Longitudinal Effects of Gendered Harassment Perpetration and Victimization on Mental Health Outcomes in Adolescence

Abstract

Gendered harassment, including sexual harassment and homophobic name-calling, is prevalent in adolescents and is linked to negative outcomes including depression, anxiety, suicidality, substance abuse, and personal distress. However, much of the extant literature is cross-sectional and rarely are perpetrators of these behaviors included in studies of outcomes. Therefore, the current study examined the effects of longitudinal changes in gendered harassment perpetration and victimization on changes in mental health outcomes among a large sample of early adolescents. Given that these behaviors commonly occur in the context of a patriarchal society (males hold power), we also investigated the impact of gender on gendered harassment. Participants included 3,549 students from four Midwestern middle schools (50.4% female, 49% African American, 34% White) at two time points (13 and 17 years old). Results indicated that increases from age 13 to 17 years in sexual harassment perpetration and victimization and homophobic name-calling perpetration and victimization predicted increases in depression symptoms and substance use. Gender did not moderate these pathways. These findings highlight that negative outcomes are associated with changes in gendered harassment among adolescents and emphasize the importance of prevention efforts. Implications for school interventions are discussed.

Keywords

harassment homophobia mental health adolescence

Sexual harassment and homophobic name-calling are relatively common acts in middle and high schools and are indicative of a larger societal force that equates masculinity with violence and strength, whereas femininity is equated with passivity and weakness (Kimmel & Mahler, 2003). In our patriarchal culture, boys frequently engage in tasks that are considered more traditionally masculine (“perform their masculinity”) to prove their heterosexuality along with their power. Students who are questioning their sexuality or identify as lesbian, gay, or bisexual are at risk for being exposed as not behaving in a properly gendered fashion. Both sexual harassment and homophobic name-calling (e.g., gay, homo, fag) fall under the broader construct of gendered harassment (Meyer, 2008). This term encompasses aggressive acts that serve to reinforce sexism and homophobia, which in turn support patriarchal power structures (Connell, 1987).

More specifically, feminist theories indicate that harassment serves as a form of social control, reinforcing women’s lower rank in society, including in employment, education, and interpersonal relationships (Benson & Thomson, 1982). In this context, sexual harassment supports male dominance and female subordination, which in turn is an aspect of maintaining a sexist and patriarchal society (Fineran, 2002; Meyer, 2008). For adolescents, the harassment that they experience as victims and perpetrators is centered on conformity to gender stereotypes, whether that is through homophobic or sexist channels (Stein, 1995).

Sexual harassment and homophobic name-calling are both aspects of reinforcing traditional masculinity and performing a form of masculinity that is antifeminine, emotionally restrictive, competitive, and aggressive (Meyer, 2008). In schools, where harassment can occur with little intervention from teachers and staff, youth can be seen as receiving not only permission but also implicit training to become harassers in their adult lives (Charmaraman, Jones, Stein, & Espelage, 2013; Meyer, 2008; Stein, 1995). Given this relation between homophobic name-calling and sexual harassment in reinforcing heterosexual masculinity for adolescents, the current study addressed the constructs of sexual harassment and homophobic name-calling as aspects of the broader concept of gendered harassment.

These behaviors are concerning not just on a societal or theoretical level; extant literature indicates that gendered harassment is indeed connected to negative mental health outcomes (Bucchianeri, Eisenberg, Wall, Piran, & Neumark-Sztainer, 2013). In terms of sexual harassment, experiencing these behaviors as a victim is associated with lower mental and physical health, life satisfaction, and educational outcomes (Bucchianeri et al., 2013; Chiodo, Wolfe, Crooks, Hughes, & Jaffe, 2009; Gruber & Fineran, 2007). Likewise, homophobic name-calling victimization is associated with lower educational outcomes and with mental health concerns, including depression, anxiety, suicidality, risky behavior including substance abuse, and personal distress (D’Augelli, Pilkington, & Hershberger, 2002; Espelage et al., 2008; Martin-Storey & Crosnoe, 2014; Poteat & Espelage, 2007).

One possible explanation for this association between gendered harassment victimization and negative health outcomes is that involvement in harassment both contributes to and is fueled by personal distress including internalizing responses and peer-relational difficulties (Hodges & Perry, 1999). Adolescents’ personal distress, as operationalized in the current study by depression symptoms and increased substance abuse, increases in response to peer victimization, which then in turn can contribute to further increases in victimization. These antecedents and consequences of peer victimization can be understood as mutually reinforcing one another and as contributing to the stability of victimization in adolescence (Hodges & Perry, 1999).

Given this potential for ongoing victimization during adolescence, this developmental time is important to consider. In the current study, students in Wave 1 were on average 13 years old (seventh grade) and in Wave 2 were 17 years old (11th grade). Sexual harassment increases throughout early adolescence and levels off by late high school; the age range analyzed here includes this period of potential increase and leveling off of harassment and will therefore more likely capture the effects of gendered harassment behaviors over time (Espelage, Basile, & Hamburger, 2012; Pepler et al., 2006; Petersen & Hyde, 2009). Pubertal changes that commonly occur during middle school have been linked to increased sexual harassment victimization in fifth through ninth grade, with both boys and girls with advanced pubertal status throughout this time period being more likely to experience harassment by ninth grade (Juvonen & Graham, 2001; Petersen & Hyde, 2009). Consequently, this study examined how students’ experiences of gendered harassment during this period affected their mental health over time.

More broadly, this period of adolescence is a time when gender roles are explored and formed, with early adolescence being a time of escalation of gender-related role expectations (J. P. Hill, 1983). Peers, family, and school climate, as well as broader societal forces including the media, all have important roles to play in creating norms of gendered behavior as students model and practice the various interactions they observe on a daily basis (Meyer, 2008). The accompanying biological and pubertal changes that occur, as well as increases in cognitive complexity and capacity, point to this period as important for the development of acceptable behaviors and adaptive responses to stress, and is a complex time of change and adjustment (J. P. Hill, 1983). Subsequently, examining the mental health outcomes of gendered harassment is an illuminating aspect of understanding adolescents’ response to difficult experiences and deserves further attention.

In considering the impact of adolescents’ gender on outcomes related to gendered harassment, one of the few longitudinal studies of homophobic name-calling and mental health outcomes reported that homophobic victimization significantly predicted higher anxiety and depression among middle school males and higher levels of withdrawal among middle school females (Poteat & Espelage, 2007). These differential outcomes across gender point to the varying meaning that homophobic name-calling has for males and females; the authors posited that homophobic epithets could (e.g., gay, homo, fag) occur regularly within male peer groups and function in part to establish dominance hierarchies within peer groups (Poteat & Espelage, 2007). For females, on the contrary, being targeted by homophobic teasing could lead to increased feelings of stigmatization and rejection because it is a less normative occurrence in the female peer social group (Poteat & Espelage, 2007). Gender differences in the association between sexual harassment victimization and mental health have also been found (Chiodo et al., 2009). Due to this variability in outcomes by gender, the current study considered gender as a moderator in the pathways between gendered harassment and mental health outcomes.

Although the association between harassment and negative mental health outcomes has been established, the literature is largely cross-sectional and often based on retrospective self-report. The absence of longitudinal research in this area means that much of the literature is lacking in both complexity and a closer consideration of which aspects of harassment lead to specific mental health concerns. In addition, much of the literature is focused solely on the effects of victimization, which ignores the important relation between perpetration and negative mental health outcomes (Bucchianeri et al., 2013). Therefore, the present study tested the link between both gender-based victimization and perpetration and negative mental health outcomes over time.

In sum, the current study addressed the lack of longitudinal research on the effect of gendered harassment on mental health outcomes by using longitudinal structural equation modeling. This analytic approach enabled simultaneous modeling of change in gendered harassment from age 13 to 17 and change in the mental health outcomes of depression and drug and alcohol use over the same period. By investigating sexual harassment and homophobic name-calling perpetration and victimization in the same study, two important aspects of gendered harassment were both addressed, which provides a fuller understanding of the phenomenon of gendered harassment. Due to the importance of gender in the experience of gendered harassment, separate but simultaneous analyses for girls and boys were modeled using multigroup analysis so that comparisons across gender could be made. Thus, we hypothesized that increases in gendered harassment perpetration and victimization would be associated with increases in mental health outcomes over time. We hypothesized that gender would be a significant moderator and that the relation between gendered harassment and negative mental health outcomes would be stronger for females than males.

Method

Participants

This study was conducted with 3,549 students from four Midwestern middle schools. These four public schools are situated in a Midwestern school district where 60.4% of the students are African American, followed by 31.5% European American, 2.6% Asian, 5.1% Hispanic, and 0.4% multiracial. Approximately, 69.3% of the student population is considered low income. The chronic truancy rate for the school district is 2.5% and the mobility rate is 30.1%. Students were included in the analyses as long as they completed a minimum of one wave of survey data. The sample was 50.4% female, 49% African American, 34% White, 6% Hispanic, and 2% Asian. In the first wave of data used, 42% of the participants were 13 years old with an almost even split between seventh and eighth grade. In the second wave of data, the mean age was 16.8 years with the majority of students aged between 16 and 18 years and in 10th through 12th grade.

Procedures

The university institutional review board and school district administration approved a waiver of active parental consent. Students were asked to assent to participate in the study through an assent procedure included on the coversheet of the survey. A 95% participation rate was achieved. Students completed surveys in fall 2009 (Wave 1) and spring 2013 (Wave 2). Six trained research assistants and the primary investigator collected data.

Measures

Depression

An eight-item version of the Orpinas Modified Depression Scale was used to assess depressive symptoms (Orpinas, 1993). Students were asked about sadness, irritability, worrying, and hopelessness. Participants were asked how often in the past 30 days they had encountered these issues. Response options were never, sometimes, often, and almost always. Higher scores indicate more depressive symptoms. The scale has demonstrated good internal consistency with an alpha coefficient of .74 when administered to adolescents aged 10 to 18 years (Orpinas, 1993). Wave 1 included an item that was not used in Wave 2 (“Did you feel like not eating or eating more than usual?”), so this item was removed to make the scales consistent across waves. Items asking about sleep, concentration, and happiness all failed to load significantly for both waves; due to this failure to add to the measurement model significantly, these three items were also removed. Alpha coefficients for Wave 1 were .79 for nine-item and .84 for the five-item, and for Wave 2 were .69 for eight-item and .90 for the five-item.

Drug and alcohol use

An eight-item scale asked students to report how many times in the past year they used alcohol or drugs in Wave 1 (Farrell, Kung, White, & Valois, 2000). The scale included statements like “drunk beer,” “smoked cigarettes,” “drunk liquor,” and “used marijuana.” Response options were never, 1 or 2 times, 3 to 5 times, 6 to 9 times, and 10 or more times. The scale correlates positively with risk behaviors like delinquency and correlates negatively with positive behaviors including school attendance (Farrell et al., 2000). Farrell et al. (2000) reported a Cronbach’s alpha of .87 with a sample of urban adolescents. The final two items (“used inhalants” and “used other drugs”) were not used in the analysis due to very low endorsement of these behaviors. The alpha coefficients for the current study were calculated both with and without these final items; Cronbach’s alpha was .83 for the eight-item scale, and improved to .86 for the six-item scale. In Wave 2, a six-item scale asked students to report how many days, out of the past 30, various substances were used (D’Amico et al., 2012). The scale asked about use of cigarettes, smokeless tobacco, alcohol and alcohol binging, marijuana, and other illegal drugs. Response options included 0 days, 1 day, 2 days, 3 to 5 days, 6 to 9 days, 10 to 19 days, and 20 to 30 days. Cronbach’s alpha for the current study was .78 for Wave 2.

Sexual harassment perpetration and victimization

The American Association of University Women (AAUW) Sexual Harassment Survey (AAUW, 2001) was used to measure the frequency with which students experienced (victimization) and perpetrated sexually harassing behaviors within the last year in Wave 1 of the data collection. Thirteen items assessed perpetration, while 13 parallel items asked about victimization. The behaviors measured include unwanted sexual jokes, comments, and touching. Response options include not sure, never, rarely, occasionally, and often. Scores are summed and higher scores indicate higher frequency of experience or perpetration of sexual harassment. In one longitudinal study, Cronbach’s alpha ranged from .68 to .75 for victimization and from .67 through .72 for perpetration (Taylor & Stein, 2007). In the current study, the response options not sure and never were collapsed, which provided the added benefit of creating more consistent scaling across waves. To create an equivalent scale across waves, the six overlapping items used in both the Wave 1 scale and in the Wave 2 scale were used. Alpha coefficients for the current study were .62 for perpetration and .70 for victimization.

A modified version of the AAUW Sexual Harassment Survey was used in Wave 2 (Espelage et al., 2012). For perpetration, participants were presented with six items to assess verbal sexual harassment (e.g., sexual comments, sexual rumor spreading, and showing sexual pictures) and forced sexual contact (e.g., touching in a sexual way, physically intimidating in a sexual way, forcing to do something sexual). Students were asked to consider how often in the current school year they had done each of these acts to other students at school, and response options were never, 1 or 2 times, 3 or 4 times, 5 or 6 times, and 7 or more times. The victimization scale asked about the same behaviors with the same response options, but instead asked students how often during the school year other students had done each act to them. The final two response options were collapsed in the current study to create more consistent scaling across waves. Alpha coefficients were .80 for perpetration and .76 for victimization.

Homophobic name-calling perpetration and victimization

The 10-item Homophobic Content Agent Target Scale was used to assess homophobic teasing perpetration and victimization (Poteat & Espelage, 2007). Students were asked how often in the past 30 days they directed homophobic epithets at other students (perpetration) or were targets of this language (victimization). For the five-item perpetration scale, students were asked, “How many times in the last 30 days did YOU say [homo, gay, lesbo, or fag] to” various categories of peers for each item. Response options included never, 1 or 2 times, 3 or 4 times, 5 or 6 times, and 7 or more times. The five-item victimization scale consisted of the same items and response options, except that students were asked how often others called them homophobic epithets. Construct validity has been supported through exploratory and confirmatory analyses and the victimization scale correlates significantly with measures of bullying victimization (Poteat & Espelage, 2007). Cronbach’s alpha coefficients were .86 for Wave 1 perpetration, .81 for Wave 2 perpetration, .82 for Wave 1 victimization, and .82 for Wave 2 victimization.

Demographic variables

Demographic information was collected, including gender, age, and race. For race, participants were given five options: African American (not Hispanic), Asian, White (not Hispanic), Hispanic, and Other (with a space to write in the preferred descriptor). Race was dummy coded into a set of dichotomous variables.

Data Analysis

Descriptive analyses, including means and standard deviations, of study variables for Waves 1 and 2 were calculated for the entire sample as well as separately by gender using SPSS Version 22.0 (Table 1). Correlations among sexual harassment and homophobic teasing were calculated as well (Table 2). Next, measurement models for depression, drug and alcohol use, sexual harassment perpetration, sexual harassment victimization, homophobic name-calling perpetration, and homophobic name-calling victimization were estimated using MPlus Version 7.0. The drug and alcohol use measurement model included six items, depression included five items, sexual harassment perpetration and victimization both had six items, and homophobic perpetration and victimization both had five items. For each construct, the factor loading for a single item was fixed to 1 to provide the scaling factor. Items were treated continuously to preserve the richness of the data (Hox & Stoel, 2005).

Table 1.

Descriptives of Measurement Models.

	Wave 1			Wave 2
	Total	Female	Male	Total	Female	Male
Depression	1.94 (.54)	2.06 (.57)	1.81 (0.49)	2.10 (.55)	2.21 (.55)	1.97 (.51)
Drug and alcohol use	1.17 (.38)	1.15 (.30)	1.18 (0.45)	1.40 (.83)	1.34 (.69)	1.49 (.95)
Sexual harassment perpetration	1.06 (.17)	1.06 (.18)	1.05 (0.16)	1.13 (.34)	1.09 (.24)	1.17 (.41)
Sexual harassment victimization	1.20 (.36)	1.28 (.41)	1.13 (0.27)	1.35 (.49)	1.38 (.47)	1.31 (.50)
Homophobic name-calling perpetration	1.73 (.91)	1.65 (.78)	1.82 (1.03)	1.53 (.79)	1.36 (.59)	1.71 (.92)
Homophobic name-calling victimization	1.36 (.65)	1.28 (.50)	1.44 (0.76)	1.39 (.71)	1.28 (.59)	1.50 (.80)

Note. Mean (standard deviation). Wave 1 of drug and alcohol use asked for use in past year and item numbers represent different substances, whereas Wave 2 asked for use in past 30 days.

Table 2.

Correlations Between Forms of Gendered Harassment.

	1	2	3	4	5	6	7
SH perpetration W1
SH perpetration W2	.06
SH victimization W1	.34**	–.03
SH victimization W2	.04	.57**	.14*
H perpetration W1	.37**	.22**	.20**	.11
H perpetration W2	.13	.41**	.04	.31**	.35**
H victimization W1	.19**	.19**	.20**	.12	.54**	.26**
H victimization W2	.02	.44**	–.05	.38**	.28**	.58**	.24**

Note. SH = sexual harassment; H = homophobic name-calling.

p < .05. **p < .01.

Next, we used longitudinal structural equation modeling to investigate the extent to which changes in gendered harassment predict changes in mental health outcomes and to determine whether these effects differ by gender. Given the modest overlap between the four types of gender-based harassment (Table 2), four separate main effects models were estimated: one for sexual harassment perpetration, one for sexual harassment victimization, one for homophobic name-calling perpetration, and one for homophobic name-calling victimization. For each predictor and outcome, we regressed Wave 2 on Wave 1, thereby creating an auto-regressive difference score that could be interpreted as an increase or a decrease in that variable across time (see Figure 1 for the hypothesized path model). We then regressed Wave 2 mental health outcomes on the Wave 2 gendered harassment variables. These regression paths represented our primary pathways of interest. Coefficients could be interpreted as the association between changes in gendered harassment and changes in mental health across time. The effects of race and gender on harassment and mental health outcomes at both waves were included and correlations among the items in the measurement models as well as among the two outcomes were also included.

Figure 1.

Gendered harassment predicting mental health main effects model.

To determine whether the association between gendered harassment and mental health differed by gender, we refit the models described above using multigroup analysis. That is, we simultaneously estimated the effects for males and females. The measurement models for gendered harassment and mental health were fixed to be equal across gender groups while the parameters of interest were freely estimated for each group. Had there been evidence of moderation, pathways would have been systematically fixed and freed to determine where those differences exist. Because the models were estimated simultaneously, using the same covariance matrix, the coefficients for each model could be compared across genders.

Parameters used to determine model fit were chi-square, root mean square error of approximation (RMSEA), comparative fit index (CFI), and Tucker–Lewis index (TLI). As chi-square is sensitive to sample size, the addition of these parameters was necessary. RMSEA values of less than .06 and CFI and TLI values greater than .90 were considered to indicate acceptable model fit. To address missing data, full information maximum likelihood (FIML), the default in MPlus, was used to ensure unbiased parameter estimates. This allowed for valid inferences from the statistical analyses. FIML results in similar information and outcomes as multiple imputation procedures and is a robust mechanism to manage missing data (Collins, Schafer, & Kam, 2001).

Results

Preliminary Data Analyses

To first assess change in various constructs over time and the differences in these constructs by gender, the means and standard deviations for each variable of interest were calculated both by gender and for the entire sample (Table 1). Higher numbers indicate higher levels of mental health concerns and harassment. Drug and alcohol use increased on average over time, as did endorsement of depressive symptoms. On average, females reported higher levels of depressive symptoms than their male counterparts, while males endorsed more substance use than did females. Both sexual harassment perpetration and victimization increased slightly over time. Homophobic name-calling victimization means stayed virtually stable across waves and endorsement of homophobic perpetration decreased over time.

Next, correlations between sexual harassment perpetration and victimization as well as homophobic teasing perpetration and victimization were calculated for the total sample (Table 2). These calculations indicated moderate levels of correlation across forms of gendered harassment, indicating less than 25% overlap across the four constructs. Of special interest were the correlations between sexual harassment perpetration and victimization across waves; most constructs were significantly correlated across waves for both perpetration and victimization. Similarly, in terms of homophobic name-calling, perpetration and victimization were both positively correlated significantly with each other across waves.

Finally, for each wave, we estimated six measurement models, representing sexual harassment perpetration and victimization, homophobic name-calling perpetration and victimization, depression, and drug and alcohol use. All factor loadings were statistically significant at p < .01 or p < .001 and model fit statistics were relatively good. Standardized and unstandardized factor loadings as well as model fit statistics indicated good fit. It is also worth noting that factor loadings for items within a given construct tended to be similar. These findings indicated that the items we hypothesized to represent each construct did, in fact, generate a single underlying construct for sexual harassment perpetration and victimization, homophobic name-calling perpetration and victimization, depression, and drug and alcohol use. (Standardized and unstandardized factor loadings as well as model fit statistics are available upon request given space constraints.)

Predictive Models Using Structural Equation Modeling

Gendered harassment and mental health outcomes

Results from the main effects models predicting changes in mental health outcomes from changes in gendered harassment are presented in Figures 2 through 5. Recall that these models control for the effects of race and gender on harassment and mental health outcomes at Wave 2. Fit indices for each model are included with the figure and, in combination, suggest satisfactory fit to the data. As hypothesized, changes from Wave 1 (age 13) to Wave 2 (age 17) in each aspect of gendered harassment significantly predicted changes in substance use and depression symptoms from age 13 to 17.

Figure 2.

Sexual harassment perpetration.

Figure 3.

Sexual harassment victimization.

Figure 4.

Homophobic name-calling perpetration.

Figure 5.

Homophobic name-calling victimization.

For example, youth who reported greater increases in sexual harassment perpetration also reported greater increases in drug and alcohol use (β = .28, p < .001) as well as greater increases in depression (β = .08, p < .01; see Figure 2). Similarly, youth who reported increases in sexual harassment victimization also reported increases in mental health problems (see Figure 3; β = .28, p < .001, for drug and alcohol use and β = .23, p < .001, for depression). Youth who reported increases in homophobic name-calling perpetration also reported increases in drug and alcohol use (β = .28, p < .001) and in depression (β = .16, p < .001; see Figure 4). Similarly, youth who reported greater increases in homophobic name-calling victimization also reported greater increases in mental health problems (see Figure 5; β = .21, p < .001, for drug and alcohol use and β = .16, p < .001, for depression).

Gender as a moderator

To investigate whether the associations between gendered harassment and mental health problems tested above differed by gender, we fit a series of multigroup models in which we simultaneously estimated the effects for male and females. As noted above, the measurement models were fixed to be equal across groups while the pathways of interest were freely estimated for male and females. In all cases, the fit of the model declined when moderation was tested, suggesting the associations identified above did not differ by gender. In other words, the negative effects of gendered harassment appear to be similar for males and females. Importantly, although we noted some differences in the parameter estimates across groups, follow-up tests of equivalence revealed that the values were functionally equivalent across gender. It should be noted that for the model that tested the effects of sexual harassment perpetration on mental health outcomes, moderation could not be tested because there was no variability in one of the items for one of the groups. More specifically, the multigroup analysis could not be fit because no females at Wave 1 reported that they had forced someone else to do something sexual. Because there was no evidence of moderation, we have not included the findings here.

Discussion

These results support findings in the extant literature that students do indeed experience sexual harassment and homophobic name-calling victimization and perpetration, and that these experiences are associated with negative mental health outcomes (Chiodo et al., 2009; Espelage et al., 2008; Gruber & Fineran, 2007). Adolescents, regardless of gender, endorsed involvement in both perpetration and victimization of sexual harassment and homophobic name-calling; however, males on average reported more homophobic name-calling perpetration and victimization than their female counterparts. This finding is consistent with an understanding of homophobic name-calling as one way that boys engage in tasks that are considered more traditionally masculine (“performing masculinity”), in that for males using homophobic teasing can be a way to indicate to others that they are straight and not part of the group that they are denigrating. In terms of sexual harassment, females reported higher levels of victimization and slightly lower amounts of perpetration than did males. Again, given the grounding of this study in feminist theory, these different involvement rates make sense given that sexual harassment, as enacted by males against females, is a form of reinforcing the dominance of masculinity and maleness and subordinate femininity and femaleness. In other words, when boys engage in sexual harassment against females, it can potentially indicate to their peers that they are appropriately masculine, whereas a girl engaging in sexual harassment does not necessarily reinforce her femininity.

For all students, their likelihood of experiencing depressive symptoms and substance use at age 17 increased if they were involved in any form of gendered harassment from age 13 to age 17. This relation existed across perpetration and victimization and for both homophobic name-calling and sexual harassment involvement. This consistent relation between increases in gendered harassment and increases in negative mental health outcomes is important in that it indicates that harassment has negative effects for all involved, regardless of gender and regardless of involvement as a victim or a perpetrator. It extends findings in the cross-sectional literature and limited longitudinal findings that document a connection between harassment victimization and negative outcomes for victims and females (Goldstein, Malanchuk, Davis-Kean, & Eccles, 2007) and reinforces the detrimental impact of these behaviors more broadly for perpetrators in addition to victims, and for males in addition to females.

Given the paucity of literature related to the outcomes for adolescent perpetrators of harassment, the finding that both sexual harassment and homophobic name-calling perpetration is associated with increased depressive symptoms and substance use is especially worth highlighting. These results point to the negative effect of performing these acts, which makes sense given that harassment is a means of devaluing and undermining the value of individual autonomy and respect. Whether or not students consciously understand harassment as a devaluation of their peers’ and subsequently their own worth, the associated decreased mental health indicates that harassment does indeed serve to diminish students’ well-being and positive sense of self.

In addition, these results reinforce that gendered harassment has negative effects over relatively long periods of time, as increases in harassment from age 13 to 17 affected mental health outcomes over the same time span. Given the long-term implications of harassment on individuals’ mental health, the high stakes of prevention and intervention efforts in middle school is highlighted. Increasing prevention efforts in early middle school could have far-reaching effects on students’ well-being, including bolstering protective factors against depression and substance abuse. These results reinforce the necessity of schools’ compliance with Title IX, given the direct mental health implications for students who are involved in sexual harassment and homophobic name-calling.

Although the current study is methodologically strong and includes a large longitudinal sample, there are some limitations. Sexual orientation, which had low response rates in the current study, and factors like childhood abuse and maltreatment, were each beyond the scope of this study but are important to consider in future research given their potential impact on the relations between gendered harassment and mental health outcomes. In addition, the data considered here are self-report and do not include teacher-report or observation. These other forms of data collection can provide nuance and alternative viewpoints that are not captured from student self-report alone. However, the use of longitudinal data and factor analysis limits this self-report bias. The sample is from the U.S. Midwest and therefore may not be generalizable to other geographical areas. In addition, the potential impact of the school climate was not included in these analyses as multilevel modeling was not employed due to the low number of schools surveyed; given the impact of the surrounding environment on individuals’ experience of harassment, school environment is important to consider (Rinehart & Espelage, 2015).

The findings presented here reinforce the negative impact of gendered harassment for all involved, from those who engage in homophobic name-calling toward friends to those who enact or experience sexual assault. For victims and perpetrators alike, these behaviors have lasting effects and need to be addressed. Despite teachers’ potential discomfort with intervening when harassment occurs, it is clear that preventing further acts is important for students’ health (Charmaraman et al., 2013; Meyer, 2008). Administrators and teachers have a legal responsibility to address sexual harassment, and these findings point to the lasting effects of shirking this responsibility.

In addition, many harassment and bullying prevention programs as well as the broader discussion about safety in schools miss the mark by ignoring sexual harassment and homophobic name-calling (Kimmel & Mahler, 2003; Klein & Chancer, 2000). Most antibullying programs overlook the topic of homophobia and homophobic harassment altogether in both policy and practice. Among 23 major antibullying intervention programs, not even one offered intervention strategies for homophobic victimization (Birkett, Espelage, & Stein, 2008). Therefore, programs that purport to prevent peer victimization and subsequently improve students’ functioning must also address the existence of gendered harassment.

Given the negative effects of harassment, students benefit when schools engage in far-reaching prevention and intervention efforts. Schoolwide policies that promote gender equity and are intolerant of gendered harassment are associated with decreased sexual harassment (Rinehart & Espelage, 2015). These policies therefore should include clear language regarding teachers’ responsibility for intervening when harassment is witnessed, and proactive teaching regarding the unacceptability of gendered harassment.

Professional development regarding the impact of these acts, which are often normalized and ignored, is needed so that teachers and other school professionals can better understand the importance of intervening when harassment occurs and can be empowered to employ prevention efforts within their classrooms (Chiodo et al., 2009). Also, extant literature consistently finds that gendered harassment during early adolescence is driven by peers or friendship groups (Birkett & Espelage, 2015), thus, prevention programs that encourage bystanders to intervene should be considered. Finally, school psychologists need to be aware of the potential involvement of gendered harassment when a student presents with depression or substance use issues; psychologists can therefore assess for gendered harassment involvement when they are working with a student with these mental health outcomes. The negative mental health outcomes for students who are involved in gendered harassment as either perpetrators or victims, or both, indicates that schools’ efforts to prevent these behaviors from ever occurring, and intervening when they do occur, is imperative for our students’ well-being.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by funding from the Centers for Disease Control (No. 1 U49 CE001268; Espelage, principal investigator [PI]) and from the National Institute of Justice (No. 2011-90948-IL-IJ; Espelage, PI).

Author Biographies

Sarah J. Rinehart, MS, is a doctoral candidate in counseling psychology in the Department of Educational Psychology at the University of Illinois, Urbana–Champaign. Her work focuses on the impact of sexual harassment and homophobia on adolescent mental health.

Dorothy L. Espelage, PhD, is professor of psychology at the University of Florida. She is the recipient of the American Psychological Association (APA) Lifetime Achievement Award in Prevention Science and the 2016 APA Award for Distinguished Contributions to Research in Public Policy, and is a fellow of American Psychological Society (APS), APA, and American Educational Research Association (AERA). Over the past 20 years, she has authored more than 140 peer-reviewed articles, five edited books, and 30 chapters on bullying, homophobic teasing, sexual harassment, dating violence, and gang violence.

Kristen L. Bub, EdD, is associate professor and child development division chair in the Department of Educational Psychology at the University of Illinois, Urbana–Champaign. Her research centers on the role that early experiences play in supporting (or hindering) positive outcomes in young children by examining risk and protective factors in the home and school contexts.

References

American Association of University Women Educational Foundation. (2001). Hostile hallways: Bullying, teasing and sexual harassment in America’s schools. Washington, DC: Author.

Benson

D. J.

Thomson

G. E.

(1982). Sexual harassment on a university campus: The confluence of authority relations, sexual interest and gender stratifications. Social Problems, 29, 236-251.

Birkett

Espelage

D. L.

(2015). Homophobic name-calling, peer-groups, and masculinity: The socialization of homophobic behavior in adolescents. Social Development, 24, 184-205. doi:10.1111/sode.12085

Birkett

M. A.

Espelage

D. L.

Stein

(2008, August). Have school anti-bullying programs overlooked homophobic bullying? Poster Presented at the American Psychological Association Annual Convention, Boston, MA.

Bucchianeri

M. M.

Eisenberg

M. E.

Wall

M. M.

Piran

Neumark-Sztainer

(2013). Multiple types of harassment: Associations with emotional well-being and unhealthy behaviors in adolescents. Journal of Adolescent Health, 54, 724-729. doi:10.1016/j.jadohealth.2013.10.205

Charmaraman

Jones

A. E.

Stein

Espelage

D. L.

(2013). Is it bullying or sexual harassment? Knowledge, attitudes, and professional development experiences of middle school staff. Journal of School Health, 83, 483-444.

Chiodo

Wolfe

D. A.

Crooks

Hughes

Jaffe

(2009). Impact of sexual harassment victimization by peers on subsequent adolescent victimization and adjustment: A longitudinal study. Journal of Adolescent Health, 45, 246-252.

Collins

L. M.

Schafer

J. L.

Kam

C.-M.

(2001). A comparison of inclusive and restrictive strategies in modern missing data procedures. Psychological Methods, 6, 330-351.

Connell

R. W.

(1987). Gender and power: Society, the person and sexual politics. Cambridge, UK: Polity Press.

10.

D’Amico

E. J.

Tucker

J. S.

Miles

J. N. V.

Zhou

A. J.

Shih

R. A.

Green

H. D.

(2012). Preventing alcohol use with a voluntary after school program for middle school students: Results from a cluster randomized controlled trial of project choice. Prevention Science, 13, 415-425. doi:10.1007/s11121-011-0269-7

11.

D’Augelli

A. R.

Pilkington

N. W.

Hershberger

S. L.

(2002). Incidence and mental health impact of sexual orientation victimization of lesbian, gay, and bisexual youths in high school. School Psychology Quarterly, 17, 148-167.

12.

Espelage

D. L.

Aragon

S. R.

Birkett

(2008). Homophobic teasing, psychological outcomes, and sexual orientation among high school students: What influence do parents and schools have? School Psychology Review, 37, 202-216.

13.

Espelage

D. L.

Basile

K. C.

Hamburger

M. E.

(2012). Bullying experiences and co -occurring sexual violence perpetration among middle school students: Shared and unique risk factors. Journal of Adolescent Health, 50, 60-65. doi:10.1016/j.jadohealth.2011.07.015

14.

Farrell

A. D.

Kung

E. M.

White

K. S.

Valois

(2000). The structure of self-reported aggression, drug use, and delinquent behaviors during early adolescence. Journal of Clinical Child Psychology, 29, 282-292. doi:10.1207/S15374424jccp2902_13

15.

Fineran

(2002). Sexual harassment between same-sex peers: Intersection of mental health, homophobia, and sexual violence in schools. Social Work, 47, 65-74.

16.

Goldstein

S. E.

Malanchuk

Davis-Kean

Eccles

J. S.

(2007). Risk factors of sexual harassment by peers: A longitudinal investigation of African American and European American adolescents. Journal of Research on Adolescence, 17, 285-300.

17.

Gruber

J. E.

Fineran

(2007). The impact of bullying and sexual harassment on middle and high school girls. Violence Against Women, 13, 627-643. doi:10.1177/1077801207301557

18.

Hill

J. P.

(1983). Early adolescence: A research agenda. The Journal of Early Adolescence, 3, 1-21.

19.

Hodges

E. V. E.

Perry

D. G.

(1999). Personal and interpersonal antecedents and consequences of victimization by peers. Journal of Personality and Social Psychology, 76, 677-685. doi:10.1037/0022-3514.76.4.677

20.

Hox

Stoel

R. D.

(2005). Multilevel and SEM approaches to growth curve modeling. In Everitt

Howell

(Eds.), Encyclopedia of statistics in behavioral science (pp. 1296-1305). Chichester, UK: John Wiley.

21.

Juvonen

Graham

(Eds.). (2001). Peer harassment in school: The plight of the vulnerable and victimized. New York, NY: The Guilford Press.

22.

Kimmel

M. S.

Mahler

(2003). Adolescent masculinity, homophobia, and violence: Random school shootings, 1982-2001. American Behavioral Scientist, 46, 1439-1458. doi:10.1177/0002764203046010010

23.

Klein

Chancer

L. S.

(2000). Masculinity matters: The omission of gender from high-profile school violence cases. In Spina

S. U.

(Ed.), Smoke and mirrors: The hidden content of violence in schools and society. Lanham, MD: Rowman & Littlefield.

24.

Martin-Storey

Crosnoe

(2014). Peer harassment and risky behavior among sexual minority girls and boys. American Journal of Orthopsychiatry, 84, 54-65. doi:10.1037/h0098854

25.

Meyer

E. J.

(2008). Gendered harassment in secondary schools: Understanding teachers’ (non) interventions. Gender and Education, 20, 555-570. doi:10.1080/09540250802213115

26.

Orpinas

(1993). Skills training and social influences for violence prevention in middle schools. A curriculum evaluation (Unpublished doctoral dissertation). University of Texas Health Science Center at Houston, School of Public Health, Houston.

27.

Pepler

D. J.

Craig

W. M.

Connolly

J. A.

Yuile

McMaster

Jiang

(2006). A developmental perspective on bullying. Aggressive Behavior, 32, 376-384. doi:10.1002/ab.20136

28.

Peterson

J. L.

Hyde

J. S.

(2009). A longitudinal investigation of peer sexual harassment victimization in adolescence. Journal of Adolescence, 32, 1173-1188. doi: 10.1016/j.adolescence.2009.01.011

29.

Poteat

V. P.

Espelage

D. L.

(2007). Predicting psychosocial consequences of homophobic victimization in middle school students. The Journal of Early Adolescence, 27, 175-191. doi:10.1177/0272431606294839

30.

Rinehart

S. J.

Espelage

D. L.

(2015). A multilevel analysis of school climate, homophobic name-calling, and sexual harassment victimization/perpetration among middle school youth. Psychology of Violence, 6, 213-222. doi:10.1037/a0039095

31.

Stein

N. D.

(1995). Sexual harassment in K-12 schools: The public performance of gendered violence. Harvard Educational Review, 65, 145-162.

32.

Taylor

Stein

(2007). Preliminary pilot research results from on-going National Institute of Justice research grant, #2005-WT-BX-002. Development and evaluation of sexual violence/ harassment prevention programs in middle schools. Unpublished manuscript.