Body Weight and Academic Achievement: The Role of Weight Diversity in Urban Middle Schools

Abstract

The current study was designed to examine one possible weight stigma-reduction mechanism: school-level weight diversity. It was hypothesized that greater weight diversity among same-sex peers at school would attenuate the negative association between weight and academic achievement. Across 26 urban public middle schools, 5,991 sixth grade students (52% girls) were included: 12% African-American/Black, 14% East/Southeast Asian, 30% Latino, 21% White, 23% from other specific ethnic groups. Weight diversity was estimated as the likelihood that two randomly selected students would be from different weight categories using Simpson’s diversity index. Standardized achievement test scores and grade point average (GPA) were used to assess academic achievement. Consistent with our contextual moderator hypothesis, high levels of weight diversity at school served a protective function. The negative association between body mass index (BMI) and achievement test scores, as well as GPA (girls only) was non-significant at schools with high levels of weight diversity. The study findings offer a potential explanation for inconsistent findings regarding body weight and achievement, and a novel methodological approach to capture weight diversity in ways that provide new insights for school-based interventions.

Among adolescents, high body weight is associated with poor physical health (Reilly & Kelly, 2011), social marginalization (Ames & Leadbeater, 2017; Puhl & Latner, 2007), as well as depression and low self-esteem (Puhl & Luedicke, 2012) – problems that can be accounted for in part by the social stigma of overweight (Puhl & Heuer, 2009). Weight stigma is also presumed to disrupt school-related adjustment by threatening the academic motivation and engagement of students with overweight (Crosnoe, 2007; Larsen, Kleinjan, Engels, Fisher & Hermans, 2014). However, the association between high weight and low achievement is inconsistent across studies (Caird et al., 2014; Martin et al., 2017). Thus, it is important to examine the role of the school context, as it may be more stigmatizing - and therefore more academically consequential - to have high body weight in some schools than in others. The present investigation specifically examines whether weight diversity of middle schools – reflecting the variablity of body shapes and sizes – moderates the association between body mass index (BMI) and academic performance in sixth grade.

The school setting is the most common context for weight-related stigmatization among youth (Neumark-Sztainer, Story & Faibisch, 1998). As targets of peer victimization and negative teacher biases, adolescents with high BMI contend with a variety of social difficulties (Ames & Leadbeater, 2017; Puhl & Latner, 2007). The negative social feedback they receive from peers can lead to emotional problems (Juvonen, Lessard, Schacter & Suchilt, 2017) and set students on a trajectory of withdrawal and school disengagement. In addition, readily endorsed stereotypical expectations by teachers and peers that overweight students are lazy, unsuccessful, and unintelligent (Puhl & Heuer, 2009) may promote an internalization of weight stigma, compromising motivation and, ultimately, lead to academic decline (Crosnoe, 2007).

The intensity of weight stigma is likely to vary across schools, however (Juvonen, Lessard, Schacter & Enders, 2018). Schools where high weight makes a student “stand out” are likely to be more stigmatizing and in turn academically consequential. Indeed, report card grades of overweight students show a more precipitous decline over time in schools with a lower average BMI, compared to schools with a higher average BMI (Crosnoe & Muller, 2004). Another study, based on the National Longitudinal Study of Adolescent Health (Add Health), found that girls with high weight attending high schools with a lower prevalence of obesity were less likely to enter college than those who attended high schools with a higher prevalence of obesity (Crosnoe, 2007). Although students with high weight may appear to be academically “protected” in schools with higher body weight norms, the social contagion of overweight may promote unhealthy habits and additional weight gain (Christakis & Fowler, 2007). Thus, it is important to investigate other potential school factors to learn about contexts that might help even the academic playing field for all students, regardless of body weight.

Drawing from literature on other types of social stigmas (e.g., racial/ethnic minorities; disability), we presume that overall diversity can function as a stigma-reducing mechanism. Specifically, we propose that observations of greater variation in body shapes and sizes are likely to widen the range of acceptable weight norms, thereby reducing the stigma of high weight. There is some indirect evidence for this assumption across studies that rely on samples with varying degrees of ethnic diversity. For example, Chomitz and colleagues (2009) found that high weight was unrelated to achievement within an ethnically diverse urban public school district. In contrast, among predominantly White samples, high weight has been identified as an academic risk factor (e.g., Castelli, Hillman, Buck & Erwin, 2007; Shore et al., 2008). Moreover, beyond the achievement domain, greater ethnic diversity has been shown to attenuate the social consequences of high weight (i.e., peer victimization; Lanza, Echols & Graham, 2018). Hence, appearance norms may be wider in ethnically heterogeneous schools as compared to more homogenous schools. While ethnic diversity might indirectly capture diversity of weight given ethnic differences in rates of obesity (Ogden et al., 2016) and body shape ideals (Gordon, Castro, Sitnikov & Holm-Denoma, 2010), it is important to directly assess body weight diversity.

Although to date no study has examined school-level weight diversity, experimental evidence has shown that repeated exposure to different body shapes influences weight preferences and attitudes by recalibrating what is considered normative. Specifically, after being presented with a series of images of overweight bodies, adults showed greater acceptance of individuals with overweight and more positive judgments about their intelligence, while presentation of low weight bodies had the opposite effect (Robinson & Christiansen, 2014). Thus, in schools with greater exposure to diverse body shapes and sizes, students with non-normative body types, including both those with high (and low) weight, may be less likely to be stigmatized in ways that compromise their ability to excel academically.

Current Study

The primary goal of the current study is to investigate whether school-level weight diversity moderates the association between BMI (as an indicator of body weight) and academic achievement in sixth grade. Presuming that weight diversity fosters or reflects a more inclusive environment, we expect the negative association between BMI and achievement to be minimized in schools with greater weight diversity. To capture weight diversity, we rely on a novel measure that varies as a function of both the number of weight categories (e.g., underweight, normal weight, overweight) present in any one school and the relative representation of weight categories in the school. Specifically, weight diversity is calculated using Simpsons Index, an indicator of heterogeneity that has been used to capture school ethnic diversity (Juvonen, Nishina & Graham, 2006). The weight diversity estimates were obtained separately for girls and boys because standardized BMI scores are based on sex-specific national norms.

The present study contributes to the existing research in several ways. First, we focus specifically on early adolescence when, despite substantial variability in body shapes and sizes related to variations in pubertal development (Loomba-Albrecht & Styne, 2009), “fitting in” is particularly challenging for youth with overweight due to heightened appearance norms (Tremblay & Lariviere, 2009) and concerns of social approval (Blakemore & Mills, 2014). Early adolescence is also a developmental period critical in shaping subsequent academic trajectories (Juvonen, Le, Kaganoff, Augustine & Constant, 2004). Second, complementing studies examining body weight averages at school (Crosnoe, 2007; Crosnoe & Muller, 2004), the current study is the first to empirically capture body weight diversity. Moreover, we examine the contextual moderator function of weight diversity while taking into account other school-level factors likely to decrease the salience of high weight: school average BMI and the ethnic diversity of the student body (Lanza et al., 2018). Finally, our large ethnically diverse sample provides an ideal context for examining weight diversity in urban public schools where students are at high risk for both obesity (Ogden et al., 2016) and low academic performance (Orfield, 2005). Because such disparities are frequently associated with socioeconomic disadvantage (Shin & Miller, 2012; Sirin, 2005), we also take into account students’ parental education level. With the diversity of the sample, our goal is to obtain findings generalizable across a wide range of demographic groups.

Method

Participants

The current study relies on a longitudinal sample recruited from 26 public middle schools in California that varied systematically in ethnic composition (N=5,991; 52% female). Based on self-report ethnicity, the sample was 32% Latino/a, 20% White, 13% Asian, 12% African American and 23% from other ethnic groups, including biracial and multi-ethnic youth.

Procedure

The study was approved by the relevant Institutional Review Board and school districts. All eligible sixth grade students and families received informed consent and informational letters. Parental consent rates averaged 81% across the schools. Data collection was conducted in schools, and students received $5 in the spring of sixth grade and $10 in seventh and eighth grade for completion of the surveys.

Measures

Academic achievement outcomes.

Two indicators were used to assess academic achievement at the spring of sixth grade: standardized achievement tests scores and grade point average.

Achievement test scores.

An annual California Standards test (CST) provides indices of language arts and mathematics proficiency based on California content standards at each grade level. Test scores from the language arts and mathematics sections were combined with possible scores ranging from 300 to 1200 (girls: M=761.03, SD=127.13; boys: M=755.42, SD=129.79).

Grade point average (GPA).

School transcripts were collected at the end of each semester. Grades for academic courses (i.e., Math, Science, Social Studies/History, Language Arts/English) from each semester were coded on a 5-point scale (A=4 and F=0) and then averaged to create a composite GPA for each student (girls: M=3.05, SD=.92; boys: M=2.71, SD=1.04).

Individual-level predictors.

In addition to relying on BMI to capture weight (relative to height), demographic data relevant to achievement were used as covariates.

Body mass index (BMI).

BMI was calculated in the spring of sixth grade based on participants’ self-reported height and weight. This methodology is acceptable given that BMI scores calculated from self-report are usually consistent with those calculated from direct measurement for this age group (Strauss, 1999; Vaughan & Halpern, 2010). Taking into account participants’ self-reported age and sex, BMI z-scores were calculated based on Centers for Disease Control and Prevention (CDC) 2000 growth charts (girls: M=.12, SD=1.08; boys: M=.31, SD=1.10).

Covariates.

Several variables known to predict academic achievement in early adolescence were used in the analyses. Students reported their sex and ethnicity in the sixth grade. In addition, parents/guardians, who completed informed consent, indicated their highest level of education on a 6-point scale (1=elementary/junior high school to 6=graduate degree) in the beginning of the study (girls: M=3.95, SD=1.55; boys: M=4.05, SD=1.52).

School-level predictors.

In addition to sex-specific weight-diversity, we took into account the sex-specific school average BMI as well as the ethnic diversity of the school.

Weight diversity.

Weight diversity was computed at each school, separately for girls and boys, as a function of the prevalence and relative representation of weight categories using age-and sex-specific cut points from the CDC 2000 growth charts. Because the CDC weight categories are asymmetrically defined with a greater number of high weight categories, a fifth “at risk for underweight” category was created (e.g., Wang, Iannotti & Luk, 2010) to differentiate weight diversity from school weight norms¹. Thus, adolescents were categorized as underweight (<5^th percentile), at risk for underweight (5^th to <15^th percentile), normal weight (15^th to <85^th percentile), overweight (85^th to <95^th percentile), or obese (≥95^th percentile). Our sample included representation across all five weight categories (girls: 6% underweight, 8% at risk for underweight, 67% normal weight, 12% overweight, 7% obese; boys: 4% underweight, 5% at risk for underweight, 65% normal weight, 16% overweight, 10% obese).

An index of weight diversity was created using a formula that captures both the number of different groups in the setting and the relative representation of each group (Juvonen et al., 2006; Simpson, 1949):

$$D_C=1-\sum _{i=1}^g{p}_i^2$$

D_C refers to the weight diversity of a given school and p indicates the proportion of students in the same grade at school belonging to a given weight category. This proportion in turn is summed across weight category groups at a school and subtracted from one, such that Simpson’s Index indicates the probability of any two students chosen at random in a school being from different weight categories, with possible values ranging from 0 to 1 (higher scores representing greater diversity). The weight diversity index, computed separately for girls and boys, ranged from 0.40 to 0.68 for girls (M=.53 SD=.08) and from 0.37 to 0.70 for boys (M=.53 SD=.07), suggesting moderate to relatively high levels of heterogeneity across the 26 middle schools.

School average BMI.

To be able to take into account the descriptive body weight norms in each school, we computed school-specific BMI averages for boys and girls. Individual sex-normed BMI z-scores (based on national standards) were summed and divided by the number of boys and girls in each school. The range of schoolwide BMI z-scores ranged from −0.30 to 0.74 (M=.15, SD=.28) for girls and −0.24 to 0.71 (M=.32, SD=.23) for boys, suggesting substantial variation across schools.

Ethnic diversity.

Given the variability of school ethnic composition, ethnic diversity was included as a school-level covariate. Data from the California Department of Education were used to compute Simpson’s diversity index (Juvonen et al., 2006). In the current study, ethnic diversity ranged from 0.48 to 0.77 (M=.64, SD=.08).

Data Analyses

The data were analyzed in Mplus 7.4 using a standard multilevel linear model to account for students nested within 26 middle schools. All analyses were conducted separately for girls and boys because BMI norms are sex-specific. The regression models accounted for student-level ethnicity (4 dummy coded variables with White youth as the reference group) and parent level of education. At the school-level, overall ethnic diversity, sex-specific BMI average as well as weight diversity were included. A cross-level interaction term between BMI and school-level weight diversity was included to test our main contextual moderator hypotheses. Initial exploratory analyses examining whether the moderating effect of weight diversity varied as a function of ethnicity were also tested with three-way interactions; however, given that the three-way interaction terms (i.e., ethnicity × BMI × weight diversity) were non-significant, they were excluded from the final regression models.

Following recommendations for models including cross-level interactions (Enders, 2013), all continuous individual-level predictors were grand-mean centered (sex-specific). For statistically significantly interactions, tests of simple slopes were conducted to compare the achievement of youth at schools one standard deviation below, at, and one standard deviation above the mean of weight diversity. All models included a random intercept, and for each outcome we first tested unconditional intercept-only models to determine the intraclass correlation, or the degree of similarity between individuals due to shared cluster membership (Hox, Moerbeek & van de Schoot, 2010). Due to between-school variability in achievement test scores (ICC_girls=.11; ICC_boys=.13) and GPA (ICC_girls=.07; ICC_boys=.09), we relied on multilevel modeling to account for the nested data.

Results

The results are divided into two main sections. First, given the novel approach to assessing school-level weight diversity, we provide descriptive information about this contextual measure by examining its correlation to the standard deviation and mean of individual BMI scores within each school. Second, we present the multilevel regression models examining the association between BMI and academic achievement, and whether these associations are moderated by weight diversity. The main findings are presented for girls followed by those for boys.

Weight Diversity and School Average BMI

To test the validity of our weight diversity index, bivariate correlations between weight diversity and the standard deviation of BMI z-scores (based on national norms) were computed separately for girls and boys across the schools. As expected, the values were positively related, such that greater school-level weight diversity was associated with greater variation in individual BMI z-scores (girls: r=.55, p=.003; boys: r=.46, p=.018). For girls, school-level weight diversity was unrelated to average BMI (r=−.03, p=.142). In contrast, for boys, greater weight diversity was associated with higher school average BMI (r=.59, p<.001), suggesting that in schools with greater body weight heterogeneity, a larger proportion of boys had high weight. To examine the role of weight diversity over and above the school average BMI, we control for sex-specific school BMI averages in the main analyses.

Girls

As shown in the upper diagonal of Table 1, BMI was negatively associated with parent level of education. In addition, as expected, BMI was negatively related to achievement test scores and GPA. With the exception of the intercorrelations between the two achievement indicators, all bivariate correlations were small to moderate in magnitude.

Table 1.

Correlations Between Continuous Individual-Level Variables.

Variable	BMI	Parent Education	Achievement Test Scores	GPA
BMI	---	−.13*	−.12*	−.13*
Parent Education	−.10*	---	.30*	.23*
Achievement Test Scores	−.09*	.28*	---	.63*
GPA	−.09*	.27*	.65*	---

Note. Values for boys are below the diagonal and values for girls are above the diagonal. All variables were group-mean centered to reflect within-school associations between the individual-level variables.

^*p<.001.

Table 2 displays the summary of the multilevel models testing our contextual moderation hypothesis. Individual level covariates revealed that compared to White girls, African American, Latina and girls of other ethnic backgrounds (e.g., biracial) received lower achievement test scores and grades, while Asian girls received higher achievement scores. In addition, higher BMI and lower levels of parent education were associated with lower achievement test scores and GPA. At the school-level, girls received lower test scores and grades in schools with higher average BMI, regardless of their weight. Controlling for these demographic differences, the hypothesized individual BMI by school weight diversity interaction was significant for achievement test scores as well as GPA. To be able to interpret these cross-level interactions, follow-up analyses were conducted by examining simple slopes between BMI and academic achievement at low (1 SD below average), average, and high (1 SD above average) levels of weight diversity. Consistent with similar studies (Sutter, Nishina, Witkow, & Bellmore, 2016), we graphed simple slopes with a focus on the right tail of the BMI distribution to capture youth with overweight and obesity.

Table 2.

Summary of Regression Models of Individual and School-level Covariates and Weight Variables on Academic Achievement Outcomes for Girls.

	Achievement Test Scores			GPA
Predictors	b	SE	95% CI	b	SE	95% CI
Individual-level
Intercept	798.28***	7.28	781.19, 810.38	3.24***	0.04	3.15, 3.31
African American	−103.05***	12.51	−127.57, −78.52	−0.65***	0.08	−0.80, −0.50
Asian	25.20**	8.99	7.59, 42.82	0.26***	0.05	0.16, 0.37
Latino	−69.94***	10.59	−90.69, −49.19	−0.33***	0.06	−0.45, −0.21
Other	−39.74***	7.36	−54.16, −25.32	−0.21***	0.04	−0.28, −0.13
Parent Education	18.34***	2.56	13.33, 23.35	0.11***	0.02	0.08, 0.14
BMI	−4.22*	2.15	−7.72, 0.39	−0.05***	0.01	−0.07, −0.02
School-level
BMI Average	−45.70**	15.00	−75.11, −16.31	−0.33*	0.17	−0.66, −0.01
Ethnic Diversity	64.75	47.75	−28.86, 158.33	−0.18	0.37	−0.89, 0.54
Weight Diversity	−101.16	52.13	−203.33, 1.00	−0.50	0.49	−1.45, 0.46
Cross-level interaction
BMI × Weight Diversity	85.43**	32.13	22.46, 148.11	0.32*	0.16	0.01, 0.63

Note.

^***p<.001,

^**p<.01,

^*p≤.05.

SE=standard error; CI=Confidence interval. Coefficients represent unstandardized estimates. Ethnicity reference group=White.

Achievement test scores.

The significant cross-level interaction between BMI and weight diversity (b=85.43, p=.008) suggests that the association between BMI and achievement scores varies across schools depending on the schools’ weight diversity among girls. Analyses of simple slopes (see left panel of Figure 1) showed that at schools with low (i.e., −1 SD; b=−11.06, p=.005) and average (b=−4.22, p=.049) levels of weight diversity, higher BMI was associated with lower test scores, whereas BMI was unrelated to test scores at schools with high levels of weight diversity (i.e., +1 SD; b=2.61, p=.315).

Figure 1.

The Moderating Role of Weight Diversity on the Association Between BMI and Achievement Test Scores for Girls (left) and Boys (right).

Note. The values presented reflect achievement test scores for individuals in schools with average, low or high weight diversity (± one SD above or below the mean), and at the mean on all other variables. ***p<.001, **p<.01, *p<.05.

GPA.

The analyses yielded also a significant cross-level interaction between BMI and weight diversity for girls’ GPA (b=.32, p=.050). Tests of simple slopes (showing a similar pattern of findings as above and therefore not graphed) revealed that at schools with low and average levels of weight diversity, higher BMI was associated with lower grades (b= −.08, p<.001, b= −.05, p<.001, respectively for low and average weight diversity). However, at schools with high levels of weight diversity, BMI was unrelated to girls’ GPA (b=−.02, p=.209).

In sum, the association between girls’ body weight and achievement varied across schools. In contrast to low and average weight diverse schools, in middle schools with high levels of weight diversity, BMI was not related to girls’ achievement.

Boys

Correlations among the individual-level variables showed the same pattern of findings as for girls (see the lower diagonal of Table 1). Coefficients and standard errors from the multilevel models for boys are shown in Table 3. Individual-level covariates revealed that compared to White boys, African American and Latino and boys of other ethnic backgrounds (e.g., biracial) received lower test scores and GPA, while Asian boys received higher achievement scores. In addition, higher parent education was related to higher achievement, while BMI was related to lower achievement.

Table 3.

Summary of Regression Models of Individual and School-level Covariates and Weight Variables on Academic Achievement Outcomes for Boys.

	Achievement Test Scores			GPA
Predictors	b	SE	95% CI	b	SE	95% CI
Individual-level
Intercept	784.71***	7.70	768.74, 798.93	3.00***	0.06	2.88, 3.12
African American	−101.46***	12.88	−126.69, −76.21	−0.83***	0.09	−1.01, −0.66
Asian	28.49**	9.44	10.00, 46.98	0.20***	0.06	0.09, 0.31
Latino	−59.13***	8.01	−74.82, −43.44	−0.49***	0.05	−0.59, −0.39
Other	−32.74***	7.70	−47.83, −17.64	−0.25***	0.05	−0.34, −0.15
Parent Education	19.94***	2.24	15.55, 24.32	0.15***	0.02	0.12, 0.19
BMI	−4.40*	1.98	−8.38, −0.61	−0.04**	0.02	−0.07, −0.01
School-level
BMI Average	−64.48*	26.98	−117.35, −11.60	−0.30	0.27	−0.83, 0.24
Ethnic Diversity	88.77	46.04	−1.47, 179.02	0.12	0.61	−1.08, 1.31
Weight Diversity	−48.61	80.38	−206.17, 108.94	0.48	0.95	−1.38, 2.35
Cross-level interaction
BMI × Weight Diversity	61.82*	29.70	3.58, 120.18	0.07	0.31	−0.54, 0.67

Note.

^***p<.001,

^**p<.01,

^*p<.05.

SE=standard error; CI=Confidence interval. Coefficients represent unstandardized estimates. Ethnicity reference group=White.

Achievement test scores.

Consistent with our contextual moderation hypothesis, a significant BMI by weight diversity interaction was obtained (b=61.82, p=.037). As shown in the right panel of Figure 1, the graphs of simple slopes revealed that higher BMI was associated lower test scores at schools with low (b=−8.73, p=.003) and average (b=−4.40, p=.026) levels of weight diversity, whereas BMI was unrelated to test scores at schools with high levels of weight diversity (b=−.07, p=.980). Also, boys received lower achievement test scores in schools with higher average BMI, regardless of their weight (b=−64.48, p=.017).

GPA.

The cross-level interaction between BMI and weight diversity was non-significant (b=.07, p=.826). When the interaction term was removed from the model, over and above individual and school-level covariates, higher BMI was related to lower GPA.

Taken together, the results suggest that for boys, weight was unrelated to achievement test scores in schools with greater weight diversity. While there was no support for weight diversity as a contextual moderator for GPA, boys with higher BMI received lower grades compared to their lower BMI peers. Also, boys received lower achievement test scores in schools with higher average BMI, regardless of their weight.

Discussion

In addition to the negative physical, social and emotional consequences of weight stigma, high body weight places overweight youth at a disadvantage in terms of college enrollment and subsequent options for employment (Crosnoe, 2007; Puhl & Heuer, 2009). In light of such findings, it is imperative to examine how body weight is related to academic achievement earlier in development. By focusing on differences across 26 middle schools, we find evidence for weight diversity as a potential stigma-reducing mechanism during early adolescence. Utilizing a novel measure of school weight diversity, the current findings underscore the need to understand the ways in which exposure to various body shapes and sizes can create a more inclusive environment and promote academic resilience among youth with non-normative body types.

Consistent with our contextual moderator hypothesis, the results suggest that the academic risk of high weight can be dialed down or turned up in any one middle school (Crosnoe, 2011). Specifically, while higher weight was a risk factor for lower standardized test performance in schools with less weight diversity, BMI was unrelated to achievement scores in schools with greater variation in body weight. Similar effects were also obtained for GPA, but only among girls. The less robust effects of weight diversity for boys may reflect greater weight stigma and a persistent thin-idealization among girls (Tang-Peronard & Heitmann, 2008; Durante, Fasolo, Mari & Mazzola, 2014). Contexts that minimize the salience of body weight ideals (e.g., weight diversity) may therefore be less impactful on boys’ weight norms (Crosnoe, 2007). In addition, the differential effects should also be interpreted in light of our descriptive findings. While the weight diversity index was not associated with school average BMI for girls, among boys, schools with more weight diversity had higher BMI averages. Thus, it may be that the protective effects of weight diversity are achieved only when the degree of heterogeneity is unrelated to normative levels of (high) weight.

We presume that the current findings regarding the protective function of weight diversity reflect a more inclusive school environment. Similar to studies that show a positive association between greater ethnic diversity and lower sense of social vulnerability (Juvonen et al., 2006; Juvonen, Kogachi, & Graham, 2017), we expect that youth with both high or low weight are less negatively stereotyped (by both peers and teachers) in schools with greater weight diversity and hence are in the position to do well academically. Although testing a mediating mechanism was beyond the scope of the current study, we expect that at schools where the body norms are not too narrow or heavily “policed” by peers (see Juvonen et al., 2018), overweight youth are more eager to attend and engage in schoolwork. It is also possible that weight diversity may reduce teachers’ stereotyped beliefs of students with overweight and in turn weight-based discriminatory grading (MacCann & Roberts, 2013).

Limitations

While the results of the present study are promising, they must be interpreted in light of a number of methodological limitations. For example, we had no objective weight and height data to compute the BMI scores. It is therefore critical that the current findings are replicated using anthropometric data. For young adolescents, it would also be important to take into account objective assessment of pubertal maturation (e.g., Tanner stages). Additionally, although we had several reasons to focus on early adolescence given the heightened social costs of not fitting in with body norms (Tremblay & Lariviere, 2009), we do not know whether our findings generalize to other developmental periods and phases of schooling (i.e., elementary or high school). Given that children as young as age three describe overweight peers as mean, ugly, stupid, lazy, and undesirable playmates (Cramer & Steinwert, 1998), future research examining the effects of exposure to diverse body shapes and sizes during early childhood is warranted. Furthermore, in our ethnically diverse California-based sample, the weight distribution was skewed with an overrepresentation of overweight and obese. Given the lack of schools with low weight norms, additional research is therefore needed to examine the effects of weight diversity in other samples. In schools where overweight is rare, and thus potentially more stigmatizing, weight diversity may be especially critical.

Consistent with a large body of research (e.g., Orfield, 2005), our descriptive findings show that African American and Latino students as well as youth from lower SES backgrounds were at increased risk for low achievement. In addition, students performed more poorly at schools with higher average BMI, regardless of their weight. As such, likely reflecting a number of structural characteristics (e.g., quality of schools, access to after-school activities), our findings underscore how health and academic risk are intertwined. While the schools in the current sample all included at least moderate levels of ethnic diversity, it is important to recognize that youth in racially isolated and predominantly low-income schools typically have less access to academic (e.g., tutoring) as well as physical health (e.g., sports, physical activity programs) resources (Basch, 2011). Thus, to understand weight-related achievement disparities, additional research addressing how race/ethnicity and SES are related to structural opportunities provided by schools is needed. Finally, our cross-sectional analyses limit causal inferences about the association between BMI and academic achievement. We presumed that obesity precedes poor academic outcomes; however, poor school performance may also serve as a stressor leading to weight gain. Such alternative – and possibly bidirectional – effects should be tested in future research with cross-lagged panel analyses.

Implications

The conceptual frameworks guiding past studies on the school-related adjustment of youth with high weight can yield some questionable implications. Although consistent with person-by-environment theories (Wright, Giammarino & Parad, 1986; Crosnoe, 2011), it would be simplistic, for example, to conclude that overweight youth would be better off in schools with heavier weight norms (i.e., higher BMI average). While higher weight may be less salient and more socially acceptable in such schools, these environments may exacerbate unhealthy habits and weight gain because of peer influence (Christakis & Fowler, 2007). In addition, school-based health interventions that focus on obesity prevention should be careful about the messages they send (Kenney, Wintner, Lee & Austin, 2017). For example, programs that underscore the importance of personal responsibility by promoting healthy eating habits and exercise (Stice, Shaw & Marti, 2006) can exacerbate weight stigma (i.e., high weight youth are blamed for not losing weight) and in turn worsen educational disparities (e.g., Russell-Mayhew, 2006).

Although weight diversity may promote a more inclusive environment and foster positive relations and attitudes across students of varying weight categories, not all schools have much weight heterogeneity within their student body. Schools lacking weight heterogeneity may therefore need to rely on direct instruction to promote acceptance of diverse body shapes and sizes. Programs targeting stereotypes about higher weight have been effective in reducing negative attitudes and teasing of overweight students in elementary school (e.g., Irving, 2000). Using developmentally sensitive methods to talk about acceptance of all individuals regardless of physical appearance may help reinforce the message of body acceptance among students and as such equalize opportunities for academic success.

Impact and Implications:

The study results suggest that while overweight may be an academic risk factor in schools with little weight heterogeneity among students, weight-based achievement disparities appear to be minimized in middle schools with greater exposure to diverse body shapes and sizes.