ABSTRACT

Recent discourse on public school reform has focused on mandatory uniform policies. Proponents of such reform measures emphasize the benefits of student uniforms on specific behavioral and academic outcomes. This research empirically tests the claims made by uniform advocates using 10th grade data from The National Educational Longitudinal Study of 1988. Our findings indicate that student uniforms have no direct effect on substance use, behavioral problems or attendance. A negative effect of uniforms on student academic achievement was found. These findings are contrary to current discourse on student uniforms. We conclude that uniform policies may indirectly affect school environment and student outcomes by providing a visible and public symbol of commitment to school improvement and reform.

INTRODUCTION

Public discourse surrounding educational reform has recently focused on the importance of uniform policies in public schools. School uniform policies have historically been restricted to the private sector and have only recently begun to be discussed as a viable policy option in public school districts. A decade of research showing the effectiveness of private schools has led some school reformers to consider various policies which are linked to private and Catholic school success. Within the Catholic school literature, school uniforms have never been asserted as a primary factor in producing the Catholic school effect. Nevertheless, public school administrators are beginning to consider uniform policies as a way to improve the overall school environment and student achievement. Due to the controversial nature of mandatory school uniform policies, educators are speaking out, both advocating and condemning the proposed reform efforts.

Uniform advocates propose several different arguments. First, uniforms are argued to positively effect student safety by: lowering student victimization (Scherer 1991), decreasing gang activity and fights (Kennedy, 1995; Loesch, 1995), and differentiating strangers from students in the school building (Department of Justice, 1996; Gursky, 1996). Second, uniforms are asserted to increase student learning and attitudes towards school through: enhancing the learning environment (Stover, 1990), raising school pride (Jarchow, 1992), increasing student achievement (Thomas, 1994), raising levels of preparedness (Thomas, 1994), and promoting conformity to organizational goals (LaPointe, Holloman, and Alleyne, 1992; Workman & Johnson, 1994). Additionally, uniforms are attributed to decreasing behavior problems by: increasing attendance rates, lowering suspension rates, and decreasing substance use among the student body (Gursky, 1996). Finally, various psychological outcomes are attributed to wearing uniforms including: increased self-esteem (Thomas, 1994), increased spirit (Jarchow, 1992), and increased feelings of "oneness" among students (LaPointe, Holoman, & Alleyne, 1992).

Opponents of adopting uniform policies stress the legal, financial, and questionable effectiveness of such policies. The legal concerns focus on the supposition that requiring a uniform violates children's individual rights (Thomas, 1994; Virginia State Department of Education, 1992). This argument is extended by opponents who argue that mandatory uniform policies are being considered largely for urban school districts, and hence are being forced upon a predominately minority and poor student population (Thomas, 1994). Financially, groups such as the American Civil Liberties Union have voiced concerns about the cost of uniforms, specifically that purchasing one is a mandatory cost which some disadvantaged parents are unable to afford (Gursky, 1996). Finally, the strongest opponents to uniform policies charge that there currently exists no empirical evidence to support the numerous and varied claims of uniform proponents (LaPointe, Holoman, & Alleyne, 1992).

The case study most often cited in the political rhetoric surrounding the uniform debate is that of the Long Beach Unified School District (LBUSD). LBUSD was one of the first large urban school districts within the United States to adopt a mandatory school uniform policy. This case provides some context for the discussion in that it serves as a prime example of a system which has recently instituted a school uniform requirement, has received national attention for it's efforts, and attributes students' behavioral changes to the mandatory uniform policy. In a press release, the Board President of LBUSD had the following to say about the uniform policy:

These schools are becoming educational workplaces. Students arrive dressed for success, ready to learn. They're getting along with one another better and experiencing significant gains. Principals and teachers tell us that students' success is taking many forms -- fewer absences, fewer tardies, fewer truancies, fewer referrals to the office for behavior problems, fewer suspensions and expulsions, better grades and, in some cases, significantly higher achievement. (Polacheck, 1996)

In this district, school uniforms are currently required from kindergarten through eighth grade in 70 schools, including approximately 60,000 students. School District press releases indicate that there is widespread parental support for the mandatory uniform policy. Although California law provides a clause allowing parents to request a uniform exemption for their students, less than 1% of parents have requested such exemptions. In efforts to aid students from financially disadvantaged families, philanthropic groups in the area have provided $160,000 in uniforms to Long Beach students. The table (Table A) in Appendix A presents the statistical evidence provided by the School District in support of their claims that school uniforms decrease crime.

It is typically assumed, as exemplified by the Long Beach case, that uniforms are the sole factor causing direct change in numerous behavioral and academic outcomes. It is these pronouncements by uniform proponents that have raised strident objections and created a political climate in which public school uniform policies have become highly contested. This ongoing public discourse is not only entrenched in controversy, but largely fueled by conjecture and anecdotal evidence. Hence, it seems critical at this point in time, for empirical analyses to be conducted to inform the school uniform debate. This paper examines the relationship between uniforms and several outcomes which represent the core elements of uniform proponents' claims. Specifically, we will examine the effect of wearing a uniform on attendance, disciplinary behavior problems, substance abuse, and academic achievement. It is the intention of the authors 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.

THEORETICAL FRAMEWORK

Nathan Joseph (1986) has formulated an analysis of clothing as communication which provides a framework within which uniform proponents claims can be better understood. He asserts that clothing can be considered a sign, which he defines as "anything that stands for something else." Clothing, as a sign, coveys information about values, beliefs and emotions. If the clothing that adolescents wear can be considered a sign, then that which they freely choose as individuals can be seen as conveying an expression of their personal identity. School uniforms, by contrast, are clothing which is selected by school officials and mandated to students. It is simple in style and color and it is intended to convey the institutional values of the school.

Joseph suggests that for clothing to be considered a 'uniform' it must fulfill the following criteria: 1) it must serve as a group emblem, 2) it must certify the institution's legitimacy by revealing an actor's status position, and 3) it must suppress individuality (1986). Within the context of an educational institution, school uniforms clearly function as a symbol of membership to the school community. The presence of a uniform in schools automatically implies a two-tiered hierarchical structure, those that wear uniforms (subordinates) and those that do not wear uniforms (superiors). School uniforms serve as a clear sign of this status distinction between students and faculty and therefore, certify the legitimacy of that distinction by all members. School uniforms act as suppressers of students' individuality by mandating standardization of appearance and removing student expression through clothing.

Given these characteristics of uniforms, it becomes clear that mandatory uniforms serve the function of maintaining social control within the school environment. The uniforms, as a sign of group membership, act as immediate cues which signal who does and does not belong to the school community. Amongst the community members themselves, uniforms seem to act as a dramaturgical device by establishing interactional boundaries between members of separate statuses (teachers and students) and promoting the internalization of organizational goals.

If uniforms are considered a sign which facilitates social control of student behavior, then it can be expected that students in uniforms will display behaviors which are consistent with the institutional goals of the school. Inconsistent attendance, disciplinary behavior problems, and substance abuse represent student behaviors which are non-representative of the values of public high schools. By contrast, high levels of academic achievement are consistent with the goals of educational institutions. The following hypotheses are provided to test the validity of the uniform advocates' statements.

H1: Student uniforms will decrease substance use

H2: Student uniforms will decrease behavioral problems

H3: Student uniforms will increase attendance

H4: Student uniforms will increase academic achievement

Within the context of the public debate on mandatory uniform policies, the mechanisms through which uniforms effect the above stated outcomes are subtly implied. They include pro-school attitudes, peer pro-school orientation, and academic preparedness. In testing each of the above stated relationships, it is expected that the direct effect of uniforms on the four outcomes will disappear when these moderating variables are added to the equation. If this is in fact the case, arguments stating uniform policies' direct effect on a given outcome should be abandoned and more attention given to the actual mechanisms which produce the sought after effects. Finally, it should be emphasized that the purpose of this paper is to test the claims made in the context of the school uniform debate using a nationally representative sample of students.

DATA AND METHODS

The National Educational Longitudinal Study of 1988 (NELS:88) is used to test the relationships outlined above. NELS:88 is a national stratified random sample of schools and students which began with eighth grade students (in 1988). It has since gone through three follow-ups with the most current one (1994) collecting data on the original eighth graders in their second year of postsecondary education. The data used for this analysis comes from the first follow-up of NELS:88 when students were in tenth grade. NELS:88 oversampled certain minority groups, private sector schools, and high performance schools. Thus, standardized weights and design effects will be applied in order to make statements about the population of tenth grade students in the United States and the effects of uniforms on them. The student component as well as the school-administrator component were used to provide data on uniform policies and the student background, peer group, achievement and behavioral characteristics needed for this analysis.

VARIABLES

NELS:88 provided a number of variables which were used to analyze the relationship between student uniforms and various student outcomes. Table 1 presents the means and standard deviations of the variables described below (see Appendix B for the specific NELS:88 variables used to construct the measures used).

Independent Variables

Several controls for student characteristics were constructed. Student minority status was measured by a dummy variable for Blacks, Asians, and Hispanics. White students remained the omitted category and all comparisons are made to them. Student gender was assessed by a dummy variable with male students as the omitted category. These categorizations resulted in a weighted distribution of 49.6% female, 50.4% male, 3.8% Asian, 12.5% black, 10.1% Hispanic, and 73.6% white students. As an indicator of student socioeconomic status a pre-computated NELS:88 socioeconomic status composite was used to control for parental education, income, and occupational prestige. The distributional properties of this scale are a mean of zero with a standard deviation of 1; thus, we obtain a scale which represents individual student family deviations from the mean of this composite.

Variables to control for school characteristics were also used. Important controls for student track placement were measured using a dummy variable for each academic, vocational-technical, and other programs with general tracked students as the omitted category. Another crucial school variable to control for in this analysis is that of school sector. Because a small percentage of public schools actually have uniform policies, the results found here could not be relegated to simply a private/public explanation. Thus, control variables were constructed for Catholic schools, private religious schools, private non-religious schools, and private schools which were not ascertained as to their affiliation. The omitted category was, of course, public high schools. This categorization revealed a weighted description of the sample as follows: 51.9% in general track, 32.3% in academic track, 9.8% in vocational/technical track, 6% in other programs, 92% Public schools, 5.6% Catholic schools, 2% Private religious schools, 1.2% Private non-religious, and a small (less than 1%) percentage of private non-ascertained schools. Type of school district (urban, rural, suburban) were also controlled for. Some 32% of schools were rural, 27.6% were urban and 40.4% were suburban.

A variable from the School Component of NELS:88 was used to ascertain whether or not a student was, due to school policy, required to wear a uniform. Some 5% of the students in the entire sample were required by policy to wear a school uniform. This can be further broken down into: 65.4% of Catholic, 16.6% of Private Non-Religious, 5.4% of Private Other-Religious, .8% of Public, and 0% of Private Non-Ascertained students are required to wear a uniform at their high school. Student uniform use is the focal independent variable of this research project.

Three scales were created to represent school preparedness, student pro-school attitudes, and the peer group's pro-school attitudes. These scales represent variables which are hypothesized to be the critical moderating variables explaining why uniforms might affect the dependent variables as opposed to uniform use having direct effects. The scales were used to measure these intervening processes (see Appendix B for scale information). An academic preparedness scale assessed the degree to which a student came to class with their books, their homework done, and school supplies (i.e., pencils, paper, etc.). A scale to assess pro-school attitudes measured the degree to which a student felt it was ok to cut class, destroy school property, fight on school grounds, etc. It taps an important dimension of attitudes towards behavior at school which would or would not cater to a positive academic atmosphere. Finally, a scale was constructed in order to assess the importance the student's peer group placed on pro-school attitudes. It included attitudes towards finishing school, getting good grades, and studying. As the analysis proceeded, interaction terms between the uniform variable and those of the moderating scales were computed to assess the special effects of these combinations on the outcomes of interest.

Dependent Variables

The debate over school uniforms suggests using several outcomes to test the effectiveness of adopting a uniform policy on how students fare on these consequences. The dependent variables chosen were student absenteeism, student behavior problems, student substance use, and student achievement. A variable was used to assess how often a student was absent from school. A behavioral scale was created from a number of variables to assess the degree to which a student has been involved in behaviorally problematic conduct in relation to school. Some of the variables in the behavior scale included whether the student: got into physical fights, got put on in-school suspension, skipped or cut classes, was suspended from school, and in general got in trouble. To assess the degree of substance usage among students, a scale was computed to reflect student use of cigarettes, alcohol, and marijuana. To assess student achievement a composite standardized achievement test (composite of reading and math tests) was used.

Regressions were also conducted on pro-school attitudes, peer pro-school attitudes and academic preparedness to observe the effect of uniforms on these characteristics. By testing the logic of claims made by advocates of school reform, several interesting findings result that have implications for the ongoing debate. The authors remain specifically interested in the relationships and the predictive power of student uniform policies on the outcomes of interest.

In all analyses a weighting procedure utilizing the population weight (F1QWT) and the appropriate design effect was computated according to the population being tested. The standardized population weight multiplied by the inverse of the appropriate design effect is used in the analyses in order to take into account the fact that NELS:88 is a clustered data set. By creating these new weights (a weight for the entire sample, for the catholic sample and for the private sample - see analysis in Table 3), we correct for clustered sample problems which stem from each case not necessarily being independent due to the sampling frame used by the collectors of NELS:88. Weighted regressions using design effects will be performed where the relationships highlighted above will be tested. These corrections provide results which are representative of U.S. tenth graders.

Variable	Mean	Standard Deviation
Female	.496	.500
Asian	.038	.191
Black	.125	.330
Hispanic	.101	.301
F1ses	3.267	17.861
Catholic	.056	.229
Privnon	.012	.110
Privnot	.001	.022
Privrel	.020	.141
Academic	.323	.468
Votech	.098	.297
Othprog	.060	.237
Rural	.320	.467
Urban	.276	.447
Uniform1	.050	.219
Absent	3.012	1.379
Behavior	1.781	2.359
Drugs	2.702	3.092
Peerpro	7.475	1.386
Prepare	9.707	1.756
Proschol	43.835	4.124

^aweighted by newghta which was computed using the NELS:88 population weight (f1qwt) and the design effect for the

entire sample, the computation is as follows: newghta=f1qwt/(mean of f1qwt) x 1/(design effect).

RESULTS AND FINDINGS

Descriptive Analysis

Table 1 gives the weighted means and standard deviations of the variables which are used in the analyses presented here. Most of the descriptive highlights have been summarized in the above section on variable construction. Appendix B gives a summary of the original NELS:88 variables which were used to create the various independent and dependent scales.

Appendix C shows the correlation matrix of the variables used in the analysis. Student uniforms are slightly (.05) correlated with standardized achievement scores indicating a possible relationship there. The debate focuses on correlations like these; however, this correlation, though not indicating a predictive nature of uniforms, is much smaller than the debate would have us believe. Student uniform use is not significantly correlated with any of the school commitment variables of absenteeism, behavior, or substance use (drugs). In addition, students who wear uniforms do not appear to have any significantly higher or lower academic preparedness, pro-school attitudes, or peer group structures with pro-school attitudes. Also, note the significant negative correlations between these attitudinal variables and the various outcomes of interest. These are correlations, hence the predictive analysis will provide more substantive results.

In order to give a slightly more rigorous test of the relationships between uniforms and the four dependent variables, t-tests comparing the means of the dependent measures by uniform use and sector were conducted. Table 2 presents the results for the weighted sector comparisons. In the first panel, means of our four dependent measures are compared between uniform and non-uniformed 10th graders for the total sample which included all school sectors. At this level, uniformed students have significantly higher achievement (p < .01). This mirrors the hypothesized character of the difference as stated in the public discourse. However, when one breaks down this type of analysis into sector, the relationships do not hold.

	Uniformd	Non-Uniform
Total Samplea (N=4578) Absent Behavior Drugs F12xcomp	2.90 1.58 2.68 52.89**	3.01 1.74 2.71 50.58
Catholic Sampleb (N=327) Absent Behavior Drugs F12xcomp	2.89* 1.49 2.73 53.51**	2.55 1.41 2.80 56.53
Private Samplec (N=80) Absent Behavior Drugs F12xcomp	2.93 1.33 2.36 56.60	2.73 1.28 2.07 56.01

*** p < .001, ** p < .01, * p < .05

^a weighted by newghta which was computed using the NELS:88 population weight (f1qwt) and the design effect for the entire sample, the computation is as follows: newghta=f1qwt/(mean of f1qwt) x 1/(design effect).

^b weighted by newghtc. See a above except this is for the catholic sample only.

^c weighted by newghtp. See a above except this is for the private school sample only.

^d t-tests of significance were conducted comparing uniform to non-uniform within each sector.

Catholic schools and uniforms go together in most people's minds, and in fact, they are the sector which utilizes uniform policies the most (65.4%). Thus, one would expect the relationship to hold here as well - uniformed Catholic students should have the desired outcomes to a greater extent then non-uniformed Catholic students (if uniforms are indeed a force behind what occurs there). The same logic should apply to other private schools, albeit to a less exact extent. In fact, panel 2 of Table 2 shows the results of a weighted comparison between uniformed students and non-uniformed students in the Catholic sector only (N=327). Only the results for absenteeism and achievement are significant and it is important to note that these relationships reverse. Uniformed Catholic students are absent more often (p < .05) and, on average, score some 3 points less (p < .01) on an achievement test than non-uniformed Catholic students. This fails to support the thesis that uniforms are related to these outcomes. None of the comparisons are significant in panel 3 where other private schools are compared most likely due to small sample size.

Student Uniforms as Predictors

So far this paper has presented somewhat weaker, though interesting, tests of the relationship between student uniforms and the various outcomes. The debate tends to imply stronger claims than simple correlations and mean comparisons: there is an implicit charge that uniforms "cause" or "impact" the outcomes with which educators and policymakers are concerned. A number of weighted regression analyses were run in order to test the predictive impact of student uniforms on absenteeism, behavior problems, substance use, and achievement. Table 3 presents the results for the regressions using the three indicators of non-commitment to school. Table 4 presents the results for the regression of achievement on uniforms and other variables.

Do uniforms have an impact on absenteeism? Model I presents the unstandardized coefficients for the impact of the control variables on absenteeism. These explain 3% of the variance in the dependent variable. In Model II, the variable for student uniforms is added. The uniform coefficient is not significantly different from zero and no statements can be made. No extra variance is explained in Model II. An interesting finding is that once the variation for uniform use is taken into account the Catholic effect actually gets stronger in decreasing absenteeism. This implies that the Catholic effect, often cited in the literature as effecting these sorts of outcomes, remains supported. However, the effect is not associated with whether the students wear uniforms or not; it is more likely due to the social relations fostered in Catholic schools. Finally, the variables added in Model III explain an extra 8% of the variance in absenteeism and are all significant predictors of decreased absenteeism indicating that academic preparedness, pro-school attitudes and peer norms significantly affect attendance at school in the desired direction. Hypothesis One, which stated that student uniforms would decrease absenteeism, is not supported by these results.

*** p < .001, ** p < .01, * p < .05

^a Coefficient was .0045 in all cases where indicated by a .00.

Do student uniforms significantly decrease behavioral problems? Again, Model I shows the results for the control variables alone on the dependent variable. These variables alone explain 8% of the variance in behavior problems. The student uniform variable is added in Model II and the insignificant effect is similar to that for absenteeism. No extra variance is explained. When the mediating variables are added in Model III, an extra 33% of the variance in behavioral problems is explained. Academic preparedness, pro-school attitudes and peer norms effectively lessen behavioral problems on average. Hypothesis Two, which stated that student uniforms will decrease behavior problems, is not supported by this analysis.

A final question of uniform's relationship to school commitment can be posed: Do student uniforms significantly decrease substance use among high school students? As in the previous results, Model I presents the control variables' effects. These explain 5% of the variance in substance use. Model II adds the student uniforms variable. The uniform variable is non-significant and it adds no extra explanatory power. Finally, academic preparedness, pro-school attitudes, and peer norms which are pro-school again effectively decrease substance use among high school students. These variables explain an extra 27% of the variance. Thus, Hypothesis Three, which stated that student uniforms will decrease substance use, is unsupported, implying that implementing uniform policies at the high school level will not effectively create the desired outcomes.

Standardized Achievement

	I	II	III
Female	.33	.40	.00
Asian	.16	.10	-.09
Black	-6.29***	-6.27***	-6.53***
Hispanic	-4.41***	-4.43***	-4.53***
F1ses	-.01	-.01	-.01
Catholic	1.40*	3.04***	2.99***
Privnon	3.30**	3.67**	3.55**
Privnot	-.08	-.14	-.68
Privrel	3.21**	3.52***	3.43***
Academic	6.60***	6.60***	6.29***
Votech	-3.95***	-3.94***	-3.97***
Othprog	-2.78***	-2.79***	-2.82***
Rural	-1.37***	-1.40***	-1.47***
Urban	.15	.17	.09
Uniform1		-2.59**	-2.50**
Prepare			.18*
Proschol			.22***
Peerpro			-.10
(Constant)	50.7***	50.6***	39.9***
R2	.23	.23	.24
Std. Err. Est	8.57	8.56	8.50
F-Value	69.1***	65.2***	57.7***
N	3286	3286	3286

*** p < .001, ** p < .01, * p < .05

Table 4 presents a similar set of models as in the previous three analyses for uniform's effect on achievement scores. Do student uniforms effect achievement? Model I presents the results for the control variables explaining some 23% of the variance in the standardized achievement test. Model II adds the dummy variable for student uniforms. Here, contrary to the expected, student uniform use actually decreases, on average, the standardized test score of these tenth graders who wear them due to mandatory school policy. It is, in fact, almost a 3-point decrease. Though it explains no more of the variance than did Model I, the coefficient for uniforms is statistically significant (p < .01) and negative. Model III adds the attitudinal variables and an extra 1% of the variance in achievement is explained with preparedness and pro-school attitudes significantly increasing achievement. Finally, Hypothesis Four, stating that student uniforms will increase student achievement, is not supported by these data. In fact, all four of the original hypotheses, derived from the public discourse surrounding the uniform debate, are not supported. Most striking is uniform's significant negative effects on achievement - an outcome of much concern to educators and policy makers.

Prepare

ProSchol

PeerPro

	I	II	I	II	I	II
Female	.59***	.59***	1.47***	1.47***	.43***	.43***
Asian	.03	.03	1.01**	1.01**	.42***	.42***
Black	-.08	-.08	1.45***	1.45***	.37***	.37***
Hispanic	-.14	-.14	.72**	.72**	.15	.15
F1ses	-.00	-.00	-.00	-.00	.00	.00
Catholic	.13	.29	-.09	.02	.01	.01
Privnon	.26	.30	.35	.38	.13	.13
Privnot	.70	.70	2.52	2.52	1.17	1.17
Privrel	.09	.12	.36	.38	.08	.08
Academic	.32***	.32***	1.29***	1.29***	.38***	.38***
Votech	.02	.02	.12	.12	-.03	-.03
Othprog	.04	.04	.23	.23	.12	.12
Rural	.14*	.13*	.22	.22	-.01	-.01
Urban	.05	.05	.25	.25	.02	.02
Uniform1		-.25		-.18		-.01
(Constant)	9.28***	9.28***	42.2***	42.2***	7.06***	7.06***
R2	.04	.04	.07	.07	.05	.05
Std. Err. Est	1.70	1.70	4.01	4.01	1.34	1.34
F-Value	11.5***	10.9***	19.4***	18.1***	14.4***	13.5***
N	3776	3776	3639	3639	3594	3594

*** p < .001, ** p < .01, * p < .05

 Uniforms and Pro-School Attitudes: Is There a Relationship?

Though the hypotheses were not borne out, the authors decided to examine whether uniforms directly impact the development of academic preparedness, pro-school attitudes, or peer structures with pro-school norms. Since these variables consistently produce the desired outcomes it is important to assess uniform's effects on these as well. Looking at Table 5, in fact, uniforms do not have any effect on the moderating variables in the analysis. Though academic preparedness, pro-school attitudes, and peer norms significantly effect the outcomes studied, uniforms have no effect on the moderating attitudinal variables either.

	Absent	Behavior	Drugs	Std. Test
uniform1*prepare	.08	.11	-.07	-.29
uniform1*proschol	-.00	.07*	.02	-.23
uniform1*peerpro	.03	.04	.01	-.05
uniform1*urban	-.30	-.17	-.47	.76
uniform1*f1ses	.00	-.00	.00	-.00
uniform1*catholic	.18	-.15	-.22	1.42

*** p < .001, ** p < .01, * p < .05

Interactions of student uniforms and each of the following variables of interest were computed and entered into a full model (like Model III's in Tables 3 and 4): academic preparedness, pro-school attitudes, peer pro-school attitudes, urbanicity, socioeconomic status, and catholic sector. The following questions apply, respectively to the tests of interactions: Do uniformed kids with high academic preparedness significantly differ in the desired direction from their counterparts on the dependent measures? Do uniformed kids with high pro-school attitudes significantly differ in the desired direction from their counterparts on the dependent measures? Do uniformed kids with strong pro-school peer groups significantly differ in the desired direction from their counterparts on the dependent measures? Do uniformed kids in urban areas significantly differ in the desired direction from their counterparts on the dependent measures? Do uniformed kids with high socio-economic status significantly differ in the desired direction from their counterparts on the dependent measures? Finally, do uniformed Catholic kids significantly differ in the desired direction from their counterparts on the dependent measures?

Looking at Table 6 only one significant coefficient is found: uniformed kids with high pro-school attitudes actually have worse behavior problems than their counterparts. This is contrary to the expected. Uniforms seemingly have no impact in tandem with those things which are proven effective.

 IMPLICATIONS AND DISCUSSION

The Discourse/Rhetoric Re-examined

Our failure to find a direct effect of uniforms on behavioral outcomes or academic achievement provide cause for a closer examination of the uniform debate. It seems that reformers have seriously considered the educational research showing outcome differentials between public and Catholic school students. However, it is equally apparent that the most superficial policies are those that have been extracted for possible reform efforts. A closer reading of the public versus private school literature would suggest that uniforms are merely symbolic of the communal organization of Catholic schools which, researchers have proposed (Bryk, Lee, & Holland, 1993; Bryk & Driscoll, 1988; Coleman & Hoffer, 1987), is the fundamental cause of the Catholic school advantage.

A reconsideration of the Long Beach case sheds light on the flawed logic of uniform proponents' assertions. The descriptive information provided by LBUSD (Appendix A) suggested that school crime was significantly reduced between the 1994-1995 and 1995-1996 school years. Between these periods a mandatory uniform policy was established district wide. Seemingly, the correlation between these two events is reason enough for Long Beach administrators to state that a causal relationship exists. While in fact, these two events may be empirically verifiable, the argument that uniforms have caused the decrease in school crime is simply not substantiated. Taking into consideration both the findings provided in this paper and the additional materials from the Long Beach public school system, we would propose an alternative interpretation.

What is omitted from the discourse on school uniforms is the possibility that, instead of directly impacting specific outcomes, uniforms work as a catalyst for change and provide a highly visible window of opportunity. It is this window which allows additional programs to be implemented. An examination of the Long Beach case shows that several additional reform efforts were simultaneously implemented with the mandatory uniform policy. These programs include a reassessment of content standards, a $1 million grant from the Edna McConnell Clark Foundation to develop alternative pedagogical strategies, and the Focused Reporting Project (Kahl, 1996). It seems curious that given these substantive reform efforts, administrators continue to insist that uniforms are the sole factor causing a variety of positive educational outcome.

Requiring students to wear uniforms is a change which not only effects students, but school faculty and parents. Instituting a mandatory uniform policy is a change which is immediate, highly visible, and shifts the environmental landscape of any particular school. This change is one that is superficial, but attracts attention because of its visible nature. Instituting a uniform policy can be viewed as analogous to cleaning and brightly painting a deteriorating building in that on the one hand, it grabs our immediate attention but on the other, is, after all, really only a coat of paint. This type of change serves the purpose of attracting attention to schools, it implies that serious problems are existent and necessitate this sort of drastic change, and it seems entirely possible that this attention renews an interest on the parts of parents and communities, and opens the possibilities for support of additional types of organizational change.

The juxtaposition of these findings and the ongoing rhetoric in the public debate on school uniforms provides a lens for viewing the effects of public opinion on school reform in general. The nature and magnitude of the support behind the mandatory uniform policies of districts such as Long Beach seem to illustrate the "quick fix" nature of school reform policies in the 1990's. A policy which is simplistic, readily understandable, cost-free (to taxpayers) and appealing to common sense is one which is politically pleasing and hence, finds great support. When challenged with broader reforms, those whose results are not immediately identifiable, those that are costly and demand energy and a willingness to change on the part of school faculty and parents are simply unacceptable.

REFERENCES

Bryk, Anthony; V. Lee; and P. Holland. (1993). Catholic Schools and the Common Good. Cambridge, MA: Harvard University Press.

Bryk, Anthony and Mary Driscoll. (1988). The High School as Community: Contextual Influences, and Consequences for Students and Teachers. National Center on Effective Secondary Schools.

Coleman, James S. and Thomas Hoffer. (1987). Private and Public High Schools: The Impact of Communities. New York: Basic Books, Inc.

Department of Justice (1996). Manual on School Uniforms. Washington, DC: Office of Elementary and Secondary Education.

Gursky, Daniel. (1996). "'Uniform' Improvement?" The Education Digest, March, 1996, 46-8.

Jarchow, Elaine. (1992). "Ten Ideas Worth Stealing from New Zealand." Phi Delta Kappan. 73:394-95.

Joseph, Nathan. (1986). Uniforms and Nonuniforms: Communication Through Clothing. New York: Greenwood.

Kahl, Kristi. (1996). "Support for Breakaway Teachers May Be the Key to LUBSD Reforms." Found at web site: http://www.lbusd.k12.ca.us/

Kennedy, Michael. (1995). "Common Denominator: Schools See Less Violence When Kids Wear Uniforms." Los Angeles Times, August, 21, 1995.

LaPoint, V., L. Holloman, and S. Alleyne. (1992). "The Role of Dress Codes, Uniforms in Urban Schools." NASSP Bulletin, October 1992, 20-6.

Loesch, Paul. (1995). "A School Uniform Program That Works." Principal. 74:28-30.

Polacheck, Karin. (1996). "Uniforms Help Solve Many School Problems." Long Beach Press-Telegram. Found at web site: http://www.lbusd.k12.ca.us/

Scherer, Marge. (1991). "School Snapshot: Focus on African-American Culture." Educational Leadership. 49:17-9.

Stover, Del (1990). "The Dress Mess." American School Board Journal. 177:26-9.

Thomas, SuSan. (1994). "Uniforms in the Schools: Proponents Say It Cuts Competition; Others Are Not So Sure." Black Issues In Higher Education, October, 20:44-7.

Virginia State Department of Education. (1992). Model Guidelines for the Wearing of Uniforms in Public Schools: Report of the Department of Education to the Governor and the General Assembly of Virginia. House Document No. 27.

Workman, Jane and Kim Johnson. (1994). "Effects of Conformity and Nonconformity to Gender-Role Expectations for Dress: Teachers Versus Students." Adolescence. 29:207-23.

APPENDIX A: LONG BEACH UNIFIED SCHOOL DISTRICT DATA

Crime Category	# of incidents in 1993-1994	# of incidents in 1994-1995	%Change
Assault/Battery	319	212	-34%
Assault with a Deadly Weapon	6	3	-50%
Fighting	1135	554	-51%
Sex Offenses	57	15	-74%
Robbery	29	10	-65%
Extortion	5	2	-60%
Chemical Substances	71	22	-69%
Weapons or Look-alikes	165	78	-52%
Vandalism	1409	1155	-18%
Dangerous Devices	46	23	-50%
Total	3242	2074	-36%

APPENDIX B: DESCRIPTION OF VARIABLE CONSTRUCTION

Dependent Measures

Absenteeism [ABSENT]

F1S13 How many days was R absent from school

Behavior Scale [BEHAVIOR]

F1S10C How many times R got in trouble

F1S10E How many times R suspended from school

F1S9D Got into a physical fight at school

Scale created by summing across values of the three variables (min.=0;

max.=14; Cronbach's alpha= .66)

Substance Use Scale [DRUGS]

F1S77 How many cigarettes does R smoke per day

F1S79 # of times R had five drinks or more in a row

F1S78B Last 12 months, # of times R drank alcohol

F1S78C Last 30 days, # of times R drank alcohol

F1S80AC Last 30 days, # of times R used marijuana

F1S80AB Last 12 months, # of times R used marijuana

Scale created by summing across values of 6 variables (min.=0; max.=22;

Cronbach's alpha= .82)

Academic Achievement

F12XCOMP Standardized Test Composite (Reading, Math)

Key Independent Measures and Moderating Variables

Student Uniform Policy [UNIFORM1]

F1C94F Student Uniforms are required (recoded 0=no; 1=yes) Taken from

School Component of NELS:88

Pro-School Attitudes [PROSCHOL]

F1S12A Feel its ok to be late for school

F1S12B Feel its ok to cut a couple of classes

F1S12C Feel its ok to skip a whole day

F1S12F Feel its ok to get into physical fights

F1S12G Feel its ok belong to gangs

F1S12J Feel its ok to steal belongings from school

F1S12K Feel its ok to destroy school property

F1S12L Feel its ok to smoke on school grounds

F1S12N Feel its ok to use drugs at school

F1S12O Feel its ok to bring weapons to school

Scale was created by summing across all reversed variables (min.=12;

max.=45; Cronbach's alpha= .81)

Peer Pro-School Attitudes [PEERPRO]

F1S70B Among friends, how important is it to study

F1S70D Among friends, how important is it to get good grades

F1S70F Among friends, how important is it to finish High School

Scale was created by summing across the three variables (min.=3;max.=9;

Cronbach's alpha=.75)

Academic Preparedness [PREPARE]

F1S40A Often go to class without paper/pencil

F1S40B Often go to class without books

F1S40C Often go to class without homework done

Scale was created by summing across three reversed variables (min.=3;

max.=12; Cronbach's alpha= .70)

Control Variables

Student Background:

Gender [FEMALE]:

F1SEX Student gender (recoded to 0=male; 1=female)

Minority Status [ASIAN, BLACK, HISPANIC]:

F1RACE Student race (recoded to 0=white; 1=black, Asian, or

Hispanic)

Socioeconomic Status:

F1SES Socioeconomic Status Composite

School Context:

School Sector [CATHOLIC, PRIVNON, PRIVNOT, PRIVREL]:

G10CTRL1 School Sector (recoded 0=public; 1=Catholic, private non-

religious, private religious, or private not ascertained)

Curricular Track [ACADEMIC, VOTECH, OTHPROG]:

F1HSPROG High School Program which R is enrolled in (recoded

0=general; 1=academic, vocational-technical, or other

program)

School District [URBAN, RURAL]:

G10URBAN Type of School District, Diocese, County (recoded

0=suburban; 1=urban or rural)

APPENDIX C: CORRELATION MATRIX OF VARIABLES IN THE ANALYSIS

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22
Female	X
Asian	ns	X
Black	ns	-.08	X
Hispanic	ns	-.07	-.13	X
F1ses	ns	ns	ns	.05	X
Catholic	ns	ns	ns	ns	.03	X
Privnon	ns	ns	-.03	ns	ns	ns	X
Privnot	ns	ns	ns	ns	ns	ns	ns	X
Privrel	ns	.06	-.04	-.04	ns	-.03	ns	ns	X
Academic	.04	.03	-.04	-.06	-.05	.12	.06	ns	ns	X
Votech	-.06	ns	.10	.03	ns	-.06	-.04	ns	-.04	-.23	X
Othprog	ns	ns	.05	.06	.05	-.04	ns	ns	ns	-.17	-.08	X
Rural	ns	-.09	-.07	-.07	-.05	-.17	ns	ns	-.07	-.06	.03	ns	X
Urban	ns	.07	.20	.15	.06	.30	.06	ns	.08	ns	.03	.04	-.42	X
Uniform1	.05	ns	ns	ns	ns	.69	.06	ns	.05	.09	-.05	-.04	-.16	.26	X
Absent	.09	-.06	-.07	.04	.04	-.05	ns	ns	-.03	-.10	ns	ns	ns	ns	ns	X
Behavior	-.19	-.03	ns	.04	.06	-.03	-.03	ns	ns	-.18	.07	.05	-.03	ns	ns	.33	X
Drugs	-.08	-.07	-.14	ns	.06	ns	ns	ns	-.04	-.12	ns	ns	ns	-.05	ns	.28	.50	X
F12xcomp	.05	.04	-.22	-.14	-.07	.10	.08	ns	.06	.39	-.22	.13	-.07	ns	.05	-.10	-.27	-.14	X

Peerpro	.16	.05	.08	ns	ns	ns	ns	ns	ns	.13	-.03	ns	ns	.03	ns	-.16	-.30	-.31	.06	X
Prepare	.17	ns	ns	-.03	-.03	ns	ns	ns	ns	.10	-.04	ns	ns	ns	ns	-.15	-.33	-.25	.12	.23	X
Proschol	.18	.03	.11	ns	ns	ns	ns	ns	ns	.14	ns	ns	ns	.05	ns	-.26	-.60	-.55	.12	.42	.34	X