School Discipline and Disruptive Classroom Behavior: The Moderating Effects of Student Perceptions

Abstract

This study examines the relationship between school discipline and student classroom behavior. A traditional deterrence framework predicts that more severe discipline will reduce misbehavior. In contrast, normative perspectives suggest that compliance depends upon commitment to rules and authority, including perceptions of fairness and legitimacy. Using school and individual-level data from the National Education Longitudinal Study of 1988 and multilevel regression modeling, the author finds support for the normative perspective. Students who perceive school authority as legitimate and teacher–student relations as positive are rated as less disruptive. While perceptions of fairness also predict lower disruptions, the effects are mediated by positive teacher–student relations. Contrary to the deterrence framework, more school rules and higher perceived strictness predicts more, not less, disruptive behavior. In addition, a significant interaction effect suggests that attending schools with more severe punishments may have the unintended consequence of generating defiance among certain youth.

ACKNOWLEDGMENTS

I would like to thank Richard Arum for his guidance and three anonymous reviewers for their insightful comments on earlier drafts of this article. This research was supported by dissertation grants from the Spencer Foundation and the American Educational Research Association, which receives funds for its “AERA Grants Program” from the National Center for Education Statistics and the Office of Educational Research and Improvement (U.S. Department of Education) and the National Science Foundation under NSF Grant RED-9980573. Any opinions, findings, or conclusions presented in this article are the author's and do not necessarily reflect those of the granting agencies.

NOTES

Notes

¹ For example, the American Federation of Teachers (AFT) advocated a “get-tough” approach to school discipline in the mid-'90s. The Texas branch campaigned for the passage of the Texas Safe Schools Act in 1995 which mandated removal, expulsion, and referral to the juvenile justice system for more serious offenses and gave teachers the authority to permanently remove students from classrooms. Past AFT president, Albert Shanker, wrote several articles in support of stricter discipline policies (Shanker 1995, 1997).

² Collected in 1990, the 10th-grade NELS survey data reflect the early phase of the changes in discipline and security policies that were to take place over the following decades. Since the goal of the analysis is to examine the underlying relationship between school discipline and student achievement and not trends in discipline and security policy, findings from the study is relevant to current issues in school discipline. While the National Center for Education Statistics has released a more recent longitudinal survey, the Educational Longitudinal Study (ELS; 2002), many of the relevant measures, including questions regarding punishment severity, were not included. In addition, since the first wave of data collection for the ELS is completed when respondents are in the 10th grade, the ability to control for students' behavior prior to entry into high school is not available.

³ I use the multiple imputation procedure available in SAS (PROC MI and PROC MIANAZE) (Guillion, Chen, and Meltesen 2008). I use all variables to create five imputed data sets. All five data sets are used to produce the multilevel regression parameter estimates which are subsequently averaged. Cases with missing dependent variables are used for imputation but are not included in the analysis (von Hippel 2007).

⁴ See Bryk and Raudenbush (1992) for a detailed description of this technique. This analysis uses the Proc Mixed procedure in SAS (Singer 1998).

⁵ The coefficients and p values were calculated for each of the four simple regression lines and are as follows: illegitimate (b = .058, p < .05), less legitimate (b = .037, p < .05), more legitimate (b = .016, p = .20), legitimate (b = −.005, p = .69).

⁶ Further analysis found a significant (p < .001) and negative zero order correlation (r = −.12) between eighth-grade socioeconomic status and frequency of classroom disruption. In addition, the socioeconomic status variable became significant and negative when the other eighth-grade controls, for test scores, misbehavior, and disruption was taken out of the model. This suggests that while overall higher socioeconomic status 10th graders tend to have lower levels of disruptive behavior, their disruption scores are relatively higher when 8th-grade behavior and achievement is taken into account.

⁷ Similar to eighth-grade socioeconomic status, a zero order correlation indicates a relationship in the opposite direction. There is a significant (p < .05) and positive relationship between percent minority of the school and frequency of classroom disorder (r = .02). There are several control variables in the analysis that could potentially be mediating racial composition effects, including the individual-level race, socioeconomic status, and prior misbehavior variables and school-level factors such as type of school and its urban, suburban, or rural location. These findings suggest that racial compositional effects are complex.