1,237 in the
year 2000 to 1,727 in 2002. According to juvenile court officials, most of
these arrests were for minor offenses or unruly student behavior while
only a very small percentage was for serious threats to school safety. A
similar escalation was reported in Miami-Dade County, Florida, where
the 2,345 school arrests in 2001 were a threefold increase over the
number of school arrests in 1999. The vast majority of these arrests
were for simple assaults and disorderly conduct. Given that both
locations utilized SROs extensively at district schools, these figures
make a compelling statement about the possible criminalization of
student behavior. The number of arrests made specifically by an SRO is
unknown, however, and such figures can be somewhat misleading
since it is unusual for all schools in a district or county to have regular
SRO involvement. In Miami-Dade County schools, for example, school
resource officers are assigned to middle schools and high schools only,
while police service is provided to elementary schools as needed.
Focusing on SROs exclusively, Johnson (1999) studied eighteen
SROs recently placed at nine high schools and eighteen middle schools
in one district in the southern United States. These officers made 145
arrests in a five-month period, including ninety-seven arrests invol-
ving drugs and forty-nine involving weapons. Without a comparison
group though, it is difficult to know if this number of arrests is high or
unusual for these schools. Similarly, Dohrn (2001) reported the
number of arrests from one Chicago-area high school with an assigned
police officer. There were 158 arrests during the 1996–1997 school
year, including sixty-one for pager possession, twenty-one for
disorderly conduct, and sixteen for non-firearm weapon possession.
Yet, it is unclear if these data from a single school generalize to other
locations since officers' and school principals' discretion as well as the
school climate will influence decisions to arrest. In contrast, however,
studies citing national statistics likewise were limited because they
included data from schools with and without an SRO.
While more empirical research is needed to evaluate school-based
arrests made by SROs, there are practical and conceptual reasons to
suggest that SROs play an important role in introducing more and
more students to the juvenile justice system. First, most crime
occurring at schools historically has not been reported to police
(Elliott, Hamburg, & Williams, 1998), yet having a police officer
available and accessible at school facilitates reporting. One likewise
would expect more crime to be witnessed by law enforcement when
they are present daily at school. Along these same lines, as SROs
assume increasingly more responsibility for handling school dis-
ciplinary problems, it is reasonable to expect that more and more
situations will be resolved with an arrest now than in the past
(Hirschfield, 2008). Finally, Bailey (2006) described SROs as having a
“quasi-law enforcement role”in the school (p. 38). This complicates
security issues and gives officers more freedom to search students
281M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
and detect contraband. Specifically, while the standard to search a
suspect for police officers patrolling the streets includes probable
cause and/or issuance of a warrant, the standard for school officials as
determined in New Jersey v. T.L.O. (1985) is reasonable suspicion only.
Therefore, an officer acting at the request of school officials—and thus
serving as an agent of the school—operates under a less stringent
standard for searching students (Bailey, 2006).
The present study
For all of these reasons, it was hypothesized that schools with an
SRO have more total arrests and more arrests for charges like
disorderly conduct and assault than schools without an SRO. To
evaluate the role of SROs in school-based arrests, this study compared
arrests occurring at middle schools and high schools with an SRO to
those occurring at schools without an SRO in the same district. While
school resource officers often are placed at all schools in a district (e.g.,
Johnson, 1999), the SRO program studied here was implemented by
one metropolitan police department within the school district's
catchment area. Thus, SROs in this district were not assigned to
schools based on a school's need, history of violence, or demographics
but rather by geography only and a school's location inside or outside
of city limits. One school resource officer therefore was assigned to
each of the seven middle schools, five high schools, and one alternative
school within the city limits regardless of the school's past experiences
with violence or delinquency. Consistent with standards promoted by
the NASRO (n.d.), these officers received extensive training in school-
based law enforcement, teaching skills, and school violence prevention
programming. This police department served the largest city in the
county with a population of nearly 200,000 residents.
The remaining seven high schools, seven middle schools, and one
alternative school in the district were outside city limits and thus did
not have an official, trained school resource officer assigned to them.
Deputies employed by the county sheriff's department were respon-
sible for these schools. Unlike the city schools, however, these
deputies focused exclusively on law enforcement duties at schools.
They received less training in school-based policing, often were
assigned to more than one school in an area, and were not expected to
make presentations to students or faculty or be a visible or proactive
presence in the schools. Instead, when present at a school, deputies
typically were stationed at the school's main office and charged with
assisting the school principal in handling disciplinary referrals as
needed. This activity contrasted markedly with the actions and level of
involvement expected from the school resource officers. Such an
organizational structure, wherein roughly half of the district's middle
and high schools had an SRO and half did not and SROs were assigned
based on school location rather than need, provided a unique
opportunity to study the alleged criminalization of students by SROs.
Methodology
Sample and study design
To evaluate the impact of school resource officers on arrests at
school, this study compared the number of arrests in three consecutive
school years at thirteen schools with an SRO and fifteen schools
without an SRO in one school district. Analyzing multiple years of data
neutralized anomalies that might arise from a single year of data, while
comparing schools in the same district controlled for variations in
policies and guidelines that might exist across different districts. The
district covered one county in the southeastern United States and
boasted almost ninety public schools, including fourteen middle
schools (grades six through eight), twelve high schools (grades nine
through twelve), and two alternative schools serving middle and high
school students with behavioral or mental health problems. These
twenty-eight schools formed the sample for this study. District schools
were located primarily in urban and suburban settings. There were
more than 53,000 students enrolled in all district schools with
approximately 13,000 middle school students and 16,000 high school
students. The majority of students district-wide were Caucasian
(81 percent), followed by African American (15 percent), and Hispanic
students (2 percent). Approximately 40 percent of all students
received a free or reduced school lunch, while 13 percent had an
accommodated disability.
Measures
Dependent variables
Seven dependent variables were analyzed here to assess differ-
ences in arrests between schools with and without an SRO. These
variables were counts of the total number of arrests at a school during
the three years, the number of arrests with a disorderly conduct
charge, the number of arrests with an assault charge, the number of
arrests involving possession of drugs or drug paraphernalia charges,
the number of arrests for possession of alcohol or public intoxication
charges, the number of arrests involving a weapon on school property,
and the number of arrests involving all other types of charges. To
collect these data, all delinquency petitions filed at the county's
juvenile court from three consecutive school years (2003–2004,
2004–2005, and 2005–2006) were reviewed to identify those arrests
occurring at district middle schools and high schools during normal
school hours or at after-school activities. Since all juvenile arrests in
the county were processed through the juvenile court regardless of
school location and departmental jurisdiction, it was an ideal place to
obtain comprehensive and consistent data about delinquency across
district schools.
During the three school years, there were 1,012 arrests involving
878 different students at district middle and high schools. To assess
differences between schools, arrests were aggregated to generate a
duplicated count by school. In a duplicated count, students with
multiple arrests are counted multiple times. While an unduplicated
count (in which students are counted only once regardless of how
many times they are arrested) is expected to underestimate the
frequency of arrests at school, duplicated counts provide the most
accurate measure of how often arrests are used to control discipline
problems. For this reason, Raffaele Mendez, Knoff, and Ferron (2002)
strongly encouraged the use of duplicated counts in school discipline
research. Similar duplicate counts were generated for each delinquent
charge of interest. Almost 90 percent of all arrests (n=893) resulted
in a single charge, while 10 percent (n=119) yielded multiple
Table 1
School and delinquency characteristics for SRO and non-SRO schools (N=28)
Schools with a school
resource officer (n=13)
Schools without a school
resource officer (n=15)
Mean±S.D. Mean ±S.D.
School characteristics
Total students 992.2 ±4 93.7 1115.9± 513.1
Percent economic disadvantage⁎60.4±23.9 30.0 ±17.7
Percent ethnic minority
students⁎⁎
33.8± 23.7 10.5± 6.3
Percent attendance 92.0± 3.4 93.9±3.7
Rates of arrests and charges per one hundred students
Total arrest rate 11.5± 25.1 3.9 ±6.9
Alcohol/public intoxication
charge rate
0.5 ±0.9 0.3 ±0.4
Assault charges 1.0± 1.7 0.7± 1.5
Disorderly conduct charges 8.5±21.1 1.8± 5.6
Drug-related charges 1.2± 2.1 0.8± 0.5
Other charges 1.1 ± 1.0 0.6 ±1.0
Weapons charges 0.1± 0.2 0.2 ± 0.3
⁎Mean difference is significant; F(1,27) = 14.87; p = .0 01.
⁎⁎Mean difference is significant; F(1,27) =13.49; p= .0 01.
282 M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
charges. For these arrests, each type of charge was counted separately,
and as a result, the number of charges exceeded the number of arrests.
Independent variables
Independent variables came from annual reports published by the
state's Department of Education. These reports are publicly available
and show summary information about each school district in the state,
as well as information on all individual schools within a district.
Variables in the present study were averages calculated from the three
years of data and included total enrollment at each school,percent of
the student body that was ethnic minority (non-Caucasian), percent of
the student body that was economically disadvantaged (a measure of
school poverty defined as the percentage of students receiving a free
or reduced lunch at school), and attendance rate (the average number
of days students attend school divided by the average number of days
the students are enrolled).
These variables were selected because they had been linked to
school discipline outcomes in other studies. In studies of school
exclusion (out-of-school suspension and expulsion), Bruns, Moore,
Stephan, Pruitt, and Weist (2005) found that the percent of students
in poverty at a school was positively correlated with the out-of-school
suspension rate, while school enrollment and mean school attendance
rate were negatively correlated with this rate. In a similar study of
rates, Raffaele Mendez et al. (2002) found that school level variables
like percent of students receiving free lunch and percent African
American were positively correlated with out-of-school suspension
rate, while percent Caucasian and percent Hispanic were negatively
correlated. Brown (2006) likewise summarized research showing a
relationship between school poverty and size and crime rates.
Data analyses
Independent variables are presented on Tabl e 1 andcomparedusing
analysis of variance (ANOVA) tests with a Bonferroni adjustment for
multiple comparisons. All data met normalityassumptions. Since school
resourceofficers were placed at schools based on geography rather than
random assignment, these comparisons were done to identify sig-
nificant differences between the two sets of schools. To better isolate the
impact of SROs on arrests, differences in the independent variables must
be controlled for in subsequent regression models. As shown in Table 1,
data suggested that schools with an SRO had more povertyand a larger
percentage of ethnic minority students. Whereas ethnic minority
students often are overrepresented in lower socioeconomic groups
(Eisenbraun, 2007), these two variables expectedly are highly correlated
(r=.81; pb.001). Therefore, to avoid multicollinearity problems that
arise when covariates are highly correlated (and given this study's
sample size), only one was included as an independent variable in the
subsequent regression models. The decision was made to use percent of
students with economic disadvantage because it represented a more
significant difference in this study, had been explicitly linked to school
problems in other studies (e.g., Bruns et al., 2005), and problems
confronting ethnic minority students at school often are embedded in
poverty and socioeconomic issues.
1
As Skiba, Michael, Nardo, and
Peterson (2002) noted in regard to school exclusion, the connection
between race and socioeconomic status (SES) in the United States is
undeniable and “increases the possibility that any finding of dispro-
portionality [in school exclusion] due to race is a by-product of
disproportionality associated with SES”(p. 321). Tabl e 1 also displays
the mean arrest and charge rates per one hundred students at schools
with and without an SRO. These rates for total arrests and all specific
charges of interest were calculated by dividing the total number of
arrests or charges in the three-yearstudy period by theaverage number
of students at school for the three years divided by one hundred.
Tables 2–5show the results of a series of negative binomial and
Poisson regression models. These types of statistical analyses are ideal
for count data (like number of arrests at school) that have nonnegative
integers, are highly skewed since some counts will be very low (i.e.,
some schools will have few arrests), and have heteroscedastic error
terms. Tests for overdispersion (the variance is greater than the mean)
showed that negative binomial regression was appropriate for all
dependent variables except the number of arrests involving weapons
charges. For this variable, Poisson regression was used.
The study's modest sample size (n= 28 schools) limited the
number of independent variables that could be included in the
regression models. Though there is still much debate about the
minimum sample size needed per independent variable in multi-
variate analysis (Knofczynski & Mundfrom, 2008), this study used the
popular rule of thumb that one independent variable per ten sample
members is appropriate (Harrel, Lee, Matchar, & Reichert, 1985;
Peduzzi, Concato, Kemper, Holford, & Feinstein, 1996; Vittinghoff &
McCulloch, 2007). Vittinghoff and McCulloch suggested this rule
might be too conservative, yet other research has found that this rule
limits bias and maintains the validity of multivariate models (Harrel
et al., 1985; Peduzzi et al., 1996). Specific to this study, three regression
Table 2
Negative binomial regression results for total arrests at schools (N= 28)
Model 1 Model 2 Model 3
Coeff. SE Coeff. SE Coeff. SE
Independent variable
SRO at school 1.091⁎⁎ .438 0.055 .404 0.328 .868
Percent economic disadvantage
at school
––0.039⁎⁎⁎ .008 0.042⁎⁎⁎ .012
SRO x percent economic
disadvantage (interaction term)
––– –−0.006 .017
Likelihood-
ratio
Likelihood-
ratio
Likelihood-
ratio
X
2
=5.66 X
2
=23.90 X
2
=24.03
p = .02 p b.001 pb.001
⁎pb.10.
⁎⁎pb.05.
⁎⁎⁎pb.001.
Table 3
Negative binomial and Poisson regression results for arrests involving assault and weapons charges at schools (N=28)
Assault Weapon on school property
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE
Independent variable
SRO at school 0.262 .468 −0.740⁎.385 0.849 .688 −0.225 .312 −1.304⁎⁎ .457 −1.295 .931
Percent economic disadvantage at school ––0.038⁎⁎⁎ .008 0.059⁎⁎⁎ .011 ––0.032⁎⁎⁎ .008 0.032⁎⁎ .013
SRO x percent economic disadvantage (interaction term) ––– –−0.037⁎⁎ .014 ––– –−0.000 .017
Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio
X
2
=0.31 X
2
=18.17 X
2
=23.79 X
2
=0.51 X
2
=15.15 X
2
=15.15
p=.58 p b.001 pb.001 p = .48 p b.001 p= .001
⁎pb.10.
⁎⁎pb.05.
⁎⁎⁎pb.001.
283M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
models were presented for each dependent variable described above.
This multi-model structure allowed for evaluating the impact of SROs
on arrests with and without controlling for other independent
variables. The first model included only one independent variable—
having an SRO at school or not (coded as SRO at school= 1, no
SRO= 0). The second model included this variable plus percent of
students with economic disadvantage. The final model then added the
interaction (SRO x school poverty) of these two variables. This term
was added to assess differences in arrests as poverty levels changed at
schools with an SRO. This was an important consideration given
speculation that the criminalization of student behavior is especially
acute at lower socioeconomic schools.
In all models, the average number of students at a school during the
three years divided by one hundred was included as an exposure
variable. This controlled for differences in the number of students
across all schools. Dividing the average by one hundred helped in
translating the output to more common and easily understood
terminology since regression coefficients then can be reported as a
percent change in the arrest rate “per one hundred students.”
Regression coefficients were interpreted using the standard formula
where a one-unit change in an independent variable equals a 100(e
b
-1)
percent change in the dependent variable (D'Alessio & Stolzenberg,
2003; DeMaris, 1995; Hannon & Cuddy, 2006). As a final comment, it is
important to note that having an SRO at school or not is a dichotomous
variable while percent of students with economic disadvantage is a
continuous variable wherein values can range from 0 to 100 percent.
Comparisons of the two variables and their resulting rates hence should
be made cautiously since the magnitude of change may vary
dramatically across the two different types of variables.
Results
Comparisons of the school characteristics presented in Table 1
show that a larger percentage of students at schools with a school
resource officer (SRO) had economic disadvantage compared to
schools without an SRO. These schools also had a larger percentage
of ethnic minority students. Regarding delinquent arrests, there were
216 more arrests at schools with an SRO (n= 614) than at comparison
schools (n=398). The most common charge at SRO schools was
disorderly conduct (n=361) followed by other charges (n= 101) and
drug-related charges (n=98). At those schools without an SRO, the
most common charges were drugs (n=138), then disorderly conduct
(n=77), and possession of alcohol and public intoxication (n=72).
Among the forty-two arrests district-wide for possessing a weapon,
twenty-three involved a knife, twelve involved a firearm, and the
remaining seven involved items like a copper pipe, metal baton, or box
cutter. Across all schools, the most common charge in the other
category was trespassing (n=38 arrests), followed by theft (n=24),
and vandalism (n=17).
Without controlling for school poverty level, the presence of an
SRO gives a 197.7 percent increase in the rate of arrests per one
hundred students (Model 1). Yet, as shown in Model 2 on Table 2,
when economic disadvantage is added to the regression equation,
having an SRO at school ceases to be a significant predictor of arrests.
Instead, for each one percentage point increase in economic
disadvantage at a school, the rate of arrests per one hundred students
increases by 3.98 percent (without interaction term) and 4.29 percent
(with interaction term). The interaction is not significant in Model 3,
indicating that the number of arrests does not change as poverty levels
change at schools with an SRO.
Regarding specific charges, though not significant when alone
(Model 1), Model 2 in Table 3 shows that having an SRO at school leads
to a 52.3 percent decrease in the rate of arrests involving assault
charges per one hundred students when controlling for the level of
economic disadvantage at school. The same model also shows that as
economic disadvantage increases by one percentage point, this rate
increases 3.9 percent. In the third model, with both independent
variables and the interaction term, each one percentage point increase
Table 4
Negative binomial regression results for arrests involving drugs and alcohol/public intoxication charges at schools (N =28)
Drugs Alcohol/public intoxication
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE
Independent variable
SRO at school 0.064 .315 −0.162 .379 0.012 .733 −0.020 .439 −0.131 .576 −1.027 .998
Percent economic disadvantage at school ––0.008 .0 08 0.011 .013 ––0.003 .011 −0.016 .022
SRO x percent economic disadvantage (interaction term) ––––−0.005 .016 ––––0.026 .025
Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio
X
2
=0.04 X
2
=0.98 X
2
=1.06 X
2
=0.00 X
2
=0.09 X
2
=1.22
p=.84 p= .61 p= .79 p= .96 p= .96 p=.75
⁎pb.10.
⁎⁎pb.05.
⁎⁎⁎pb.001.
Table 5
Negative binomial regression results for arrests involving disorderly conduct and other charges at schools (N= 28)
Disorderly conduct Other
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE Coeff. SE
Independent variable
SRO at school 1.614⁎⁎ .703 0.825⁎.482 3.034⁎⁎ 1.249 0.798⁎⁎ .373 −0.242 .343 0.178 .668
Percent economic disadvantage at school ––0.070⁎⁎⁎ .011 0.098⁎⁎⁎ .020 ––0.031⁎⁎⁎ .007 0.038⁎⁎ .011
SRO x percent economic disadvantage (interaction term) ––– –−0.049⁎.026 ––– –−0.010 .014
Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio Likelihood-ratio
X
2
=4.64 X
2
=30.72 X
2
=34.83 X
2
=4.04 X
2
=20.60 X
2
=21.13
p =.03 p b.001 pb.0 01 p= .04 p b.001 pb.0 01
⁎pb.10.
⁎⁎pb.05.
⁎⁎⁎pb.001.
284 M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
in economic disadvantage at a school increases the rate of arrests
involving assault charges by 6.1 percent while a rise in economic
disadvantage at schools with an SRO decreases this rate by 3.6 percent.
Similar patterns exist regarding arrests involving possession of a
weapon on school property. For this charge, when controlling for
economic disadvantage, schools with an SRO have a 72.9 percent
decrease in the rate of arrests per one hundred students. Conversely,
each one percentage point climb in school poverty increases this rate
of arrest by 3.3 percent. This same effect is evident in the full model
with the interaction term.
Table 4 shows that neither school resource officers nor poverty
predicts changes in the rate of arrests involving drug or alcohol and
public intoxication charges. The regression coefficients associated
with having an SRO at school generally are negative, but none
approach a level of statistical significance. The interaction term also is
not significant for either dependent variable.
Finally, results presented in Table 5 show that school resource
officers dramatically increase the rate of arrests with disorderly
conduct charges with and without controlling for school poverty.
Specifically, without controlling for economic disadvantage at schools
(Model 1), having an SRO yields a 402.3 percent increase in this arrest
rate per one hundred students. This percent increase remains large
even after controlling for poverty and the interaction of SROs and
poverty. As Models 2 and 3 illustrate, the presence of an SRO at school
increases the rate of arrests involving disorderly conduct charges by
128.2 percent and 1978.0 percent, respectively. These two models also
show that a one percentage point rise in economic disadvantage
increases the arrest rate by 7.3 percent when controlling for the
presence of an SRO, and 10.3 percent when controlling for having an
SRO and the interaction term. Interestingly, regarding the interaction
term, a one-percentage point increase in poverty at schools with an
SRO equals a 4.8 percent decrease in the arrest rate per one hundred
students.
Schools with a resource officer have a 122.1 percent increase in the
rate of arrests involving other charges per one hundred students when
analyzed without other independent variables. When economic
disadvantage is added to the regression models (Models 2 and 3),
however, the impact of SROs ceases to be significant. Instead, school
poverty emerges as the only significant predictor. A one percent
increase in this variable raises the rate of arrests with other charges
per one hundred students by 3.1 percent. When controlling for SROs
and the interaction term, a one-percentage point increase in economic
disadvantage increases this arrests rate by 3.9 percent.
Discussion
Evidence of criminalization
While it was hypothesized that having an SRO at school predicts
more total arrests, this hypothesis received only limited support here.
While the data presented in Table 1 implied significant differences in
the total number of arrests between SRO and non-SRO schools, such
differences were not as robust as expected. Though the presence of
SROs did predict a dramatic increase in the rate of arrest per one
hundred students independent of other variables, this variable ceased
to be significant when controlling for school-level poverty. Such
mixed results might be a function of the study's sample size since
smaller samples limited the detection of smaller effect sizes.
On the other hand, however, this potential limitation makes the
observed differences in types of charges all the more noteworthy. The
analyses revealed several interesting findings that, when considered
together, show an interesting pattern regarding the role of SROs in
school-based arrests. Primarily, the high number of disorderly conduct
incidences at SRO schools compared to non-SRO schools was consistent
with the belief that SROs contribute to criminalizing student behavior.
Havingan SRO at school significantly increased the rate of arrests for this
charge by over 100 percent even when controlling for school poverty.
Given that disorderly conduct was the most common charge in this
study, these results have serious implications for schools, law enforce-
ment agencies, and juvenile courts.
Clearly, disorderly conduct is the most subjective, situational, and
circumstantial of the charges studied here. Compared to more
objective situations like finding a youth in possession of a knife or
narcotics, the decision to interpret disruptive behavior as criminal is
done at the officer's discretion. Thus, one strategy to reduce the
number of school-based arrests is to change how officers approach
such situations. When approaching a disruptive student, for example,
an arrest should be the least preferred outcome and done only in
agreement with the teacher and school principal. Likewise, it also is
important to change teachers' and school administrators' expectations
of SRO interventions. As Dohrn (2001) described, teachers more often
are turning to police officers to handle difficult students. Teachers and
principals are ignoring the “teachable moments”that come from
student misbehavior and failing to take advantage of opportunities to
work with adolescents in need (p. 95). This is truly unfortunate since
quality education is a path to success in adulthood. Given the long-
term negative consequences that can follow removing a child from the
classroom and denying them educational opportunities, improved
classroom management skills and appropriate behavioral training for
students would seem preferable to arrest and other more punitive
outcomes.
For the remaining, more objective charges studied here, having an
SRO at school was insignificantly or negatively associated with these
outcomes. This latter result was true for assault and weapons charges,
wherein the presence of an SRO decreased the rates of arrest involving
these charges per one hundred students by 52.3 percent and 72.9
percent respectively. Such findings were counterintuitive since better
detection of weapons was expected at schools with an SRO and it was
hypothesized in the extant literature that SROs criminalize fighting by
pressing assault charges (e.g., Beger, 2003; Dohrn, 2001).
While it was not possible to determine the exact reasons for these
unexpected findings, one possible explanation is that the presence of
SROs at schools might deter certain behaviors. For instance, students
might be less inclined to carry a weapon into the school building
knowing that a law enforcement officer will be there. Likewise,
students might be less likely to fight knowing that an officer is present
and such behaviors could lead to being arrested. They might delay the
fight until after school or decide to move it away from school grounds.
Along the same lines, Astor, Meyer, and Behre (1999) found that most
school violence occurs in “unowned”places, or those locations like
hallways and parking lots that usually lack adult supervision (p. 3).
Thus, having regular police patrol in these areas might be preventing
some acts of violence and crime. Alternately, the presence of SROs at
schools might make students feel safer and thus less likely to feel the
need to carry a weapon for protection. These enhanced feelings of
safety also might contribute to better feelings about school in general, a
stronger sense of connection to school, and a better school environ-
ment that could then lead to decreased aggression and fewer fights
among students.
Arrests and economic disadvantage
The significance of school poverty to predict number of arrests was
noteworthy, especially given its high correlation with ethnicity. This
study showed that students at schools with greater economic
disadvantage had a higher number of total arrests as well as more
arrests for assault, weapons possession, disorderly conduct, and other
charges than schools with less poverty. While it has been suggested
that poverty might play a role in school-based arrests (Brown, 2006;
Dohrn, 2001), this association has not been explicitly studied. Yet,
such results were consistent with research finding that poverty is a
strong predictor of school exclusion (Cameron & Sheppard, 2006;
285M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
Raffaele Mendez et al., 2002), as well as research finding that poor and
ethnic minority youth are disproportionately involved with the
juvenile and criminal justice systems (see Hirschfield, 2008; Laub,
2002; Sampson & Lauritsen, 1997).
Reasons explaining poverty's role in predicting school-based arrests
specifically are unclear and warrant further investigation. When
examining school violence generally, however, Khoury-Kassabri, Ben-
benishty, Astor, and Zeira (2004) described the importance of assuming
an ecological perspective that considers school violence within the
context of student, school, family, and neighborhood factors. As Chen
(2008) stated, “schools are extensions of the community”(p. 302). It is
not surprising then that previous research had found links between
higher levels of community poverty, crime rates, and unemployment
and greater school crime and disorder (Chen, 2008; Khoury-Kassabri
et al., 2004; Welsh, 2001, 2003; Welsh, Stokes, & Greene, 2000). Other
suggested explanations offered in the published literature emphasized
the difficulties associated with living in poverty. Dohrn (2001),for
example, suggested that parents from lower socioeconomic back-
grounds lack the resources and influence needed to protect their
children from the juvenile justice system. Moreover, while many
families in lower socioeconomic neighborhoods have a single parent
only, two-parent families are a protective factor against delinquency
(Farrington & Loeber, 2000; D. M. Gottfredson & Snyder, 2005).
It also has been suggested that discrepancies in school discipline
result from the clash between middle-class school systems and low
socioeconomic status students. For instance, Caucasian teachers and
principals might misunderstand or misconstrue the physical commu-
nication style common among ethnic minority youth, particularly
African American youth (Raffaele Mendez & Knoff, 2003; Skiba et al.,
2002). This could lead to an unnecessarily harsh response from
teachers, school administrators, or security officers. It should be
recognized, however, that data in this study did not support that SROs
discriminate against lower socioeconomic status students. In fact,
when significant in the analyses, regression coefficients for the
interaction term showed that arrest rates declined as poverty
increased at schools with an SRO. This was somewhat counterintuitive
since research had found that lower socioeconomic status juveniles
and minority youths often had poorer attitudes toward the police and
legal system (see Hurst & Frank, 2000).
As a final comment, it was interesting to compare the types of
charges here with those reported in other studies. Consistent with
reports from Ohio and Florida summarized by Rimer (2004),
disorderly conduct was the most common charge in the present
study followed by other, miscellaneous nonviolent charges. While the
majority of arrests at the Chicago-area high school studied by Dohrn
(2001) were for pager possession, there were no such arrests during
the three years studied here. This most likely was a byproduct of the
different time periods when data were collected. During the 1996–
1997 school year (Dohrn's study), pagers were relatively new and
novel. Now, though, cellular phones are pervasive on school campuses,
most students possess at least one, and schools cannot regulate
possession like they used to. Regardless, Dohrn's point that the
majority of arrests at the school were for relatively minor, nonthreat-
ening behaviors was true in the present study too.
Limitations and future research
A critical avenue for future research is to compare the number of
arrests at a school before and after the arrival of SROs. While such
within-school comparisons are critical for understanding the impact
of SROs on a school's arrest rate, data limitations and availability in the
present study precluded these types of comparisons. Specifically, the
juvenile court providing data for this project updated its data
management system in April 2003. This involved changing manage-
ment system software, adjusting data-entry procedures, and addingor
modifying system variables. Efforts to compare arrests by school
between the old and updated systems thus may yield complicated and
unreliable results since data coding and categories varied. Along
similar lines, since this study compared school-based arrests across
schools in one school district, more research is needed to determine
how the findings generalize to other districts and regions.
Furthermore, the present sample was not sufficient to detect small
effect sizes in the data. Future research therefore should seek to
compare data from more schools located in multiple districts. Analysis
done with a larger sample would help to clarify associations in the
data, including the role of SROs to predict more total arrests at schools.
The sample size also limits the number of independent variables that
can be appropriately included in multivariate tests; so, a larger sample
size would allow for evaluating the impact of SROs on arrests while
controlling for more descriptive and demographic characteristics of
the schools. Given the few observed differences between schools in
this study combined with the fact that SROs were assigned based on
geography rather than school demographics, it was unlikely this
would have meaningfully altered the results but it is an important
consideration in future research. Nevertheless, since schools with and
without an SRO in this study did differ in characteristics that are often
associated with arrests, specifically having higher levels of poverty
and more ethnic minority students, future research is needed that
continues exploring these key variables, their relationship with other
school characteristics, and the link between these variables and higher
or lower school arrest rates.
In building on this study, future research evaluating the long-term
consequences of school-based arrests is needed. Classic labeling
theory, for example, postulates that involvement with the juvenile
justice system increases the likelihood of future delinquency (Becker,
1963). If valid for students arrested at school, such findings would have
tremendous implications for how behavior problems arehandled. This
is especially true for those juveniles arrested for relatively minor
offenses since arresting them might be creating a delinquent where
none existed before. Differences in the long-term consequences of
school-based arrest by gender, ethnicity, socioeconomic status, and
other student characteristics should be investigated. Besides testing
labeling theory, future research also should seek to clarify the role of
poverty in arrests at schools, particularly at those with an SRO.
Additional areas for future research include investigating how SROs
make the decision to arrest, typical circumstances leading to arrests, and
if there are demographic or behavioral differences in problematic
studentswho get arrested and those who donot. Finally, research shows
that school culture is related to both school violence and successful
violence prevention program implementation (D. C. Gottfredson, 2001).
It therefore is critical to evaluate the relationships between school
culture, arrests at school, and SRO activities. Researching these issues is
not possible with juvenile court records and consequently requires
additional data from schools, including observational data from police-
student encounters as well as surveys of students.
Conclusions
Concerning the role of SROs in criminalizing student behavior, this
study yielded mixed results. The findings showing that SROs were not
associated with an increase in total arrests when controlling for school
povertyand that schools with an SRO had fewer arrests for weapons and
assault charges are encouraging. Such results are contrary to the
criminalization hypothesis and may even signify that SROs have a
positive impact at schools. Nonetheless, the number of arrests involving
disorderly conduct charges at schools with an SRO is troubling. As police
and school security become more and more omnipresent at schools,
school resource officers, teachers, principals, and all school staff need to
be mindful of the negative consequences associated with punitive
disciplinary strategies and criminal arrests. For most youth, especially
those from lower socioeconomic neighborhoods, education is an invalu-
able resource to insure a brighter future. To deny them an education
286 M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
because of a minor classroom disturbance or hallway disruption is
unacceptable, unfair, and may permanently limit their prospects for a
better life.
Acknowledgements
The author wishes to gratefully acknowledge the contributions of
Judge Tim Irwin, Darrell Smith, Laurence Gibney, Heidi Garrett, Paul
Lewis, and all members of the Juvenile Court Assistance Board. The
author also thanks John G. Orme, Professor, University of Tennessee
College of Social Work, for his consultation regarding this project and
Andrea Prince, BSSW student, University of Tennessee College of
Social Work, for her assistance preparing the data.
Note
1. To fully understand the relationship between the relevant independent variables
and the dependent variables, all regression models were re-estimated with percent of
the student body that is ethnic minority replacing percent of students with economic
disadvantage as the independent variable (including the interaction term). These two
sets of models show consistent results and strikingly similar relationships between the
independent and dependent variables. A copy of this additional analysis is available by
contacting the author.
References
Astor, R. A., Meyer, H. A., & Behre, W. J. (1999). Unowned places and times: Maps and
interviews about violence in high schools. American Educational Research Journal,
36,3−42.
Bailey, K. A. (2006). Legal knowledge related to school violence and school safety. In M. R.
Randazzo, S. R. Jimerson, & M. J. Furlong (Eds.), Handbook of school violence and
school safety: From research to practice (pp. 31−49). Mahwah,NJ: Lawrence Erlbaum.
Becker, H. (1963). Outsiders: Studies in the sociology of deviance. New York: Free Press.
Beger, R. R. (2003). The “worst of both worlds”: School security and the disappearing
Fourth Amendment rights of students. Criminal Justice Review,28, 336−354.
Briers, A. N. (2003). School-based police officers: What can the UK learn from the USA?
International Journal of Police Science and Management,5,129−142.
Brown, B. (2005). Controlling crime and delinquency in the schools: An exploratory
study of student perceptions of school security measures. Journal of School Violence,
4,105−125.
Brown, B. (2006). Understanding and assessing school police officers: A conceptual and
methodological comment. Journal of Criminal Justice,34,591−604.
Bruns, E. J., Moore, E., Stephan, S. H., Pruitt, D., & Weist, M. D. (2005). The impact of
mental health services on out-of-school suspension rates. Journal of Youth and
Adolescence,34,23−30.
Cameron, M., & Sheppard, S. M. (2006). School discipline and social work practice:
Application of research and theory to intervention. Children and Schools,28,15−22.
Center on Juvenile and Criminal Justice. (20 00, April). School house type: Two years later.
San Francisco: Author.
Chen, G. (2008). Communities, students, schools, and school crime: A confirmatory
study of crime in U.S. high schools. Urban Education,43,301−318.
D'Alessio, S. J., & Stolzenberg, L. (2003). Race and the probability of arrest. Social Forces,
81, 1381−1397.
DeMaris, A. (1995). A tutorial in logistic regression. Journal of Marriage and the Family,
57, 956−968.
Dohrn, B. (2001). “Look out kid/It's something you did”: Zero tolerance for children. In
W. Ayers, B. Dohrn, & R. Ayers (Eds.), Zero tolerance: Resisting the drive for
punishment in our schools (pp. 89−113). New York: The New Press.
Dohrn, B. (2002). The school, the child, and the court. In M. K. Rosenheim, F. E. Zimring,
D. S. Tanenhaus, & B. Dohrn (Eds.), A century of juvenile justice (pp. 267−309).
Chicago: University of Chicago Press.
Eisenbraun, K. D. (2007). Violence in schools: Prevalence, prediction, and prevention.
Aggression and Violent Behavior,12, 459−469.
Elliott, D. S., Hamburg, B. A., & Williams, K. R. (1998). Violence in American schools: A new
perspective. New York: Cambridge University Press.
Farrington, D. P., & Loeber, R. (2000). Some benefits of dichotomization in psychiatric
and criminological research. Criminal Behaviour and Mental Health,10 ,100−122.
Finn, P., Shively, M., McDevitt, J., Lassiter, W., & Rich, T. (2005). Comparison of program
activities and lessons learned among 19 school resource officer (SRO) programs.
Cambridge, MA: Abt Associates.
Gottfredson, D. C. (2 001). Schools and delinque ncy. Cambridge, UK: Cambridge
University Press.
Gottfredson, D. M., & Snyder, H. N. (2005, July). The mathematics of risk classification:
Changing data into valid instruments for juvenile courts (NCJ 209158). Washington,
DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency
Prevention.
Green, M. W. (1999, September). The appropriate and effective use of security technology
in U.S. schools: A guide for schools and law enforcement agencies (NCJ 178265).
Washington, DC: U.S. Department of Justice, National Institute of Justice, Office of
Justice Programs.
Hannon, L., & Cuddy, M. M. (2006). Neighborhood ecology and drug dependence
mortality: An analysis of New York City census tracts. American Journal of Drug and
Alcohol Abuse,32, 453−463.
Harrel, F., Lee, K. L., Matchar, D. B., & Reichert, T. A. (1985). Regression models for
prognostic prediction: Advantages, problems, and suggested solutions. Cancer
Treatment Reports,69, 1071−1077.
Hirschfield, P. J. (2008). Preparing for prison? The criminalization of school discipline in
the USA. Theoretical Criminology,12,79−101.
Hopkins, N. (1994). School pupils' perceptions of the police that visit schools: Not all
police are “pigs.”Journal of Community and Applied Social Psychology,4,189−207.
Hopkins, N., Hewstone, M., & Hantzi, A. (1992). Police-school liaison and young people's
image of the police: An intervention evaluation. British Journal of Psychology,83,
203−220.
Hurst, Y. G., & Frank, J. (2000). How kids view the cops: The nature of juvenile attitudes
toward the police. Journal of Criminal Justice,28,189−202.
Hyman, I. A., & Perone, D. C. (1998). The other side of school violence: Educator policies
and practices that may contribute to student misbehavior. Journal of School
Psychology,36,7−27.
Jackson, A. (2002). Police-school resource officers' and students' perception of the
police and offending. Policing: An International Journal of Police Strategies and
Management,25,631−650.
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.
Khoury-Kassabri, M., Benbenishty, R., Astor, R. A., & Zeira, A. (2004). The contributions
of community, family, and school variables to student victimization. American
Journal of Community Psychology,34,187−204.
Knofczynski, G. T., & Mundfrom, D. (2008). Sample sizes when using multiple linear
regression for prediction. Educational and Psychological Measurement,68,431−442.
Kovac, R. (2006, February 17). Beefing up school safety. Knoxville News Sentinel,pp.B1,B7.
Laub, J. H. (2002). A century of delinquency research and delinquency theory. In M. K.
Rosenheim, F. E. Zimring, D. S. Tanenhaus, & B. Dohrn (Eds.), A century of juvenile
justice (pp. 179−205). Chicago: University of Chicago Press.
Lawrence, R. (2007). School crime and juvenile justice (2nd ed.). New York: Oxford
University Press.
May, D. C., Fessel, S. D., & Means, S. (2004). Predictors of principals' perceptions of
school resource officer effectiveness in Kentucky. American Journal of Criminal
Justice,29,75−93.
Mayer, M. J., & Leone, P. E. (1999). A structural analysis of school violence and
disruption: Implications for creating safer schools. Education and Treatment of
Children,22,333−356.
Miller, J. M., Gibson, C., Ventura, H. E., & Schreck, C. J. (2005). Reaffirming the
significance of context: The Charlotte School Safety Program. Journal of Criminal
Justice,33,477−485.
National Association of School Resource Officers. (n.d.). About NASRO. Retrieved June 29,
2007, from http://www.nasro.org/about_nasro.asp
Patterson, G. T. (2007). The role of police officers in elementary and secondary schools:
Implications for police-school social work collaboration. School Social Work Journal,
31,83−99.
Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996). A simulation
study of the number of events per variable in logistic regression analysis. Journal of
Clinical Epidemiology,49,1373−1379.
Raffaele Mendez, L. M., & Knoff, H. M. (2003). Who gets suspended from school and
why: A demographic analysis of schools and disciplinary infractions in a large
school district. Education and Treatment of Children,26,30−51.
Raffaele Mendez, L. M., Knoff, H. M., & Ferron, J. M. (2002). School demographic variables
and out-of-school suspension rates: A quantitative and qualitative analysis of a
large, ethnically diverse school district. Psychology in the Schools,39, 259−277.
Rich, T., & Finn, P. (2001). School COP: A software package for enhancing school safety.
Cambridge, MA: Abt Associates.
Rimer, S. (2004, January 4). Unruly students facing arrest not detention. New York Times,
p. 1.
Sampson, R. J., & Lauritsen, J. L. (1997). Racial and ethnic disparities in crime and
criminal justice in the United States. In M. Tonry (Ed.), Ethnicity, crime, and
immigration: Comparative and cross-national perspectives (pp. 311−374). Chicago:
University of Chicago Press.
Scheffer, M. W. (1987). Policing from the schoolhouse: Police-school liaison and resource
officer programs. Springfield, IL: Charles C. Thomas.
Schreck, C. J., Miller, J. M., & Gibson, C. L. (2003). Trouble in the school yard: A study of
the risk factors of victimization at school. Crime and Delinquency,49, 460−484.
Skiba, R. J., Michael, R. S., Nardo, A. C., & Peterson, R. L. (2002). The color of discipline:
Sources of racial and gender disproportionality in school punishment. Urban
Review,34,317−342.
Vittinghoff, E., & McCulloch, C. E. (2007). Relaxing the rule of ten events per variable in
logistic and Cox regression. American Journal of Epidemiology,165,710−718.
Welsh, W. N. (2001). Effects of student and school factors on five measures of school
disorder. Justice Quarterly,18,911−947.
Welsh, W. N. (2003). Individual and institutional predictors of school disorder. Youth
Violence and Juvenile Justice,1, 346−368.
Welsh, W. N., Stokes, R., & Greene, J. R. (2000). A macro-level model of school disorder.
Journal of Research in Crime and Delinquency,37, 243−283.
Case cited
New Jersey v. T.L.O., 469 U.S. 325 (1985).
287M.T. Theriot / Journal of Criminal Justice 37 (2009) 280–287
