Temporal Differences in Crime Committed During School Hours

October 2013


School administrators and educators are responsible for the safety and security of their respective school districts and campuses. A key component to providing a satisfactory level of protection is to understand crime trends that occur within school districts and campuses. The Texas School Safety Center (TxSSC) understands the need to provide this level of information. As such, the TxSSC has undertaken a series of descriptive research articles to offer educators and school administrators introductory information relating to topics of safety and security. For instance, the TxSSC recently released an article titled Prevalence of School Crime that informed readers of crime rates experienced by schools across the nation. This article aims to build upon the knowledge gained from the prior article by addressing two separate goals.

A key component to providing a satisfactory level of protection is to understand crime trends that occur within school districts and campuses.

First, this article will explore the temporal differences in crime committed during school hours of operation. It is estimated that certain crimes are committed at different times throughout the day. The analyses will address how crimes occur throughout the day as well as parse out any differences between crimes and when they occur. Secondly, this research will perform a cursory examination into the distribution of co-offending at schools (i.e., two or more offenders acting together). This analysis will focus on both temporal trends and on the types of crime committed by multiple offenders. Namely, the analysis will address if certain crimes are committed by multiple offenders at higher rates than solo offenders. The analysis undertaken to address both goals will lead to a discussion regarding potential implications for educators and administrators.

Conceptual Framework

It is well established that school crime negatively influences students, school personnel, and the surrounding community2,3. This concept is further complicated once one considers that crimes can occur at different times and frequencies throughout the day. The basis of the current inquiry is grounded on criminological theory to provide a logical/conceptual foundation. Routine Activity Theory (RAT) provides the framework to explain how crime occurs in time and space1. Routine Activity Theory posits that for a crime to be committed there must be a motivated offender, a suitable target, and the absence of a capable guardian to converge in the same location at the same time.

It is well established that school crime negatively influences students, school personnel, and the surrounding community.

Motivated offenders are simply individuals capable and willing to commit a crime. A suitable target can be an object (e.g., wall to graffiti) or a person (e.g., simple assault). Targets are not universally suitable; rather, individual motivated offenders will find certain targets more suitable than others. Lastly, a capable guardian is any person or object (e.g., surveillance camera) that reduces the suitability of the proposed target. This does not necessarily mean the presence of a guardian equates to a capable guardian. A motivated offender may not feel the guardian reduces the suitability of the target1.

This framework is readily applied to school-based situations to form the foundation of this analysis. For instance, throughout the school day motivated offenders are within the target space (i.e., the school). Also located within the space are targets. The targets and offenders continuously intersect until the time when a capable guardian is no longer present. Conceptually the convergence of all three RAT requirements occurs throughout the day. Students walk throughout the school premises between classes. Additionally, students spend meal breaks in large communal areas without much supervision. For these reasons, the authors of this report believe crimes will vary throughout the day according to the convergence of the three pillars of RAT. Additionally, solo and multiple offenders may vary according to their motivation and the crime committed. The methodology and findings of the analysis follow.


National Incident-Based Reporting System (NIBRS) data for 2010 were utilized for this report. The FBI collects NIBRS data annually. Data are collected along multiple conditions known as segments. These segments include information relating to the offense, property, victim, offender, and arrestee. This report is concerned with temporal trends of school crime as they relate to both solo and co-offenders. Therefore, offense and offender data are utilized. Arrestee data are not reliable for this analysis. Arrestee data assume probable cause was established and an actual arrest was made. This process removes crimes that were committed but not pursued through the criminal justice system. For this reason, offender data were considered to exhibit higher validity for the analysis undertaken.

The data were reduced to only crimes committed both in Texas and on school property. The data were further reduced to crimes committed by school-aged individuals (i.e., 6-18). As this report is also concerned with co-offending, this same process was undertaken to only include school-aged co-offenders. These procedures reduced the overall criminal incidents from 4,998,914 to 4,291. In order to capture crime trends during school hours of operation, the data were further reduced to crimes committed between 7am and 4pm. This reduced the sample from 4,261 to 3,780.

No inferential analysis is to be conducted during this analysis. Therefore, randomized sampling techniques were not necessary. Rather, the use of frequencies and cross tabulations were undertaken to parse out trends exhibited throughout the school day. It should be further noted that only crimes known to law enforcement are reported. Furthermore, police departments are not required to submit data to NIBRS. This results in some geographic locations not being represented. The upcoming sections will detail the sample demographics in order to illustrate its representativeness.


As previously mentioned, the sample is made up entirely of school-aged, Texas students. Table 1 illustrates the breakdown of the sample by student demographics and Table 2 depicts the size of cities that reported NIBRS data.

As presented in Table 1, the first offender and second offender are both represented in order to show the similarities. As in the Prevalence of School Crime article, African-American students are over-represented with approximately 33% of the crimes committed even though African-American students only compose approximately 15% of students. Additionally, the Caucasian category includes Hispanic students. Hispanic is considered an ethnicity; therefore, Hispanic was included with its primary racial affiliation. The only change exhibited between the first offender and co-offenders is the ratio of males and females. The co-offenders presented have a higher ratio of female offenders.

As previously mentioned, departments are not required to submit information to NIBRS. To illustrate the diversity of the sample, a descriptive analysis was performed on city size. Therefore, Table 2 shows the breakdown of the number of crimes reported by various city populations. Even though not all cities report NIBRS data, a wide variety of cities do. A majority of offenses presented stem from cities with 50,000 to 99,999 residents. Next, findings are presented.


The initial findings presented are intended to descriptively depict the offenses. It is important to understand the components of the crimes before moving on to a more thorough analysis of the crime trends. The frequencies of committed crimes are presented below in Table 3. The two most prevalent crimes committed in the sample are simple assault (e.g., pushing; 43.3%) and drug offenses (23.5%).

Time of Day Information

Figure 1 illustrates the frequency of crimes by time of day. Since this report is primarily concerned with crimes committed during operational hours, only 7am to 4pm hours are shown. This reduced the overall sample to 3,780 offenses. The temporal distribution of offenses is normally distributed. This is readily apparent due to the gradual increase in offenses until the 12pm hour. After the 12pm hour there is a decline in crime until it holds constant through the rest of school hours until the 4pm hour. At this time most school have dismissed for the day. There is a rapid decline in crime at this point.

The numeric values depicted are percentages of crimes committed for each hour. A percentage was utilized to standardize the raw value. Additionally, percentages of crimes committed will be used in later analyses. The statistical software calculated a line of best fit for the histogram. This line depicts a normal curve to illustrate the normal distribution of crimes performed per hour.


A variable was created to capture offenses committed by more than one offender. This allowed for the analysis to discover if co-offending exhibited different trends than solo offenders. As presented in Figure 2, over three-quarters of offenses during school operational hours were committed by solo offenders (78.5%).

Cross tabulations

The primary analysis consists of cross tabulations. A cross tabulation is a tool that allows readers to easily recognize trends present in the data. A cross tabulation allows the reader to understand what percentage of one categorical variable (e.g., crime committed) can be attributed to another categorical variable (e.g., time of day, co-offender). Furthermore, a chi-square test of significance was performed for each cross tabulation. Simply put, a chi-square test of significance merely informs the reader if any observed difference in variables is by random error or not. A non-significant finding would inform the reader that any observed difference in variables is simply random error (i.e., there isn’t actually a difference between the variable in question). A significant finding would allow the reader to state that the observed difference in the variable is in fact a difference. For example, Table 6 examines the presence of co-offenders cross tabulated with the school’s operational hours. The results were statistically significant; meaning the observed differences between solo offenders and co-offenders is not by random chance. There is, statistically, a difference between the two.

Co-offending by Time of Day

Table 6 displays the frequency and percentages of crimes committed by co-offenders and solo offenders throughout the schools’ operational hours. These findings show a similar pattern. However, they are slightly different. Solo offenders show little change throughout the day. At most, there is a two percentage point difference between 8am and 11am - 12pm. However, co-offending nearly doubles in this same time period—9.3% at 8am to 16% at 12pm. As stated in the above example, this difference is statistically significant.

Crimes by Time of Day

As shown in Table 7, the various crimes reported are committed at different rates throughout the day. Starting from the left side of the table, one can track each crime throughout the operational hours of schools. The findings are presented in standardized percentages for ease of reading. Furthermore, recall Figure 1. The histogram illustrated a normal distribution of crimes throughout the day when aggregated. Thirteen percent of crimes were committed during the noon hour with slightly less before (2-11%) and after (11-6%). However, this is not always the case once crimes are disaggregated for further analysis. In fact, much variance between crimes is exhibited.

Drug offenses and assaults, as a whole, are normally distributed. However, aggravated assaults spike at the beginning of the school day, during the lunch period, and at the conclusion of the school day. Much of the remaining crimes are bi-modal (i.e., there is a spike in the morning, a lull during the middle of the day, and another spike in the afternoon). For instance, destruction of property has a clear spike during the 9am hour. However, only 6.7% of these crimes are committed during the 12pm hour. Directly after the 12pm hours, the destruction of property crimes begins to rise again. Explore Table 7 for detailed results relating to the remaining offenses. These findings were also found to be statistically significant. Therefore, the observed differences are not by chance.

Crimes Committed by Solo and Co-offenders

The final analysis consists of cross tabulating crimes on the presence of single or multiple offenders. Recall in Figure 2 that co-offenders are present in 21.5% of the sample offenses. Therefore, any value above relates to an overrepresentation of multiple offenders, and any value below relates to an underrepresentation of multiple offenders.

The only two offenses in which co-offenders are simple assault (26.2%) and both classifications of theft (24.2% and 23.6%). Based on the sample characteristics, one would expect solo offenders to compose 78.5% of offenses. However, Table 8 clearly shows most offenses in this sample are committed by solo offenders. In fact, only Destruction of Property is closely related to the sample distribution. These findings were also found to be statistically significant. This indicates that the observed difference between single and multiple offenders actually exists and is not by chance.

Summary of Findings

The goal of this analysis was two-fold; (1) explore temporal differences in crime committed during school hours of operation, and (2) to perform a cursory examination into the distribution of co-offending at schools both temporally and in regards to types of crimes committed. The analyses showed a normal distribution of crime throughout the sample schools’ operational hours. That is, percentages of crimes committed slowly increased until midday when the percentages of crimes committed slowly decreased (see Figure 1). The rate of solo and co-offenders follows the same pattern. However, while solo offenders exhibit little variance throughout the day, co-offending showed a dramatic spike during lunch hours (see Table 6). Furthermore, once disaggregated, individual crime categories did not exhibit the same distribution as the overall rate. Several crimes showed multiple spikes throughout the day, while simple assault and drug offenses followed a pattern similar to the overall distribution (see Table 7). The final important findings were presented in Table 8. Certain crimes exhibit higher rates of solo offenders than expected based on the overall sample rate. Further, some crimes (e.g., theft) indicated higher rates of co-offending than expected. Implications for educators and administrators will follow.

Co-offending showed a dramatic spike during lunch hours


These important findings illustrate key, potential implications for educators and administrators. For instance, these data suggest an increased presence of staff during communal periods may be warranted. Recall that assaults and drug offenses accounted for 66.8% of the crimes committed on school property. Assaults and drug offenses also peak during 11am to 12pm block. This time period is generally reserved for lunch; therefore it is plausible the increase in assaults and drug use can be attributed to the foundations of routine activities (i.e., motivated offender, opportunity, and lack of guardianship).

Drug offenses accounted for 66.8% of the crimes committed on school property

Administrators may want to also consider the period of time between classes. Various offenses occur at different rates throughout the school day. The variance exhibited suggests administrators should conduct similar trend analyses for their specific school districts and/or campuses. This procedure would allow administrators to make data/evidence backed staffing decisions in regards to student oversight. Furthermore, crime specific measures can be targeted if trends are exhibited at a particular school/district (e.g., drug prevention measures to combat peak offense times).

Crime specific measures can be targeted if trends are exhibited at a particular school/district

The findings regarding co-offenders also present potential policy implications. Recall that co-offending experiences a dramatic spike in frequency during the 12pm hour. Furthermore, certain offenses (i.e., simple assault and theft) are perpetuated by more than one offender at a higher rate than expected based on the sample. An increased adult presence and the removal of opportunity may successfully combat these co-offending trends in keeping with routine activities theory doctrine.

Administrators and educators should be made aware of the prevalence of crime as well as the temporal trends exhibited on school premises. These analyses were undertaken at a state-wide level to help administrators and educators make evidence-based decisions regarding the safety and security of their students. However informative these findings are, school districts should be proactive by performing additional trend analyses on their specific districts/campuses. This process begins by collecting accurate crime data, conducting thorough analysis, and implementing policies to offset any discovered crime trends.


1Cohen, L. & Felson, M. (1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44(4), pp. 588–608

2Crews, K., & Crews, J., Turner, F. (2008). School violence is not going away so proactive stepsare needed. College Teaching Methods and Styles Journal, 4(1), 25-28.

3Hart, C., & Mueller, C. (2013). School delinquency and social bond factors: Exploring gendered differences among a national sample of 10th graders. Psychology in the Schools, 50(2), 116-133.