ORIGINAL ARTICLE

Navigating Middle Grades: Role of Social Contexts in MiddleGrade School Climate

Ha Yeon Kim • Kate Schwartz • Elise Cappella •

Edward Seidman

Published online: 16 May 2014

� Society for Community Research and Action 2014

Abstract During early adolescence, most public school

students undergo school transitions, and many students

experience declines in academic performance and social-

emotional well-being. Theories and empirical research

have highlighted the importance of supportive school

environments in promoting positive youth development

during this period of transition. Despite this, little is known

about the proximal social and developmental contexts of

the range of middle grade public schools US students

attend. Using a cross-sectional dataset from the eighth

grade wave of the Early Childhood Longitudinal Study—

Kindergarten Cohort 1998–1999, the current study exam-

ines the middle grade school social context from the per-

spectives of administrators and teachers in public schools

with typical grade configurations (k–8 schools, middle

schools, and junior high schools) and how it relates to

students’ perceptions of school climate. We find that

administrators and teachers in k–8 schools perceive a more

positive school social context, controlling for school

structural and demographic characteristics. This school

social context, in turn, is associated with students’ per-

ceptions of their schools’ social and academic climate.

Implications for educational policy and practice are

discussed.

Keywords Middle grade schools � School climate �School transitions � Adolescent development � School

reform � Educational policy

Introduction

Early adolescence is a vulnerable period for both academic

and social-emotional development (Eccles and Midgley

1989; Simmons and Blyth 1987). Many early adolescents

experience declines in their academic performance (Barber

and Olsen 2004), motivation (Maehr and Midgley 1996),

and engagement (Archambault et al. 2009). Social-emo-

tional struggles ensue during this period, including

decreases in students’ sense of belonging in school (Maehr

and Midgley 1996; Wang and Eccles 2012) and self-esteem

(Archambault et al. 2010), as well as increases in anxiety,

depression, and behavioral difficulties (Grills-Taquechel

et al. 2010; Way et al. 2007). These struggles may be

magnified when students are undergoing a transition to a

new school during the early adolescent period (Eccles

2004; Eccles and Midgley 1989; Seidman et al. 2004).

Ecological theories and empirical research have high-

lighted the importance of supportive developmental social

settings, such as family, classrooms, schools, and neigh-

borhood community contexts, during early adolescence

(Bronfenbrenner and Morris 2006; Cappella et al. 2013;

Delany-Brumsey et al. 2014; Eccles and Midgley 1989;

McCoy et al. 2013; Smith et al. 2013). In particular,

researchers have focused on the central role of school

social contexts in promoting positive youth adjustment and

development (Centers for Disease Control and Prevention

2009; National Research Council 2004; Simmons and

Blyth 1987; Trickett and Rowe 2012). Increasingly, evi-

dence suggests that the middle grade school climate plays a

significant role in young adolescents’ academic and social-

emotional adjustment (Brand et al. 2003; Jia et al. 2009;

Way et al. 2007). Despite growing interest in the social

context of middle grade schools, and evidence for its

influence on student adjustment, current understanding of

H. Y. Kim (&) � K. Schwartz � E. Cappella � E. Seidman

Department of Applied Psychology, New York University, 246

Greene Street, New York, NY 10003, USA

e-mail: hayeon@nyu.edu

123

Am J Community Psychol (2014) 54:28–45

DOI 10.1007/s10464-014-9659-x

middle grade school social context is limited in scope.

Most studies focus on students’ perception of school cli-

mate (e.g., Brand et al. 2003; Way et al. 2007) or the

structural characteristics of middle grade schools, particu-

larly around grade span configuration (e.g., Benner and

Graham 2009; Kieffer 2013). Far less consideration has

been given to other perspectives or dimensions of the

school social setting, such as administrators’ perception of

student conduct and teachers’ reports of stress and pro-

fessional climate. In addition, to our knowledge, no

empirical literature has examined school social context

across a national sample with grade spans typical of US

public middle grade schools (e.g., k–8 schools, middle

schools, junior high schools).1

In this article, we aim to develop a multi-reporter

understanding of the middle grade social context across

these typical school grade spans. In addition, we examine

the potential role of middle grade social context as a link

between school grade span and students’ perceptions of

their schools. Our ultimate goal is to provide theoretically-

and empirically-based evidence for structural and contex-

tual school reform and policy change toward best sup-

porting public middle grade school students’ academic and

social-emotional well-being.

School Climate and Grade Configurations in the Middle

Grades

Middle grade school context has been of increasing interest

among developmental and educational researchers. Envi-

ronment-stage fit theory (Eccles and Midgley 1989) sug-

gests certain types of middle grade schools may not be

supportive of early adolescents’ development. Testing such

assumptions, researchers have directly and indirectly

examined middle grade school social contexts, with the

majority of the literature examining students’ perceptions

of the school social climate (e.g., Eccles et al. 1993;

Seidman et al. 1994). Researchers have also examined a

broader range of school climate dimensions, including

interpersonal (e.g., teacher support, peer support), organi-

zational (e.g., safety, chaos), and instructional/academic

(e.g., student commitment to achievement, instructional

innovation) aspects that influence students’ adjustment in

multiple domains (Brand et al. 2003; Jia et al. 2009; Ku-

perminc et al. 2001; Rutter and Maughan 2002; Skinner

and Wellborn 1994). Across these studies, school climate is

often conceptualized and measured using different dimen-

sions, reflecting the difficulties of defining school climate

(Anderson 1982). Interpersonal and academic dimensions

of student-perceived school climate seem to be the most

common and consistent predictors of multiple domains of

student outcomes (Brand et al. 2003; Jia et al. 2009; Way

et al. 2007).

The current literature on school climate is incomplete.

First, with a few exceptions (e.g., Brand et al. 2003), most

studies using student perceptions involve a small number

of schools. Replication of these studies with a large

national sample of schools is necessary to generalize the

findings across the diversity of schools in the United States.

Second, these studies conduct analyses on the individual-

level, focusing on how individual student perceptions are

linked to student outcomes. Given that individual charac-

teristics influence perceptions of social phenomena

(Mitchell et al. 2010), school-level analyses of social

context are needed to adequately assess school social

context. Third, most studies rely on students’ reports of

school social climate, and do not include other participants

in these settings, such as teachers and administrators. This

is despite findings suggesting that teachers may be better

raters of school climate than students when comparing

social climate between (as opposed to within) schools

(Bryk et al. 2010; Nathanson et al. 2013). Lastly, despite

calls to improve school climate (National School Climate

Council 2007), we know relatively little about how the

broader school context relates to students’ perceptions of

school climate. Given evidence linking student perceptions

to outcomes, a better understanding of how school level

factors influence student perceptions could be used to aid

policy makers and educators in strategically improving

school climate in the middle grades.

Another major focus of research on middle grade school

contexts has been on structural aspects of middle grade

schools, with a particular focus on grade span configura-

tions. This line of research suggests that the transition to

middle grade schools, compared to continuing attendance

in k–8 schools, is negatively associated with student aca-

demic and social-emotional outcomes in both correlational

(e.g., Byrnes and Ruby 2007; Rockoff and Lockwood

2010; Weiss and Kipnes 2006) and causal examinations

(e.g., Kieffer 2013; Schwerdt and West 2013). For exam-

ple, Schwerdt and West (2013) find that students moving

from elementary to middle grade schools suffer a sharp

drop in academic achievement, experience an increase in

school absences, and are less likely to still be enrolled in

school by grade 10 as compared with students who do not

transition. There is no clear consensus on whether there are

differences in student outcomes between transitioning to

middle schools versus junior high schools. Some studies

1 These are the terms that are typically used in the middle grade

school literature. Generally, k–8 schools include schools that serve

pre-kindergarten or kindergarten to 8th grade; middle schools include

schools that primarily serve 6th, 7th, and 8th graders; and junior high

schools serve 7th, 8th, or 7th, 8th, and 9th graders. Some of these

schools may serve higher grade levels as well (e.g., schools serving

7th–12th grade). This study uses the term k–8, middle, and junior high

schools liberally, including those schools serving higher grades.

Am J Community Psychol (2014) 54:28–45 29

123

demonstrate that, for students who experience a school

transition in the middle years, a transition in 6th grade (to

middle schools) may be more detrimental than a transition

in 7th grade (to junior high schools: e.g., Cook et al. 2008;

Rockoff and Lockwood 2010). Others find transitions in

6th or 7th grade equally detrimental to student achievement

(e.g., Schwartz et al. 2011; Seidman et al. 1994). Thus far,

this research has been conducted with a relatively small

number of schools or schools in specific states or cities

(e.g., Rockoff and Lockwood 2010; Schwerdt and West

2013), precluding generalizations across districts and

regions.

Taken together, little, if any, research has directly

compared the social contexts of schools with different

grade configurations. Although studies seem to suggest k–8

schools may better support students’ needs than middle

grade schools (for a review, see Seidman et al. 2004), it is

unclear whether there are systematic differences in social

context between middle grade schools with varying grade

span configurations.

Social Contexts of Middle Grade Schools

The school setting, like other social settings, has a set of

psychological and institutional attributes that give it a

distinctive interpersonal context (Kuperminc et al. 1997).

Within a school building, each individual—including stu-

dents, administrators, and teachers—are participants in

social interactions, both dyadic and school-wide, with

students and adults in different roles with different inter-

ests. Each individual contributes to the overall school

social context as a member of the school community, and

also has a unique perspective on the school context. For

example, teachers’ perceptions of teaching burdens are

significantly associated with the overall school climate and

student outcomes (Jones et al. 2011; Thijs et al. 2008).

Despite this, the school climate literature has focused

predominantly on students’ perspectives, overlooking the

views and processes of other participants of school social

settings, such as teachers and administrators. Educational

and organizational health research suggests social pro-

cesses that do not directly involve students, for example

teacher–teacher and teacher–administrator relationships, as

well as teacher and administrator perceptions and attitudes,

contribute to overall school health or climate (Hoy and

Hannum 1997; Marks and Printy 2003; Sweetland and Hoy

2000). In addition, students’ perceptions of school climate

are influenced by individual experiences and relationships

(Baker 1999; Goldstein et al. 2008). Examining the rela-

tions between administrators’ and teachers’ perspectives

and student perceptions will provide a more complete

picture of middle grade school social contexts.

Key dimensions of school-wide social contexts that have

not been fully integrated in the school climate literature

include school disorder and stress. Specifically, chaotic

school environments with frequent and uncontrollable

interruptions and disturbances due to high levels of vio-

lence, teacher mobility, and racial tensions may be highly

stressful for students (Bellmore et al. 2012; Birnbaum et al.

2003; Grannis 1992; Wang and Gordon 2012). In addition,

school culture in which noncompliance and disruptive

behavior is prevalent not only contributes to negative

school climate but also affects students’ academic

achievement (Flannery et al. 2009; Mitchell et al. 2010;

Warren et al. 2003). Successful interventions focused on

positive management of school-wide student behaviors

have a positive influence on overall school climate (e.g.,

Bradshaw et al. 2009; Mitchell et al. 2010).

Teacher perceptions and experiences—such as staff

professional climate, agency, and teaching burdens—are

also important dimensions of school social context.

Teachers engage in social and professional interactions

with other staff in their unique school settings. There is

growing recognition that a positive professional climate

among staff is critical to positive school-wide climate and

successful school reforms (Edgerson et al. 2006; Rhodes

et al. 2009). Teachers’ feelings of satisfaction and control

have been positively associated with teaching and

instructional behaviors (Skaalvik and Skaalvik 2007;

Tschannen-Moran and Hoy 2001). Additionally, some

evidence suggests middle grade school teachers feel less

efficacious than elementary or high school teachers (Eccles

et al. 1993; Midgley et al. 1995). Moreover, few training

programs or certification requirements are geared toward

specializing in the learning style and needs of middle grade

children. Lastly, teachers’ perception of the barriers to or

burdens of teaching may play an important role in teachers’

daily interactions with students and other members of the

school community (Raver et al. 2009; Wasik et al. 2006).

In sum, school-level disorder and stress, as well as

teacher experiences, perceptions, and training contribute to

the social context of middle grade schools and should be

considered in combination as potential influences on stu-

dents’ middle grade school experiences. By comparing the

overall social context of middle grade schools and exam-

ining its association with students’ perception of school

climate, we may better determine which types of middle

school grade configurations are most supportive of early

adolescent students’ development.

Current Study

Informed by the current literature, this study examines

social contexts across middle grade schools with different

grade span configurations and the associations between

30 Am J Community Psychol (2014) 54:28–45

123

school-level social context and students’ perceptions of

school climate. By doing so, we aim to bridge the gap

between the school climate literature focused on students’

individual perspectives and the middle school transition

literature testing impacts of varying grade span configura-

tions. Incorporating the perspectives and experiences of

administrators and teachers, we focus on understanding

school-wide social contexts and processes as an overarch-

ing construct that collectively influences students’ percep-

tions of their schools. Specifically, utilizing the eighth

grade cross-sectional data from a large national dataset—

the Early Childhood Longitudinal Study, Kindergarten

Class of 1998–1999 (ECLS-K)—we (1) describe the factor

structure of middle grade school social context as reported

by administrators and teachers; (2) examine variation in

social context among middle grade schools with different

grade span configurations (k–8 schools; 6–8 middle

schools; 7–9 junior high schools), controlling for school

demographic and structural characteristics; and (3) test

middle grade school social context as a potential link

between school grade span configuration and students’

perceptions of school climate. Using a multiple indicator

multiple causes (MIMIC) approach with multi-level

structural equation models (MSEM), this study aims to

enhance understanding of school-level differences in social

contexts between different types of middle grade schools

and how these differences may influence the experiences

and development of students within those schools. By

doing so, this study seeks to inform the development of

more supportive, positive school environments for all

public school students.

Method

Sample

The data used in this study are drawn from the Early

Childhood Longitudinal Study, Kindergarten Class

1998–1999 (ECLS-K). The ECLS-K followed a nationally

representative sample of 21,260 kindergarteners from the

1998–1999 school year through the 2006–2007 school

year, at which time the majority of participants were

enrolled in eighth grade. It is a multi-method and multi-

source study that includes: interviews with parents, prin-

cipal and teacher surveys, student records, direct student

assessments, and student self-reports. For this study, the

sample consists of the 5,754 students from 1,712 schools

who participated in the 2007 spring (Wave 7) data col-

lection; and were: (1) in eighth grade; (2) attending a k–8

school, middle school, or junior high school; and (3)

attending a regular public school.

The data used in the present analysis were provided by

students, teachers, and school administrators. Students

were evenly split by gender (51 % female) with an average

eighth grade student age of 14 years. They are predomi-

nately white (62 %), followed by Hispanic (18 %), African

American (9 %), Asian (6 %), and other (5 %). Students

represent an even distribution of the range of socioeco-

nomic statuses (SES) with the exception of an underrep-

resentation of the lowest quintile (22 % are in the top

quintile followed by 20, 21, 20, and 15 %). Teacher par-

ticipants include 5,085 unique English (n = 2,778) and

math or science teachers (n = 3,307). The average teacher

completed a bachelor’s degree, had 14 years of teaching

(SD = 10.40), and was 44 years old (SD = 11.76) at this

wave of data collection. School administrator surveys were

reported by the principal (N = 1,712). On average, the

principals reported 12 years teaching (SD = 6.28), eight

years as a principal (SD = 6.54), and five years as a

principal at their current school (SD = 4.55). The majority

had a master’s degree or higher.

Measures

School Social Context: Administrator- and Teacher-

Reports

To examine the factor structures of middle school social

contexts, we identified 25 social context items from the

administrator-report (8 items) and teacher-report (17 items)

questionnaires (Table 2). These items were chosen from

within the constraints of the ECLS-K data with a focus on

capturing school-wide stressors as well as teachers’ expe-

riences in, and perceptions about, the school context.

In order to measure school-wide stressors, administra-

tors were asked the degree to which teacher turnover; gang

activities; and racial tensions are problems in their schools

(five-point scale: 1 = strongly disagree and 5 = strongly

agree). They also reported on the frequency of student

conduct problems in school, including class cutting, theft,

vandalism, and bullying on a (five-point Likert scale:

1 = happens daily and 5 = never happens).

Teachers’ experiences and perceptions were captured

using questions from the eighth grade teacher question-

naire. The first three questions asked teachers about their

own teaching experience (e.g., ‘‘I really enjoy my present

teaching job.’’). The second set of questions included 14

items asking teachers about their perceptions of the various

aspects of the school social context, including staff climate,

difficulties in teaching, and their own attitude toward

teaching (e.g., ‘‘Staff members in this school generally

have school spirit’’; ‘‘The level of misbehavior in this

school interferes with my teaching’’; ‘‘I feel that it’s part of

my responsibility to keep students from dropping out of

Am J Community Psychol (2014) 54:28–45 31

123

school’’). For all items, teachers responded on a five-point

Likert scale ranging from 1 = strongly disagree to

5 = strongly agree (see Table 2 for all items).

For this study, we calculated and used the mean teacher

response for each school. We also ran intra-class correla-

tions (ICCs) for all of the teacher items, in order to assess

how well items distinguish between schools as opposed to

between teachers within the same school. ICCs were suf-

ficiently large to confidently support interpreting teacher

responses as school level context variables for all but four

items where the ICC B 0.032 (see Table 2 notes for items).

Student Perceptions of School Climate

Thirteen items from the 8th grade student survey were used

to measure student perceptions of school climate, including

school attachment, peer academic values, and peer support.

Specifically, the five items on school attachment included:

how often students feel like they fit in; feel close to

classmates; feel close to teachers; enjoy being at school;

and feel safe at school. Responses were on a four-point

scale from 1 (never) to 4 (always). Students’ perception of

peers’ academic values were measured using three items:

how important is it to their close friends that they attend

classes regularly; get good grades; continue their education

past high school. The responses ranged from 1 (not

important) to 3 (very important). A questionnaire on stu-

dents’ perception of peer support among their classmates

consisted of five items that asked whether their classmates

think it is important to be their friend; like them the way

they are; care about their feelings; like them as much as

they like others; and really care about them (five-point

Likert scale: 1 = never to 5 = always).

Both exploratory and confirmatory factor models of

these 12 items revealed and confirmed the three construct

model—school attachment, peer academic values, peer

support—as the best fitting model (for details on factor

analysis procedure and results, see ‘‘Appendix 1’’).

School Structure

Based on our literature review, we hypothesized that the

grade span configurations of a school as well as key

demographics would influence the school context. In order

to both test and control for this, we included the following

variables in our analyses.

Grade Span Schools included in the ECLS-K dataset

have various school grade span configurations. For mean-

ingful comparison, school grade span was coded in three

categories based on transition timing: k–8, middle, and

junior high schools. Specifically, schools that started in

either pre-kindergarten or kindergarten and extended

through 8th or 12th grade were coded as k–8 schools

(11 %); schools beginning in 6th grade and extending

through 8th or 12th grade were coded as middle schools

(64 %); and lastly, schools beginning in 7th grade and

extending through 8th, 9th, or 12th grade were coded as

junior high schools (25 %).

Structure and Demographics Past work suggests that

school urbanicity, school size, student composition, and

facility quality may contribute to school social context

(Anderson 1982; Buckley et al. 2005; Hannaway and

Talbert 1993; Opdenakker and Damme 2007). We selected

one variable—whether the school is in an urban, suburban,

or rural location—as an indicator of school location urba-

nicity. For student composition, four variables were

selected: number of students enrolled, % Hispanic students,

% black students, and % students eligible for free lunch.

Facility quality was measured using a composite score of

Table 1 Descriptives of school structural and demographic variables

N %

Grade span

k–8 schools (PK/K – 8/12) 194 11

Middle schools (6 – 8/12) 1,087 63

Junior high schools (7 – 8/9/12) 431 25

School location urbanicity

Urban 562 33

Suburban 710 42

Rural 408 24

% Hispanic students

Less than 1 % 91 5

1 % to less than 5 % 533 33

5 % to less than 10 % 242 15

10 % to less than 25 % 310 19

25 % or more 447 28

% African American students

Less than 1 % 80 5

1 % to less than 5 % 606 37

5 % to less than 10 % 244 15

10 % to less than 25 % 316 20

25 % or more 375 23

N Mean SD Min Max

School size 1,700 783.53 341.13 100 5,000

% free lunch 1,712 34.56 26.05 0 95

Facility quality 1,689 60.81 28.92 0 100

2 We initially included these items in our analyses, eventually

dropping the final model which had low ICCs and did not load onto

the final factor structure model.

32 Am J Community Psychol (2014) 54:28–45

123

nine items reported by the administrator on the adequacy of

the following facilities: cafeteria, computer lab, library/

media center, art room, gymnasium, music room, play-

ground/school yard, classrooms, auditorium/multi-purpose

room. Responses were originally coded on a five-point

scale. However, due to the skewed distribution of the data

with the majority of respondents answering always ade-

quate for each question, the responses were dichotomized

at the item level (0 = do not have, never adequate, often

not adequate, and sometimes not adequate and 1 = always

adequate) Descriptive statistics for school structure and

demographic variables are listed in Table 1.

Student Covariates

We hypothesized that a number of student level demo-

graphics would influence students’ perceptions of school

social processes. Therefore, we included the following

covariates at the student level: gender; race/ethnicity;

socioeconomic status (SES); seventh grade self-perception

of math and reading ability; internalizing problems, and

locus of control.

Results

Prior to data analysis, we assessed missing information. For

the majority of variables included in the analysis, no more

than 3 % of schools had missing data. The exceptions were

% Hispanic (5 %), % African-American (5 %), % free

lunch status (9 %), and school urbanicity (10 %). Our

analysis, nevertheless, included data for all 1,712 schools.

School-level weights are not available for the schools

included in Wave 7 (8th grade) of the ECLS-K. In addition,

schools in this 7th wave are not representative of the initial

national school sample.

To address the research aims, we utilized a series of

single- and multi-level exploratory and confirmatory factor

analyses as well as multi-level structural equation models

(MSEM). All models were estimated using maximum

likelihood estimation with robust standard errors and a

mean-adjusted chi square statistic test (Asparouhov and

Muthen 2006) on Mplus 7.0 software (Muthen and Muthen

2012). In all SEM models, factor variance was fixed at 1

and all item factor loadings were freely estimated. As

suggested by Kline (2011), we evaluated the fit of all multi-

level confirmatory analysis and MSEM using multiple

indices of model fit, including both level-specific indices

such as standardized root mean square residuals for within-

model (SRMR-W) and between-level (SRMR-B); and

overall model fit indices such as the model chi square

statistic, the root mean square error of approximation

(RMSEA), and comparative fit index (CFI). There are no

clear guidelines for model fit evaluation of multilevel

models (Marsh et al. 2012). Following suggestions of

recent studies, we used a strict cut-off value for SRMR-

B B 0.06 to evaluate school-level models (Hsu 2009), and

liberal criteria for RMSEA (B0.08) and CFI (C0.80) given

the large ECLS-K sample size (Marsh et al. 2004a, b).

Question 1: Factor Structure of School Social Contexts

To examine the factor structure of school social contexts,

the first stage of analysis was to determine the best-fitting

measurement model. A series of preliminary exploratory

and confirmatory factor analyses were run. First, we used

40 % of the randomly-split school sample to conduct

exploratory factor analyses with administrator- and tea-

cher-report items, using a geomin rotation solution. We

extracted two factors for administrator-report items (school

chaos and student conduct problems) and four factors for

teacher-report items (teacher professional climate, agency,

teacher burden, commitment), considering the model fits

and related theories and prior studies on school social

context. Second, we confirmed these within-reporter factor

structures using the remaining 60 % of the sample. In the

next step, we combined teacher-report and administer-

report items and performed a confirmatory factor analysis

at the school level using the whole school sample. The

findings confirmed the six social context factors with a

reasonable fit (v2(237) = 1314.85, p \ .001; RMSEA =

0.03; CFI = 0.88; SRMR-B = 0.06). The six social con-

text factors were strongly correlated with each other in the

expected directions. Interestingly, the teacher commitment

factor was not correlated with these factors: administrators’

reports of school chaos and student conduct (see Fig. 1 for

details).

Although school social context comprises six different

factors, we expected the various dimensions to be related

and collectively influence student perceptions of school

social climate. Thus, we subjected the six factors to a

second-order factor analysis to see if these factors reflected

an overarching construct. The residual variances between

school chaos and student conduct problems were allowed

to correlate because these measures were completed by the

same source (administrator), and allowing the errors to

correlate significantly improved the model fit (Bentler

2000). The initial second order model with six first-order

factors had a significantly better fit than the first-order

factor model, as indicated by the Satorra-Bentler scaled chi

square difference test results (Dv2(8) = 255.82, p \ .001)

but was not fully satisfactory with SRMR-B = 0.07. In

addition, we found only a modest factor loading of teacher

commitment (0.32) on the overarching secondary factor.

To improve the model fit, we omitted the teacher com-

mitment factor and its items from the analysis. The

Am J Community Psychol (2014) 54:28–45 33

123

modified model fit the data well with 5 latent factors of

social context: school chaos, student conduct problems,

staff professional climate, teacher agency, and teaching

burden (v2(164) = 950.10, p \ .001; RMSEA = 0.03;

CFI = 0.90; SRMR-B = 0.06). Factor loadings of each

item are presented in Table 2. School chaos, student con-

duct problems, and teaching burden factors negatively

loaded on the overarching social context factor; and staff

professional climate and teacher agency factors had posi-

tive loadings on the overarching social context factor.

Question 2: Difference in School Social Context

by Grade Spans

To determine the extent to which the social context factor

varied among k–8 schools, middle schools, and junior high

schools we used a multiple indicators and multiple causes

(MIMIC) model. MIMIC models allow for simultaneous

factor analysis and regression of factor scores on covariates

in order to test for heterogeneity in factor means and mea-

surement invariance across groups (Joreskog and Goldber-

ger 1975; Muthen 1989). In order to compare three different

grade span configurations, we created two binary variables

that indicated school membership for k–8 schools and

middle schools with junior high schools as the reference

group. Then, we regressed the second-order school context

factor on these two binary variables. In addition, we inclu-

ded a host of school level covariates to control for school

demographic and structural characteristics such as urba-

nicity, school size, free lunch %, Hispanic and black student

%, and school facility quality. The results of this MIMIC

model are presented in Fig. 2. The best-fitting MIMIC

model provided a good fit to the data, v2(202) = 1,884.02,

p \ .001; CFI = 0.82; RMSEA = 0.03; SRMR-B = 0.06.

Further modifications did not improve the overall fit.

When school covariates were entered into the model, the

direction of loadings of the social context factors reversed.

School chaos, student conduct problems, and teaching

burdens had positive loadings on overall school context;

and staff professional climate and teacher satisfaction had

negative loadings. Thus, a higher score of overall school

social context indicates a more negative social context.

The SEM results (see Fig. 2) suggest that school social

context varies by school grade span configurations. Spe-

cifically, k–8 schools have a more positive social context

compared to middle schools and junior high schools con-

trolling for school demographic and structural character-

istics. There was no significant difference between middle

and junior high schools.

Question 3: Predicting Student Perceptions of School

Climate

To account for the nested structure of the data (students

nested in schools), we conducted two-level structural

equation modeling, which accounts for the dependency

among student reports within schools and allows school-

and student-level variables to be modeled distinctly. The

use of different reporters and methods to measure school

social context and student perceptions of school social

climate constructs reduces potential bias due to shared

reporters (Bank et al. 1990).

Prior to testing the hypothesized model, we examined an

unconditional model to decompose the amount of variance

that existed between student- and school-levels. In this

model, student reports of climate were allowed to be cor-

related at both student (r = 0.33–0.76, p \ .001) and

school levels (r = 0.20–0.23, p \ .001), with random

intercepts to consider mean level differences between

schools. Student reports of school climate in the uncondi-

tional model had low intraclass correlations (school

attachment = 0.03; peer support = 0.03; and peer aca-

demic values = 0.05), suggesting most of the variance in

student reports could be explained by individual differ-

ences within schools rather than between schools. Typi-

cally such low ICCs are considered indicators of

independency of individual scores from the cluster, and

used for a rationale for ignoring the nested structure.

However, given our research focus on between-school

differences in student outcomes, it remains important to

consider the nested structure of the data. In addition, given

the highly significant correlations and shared respondents,

error variance of the student-report of school climate

variables was allowed to covary for further analyses.

The next step was to include student-level covariates in

the student-level models to examine within-school associ-

ations. Students’ individual characteristics significantly

explained the within-school individual variations in per-

ceptions of school climate—33 % of school attachment

(p \ .001, 95 % CI [0.31, 0.35]), 29 % of peer support

(p \ .001, 95 % CI [0.13, 0.17]), and 15 % of peer aca-

demic values (p \ .001, 95 % CI [0.13, 0.17]).

Lastly, the hypothesized model was specified as shown in

Fig. 3. The bolded arrows indicate the hypothesized path

between school grade span configurations, school social con-

text, and student-report of school climate. School and student

covariates were included to control for differences between

schools and students. The hypothesized model provided a good

fit to the data, v2(413) = 1,906.26, p\ .001, CFI = 0.92,

RMSEA = 0.03, SRMR-W = 0.001, SRMR-B = 0.06. To

34 Am J Community Psychol (2014) 54:28–45

123

Ta

ble

2S

tan

dar

diz

edfa

cto

rlo

adin

gs

of

sch

oo

lso

cial

con

tex

tv

aria

ble

sfo

rth

ese

con

d-o

rder

con

firm

ato

ryfa

cto

ran

aly

sis

mo

del

Sec

on

dO

rder

Fac

tor:

Sch

oo

lS

oci

alC

on

tex

t

Fac

tor

1:

Fac

tor

2:

Fac

tor

3:

Fac

tor

4:

Fac

tor

5:

Sch

oo

lch

aos

Stu

den

tco

nd

uct

pro

ble

ms

Sta

ffp

rofe

ssio

nal

clim

ate

Tea

cher

agen

cyT

each

ing

bu

rden

sa

Fac

tor

load

ing

[95

%C

I]

-0

.46

[-0

.49

,-

0.4

3]

-0

.41

[-0

.44

,-

0.3

8]

0.6

8[0

.65

,0

.71

]0

.73

[0.7

0,

0.7

6]

-0

.92

[-0

.95

,-

0.8

9]

Mea

n(S

D)

Fac

tor

load

ing

(SE

)

Mea

n(S

D)

Fac

tor

load

ing

(SE

)

Mea

n(S

D)

Fac

tor

load

ing

(SE

)

Mea

n(S

D)

Fac

tor

load

ing

(SE

)

Mea

n(S

D)

Fac

tor

load

ing

(SE

)

Tea

cher

turn

ov

er4

.28

(0.9

0)

0.4

7(0

.03

)

Gan

gac

tiv

ity

3.5

7(1

.09

)0

.80

(0.0

2)

Rac

ial

ten

sio

n1

.79

(0.9

7)

0.6

7(0

.02

)

Cla

sscu

ttin

g3

.51

(1.1

0)

0.7

0(0

.02

)

Ph

ysi

cal

con

flic

t3

.36

(0.8

9)

0.7

6(0

.02

)

Th

eft

3.6

5(0

.69

)0

.67

(0.0

2)

Van

dal

ism

3.8

1(0

.65

)0

.64

(0.0

2)

Stu

den

tb

ull

yin

g2

.85

(0.9

9)

0.5

5(0

.02

)

Tea

cher

sin

this

sch

oo

lar

eco

nti

nu

ally

lear

nin

gan

dse

ekin

gn

ewid

eas

4.0

4(0

.53

)0

.68

(0.0

3)

Sta

ffm

emb

ers

inth

issc

ho

ol

gen

eral

lyh

ave

sch

oo

lsp

irit

3.8

0(0

.61

)0

.70

(0.0

3)

Ife

elac

cep

ted

and

resp

ecte

das

aco

llea

gu

eb

ym

ost

staf

fm

emb

ers

4.3

6(0

.46

)0

.61

(0.0

3)

Ire

ally

enjo

ym

yp

rese

nt

teac

hin

gjo

b4

.38

(0.5

4)

0.8

7(0

.02

)

Iam

cert

ain

Iam

mak

ing

ad

iffe

ren

cein

the

liv

eso

fth

est

ud

ents

Ite

ach

4.3

6(0

.43

)0

.61

(0.0

3)

IfI

cou

ldst

art

ov

er,

Iw

ou

ldch

oo

sete

ach

ing

agai

nas

my

care

er4

.12

(0.6

5)

0.6

1(0

.02

)

Th

ele

vel

of

mis

beh

avio

rin

this

sch

oo

lin

terf

eres

wit

hm

yte

ach

ing

2.6

7(0

.90

)0

.70

(0.0

2)

Man

yo

fth

est

ud

ents

Ite

ach

are

no

tca

pab

leo

fle

arn

ing

the

mat

eria

lI

amsu

pp

ose

dto

teac

hth

em2

.23

(0.7

1)

0.5

4(0

.02

)

Par

ents

are

sup

po

rtiv

eo

fsc

ho

ol

staf

f3

.50

(0.6

6)

-0

.61

(0.0

2)

Th

eat

titu

des

and

hab

its

stu

den

tsb

rin

gto

my

clas

sg

reat

lyre

du

ceth

eir

chan

ces

for

acad

emic

succ

ess

3.3

0(0

.71

)0

.65

(0.0

2)

My

succ

ess

or

fail

ure

inte

ach

ing

isd

ue

pri

mar

ily

tofa

cto

rsb

eyo

nd

my

con

tro

lra

ther

than

my

ow

nef

fort

or

abil

ity

2.5

4(0

.71

)0

.51

(0.0

3)

Ife

elso

met

imes

itis

aw

aste

of

my

tim

eto

try

tod

om

yb

est

aste

ach

er1

.72

(0.6

1)

0.4

9(0

.03

)

All

fact

or

load

ing

sw

ere

sig

nifi

can

tat

p\

.00

1

Fo

ur

item

s,in

clu

din

g‘‘

Ife

elth

atit

’sp

art

of

my

resp

on

sib

ilit

yto

kee

pst

ud

ents

fro

md

rop

pin

go

ut

of

sch

oo

l,’’

‘‘If

som

est

ud

ents

inm

ycl

ass

no

td

oin

gw

ell,

Ife

elth

atI

sho

uld

chan

ge

my

app

roac

hto

the

sub

ject

,’’

‘‘B

ytr

yin

gd

iffe

ren

tm

eth

od

,I

can

sig

nifi

can

tly

affe

cta

stu

den

t’s

ach

iev

emen

t,’’

and

‘‘I

wo

rkto

crea

tele

sso

ns

som

yst

ud

ents

wil

len

joy

lear

nin

gan

db

eco

me

ind

epen

den

tth

ink

ers’

’w

ere

fact

ore

din

to‘‘

teac

her

com

mit

men

t’’

fact

or—

this

fact

or

and

the

item

sw

ere

dro

pp

edd

ue

tolo

wfa

cto

rlo

adin

g(\

0.4

0)

on

the

seco

nd

-ord

ersc

ho

ol

soci

alco

nte

xt

fact

or.

Th

ese

item

sh

adv

ery

low

ICC

s(B

0.0

3),

sug

ges

tin

gth

atth

ese

item

sar

eb

ette

rm

easu

res

of

ind

ivid

ual

dif

fere

nce

san

dn

ot

of

sch

oo

lle

vel

dif

fere

nce

sa

On

eit

em,

‘‘R

ou

tin

ead

min

istr

ativ

ed

uti

esan

dp

aper

wo

rkin

terf

ere

wit

hm

yjo

bo

fte

ach

ing

,’’

was

dro

pp

edfr

om

the

fin

alC

FA

du

eto

the

low

fact

or

load

ing

(\0

.40

)o

nth

ete

ach

ing

bu

rden

s

fact

or

Am J Community Psychol (2014) 54:28–45 35

123

avoid possible model misidentification due to the low ICC, we

ran the model with twenty different starting values. Each

yielded the same converged model. A full model with the

standardized estimates of school level path coefficients is

illustrated in Fig. 3 and the total indirect and total effects of the

main variables, along with standardized coefficients of covar-

iates, are reported in Table 3. Coefficients of school- and stu-

dent-level covariates are reported in ‘‘Appendix 2’’.

There was no direct association between school grade

span and any of the student perceptions of school climate

variables, with social context in the model. However, the

quality of the school social context varied by school grade

configuration and negative school social context was itself

directly associated with lower levels of students’ school

attachment, peer support, and peer academic values. Spe-

cifically, k–8 schools had a more positive school climate

compared to junior high schools; and as well as to middle

schools (k–8 vs. middle school: Wald v2(1) = 5.91,

p \ .05). There was no significant difference between

middle school and junior high school.

We further investigated the significance of the simple

indirect effects of grade span on students’ perceptions of

school climate using a 2-2-1 multi-level mediation

approach (Bauer et al. 2006). Testing indirect effects with

clustered data using the MSEM method provides more

efficient and unbiased estimates compared to tests using

multi-level modeling (Preacher et al. 2011). Simple indirect

effects by social context of the effect of k–8 schools on

student school attachment (unstandardized estimate = 0.02;

standardized = 0.05, p \ .05, 95 % CI [0.003, 0.100]) and

peer academic values (unstandardized estimate = 0.01;

standardized = 0.03, p = .05, 95 % CI [-0.001, 0.068])

were marginally significant after controlling for school- and

child-level covariates. The indirect effect on peer support

was not significant (unstandardized estimate = 0.01; stan-

dardized = 0.04, p = .08, 95 % CI [-0.008, 0.084]).

However, peer support was highly correlated with school

attachment and peer academic values, and was linked to

more complex indirect paths (e.g., k–8 schools ? school

social context ? school attachment ? peer academic

values). The significance of these complex indirect effects

through two mediators cannot be tested. However, if all

paths are significant for a given a level, such complex

indirect effects can be accepted to be significant (Kline

2011). Total indirect effects by school social context of the

effect of k–8 schools on student-report school climate

variables are presented in Table 3. The school-level pre-

dictors included in the final model explained 30 % of the

variance in school attachment, 15 % in peer supports, 10 %

in peer academic values at the school level. The student-

level model explained 33 % of the variance in school

attachment; 30 % in peer support; and 15 % in peer aca-

demic values.

Discussion

Early adolescence is a time of significant developmental

change. Despite an increasing focus on supportive school

environments during these years as a means to foster

positive development, little is known about the school

context of middle grade schools. Similarly, there is little

research on whether school social context varies by grade

span configuration or relates to student perceptions and

experiences. In this study, we developed a measurement

model for school social context that incorporated teacher

and administrator perspectives in a large national sample.

We found that administrators and teachers in k–8 schools

reported a more positive school social context than their

colleagues in middle and junior high schools, controlling

for school demographic and structural characteristics. This

positive school social context, in turn, explained the asso-

ciation between grade span configuration and students’

perceptions of their schools’ social and academic climates.

These findings are consistent with research literature and

policy recommendations suggesting k–8 schools best serve

the academic and socio-emotional needs of middle grade

students (Seidman et al. 2004). Furthermore, these findings

suggest school social context is likely of critical impor-

tance to students’ perceptions of school climate, which in

turn play a central role in academic and social-emotional

development following the middle grade school transition

(Brand et al. 2003; Jia et al. 2009; Way et al. 2007).

Our measurement model of school social context identi-

fied five interrelated factors. Two of these factors—admin-

istrator reports of school chaos and conduct problems—

indicate the levels of disorder, such as racial tension, turn-

over, and student behavior problems, present in the school.

The remaining three factors—staff professional climate,

teacher agency, and teacher burden—reflect a range of

teachers’ perceptions, from teaching challenges and teaching

efficacy to social and professional interactions among school

staff. Collectively, these five factors form a single over-

arching factor that encompasses school social context, to

which increased chaos, conduct problems, and teacher bur-

den negatively contribute, while feelings of a supportive

professional climate and agency positively contribute.

Incorporating teacher and administrator perspectives in

the measurement of school social context provides a more

nuanced, multidimensional view than in prior work, which

has frequently relied on student perceptions. While student

perceptions are significant predictors of student achieve-

ment (Brand et al. 2003; Jia et al. 2009; Way et al. 2007),

they show greater differences between individuals than

they do differences between schools (Bryk et al. 2010;

Nathanson et al. 2013). Such individual-level variation

provides little information on overall school contexts and

makes it difficult to identify efficient targets of educational

36 Am J Community Psychol (2014) 54:28–45

123

interventions and policies. For example, students’ percep-

tions of low peer academic values are likely to be more

greatly influenced by their direct interactions with their

specific friend groups than their experience in the school as

a whole. Targeting individual-level social experience

requires identifying the primary contributor of individual

variance for each child and tailoring interventions

accordingly, which may be labor- and time-intensive as

well as difficult to implement on a large scale. In contrast,

if it were found that decreasing chaos in the school as a

whole has an average effect of raising everyone’s peer

academic values, this would present a more efficient, set-

ting-level, primary prevention solution. In this way, iden-

tifying more global measures of school context, may allow

us to identify better targets of policies and strategies to

improve middle grade school climate.

Using this measurement model, our findings suggest k–8

schools have a more positive school social context than

Fig. 1 School social context factor structure, standardized factor

loadings, and correlations between factors. School chaos and student

conduct (Conduct) factors were driven from school administrators’

reports; and professional climate (Prof. Climate), teacher agency (T

Agency), teacher burden (T Burden), and teacher commitment (T

Commit) factors were identified from teacher reports of school

context. Solid lines indicate significant correlations or factor loadings;

and dotted lines indicate non-significant correlations. ***p \ .001

Fig. 2 MIMIC model testing school social context varying by grade

configurations with standardized coefficients and factor loadings.

School structural and demographic characteristics were controlled for.

Solid lines indicate significant correlations or factor loadings; and

dotted lines indicate non-significant correlations. ***p \ .001;**p \ .01

Am J Community Psychol (2014) 54:28–45 37

123

middle and junior high schools, controlling for other

demographic and structural characteristics. We found no

significant difference in social context between middle and

junior high schools. These results parallel prior research

showing declines in academic achievement and engage-

ment among youths who transition to middle grade schools

as compared with those who do not (Byrnes and Ruby

2007; Cook et al. 2008; Rockoff and Lockwood 2010;

Seidman et al. 2004). This literature does not generally

distinguish between the middle and junior high school

transitions, an approach supported by our findings. The

lack of a significant difference between middle and junior

high schools, which are intended to be rooted in very dif-

ferent pedagogical approaches, is an interesting finding on

its own. One possible explanation is that the majority of

middle schools have failed to implement true middle

school pedagogy, instead utilizing distinct subject organi-

zation as defined by junior high school pedagogy (Wil-

liamson and Johnston 1999).

While most of the studies on the transition between

elementary and middle grade schools examine only the

relations between the act of transitioning or not and student

outcomes, the theoretical explanation given generally

focuses on the middle and junior high school setting.

Ecological systems theorists explain the declines in student

outcomes during the transition to middle grade schools as a

result of students experiencing less supportive social con-

texts in middle and junior high schools (Eccles et al. 1993;

Seidman et al. 2004; Trickett and Rowe 2012). This study

extends and supports this literature in two ways. First, our

findings confirm prior research showing more negative

student perceptions of school climate in middle and junior

Fig. 3 Multi-level structural equation model (MSEM) testing the role

of school social context with standardized coefficients of main paths

of interests. Solid lines indicate significant correlations or factor

loadings; and dotted lines indicate non-significant correlations. For

full results, including unstandardized and understandized coefficients

of covariates and factor loadings for school social context factors, see

Appendix 2. ***p \ .001; **p \ .01; *p B .05

Table 3 Unstandardized and standardized total indirect effects and total effects of k–8 schools and school social context predicting student-

report school attachment, peer support, and peer academic values

k–8 schools predicting School social context predicting

Total indirect Total Total indirect Total

Student-report school climate Estimate Std. estimate Estimate Std. estimate Estimate Std. estimate Estimate Std. estimate

School attachment 0.01 0.11 0.01 0.11 0.00 -0.62 -0.04 -1.15

Peer support 0.01 0.11 0.01 0.11 0.00 -0.74 -0.03 -0.42

Peer academic values 0.01 0.12 0.01 0.12 0.00 -0.83 -0.04 -1.17

38 Am J Community Psychol (2014) 54:28–45

123

high schools compared to elementary schools (Seidman

et al. 2004) through teacher- and administrator-reports of

school-level social context. Second, this study provides

empirical evidence in support of the theoretical explanation

that school social context may be a central mechanism

through which to explain differences in student experiences

by grade span configuration (Eccles et al. 1993; Seidman

et al. 2004). Indeed, once school social context was taken

into account, we found no independent association between

school grade span configuration and school attachment,

peer support, or peer academic values.

Overall, the findings from this study suggest a central role

for school social context in explaining the relations between

the middle grade schools’ grade span configuration and

student level experience and suggest support for existing

theories surrounding the mechanisms at work across this

transition (Eccles and Midgley 1989; Seidman et al. 2004).

Specifically, they indicate that it may not be the timing of the

middle grade transition or even the act of the transitioning

that leads to declines in performance and well-being, but

rather the environment into which youth transition. Struc-

tural changes may influence the resulting social context—for

example, through the presence/absence of younger students,

many versus one or two teachers, or the same/different

constellations of classmates. However, structural change

alone may not be sufficient in reversing the declining pat-

terns of student academic and social-emotional develop-

ment. Recent findings that older k–8 schools perform better

than newly-converted k–8 schools (Byrnes and Ruby 2007)

support this in suggesting schools may need additional

resources and strategies in order to best support adolescents’

needs and that the restructuring of grade configuration alone

may be insufficient. While further research is warranted as to

the factors underlying the relations between school grade

span configuration and school social context, our findings

suggest that targeting social context directly, with or without

shifts in grade span, may be an efficient way to support

positive student development during these years.

Study Strengths, Limitations, and Future Directions

This study has methodological and conceptual strengths.

First, it used a large national dataset with a range of public

schools represented, enabling between-school analyses on

social context and climate in the middle years. Second, it

included child, teacher, and administrator reports for a

nuanced conceptualization of school social context and an

examination of the association between school-level social

context and individual student perceptions. Third, it includes

an array of covariates at school and child levels to increase

our ability to draw valid conclusions from these findings.

Lastly, this study separated school context from students’

perceptions. This adds to our understanding of the processes

by which student perceptions of school climate might be

decreasing across the transition to middle or junior high

school.

The study is also limited in a few ways. First, the use of

secondary data, albeit from a rich national dataset, restricts

the variables at our disposal. We were unable to examine all

aspects of school social context identified in the literature,

such as principal leadership and principal–teacher relation-

ships (Rhodes et al. 2009). In addition, we were unable to

capture classroom (Cappella et al. 2013), afterschool pro-

grams (Smith et al. 2013), and neighborhood setting social

contexts (Delany-Brumsey et al. 2014; Duke et al. 2011;

McCoy et al. 2013) that may be critical to youths’ school

experiences and adjustment. Additional work with different

dimensions of school social context, and examination of

associations with other developmental settings, should be

undertaken to affirm and extend the findings. Second, due to

the unique data structure of the ECLS-K, we were unable to

consider teachers nested within schools. Replicating our

measurement model with multiple teachers and administra-

tors per school in a two-level model would help isolate

school social context from individual teachers’ or principals’

experiences. Third, due to the limited data points in the

ECLS-K during the middle grade years, this study examined

cross-sectional relations between school social contexts and

students’ perceptions of school climate for the middle grade

schools (6th, 7th, and 8th grade). While our findings suggest

a model in which school social context mediates the relation

between grade span configuration and student outcomes, the

directionality of such a model cannot be evaluated using this

cross-sectional and correlational study. Intervention

research that targets school social context and assesses the

impact on student perceptions would lend support and clarity

to the directionality of this association. Lastly, consistent

with past research indicating students vary more within

school than between schools (Bryk et al. 2010; Nathanson

et al. 2013), and given our low average number of students

per school, we found low intraclass correlations (ICCs) for

student responses. This may relate to the wording of items,

which asked about individual experiences rather than school

perceptions. Further research should draw from data with

more students per school in order to increase the between-

level association of student responses with questionnaires

designed to measure school-level processes.

Conclusion and Implications

This study identified school social context as a potentially

critical avenue of intervention toward supporting students’

social and academic development in the middle grades.

Our findings suggest that researchers and practitioners

should target school social context during these years and

Am J Community Psychol (2014) 54:28–45 39

123

point to avenues through which this might be accom-

plished. One avenue is structural change. Consistent with

past work, this study identified k–8 schools as more sup-

portive contexts for youth. A second avenue, not incom-

patible with the first, is implementation of policies directly

targeting key aspects of social context. This could take the

form of increasing relational support among teachers,

administrators, and parents (Jia et al. 2009); implementing

interventions to reduce conduct problems (Bradshaw et al.

2009); reducing teacher burden through streamlined

administrative demands; or otherwise improving profes-

sional climate. Further research as to the characteristics of

k–8 schools most critical to their positive social context

would greatly aid this intervention work through informing

which areas of social context to target and how to target

these areas. Whether targeting social context along with, or

in the absence of, structural change, our findings indicate

that in order to improve student perceptions of social and

academic climate and, through those, student outcomes,

educators may need to focus on improving teacher and

administrative social processes as well as supporting indi-

vidual students.

School social context is critical to student learning and

well-being. It may also help to explain why many youth

experience declines in achievement and adjustment across

the transition to middle or junior high school. Targeting

school social context may be a particularly effective way of

addressing these declines while, preventatively, benefiting

a large number of youth.

Acknowledgments This research was conducted with support

from Spencer Foundation (#201300077, PI: Elise Cappella, Co-PI:

Edward Seidman), the NYU Institute of Human Development and

Social Change, and the NYU Predoctoral Interdisciplinary Research

Training fellowship. All procedures and restricted data use were

approved by the IES Data Security Office (#12040005) and the

New York University Committee on Activities Involving Human

Subjects.

Appendix 1: Student Perceptions of Social

and Academic Climate Exploratory and Confirmatory

Factor Model Testing

Procedures

Student perceptions of school climate were examined using

13 items from the Wave 7 ECLS-K student questionnaires.

First, we used student reports from 40 % of the randomly

selected schools to conduct an exploratory factor analysis

using a geomin extraction approach with varimax rotation.

We found a three-factor model to be the most reasonable

solution considering the eigen value ([1), scree plot, and

model fit (v2(42) = 0.422, p \ .001, RMSEA = 0.06:

further details available from primary author). Second, we

confirmed the factor structure using the rest of the student

sample from the remaining 60 % of the schools. In these

analyses, we fixed the factor loadings to 1 and allowed the

factor loadings to be freely estimated. Using a combination

of theory, face validity, and modification indices, we then

Fig. 4 Student perceptions of school social and academic climate. v2(59) = 293.87, p \ .001 RMSEA = 0.03; CFI = 0.97; TLI = 0.97;

SRMR = 0.03

40 Am J Community Psychol (2014) 54:28–45

123

modified the measurement model until the best fitting

measurement model was obtained. In this process, we

allowed the residual variances of some of the items within

factors to correlate. These measures are completed by the

same source (student), and allowing the errors to correlate

significantly improved the model fit (Bentler 2000). Lastly,

we applied the same factor model to the whole student

sample. Factor loadings and fit statistics for the final

measurement model are presented in Fig. 4, below.

Appendix 2

See Table 4.

Table 4 Multi-level structural equation model (MSEM) testing the role of school social context with unstandardized and standardized coef-

ficients, and factor loadings

Unstandardized estimates Standardized estimates

Coefficient 95 % CI Coefficient 95 % CI

Within level

School attachment on

Gender 0.28*** [0.23, 0.33] 0.16 [0.13,0.18]

White -0.08* [-0.16, 0.00] -0.04 [-0.09, 0.00]

Black -0.19*** [-0.29, -0.09] -0.06 [-0.09, -0.03]

Hispanic -0.02 [-0.13, 0.09] -0.01 [-0.04, 0.03]

Asian -0.03 [-0.15, 0.09] -0.01 [-0.03, 0.02]

Other -0.07 [-0.18, 0.04] -0.02 [-0.04, 0.01]

Age 0.01* [0.01, 0.01] 0.03 [0.00, 0.05]

SES 0.00 [-0.02, 0.02] 0.00 [-0.03, 0.03]

Reading interest/competence 0.02 [-0.01, 0.05] 0.01 [-0.02, 0.04]

Math interest/competence 0.09*** [0.06, 0.12] 0.09 [0.06, 0.12]

Internalizing problems -0.12*** [-0.17, -0.07] -0.07 [-0.10, -0.04]

Locus of control 0.04 [-0.01, 0.09] 0.03 [-0.01, 0.06]

Self-concept 0.63*** [0.58, 0.68] 0.49 [0.45, 0.52]

Peer academic values on

Gender 0.20*** [0.15, 0.24] 0.12 [0.09, 0.15]

White 0.01 [-0.08, 0.09] 0.00 [-0.05, 0.06]

Black 0.13* [0.02, 0.24] 0.04 [0.01, 0.08]

Hispanic -0.02 [-0.15, 0.11] -0.01 [-0.05, 0.04]

Asian 0.19** [0.07, 0.30] 0.05 [0.02, 0.08]

Other 0.01 [-0.12, 0.14] 0.00 [-0.03, 0.04]

Age 0.00 [0.00, 0.01] 0.02 [-0.01, 0.05]

SES 0.03*** [0.01, 0.05] 0.05 [0.02, 0.08]

Reading interest/competence 0.10*** [0.07, 0.14] 0.09 [0.06, 0.12]

Math interest/competence 0.08*** [0.06, 0.11] 0.09 [0.06, 0.12]

Internalizing problems 0.12*** [0.07, 0.17] 0.08 [0.05, 0.11]

Locus of control 0.07** [0.02, 0.12] 0.05 [0.02, 0.09]

Self-concept 0.30*** [0.25, 0.34] 0.25 [0.21, 0.29]

Peer support on

Gender 0.33*** [0.28, 0.38] 0.18 [0.15, 0.20]

White -0.10* [-0.19, -0.02] -0.05 [-0.10, -0.01]

Black -0.19*** [-0.31, -0.07] -0.06 [-0.09, -0.02]

Hispanic -0.03 [-0.15, 0.09] -0.01 [-0.04, 0.03]

Asian 0.00 [-0.13, 0.13] 0.00 [-0.03, 0.03]

Other -0.04 [-0.16, 0.08] -0.01 [-0.04, 0.02]

Age 0.01*** [0.00, 0.01] 0.04 [0.02, 0.06]

SES 0.01 [-0.01, 0.03] 0.01 [-0.02, 0.04]

Reading interest/competence 0.00 [-0.04, 0.03] 0.00 [-0.03, 0.03]

Am J Community Psychol (2014) 54:28–45 41

123

Table 4 continued

Unstandardized estimates Standardized estimates

Coefficient 95 % CI Coefficient 95 % CI

Math interest/competence 0.07*** [0.04, 0.10] 0.07 [0.04, 0.09]

Internalizing problems -0.08** [-0.13, -0.02] -0.04 [-0.08, -0.01]

Locus of control 0.04 [-0.01, 0.09] 0.03 [-0.01, 0.06]

Self-concept 0.64*** [0.59, 0.69] 0.47 [0.44, 0.51]

School attachment with

Peer academic values 0.19*** [0.17, 0.21] 0.35 [0.32, 0.38]

Peer support 0.51*** [0.48, 0.53] 0.89 [0.88, 0.90]

Peer academic values with

Peer support 0.19*** [0.17, 0.21] 0.33 [0.30, 0.36]

Between level

School chaos by

Racial tension 0.49*** [0.43, 0.55] 0.66 [0.61, 0.71]

Gang activity 0.67*** [0.60, 0.73] 0.80 [0.76, 0.85]

Teacher turnover 0.40*** [0.34, 0.45] 0.48 [0.42, 0.54]

Student conduct by

Theft 0.41*** [0.37, 0.45] 0.67 [0.62, 0.71]

Class cutting 0.68*** [0.63, 0.73] 0.70 [0.66, 0.74]

Physical conflict 0.59*** [0.54, 0.63] 0.75 [0.71, 0.78]

Vandalism 0.36*** [0.31, 0.40] 0.63 [0.58, 0.68]

Student bullying 0.47*** [0.42, 0.52] 0.53 [0.49, 0.58]

Teaching burden by

Misbehavior interferes -0.20*** [-0.30, -0.11] -0.72 [-0.75, -0.68]

Students not capable -0.12*** [-0.18, -0.06] -0.54 [-0.59, -0.49]

Parents support 0.13*** [0.07, 0.19] 0.64 [0.59, 0.68]

Student attitudes -0.14*** [-0.21, -0.07] -0.62 [-0.67, -0.58]

Factors beyond control -0.11*** [-0.17, -0.06] -0.50 [-0.55, -0.44]

Waste of time -0.09*** [-0.14, -0.05] -0.48 [-0.54, -0.41]

Teacher agency by

I am making difference -0.20*** [-0.23, -0.16] -0.61 [-0.67, -0.55]

Choose teaching again -0.30*** [-0.35, -0.26] -0.62 [-0.67, -0.56]

Enjoy teaching -0.35*** [-0.40, -0.31] -0.87 [-0.92, -0.82]

Professional climate by

Teachers continue to learn -0.27*** [-0.31, -0.22] -0.67 [-0.73, -0.60]

School spirit -0.34*** [-0.37, -0.30] -0.72 [-0.79, -0.65]

Accept me -0.20*** [-0.24, -0.17] -0.59 [-0.66, -0.52]

School social context by

Teaching burden -2.56*** [-3.94, -1.17] -0.95 [-1.00, -0.90]

School chaos 0.55*** [0.45, 0.66] 0.54 [0.47, 0.62]

Student conduct 0.44*** [0.36, 0.51] 0.45 [0.38, 0.52]

Teacher agency 0.73*** [0.59, 0.87] 0.65 [0.58, 0.72]

Professional climate 0.72*** [0.58, 0.85] 0.64 [0.58, 0.71]

School social context on

k–8 -0.36** [-0.62, -0.10] -0.10 [-0.17, -0.03]

Middle -0.07 [-0.22, 0.08] -0.03 [-0.09, 0.03]

Urban 0.13 [-0.04, 0.29] 0.05 [-0.01, 0.11]

Rural 0.37*** [0.17, 0.56] 0.11 [0.05, 0.17]

Enrollment 0.10** [0.02, 0.18] 0.09 [0.02, 0.15]

42 Am J Community Psychol (2014) 54:28–45

123

References

Anderson, C. S. (1982). The search for school climate: A review of

the research. Review of Educational Research, 52, 368–420.

doi:10.3102/00346543052003368.

Archambault, I., Eccles, J. S., & Vida, M. N. (2010). Ability self-

concepts and subjective value in literacy: Joint trajectories from

grades 1 through 12. Journal of Educational Psychology, 102,

804–816. doi:10.1037/a0021075.

Archambault, I., Janosz, M., Morizot, J., & Pagani, L. (2009).

Adolescent behavioral, affective, and cognitive engagement in

school: Relationship to dropout. Journal of School Health, 79,

408–415. doi:10.1111/j.1746-1561.2009.00428.x.

Asparouhov, T., & Muthen, B. (2006). Robust Chi square difference

testing with mean and variance adjusted test statistics. Retrieved

October 23, 2013, from http://statmodel2.com/download/web

notes/webnote10.pdf.

Baker, J. A. (1999). Teacher–student interaction in urban at-risk

classrooms: Differential behavior, relationship quality, and

student satisfaction with school. The Elementary School Journal,

100, 57–70. doi:10.1086/461943.

Bank, L., Dishion, T. J., Skinner, M., & Patterson, G. R. (1990).

Method variance in structural equation modeling: Living with

‘‘glop.’’ In G. R. Patterson (Ed.), Depression and aggression in

family interaction (pp. 247–279). Hillsdale, NJ: Lawrence

Erlbaum Associates, Inc.

Barber, B. K., & Olsen, J. A. (2004). Assessing the transitions to

middle and high school. Journal of Adolescent Research, 19,

3–30. doi:10.1177/0743558403258113.

Bauer, D. J., Preacher, K. J., & Gil, K. M. (2006). Conceptualizing and

testing random indirect effects and moderated mediation in multi-

level models: New procedures and recommendations. Psychological

Methods, 11, 142–163. doi:10.1037/1082-989X.11.2.142.

Bellmore, A., Nishina, A., You, J., & Ma, T.-L. (2012). School

context protective factors against peer ethnic discrimination

across the high school years. American Journal of Community

Psychology, 49, 98–111. doi:10.1007/s10464-011-9443-0.

Benner, A. D., & Graham, S. (2009). The transition to high school as a

developmental process among multiethnic urban youth. Child

Development, 80, 356–376. doi:10.1111/j.1467-8624.2009.01265.x.

Bentler, P. M. (2000). Rites, wrongs, and gold in model testing.

Structural Equation Modeling: A Multidisciplinary Journal, 7,

82–91. doi:10.1207/S15328007SEM0701_04.

Birnbaum, A. S., Lytle, L. A., Hannan, P. J., Murray, D. M., Perry, C.

L., & Forster, J. L. (2003). School functioning and violent

behavior among young adolescents: A contextual analysis.

Health Education Research, 18, 389–403. doi:10.1093/her/

cyf036.

Table 4 continued

Unstandardized estimates Standardized estimates

Coefficient 95 % CI Coefficient 95 % CI

Free lunch % 0.02*** [0.01, 0.02] 0.40 [0.33, 0.47]

Hispanic % 0.05 [-0.02, 0.11] 0.05 [-0.02, 0.12]

Black % 0.16*** [0.10, 0.22] 0.18 [0.11, 0.24]

Facility quality 0.00 [0.00, 0.00] -0.05 [-0.10, 0.01]

School attachment on

School social context -0.04*** [-0.07, -0.02] -0.53 [-0.88, -0.17]

Peer academic values on

School social context -0.04** [-0.07, -0.01] -0.34 [-0.59, -0.08]

Peer support on

School social context -0.03* [-0.06, 0.00] -0.39 [-0.79, 0.01]

School attachment on

k–8 0.01 [-0.07, 0.09] 0.04 [-0.23, 0.31]

Middle 0.00 [-0.05, 0.06] 0.02 [-0.24, 0.28]

Peer academic values on

k–8 -0.05 [-0.14, 0.04] -0.12 [-0.34, 0.10]

Middle 0.02 [-0.03, 0.08] 0.09 [-0.12, 0.29]

Peer support on

k–8 -0.05 [-0.14, 0.04] -0.18 [-0.52, 0.16]

Middle -0.01 [-0.07, 0.04] -0.07 [-0.37, 0.22]

Student conduct with

School chaos 0.53*** [0.44, 0.61] 0.53 [0.44, 0.61]

Peer academic values with

School attachment 0.01* [0.00, 0.02] 0.88 [0.19, 1.57]

Peer support with

School attachment 0.01 [0.00, 0.01] 0.81 [0.53, 1.09]

Peer academic values 0.01* [0.00, 0.02] 0.92 [0.08, 1.76]

* p \ .05; **p \ .01; ***p \ .001

Am J Community Psychol (2014) 54:28–45 43

123

Bradshaw, C., Koth, C., Thornton, L., & Leaf, P. (2009). Altering

school climate through school-wide positive behavioral inter-

ventions and supports: Findings from a group-randomized

effectiveness trial. Prevention Science, 10, 100–115. doi:10.

1007/s11121-008-0114-9.

Brand, S., Felner, R., Shim, M., Seitsinger, A., & Dumas, T. (2003).

Middle school improvement and reform: Development and

validation of a school-level assessment of climate, cultural

pluralism, and school safety. Journal of Educational Psychology,

95, 570–588. doi:10.1037/0022-0663.95.3.570.

Bronfenbrenner, U., & Morris, P. A. (2006). The bioecological model of

human development. In R. M. Lerner & W. Damon (Eds.), Handbook

of child psychology (Vol. 1, pp. 793–828). Hoboken, NJ: Wiley.

Bryk, A., Sebring, P., Allensworth, E., Luppescu, S., & Easton, J.

(2010). Organizing schools for improvement: Lessons from

Chicago. Chicago: University of Chicago Press.

Buckley, J., Schneider, M., & Shang, Y. (2005). Fix it and they might

stay: School facility quality and teacher retention in Washington,

DC. Teachers College Record, 107, 1107–1123.

Byrnes, V., & Ruby, A. (2007). Comparing achievement between K-8

and middle Schools: A large-scale empirical study. American

Journal of Education, 114, 101–135.

Cappella, E., Kim, H. Y., Neal, J. W., & Jackson, D. R. (2013).

Classroom peer relationships and behavioral engagement in

elementary school: The role of social network equity. American

Journal of Community Psychology, 52, 367–379. doi:10.1007/

s10464-013-9603-5.

Centers for Disease Control and Prevention. (2009). School connect-

edness: Strategies for increasing protective factors among youth.

Atlanta, GA: US Department of Health and Human Services.

Cook, P. J., MacCoun, R., Muschkin, C., & Vigdor, J. (2008). The

negative impacts of starting middle school in sixth grade.

Journal of Policy Analysis and Management, 27, 104–121.

doi:10.1002/pam.20309.

Delany-Brumsey, A., Mays, V., & Cochran, S. (2014). Does

neighborhood social capital buffer the effects of maternal

depression on adolescent behavior problems? Journal of Com-

munity Psychology,. doi:10.1007/s10464-014-9640-8.

Duke, N. N., Borowsky, I. W., & Pettingell, S. L. (2011). Adult

perceptions of neighborhood: Links to youth engagement. Youth

& Society, 44, 408–430. doi:10.1177/0044118X11402852.

Eccles, J. S. (2004). Schools, academic motivation, and stage-

environment fit. In R. M. Lerner & L. D. Steinberg (Eds.),

Handbook of adolescent psychology (2nd ed., pp. 125–153).

Hoboken, NJ: Wiley.

Eccles, J. S., & Midgley, C. (1989). Stage-environment fit: Devel-

opmentally appropriate classrooms for young adolescents.

Research on Motivation in Education, 3, 139–186.

Eccles, J. S., Wigfield, A., Midgley, C., Reuman, D., Iver, D. M., &

Feldlaufer, H. (1993). Negative effects of traditional middle

schools on students’ motivation. The Elementary School Journal,

93, 553–574. doi:10.1086/461740.

Edgerson, D. E., Kritsonis, W. A., & Herrington, D. (2006). Analysis

of the influence of principal–teacher relationships on student

academic achievement: A national focus. National Journal for

Publishing and Mentoring Doctoral Student Research, 1, 1–5.

Flannery, K., Sugai, G., & Anderson, C. (2009). School-wide positive

behavior support in high school. Journal of Positive Behavior

Interventions, 11, 177.

Goldstein, S. E., Young, A., & Boyd, C. (2008). Relational aggression

at school: Associations with school safety and social climate.

Journal of Youth and Adolescence, 37(6), 641–654. doi:10.1007/

s10964-007-9192-4.

Grannis, J. C. (1992). Students’ stress, distress, and achievement in an

urban intermediate school. The Journal of Early Adolescence,

12, 4–27. doi:10.1177/0272431692012001001.

Grills-Taquechel, A. E., Norton, P., & Ollendick, T. H. (2010). A

longitudinal examination of factors predicting anxiety during the

transition to middle school. Anxiety, Stress & Coping, 23,

493–513. doi:10.1080/10615800903494127.

Hannaway, J., & Talbert, J. E. (1993). Bringing context into effective

schools research: Urban-suburban differences. Educational

Administration Quarterly, 29, 164–186. doi:10.1177/

0013161X93029002004.

Hoy, W. K., & Hannum, J. W. (1997). Middle school climate: An

empirical assessment of organizational health and student

achievement. Educational Administration Quarterly, 33,

290–311. doi:10.1177/0013161X97033003003.

Hsu, H.-Y. (2009). Testing the effectiveness of various commonly

used fit indices for detecting misspecifications in multilevel

structure equation models (Doctoral dissertation). Retrieved

from http://hdl.handle.net/1969.1/ETD-TAMU-2009-12-7411.

Jia, Y., Way, N., Ling, G., Yoshikawa, H., Chen, X., Hughes, D.,

et al. (2009). The influence of student perceptions of school

climate on socioemotional and academic adjustment: A com-

parison of Chinese and American adolescents. Child Develop-

ment, 80, 1514–1530. doi:10.1111/j.1467-8624.2009.01348.x.

Jones, S. M., Brown, J. L., & Aber, J. L. (2011). Two-year impacts of a

universal school-based social-emotional and literacy intervention:

An experiment in translational developmental research. Child

Development, 82, 533–554. doi:10.1111/j.1467-8624.2010.01560.x.

Joreskog, K. G., & Goldberger, A. S. (1975). Estimation of a model

with multiple indicators and multiple causes of a single latent

variable. Journal of the American Statistical Association, 70,

631–639. doi:10.1080/01621459.1975.10482485.

Kieffer, M. J. (2013). Development of reading and mathematics skills

in early adolescence: Do K-8 public schools make a difference?

Journal of Research on Educational Effectiveness, 6, 361–379.

doi:10.1080/19345747.2013.822954.

Kline, R. B. (2011). Principles and practice of structural equation

modeling. New York: Guilford press.

Kuperminc, G., Leadbeater, B., & Blatt, S. (2001). School social

climate and individual differences in vulnerability to psychopa-

thology among middle school students. Journal of School

Psychology, 39, 141–159. doi:10.1016/S0022-4405(01)00059-0.

Kuperminc, G. P., Leadbeater, B. J., Emmons, C., & Blatt, S. J.

(1997). Perceived school climate and difficulties in the social

adjustment of middle school students. Applied Developmental

Science, 1, 76–88. doi:10.1207/s1532480xads0102_2.

Maehr, M. L., & Midgley, C. (1996). Transforming school cultures.

Boulder, CO: Westview Press.

Marks, H. M., & Printy, S. M. (2003). Principal leadership and school

performance: An integration of transformational and instruc-

tional leadership. Educational Administration Quarterly, 39,

370–397. doi:10.1177/0013161X03253412.

Marsh, H. W., Hau, K.-T., & Wen, Z. (2004a). In search of golden

rules: Comment on hypothesis-testing approaches to setting

cutoff values for fit indexes and dangers in overgeneralizing Hu

and Bentler’s (1999) findings. Structural Equation Modeling: AMultidisciplinary Journal, 11, 320–341. doi:10.1207/

s15328007sem1103_2.

Marsh, H. W., Ludtke, O., Nagengast, B., Trautwein, U., Morin, A.

J. S., Abduljabbar, A. S., et al. (2012). Classroom climate and

contextual effects: Conceptual and methodological issues in the

evaluation of group-level effects. Educational Psychologist, 47,

106–124. doi:10.1080/00461520.2012.670488.

Marsh, H. W., Wen, Z., & Hau, K.-T. (2004b). Structural equation

models of latent interactions: Evaluation of alternative estima-

tion strategies and indicator construction. Psychological Meth-

ods, 9, 275–300. doi:10.1037/1082-989X.9.3.275.

McCoy, D. C., Roy, A. L., & Sirkman, G. M. (2013). Neighborhood

crime and school climate as predictors of elementary school

44 Am J Community Psychol (2014) 54:28–45

123

academic quality: A cross-lagged panel analysis. American

Journal of Community Psychology, 52, 128–140. doi:10.1007/

s10464-013-9583-5.

Midgley, C., Anderman, E., & Hicks, L. (1995). Differences between

elementary and middle school teachers and students: A goal

theory approach. The Journal of Early Adolescence, 15, 90–113.

doi:10.1177/0272431695015001006.

Mitchell, M. M., Bradshaw, C. P., & Leaf, P. J. (2010). Student and

teacher perceptions of school climate: A multilevel exploration

of patterns of discrepancy. Journal of School Health, 80,

271–279. doi:10.1111/j.1746-1561.2010.00501.x.

Muthen, B. O. (1989). Latent variable modeling in heterogeneous

populations. Psychometrika, 54, 557–585. doi:10.1007/

BF02296397.

Muthen, L. K., & Muthen, B. O. (1998–2012). Mplus user’s guide

(7th ed.). Los Angeles, CA: Muthen & Muthen.

Nathanson, L., McCormick, M., & Kemple, J. (2013). Strengthening

assessments of school climate: Lessons from the NYC school survey.

New York: The Research Alliance for New York City Schools.

Retrieved from http://files.eric.ed.gov/fulltext/ED543180.pdf.

National Research Council. (2004). Engaging schools: Fostering high

school students’ motivation to learn. Washington, DC: National

Academies Press. Retrieved from http://www.loc.gov/catdir/toc/

ecip047/2003017626.html.

National School Climate Council. (2007). The school climate

challenge: Narrowing the gap between school climate research

and school climate policy, practice guidelines and teacher

education policy. Retrieved from http://nscc.csee.net/.

Opdenakker, M., & Damme, J. V. (2007). Do school context, student

composition and school leadership affect school practice and

outcomes in secondary education? British Educational Research

Journal, 33, 179–206. doi:10.1080/01411920701208233.

Preacher, K. J., Zhang, Z., & Zyphur, M. J. (2011). Alternative

methods for assessing mediation in multilevel data: The

advantages of multilevel SEM. Structural Equation Modeling:

A Multidisciplinary Journal, 18, 161–182. doi:10.1080/

10705511.2011.557329.

Raver, C. C., Jones, S. M., Li-Grining, C., Zhai, F., Metzger, M. W.,

& Solomon, B. (2009). Targeting children’s behavior problems

in preschool classrooms: A cluster-randomized controlled trial.

Journal of Consulting and Clinical Psychology, 77, 302–316.

doi:10.1037/a0015302.

Rhodes, J. E., Camic, P. M., Milburn, M., & Lowe, S. R. (2009).

Improving middle school climate through teacher-centered

change. Journal of Community Psychology, 37, 711–724.

doi:10.1002/jcop.20326.

Rockoff, J. E., & Lockwood, B. B. (2010). Stuck in the middle:

Impacts of grade configuration in public schools. Journal of

Public Economics, 94, 1051–1061. doi:10.1016/j.jpubeco.2010.

06.017.

Rutter, M., & Maughan, B. (2002). School effectiveness findings

1979–2002. Journal of School Psychology, 40, 451–475. doi:10.

1016/S0022-4405(02)00124-3.

Schwartz, A. E., Stiefel, L., Rubenstein, R., & Zabel, J. (2011). The

path not taken: How does school organization affect eighth-

grade achievement? Educational Evaluation and Policy Ana-

lysis, 33, 293–317. doi:10.3102/0162373711407062.

Schwerdt, G., & West, M. R. (2013). The impact of alternative grade

configurations on student outcomes through middle and high

school. Journal of Public Economics, 97, 308–326. doi:10.1016/

j.jpubeco.2012.10.002.

Seidman, E., Aber, J. L., & French, S. E. (2004). The organization of

schooling and adolescent development. In K. I. Maton, C.

J. Schellenbach, B. J. Leadbeater, & A. L. Solarz (Eds.),

Investing in children, youth, families, and communities:

Strengths-based research and policy (pp. 233–250). Washington,

DC: American Psychological Association.

Seidman, E., Allen, L., Aber, J. L., Mitchell, C., & Feinman, J.

(1994). The impact of school transitions in early adolescence on

the self-system and perceived social context of poor urban youth.

Child Development, 65, 507–522. doi:10.2307/1131399.

Simmons, R. G., & Blyth, D. A. (1987). Moving into adolescence:

The impact of pubertal chance and school context. New York:

Aldine Transaction.

Skaalvik, E. M., & Skaalvik, S. (2007). Dimensions of teacher self-

efficacy and relations with strain factors, perceived collective

teacher efficacy, and teacher burnout. Journal of Educational

Psychology, 99, 611–625. doi:10.1037/0022-0663.99.3.611.

Skinner, E. A., & Wellborn, J. G. (1994). Coping during childhood

and adolescence: A motivational perspective. In D. L. Feather-

man, R. M. Lerner, & M. Perlmutter (Eds.), Life-span develop-

ment and behavior (Vol. 12, pp. 91–133). Hillsdale, NJ:

Lawrence Erlbaum.

Smith, E., Osgood, D., Caldwell, L., Hynes, K., & Perkins, D. (2013).

Measuring collective efficacy among children in community-based

afterschool programs: Exploring pathways toward prevention and

positive youth development. American Journal of Community

Psychology, 52, 27–40. doi:10.1007/s10464-013-9574-6.

Sweetland, S. R., & Hoy, W. K. (2000). School characteristics and

educational outcomes: Toward an organizational model of student

achievement in middle schools. Educational Administration

Quarterly, 36, 703–729. doi:10.1177/00131610021969173.

Thijs, J. T., Koomen, H. M. Y., & van der Leij, A. (2008). Teacher–

child relationships and pedagogical practices: Considering the

teacher’s perspective. School Psychology Review, 37, 244–260.

Trickett, E. J., & Rowe, H. L. (2012). Emerging ecological

approaches to prevention, health promotion, and public health

in the school context: Next steps from a community psychology

perspective. Journal of Educational and Psychological Consul-

tation, 22, 125–140. doi:10.1080/10474412.2011.649651.

Tschannen-Moran, M., & Hoy, A. W. (2001). Teacher efficacy:

Capturing an elusive construct. Teaching and Teacher Educa-

tion, 17, 783–805. doi:10.1016/S0742-051X(01)00036-1.

Wang, M.-T., & Eccles, J. S. (2012). Adolescent behavioral, emotional,

and cognitive engagement trajectories in school and their differ-

ential relations to educational success. Journal of Research on

Adolescence, 22, 31–39. doi:10.1111/j.1532-7795.2011.00753.x.

Wang, M. C., & Gordon, E. W. (2012). Educational resilience in

inner-city America: Challenges and prospects. Mahwah, NJ:

Routledge.

Warren, J. S., Edmonson, H. M., Griggs, P., Lassen, S. R., McCart,

A., Turnbull, A., et al. (2003). Urban applications of school-wide

positive behavior support critical issues and lessons learned.

Journal of Positive Behavior Interventions, 5, 80–91. doi:10.

1177/10983007030050020301.

Wasik, B. A., Bond, M. A., & Hindman, A. (2006). The effects of a

language and literacy intervention on Head Start children and

teachers. Journal of Educational Psychology, 98, 63–74. doi:10.

1037/0022-0663.98.1.63.

Way, N., Reddy, R., & Rhodes, J. (2007). Students’ perceptions of

school climate during the middle school years: Associations with

trajectories of psychological and behavioral adjustment. Amer-

ican Journal of Community Psychology, 40, 194–213. doi:10.

1007/s10464-007-9143-y.

Weiss, C. C., & Kipnes, L. (2006). Reexamining middle school effects:

A comparison of middle grades students in middle schools and k-8

schools. American Journal of Education, 112, 239–272.

Williamson, R., & Johnston, J. H. (1999). Challenging orthodoxy: An

emerging agenda for middle level reform. Middle School

Journal, 30, 10–17.

Am J Community Psychol (2014) 54:28–45 45

123