University of Arkansas Office for Education Policy

School Discipline in Arkansas

In The View from the OEP on July 16, 2014 at 11:45 am

School safety is, without dispute, an important issue (in fact, the OEP named school safety concerns as the #4 story in our top 10 education stories of 2013).  According to Maslow’s hierarchy of needs and the common sense of any teacher or parent, students need a safe environment in order to learn at an optimal level. But sometimes creating a safe environment that is equitable to all students isn’t simple, or else we wouldn’t have issues like those highlighted in a report by the US Department of Education’s Office for Civil Rights earlier this year.  The report showed large racial disparities in discipline rates, which are often viewed as contributing to a “school-to-prison pipeline” that disproportionately affects students of color and limits their educational opportunities.

aacf reportSo this raises the question: what is the state of discipline and school culture in Arkansas? According to a report by Arkansas Advocates for Children and Families (AACF), Arkansas ranks
15th in the country in the use of out-of-school suspension (OSS) for all students and 13th in the disparity between the use of OSS for black and white students. The AACF report also found that black students in Arkansas were suspended about 3.5 times as often as white students.

On Friday, July 11th, OEP Director Gary Ritter presented the results of a study on school discipline rates in Arkansas to the Arkansas State Board of Education. This report was in response to Act 1329 of the Arkansas legislature, which required a report on school discipline by July 1st of each year starting in 2014. Using student-level data provided by the Arkansas Department of Education, with all personal identifiers removed in accordance with the Family Educational Rights and Privacy Act (FERPA), we were able to analyze the relationship between discipline rates, student demographics, and academic achievement.

Before moving to the results, it’s important to consider some limitations of the data.

     1) Lack of transparency of discipline measures

Discipline data are reported by school districts in systems with different codes than the state level codes, so when aggregated up to the state level, some district actions are lumped into an “other” category. Therefore, we lose transparency between the local and state level. In addition, the data did not include categories such as referrals to law enforcement authorities (a key indicator of the “school-to-prison” pipeline), but the ADE will start reporting this measure as of the 2014-15 school year.

     2) Each district has has different discipline policies and resources

The use of resources like deans or other administrators that have time to focus on discipline or the use of School Resource Officers (SROs) differ among schools and districts.

     3) It is not clear how to interpret discipline rates

For instance, both high and low discipline rates could be considered to be good or bad. A district with high discipline rates could be viewed either as positive (the school is not letting anything slide and is doing a great job handling, reporting, and tracking issues) or negative (kids are unruly, and there are a lot of behavior issues).  On the other hand, a district with low discipline rates could be viewed as positive (the school culture is positive, and there are little to no behavior issues) or as negative (the district isn’t reporting the issues that it has, or is letting too many behaviors go unpunished).

Now that we have listed the appropriate caveats to the interpretation of discipline data, let’s move on to the results. In our report, we looked at three-year average discipline rates for seven different actions: In School Suspension (ISS), Out of School Suspension (OSS), Expulsion, Corporal Punishment, referrals to an Alternative Learning Environment (ALE), No Action, and Other.  Act 1329 required a report based on a discipline rate calculated as the number of students who receive a discipline measure divided by the total number of students. The discipline rate using this method was 7.6% for ISS, 4.7% for OSS, and 5.1% for Corporal Punishment. In order to account for the fact that some students may have repeated discipline actions, we have also provided the number of incidences per 100 students. At the state level, there were about 19.5 ISS incidents per 100 students, 13.1 OSS incidents per 100 students, and 7.0 Corporal Punishment incidents per 100 students.  The rates for the other categories were relatively small.

discipline image

So what does this mean? Perhaps for those who don’t believe corporal punishment should still exist in schools (Arkansas is one of 19 states in which corporal punishment is still legal), any rate here is “bad.”  Otherwise, it’s a bit unclear whether high rates are good or bad. To add some clarity, we can at least look at disparities in the rates between various subgroups and try to answer questions related to equity.  This data showed that over the past three years, the ISS rate for non-white students (30.8 incidences per 100 students) was more than double the rate for white students (13.4 incidences per 100 students). The biggest disparity in ISS, however, that we were able to find was between students who had scored basic or below basic on a standardized test in a given year (45.3 incidences per 100 students) compared to students who hadn’t scored basic or below on their exam (13 incidences per 100 students).

ISS disparity rates

The story in OSS rates is similar, though even more striking.  Non-white students (24.8 incidences per 100 students) received OSS at a rate of over 3.6 times as high as white students (6.8 incidences per 100 students), and students who had scored basic or below basic on a standardized test in a given year (31.2 incidences per 100 students) received OSS at a rate of 4.75 times as high as students who hadn’t scored basic or below on their exam (6.6 incidences per 100 students).

At this point, it’s important to reiterate the limitations of the data and methodology used when interpreting this information.  We are only showing correlation at this point between certain demographic factors and discipline rates and are not attempting in any way to show causationIt is unclear for example, whether low-achieving students start out low-achieving and then misbehave because of it or whether students who miss school due to suspension score lower on their tests due to missing instructional time.

Despite the inability to show causation, however, we can come to an important conclusion about the importance of school culture and positive discipline policies. We found that districts with lower discipline rates have higher test scores, which is not surprising. We also found that districts with lower discipline rate disparities between students who scored basic or below basic and those who did not generally also had higher test scores. In other words, not only were the discipline rates lower in higher performing schools but the differences in disciplinary actions between groups were smaller.

The OEP will continue to work with the Arkansas Department of Education on asking and attempting to answer more questions about the relationship between school discipline and student achievement in Arkansas. In the future, the ADE will likely want to look for schools and districts who seem to be getting it right – having positive school culture, low disparity in these rates, and high levels of academic performance among all subgroups, and then identify best practices and resources that can be shared with other districts and schools. Act 1329 cited evidence-based strategies, such as restorative justice and Positive Behavioral Intervention and Supports (PBIS), as ways to improve school culture and reduce behavioral problems. In a variety of studies, PBIS in schools has been linked to lower discipline referrals, higher test scores, lower truancy rates, and improved relationships between students.

For now, school districts should know that there are plenty of resources available. The Arkansas Department of Education provides tools and resources related to PBIS on its website. In addition, the US Department of Education (USDE) has already compiled a fantastic list of resources.

  1. Characterizing of the findings of the study at issue here, as well as further study of the issue, must be cognizant of the ways measure of differences between outcome rates tend to be affected by the frequency of an outcome. The rarer an outcome the greater tends to be the relative difference in experiencing it and the smaller tends to be the relative difference in avoiding it. Further, for the rate ranges of the discipline outcomes at issue here, absolute (percentage point) differences tend to change in opposite directions of relative differences as the frequency of an outcome changes.

    I explain this most recently in the following article.

    “Race and Mortality Revisited,” Society (July/Aug. 2014)

    Click to access Race_and_Mortality_Revisited.pdf

    Particularly pertinent is the article’s Table 8, which show that relative racial differences in multiple suspensions are larger in preschool (where such outcomes are less common) than in K-12, while absolute differences are larger in K-12 than in preschool. The table also shows that the strength of the forces causing the rates of blacks and whites to differ is essentially the same in the two settings. The following web page is also especially pertinent.

    IDEA Data Center Disproportionality Guide subpage of the Discipline Disparities page of

    Presumably the finding that districts with lower discipline rate disparities between students who scored basic or below basic and those who did not generally had higher test scores is based on absolute differences between rates (as seems to be the measure mandated by the state). Relative differences between rates would tend to show the opposite pattern. See the following web page.
    Suburban Disparities subpage of the Discipline Disparities page of

    But the findings that disparities were more striking for OSS than ISS were based on relative differences. And, for example, while the relative difference between rates for low and high achiever was larger for OSS than ISS, the absolute difference was larger for ISS than OSS. The racial differences were larger in OSS than ISS in both relative and absolute terms. But the referenced patterns of correlations of measures with frequency of an outcome are merely tendencies and will not be evident in every case. The patterns must be understood, however, in order to effectively interpret the data.

Comments are closed.

%d bloggers like this: