Brunsma and Rockquemore wanted to investigate the extraordinary claims being made about how wonderful school uniforms are, particularly from the Long Beach California. It was being claimed that mandatory uniform policies were resulting in massive decreases (50 to 100 percent) in crime and disciplinary problems.
 
It is typically assumed, as exemplified in Long Beach, that uniforms are the sole factor causing direct change in numerous behavioral and academic outcomes. Those pronouncements by uniform proponents have raised strident objections and created a political climate in which public school uniform policies have become highly contested. The ongoing public discourse is not only entrenched in controversy but also largely fueled by conjecture and anecdotal evidence. Hence, it now seems critical that empirical analysis should be conducted to inform the school uniform debate. In this study, we investigated the relationship between uniforms and several outcomes that represent the core elements of uniform proponent's claims. Specifically, we examined how a uniform affects attendance, behavior problems, substance abuse, and academic achievement. We believe that a thorough analysis of the arguments proposed by uniform advocates will add critical insight to the ongoing debate on the effects of school uniform policies. (Brunsma and Rockquemore, 1998, pg. 54)

The authors point out that if uniforms work, they should see some of the following trends in schools with uniforms:

    1. Student uniforms decrease substance use (drugs).
    2. Student uniforms decrease behavioral problems.
    3. Student uniforms increase attendance.
    4. Student uniforms increase academic achievement.

They suspected that when other variables affecting these four items were accounted for, it would be shown that uniforms were not the cause for improvement.

How They Did Their Study

They used data from the National Educational Longitudinal Study of 1988 (NELS:88), and three follow-up studies. These studies tracked a national sample of eighth graders (in 1988) from a wide variety of public and private schools and followed their academic careers through college. Some of the data collected in the studies included uniform policies, student background (economic and minority status), peer group (attitudes towards school and drug use), school achievement, and behavioral characteristics (how often did each student get into trouble, fights , suspensions, etc.). The authors concentrated on data from the students 10th grade year.
 
Some of the independent variables they considered were sex, race, economic status, public or private school, academic or vocational "tracking", rural or urban district, peer proschool attitudes, academic preparedness, the student's own proschool attitudes, and most importantly, whether or not the students wore uniforms. The researchers wanted to determine if there was a tie between these variables and desirable behavior by the students. The areas that they were looking for improvement as a result of the previous variables included reduced absenteeism, fewer behavioral problems, reduced illegal drug use, and improved standardized test scores. The researchers considered this second group of variables to be the dependent variables. The goal of their study was to determine if there was any relationship between the independent variables (particularly uniforms) and the dependent variables.

The authors took all of the data for these variables from the NELS:88 study and performed a regression analysis to see if any of the independent variables were predictors of any of the dependent variables. If there was a strong tie in the data between any two variables ( uniforms and absenteeism, for example), it would show up in the study as a correlation coefficient close to 1 or -1. A correlation coefficient near 0 indicates no relationship between the two variables. So, if wearing uniforms had a large effect on behavior, we would expect to see a correlation coefficient of say 0.5 between uniforms and measures of good behavior. If we see a very low correlation coefficient between these two, then we know that wearing uniforms has no real effect on behavior.

Results

The only positive result for uniforms that the study showed was a very slight relationship between uniforms and standardized achievement scores. The correlation coefficient was 0.05, indicating a very slight possible relationship between the two variables, but showing that uniforms are a very poor predictor of standardized test scores and that the relationship is much weaker than has been indicated in the uniform debate. Notice that 0,05 is much closer to 0 than to 1. Other than this one weak, possible relationship, uniforms struck out. In the authors own words:

Student uniform use was not significantly correlated with any of the school commitment variables such as absenteeism, behavior, or substance use (drugs). In addition, students wearing uniforms did not appear to have any significantly different academic preparedness, proschool attitudes, or peer group structures with proschool attitudes than other students. Moreover, the negative correlations between the attitudinal variables and the various outcomes of interest are significant; hence, the predictive analysis provides more substantive results.

In other words, the authors saw no relationship between wearing uniforms and the desirable behavior (reduced absenteeism, reduced drug usage, improved behavior). They did, however, see a strong relationship between academic preparedness, proschool attitudes, and peers having proschool attitudes and the desirable behaviors. Furthermore, they saw no relationship between wearing uniforms and the variables that do predict good behavior (academic preparedness, proschool attitudes, and peers having proschool attitudes).

Conclusion

Based upon this analysis, the authors were forced to reject the ideas that uniforms improved attendance rates, decreased behavioral problems, decreased drug use, or improved academic achievement. The authors did find that proschool attitudes from students and their peers and good academic preparedness did predict the desired behavior. They saw that wearing uniforms did not lead to improvements in proschool attitudes or increased academic preparation.

David L. Brunsma, D.L. and Rockquemore, K.A. (1998)  Effects of Student Uniforms on Attendance, Behavior Problems, Substance Abuse, and Academic Achievement, The Journal of Education Research Volume 92, Number 1, Sept./Oct. 1998, pp. 53-62