4/2/23

Authoritative School Climate and Sexual Harassment: A Cross-Sectional Multilevel Analysis of Student Self-Reports

Brittany Z. Crowley, Pooja Datta, Shelby Stohlman, Dewey Cornell, and Tim Konold University of Virginia

School sexual harassment (SH) is defined as unwelcome behavior of a sexual nature that interferes with a student’s ability to learn. There is an important need for schools to assess the prevalence of SH and its relation to school climate to guide intervention efforts. This study investigated 3 research questions: (a) Is there psychometric support for a 4-item multilevel measure of SH? (b) What is the prevalence of SH in a statewide high school sample, and how does SH vary across gender, grade level, race–ethnicity, and socioeconomic status? (c) Is an authoritative school climate—characterized by strict but fair discipline and supportive teacher–student relationships—associated with lower levels of SH for students? A statewide sample of high school students (N � 62,679) completed a school climate survey that included a new 4-item measure of SH. Results of a multilevel confirmatory factor analysis indicated good fit for a single SH factor at both student and school levels. A multiway analysis of variance demonstrated the high prevalence of SH and variations across demographic groups. Multilevel hierarchical regression analyses indicated that an authoritative school climate accounted for 5.7% of the student-level variance and 38.3% of the school-level variance in SH scores. Routine assessment of SH can help school psychologists bring attention to this underrecognized problem.

Impact and Implications Sexual harassment is an important national concern that has not been adequately recognized in high schools. Approximately 38% of high school students in a large statewide sample reported at least one incident of sexual harassment in the past year, with differences across gender, grade level, race– ethnicity, and socioeconomic status. School psychologists can bring attention to the problem and advocate routine assessment of sexual harassment with a four-item scale. An authoritative school climate is associated with lower levels of sexual harassment.

Keywords: sexual harassment, measurement, authoritative school climate, disciplinary structure, student support

School sexual harassment (SH) is defined by the U.S. Depart- ment of Education (USDOE) as any unwelcome behavior of a sexual nature that interferes with a student’s ability to learn, study, work, or participate in school activities. SH can be verbal (e.g., unwanted sexual humor, homophobic slurs), nonverbal (e.g., un- wanted written sexual communications, gestures), or physical

(e.g., unwanted sexual touching, kissing; USDOE, Office for Civil Rights, 2008). Evidence suggests that SH is prevalent in schools and associated with negative outcomes of student health and well-being (Eom, Restaino, Perkins, Neveln, & Harrington, 2015; Gruber & Fineran, 2016). Schools have a legal obligation to stop SH when it is severe enough to interfere with learning (Title IX, 1972).

School psychologists should encourage routine assessment of the prevalence of SH in their schools. Such assessment should meet high psychometric standards yet be brief enough that it can be incorporated into routine school climate measures. School cli- mate is broadly defined as the quality of school life that reflects the norms, goals, values, and practices of a school (Cohen, McCabe, Michelli, & Pickeral, 2009). Extensive research has demonstrated that school climate has a substantial association with various types of peer victimization (Bradshaw, 2015; Guerra, Williams, & Sadek, 2011). Although there is some evidence supporting the association between school climate and SH victimization, further study is needed to understand the nature and prevalence of SH across students and schools of varying demographics (Gruber & Fineran, 2016). The focus of this study is on peer-to-peer SH in

This article was published Online First December 27, 2018. Brittany Z. Crowley, Pooja Datta, Shelby Stohlman, Dewey Cornell, and

Tim Konold, Curry School of Education, University of Virginia. This project was supported by Grants NIJ 2014-CK-BX-0004 and NIJ

2017-CK-BX-007 awarded to the University of Virginia by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and recommendations expressed in this report are those of the researchers at the Youth Violence Project of the University of Virginia and do not necessarily reflect those of the U.S. Department of Justice or the Center for School and Campus Safety at the Virginia Department of Criminal Justice Services.

Correspondence concerning this article should be addressed to Brittany Z. Crowley, Curry School of Education, University of Virginia, 205 Ruffner Hall, Charlottesville, VA 22903. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

School Psychology © 2018 American Psychological Association 2019, Vol. 34, No. 5, 469–478 1045-3830/19/$12.00 http://dx.doi.org/10.1037/spq0000303

469

schools. Student-to-adult and adult-to-student SH, although impor- tant, are beyond the scope of this study.

Measuring School Sexual Harassment

In the most recent national survey of peer-to-peer SH in schools, the American Association of University Women (AAUW, 2011) found that 48% of adolescents experienced one of 10 types of SH in the school year. These rates differ from those reported in smaller scale, regional samples, such as a study by Eom and colleagues that found that 59% had experienced one of six different types of SH over the past year (Eom et al., 2015).

It is typical for studies of school SH to survey students about how often they have experienced verbal SH, physical SH, and spreading of sexual rumors. For example, the AAUW (2011) survey asked students about 10 different harassment behaviors, such as (a) having someone make unwelcome sexual comments, jokes, or gestures in person; (b) being touched in an unwelcome sexual way; and (c) having someone spread unwelcome sexual rumors about you. They found that students were more likely to report experiencing nonphysical harassment (33%) than being touched in an unwelcome sexual away (8%; AAUW, 2011).

Demographic Differences

There is some evidence that rates of SH differ across gender, age, race–ethnicity, and socioeconomic status (SES). Studies on the prevalence of SH have consistently reported that girls experi- ence more SH than do boys (AAUW, 1993, 2001, 2011). For example, in the AAUW (2011) study, 56% of girls reported experiencing SH, compared to 40% of boys.

With respect to age, studies have generally suggested that school SH begins in early adolescence and increases through adolescence. At the high school level, students in all grades report SH experi- ences, although studies have been mixed on grade trends (AAUW, 2001, 2011).

Racial–ethnic differences in prevalence were also noted in some studies (AAUW, 2001, 2011; Espelage, Hong, Rinehart, & Doshi, 2016). In 2001, the AAUW found that African American girls were more likely to report unwelcome touching (67%) than were Hispanic and White girls (51% for each; AAUW, 2001). However, in 2011 the AAUW found no significant differences in prevalence among racial–ethnic groups, though this may have been due to having few African American and Hispanic students in their sam- ple (AAUW, 2011). Additional research is needed to detect po- tential racial–ethnic differences in SH victimization.

In 2011, the AAUW found that students from high- versus low-income households were equally likely to report most types of SH, although lower income students were more likely to report unwelcome touching (AAUW, 2011).

School Sexual Harassment Measures

The majority of studies of school SH used items from the AAUW national surveys. The 2001 AAUW study contained 14 items, and the 2011 study contained 10 items (AAUW, 2001, 2011). The AAUW reports provide no information about psycho- metric properties of their measures, though there is some psycho- metric information reported in other studies that use these items.

For example, Espelage, Basile, and Hamburger (2012) examined psychometric properties of the 14-item AAUW (2001) survey in a sample of middle school students. The scale was analyzed using exploratory factor analysis, and results suggested the existence of a two-factor solution. Factor 1 (sexual harassment) contained nine items (e.g., making sexual comments, spreading rumors, and pull- ing at clothing of another student) and demonstrated internal consistency (� � .81). Factor 2 (forced sexual contact) contained three items (i.e., forcing someone to kiss you, forcing someone to do something sexual besides kissing, and forcing someone to touch your private parts) and demonstrated internal consistency (� � .73). Items that cross-loaded or had loadings lower than .30 on their primary factor loading were deleted. Gruber and Fineran (2016) reported a similar factor analysis of the AAUW (2011) items but with somewhat different results. The 14 sexual harass- ment items were factor-analyzed using varimax rotation, and re- sults suggested the existence of a one-factor solution (sexual harassment) with high reliability of items (� � .90). These studies examined SH at the individual level but did not test it at the school level.

Studies using different items or fewer items reported more limited psychometric results. For example, Clear and colleagues (2014) surveyed high school students using three items from the Sexual Experiences Questionnaire (SEQ; Fitzgerald, Magley, Drasgow, & Waldo, 1999). Although the full-length SEQ (versions range from 23 to 26 items) demonstrated good psychometric properties (Fitzgerald et al., 1999; Ormerod, Collinsworth, & Perry, 2008), the only psychometric information reported for the three items used by Clear and colleagues was moderate internal consistency (� � .66; Clear et al., 2014). In conclusion, there is a need for a brief SH scale that can be used at both student and school levels with good psychometric properties.

School Climate and Sexual Harassment

School climate is broadly defined as the quality of school life that reflects the norms, goals, values, and practices of a school (Cohen et al., 2009). A positive school climate is commonly acknowledged as a factor influencing school victimization, but there is little agreement on the composition or dimensional struc- ture of school climate (Cohen et al., 2009; Cornell & Mayer, 2010). For example, past studies of the association between school climate and SH have operationalized positive school climate as school satisfaction and student academic performance (Gruber & Fineran, 2016); presence of school supports (Kosciw, Palmer, Kull, & Greytak, 2013); perceptions that one’s school climate is not tolerant of SH (Ormerod et al., 2008); and positive perceptions of school policy, teacher support, and student participation in decisions (Attar-Schwartz, 2009).

Some studies have reported associations between SH experi- ences and school climate at the individual level but not at the school level. For example, Gruber and Fineran (2016) found that SH experiences were associated with lower satisfaction with school and academic performance. Similarly, Kosciw et al. (2013) found that in a sample of sexual-minority adolescents, SH was associated with poorer academic outcomes and psychological well-being, but inclusion of school-based supports was associated with lower victimization rates and improved outcomes.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

470 CROWLEY, DATTA, STOHLMAN, CORNELL, AND KONOLD

Some studies have demonstrated associations between SH and school climate factors at the individual and school levels. For example, Ormerod et al. (2008) examined peer SH victimization, school climate, and student outcomes in a sample of girls and boys from seven high schools. For girls, at the individual level SH was associated with greater psychological distress, negative body im- age, and poorer perceptions of school safety. Similarly, percep- tions that one’s school climate is tolerant of SH (poor school climate) were associated with more negative body image, poorer perceptions of safety, and greater withdrawal from school. For boys, at the individual level SH was associated with negative body image. Poor climate perceptions were associated with greater psychological distress and withdrawal from school and poorer self-esteem, health satisfaction, and perceptions of safety. At the school level, student perception of school climate was a significant predictor of increased SH (Ormerod et al., 2008).

Similarly, Attar-Schwartz (2009) examined peer SH victimiza- tion and school climate perceptions in a sample of 7th- through 11th-grade students in Israel. School climate was measured by three scales addressing perceptions of school policy, teacher sup- port, and student participation in decision-making. At the individ- ual level, student perceptions of school climate explained 5.8% of the variance in reports of SH victimization, such that students with greater perceptions of school climate reported less SH victimiza- tion. At the school level, school climate factors explained an additional 5% of variance in school SH rates (Intraclass Correla- tion Coefficient [ICC] � 0.07). Schools with greater average teacher support and school policy ratings had lower average SH victimization rates (Attar-Schwartz, 2009).

Authoritative School Climate Theory

Authoritative school climate (ASC) theory proposes two critical components of positive school climate: high disciplinary structure (i.e., perception that rules are strict but fair) and high student support (i.e., perception that students feel supported and respected by school personnel; Gregory & Cornell, 2009; Gregory et al., 2010).

Authoritative school climate is associated with a variety of positive outcomes for schools, particularly reduced student aggression and victimization. Gregory et al. (2010) found that in a sample of high school students and teachers, high structure and high support were associated with lower levels of peer victimization such as bullying, theft, and property damage. Similarly, Cornell, Shukla, and Konold (2015) found that higher structure and support were associated with lower student and school-level victimization. Higher disciplinary structure was associated with lower levels of teasing and bullying, bul- lying victimization, theft, and property damage. Higher student support was associated with less teasing and bullying, theft, and property damage (Cornell et al., 2015).

Furthermore, Gregory, Cornell, and Fan (2012) found that high structure and high support were associated with less victimization of teachers by students and that high support was associated with fewer threats made by students against staff (Gregory et al., 2012). Similarly, Berg and Cornell (2016) found that teachers in author- itative schools reported less student aggression and feeling safer and less distressed.

Present Study

There is need for schools to have access to a brief measure of SH that demonstrates good psychometric properties at both student and school levels so that it can be used in comprehensive climate surveys. A brief measure can be incorporated in longer school climate surveys and thereby generate greater awareness of sexual harassment and its relation to other aspects of school climate and student well-being. Furthermore, there is need for investigation into the prevalence of different types of SH across different de- mographic groups. Finally, substantial research linking authorita- tive school climate to other similar forms of peer victimization has supported an investigation of the association between authoritative school climate and SH victimization (e.g., Cornell et al., 2015; Gregory et al., 2010). The present study examined three research questions: (a) Is there evidence to support a four-item multilevel measure of school SH? (b) What is the prevalence of SH in a statewide high school sample, and how does SH vary across student characteristics of gender, grade level, race–ethnicity, and SES? And (c) Is an authoritative school climate associated with lower levels of SH for high school students at the individual and/or school levels? For the first question, we hypothesized that an analysis of our four SH items, derived from longer, widely used measures (AAUW, 2011; Espelage & Holt, 2001), would demon- strate support for a one-factor scale measuring student experiences of SH. Although four items cannot encompass the full range of sexual harassment behaviors, we constructed items that covered multiple forms of sexual harassment that would be most relevant to school settings. For the second question, we hypothesized that rates of SH would differ across demographic groups. For the third question, we hypothesized that students who view their school as supportive and fair would report less SH and that schools charac- terized by high structure and high support would have lower levels of SH.

Method

Sampling and Participants

The sample of schools was obtained from a statewide survey of Virginia high schools conducted in spring 2016 as part of the state’s annual School Safety Audit program (Cornell et al., 2016). The school participation rate was 99% based on 320 out of 322 eligible high schools. This high rate was accomplished in cooper- ation with the Virginia Department of Education and the Virginia Department of Criminal Justice Services, which endorsed the study and encouraged participation. The study was approved by the University of Virginia Institutional Review Board.

Schools were given the option to either (a) invite all ninth through 12th graders to participate in the study (whole-grade option) or (b) randomly select 25 students from each grade to participate (random sample option). Schools choosing the random sample option were provided with a random number list for each grade (calibrated to the enrollment in each grade), along with instructions for selecting students by matching the random num- bers to an alphabetical list of all students in that grade. Details of the selection process and response rate analyses are available in a technical report (Cornell et al., 2016).

Approximately 18% of schools selected the whole-grade option (Cornell et al., 2016). The schools that surveyed their whole grade

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

471SCHOOL CLIMATE AND SEXUAL HARASSMENT

tended to be smaller (mean enrollment � 944) than the schools that used random sampling (mean enrollment � 1,234), t(248) � 2.56, p � .05. However, there were no statistically significant differences in the percentage of students eligible for free or reduced-price meals (FRPM) for schools using the whole grade option (FRPM � .38) versus the random sample (FRPM � .40), t(248) � .72, p � .47. For urbanicity (i.e., urban, suburban, town, rural), there were no statistically significant differences as well, �2(3, N � 250) � 4.30, p � .23. All students were eligible to participate except those unable to complete the survey because of limited English proficiency or an intellectual or physical disability. Surveys were completed anonymously online using Qualtrics soft- ware (Smith, Smith, Smith, & Orgill, 2002), and students com- pleted the survey in classrooms with staff supervision using a standard set of instructions.

A multistage validity screening procedure dropped 6,272 (9.1%) of the participants for two reasons: (a) completion time was lower than an empirically derived cutoff of approximately 6 min (2.4%) and (b) students reported not answering truthfully on two screen- ing items (6.7%). The time cutoff was derived from examination of the distribution of response times on prior administrations of this survey (Cornell, Huang, et al., 2014). The first validity-screening item was “I am telling the truth on this survey” (1 � strongly disagree, 2 � disagree, 3 � agree, and 4 � strongly agree). Students who answered strongly disagree or disagree were ex- cluded from the sample. The second item was “How many of the questions on this survey did you answer truthfully?” (1 � all of them, 2 � all but 1 or 2 of them, 3 � most of them, 4 � some of

them, and 5 � only a few or none of them). Students who answered some of them or only a few or none of them were omitted from the sample. Previous research with middle and high school samples found that these items can identify students who give exaggerated reports of risk behaviors and endorse views of school conditions that are substantially more negative than those of other students (Cornell, Klein, Konold, & Huang, 2012; Cornell, Lovegrove, & Baly, 2014).

The racial–ethnic breakdown of the final sample was 54.5% White, 17.8% Black, 11.8% Hispanic, 5.8% Asian, 1.0% Ameri- can Indian or Alaska Native, and .5% Native Hawaiian or Pacific Islander, with an additional 8.7% of students identifying their background as two or more races. Approximately 23.0% reported speaking a language other than English at home. The racial–ethnic breakdown of the final sample was similar to the demographic estimates for all full-time students in the state, as estimated by the Virginia Department of Education. The racial–ethnic breakdown of students in the state for 2016–2017 was 49.7% White, 22.6% Black, 6.8% Asian, .3% American Indian or Alaska Native, .2% Native Hawaiian or Pacific Islander, 5.3% two or more races, and 15.1% Hispanic of any race. It should be noted that the survey demographics were based on student self-report and the state demographics were based on parent reports to school authorities. In the survey sample, more students reported being White and belonging to two or more races than in the state records. The distributions of parental education and free or reduced-price meals (FRPM) for the final sample are displayed in Table 1.

Table 1 Demographic Characteristics for Students Who Did and Did Not Experience Sexual Harassment

Characteristic

Students with no SHa

(n � 38,583) Students with SH

(n � 24,096) Total

(N � 62,679)

N % N % N %

Gender Female 16,508 42.8 15,612 64.8 32,120 51.2 Male 22,075 57.2 8,484 35.2 30,559 48.8

Grade Ninth 10,258 26.6 6,825 28.3 17,083 27.3 10th 9,880 25.6 6,536 27.1 16,416 26.2 11th 9,588 24.9 5,880 24.4 15,468 24.7 12th 8,857 23.0 4,855 20.1 13,712 21.9

Race–ethnicity White 20,467 53.0 13,681 56.8 34,148 54.5 Black 7,485 19.4 3,684 15.3 11,169 17.8 Hispanic 4,449 11.5 2,926 12.1 7,375 11.8 Asian 2,629 6.8 1,015 4.2 3,644 5.8 Other 3,553 9.2 2,790 11.6 6,343 10.1

Parental education No high school 2,894 7.5 1,849 7.7 4,743 7.6 High school 10,686 27.7 6,099 25.3 16,785 26.8 2-year college or technical school 5,445 14.1 3,623 15.0 9,068 14.5 4-year college 10,172 26.4 6,339 26.3 16,511 26.3 Postgraduate studies 9,386 24.3 6,186 25.7 15,572 24.8

SH Unwelcome sexual comments, jokes, or gestures — — 17,032 70.7 17,032 27.2 Unwelcome sexual rumors — — 11,144 46.2 11,144 17.8 Unwelcome sexual touch — — 8,258 34.3 8,258 13.2 Repeatedly being bothered to go out — — 10,220 42.4 10,220 16.3

Note. Dashes indicate effects that were purposely not estimated. SH � sexual harassment. a Refers to students who indicated never on all four SH items, compared to those who scored at least once on any of the items.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

472 CROWLEY, DATTA, STOHLMAN, CORNELL, AND KONOLD

Measures

The Virginia Secondary School Climate Survey consisted of 94 questions about school climate characteristics and student experi- ences (Cornell et al., 2016). This study was concerned with a subset of items measuring SH experiences, perceptions of student support, and perceptions of disciplinary structure. The survey was completed online and anonymously.

Sexual harassment. The SH scale consisted of four items (� � .78 in the current sample) derived from the AAUW’s SH survey (AAUW, 2011), as well as the University of Illinois Vic- timization Scale (Espelage & Holt, 2001). To make the scale as brief as possible, we had each question cover multiple related forms of SH. Each question began with the root “During the past 12 months how many times did another student. . . .” The four items (see Figure 1) assessed student’s experiences of (a) multiple forms of unwelcome sexual comments or jokes, (b) sexual rumors being spread about them, (c) being repeatedly bothered to go out with someone, and (d) multiple forms of unwanted physical con- tact of a sexual nature. Students responded to these items on a 5-point Likert scale (1 � never, 2 � once, 3 � twice, 4 � three times, 5 � four or more times).

Student support. The first dimension of authoritative school climate, supportiveness of teacher–student relationships, was mea- sured with an eight-item scale. Konold et al. (2014) demonstrated psychometric support for this scale through multilevel confirma- tory factor analysis. Each item was answered on a 4-point Likert- scale (1 � strongly disagree, 2 � disagree, 3 � agree, and 4 � strongly agree). Representative items included “Most teachers and

other adults at this school want all students to do well” and “There are adults at this school I could talk with if I had a personal problem” (range � 8–32; � � .86 in the current sample).

Disciplinary structure. The second dimension of authorita- tive school climate, perceived fairness and strictness of school discipline, was measured with a seven-item scale (Konold et al., 2014). Each item was answered on a 4-point Likert-scale (1 � strongly disagree, 2 � disagree, 3 � agree, and 4 � strongly agree). Representative items included “The adults at this school are too strict” and “Students are treated fairly regardless of their race or ethnicity” (range � 7–28; � � .79 in the current sample).

Covariates. Student demographic variables obtained from the survey included grade level, gender, race–ethnicity, and SES. SES was represented by parent educational level (1 � did not graduate from high school, 2 � graduated from a high school, 3 � grad- uated from a 2-year college or technical school, 4 � graduated from a 4-year college, 5 � completed postgraduate studies) and FRPM.

Analytic Strategy

Analyses were conducted in three phases. The first phase tested the hypothesized one-factor structure of the SH scale using Mplus 7.0 (Muthén & Muthén, 2012). Horn’s parallel analysis was used to identify the number of factors underlying the SH items, and the structure of the items was evaluated within a multilevel confirma- tory factor analysis (MCFA). The quality of the resulting model was gauged with the Tucker–Lewis index (TLI), the comparative fit index (CFI), the root-mean-square error of approximation

SH between group

…make

unwelcome

sexual

comments,

jokes, or

gestures…?

…spread

sexual rumors

about you?

…touch, brush

up against

you, grab or

pull your

clothing?

…repeatedly

asking you to

go out or do

something you

did not want

to?

b1 b2 b3 b4

SH within group

.672 .752 .600 .741

.587 .681 .838 1.025

Figure 1. Multilevel confirmatory factor analytic model for the four-item sexual harassment (SH) scale. Values indicate standardized factor loadings for within and between effects.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

473SCHOOL CLIMATE AND SEXUAL HARASSMENT

(RMSEA), and the standardized root-mean-square residuals (SRMRbetween and SRMRwithin). The first two measures generally range between 0 and 1.0, with larger values reflecting better fit. Values of .90 or greater have been taken as evidence of good- fitting models (Hu & Bentler, 1999). By contrast, smaller RMSEA values are reflective of better fitting models; values less than .08 are considered reasonable, and values greater than .10 are consid- ered poor (Huang, 2016). SRMR indicates acceptable fit when values are below .10 and indicates good fit when values are below .05 (Cangur & Ercan, 2015).

The second phase examined the prevalence of different types of SH reported by students in the sample, as well as how prevalence rates differed across demographic groups. A multiway analysis of variance (ANOVA) model was run using SPSS Statistics 24.0.

In the third phase, hierarchical multilevel regression investi- gated the association of SH with authoritative school climate (ASC). Two models were used to predict the prevalence of SH at the individual and school level: The first model included only covariates, and the second model included covariates with the two constructs of ASC (student support and disciplinary structure). Higher scores on the continuous variables of student support and disciplinary structure indicated more authoritative climates. Statis- tical analyses were performed with Mplus 7.0 using a maximum likelihood estimator with robust standard errors.

Results

Descriptive statistics for student- and school-level variables are presented in Table 1. The first research question tested the hypoth- esized one-factor structure of the SH scale. All SH items were moderately correlated with one another (rs � .35–.68, all ps � .001), and Horn’s parallel analysis suggested the presence of a single factor. Results of a single factor MCFA analysis suggested good fit as gauged by the CFI (.99), RMSEA (.03), TLI (.98), SRMRbetween (.04), and SRMRwithin (.01). Standardized factor loadings for the MCFA were large and statistically significant, ranging between 0.587 and 1.025 (see Figure 1).

For the second research question, prevalence rates for the indi- vidual SH items are presented in Table 1. In the full student sample, 38.4% of students reported experiencing at least some type of harassment at least once in the past school year. Among students who experienced SH, 70.7% reported unwelcome sexual com- ments or jokes, 46.2% reported having sexual rumors spread about them, 34.3% reported being repeatedly bothered to go out with someone, and 42.4% reported unwanted physical contact of a sexual nature. To investigate differences in victimization rates across demographic groups, we used a multiway ANOVA model to examine the average sum of SH experiences across gender, race–ethnicity, grade level, parent education level, and FRPM (see Table 2). In comparison to female students, male students reported less SH, F(1, 62,666) � 1,918.022, p � .001. Students in younger grades reported experiencing more SH, F(3, 62,666) � 12.857, p � .001, as did students whose parents were more highly edu- cated, F(4, 62,666) � 14.899, p � .001. Post hoc Dunnett’s pairwise comparisons indicated that non-Hispanic White students reported lower SH scores (M � 5.76) in comparison to Hispanic students (M � 5.89, p � .001) and “Other” students (M � 6.23, p � .001). Non-Hispanic White students reported higher SH scores than did Asian students (M � 5.03, p � .001) and Black students

(M � 5.42, p � .001). There were no significant differences in SH between students who qualified or did not qualify for FRPM, F(1, 62,666) � 10.390, p � .306. Effect sizes for significant results were small, with �p

2 values ranging from .001 to .03. The third research question examined regression analyses test-

ing the associations between ASC and SH (see Table 3). Model 1 presents the relations among SH and all control variables at the student level (gender, grade level, FRPM, parent education, and race–ethnicity) and school level (percentage of students per school receiving FRPM, school size, percentage of White students per school). At the student level, the control variables explained 3.7% of the variance in SH. At the school level, the control variables explained 17.7% of the variance in SH.

Model 2 built on Model 1 by introducing the ASC predictors of student support and disciplinary structure. At the student level, support and structure explained an additional 5.7% of the variance in SH, with an overall explained variance of 9.4%. Higher levels of perceived support and disciplinary structure were associated with fewer reported SH victimization experiences. At the school level, support and structure explained an additional 38.3% of the between-schools variance in SH, with an overall explained vari- ance of 56% in the final between-schools model. Approximately 1% of the of student-level variance in SH was accounted for by the school-level clusters (ICC � .009). High schools with higher support and higher structure had less reported SH than did schools with lower support and structure. To provide a sense of the size of the differences in SH between groups, we compared schools that were 1 SD above the statewide mean with those that were 1 SD below the statewide mean for each of the school climate measures. The prevalence of sexual harassment was 34% in high-structure schools and 41% in low-structure schools. Sim- ilarly, the prevalence of sexual harassment was 36% in high- support schools and 41% in low-support schools. This indicates that high structure was associated with a 17% reduction in school SH (7/41 � .17) and high support was associated with a 12% reduction in SH (5/41 � .12) in the previously mentioned subset of schools.

Table 2 Analysis of Variance for Frequency of Sexual Harassment by Demographic Variable

Variable F p �p 2

Gender 1,918.780 �.001 .030 Grade 12.966 �.001 .001 Race–ethnicitya 116.888 �.001 .007

White vs. Black — �.001 — White vs. Hispanic — �.001 — White vs. Asian — �.001 — White vs. Other — �.001 —

Parent educationb 15.121 �.001 .001 FRPMb 10.3901 .306 .000

Note. Dashes indicate effects that were purposely not estimated. FRPM � free or reduced-price meals. a Post hoc analysis of Dunnett’s pairwise comparison for White students versus Hispanic, Black, Asian, and “Other.” b Socioeconomic status mea- sured by highest education level of either parent and whether students qualified for free or reduced-price meals.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

474 CROWLEY, DATTA, STOHLMAN, CORNELL, AND KONOLD

Discussion

The current study focused on development of a four-item scale that can be used to measure the prevalence of sexual harassment (SH) in high schools with limited cost of time. To our knowledge, this is the first such scale to demonstrate good psychometric properties at both the student and school levels. Schools have a compelling need to measure SH in order to increase awareness of student experiences of victimization and inform policies and pre- vention measures to ameliorate the problem. Psychometric support for the four-item scale to measure student SH was demonstrated in a variety of ways, including good model fit and evidence of a single SH factor. Furthermore, other measures are substantially longer, such as the AAUW scale with versions ranging from 10 to 14 items (AAUW, 2001, 2011) and the SEQ with versions ranging from 23 to 26 items (Fitzgerald et al., 1999). As a result, it provides schools with a convenient and brief tool for assessing potential SH problems. Schools are more likely to monitor SH if it can be incorporated in a longer survey. Furthermore, inclusion of a brief SH scale in longer surveys will help schools recognize the relations between SH and other aspects of school climate and student well-being.

School psychologists should be aware of the high prevalence of SH in high schools and bring this to the attention of school authorities and the staff as a whole. The current study found that 38.4% of students in our statewide high school sample reported experiencing some type of SH at least once in the past school year. The most common form was experiencing unwelcome sexual comments, jokes, or gestures (70.7%), though sexual rumors (46.2%), physical sexual contact (42.4%), and being repeatedly bothered to go out with someone (34.3%) were all common.

Consistent with the majority of other studies of SH (see e.g., AAUW, 1993, 2001, 2011), girls reported substantially more SH than did boys on all four SH items. Students in higher grades reported less SH. Although previous studies with smaller and less diverse samples did not detect racial–ethnic differences, we found significant differences between all of the racial–ethnic groups in this study. However, the findings were complex and deserve further examination: Hispanic students and “Other” race students reported more SH than did non-Hispanic White students, but Asian and Black students reported less SH than did non-Hispanic White students. One unexpected finding is that students from higher income households reported more SH than did lower income students. It should be noted that the effect sizes of these demo- graphic differences were small, and therefore the results should be interpreted with caution.

The results extend previous findings that an authoritative school climate is associated with lower levels of peer victimization. High disciplinary structure and high student support were both associ- ated with lower levels of student-reported SH experiences. These associations fit well into the theoretical basis of authoritative school climate theory. In schools where rules are strict (high structure) and school staff are supportive and willing to help students (high support), it is likely that school personnel would be more willing to intervene in SH situations and students would be more likely to turn to school personnel for help. Furthermore, it is probable that schools with high structure and high support also have cultures where peer victimization such as SH is not tolerated and would therefore be lower. The present study cannot demon- strate such causal effects, and there may be bidirectional effects and other influences that were not measured.

At the student level, findings indicate that students’ individual perceptions of higher structure and higher support are associated with fewer self-reported experiences of SH, controlling for student gender, grade level, race, and SES. These findings are consistent with results reported in other studies, which indicated that schools where students perceive that there are “clear and known rules against violence” (Attar-Schwartz, 2009, p. 412) and feelings that “my teachers really care about me” (Gruber & Fineran, 2016, p. 119) have lower levels of SH. The magnitude of the individual- level effects is also comparable to those in previous studies. In our study, school climate (structure and support) explained 5.7% of the variance in SH, compared to the result in Attar-Schwartz (2009), who found that school climate (school policy, teacher support, and student participation) explained 5.8% of the variance in SH at the individual level. Though school climate was measured differently in these two studies, the findings suggest that there is a significant effect of school climate on individual SH experiences. The small size of this effect is likely because there are many individual differences in student experiences that should also be considered in association with SH experiences, such as parent and peer influ- ences and individual attitudes about harassment experiences. Fur- thermore, because many students report no SH experiences, the effects of school climate at the individual level may be limited.

On the other hand, school climate had a substantial association at the school level. School-level analyses demonstrated that schools with higher overall ratings of support and structure had lower reported rates of SH. The magnitude of these findings was substantial, explaining 38.3% of the variance in SH across schools. Previous studies of the ASC model have found comparable effects

Table 3 Standardized (STDy) Regression Coefficients for Multi-Level Covariate and Predictor Models

Predictor

Sexual harassment

Model 1 Model 2

Student level Male �.345��� �.328���

Grade level �.017��� �.026���

FRPM �.018 �.017 Parent education .004 .016���

Hispanic .078��� .048���

Non-Hispanic Black �.129��� �.17���

Non-Hispanic Asian �.182��� �.147���

Non-Hispanic Other .144��� .08���

Support — �.198���

Structure — �.31���

R2 .037 .094 �R2 — .057

School level FRPM .971 .904�

School size �.001�� �.001�

% White �.408 .122 Support — �3.118�

Structure — �2.005�

R2 .177 .56 �R2 — .383

Note. Dashes indicate effects that were purposely not estimated. FRPM � free or reduced-price meals. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

475SCHOOL CLIMATE AND SEXUAL HARASSMENT

on other forms of peer harassment; for example, measures of authoritative school climate explained 34% of the variance in school-level bullying victimization rates in a statewide middle school sample (Cornell et al., 2015).

Limitations

The main limitation of this study is the correlational nature of our findings. Correlational findings are open to multiple interpre- tations, which leaves room for the possibility that there are bidi- rectional causal effects between school climate and SH victimiza- tion. It is possible that a more positive school climate (higher levels of structure and support) leads to lower levels of SH, but it is also possible that levels of SH affect student perceptions of school climate (Ormerod et al., 2008). Future longitudinal and intervention studies would be beneficial to test how improvements to school climate affect SH.

Another limitation of this study is that all items are based on student self-report and share method variance. Although the vali- dation of our four-item scale showed good fit as a measure of student SH victimization, it would be beneficial to examine inde- pendent measures of SH that do not share method variance with the survey, such as cases reported to school authorities.

Another limitation is that our SH measure was constrained in length so that it could be included in a larger statewide survey. We recognize that a four-item measure might not detect as high a prevalence of SH as would measures with many more items. To help compensate for its brevity, we constructed items that included multiple forms of SH that would be most relevant to schools.

Finally, it is possible that SH is associated with climate-related variables that are not included in this analysis. For example, the model of school climate developed by the U.S. Department of Education has 13 components organized into three domains: en- gagement, safety, and environment (Bradshaw, Waasdorp, Deb- nam, & Johnson, 2014). Several of their components align with authoritative school climate theory, but future research could in- vestigate other components as they relate to SH victimization. Future studies could explore other variables influencing the school climate–SH relationship, such as the extent and quality of student support services in the school. Nevertheless, our study controls for many student-level and school-level demographic variables to account for factors beyond the predictors of interest.

Implications

The psychometric properties of the four-item SH scale suggest that it is a potentially valuable tool for school psychologists to use in school surveys and research. Whereas most studies have col- lected SH data using longer measures, a brief measure of SH has more practical value for routine use in school climate surveys.

Furthermore, these results indicate that the most common form of SH that students experience is nonphysical harassment, such that unwelcome sexual comments, jokes, and gestures contribute most to the overall rates of reported SH in our sample. This finding has implications for intervention and prevention in schools, be- cause it illustrates that SH is not just sexual touching but rather a variety of different behaviors. Education about SH in schools should be comprehensive and incorporate many different types of experiences to foster understanding for both students and school

personnel. We recommend that schools measure SH to increase awareness of this pervasive problem, lay the foundation for edu- cational efforts, and evaluate intervention efforts. Information about the prevalence of SH can be shared with students, staff, and parents to increase awareness, stimulate discussion, and encourage reporting. Victims of SH will be more likely to come forward and seek help when they realize that it is a common experience and that their schools and communities are concerned about it.

Studies of SH have indicated that it is prevalent in schools and is associated with negative student outcomes (see e.g., Eom et al., 2015; Ormerod et al., 2008). However, examination of research literature and the National Registry of Evidence-Based Programs and Practices found few effective intervention or prevention pro- grams specifically for SH in schools (Espelage, Low, Polanin, & Brown, 2013; Substance Abuse and Mental Health Services Ad- ministration, n.d.; Taylor, Stein, Mack, Horwood, & Burden, 2008). One program with some demonstrated effectiveness is Shifting Boundaries, a classroom-based (e.g., education sessions) and building-based (e.g., awareness posters) program for middle school students, which demonstrated 26% less SH victimization compared to a control group (Taylor, Stein, & Burden, 2010; Taylor, Stein, Mumford, & Woods, 2013).

The results from the current study suggest that authoritative school climate theory could inform future prevention programs to decrease student SH experiences. Given the demonstrated associ- ation between ASC and SH rates, it is possible that school climate interventions could contribute to schoolwide efforts to reduce SH and enhance the effectiveness of existing educational programs. For example, students may know what SH is and be told to report it, but they might be less likely to come forward if they do not trust school authorities and policies or feel respected and treated fairly. Furthermore, it is possible that intervention programs addressing SH would be more effective in schools with more authoritative school climates. For example, in schools with high structure, it is likely that when students believe that school rules are clear and fairly enforced, they will be more likely to comply with policies regarding school SH. Furthermore, if students feel that teachers and staff are supportive of them, then they may be more likely to engage in interventions, report SH experiences, and seek help after experiencing victimization. School psychologists and other school personnel should be cognizant of the way their policies, expecta- tions, and relationships with students emulate an authoritative climate, taking care to offer help to students, encourage disclosure of victimization experiences, and demonstrate trust in and respect for students who do report victimization. School administrators should ensure that policies are appropriately created and enforced to identify, resolve, and prevent SH.

Conclusion

With the onset of the #MeToo movement in October 2017, there has been an outpouring of reports of SH in professional settings, signaling a cultural shift toward increased awareness of SH and support for those who report it (Camera, 2018). In response to heightened public awareness, there should be a greater focus on SH in schools, recognizing that the problem of SH does not begin in adulthood but is clearly evident in adolescence. States are making legislative efforts to respond to the school SH problem—in 2018, the Virginia General Assembly mandated that family life

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

476 CROWLEY, DATTA, STOHLMAN, CORNELL, AND KONOLD

curricula in elementary, middle, and high schools incorporate age-appropriate and evidence-based programs on personal privacy and personal boundaries and established that “Any high school family life education curriculum offered by a local school division shall incorporate age-appropriate elements of effective and evidence-based programs on the prevention of dating violence, domestic abuse, sexual harassment, and sexual violence” (Code of Virginia, 2018). We recommend that school psychologists help their schools to routinely assess student SH in order to identify problems; increase awareness among students, staff, and parents; and monitor the effectiveness of interventions. Furthermore, we suggest that generating a school climate where students feel re- spected and supported might be helpful in reducing SH. With increased efforts to monitor and intervene in SH victimization, schools can better provide safe and healthy learning environments for their students.

References

American Association of University Women (AAUW). (1993). Hostile hallways: The AAUW survey on sexual harassment in America’s schools. Washington, DC: AAUW Educational Foundation.

American Association of University Women (AAUW). (2001). Hostile hallways: Bullying, teasing, and sexual harassment in school. Washing- ton, DC: AAUW Educational Foundation.

American Association of University Women (AAUW). (2011). Harassment-free hallways: How to stop sexual harassment in America’s schools. Washington, DC: AAUW Educational Foundation.

Attar-Schwartz, S. (2009). Peer sexual harassment victimization at school: The roles of student characteristics, cultural affiliation, and school factors. American Journal of Orthopsychiatry, 79, 407–420. http://dx .doi.org/10.1037/a0016553

Berg, J. K., & Cornell, D. (2016). Authoritative school climate, aggression toward teachers, and teacher distress in middle school. School Psychol- ogy Quarterly, 31, 122–139. http://dx.doi.org/10.1037/spq0000132

Bradshaw, C. P. (2015). Translating research to practice in bullying pre- vention. American Psychologist, 70, 322–332. http://dx.doi.org/10.1037/ a0039114

Bradshaw, C. P., Waasdorp, T. E., Debnam, K. J., & Johnson, S. L. (2014). Measuring school climate in high schools: A focus on safety, engage- ment, and the environment. Journal of School Health, 84, 593–604. http://dx.doi.org/10.1111/josh.12186

Camera, L. (2018, January 8). #Me too goes to school. U.S. News and World Report. Retrieved from https://www.usnews.com/news/education- news/articles/2018-01-08/the-metoo-movement-goes-to-school

Cangur, S., & Ercan, I. (2015). Comparison of model fit indices used in structural equation modeling under multivariate normality. Journal of Modern Applied Statistical Methods, 14, 152–167. http://dx.doi.org/10 .22237/jmasm/1430453580

Clear, E. R., Coker, A. L., Cook-Craig, P. G., Bush, H. M., Garcia, L. S., Williams, C. M., . . . Fisher, B. S. (2014). Sexual harassment victim- ization and perpetration among high school students. Violence Against Women, 20, 1203–1219. http://dx.doi.org/10.1177/1077801214551287

Cohen, J., McCabe, L., Michelli, N. M., & Pickeral, T. (2009). School climate: Research, policy, practice, & teacher education. Teachers College Record, 111, 180 –213. Retrieved from https://www.researchgate.net/profile/ Jonathan_Cohen4/publication/235420504_School_Climate_Research_ Policy_Teacher_Education_and_Practice/links/59d67f050f7e9b42a6aa0145/ School-Climate-Research-Policy-Teacher-Education-and-Practice.pdf

Cornell, D., Huang, F., Datta, P., Malone, M., Jia, Y., Burnette, A. G., . . . Meyer, P. (2016). Technical report of the Virginia secondary school climate survey: 2016 results for 9th-12th grade students and school staff. Charlottesville, VA: Curry School of Education, University of Virginia.

Retrieved from https://curry.virginia.edu/uploads/resource Library/2016_HS_Technical_Report_July_1_2016_for_release.pdf

Cornell, D., Huang, F., Konold, T., Meyer, P., Shukla, K., Heilbrun, A., . . . Nekvasil, E. (2014). Technical report of the Virginia secondary school climate survey: 2014 results for 9th–12th grade students and teachers. Charlottesville, VA: Curry School of Education, University of Virginia. Retrieved from https://curry.virginia.edu/uploads/resourceLibrary/ State_Technical_Report_for_2014_high_school_survey_8-14-14.pdf

Cornell, D., Klein, J., Konold, T., & Huang, F. (2012). Effects of validity screening items on adolescent survey data. Psychological Assessment, 24, 21–35. http://dx.doi.org/10.1037/a0024824

Cornell, D. G., Lovegrove, P. J., & Baly, M. W. (2014). Invalid survey response patterns among middle school students. Psychological Assess- ment, 26, 277–287. http://dx.doi.org/10.1037/a0034808

Cornell, D. G., & Mayer, M. J. (2010). Why do school order and safety matter? Educational Researcher, 39, 7–15. http://dx.doi.org/10.3102/ 0013189X09357616

Cornell, D., Shukla, K., & Konold, T. (2015). Peer victimization and authoritative school climate: A multilevel approach. Journal of Educa- tional Psychology, 107, 1186 –1201. http://dx.doi.org/10.1037/ edu0000038

Eom, E., Restaino, S., Perkins, A. M., Neveln, N., & Harrington, J. W. (2015). Sexual harassment in middle and high school children and effects on physical and mental health. Clinical Pediatrics, 54, 430–438. http://dx.doi.org/10.1177/0009922814553430

Espelage, D. L., Basile, K. C., & Hamburger, M. E. (2012). Bullying perpetration and subsequent sexual violence perpetration among middle school students. Journal of Adolescent Health, 50, 60–65. http://dx.doi .org/10.1016/j.jadohealth.2011.07.015

Espelage, D. L., & Holt, M. K. (2001). Bullying and victimization during early adolescence: Peer influences and psychosocial correlates. Jour- nal of Emotional Abuse, 2, 123–142. http://dx.doi.org/10.1300/ J135v02n02_08

Espelage, D. L., Hong, J. S., Rinehart, S., & Doshi, N. (2016). Under- standing types, locations, & perpetrators of peer-to-peer sexual harass- ment in U.S. middle schools: A focus on sex, racial, and grade differ- ences. Children and Youth Services Review, 71, 174–183. http://dx.doi .org/10.1016/j.childyouth.2016.11.010

Espelage, D. L., Low, S., Polanin, J. R., & Brown, E. C. (2013). The impact of a middle school program to reduce aggression, victimization, and sexual violence. Journal of Adolescent Health, 53, 180–186. http://dx .doi.org/10.1016/j.jadohealth.2013.02.021

Fitzgerald, L. F., Magley, V. J., Drasgow, F., & Waldo, C. R. (1999). Measuring sexual harassment in the military: The Sexual Experiences Questionnaire (SEQ-DoD). Military Psychology, 11, 243–263. http://dx .doi.org/10.1207/s15327876mp1103_3

Gregory, A., & Cornell, D. (2009). “Tolerating” adolescent needs: Moving beyond zero tolerance policies in high school. Theory Into Practice, 48, 106–113. http://dx.doi.org/10.1080/00405840902776327

Gregory, A., Cornell, D., & Fan, X. (2012). Teacher safety and authorita- tive school climate in high schools. American Journal of Education, 118, 401–425. http://dx.doi.org/10.1086/666362

Gregory, A., Cornell, D., Fan, X., Sheras, P., Shih, T.-H., & Huang, F. (2010). Authoritative school discipline: High school practices associated with lower bullying and victimization. Journal of Educational Psychol- ogy, 102, 483–496. http://dx.doi.org/10.1037/a0018562

Gruber, J., & Fineran, S. (2016). Sexual harassment, bullying, and school outcomes for high school girls and boys. Violence Against Women, 22, 112–133. http://dx.doi.org/10.1177/1077801215599079

Guerra, N. G., Williams, K. R., & Sadek, S. (2011). Understanding bullying and victimization during childhood and adolescence: A mixed methods study. Child Development, 82, 295–310. http://dx.doi.org/10 .1111/j.1467-8624.2010.01556.x

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

477SCHOOL CLIMATE AND SEXUAL HARASSMENT

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alterna- tives. Structural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/ 10705519909540118

Huang, F. L. (2016). Alternatives to multilevel modeling for the analysis of clustered data. Journal of Experimental Education, 84, 175–196. http:// dx.doi.org/10.1080/00220973.2014.952397

Konold, T., Cornell, D., Huang, F., Meyer, P., Lacey, A., Nekvasil, E., . . . Shukla, K. (2014). Multilevel multi-informant structure of the Authori- tative School Climate Survey. School Psychology Quarterly, 29, 238– 255. http://dx.doi.org/10.1037/spq0000062

Kosciw, J. G., Palmer, N. A., Kull, R. M., & Greytak, E. A. (2013). The effect of negative school climate on academic outcomes for LGBT youth and the role of in-school supports. Journal of School Violence, 12, 45–63. http://dx.doi.org/10.1080/15388220.2012.732546

Muthén, L. K., & Muthén, B. O. (1998–2012). Mplus User’s Guide (7th Edition). Los Angeles, CA: Author. Retrieved from https://www .statmodel.com/download/usersguide/Mplus%20user%20guide%20 Ver_7_r3_web.pdf

Ormerod, A. J., Collinsworth, L. L., & Perry, L. A. (2008). Critical climate: Relations among sexual harassment, climate, and outcomes for high school girls and boys. Psychology of Women Quarterly, 32, 113–125. http://dx.doi.org/10.1111/j.1471-6402.2008.00417.x

Smith, S., Smith, R., Smith, J., & Orgill, S. (2002). Qualtrics [Computer software]. Retrieved from https://www.qualtrics.com

Substance Abuse and Mental Health Services Administration. (n.d.). Na- tional Registry of Evidence-Based Programs and Practices (NREPP). Retrieved June 3, 2018, from https://www.samhsa.gov/nrepp

Taylor, B., Stein, N., & Burden, F. (2010). The effects of gender violence/ harassment prevention programming in middle schools: A randomized experimental evaluation. Violence and Victims, 25, 202–223. http://dx .doi.org/10.1891/0886-6708.25.2.202

Taylor, B., Stein, N., Mack, A. R., Horwood, T. J., & Burden, F. (2008). Experimental evaluation of gender violence/harassment prevention pro- grams in middle schools. Advance online publication. http://dx.doi.org/ 10.1037/e440442008-001

Taylor, B. G., Stein, N. D., Mumford, E. A., & Woods, D. (2013). Shifting Boundaries: An experimental evaluation of a dating violence prevention program in middle schools. Prevention Science, 14, 64–76. http://dx.doi .org/10.1007/s11121-012-0293-2

Title IX, 20 U.S.C. § 1681 et seq. (1972) U.S. Department of Education, Office for Civil Rights. (2008). Sexual

harassment: It’s not academic. Retrieved from https://www2.ed.gov/ about/offices/list/ocr/docs/ocrshpam.pdf

Code of Virginia § 22.1-207.1:1 (2018).

Received July 15, 2018 Revision received October 27, 2018

Accepted November 1, 2018 �

E-Mail Notification of Your Latest Issue Online!

Would you like to know when the next issue of your favorite APA journal will be available online? This service is now available to you. Sign up at https://my.apa.org/portal/alerts/ and you will be notified by e-mail when issues of interest to you become available!

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

478 CROWLEY, DATTA, STOHLMAN, CORNELL, AND KONOLD