Journal of School Psychology 53 (2015) 447–461

Contents lists available at ScienceDirect

Journal of School Psychology

journal homepage: www.elsevier.com/locate/jschpsyc

Measurement and structural relations of an authoritative school

climate model: A multi-level latent variable investigation☆
Timothy R. Konold ⁎, Dewey Cornell
Curry School of Education, University of Virginia, USA

a r t i c l e i n f o a b s t r a c t

Article history: This study tested a conceptual model of school climate in which two key elements of an authori-
Received 8 January 2015 tative school, structure and support variables, are associated with student engagement in school
Received in revised form 2 September 2015 and lower levels of peer aggression. Multilevel multivariate structural modeling was conducted
Accepted 8 September 2015
in a statewide sample of 48,027 students in 323 public high schools who completed the Authori-
Available online 18 October 2015
tative School Climate Survey. As hypothesized, two measures of structure (Disciplinary Structure
and Academic Expectations) and two measures of support (Respect for Students and Willingness
Keywords: to Seek Help) were associated with higher student engagement (Affective Engagement and
School Climate
Cognitive Engagement) and lower peer aggression (Prevalence of Teasing and Bullying) on both
multi-level factor analysis
student and school levels of analysis, controlling for the effects of school demographics (school
measurement
size, percentage of minority students, and percentage of low income students). These results
support the extension of authoritative school climate model to high school and guide further
research on the conditions for a positive school climate.
© 2015 Society for the Study of School Psychology. Published by Elsevier Ltd. All rights reserved.

School climate is broadly defined as the “quality and character of school life” and is “based on patterns of people's experiences of
school life and reflects norms, goals, values, interpersonal relationships, teaching and learning practices, and organizational
structures” (Cohen, McCabe, Michelli, & Pickeral, 2009 p. 182). Many studies have identified a positive school climate as an important
condition for favorable student outcomes ranging from academic achievement to healthy socio-emotional development (Cohen,
McCabe, Michelli, & Pickeral, 2009; Espelage, Low, & Jimerson, 2014; Kidiger, Araya, Donovan, & Gunnell, 2012; Thapa, Cohen,
Guffey, & Higgins-D'Alessandro, 2013). For example, Barile et al. (2012) found that favorable student-teacher relationships were
associated with higher academic achievement and lower dropout rates. A study of a nationally representative sample of secondary
schools found that a school climate characterized by clear and fair school rules had less delinquent behavior and student victimization
(Gottfredson, Gottfredson, Payne, & Gottfredson, 2005). Another study found that a positive school climate was associated with
greater teacher satisfaction and teaching efficacy (Collie, Shapka, & Perry, 2012).
Based on this diverse body of research, there are numerous state and national initiatives to make positive school climate a guiding
principle of school improvement (Dary & Pickeral, 2013; Piscatelli & Lee, 2011; U.S. Department of Education, 2014). In its “Guiding
Principles” resource guide, the U.S. Department of Education (2014) specifically urged schools to “engage in deliberate efforts to
create positive school climates” (p. 5) as a means to engage all students in learning, prevent problem behaviors, and support

☆ We thank Donna Michaelis and Jessica Smith of the Virginia Department of Criminal Justice Services for their support of the Virginia Secondary School Climate
Study. This project was supported by Grant #2012-JF-FX-0062 awarded by the Office of Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S.
Department of Justice. The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect
those of the Department of Justice.
⁎ Corresponding author at: Curry School of Education, University of Virginia, 417 Emmet Street South, P.O. Box 400265, Charlottesville, VA 22904-4265, USA.
URL: Konold@Virginia.edu (T.R. Konold).
Action Editor: Andy Garbacz

http://dx.doi.org/10.1016/j.jsp.2015.09.001
0022-4405/© 2015 Society for the Study of School Psychology. Published by Elsevier Ltd. All rights reserved.
448 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

struggling or at-risk students. High schools have been targeted for school climate improvement because of concern about the
unfairness of school disciplinary practices and the associated risk of school failure and dropout (Morgan, Salomen, Plotkin, &
Cohen, 2014).
Several studies have found that a positive school climate is associated with successful implementation of prevention programs
(Bradshaw, Koth, Thornton, & Leaf, 2009; Low & Van Ryzin, 2014). Although there is general agreement with the theory that school
climate influences student learning and behavior (Cohen et al., 2009), there is relatively little consensus on the key dimensions of
school climate and how to measure them (American Institutes for Research, 2013; Hung, Luebbe, Flaspohler, 2014). The purpose of
the present study is to investigate the authoritative school climate model as a framework for measuring and testing relations
among key elements of school climate at the high school level.

1. Authoritative school climate model

The authoritative school climate model is derived from Baumrind's (1968) work on authoritative parenting which continues to
guide a substantial body of child development research (Larzelere, Morris, & Harrist, 2013). This work identified two dimensions of
parenting: one dimension concerned with the parent's high expectations and demands for the child and the other concerned with
how warm and supportive the parent is toward the child. Research has found that parents are most effective when they apply a
combination of strict discipline and emotional support for their children, and are less effective when they are highly structured and
demanding but not supportive (authoritarian), emotionally supportive but lacking in structure (permissive), or lacking in both
structure and support (disengaged or neglectful).
The authoritative school climate model presents a promising theoretical framework for conceptualizing a positive school climate
(Gregory & Cornell, 2009; Gregory et al., 2010; Konold et al., 2014). According to this model, two key dimensions of school climate are
structure and support, which are similar, but not equivalent, to the corresponding dimensions of high expectations and supportive-
ness in the parenting literature. Two components of structure or high expectations have been used in the literature. Some studies
have measured high expectations in the form of disciplinary structure, which refers to strict but fair enforcement of school rules
(e.g., Gregory et al., 2010; Konold et al., 2014), while other studies have included the academic expectations or academic press that
teachers demand of their students as another aspect of a structured or demanding school (Gill, Ashton, & Algina, 2004; Pellerin,
2005; Gregory, Cornell, & Fan, 2011). Studies have more consistently measured support (sometimes called responsiveness) as the
degree to which students perceive their teachers to be supportive, respectful, and willing to help them.
It has been theorized that both structure and support dimensions are important because students are more willing to comply with
the expectations of school authorities when they feel supported and respected by them (Gregory & Cornell, 2009). Measures of
tructure and support are found on other school climate surveys with various scale names (Bear, Gaskins, Blank, & Chen, 2011;
Bradshaw, Waasdrop, Debnam, & Johnson, 2014; Brand, Felner, Shim, Seitsinger, & Dumas, 2003), but the authoritative model places
special emphasis on these domains as integrally related to overall school climate and successful student outcomes. Other models of
school climate do not make specific claims for the relations among their school climate scales.

1.1. Initial studies of authoritative school climate

Three studies operationalized an authoritative school climate as characterized by independent variables of high demandingness in
the form of high academic expectations for students, and high responsiveness, which was defined by student perceptions that their
teachers were supportive and interested in their students. These studies used pre-existing national databases with student and/or
administrator survey questions that subsequently could be used to construct their measures of school climate. Because these
databases were not constructed to measure authoritative school climate, in each study the authors constructed indicators of
authoritative school climate qualities using available survey items or scales. A study using National Education Longitudinal Study
(NELS) data tested whether an authoritative school climate was associated with mathematics achievement, internal control, and
student engagement (Gill, Ashton, & Algina, 2004). Gill et al. (2004) measured authoritative qualities they termed “responsiveness”
and “demandingness” with selected student and principal survey items that were factor analyzed and labeled “Student Perceived
School Responsiveness,” “Principal Perceived School Responsiveness,” and “Principal Perceived School Demandingness.” Of most
relevance to the present study, a hierarchical multilevel modeling analysis found that Students' Perceived Responsiveness was
significantly related to a three-item student-report measure of engagement in learning (β = .35).
Pellerin (2005) also used NELS survey data and constructed measures of responsiveness, academic demandingness, and disciplin-
ary demandingness using a combination of administrator (principal) and student survey items. Pellerin (2005) constructed overall
measures of responsiveness and demandingness for each school and classified schools into authoritative, authoritarian, permissive,
and indifferent categories. A series of ordinary least regression analyses found that authoritative schools had less truancy and fewer
dropouts than other schools. This study did not conduct multilevel analyses or provide information to distinguish the relative
contributions of authoritative school qualities.
Lee (2012) used student survey data from the Program for International Student Assessment (PISA) for U.S. schools. Student
survey items measured independent variables of “Teacher-student Relationship” and “Academic Press,” and dependent variables of
“Behavioral Engagement,” “Emotional Engagement,” and “Academic Performance.” A multilevel analysis at student and school levels
found that a supportive teacher-student relationship was associated with all three student outcomes (specifically, a one-unit increase
in teacher–student relationship was associated with a 0.33 unit increase in behavioral engagement, 0.29 increase in emotional
engagement, and 10.96-unit increase in reading performance). Academic press was associated with behavioral engagement
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 449

(β = 0.13) and emotional engagement (β = .0.06), but not reading performance. Interaction effects for teacher–student relationship
and academic press were not significant.
A fourth study used a nationally representative sample of 23,974 student surveys from the School Crime Supplement to the
National Crime Victimization Survey (Gerlinger & Wo, 2014). This study identified four student survey items to construct an index
of structure and support (school rules were fair, school rules were strictly enforced, punishment for breaking school rules was the
same for all students, and teachers treated students with respect). This somewhat limited measure of authoritative school climate
was associated with lower student reports of physical, verbal, and relational bullying in school, which were measured with single
items. Logistic regression analyses found that, after controlling for other school characteristics, the authoritative school climate
measure generated odds ratios of .35 for physical bullying, .39 for verbal bullying, and .33 for relational bullying.

1.2. Authoritative school climate survey of ninth grade

Although the aforementioned four studies had the advantage of using nationally representative samples, they were not designed
to measure authoritative school climate and relied on a limited number of available survey items to construct measures of school
climate as well as associated outcomes of student engagement and bullying. Another group of studies prospectively measured
structure and support dimensions of school climate using items selected to measure those constructs. These studies all made use of
the same statewide sample of 7318 ninth grade students and 2922 teachers selected from 290 high schools. In the first study,
Gregory et al. (2010) conducted exploratory and confirmatory factor analyses to identify a series of items concerned with student
perceptions that school rules are strict but fair that was labeled School Structure and another series of items concerned with student
perceptions that teachers are supportive and willing to help their students that was labeled School Support. This study examined three
outcome measures, student and teacher reports of the prevalence of teasing and bullying (Bandyopadhyay, Cornell, & Konold, 2009)
and student reports of general victimization, such as being threatened or physically attacked, derived from Gottfredson (1999). A
hierarchical linear modeling analysis found that school structure and support accounted for an additional 45% of the between-schools
variance in student-reported bullying, 8% of the variance in teacher-reported bullying, and 50% of the between-school variance in
student-reported victimization beyond the effects of school size, proportion of minority students, and proportion of students who
qualified for free- or reduced-price meals.
A second study examined the suspension rates in a somewhat smaller subgroup of 199 schools (limited to schools with sufficient
numbers of Black and White students and available data on schoolwide suspension rates for both Black and White students; Gregory,
Cornell, & Fan, 2011). This study examined one indicator of support, Supportive Climate, and two indicators of structure, Experience of
School Rules and Academic Press, and their interactions. Hierarchical multiple regression analyses found that schools lowest on
academic press and support had the highest suspension rates, accounting for 7% of the variance in Black suspension rates and 5% of
the variance in White suspension rates after controlling for school demographics (Gregory, Cornell, & Fan, 2011).
A third study examined student aggression against teachers in 280 high schools with both student and teacher survey data
(Gregory, Cornell, & Fan, 2012). Teacher victimization was measured by teacher-reports on seven victimization items, ranging from
receiving obscene remarks from a student to being physically assaulted (Gottfredson, 1999). As an alternative source of information
about aggression against teachers, school records of serious disciplinary infractions were obtained from a state database, but only
infractions for “threats or intimidation against school staff” occurred with sufficient frequency to compare across schools. Both
structure, as measured by student- and teacher-reported Experience of School Rules, and support, as measured by teacher-reported
but not student-reported Help Seeking, were associated with less teacher victimization, after controlling for school and neighborhood
demographics. These characteristics of an authoritative school climate explained 19% of the variance in teacher reports of victimization
and 6% of the variance in school records of disciplinary infractions involving threats against school staff.

1.3. Authoritative school climate survey in middle school

Two recent studies extended the line of research with ninth grade to a middle school population (7th and 8th grades; Konold et al.,
2014; Huang et al., in press). These studies revised the survey used with ninth grade (renamed the survey to the Authoritative School
Climate Survey) and employed more complex, multi-level factor analyses to examine scales for use at both individual and school
levels of analysis. The first study conducted a multi-level structural analysis to evaluate the factor structure and construct validity
of four scales (structure, support, engagement, and prevalence of teasing and bullying) in a Virginia statewide sample of 39,364
students attending 423 schools (Konold et al., 2014). Disciplinary structure was based on the prior scale measuring perceptions
that school disciplinary practices were strict and fair (Gregory et al., 2010). Support was based on the prior scale measuring
perceptions that school staff were supportive and respectful, and that students were willingness to seek help from them (Gregory
et al., 2010). The factor analyses reduced the Disciplinary Structure scale to seven items and the Student Support scale to eight
items that functioned on both student and school levels. Prevalence of Teasing and Bullying was increased from four items to five
items, and Engagement was decreased to six items. This study established that scales for structure and support had good model fit
on both student and school levels, and that the measures could be used for evaluating student perceptions of structure and support
as well as assessing the degree to which these characteristics were present for a given school. Moreover, this work also demonstrated
that more authoritative schools manifested higher student engagement and lower prevalence of teasing and bullying.
A second study examined a teacher version of the Authoritative School Climate Survey in a statewide sample of 9099 7th and 8th
grade teachers (Huang et al., in press). Although this study was confined to teacher reports, it also provided support for the construct
validity of the authoritative school climate model through multilevel structural modeling with results consistent with the findings for
450 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

students. Similar measures of structure and support were identified on both teacher and school levels, and an authoritative school
climate was associated with higher levels of student engagement and lower levels of student teasing and bullying.

2. Present study

The current study extends the line of research on authoritative school climate in both methodological and substantive directions
through investigation of two guiding research questions. First, does the factor structure of items on the authoritative school climate
survey maintain for high school students and their schools? Second, are high levels of structure and support variables associated
with: a) lower prevalence rates of teasing and bullying, and b) greater cognitive and affective engagement? The current study
examined the authoritative school climate measures in a large and diverse statewide sample of high schools encompassing grades
9–12, whereas previous investigations of this model have focused on earlier grades (i.e., grades 7–9). It tested the same measures
of structure, support (respect for students and willingness to seek help), engagement (affective and cognitive), and prevalence of
teasing and bullying used in the middle school studies. Here however, we broadened the concept of school structure to include not
only disciplinary structure, but also items designed to measure academic expectations (i.e., academic press) that teachers have for
students. Because high school student perspectives and responses to these items may differ in important ways from those obtained
by students in middle school grades, and because additional items have been added to this survey to assess additional constructs of
authoritative school climate, we employed multilevel confirmatory factor analysis to test the measurement structure of these items.
In examining the measurement characteristics of these items, we also focused on the extent to which the resulting scales can be
used to assess features of both the student and the school in order to ensure they maintain the same properties when used for
evaluating these different targets (Bliese, 2000). Although some constructs might only be meaningful on an individual level
(e.g., personality) or a school level (e.g., racial diversity), the constructs of authoritative school climate are likely to hold importance
for both students and the schools they attend. For example, student perceptions of school structure and support are likely to be
influenced by the overall levels of these features within a given school, and both student and school measures of these dimensions
may have different causes and influences. We employed multilevel factor analysis because it provides a valuable means of
determining the degree to which items are useful for measuring constructs at different levels, and whether the measurement of
these constructs is consistent across levels (Muthén, 1991). Examination of authoritative school climate constructs on both students
and schools is an important methodological contribution of this study, because the development of many school climate scales fails to
model distinctions between these levels of measurement, which can result in a variety of substantive misinterpretations and
erroneous conclusions (Dedrick & Greenbaum, 2011; Dyer, Hanges, & Hall, 2005).
Our first research question was a preliminary step to investigating the manner in which the constructs of authoritative school
climate are related to one another (i.e., research question two). The first question focused on the measurement of these constructs
and it is hypothesized that the number and nature of constructs underlying the investigated items will be similar to that which
was found when investigations were focused on middle school samples (Konold et al., 2014). However, the newly revised Standards
for Educational and Psychological Testing (American Educational Research Association [AERA], American Psychological Association
[APA], & National Council on Measurement in Education [NCME], 2014) place even greater emphasis on ensuring that score interpre-
tations are valid “for the widest possible range of individuals and relevant subgroups” (p. 63). As a result, our first research question
focused specifically on testing whether items on the authoritative school climate survey are operating in the same manner when
employed with samples of high school students.
High schools differ in important ways from middle schools in terms of expectations of students, increased autonomy, and greater
demands on the student's time. In fact, ninth grade in particular represents a particularly challenging time as the national failure rate
of students in grade nine is greater than any other grade level (Wheelock & Miao, 2005). The differences between middle school and
high school are underscored by a meta-analysis finding that bullying prevention programs that are successful in elementary and
middle school are not effective in high school (Yeager, Fong, Lee, & Espelage, 2015). There seems to have been a general assumption
that bullying programs developed for younger students could be applied to older students with only minor modifications. Yeager and
colleagues (2015) pointed out that high school-age adolescents have markedly different social skills and interact in a more complex
social environment than younger students. Student perceptions of middle and high school climates are likely to be influenced by
environmental differences that vary as a function of differences in norms, values, organizational structures and relationships: the
very characteristics that define school climate. Based on these considerations, it seems necessary to examine the factor structure
and scale relations of the Authoritative School Climate Survey in a high school environment rather than assume that evidence from
middle school is sufficient.
The second research question examined whether high levels of structure and support variables are associated with positive
student outcomes, extending previous work with middle school students (Konold et al., 2014) to a population of high school students.
Equally important, the present study also examines a more comprehensive authoritative school climate model that includes the
influences of school structure and support on student engagement (separately considering affective and cognitive components)
and prevalence of teasing and bullying at the school level with the incorporation of important school level control variables that
have not been previously considered in the prior research on middle schools. These control variables include school size, enrollment
of disadvantaged minority students, and student socioeconomic status that are consistently associated with lower levels of student
engagement and weaker academic performance (Gottfredson et al., 2005; Leithwood & Jantzi, 2009). Moreover, students from
lower income households typically report more bullying victimization (Tippet & Wolke, 2014). On the basis of similar investigations
with younger children (Cornell et al., 2015; Gregory et al., 2010) and research that operationalized authoritative school climate
somewhat differently (Lee, 2012), we hypothesized that high levels of structure and support variables would be associated with
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 451

greater levels of engagement and lower prevalence of teasing and bullying. In summary, the present study represents an upward
extension of the Authoritative School Climate Model to high school grades. In addition, this study for the first time evaluated the
multi-level structural relations among the four key school climate measures while controlling for important school level
demographics that were not considered in prior studies of middle school.

3. Method

3.1. Participants and settings

Study schools were obtained from the 2014 Virginia Secondary School Climate Survey, which is part of the state's annual School
Safety Audit program. The survey was administered to students in Virginia public schools with grades 9–12 (including some schools
without a grade 9). The survey did not include schools located in juvenile detention and correctional facilities, centers providing
part-time or temporary services such as suspension centers, facilities exclusively serving students with disabilities, or programs
specifically for adults. A total of 323 of 324 eligible schools participated in the survey. The school participation rate of 99.7% was
achieved with the cooperation of the Virginia Department of Education and the Virginia Department of Criminal Justice Services,
who endorsed the study and encouraged participation.
Student enrollments in the 323 schools ranged from N = 66 to 4072 (M = 1178 students; http://www.doe.virginia.gov). Based on
U.S. census classification, schools were located in urban (20.7%), suburban (31.6%), town (4.6%), and rural (43.0%) regions (Virginia
Department of Education, n.d.). Virginia Department of Education records indicated that the percentage of students eligible for free
or reduced price meals (FRPM) in the participating schools ranged from 1.9% to 94.3% (M = 38.0%). School enrollments obtained
from the Virginia Department of Education were approximately 60.5% Caucasian, 23.0% African American, 8.7% Hispanic, 4.1% Asian
American, and 3.8% other minority groups.
Student participation rate was defined as the total number of students across all schools who participated in the survey divided by
the total number invited to take the survey. The overall student participation rate was estimated to be 88.7% (52,012 student
participants from a pool of 58,613 students asked to participate). Participation rates for schools choosing to invite all students to
participate and those electing the random sampling option were 82.9% and 93.4%, respectively. Principals from 299 of 323 schools
(92.6%) completed an additional survey. According to their reports, 45 school principals invited all students to take the student survey
and 254 school principals elected to use the random sampling option for surveying students. Operating within the constraints of the
state agency charged with administering the survey, there was no procedure to confirm that the principals followed the random
selection directions.
After survey screening (described below), the analytic sample consisted of N = 48,027 (51.4% female) participants in ninth
(26.1%), tenth (26%), eleventh (24.9%) and twelfth (23.1%) grade. The racial/ethnic breakdown was 59.1% European American,
18.6% African American, 10.5% Hispanic, 4% Asian American, 1.6% American Indian or Alaska Native, and 0.9% Native Hawaiian or
Pacific Islander, with an additional 15.8% of students identifying themselves with having more than one race. Approximately 18.9%
reported speaking a language other than English at home. The distribution of parental education was 19.9% completed
post-graduate studies, 24.1% completed a four-year college degree, 16% completed a two-year college or technical education degree,
31.2% graduated from high school, and 8.8% did not graduate from high school.

3.2. Procedures

Students completed the survey in classrooms under teacher supervision using a standard set of instructions. Each student was
provided with a password unique to the school. Surveys were administered anonymously online using Qualtrics software. Schools
were given two options for sampling students: (1) invite all students in each grade to take the survey, with a goal of surveying at
least 70% of all eligible students; (2) randomly select at least 25 students from each grade to take the survey. Schools were given
these options in order to provide administrators flexibility in choosing a more or less comprehensive assessment of their students.
Schools choosing the random sample option were provided with a random number list that uniquely corresponded to the size of
their school along with instructions for selecting students. The instructions indicated that the random numbers should be applied
to an alphabetized list of student names for purposes of selecting students for participation. Principals were advised to invite up to
50 students in each grade to take the survey in order to have a pool of alternates in the event that any of the first 25 selected students
were unable or unwilling to participate. All students were eligible to participate except those unable to complete the survey because of
limited English proficiency or an intellectual or physical disability.
On average, participants completed the 80-item survey in 14.2 min (SD = 9.2). To improve data quality, a multi-stage screening
procedure resulted in the removal of student surveys (N = 649 or 1.3% of the sample) for completing the survey too rapidly (i.e., less
than six minutes) or not being truthful as revealed through responses to two validity questions (an additional N = 3,336 or 6.4% of the
sample).

3.3. Measures

The complete online survey consisted of approximately 100 items that were primarily intended to elicit student perceptions of
climate and safety within their schools. The survey also included items asking about student activities, values, disciplinary infractions,
and demographics. The focus of the current study was on items designed to measure aspects of school climate in relation to structure
452 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

(i.e., discipline and academic expectations) and support (i.e., respect for students and willingness to seek help), as well as items to
measure the prevalence of teasing and bullying, and affective and cognitive engagement. The headings below are used to organize
our descriptions of each of these variables. Responses to all items were recorded on a four-point scale: “strongly disagree,” “disagree,”
“agree,” and “strongly agree.” All items for each measure are shown in Table 1.

3.3.1. Disciplinary structure

The perception that school discipline is strict but fair was measured with a seven-item scale that was derived from previous
research on the concept of school disciplinary structure (Gregory et al., 2010) and the Experience of School Rules scale developed
by federal authorities for use in the School Crime Supplement to National Crime Victimization Survey (National Center for
Education Statistics, 2005). Example items include: “Students at this school only get punished when they deserve it” and “The school
rules are fair.” Previous research on these items in middle school samples revealed school level factor loadings that ranged from .77 to
.95, and a school level alpha reliability estimate for the scale of .70 (Konold et al., 2014).

3.3.2. Academic expectations

Student perceptions that their teachers held expectations of high academic standards for them were measured with a five-item
scale. This scale was derived from the work of Midgley, Maehr, Hruda, Anderman, and Freeman (2000) on a scale called “Academic
Press” and also was used as an indicator of structure by Gregory, Cornell, and Fan (2011). The scale in the present study was intended

Table 1
Multilevel confirmatory factor analysis completely standardized loadings for students (W) and schools (B).

Structure (W) B

Disciplinary Structure (reliability: student level = .78 and school level = .95)
STR 1 The punishment for breaking school rules is the same for all students (.71⁎) .89⁎
STR 2 Students at this school only get punished when they deserve it (.71⁎) .97⁎
STR 3 Students are treated fairly regardless of their race or ethnicity (.72⁎) .89⁎
STR 4 Students get suspended without good reason (reverse scored) (.36⁎) .88⁎
STR 5 The adults at this school are too strict (reverse scored) (.40⁎) .74⁎
STR 6 The school rules are fair (.75⁎) .93⁎
STR 7 When students are accused of doing something wrong, they get a chance to explain it (.65⁎) .89⁎
Academic Expectations (Reliability: Student Level = .72 and School Level = .86)
STR 8 My teachers expect me to work hard (.82⁎) .99⁎
STR 9 My teachers really want me to learn a lot (.93⁎) .97⁎
STR 10 My teachers expect a lot from students (.48⁎) .79⁎
STR 11 My teachers do not really care how much I learn (reverse scored) (.55⁎) .78⁎
STR 12 My teachers expect me to attend college (.57⁎) .65⁎

Support

Respect for Students (reliability: student level = .87 and school level = .90)
SPRT 1 Most teachers and other adults at this school care about all students (.87⁎) .98⁎
SPRT 2 Most teachers and other adults at this school want all students to do well (.87⁎) .95⁎
SPRT 3 Most teachers and other adults at this school listen to what students have to say (.85⁎) .97⁎
SPRT 4 Most teachers and other adults at this school treat students with respect (.86⁎) .99⁎
Willingness to Seek Help (Reliability: Student Level = .73 and School Level = .80)
SPRT 5 There are adults at this school I could talk with if I had a personal problem (.63⁎) .67⁎
SPRT 6 If I tell a teacher that someone is bullying me, the teacher will do something to help (.81⁎) 1.00
SPRT 7 I am comfortable asking my teachers for help with my school work (.74⁎) .93⁎
SPRT 8 At least one teacher or another adult at this school really wants me to do well (.68⁎) .67⁎

Prevalence of Teasing and Bullying (reliability: student level = .85 and school level = .93)

PTB 1 Students here often get teased about their clothing or physical appearance (.78⁎) .95⁎
PTB 2 Students get teased or put down because of their race or ethnicity at this school (.77⁎) .87⁎
PTB 3 There is a lot of teasing about sexual topics at this school (.76⁎) .88⁎
PTB 4 Students here get teased or put down about their sexual orientation (.79⁎) .87⁎
PTB 5 Bullying is a problem at this school (.74⁎) .95⁎

Engagement

Affective (reliability: student level = .89 and school level = .95)

ENG 1 I like this school (.93⁎) 1.00
ENG 2 I am proud to be a student at this school (.93⁎) .97⁎
ENG 3 I feel like I belong at this school (.84⁎) .99⁎
Cognitive (reliability: student level = .71 and school level = .73)
ENG 4 I usually finish my homework (.68⁎) .83⁎
ENG 5 Getting good grades is very important to me (.81⁎) .67⁎
ENG 6 I want to learn as much as I can at school (.80⁎) .35⁎

W = within schools or student, B = between schools.

⁎ p b .05.
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 453

to cover similar content but to use simple wording about teacher expectations (e.g., “My teachers expect me to work hard” and “My
teachers expect me to attend college”).

3.3.3. Student support

The perception that teachers and other school staff members are supportive of students (e.g., “Most teachers and other adults at
this school want all students to do well”) was measured with an eight-item scale. Four items were taken from the Help Seeking
scale (Bandyopadhyay et al. 2009) used in the School Climate Bullying Survey (Cornell, 2015). The other four items were adapted
from the Learning Environment scale (Austin & Duerr, 2005) designed to measure how much students perceive that adults in their
school are supportive and respectful of students (e.g., “At least one teacher or another adult at this school really wants me to do
well”). Prior research combining these scales found that high schools characterized by higher levels of student support had less bul-
lying and peer victimization as reported by ninth grade students and their teachers (Gregory et al., 2010). School level factor loadings
across all items were previously found to range from .67 to .98, and alpha reliability estimates were .72 and .61 for the two four-item
scales, respectively (Konold et al., 2014).

3.3.4. Prevalence of teasing and bullying (PTB)

The prevalence of teasing and bullying in a school was measured with five items that assessed the extent of bullying and teasing
observed at school as distinguished from personal victimization. Similar to other bullying measures (e.g., Juvonen et al., 2000; Olweus,
2007), item content was not limited to use of the term “bullying,” but included general forms of peer harassment associated with bul-
lying. Previous research supported internal and external validity of this scale across multiple samples spanning grades seven through
twelve (Bandyopadhyay et al., 2009; Klein, Cornell, & Konold, 2012; Konold et al., 2014). For example, school-level factor loadings
ranged from .81 to .97 and alpha reliability was estimated to be .88 (Konold et al., 2014). Illustrative items include: “Students here
often get teased about their clothing or physical appearance” and “Bullying is a problem at this school.”

3.3.5. Student engagement

Student engagement in school was measured with six items derived from the Commitment to School scale (Thornberry, Lizotte,
Krohn, Farnworth, & Jang, 1991) that have been found to be useful in measuring both affective and cognitive aspects of engagement,
with alpha reliability estimates of .96 and .87 for the two scales, respectively (Konold et al., 2014). Mehta, Cornell, Fan, and Gregory
(2013) found that a nine-item version of this scale was negatively associated with student reports of the prevalence of teasing and
bullying in school. Konold et al. (2014) found support for the factor structure and concurrent validity of the scale in a sample of middle
school students. Example items include: “I am proud to be a student at this school” and “I want to learn as much as I can at school.”

3.4. Analytic plan

All analyses were conducted using multilevel modeling approaches to account for the nested nature of the data. This approach
recognizes that student (i.e., level 1) reports within a school are not independent and are likely influenced by the school (i.e., level
2) they attend. Multilevel modeling accounts for violation of the independence assumption by portioning the total variance into
that which can be attributed to differences between level 1 (e.g., students) and level 2 (e.g., school) units, and allows for examination
of associations among variables that may exist at different levels of the design (O'Connell & McCoach, 2008; Raudenbush & Bryk,
2002). By contrast, ordinary regression techniques can only be employed on one level of the design, and the resulting parameter
estimates and standard errors will be biased when the independence assumption is violated (McCoach, 2010).
The extent to which individual student responses vary across schools can be quantified with the intraclass correlation (ICC)
coefficient. Values range from 0 to 1.0 and measure the amount of level 1 variance (e.g., differences in student responses) that can
be explained by the level 2 units (e.g., schools in which the students reside). An ICC value of 0 indicates that all score variation occurs
at level 1 of the design, and ICC values greater than 0 represent “notable” variation across level 2 units (Peugh, 2010, p. 89). For
example, an ICC value of .08 can be interpreted to indicate that 8% the variance in student responses occurs between schools. Related,
design effects measure the extent to which standard errors would be underestimated in the absence of employing multilevel
modeling techniques when the independence assumption is violated. For example, a design effect of 2.0 would indicate that the
standard errors in the model would be underestimated by a factor of 2.0 (McCoach, 2010).
The 31 items hypothesized to measure seven dimensions of Structure (Disciplinary Structure and Academic Expectations),
Support (Respect for Students and Willingness to Seek Help), Prevalence of Teasing and Bullying (PTB), and Engagement (Affective
and Cognitive) were jointly examined through a multilevel confirmatory factor analytic (CFA) model. This analysis answered the
first research question: Does the factor structure of items on the Authoritative School Climate Survey (ASCS) maintain for high school
students and their schools? Each item was specified to load only on its hypothesized factor as shown in Table 1. Notably, the sample of
N = 48,027 students from 323 high schools was used to evaluate the degree to which the hypothesized latent measurement
structures were defensible across both students and schools through specification of a common structure at both levels of analysis.
The model was evaluated through weighted least squares (MV) to account for the ordinal nature of the items.
Two forms of reliability estimates were calculated for each of the CFA factors. The first was the well-known Cronbach's alpha that
was computed at the student level. These estimates, however, do not account for the nested data structure and do not accurately
reflect expected reliability at the school level (Dedrick & Greenbaum, 2011; Raudenbush, Rowan,& Kang, 1991). As a result,
a second set of reliability estimates employed the Spearman-Brown Formula to estimate factor reliabilities at the school
454 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

level: [K(ICC)]/[(K − 1)(ICC) + 1], where K = the average cluster (or school) size, and ICC = intraclass correlation for a factor
[σ2B/(σ2B + σ2W)] (Muthén 1991).
Our second research question focused on whether high levels of structure and support are associated with: a) lower prevalence
rates of teasing and bullying, and b) greater cognitive and affective engagement. Two sets of analyses were conducted to address
this question, both of which examined these associations for students and schools. First, convergent associations were examined
through multilevel models in which factor correlations were evaluated at both the student and school levels. These associations
were examined to assess the degree to which the resulting factors were related in substantively meaningful ways, and the extent
to which these associations maintained across different schools. Secondly, a multilevel structural model evaluated how the full battery
operated in concert through examination of a comprehensive model of authoritative school climate with particular focus on school
level effects. Fig. 1 presents a graphic representation of the model in which two dimensions of structure and support were modeled
to have a direct influence on the outcomes of PTB, affective engagement, and cognitive engagement. School level control variables in-
cluded ethnic composition as operationalized by the percentage of White students enrolled in the school, school poverty as measured
by the percentage of students in the school receiving free or reduced priced meals (FRPM), and school size. All school level covariates
were grand mean centered as recommended when the focus is on the higher level (e.g., schools) of a multilevel model (Peugh, 2010).
In order to assess the unique contributions of the school climate variables on the three school level outcomes, the multilevel
structural model was examined in two steps. The first step specified that the paths linking the four school climate variables to the
three school outcome variables were fixed to zero. Here, only the three school characteristic control variables were freely estimated.
Contrasts between squared multiple correlations (SMCs) of the outcome variables from this step with SMCs obtained from a second
model in which all paths were freely estimated allowed for quantification of the amount of variance in school level outcomes that
could accounted for by measures of school climate, beyond that which could be explained by the school characteristic models.
Student level influences of school climate on the outcomes of PTB and engagement (i.e., affective and cognitive) were also
examined within the context of the multilevel structural model displayed in Fig. 1. However, because the school level control variables
were not available at the student level, interpretations of these influences are more limited with respect to evaluation of a theoretical
model. Here, the resulting associations are important from the perspective of evaluating the psychometric properties of the instru-
ment within a concurrent validity framework in a way that can be useful to schools for evaluating climate.
Four measures of fit were considered in evaluating model quality: the Tucker–Lewis index (TLI), comparative fit index (CFI), root
mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR; Browne & Cudeck, 1993; Hu &
Bentler, 1995). These four measures generally range between 0 and 1.0. Values of .90 or greater (Bentler & Bonett, 1980) or .95 or
greater (Hu & Bentler, 1999) on the first two measures are often taken as evidence of good fitting models. Alternatively, smaller

Fig. 1. Multi-level structural model linking school-level structure and support scales with prevalence of teasing and bullying and engagement scales, with school level
controls.
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 455

RMSEA and SRMR values support better fitting models, with RMSEA values of .06 or less and SRMR values “close to .08” (Hu & Bentler,
1999, p. 1) representing good fit. We also reported the popular χ2 statistic for all models, but place little emphasis on this measure for
evaluating the quality of stand-along models as it well known to reject reasonably specified models when estimated on large samples
(Cheung & Rensvold, 2002; Gerbing & Anderson, 1992; Hu & Bentler, 1995; Keith, 1997; Schumacker & Lomax, 2010). All models were
estimated with MPlus version 7.1 where categorical indicators were handled through WSLMV estimation.

4. Results

Preliminary analyses to answer the first research question found that the average cluster size of students across the 323 schools
was 148.7. Schools accounted for 2% to 14% of the variance in student responses (ICCs ranged from .02 to .14; MICC = .06) across
the 31 items, and design effects ranged from 1.99 to 4.66. As shown in Table 1, three of the 31 items were reverse-coded prior to
analyses so that higher scores reflected more favorable dispositions. PTB was an exception where higher scores reflected more
negative dispositions.

4.1. First research question

Estimation of the multilevel confirmatory factor (CFA) model resulted in negative error variances for two of the 31 items. Because
the confidence intervals of these estimates captured zero, these estimates were sequentially fixed to zero in subsequent estimations.
With these two model modifications, all remaining parameter estimates were within acceptable thresholds.
Measures of model fit for this comprehensive multilevel measurement model were favorable (CFI = .964, TLI = .959,
RMSEA = .023). Fit for the student level portion of the model was also good (SRMR within = .040) and that for the school
level portion of the model was acceptable (SRMRbetween = .096). As expected (Schumacker & Lomax, 2010) given the size of
the sample, chi-square was statistically significant, χ2 (828) = 22,780.32, p b .05.
Multilevel CFA results for the jointly estimated items and hypothesized scales are presented for both students (W) and schools
(B) in Table 1. All factor loadings were large and statistically significant and supportive of the construct validity of these scales in
indicating that the items provide good measures of their respective factors when used for both students and schools. Across all factors,
school level item-factor associations were generally larger at the school level than at the student level. This likely reflects the fact that
school level loadings are based on means from multiple informants where much of the measurement error has been removed (Byrne,
2012). Student-level Cronbach's alpha reliability estimates across the seven scales were all greater than .70, ranging from .71 to .89.
School-level Spearman-Brown reliability estimates were also all greater than .70, with all but one scale (i.e., Willingness to Seek
Help) yielding estimates of .80 or greater.

4.2. Second research question

Correlations among the seven latent factors are shown for both students (above diagonal) and schools (below diagonal) in Table 2.
All correlations were reflective of good convergent validity in that they were within expectation, statistically significant, and with few
exceptions were moderate to large in magnitude. Associations among scales were greater at the school level than at the student
level. The sole exception to this was the correlation between Discipline and Cognitive Engagement where student (r = .45) and school
(r = .44) estimates were similar. Notably, associations between the PTB scale and all other measures were materially stronger at the
school level than at the student level. Differences between these school and student level estimates ranged from .32 to .54.
Fig. 1 presents a graphic representation of the full multilevel structural model that investigated associations among measures of
structure and support with the prevalence of teasing and bullying and two measures of engagement, while controlling for school
level demographic features. ICC values for the three outcome measures ranged from .018 to .113, see Table 3; and design effects
ranged from 1.91 to 4.21. Two specifications of this model were examined. The first model constrained relations between the four
exogenous school climate predictors and three outcomes to zero in order to estimate the amount of variance in the three school
outcome variables that could be accounted for by school characteristics alone. This served as a baseline model. The second model

Table 2
Multilevel model latent factor correlations for students (upper diagonal) and schools (lower diagonal).

1 2 3 4 5 6 7

Structure
1: Disciplinary Structure – .55⁎ .77⁎ .60⁎ −.45⁎ .60⁎ .45⁎
2: Academic Expectations .71⁎ – .66⁎ .72⁎ −.20⁎ .48⁎ .53⁎
Support
3: Respect for Students .93⁎ .78⁎ – .68⁎ −.37⁎ .54⁎ .41⁎
4: Willingness to Seek Help .82⁎ .90⁎ .90⁎ – −.27⁎ .55⁎ .52⁎
Teasing and Bullying
5: Prevalence of Teasing and Bullying −.77⁎ −.74⁎ −.80⁎ −.69⁎ – −.34⁎ −.10⁎
Engagement
6: Affective Engagement .87⁎ .73⁎ .88⁎ .80⁎ −.74⁎ – .39⁎
7: Cognitive Engagement .44⁎ .77⁎ .53⁎ .73⁎ −.42⁎ .39⁎ –
⁎ p b .001.
456 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

Table 3
Completely standardized (student) and school-level structure coefficients and fit statistics from structural model of student and school level outcomes.

School characteristics model: Full model:

(student) and school level outcomes (student) and school level outcomes

Prevalence of Affective Cognitive Prevalence of Affective Cognitive

Teasing and Bullying Engagement Engagement Teasing and Bullying Engagement Engagement

Authoritative School Climate Predictors

Disciplinary Structure – – – (−.35⁎) −.17 (.32⁎) .26⁎ (.21⁎) −.14
Academic Expectations – – – (.13⁎) −.45⁎ (.03⁎) .15 (.34⁎) .44
Respect for Students – – – (−.16⁎) −.76⁎ (.10⁎) .27⁎ (−.13⁎) −.30
Willingness to Seek Help – – – (−.06⁎) .62⁎ (.28⁎) .14 (.24⁎) .65⁎

School Characteristics
Percent White .35⁎ .23⁎ −.29⁎ .34⁎ .23⁎ −.29⁎
School Poverty .61⁎ −.32⁎ .01 .61⁎ −.32⁎ .01
School Size .34⁎ .19⁎ b.01 .35⁎ .19⁎ −.01

ICC .086 .113 .018 – – –

Residuals (1.0) .80⁎ (1.0) .74⁎ (1.0) .92⁎ (.80) .19⁎ (.59) .13⁎ (.65) .45⁎
SMC (0%) 20% (0%) 26% (0%) 8% (20%) 81% (41%) 87% (35%) 55%
Δ SMC – – – (20%) 61% (41%) 61% (35%) 47%

Model Fit Statistics

χ2 19,2202.70 26,201.95
df 936 914
RMSEA .065 .024
CFI .712 .962
TLI .685 .957
SRMR
Within .176 .040
Between .437 .064

ICC = Intraclass Correlation, SMC = Squared Multiple Correlation, χ2 = Chi-Square Statistic, RMSEA = Root Mean Square Error of Approximation, CFI = Comparative
Fit Index, TLI = Tucker–Lewis Index, SRMR = Standardized Root Mean Square Residual.
⁎ p b .05.

freely estimated the paths between the four authoritative school climate predictors and the three outcome variables. Estimation of
these models resulted in negative error variances for four of the 31 items. Because the confidence intervals of these estimates captured
zero, these estimates were sequentially fixed to zero.
Results of the two model estimations are presented in Table 3. Estimates for the first model specification are shown in columns two
through four of Table 3, and estimates for the second model in which the associations of the authoritative school climate predictors
were estimated in the model are shown in columns five through seven of Table 3. Fit for the first model in which only the school
characteristic variables were estimated were mixed. Although the RMSEA (.065) indicated an acceptable model, other estimates
(i.e., CFI = .712, TLI = .685, SRMRwithin = .176, and SRMRbetween = .437) indicated otherwise; see left side of Table 3. By contrast,
fit statistics for the second model in which all paths were freely estimated were suggestive of good fit across all standalone measures
(i.e., RMSEA = .024, CFI = .962, TLI = .957, SRMRwithin = .040, and SRMRbetween = .064).
At the school level, the three school characteristic variables of ethnic composition, school poverty, and school size explained 20%,
26%, and 8% of the variance in PTB, Affective Engagement, and Cognitive engagement, respectively. Estimation of the associations
between the four school climate predictors and the three school outcome variables indicated that the school climate variables jointly
accounted for an additional 61%, 61%, and 47% of the variance in PTB, Affective Engagement, and Cognitive Engagement, respectively
while controlling for school characteristics. It is important to note that these variance percentages pertain to the 8.6% (ICC = .086) of
school level PTB variance, 11.3% (ICC = .113) of school level Affective Engagement variance, and 1.8% (ICC = .018) of school level
Cognitive Engagement variance. With the exception of Cognitive Engagement where relatively little between-school variation was
observed, the ICC values for PTB and Affective Engagement were within expectation of values that typically range from.10 to .20 in
school effect research (McCoach, 2010). Academic Expectations (β = −0.45, p b .05), Respect for Students (β = − 0.76, p b .05),
and Willingness to Seek Help (β = 0.62, p b .05) were all found to be statistically related to PTB; Discipline (β = 0.26, p b .05) and
Respect for Students (β = 0.27, p b .05) were statistically significant predictors of Affective Engagement; and Willingness to Seek
Help (β = 0.65, p b .05) was statistically associated with Cognitive Engagement, see Table 3.
Student-level criterion-related validity coefficients in the form of standardized structure coefficients are shown in parentheses in
Table 3. It is important to note that because the sampling distributions of the within (i.e., student) and between (i.e., school) portions
of the models are different, comparisons of path coefficients magnitudes across levels are not meaningful (Byrne, 2012). That said, the
four school climate variables of Disciplinary Structure (DS), Academic Expectations (AE), Respect for Students (RS), and Willingness to
Seek Help (WS) were all statistically associated (p's b .05) with the Prevalence of Teasing and Bullying (βDS = −0.35, βAE = 0.13,
βRS = − 0.16, and βWS = − 0.06), Affective Engagement (βDS = 0.32, βAE = 0.03, βRS = 0.10, and βWS = 0.28), and Cognitive
Engagement (βDS = 0.21, βAE = 0.34, βRS = −0.13, and βWS = 0.24). In addition, the four school climate variables accounted for
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 457

20% of the student level variance in PTB, and 41% and 35% of the variance in Affective and Cognitive Engagement, respectively. These
results suggest that student level dispositions on the two Structure and two Support factors are meaningfully related to their percep-
tions of Prevalence of Teasing and Bullying, Affective Engagement, and Cognitive Engagement.

5. Discussion

A conceptual model of authoritative school climate was empirically investigated to evaluate the hypothesis that higher levels of
structure and support variables were associated with greater levels of engagement and lower levels of peer aggression in a sample
of high school students. As a preliminary step in this examination, the first research question addressed whether the factor structure
of items on the ASCS maintained for high school students and their schools. Multilevel procedures were employed because students
were nested within schools and because we were interested in determining whether survey items were appropriate for measuring
their intended constructs for both students and schools (Muthén, 1991).
With regard to the first research question, multilevel factor analyses demonstrated a stable factor structure for the ASCS on both
student and school levels of analysis. This finding is noteworthy because many instruments developed to measure individual student
perceptions do not necessarily maintain their psychometric properties when aggregated to school level scores (Muthén, 1991). The
present study extends prior work on middle schools (Konold et al., 2014) to a large and demographically diverse sample of high
schools, and supports use of the survey scales both to analyze student perceptions of school climate within schools and to compare
schools using aggregated scores that represent the school. The development of such instruments is important because of the increased
national interest in the assessment of school climate to monitor conditions within a school and to study differences across schools in
order to identify effective practices (American Institutes for Research, 2013; Thapa et al., 2013).
To answer the second research question, the multi-level structural analysis (see Fig. 1) provided new evidence of convergent
validity for the constructs of structure and support that are the cornerstone of the authoritative school climate model. School climates
characterized by high structure and support are expected to have students who are more engaged in school and treat one another
with respect, resulting in less teasing and bullying. In the present study, schools where students reported fair disciplinary practices
tended to have higher levels of student affective engagement characterized by feelings of belonging and pride in their school.
These schools also showed lower levels of teasing and bullying of peers, results that are consistent with prior studies of authoritative
school climate and peer aggression in younger students (Cornell, Shukla & Konold, in press; Gregory et al., 2010). Other studies using
different measures of authoritative school climate also found higher engagement (Lee, 2012) and less bullying (Gerlinger & Wo,
2014). This body of literature extends the authoritative model of parenting (Larzelere et al., 2013) to the broader social context of
the school as an environment intended to enhance youth development.
Disciplinary structure is an important construct in the context of national interest in the fairness of school disciplinary practices,
emphasized in a Dear Colleague letter from the U.S. Department of Justice and U.S. Department of Education (2014) directing schools
to review their disciplinary practices for potential biases that result in disproportionate suspension of minority students. Previous
research found that high disciplinary structure was associated with lower suspension rates for both Black and White high school
students (Gregory et al., 2011). This finding is consistent with Pellerin's (2005) report that high schools using an authoritarian
approach to school discipline had less truancy and fewer dropouts, although additional studies examining suspension rates are
needed. In addition, research on strategies to reduce the disproportionately high rate of suspensions among minority students has
recommended that schools should concentrate on improving student academic engagement (Toldson, McGee, & Lemmons, 2015),
using restorative principles that bring a sense of justice and fairness to disciplinary practices (González, 2015), and coaching teachers
to make their interactions with students more sensitive and supportive (Gregory et al., 2015). All of these recommendations reflect
dimensions of school climate tapped by an authoritative model.
From a theoretical perspective, the high structure of an authoritative school climate includes high academic expectations as well as
strict but fair discipline (Gregory & Cornell, 2009). Accordingly, the present study included a new scale to assess academic expecta-
tions. Factor analyses of this scale revealed good psychometric properties and it correlated .71 with disciplinary structure at the school
level. Schools with high academic expectations had lower levels of teasing and bullying among students at both the student and school
levels. High academic expectations were associated with higher affective and cognitive engagement at the student level, but not at the
school level. These findings are consistent with results of two previous studies with nationally representative samples which found
that high academic press was associated with higher student engagement (Gill et al., 2004; Lee, 2012). The current findings in support
of high academic expectations are especially noteworthy because these analyses controlled for the school demographic variables of
student poverty and race/ethnicity that are often associated with lower school academic performance (Gottfredson et al., 2005).
Authoritative schools also show high student support, which was measured by two scales, Respect for Students and Willingness to
Seek Help that have been previously combined into a single measure (Gregory et al., 2010). The two scales were strongly correlated in
the current study (.68 at the student level and .90 at the school level), and both were negatively associated with the prevalence of
teasing and bullying as expected. However, in the structural analysis, Respect for Students was negatively associated with teasing
and bullying while Willingness to Seek Help was unexpectedly positively associated with teasing and bullying. One possibility is
that, after the influence of other variables has been removed, there is a unique correlation between Willingness to Seek Help and
prevalence of teasing and bullying among students who have observed peer aggression and are willing to seek help for it from
their teachers. There also may be schools with active anti-bullying programs where there is a climate of encouraging help-seeking
and increased student sensitivity to recognizing teasing and bullying.
Both measures of student support were associated with higher affective and cognitive engagement at the student level. Findings
for school level engagement were mixed, with Respect for Students predictive of higher affective engagement but not cognitive
458 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

engagement, while Willingness to Seek Help was associated with higher cognitive engagement but not affective engagement. The
inconsistent findings for the two engagement measures may reflect some weakness in the factor structure of this construct at the
school level, an observation noted by Konold et al. (2014). The concept of student engagement, as operationalized in these scales,
may be better conceptualized as an individual characteristic than a group (school) characteristic.
Overall, the structural analysis showed support for a model of authoritative schools as indicative of a positive school climate where
students are more engaged in school and less involved in peer aggression in the form of teasing and bullying one another. These
relations were maintained in a model that controlled for school demographic variables which themselves accounted for considerable
variability in student engagement and peer aggression. For example, although the school demographic variables of ethnic composition,
student poverty, and school size accounted for 20% of the variance in the prevalence of teasing and bullying, the four measures of
authoritative school climate accounted for an additional 61% of the school level variance.

5.1. Study limitations and directions for future study

This study examined structure and support variables as two key components of an authoritative school climate, but there are other
aspects of school climate to consider. The U. S. Department of Education's Safe and Supportive Schools model of school climate has 13
components organized into domains of engagement, safety, and environment (Bradshaw, Waasdorp, Debnam, & Johnson, 2014). How
these components map onto the scales used in this study, as well as how these components are related to one another should be a
direction for future research. Furthermore, a comprehensive assessment of school climate should consider whether there are
differential perceptions and relations among different student subgroups as a function of gender, race, social background, grade
level, and other factors (Shirley & Cornell, 2012).
Correlational findings cannot establish causal relationships and there may be bidirectional causal effects to consider. For example,
highly engaged students may stimulate positive teacher-student relations and conversely, high levels of student teasing and bullying
could adversely affect school disciplinary practices. One way to test for causal effects is to conduct interventions that improve school
discipline practices or teacher-student relationships. It may be useful to measures authoritative school climate characteristics before
and after implementation of schoolwide improvement efforts. For example, Flannery, Fenning, Kato, and McIntosh (2014) reported
that implementation of School-Wide Positive Behavioral Interventions and Supports (SWPBIS) in eight high schools resulted in
decreased office referrals. It would be informative to assess whether changes in student perceptions of the school climate, especially
the fairness of school discipline and supportiveness of teacher-student relationships, preceded the decreases in office referrals.
Nevertheless, the findings from this study were consistent with our hypothesized model of authoritative school climate by
demonstrating a statistical effect of school structure and support measures on student engagement and peer aggression, after controlling
for known demographic risk factors and using a model that considered the nesting of students within schools. Few studies have
conducted a multivariate multilevel analysis that can provide such a robust and comprehensive analysis.
Another study limitation is that the primary school climate measures were based on student self-report, and it would be useful to
incorporate independent observations such as teacher reports. Several studies have used principal reports, but these are limited to a
single observer and have not been compared to other informants (Gill et al., 2004; Pellerin, 2005). There is promising evidence that
the teacher version of the Authoritative School Climate Survey yields similar findings to the student version in middle schools
(Huang et al., in press), but a high school comparison is needed. In addition, the study did not have a provision for confirming that
the principals followed the random selection procedure and other survey instructions. This would be a desirable, but logistically
difficult, improvement, but it can be noted that the high participation rate (323 of 324 high schools in the state) and the positive
feedback from school principals in a follow-up survey provides evidence of cooperation and engagement.
These findings were based on high schools in a single state. Although it was a strength of the study to include nearly all public high
schools in a large and demographically diverse state, investigations in other states are needed.
Finally, it is also important to examine differences in perceptions of school climate associated with race, ethnicity, and gender.
Most of the research in this area concerns peer victimization, but some evidence that students from racial/ethnic minority groups
might have a less positive view of school climate than majority group classmates (Shirley & Cornell, 2012). There is mixed evidence
regarding whether racial or ethnic minority students experience more bullying than non-minority students (Fisher et al., 2014;
Goldweber, Waasdorp, & Bradshaw, 2013), but a student's race or ethnicity may interact with the racial and ethnic diversity of the
student body. For example, studies have found that having a greater number of students of the same ethnicity was associated with
less victimization (Felix & You, 2011; Juvonen & Graham, 2014). There may be important gender differences that interact with race
or ethnicity (Fisher et al., Juvonen & Graham, 2014; Mehta et al., 2013). Although boys have greater involvement than girls in physical
bullying, involvement in other forms of bullying are similar for boys and girls (Juvonen & Graham, 2014).

5.2. Implications for practice

The Authoritative School Climate Survey provides school psychologists and other educators with a tool for the assessment of
school climate at both student and school levels of analysis. This avoids the potential errors and problems of measurement that
may exist when an instrument developed from student-level analyses is used to infer school-level characteristics, and vice versa.
Users can also be confident that the survey can be used in high school as well as middle school populations (see Konold et al.,
2014), since instruments may be developed in one population but not examined for validity in the other.
The ASCS provides a more theoretically coherent assessment of school conditions that recognizes the importance of both high
expectations for students and supportive teacher-student relationships. Especially noteworthy is that high expectations include
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 459

high academic expectations and high disciplinary structure, and that both domains are associated with greater student engagement in
school. Based on findings from this study, these relationships appear to hold regardless of school size, percentage of minority students,
or percentage of low income students.
The Prevalence of Teasing and Bullying Scale provides educators with an alternative to self-reported victimization as a way to
assess common forms of peer aggression (Cornell et al., 2013). By asking students to report on their observations of teasing and
bullying among their peers, the survey avoids some of the limitations that arise when students are asked to identify and label
themselves as victims (Cornell et al., 2013). It is noteworthy that schools characterized by high expectations and high support
show consistently lower levels of peer aggression as measured by this scale. Moreover, previous studies have found that lower
Prevalence of Teasing and Bullying is associated with important school outcomes, including higher student engagement (Mehta
et al., 2013), higher schoolwide performance on standardized achievement testing (Lacey & Cornell, 2013), and lower dropout
rates (Cornell et al., 2013).

6. Conclusions

The U.S. Department of Education has placed increasing emphasis on the need for schools to monitor their school climate in order
improve conditions for learning and assure a safe and supportive environment for all students (U.S. Department of Education, 2014;
U.S. Department of Justice and U.S. Department of Education, 2014). There has been a proliferation of school climate measures, but the
field lacks a conceptual framework for a positive school climate (Dary & Pickeral, 2013, U.S. Department of Education, 2014). In
addition, many school climate measures have not been developed to measure constructs on both student and school levels of analysis.
The authoritative school climate model identifies two key elements of school climate, structure and support, that have been associated
with positive student behavior and favorable school conditions across a series of studies (Gerlinger & Wo, 2014; Gill et al., 2004;
Gregory et al., 2010; 2011; 2012; Pellerin, 2005). The present study extended the authoritative school climate model to a large sample
of high schools and found that measures of structure and support were associated with higher student engagement and lower peer
aggression. In addition, findings from this study provide measurement tools for use in high school settings that could be valuable
to school psychologists, administrators, and teachers for purposes of gauging student- and school-level elements of school climate.

References

American Educational Research Association, American Psychological Association, & National Council of Measurement in Education (2014c). Standards for Educational
and Psychological Testing. Washington, DC: AERA.
American Institutes for Research (2013). School Climate Survey Compendia. 2013 Retrieved from http://safesupportivelearning.ed.gov/topic-research/school-climate-
measurement/school-climate-survey-compendium
Austin, G., & Duerr, M. (2005). Guidebook for the California Healthy Kids Survey. Part III School Climate Survey for teachers and other staff (2005-2006 edition ). WestEd
Retrieved January 5, 2006, from http://www.wested.org/pub/docs/chks_surveys_summary.html.
Bandyopadhyay, S., Cornell, D., & Konold, T. (2009). Internal and external validity of three school climate scales from the School Climate Bullying Survey. School
Psychology Review, 38, 338–355.
Barile, J. P., Donohue, D. K., Anthony, E. R., Baker, A. M., Weaver, S. R., & Henrich, C. C. (2012). Teacher-student relationship climate and school outcomes: Implications
for educational policy initiatives. Journal of Youth & Adolescence, 41, 256–267.
Baumrind, D. (1968). Authoritarian vs. authoritative parental control. Adolescence, 3, 255–272.
Bear, G. G., Gaskins, C., Blank, J., & Chen, F. F. (2011). Delaware School Climate Survey-Student: its factor structure, concurrent validity, and reliability. Journal of School
Psychology, 49, 157–174. http://dx.doi.org/10.1016/j.jsp.2011.01.001.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness-of-fit in the analysis of covariance structures. Psychological Bulletin, 88, 588–606.
Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: implications for data aggregation and analysis. In K. J. Klein, & W. Kozlowski (Eds.),
Multilevel Theory, Research, And Methods In Organizations: Foundations, Extensions And New Directions (pp. 349–381). San Francisco, CA: Jossey-Bass.
Bradshaw, C. P., Waasdrop, T. E., Debnam, K. J., & Johnson, S. L. (2014). Measuring school climate in high schools: a focus on safety, engagement, and the environment.
Journal of School Health, 84, 593–604.
Bradshaw, C. P., Koth, C. W., Thornton, L. A., & Leaf, P. J. (2009). Altering school climate through school-wide positive behavioral interventions and supports: findings
from a group-randomized effectiveness trial. Prevention Science, 10, 100–115.
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.
Browne, M., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. Bollen, & J. Long (Eds.), Testing Structural Equation Models (pp. 136–162). Newbury Park,
CA: Sage.
Byrne, B. M. (2012). Structural equation modeling with MPlus: Basic concepts, applications, and programming. New York, NY: Routledge.
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255.
Cohen, J., McCabe, L., Michelli, N. M., & Pickeral, T. (2009). School climate: research, policy, practice, & teacher education. Teachers College Record, 111, 180–213.
Collie, R. J., Shapka, J. D., & Perry, N. E. (2012). School climate and social-emotional learning: Predicting teacher stress, job satisfaction, and teaching efficacy. Journal of
Educational Psychology, 104, 1189–1204.
Cornell, D. (2015). The Authoritative School Climate Survey and the School Climate Bullying Survey: Description and research summary. Charlottesville, Virginia: Curry
School of Education, University of Virginia.
Cornell, D., Gregory, A., Huang, F., & Fan, X. (2013). Perceived prevalence of teasing and bullying predicts high school dropout rates. Journal of Educational Psychology,
105, 138–149.
Cornell, D., Shukla, K., & Konold, T. (2015). Peer victimization and authoritative school climate: a multilevel multivariate approach. Journal of Educational Psychology in
press.
Dary, T. & Pickeral, T. (Eds.) (2013). School Climate: Practices for Implementation and Sustainability. A School Climate Practice Brief, Number 1. New York, NY: National
School Climate Center.
Dedrick, R. F., & Greenbaum, P. E. (2011). Multilevel confirmatory factor analysis of a scale measuring interagency collaboration of children's mental health agencies.
Journal of Emotional and Behavioral Disorders, 19, 27–40. http://dx.doi.org/10.1177/1063426610365879.
Dyer, N. G., Hanges, P. J., & Hall, R. J. (2005). Applying multilevel confirmatory factor analysis techniques to the study of leadership. Leadership Quarterly, 16, 149–167.
Espelage, D. L., Low, S. K., & Jimerson, S. R. (2014). Understanding school climate, aggression, peer victimization, and bully perpetration: contemporary science,
practice, and policy. School Psychology Quarterly, 29, 233–237.
Felix, E. D., & You, S. (2011). Peer victimization within the ethnic context of high school. Journal of Community Psychology, 39, 860–875.
460 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461

Fisher, S., Middleton, K., Ricks, E., Malone, C., Briggs, C., & Barnes, J. (2014). Not just black and white: peer victimization and the intersectionality of school diversity and
race. Journal of Youth and Adolescence, 44, 1–10.
Flannery, K. B., Fenning, P., Kato, M. M., & McIntosh, K. (2014). Effects of school-wide positive behavioral interventions and supports and fidelity of implementation on
problem behavior in high schools. School Psychology Quarterly, 29, 111–124.
Gerbing, D., & Anderson, J. (1992). Monte Carlo evaluations of goodness of fit indices for structural equation models. Sociological Methods and Research, 21, 132–160.
Gerlinger, J., & Wo, J. (2014). Preventing school bullying: should schools prioritize an authoritative school discipline approach over security measures? Journal of School
Violence. http://dx.doi.org/10.1080/15388220.2014.956321.
Gill, M. G., Ashton, P., & Algina, J. (2004). Authoritative schools: a test of a model to resolve the school effectiveness debate. Contemporary Educational Psychology, 29,
389–409.
Goldweber, A., Waasdorp, T. E., & Bradshaw, C. P. (2013). Examining associations between race, urbanicity, and patterns of bullying involvement. Journal of Youth and
Adolescence, 42, 206–219.
González, T. (2015). Socializing schools: addressing racial disparities in discipline through restorative justice. In D. J. Losen (Ed.), Closing the School Discipline Gap
(pp. 151–165). New York, NY: Teachers College Press.
Gottfredson, G. D. (1999). User's manual for the Effective School Battery. Ellicott City, MD: Gottfredson Associates.
Gottfredson, G. D., Gottfredson, D. C., Payne, A. A., & Gottfredson, N. C. (2005). School climate predictors of school disorder: results from a national study of delinquency
prevention in schools. Journal of Research in Crime and Delinquency, 42, 412–444. http://dx.doi.org/10.1177/0022427804271931.
Gregory, A., Allen, J. P., Mikami, A. Y., Hafen, C. A., & Pianta, R. C. (2015). The promise of a teacher professional development program in reducing racial disparity in
classroom exclusionary discipline. In D. J. Losen (Ed.), Closing the School Discipline Gap (pp. 166–179). New York, NY: Teachers College Press.
Gregory, A., & Cornell, D. (2009). “Tolerating” adolescent needs: moving away from Zero Tolerance policies in high school. Theory Into Practice, 48, 106–113.
Gregory, A., Cornell, D., & Fan, X. (2011). The relationship of school structure and support to suspension rates for Black and White high school students. American
Educational Research Journal, 48, 904–934. http://dx.doi.org/10.3201/002831211398531.
Gregory, A., Cornell, D., & Fan, X. (2012). Teacher safety and authoritative school climate in high schools. American Journal of Education, 118, 401–425.
Gregory, A., Cornell, D., Fan, X., Sheras, P., Shih, T., & Huang, F. (2010). Authoritative school discipline: high school practices associated with lower student bullying and
victimization. Journal of Educational Psychology, 102, 483–496.
Hu, L., & Bentler, P. M. (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural Equation Modeling: Concepts, Issues, and Applications (pp. 76–99). Thousand Oaks, CA: Sage.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structual Equation Modeling,
6, 1–55. http://dx.doi.org/10.1080/10705519909540118.
Huang, F., Cornell, D., Konold, T., Meyer, P., Lacey, A., Nekvasil, E., ... Shukla, K. (2015). Multilevel factor structure and concurrent validity of the teacher version of the
Authoritative School Climate Survey. Journal of School Health (in press).
Hung, A. H., Luebbe, A. M., & Flaspphler, P. D. (2014). Measuring school climate: factor analysis and relations to emotional problems, conduct problems, and victimization in
middle school students. School Mental Health, Advance online publication. http://dx.doi.org/10.1007/s12310-014–9131-y http://download.springer.com/static/pdf/
265/art%253A10.1007%252Fs12310-014-9131-y.pdf?auth66=1414801331_1f89d25a60412139b684fc5f60a5bd85&ext=.pdf.
Juvonen, J., & Graham, S. (2014). Bullying in schools: the power of bullies and the plight of victims. Annual Review of Psychology, 65, 159–185.
Juvonen, J., Nishina, A., & Graham, S. (2000). Peer harassment, psychological adjustment, and school functioning in early adolescence. Journal of Educational Psychology,
92, 349–359.
Keith, T. Z. (1997). Using confirmatory factor analysis to aid in understanding the constructs measured by intelligence tests. In D. P. Flanagan, J. L. Genshaft, & P. L.
Harrison (Eds.), Contemporary Intellectual Assessment: Theories, Tests, and Issues (pp. 373–402). New York, NY: Guilford Press.
Kidiger, J., Araya, R., Donovan, J., & Gunnell, D. (2012). The effect of the school environment on the emotional health of adolescents: a systematic review. Pediatrics, 129,
925–949. http://dx.doi.org/10.1542/peds.2011-2248.
Konold, T. R., Cornell, D., Huang, F., Meyer, P., Lacey, A., Nekvasil, E., ... Shukla, K. (2014). Multi-level Multi-informant Structure of the Authoritative School Climate
Survey. School Psychology Quarterly, 29, 238–255.
Klein, J., Cornell, D., & Konold, T. (2012). Relationships between bullying, school climate, and student risk behaviors. School Psychology Quarterly, 27, 154–169.
Lacey, A., & Cornell, D. (2013). The impact of teasing and bullying on schoolwide academic performance. Journal of Applied School Psychology, 29, 262–283.
Larzelere, R. E., Morris, A. S., & Harrist, A. W. (2013). Authoritative Parenting: Synthesizing Nurturance and Discipline for Optimal Child Development. Washington, DC:
American Psychological Association.
Lee, J. S. (2012). The effects of the teacher-student relationship and academic press on student engagement and academic performance. International Journal of
Educational Research, 53, 330–340.
Leithwood, K., & Jantzi, D. (2009). A review of empirical evidence about school size effects: a policy perspective. Review of Educational Research, 79, 464–490. http://dx.
doi.org/10.3102/0034654308326158.
Low, S., & Van Ryzin, M. (2014). The moderating effects of school climate on bullying prevention efforts. School Psychology Quarterly, 29, 306–319.
McCoach, B. D. (2010). Hierarchical linear modeling. In G. R. Hancock, & R. O. Mueller (Eds.), The Reviewer's Guide to Quantitative Methods in the Social Sciences
(pp. 123–140). New York, NY: Routledge.
Mehta, S., Cornell, D., Fan, X., & Gregory, A. (2013). Bullying climate and school engagement in ninth grade students. Journal of School Health, 83, 45–54.
Midgley, C., Maehr, M., Hruda, L., Anderman, E., & Freeman, K. (2000). Manual for Patterns of Adaptive Learning. Ann Arbor, MI: University of Michigan.
Morgan, E., Salomen, N., Plotkin, M., & Cohen, R. (2014). The School Discipline Consensus Report: Strategies from the Field to Keep Students Engaged In School and Out of the
Juvenile Justice System. The Council of State Governments Justice Center. New York: NY Available from: http://csgjusticecenter.org/wp-content/uploads/2014/06/The_
School_Discipline_Consensus_Report.pdf.
Muthén, B. O. (1991). Multilevel factor analysis of class and student achievement components. Journal of Educational Measurement, 28, 338–354.
National Center for Education Statistics (2005). School crime supplement to the national crime victimization survey. , 2005 Retrieved from http://nces.ed.gov/Programs/
Crime/surveys.asp.
O'Connell, A. A., & McCoach, B. D. (2008). Multilevel Modeling of Educational Data. Charlotte, NC: Information Age Publishing.
Olweus, D. (2007). The Olweus Bullying Questionnaire. Center City, MN: Hazelden.
Pellerin, L. A. (2005). Applying Baumrind's parenting typology to high schools: toward a middle range theory of authoritative socialization. Social Science Research, 34,
283–303.
Piscatelli, J., & Lee, C. (2011). State Policies on School Climate and Bully Prevention Efforts: Challenges and Opportunities for Deepening State Policy Support for Safe and Civil
Schools. National School Climate Center Retrieved from http://schoolclimate.org/climate/policy.php.
Peugh, J. L. (2010). A practical guide to multilevel modeling. Journal of School Psychology, 85-112.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical Linear Models: Applications and Data Analysis Methods (2nd ed.). Thousand Oaks, CA: Sage.
Raudenbush, S. W., Rowan, B., & Kang, S. J. (1991). A multilevel, multivariate model for studying school climate with estimation via the EM algorithm and application to
U.S. high school data. Journal of Educational Statistics, 16, 295–330.
Schumacker, R. E., & Lomax, R. G. (2010). A Beginner's Guide to Structural Equation Modeling (3rd ed.). New York: Routledge.
Shirley, E., & Cornell, D. (2012). The contribution of student perceptions of school climate to understanding the disproportionate punishment of African American
students in a middle school. School Psychology International, 33, 115–134.
Thapa, A. S., Cohen, J., Guffey, S., & Higgins-D'Alessandro, A. (2013). A review of school climate research. Review of Educational Research, 83, 357–385. http://dx.doi.org/
10.3102/0034654313483907.
Thornberry, T. P., Lizotte, A. J., Krohn, M. D., Farnworth, M., & Jang, S. J. (1991). Testing interactional theory: an examination of reciprocal causal relationships among
family, school, and delinquency. J Crim Law & Criminology, 82, 3–35.
Tippett, N., & Wolke, D. (2014). Socioeconomic status and bullying: a meta-analysis. American Journal of Public Health., 104, e48–e59. http://dx.doi.org/10.2105/AJPH.
2014.301960.
 T.R. Konold, D. Cornell / Journal of School Psychology 53 (2015) 447–461 461

Toldson, I. A., McGee, T., & Lemmons, B. P. (2015). Reducing suspension by improving academic engagement among school-age black males. In D. J. Losen (Ed.), Closing
the School Discipline gap (pp. 107–117). New York, NY: Teachers College Press.
U.S. Department of Education (2014). Directory of federal school climate and discipline resources. D.C.: Washington Retrieved from http://www2.ed.gov/policy/gen/guid/
school-discipline/appendix-1-directory.pdf.
U.S. Department of Justice and U.S. Department of Education (2014). Dear colleague letter on nondiscriminatory administration of school discipline. DC: Washington
Retrieved from http://www2.ed.gov/about/offices/list/ocr/letters/colleague-201401-title-vi.pdf.
Virginia Department of Education (n.d.). Statistics and reports. Retrieved from http://www.doe.virginia.gov/statistics_reports/index.shtml
Wheelock, A., & Miao, J. (2005). The ninth-grade bottleneck: an enrollment bulge in a transition year that demands careful attention and action. School Administrator,
62, 36–38.
Yeager, D. S., Fong, C. J., Lee, H. Y., & Espelage, D. L. (2015). Declines in efficacy of anti-bullying programs among older adolescents: theory and a three-level
meta-analysis. Journal of Applied Developmental Psychology, 37, 36–51.