Perceptions of School Uniforms in Relation to Socioeconomic Statuses

Abstract The purpose of this study was to examine any relationships between students’ perceptions of the effects of school uniforms and student socioeconomic status. A survey was administered to 182 students in a charter school to gather perception information, and a separate survey was administered to parents to gather socioeconomic status information. An examination of individual survey items revealed older students were more likely to report that school uniforms help to reduce bullying and teasing, and some students of high socioeconomic status reported that uniforms help reduce arguments with parents about clothing. Student responses on 10 survey items were grouped into one factor called School Climate, reflecting student perceptions on how uniforms affect the school’s climate. Analyses revealed no significant relationships between the School Climate factor and socioeconomic status. However, Hispanic students reported a significantly more positive response overall than non-Hispanic students. These findings suggest students of various socioeconomic status perceive school uniforms similarly, but older students could be more likely to associate uniforms with a reduction in bullying.


2002) and social status
. Consequentially, the clothing students wear can have great effects on their perception of the school climate and their view of themselves as learners (Francis, 1992;Murray, 1997). Middle grade students need to feel a sense of belonging in school in order to have less anxiety and increased resiliency and academic engagement (Bishop & Pflaum, 2005;Duquette, Akos, & Harrison, 2016). This belonging can be affected by the school climate and school community, and Kommer (1999) noted that an important element that can contribute to a positive school community is the addition of a uniform policy.
Because of the potential effect school uniforms can have on students and their environment, researchers continue to explore schools that implemented uniform policies. Unfortunately, the research on school uniforms reveals a fractured conversation, with emphases including potential impacts on safety and behavior, academic success, school climate, and family dynamics and results varying from uniforms having positive to negative effects in various educational contexts. In addition, much of the research that has been done reports socioeconomic status (SES), but fails to interpret any relationship between SES and results from research on school uniforms. The purpose of the present study was to explore the following research question: What is the relationship between student's perceptions of the effects of school uniforms and students' SES?

Literature Review
The Effects of Uniforms One of the major claims in previous research has been that school uniforms help improve behavior (defined by attendance rates, reported infractions, and bullying) and safety (defined by measure of gang presence, violence, and bullying) in a school. Many researchers have examined whether uniforms affect bullying, gang violence, and other disciplinary issues (Brunsma, 2004;Chime, 2010;Gentile & Imberman, 2012;Pate, 1999;Wade & Stafford, 2003;Walker, 2007;Woods & Ogletree, 1992). Some researchers reported that students had better attendance and fewer suspensions (Pate, 1999;Sowell, 2012;Stockton, Gullatt, & Parke, 2002), while others reported no increase in positive behavior or attendance or change in in-school infractions (Brunsma, 2004;Brunsma & Rockquemore, 1998;Sowell, 2012).
However, the perceptions of parents, teachers, and administrators seem to support the notion that school uniform policies can increase the safety of a school (Bodine, 2003a;Chime, 2010;Hawkins, 2013;Woods & Ogletree, 1992). Students, on the other hand, might not always share the same perceptions as adults. Bodine's (2003b) research indicated that 88% of parents and 86% of teachers felt uniforms increase the safety of the school, but only 43% of students felt the same.
Another claim by proponents of school uniforms is that a uniform policy can increase students' academic success (Sowell, 2012). Once again, this research reported mixed results. Baumann and Krskova (2016) and Gentile and Imberman (2012) saw an increase in academic achievement, while Brunsma (2004), Sowell (2012), and Stockton and colleagues (Stockton et al., 2002) found no statistically significant increase in academic achievement with schools that have uniform policies. Another researcher reported gains in reading scores for elementary boys but not for girls (Pate, 1999). Gains in academic performance may be difficult to correlate with school uniforms because most schools implement a uniform policy along with many other policies designed to improve academics (Bodine, 2003b;Sowell, 2012). Regardless of these difficulties, there is some research suggesting that teachers, parents, administrators, and even students perceive that school uniforms contribute to improved academic success (Hawkins, 2013;Murray, 1997).
Perhaps because it is largely based on the perceptions of research participants, school climate is an area in which the results of research have been a little more consistent. Many researchers have reported positive perceptions from parents, students, teachers, and administrators regarding the effect of uniforms on a school's climate (Baumann & Krskova, 2016;Bodine, 2003b;Chime, 2010;DeCosta, 2014;Hawkins, 2013;Murray, 1997). In some of the research, students reported benefits such as increased safety, more support from teachers, better relationships with peers and teachers, students who listen better, and less bullying and teasing about clothing (Baumann & Krskova, 2016;Bodine, 2003b;Murray, 1997). One of the primary justifications by parents who favor school uniforms was that uniforms help create a protected space where students are free from "markers of economic disparity" (Bodine, 2003a, p. 55), which reduces social exclusion. However, some students and parents feel this equalization through controlling students' clothing restricts students' freedom of expression and denies students the opportunity to learn to work with people who are different (Walker, 2007).
Research on how school uniforms affect the family also has appeared more consistent than research on academics or behavior. Researchers have found that most parents tend to see the use of school uniforms as reducing preparation time in the morning, arguments about clothing, and financial burdens on families (Andersen et al., 2002;Bodine, 2003a;Brunsma, 2004;Chime, 2010;Walker, 2007). Some of this research includes hints that SES might influence parents' perceptions. One study claimed that parents with more education and more income felt uniforms would be more financially burdensome (West, Tidwell, Bomba, & Elmore, 1999). However, this research was conducted in schools that did not require school uniforms. Other research reported parents opting into a uniform program in an effort to reduce their children's concern with clothing ( Bodine, 2003a). Such potential connections between SES and perceptions about school uniforms has received comparatively little attention in the literature. However, according to recent research, 74% of schools with a uniform policy also have a student population where 76% or more qualify for free or reduced lunch (Musu-Gillette et al., 2017).

Clothing and Socioeconomic Status
The social group that accepts a student, which can be marked by clothing, can affect the student's feelings of belonging and autonomy, which in turn can have possible effects on perceptions of academic ability. Caldas and Bankston (1997) found that students are drawn to social groups of similar social status. When looking for a peer group, students search for groups that will meet their need for belonging, autonomy, and competence, and these are typically groups of peers who share a similar SES with them (Battistich, Solomon, Kim, Marilyn, & Schaps, 1994;Francis, 1992). Francis (1992) noted that students may begin this search for a peer group by observing how other students dress and then make judgment calls about SES according to clothing. Another researcher suggested that school uniforms may help conceal a student's SES, thus requiring students to rely on other clues besides clothing to make decisions about socialization (Andersen et al., 2002).
A consequence of socializing with similar peer groups is that students tend to share the peer group's perception of their "academic value" (Wiederkehr et al., 2015, p. 771). Weiderkehr and colleagues found that students have perceptions of their SES, whether conscious or unconscious, that affect perceptions of their ability to succeed. They also found that social class differences in academic achievement can be internalized so that lower SES students have lower self-efficacy and believe they have a lower academic value. Additionally, some researchers have reported that students in peer groups of higher SES achieve more academically while students in peer groups of lower SES achieve less in their academic pursuits (Caldas & Bankston, 1997;Wiederkehr et al., 2015).
The value students place on clothing may come, in part, from the values of their parents (Pilcher, 2011). Parents may be concerned with fashion and fitting in or, even more so, with functionality, and parents tend to pass these values to their children. Researchers have found that families more concerned with name brand clothing and fashion were generally those in higher SES groups, while those who favored function over fashion tended to be in lower SES groups (Bodine, 2003a;Pilcher, 2011).
In most of the research on school uniforms, researchers have used free and reduced lunch as a primary indicator of SES. This research has generally indicated that schools requiring uniforms have a high percentage of low SES students. However, most of this research includes no specific analysis relating to SES. As noted, one study made an attempt to learn about uniforms and perceptions according to SES, but none of the participants were involved in a school that currently required the use of school uniforms (West et al., 1999). The present study aimed to explore the relationship between students' perceptions of the effects of uniforms and such SES variables as whether a student qualifies for free or reduced lunch and the academic attainment of students' parents. Grade, gender, and ethnicity were also included in the study to examine whether they interacted with SES in predicting perceptions of school uniforms. It was hypothesized that students of lower SES would generally view uniforms as having a more positive affect on a student's experience in school.

Participants
The participants in this study were third through eighth grade students from one charter school in Utah. A charter school is a unique setting for a study of this sort because they can potentially draw students from beyond district boundaries. Although some researchers argue that charter schools contribute to segregation of students (Lacireno-Paquet, Holyoke, Moser, & Henig, 2002), the fact that charter schools allow enrollment beyond an immediate surrounding population may allow them to draw families with a larger variety of SES. In addition, in the local context of the present study, charter schools are the only public schools that require the use of uniforms. At the school selected for the present study, the student body consisted of 671 students with a fairly equal ratio of male to female students. Approximately 92% of the students registered at the school identified as White, and about 8% who registered identified as a race other than White (Black, Pacific Islander, Native America, Asian, and any race other than White). The total percentage of students who indicated a Hispanic ethnicity was 14%, with all but one registering as White Hispanic. Approximately one third of the students at the school qualified for free or reduced lunch.

Data Collection
Data were collected through a survey adapted from one designed by Chime (2010). Chime's survey was designed to assess perceptions held by teachers and administrators of the effect school uniforms have on school climate. His survey questions were computed as one score that was analyzed along with varying characteristics of the teachers and administrators who participated. Most of the questions met the needs of this study, but some questions needed to be deleted and new questions added to address additional areas discussed in the research. Questions that were not deleted were simplified to increase comprehensibility by students as young as third grade. Teachers of students in grades three and four were consulted on the wording of the survey and procedures for administration to make sure their students would be able to understand every step of the process.
There were two phases of data collection: (a) a survey on student demographics and parents' education attainment level that was to be filled out by parents and collected along with the parental consent form and (b) a survey with items assessing students' perceptions of the effects of school uniforms. Information on student grade level and qualification for free or reduced lunch was obtained from school administrators. Survey items for students were written in a random order rather than grouped by topic (e.g. academics, school safety, etc.). Each item was a statement, such as "A school uniform helps reduce being bullied or teased" or "A school uniform helps students focus in class." Students responded using a 4-point Likert-type scale that included the following response categories: disagree, somewhat disagree, somewhat agree, and agree.
Before administering the survey, teachers obtained the parental consent forms and demographic surveys from the students. Teachers of students in grades five through eight administered the survey using Google Forms. Teachers of students in grades three and four administered a paper survey in an effort to increase usability and response accuracy for these younger students. All data from Google Forms and paper surveys were transferred to a data analysis program. Entries were double checked to make sure participants completed the parental consent form.

Data Analysis
The survey questions were originally organized into four conceptual categories: safety and behavior, academic well-being, school climate, and family stress. However, factor analyses indicated a poor model fit for the four categories, so-as in Chime's (2010) study-the items were subsequently combined into one factor. Demographic and SES variables were considered predictor variables. The two SES variables, free or reduced lunch and education attainment, were treated as categorical variables. Grade level was treated as a continuous variable, while gender, race, and ethnicity were considered categorical variables. Grade, ethnicity, and gender were first analyzed independently and then used as controls in regression models that examined the variance in student perceptions related to the SES variables.

Demographic Data
Of the 440 third through eighth grade students who were asked to participate, 184 students completed and returned all the forms and surveys (approximately 42% response rate). Not all demographic groups are representative of the school populations as reported above. More female students participated (56%) than males (44%). An average of about 30 students participated in each grade, with fifth grade having the largest number of participants (n = 43) and eighth grade having the lowest number of participants (n = 17). The difference in response rates may be due in part to the fact that the upper grades (grades 5-8) completed the survey online and outside of the school environment, while the lower grades (grades 3-4) completed the survey in class. Ethnicity was reported as Hispanic or not Hispanic, and race was reported as White, Black, Asian, Pacific Islander, Native American, or other. Most participants indicated their race as White and ethnicity as non-Hispanic (about 85%). White Hispanic students represented about 5%, with another 5% nonwhite and Hispanic, and another 5% indicated a race other than White and ethnicity as non-Hispanic. The percentage of students who indicated nonwhite and Hispanic on survey forms for the study does not agree with the percentage of nonwhite Hispanic students reported from the administration as having registered with the school (only one student). This disparity might be due to parents not reading registration or survey forms correctly or thoroughly before submitting. Participants indicating a race other than White were over-represented (about 10%) in comparison to the schoolwide percentage (about 8%). The number of participants indicating Hispanic ethnicity was about 11%, which was smaller than the school wide percentage of 14%. Of the students who participated, about 21% qualified for free or reduced lunch, which is also a smaller percentage than reported schoolwide (33%).

Factor Analysis
We began with descriptive analyses of all the questions on the student perceptions survey (means, variance, and correlations among the items). A factor analysis was then conducted which revealed a poor model fit for the four separate categories. As a result, six questions were not included in the regression analyses (they were removed for conceptual dissimilarity or lack of clarity), and the remaining ten questions were considered as a single factor called "School Climate." Descriptive statistics were then obtained for each predictor variable in relation to the School Climate factor. Finally, a stepwise regression was conducted using grade, gender, ethnicity, and SES variables as predictors of the School Climate factor.
Initial descriptive analyses prior to the factor analysis indicated that most students viewed uniforms somewhat positively. That is, most students responded with "agree" or "somewhat agree" to most items. All 16 questions combined showed a mean score of 2.73 (SD =.63). Items with means above 2.50 indicate a positive response (see Table 1). Items Q3, Q6, Q7, and Q8 showed a more negative response (M < 2.50). There were two questions with a mean above 3.00, Q2 and Q16. It is also interesting to note that Q2 and Q16 also had the smallest standard deviations, indicating less variation in students' responses.
Of the four perception categories, academic wellbeing had the lowest mean responses with means for the questions being mostly negative, except for Q5. This suggests that most students did not perceive uniforms as positively influencing academic wellbeing. The questions in the family stress category showed the highest means, suggesting that students perceived uniforms as reducing family stress. The school climate category also resulted in comparatively high mean scores. Questions in the category of safety and behavior had the largest range in mean scores (low = 2.31, high = 3.37). This could have been because of a larger variation of student interpretations of each question.
Prior to conducting regression analyses, we also examined relationships between separate demographic variables and survey items. There was a statistically significant relationship between grade level and responses to Q11, r = .39, p < .01. Responses to this question reveal a pattern of increasing scores that could be interpreted to mean as students get older, they agree more that school uniforms reduce the amount of teasing or bullying that occurs (see Table 2). In addition, the standard deviation got smaller in higher grade levels, indicating that as students get older, their perceptions on this variable are more similar with less variation in responses.
Question 16 and qualification for free or reduced lunch also showed a statistically significant difference between students, t(182) = 2.66, p < .01. Those who did not qualify for free or reduced lunch had a higher mean score (M = 3.30, SD = .92) than students who did qualify for free or reduced lunch (M = 2.85, SD = 1.07). This may suggest that students who do not qualify for free or reduced lunch are more likely to experience fewer arguments about clothing when students wear a school uniform.
To examine the construct validity of the four perception categories found while reviewing previous studies, a confirmatory factor analysis was conducted. Fit indices for the confirmatory factor analysis suggested poor model fit for the purported factors. For example, the Tucker Lewis Index indicated a value of .81 which does not meet the minimum value of .95 or above (Schreiber et al., 2006). Also, the Root Mean Square Error of Approximation (RMSEA) was .11 where a RMSEA value above .06 is typically considered suggestive of poor model fit (Schreiber et al., 2006). This suggests the four-factor model previously discussed was not supported by the data.
With the new School Climate factor (10 items), descriptive statistics were re-analyzed to explore relationships with demographic variables (see Table 3). The mean student response on the overall factor was 2.57 (SD = .76), which is lower than the previous overall mean with all 16 items. An explanation for this discrepancy might be because four of the six survey items that were not included in the new School Climate factor had mean scores of 2.95 or higher (Q2, Q13, Q14, and Q16). Table 4 includes descriptive statistics of the School Climate factor across all demographic categories used in this analysis-grade level, gender, race, ethnicity, and SES variables. One interesting value is the mean difference in the variable of ethnicity. The variable of ethnicity indicates whether or not a student identified as Hispanic. This variable has the largest mean difference (−.41), with Hispanic students showing a more positive response (M = 2.61, SD = .77) than non-Hispanic students (M = 2.20, SD = .61). Also, the non-Hispanic students showed a lower standard deviation which indicates that more of these students reported similar perceptions with less variance. However, participation among Hispanic students was lower than preferred for statistical analyses.

Regression Analysis
A series of multiple regression models (see Table 4) were conducted to examine the effects of parent education and free or reduced lunch on the dependent variable above and beyond grade, gender, race, and ethnicity. The first multiple regression model included only these other demographic variables as control variables and School Climate as the dependent variable (control model). This regression model had an R 2 of .062, meaning that more than six percent of the overall variation in student perceptions of school uniforms was explained by the demographic variables included in the model. Grade and ethnicity were statistically significant predictors in the model. Higher grades were associated with more positive views of school uniforms, and Hispanic students perceived uniforms more positively than non-Hispanic students.
The second model was conducted to assess whether the SES variables (parents' education and free or reduced lunch) predicted variation in the dependent variable (School Climate) above and beyond the effects of the other demographic variables. In this model, an R 2 = .062 with an R 2 change < .001 was reported, indicating that a parents' education attainment level and whether a student qualifies for free or reduced lunch did not have a significant effect on the students' perceptions of the effects of school uniforms on School Climate above and beyond the effects of the other demographic variables. Grade and ethnicity were also statistically significant predictors in the second model. Higher grades were associated with more positive views of school uniforms, and Hispanic students perceived uniforms more positively than non-Hispanic students.
The third model explored interactions between grade and parent education, with School Climate as the dependent variable. This was done to find the extent to which student perceptions at different grade levels might be impacted by different levels of parent education. The only variables of interest and with enough participants to divide into dichotomous variables for analysis were grade level and parents' education level. A dichotomous variable was created within the grade category such that grades 3-4 were one group and grades 5-8 the other. This grouping was chosen because of the possible effect the physical placement of grades within the school might have had on the students. In the school, kindergarten through fourth grade were on one side of the school and fifth grade through eighth grade were together on the other. This grouping could have effects on students' perceptions as students interacted more with students in their surrounding grade levels, while interactions between the lower grades and upper grades were limited.
New variables were created using the parent education variable to create four interaction terms. The four terms were: (a) grade and mothers without a college degree, (b) grade and mothers with a college degree, (c) grade and fathers without a degree, and (c) grade and fathers with a college degree. Before calculating the interaction, dummy codes were created for the interaction with mothers' education and grade, with mothers without a college degree being omitted and replaced with the mean of that dependent variable. This allowed for examination of just the interaction with grade and mothers without a college degree. The interaction with fathers' education and grade also used dummy codes with fathers without education being omitted and replaced with the mean of that dependent variable. This third model reported an R 2 = .074 with an R 2 change of .012, meaning that 1.2 percent of the variation in School Climate is explained by the interaction terms as explained, above and beyond the other demographic and SES variables. However, neither of the interaction terms were statistically significant. The effect of ethnicity in the third model was still statistically significant, in the same direction as before, though the effect of grade in the third model was no longer significant.

Discussion
This is the only study to our knowledge that has addressed student perceptions of school uniforms in a charter school. Previous studies have examined regular public schools (which we refer to as district schools to separate and clarify that charter schools are also public, but not part of a district), private schools, international schools, and even special behavior schools; but no charter schools. This study offers a unique view on the issue of school uniforms as charter schools involve a different context than most district schools. In addition to being the first study examining the differences of perceptions of the effects of school uniforms by SES in a charter school, this study also included the additional SES variable of a parents' education level. SES is determined by many things, so having more than one variable assessing a student's SES only enhances the results of the study.

Perceptions of Uniforms and SES
Students' perceptions of school uniforms, measured as a single factor, failed to show any significant relationship to SES variables. When controlling for demographic variables, the correlation was even less. This suggests that the two factors of SES, parent education and qualifying for free or reduced lunch, do not affect a students' perception of school uniforms. One reason there was no statistical significance with SES variables might be because of the enrollment process of the charter school. When parents decided to have their children go to this charter school instead of their local district school, they were doing so fully aware that the choice meant their children would have to wear a uniform. Pilcher (2011)  However, the lack of difference in perception by SES may be a positive result for students of low SES. Wiederkehr and colleagues (Wiederkehr et al., 2015) suggested that uniforms may help hide some signs of low SES, which may encourage students to not classify themselves with low academic value. The fact that the present study showed little variations by SES in the perceptions of the effects uniforms had on academics might support Wiederkehr's ideas that students of high and low SES who wear uniforms do not perceive their socioeconomic differences and thus perceive their academic value as equal. Many of these students had been in this school since kindergarten and had not had as much experience searching for peers to socialize with by assessing another's clothing. Because students' SES is masked through the school uniform, it helps make students all appear equal in this regard. Andersen et al. (2002) suggested students look for peer groups of similar SES, which they first perceive by clothing. Again, the fact that this study had similar results from students of various SES might suggest students were socializing with peers with a wider variety of SES. This wide range of socialization could increase their sense of belonging at school because there is less fear of unacceptance from social groups of peers of different SES.

Other Observations
The three items that showed statistical significance are worth discussing even though they are not all directly related to our research question. These items may point to areas where future research may be more fruitful. The first item that showed a relationship between the question about bullying and the variable of grade was interesting. As students got older, they felt more consistently that uniforms did help reduce being teased and bullied. This result was similar to Bodine's (2003a) finding that students felt they were teased less while wearing a school uniform. Bodine said, "When I asked children (including those who oppose common school dress) the best thing they have found about wearing uniforms, virtually all pointed to a reduction in teasing" (p. 55). Maybe as students get older, they begin to notice and value clothing more, and see it as a symbol of wealth (Battistich et al., 1994) or value (Wiederkehr et al., 2015). This new recognition may cause other students to tease those who are different, and older students who have worn school uniforms longer or experienced the younger grades in other schools without uniforms recognize that teasing and bullying about clothing is mitigated by the use of a school uniform. This observation may be valuable to administrators who are constantly searching for ways to reduce bullying in their schools.
The second item with statistical significance was the relationship between students arguing with their parents about clothing choices and free or reduced lunch. This was the only statistically significant relationship with an SES variable. Students who did not qualify for free or reduced lunch agreed more that school uniforms reduce conflicts between parents and child when choosing appropriate clothing. This finding is similar to Bodine's (2003a) observation that the majority of students in her study also reported fewer arguments with parents about clothing when getting ready for school. A possible explanation could be that parents of students who do not qualify for free or reduced lunch have enough income that they do not need assistance to buy lunch, which also means they may have the ability to purchase more stylish clothes for their children. This explanation supports Bodine's (2003a) and Pilcher's (2011) claims that students of higher SES are more concerned with name brand clothing and fashion. These stylish clothes may ignite tensions when the children do not get to wear them to school to show their friends their stylish clothes because one part of the clothing item may be inappropriate for school. Having school uniforms simplifies the choices of clothing for parents and students, thus reducing conflict over clothing. On the other hand, children whose parents qualify for lunch assistance may not have the income to spend on stylish clothes. Recognizing that their parents might not have money for more expensive or stylish clothing purchases, these children would have less reason to argue with their parents over clothes. For these students, wearing a school uniform would not reduce conflict as much if the conflict was less to begin with.
The third statistically significant item showed that, overall, Hispanic students viewed uniforms more positively. There could be numerous possibilities as to why Hispanic students could perceive uniforms as more helpful. It is possible that these students or their parents viewed their personal family situation in that they are somewhat new to the area, and feel conformity is more important than popularity or fighting the status quo. Uniforms may also contribute to Hispanic students feeling less separated from the student body which creates a sense of belonging with the school (Hurd, 2012). This phenomenon has intriguing implications for further research. Why do Hispanic students have different perspectives on school uniforms than non-Hispanic students? With such a small number of Hispanic participants, would similar results be seen if the same survey was done on a larger scale? In this study, Hispanic students represented a small percentage of the participants. How do Hispanic students view uniforms in schools where they are the majority instead of the minority?
Generally, students' responses were positive, suggesting a positive perception of uniforms by most students. This positive perception may contribute to the school climate, creating a positive atmosphere which, as Cohen, Mccabe, and Michelli (2009) explained, is what is needed for students to learn what is necessary to have a "productive, contributive, and satisfying life" (p. 182). This conclusion agrees with other researchers who have claimed that school uniforms have a positive effect on a school's climate (Chime, 2010;Hawkins, 2013;Murray, 1997). However, because the study took place in a charter school, it is hard to determine how much of the difference in positivity can be determined by the different demographics or culture created by the methods, policies, and staff of the charter school.

Limitations
Because parents chose to enroll their children in the charter school, they were fully aware of the uniform policy and may have some level of prior acceptance of or bias toward uniforms, and this bias may have transferred to the students. Also, the extent to which findings in the charter school setting can be generalized to the surrounding district school settings is limited by dissimilarities in demographics, school policies, and instructional philosophies. Finally, the use of free or reduced lunch as a proxy for SES has limitations because families can qualify for free or reduced lunch with a wide range of incomes, education attainment, and employment depending on the size of their family. We added the variable of parent educational attainment to provide additional information and offset this limitation. Future studies might include family size, regional cost of living, or occupation.

Conclusion
In conclusion, this study did not reveal overall differences in students' perceptions on the effects of school uniforms by SES. However, with more research this conclusion may help support some of the purposes stated by proponents for school uniforms-that uniforms may help equalize the playing field (Andersen et al., 2002) among students of various SES and help improve a school's climate. This study also supports the claim that uniforms help to reduce bullying and may reduce conflicts between parents and children regarding choice of clothing. Finally, this study and the current research seem to support the idea that school uniforms contribute to more positive school climates even with a relatively high SES population. Currently, most schools that have uniform policies have a larger low-SES population (Musu-Gillette et al., 2017), but the present study, which was done in a school with a lower population of low-SES students, suggests that communities with a higher-SES population may also be supportive of a uniform policy.
Future research might explore school uniforms in multiple school contexts (e.g., comparing student perceptions in charter, public, and private schools) or in schools with greater demographic diversity than the present study. Researchers might also employ qualitative or mixed methods designs to gain greater insights into students' perceptions and focus their research on other key variables, such as parent involvement or variations in perceptions of students who identify as Hispanic compared to other populations.