Corresponding Author: Leila Wood, Research Assistant Professor, Institute on Domestic Violence & Sexual Assault, Steve Hicks School of Social Work, The University of Texas at Austin, 1925 San Jacinto Blvd.,Stop D3500, Austin, TX 78712-1405, USA. ude.saxetu.nitsua@doowaliel
The publisher's final edited version of this article is available at J Interpers ViolenceSexual harassment is a pervasive problem on college campuses. Across eight academic campuses, 16,754 students participated in an online study that included questions about sexual harassment victimization by a faculty/staff member or by a peer since enrollment at their Institution of Higher Education (IHE). Utilizing an intersectional theory and hurdle models, this study explored the effects of gender, race/ethnicity, sexual orientation, age at enrollment, student status, and time spent at institution on students’ risk for peer- and faculty/staff-perpetrated sexual harassment victimization, as well as the extent of victimization for students who experience harassment. Across institutions, 19% of students reported experiencing faculty/staff-perpetrated sexual harassment and 30% reported experiencing peer-perpetrated sexual harassment. Hypotheses related to intersectional impacts were partially supported, with most significant findings in main effects. Time at institution was found to increase both risk and extent of victimization of both types of harassment. Traditional undergraduate students, non-Latinx White students, female students, and gender and sexual minority students were found to be at increased risk for harassment. Being female increases the odds of experiencing both faculty/staff and peer sexual harassment by 86% and 147%, respectively. Latinx students and students with an ethnicity other than White reported less victimization, but those who reported sexual harassment faced greater extent of harassing behaviors. A discussion of these findings for institutional program planning and policy is explored.
Keywords: sexual harassment, GLBT, anything related to sexual harassmentMany students at Institutions of Higher Education (IHEs) experience sexual harassment and its deleterious mental health, physical health, and academic outcomes. Evidence indicates that 50% to 90% of undergraduate women experience sexual harassment while in school (Cantor et al., 2015; Dziech, 2003; Hill & Silva, 2005; Yoon, Funk, & Kropf, 2010). Students may face harassment from a variety of perpetrators, including peers and IHE faculty and staff (Cortina, Swan, Fitzgerald, & Waldo, 1998; Dziech, 2003; Fitzgerald et al., 1998; Rosenthal, Smidt, & Freyd, 2016). Negative outcomes of sexual harassment include depression, posttraumatic stress disorder (PTSD), shame, alcohol use, and disruptions to academic experiences (Avina & O’Donahue, 2002; Shinsako, Richman, & Rospenda, 2001; Street, Gradus, Stafford, & Kelly, 2007). Under federal guidance provided by Title IX, IHEs that receive federal funding must investigate sexual harassment incidents and remedy resulting sex discrimination (Block, 2012). When IHEs fail to respond adequately, victims’ trauma symptoms may be exacerbated (Smith & Freyd, 2013). Sexual harassment research has historically paid little attention to overlapping identities of gender, race/ethnicity, and sexual orientation (Buchanan, Bergman, Bruce, Woods, & Lichty, 2009; Calafell, 2014; Lundy-Wagner & Winkle-Wagner, 2013). Emerging evidence and theory indicate these identities and their intersections may influence sexual harassment risk and outcomes (Cantor et al., 2015; Hill & Silva, 2005; McGinley, Wolff, Rospenda, Liu, & Richman, 2016; Yoon et al., 2010). Using intersectional theory to inform analysis, this study uses data from a survey of students across eight academic campuses in one state to examine differences in rates, risk, and extent of faculty/staff- and peer-perpetrated sexual harassment victimization.
Initially conceived to describe women’s experiences with unwanted workplace sexual attention, “sexual harassment” includes a variety of inappropriate behaviors and circumstances that can take place in a number of settings, including educational institutions. At American IHEs, discussions of sexual harassment are informed by Title IX of the Educational Amendments of 1972, which prohibits sex discrimination at educational institutions that receive federal funds (Huerta, Cortina, Pang, Torges, & Magley, 2006). The Office for Civil Rights (OCR) in the Department of Education oversees Title IX implementation. OCR outlines two types of sexual harassment: quid pro quo, or requesting sexual favors in exchange for a benefit, and hostile environment harassment, when an educational environment is disrupted by severe or persistent offensive behavior (Hill & Silva, 2005). Sexual harassment can also be based on gender or perceived gender.
Sexual harassment prevalence estimates at IHEs vary, primarily based on gender and student status. An Association of American Universities (AAU) study found that, among female undergraduates at 27 IHEs, 61.9% had been sexually harassed by a student or IHE employee since enrolling (Cantor et al., 2015). However, the AAU study differs methodologically from the present study in important ways, most notably Cantor et al.’s exclusion of gender-based sexist harassment behaviors. In Yoon et al.’s (2010) sample, 50% of the undergraduate women had been harassed at least once, and 22.2% had been harassed multiple times. Peer-perpetrated harassment is more common than harassment by faculty or staff (Dziech, 2003; Hill & Silva, 2005; Hoewing & Rumburg, 2005; Rosenthal et al., 2016). Rosenthal et al. (2016) found that 38% of female graduate students and in their sample had experienced faculty/ staff-perpetrated sexual harassment, while 57.7% had experienced peer-perpetrated harassment. The most frequent types of sexual harassment experienced by students at IHEs are verbal comments and jokes and nonverbal gestures (Clodfelter, Turner, Hartman, & Kuhns, 2010; Hill & Silva, 2005). Individuals with different gender, racial, and cultural identities often label harassment behaviors different, complicating efforts to establish harassment prevalence (Bursik & Gefter, 2011).
Sexual harassment is correlated with many negative outcomes, including depression, posttraumatic stress disorder (PTSD) symptoms, diminished mental health, perceived isolation and helplessness, internalized shame, disordered eating, problematic alcohol use, nausea, and sleeplessness (Avina & O’Donahue, 2002; Huerta, Cortina, Pang, Torges, & Magley, 2006; Shinsako et al., 2001; Street et al., 2007). Many harassed students experience negative academic effects, such as decreased academic satisfaction, perceptions of faculty, engagement, and performance (Cortina et al., 1998; Huerta et al., 2006; Rosenthal et al., 2016). Cultural context, gender, race, and ethnicity also influence postharassment coping styles (Hill & Silva, 2005; Wasti & Cortina, 2002). For example, among female students who experienced harassment, 48% indicated avoiding the harasser, 27% indicated staying away from particular buildings, and 9% indicated skipping or dropping a class, compared to 26%, 11%, and 4% of male students, respectively (Hill & Silva, 2005). Students’ efforts to protect themselves, like dropping classes, changing advisors, switching majors, skipping class, and dropping out of school, may further disrupt their educational experience (Fitzgerald et al., 1988; Hill & Silva, 2005; Huerta et al., 2006).
Identity positions related to age, gender, sexual orientation, and race/ethnicity may affect sexual harassment victimization experiences. Women experience sexual harassment at higher rates than men (Berdahl, Magley, & Waldo, 1996; McGinley et al., 2016; Rosenthal et al., 2016). Both gender and sexual minority students and younger students also face elevated harassment rates (Cantor et al., 2015; Hill & Silva, 2005; Silverschanz, Cortina, Konik, & Magley, 2007).
Evidence regarding race and sexual harassment is complex and often contradictory. While some research has found that non-Latinx White women experience harassment more frequently than other groups (Hill & Silva, 2005; Kearney & Gilbert, 2012), other evidence shows greater risk or severity of harassment for women of color, particularly Black and multiracial women (Buchanan et al., 2009; Yoon et al., 2010). Moylan and Wood (2016) found that Latinas reported elevated rates of harassment at their field placements than did other social work students. Students of color are marginalized in many ways that interact with harassment experiences. When Black students face interracial sexual harassment, it is likely to include racialized harassment and associated with greater posttraumatic stress (Woods, Buchanan, & Settles, 2009). Students of color face barriers accessing services, including racism, language barriers, and cultural norms related to gender roles, sexuality, and mental health (Angelone, Mitchell, & Hirschman, 2006). Building on the work of Crenshaw (1991), Lockhart and Mitchell (2010) apply intersectional theory to understand that individuals many simultaneously held identities like race, gender, sexual orientation, class, age, and culture, interact with each other and with environmental factors to impact experiences, including violence and oppression. Else-Quest and Hyde (2016) describe intersectional research as that which recognizes that people are defined by multiple and connected social categories, which have inherent aspects of power and inequality that vary with social context and environment.
Informed by this intersectional lens, this study examines ways that identity positions, individually and in interaction, affect risk for harassment among IHE students. This study uses a diverse sample of students at eight campuses in a Southwestern state to analyze rates of faculty/staff-perpetrated and peer-perpetrated sexual harassment. The study further explores how gender, race/ethnicity, age, time at institution, and sexual orientation interact to affect risk and the extent of victimization for sexual harassment experiences among IHE students. The hypotheses guiding the study are the following:
Participant gender, age, sexual orientation, race/ethnicity, time at institution, and student status will interact to predict differences in risk for sexual harassment perpetrated by a faculty/staff member.
Participant gender, age, sexual orientation, race/ethnicity, time at institution, and student status will interact to predict differences in risk for sexual harassment perpetrated by a peer.
Participant gender, age, sexual orientation, race/ethnicity, time at institution, and student status will interact to predict differences in extent of sexual harassment perpetrated by a faculty/staff member.
Participant gender, age, sexual orientation, race/ethnicity, time at institution, and student status will interact to predict differences in extent of sexual harassment perpetrated by a peer.
Data for this study came from a larger research study conducted across a university system in the Southwest United States, which included an online survey of prevalence and perceptions, or climate survey. The larger survey measured the prevalence of (a) sexual harassment, (b) stalking, (c) dating/domestic violence, and (d) sexual assault. The university system includes eight academic IHEs, ranging between 4,000 and over 50,000 enrolled students. Depending on campus size, a representative random sample or census sample was drawn with assistance from the registrar to recruit student participants. Students were eligible to participate if they were currently enrolled at a participating academic IHE and 18 or older. Invitations to participate in the anonymous online survey were sent via email, with four reminders. The survey was open 5 weeks in the fall of 2015, and was designed and distributed via Qualtrics (2016). Participants were entered into an incentive drawing. The survey took 20 min to 40 min to complete. Across all IHEs, 186,790 students were invited to participate and 26% of those invited began the survey. The total sample across all IHEs was 26,417 for an average completed response rate of 14.1%. Surveys were considered complete if the participant finished victimization measures, about 75% of the way through the survey. Participants were randomized to different paths through the survey to reduce length, resulting in a sample of 17,406 participants who were asked questions about sexual harassment. Additional sample criteria of answering specific demographic questions produced a final sample of 16,754 participants.
The study used a modified version of the Administrator Research Campus Climate Consortium (ARC3) survey, a collection of reliable and valid measures of violence in the lives of IHE students. The ARC3 instrument was developed collaboratively by a national panel of experts (Administrator Researcher Campus Climate Consortium [ARC3], 2016).
Sexual harassment was measured with the Department of Defense Sexual Experiences Questionnaire (SEQ-DoD; Fitzgerald, Gelfand, & Drasgow, 1995; Fitzgerald, Magley, Drasgow, & Waldo, 1999). Two versions of the SEQ were administered, one asking about faculty- and staff-perpetrated sexual harassment, and one asking about peer-perpetrated sexual harassment. On both measures, participants were asked to indicate the number of times since their enrollment at the IHE when, respectively, a faculty, instructor, or staff member or a student did any of the listed behaviorally specific items. The peer measure assesses for peer behaviors both within and outside educational settings, meaning the survey captured peer harassment incidents that occurred in social or other settings. The faculty- and staff-perpetrated SEQ was administered in 16 questions with four subscales. The peer-perpetrated SEQ was administered in 12 questions with four subscales. For each scale, follow up questions were asked about the most impactful incident, including incident location and the gender, age, and professional or academic status of the perpetrator(s). Achieved reliability of the faculty and staff SEQ was 0.947 and achieved reliability of the peer SEQ was 0.914 for this study.
Analyses were conducted in SPSS 24 and with the glmmADMB package 0.8.33 (Fournier et al., 2012) in R 3.4.0 within RStudio 1.1.383. Descriptive statistics and bivariate inferential (t and z tests) were calculated using Custom Tables in SPSS version 24 (IBM Corporation, 2016) with the test chosen as appropriate for the scale of the variables involved with a significance threshold of 0.05.
Hurdle models with random intercepts were used to test hypotheses about the risk and extent of sexual harassment victimization. Models were fit using the glmmADMB package with random intercepts specified for individual campuses. The fixed effect coefficients were interpreted with a significance threshold of 0.05. Hurdle models are appropriate for processes wherein certain dynamics impact some members of the population more so than others, and once affected, those dynamics contribute to the extent of the impact (Loeys, Moerkerke, De Smet, & Buysse, 2012). Expressed nonmathematically, individuals need to cross a “hurdle” before impact is experienced; once the hurdle is crossed, the impact can accumulate with additional events. The likelihood of crossing the hurdle is typically modeled with a binary logistic regression model and the extent of impact is modeled with a count model, for example, a Poisson or negative-binomial regression model. Hurdle models incorporate the likelihood of experiencing any events and, given that the first one is experienced, the expected rate at which additional events will be experienced (Loeys et al., 2012).
In the present study, the response variable of interest is sexual harassment victimization, and the hurdle model predicted risk and extent of victimization. A logistic regression model (binomial distribution and logit link) was specified for the hurdle portion of the model, referred to below as the risk portion of the model. A zero-censored negative-binomial regression (zero-censored negative-binomial distribution with logit link function) was specified for the count portion of the model, referred to as the extent of victimization portion of the model. The predictors included gender, sexual orientation, race/ethnicity, age at time of enrollment, time in the program, and student status (see Equation 1). The same set of predictors were used for both the risk and extent of victimization portions of the hurdle models:
Y i t = β o i t + β 1 i t time i + β 2 i t status i + β 3 i t time i × status i + β 4 age 0 i × status i + β 5 i gender i + β 6 i t sexOrient i + β 7 i t ethnicity i + β 8 i t gender i × sexOrient i + β 9 i t gender i × ethnicity i + β 10 gender i × status i ,
where Y i t is the hurdle model response (risk or count), i indexes participants, t is the type of sexual harassment (faculty/staff or peer), and a g e 0 is the age at enrollment in the current program.
Some demographic categories were collapsed for analysis due to small sample sizes. Gender was represented with three levels: male, female, and gender minority, which includes transgender male, transgender female, nonbinary, genderqueer, and gender nonconforming participants. 1 Sexual orientation was collapsed to two levels: heterosexual and sexual minority. 2 Race and ethnicity categories have been collapsed to non-Latinx White, Latinx, and additional races/ethnicities due to demographic composition of the area. 3 Respondents were asked about their age (in years) at the time of the survey and the year that they enrolled at the IHE for their current program. The respondent’s age at enrollment in their current program and the number of years since enrolling were calculated from the responses to these questions. Respondents also provided information about whether they were caring for children or other adults, their residence type, and whether they were classified as undergraduate students or graduate/professional students. These responses were combined with respondent age at enrollment to classify them as a traditional undergraduate student (enrolled when they were 24 years of age or younger and not caring for children or other adults), a nontraditional undergraduate student (25 or older when enrolled or caring for children or older adults), or a graduate/professional student (age and caring for others not considered).
Achieved sample size provided power of greater than or equal to .8 for small effect sizes in the analyses planned for the current study. Gender minority participants make up 1% of the sample; while this group is underpowered in the present study, the ability to provide insights even at a reduced level of confidence is one of the contributions of this study. Accordingly, summary tables for the models include estimates and significance for all effects. Data results are aggregated across campuses to enhance analysis and protection for participants.
Table 1 summarizes the demographics for the sample in total. The total sample included 16,754 participants. For all participants, mean age at enrollment was 23.3 years and mean time at institution was 1.2 years
Demographics: Total Sample.
| Age at enrollment (Categories) | |
| 18 | 34.2% |
| 19 | 14.9% |
| 20–21 | 10.3% |
| 22–24 | 13.3% |
| 25–29 | 11.8% |
| 30–34 | 5.9% |
| 35–39 | 3.5% |
| 40 or more | 5.9% |
| Status | |
| Traditional undergraduate | 57.2% |
| Nontraditional undergraduate a | 16.5% |
| Graduate/professional | 26.4% |
| Current gender | |
| Male | 35.2% |
| Female | 63.7% |
| Gender minority | 1.0% |
| Sexual orientation | |
| Heterosexual | 88.8% |
| Sexual minority | 11.2% |
| Race/ethnicity | |
| Non-Latinx White | 33.8% |
| Latinx | 36.7% |
| Additional ethnicities | 29.5% |
a Participants were classified as nontraditional undergraduate students if they were 25 years or older at the time of enrollment or were caring for children or older adults.
Across eight campuses, 19% of all students had experienced faculty/staff-perpetrated sexual harassment (range 27.3%−11.8%) and 30% had experienced peer-perpetrated sexual harassment (range 46.4%−17.1%). See Table 2 for more details.
Faculty/Staff- and Peer-Perpetrated Sexual Harassment Victimization Rates.
| Faculty/Staff (%) | Peer (%) | |
|---|---|---|
| Total sample | 19.3 | 29.9 |
| Status | ||
| Traditional undergraduate | 21.2 | 36.5 |
| Nontraditional undergraduate | 13.2 | 19.7 |
| Graduate/professional | 19.1 | 21.9 |
| Gender | ||
| Male | 14.6 | 20.9 |
| Female | 21.5 | 34.3 |
| Gender minority | 48.6 | 65.9 |
| Sexual orientation | ||
| Heterosexual | 17.8 | 27.7 |
| Sexual minority | 31.4 | 47.4 |
| Race/ethnicity | ||
| Non-Latinx White | 24.0 | 37.3 |
| Latinx | 15.8 | 23.7 |
| Additional ethnicities | 18.4 | 28.9 |
Of people that experienced victimization by a faculty or staff member, 75.9% have experienced two or more behaviors since enrolling. Half of the victims have experienced three behaviors by faculty or staff member. Of the people that experienced victimization by a student, 86.1% experienced two or more behaviors. Half had experienced six behaviors since enrolling. Among undergraduates, freshman report a rate of faculty harassment of 12.5%, while 24.9% of seniors reported faculty harassment, indicating rates increasing overtime. Likewise, 28.6% of freshman reported peer harassment, while 35.7% of seniors indicate peer harassment experiences.
Perpetrators 4 of faculty/staff sexual harassment incidents were 78% male-identified, 15% were female-identified and 7% were of a gender unknown to the participant. A large majority of incidents, 84%, occurred on campus. Totally, 61% of these perpetrators were faculty members, 27% were student employees, and 12% were staff. Of perpetrators of peer sexual harassment, 86% were male-identified, 11% were female-identified, and 3% were of a gender unknown to the participant. Overall, 85% of peer perpetrators were undergraduates, while 15% were graduate/professional students. Totally, 68% of incidents of peer-perpetrated harassment occurred on campus.
Both the faculty/staff and the peer sexual harassment hurdle models with random intercepts are highly significant relative to their respective intercept-only model and their baseline time/exposure-controlled model. See Table 3 for details.
| Faculty/Staff Sexual Harassment | Peer Sexual Harassment | |||
|---|---|---|---|---|
| Model | Risk | Extent | Risk | Extent |
| Controlling for campus differences as random intercepts | ||||
| Delta χ 2 (1) | 265 | 1.54 | 898 | 19 |
| sig. | < 0.000 | 0.215 | < 0.000 | < 0.000 |
| Controlling for years of exposure | ||||
| Delta χ 2 (1) | 330 | 16 | 236 | 82 |
| sig. | < 0.000 | < 0.000 | < 0.000 | < 0.000 |
| Predictor significance | ||||
| Delta χ 2 (22) | 577 | 43 | 1411 | 196 |
| sig. | < 0.000 | < 0.000 | < 0.000 | < 0.000 |
A hurdle model was used to understand the risk and extent of faculty/staff sexual harassment and test hypothesis 1 and 3. The reference case for this model is a traditional undergraduate, heterosexual, non-Latinx White male who was 18 and in his first year of school at the time of the survey. The coefficient estimates are provided as logits with the implication that a positive estimate corresponds to increased risk and a negative estimate corresponds to decreased risk.
As shown in Table 4 , time plays a critical role relative to risk of experiencing faculty/staff-perpetrated sexual harassment with increased exposure (time) increasing the likelihood of victimization (estimate = .25, sig < 0.000). Being female (estimate = .62, sig < 0.000), a sexual minority (estimate = .69, sig < 0.000), or gender minority (estimate = 1.17, sig < 0.029) also increases the risk.
Faculty/Staff Risk and Extent.
| Risk of Victimization Model | Extent of Victimization Model | |||
|---|---|---|---|---|
| Estimate | Pr(>|z|) | Estimate | Pr(>|z|) | |
| (Intercept) | −1.25 | 0.002 ** | 0.68 | 0.282 |
| Time | 0.25 | 0.000 *** | 0.04 | 0.165 |
| Nontraditional undergraduate | 0.28 | 0.536 | −0.09 | 0.902 |
| Graduate/professional | 0.03 | 0.953 | −0.51 | 0.447 |
| Female | 0.62 | 0.000 *** | 0.18 | 0.175 |
| Gender minority | 1.17 | 0.029 * | 0.51 | 0.510 |
| Sexual minority | 0.69 | 0.000 *** | 0.22 | 0.187 |
| Latinx | −0.45 | 0.000 *** | 0.22 | 0.139 |
| Additional ethnicity | −0.37 | 0.000 *** | 0.64 | 0.000 *** |
| Time × Nontraditional Undergraduate | −0.06 | 0.077 | 0.01 | 0.859 |
| Time × Graduate/Professional | −0.03 | 0.353 | 0.12 | 0.013 * |
| Age0 × Traditional Undergraduate | −0.04 | 0.045 * | 0.00 | 0.959 |
| Age0 × Nontraditional Undergraduate | −0.04 | 0.000 *** | 0.00 | 0.805 |
| Age0 × Graduate/Professional | −0.03 | 0.000 *** | 0.01 | 0.285 |
| Female sexual minority | −0.07 | 0.599 | 0.01 | 0.941 |
| Gender and sexual minority | 0.06 | 0.902 | −0.02 | 0.981 |
| Female Latinx | −0.25 | 0.019 * | −0.33 | 0.050 |
| Gender minority Latinx | −0.34 | 0.386 | −0.45 | 0.387 |
| Female additional ethnicity | 0.02 | 0.891 | −0.50 | 0.004 ** |
| Gender minority additional ethnicity | 0.43 | 0.297 | −1.23 | 0.018 * |
| Nontraditional undergraduate female | −0.33 | 0.018 * | 0.17 | 0.463 |
| Graduate/professional female | 0.11 | 0.316 | 0.06 | 0.704 |
| Nontraditional undergraduate gender minority | −0.60 | 0.198 | 0.01 | 0.983 |
| Graduate/professional gender minority | 0.08 | 0.868 | 0.24 | 0.675 |
Taken together, both portions of the faculty/staff model quantify the degree to which additional time increases risk and differentially increases the expected rate of experiencing harassing behaviors by graduate/professional students. Being female increases risk of experiencing sexual harassment but does not differentially increase the rate of harassing behaviors experienced. In contrast, while being Latinx or an additional ethnicity reduces risk of experiencing sexual harassment, if harassment is experienced, the rate of behaviors experienced is predicted to be higher. Increased age at enrollment decreases the likelihood of experiencing sexual harassment, but age at enrollment does not moderate the rate of experiencing harassing behaviors by status.
A hurdle model was used to understand the risk and extent of peer sexual harassment and test Hypotheses 2 and 4. As before, the reference case for this model is a traditional undergraduate, heterosexual, non-Latinx White male who was 18 and in his first year of school at the time of the survey.
Peer Risk and Extent.
| Risk of Victimization Model | Extent of Victimization Model | |||
|---|---|---|---|---|
| Estimate | Pr(>|z|) | Estimate | Pr(>|z|) | |
| (Intercept) | 1.15 | 0.002 ** | 1.86 | 0.000 *** |
| Time | 0.23 | 0.000 *** | 0.13 | 0.000 *** |
| Nontraditional undergraduate | −1.10 | 0.007 ** | 0.39 | 0.345 |
| Graduate/professional | −1.71 | 0.000 *** | 0.26 | 0.499 |
| Female | 0.90 | 0.000 *** | 0.44 | 0.000 *** |
| Gender minority | 1.17 | 0.029 * | 0.17 | 0.713 |
| Sexual minority | 0.98 | 0.000 *** | 0.17 | 0.054 |
| Latinx | −0.48 | 0.000 *** | −0.06 | 0.509 |
| Additional ethnicity | −0.55 | 0.000 *** | 0.01 | 0.867 |
| Time × Nontraditional Undergraduate | −0.12 | 0.000 *** | −0.04 | 0.236 |
| Time × Graduate/Professional | −0.09 | 0.001 ** | −0.04 | 0.126 |
| Age0 × Traditional Undergraduate | −0.14 | 0.000 *** | −0.01 | 0.581 |
| Age0 × Nontraditional Undergraduate | −0.06 | 0.000 *** | −0.02 | 0.001 ** |
| Age0 × Graduate/Professional | −0.05 | 0.000 *** | −0.01 | 0.030 * |
| Female sexual minority | −0.28 | 0.017 * | 0.11 | 0.301 |
| Gender and sexual minority | 0.32 | 0.534 | 0.48 | 0.322 |
| Female Latinx | −0.20 | 0.043 * | −0.33 | 0.001 *** |
| Gender minority Latinx | 0.02 | 0.970 | −0.17 | 0.550 |
| Female additional ethnicity | 0.13 | 0.178 | −0.09 | 0.333 |
| Gender minority additional ethnicity | −0.07 | 0.878 | 0.34 | 0.251 |
| Nontraditional undergraduate female | −0.32 | 0.009 ** | −0.02 | 0.862 |
| Graduate/professional female | −0.18 | 0.062 | −0.26 | 0.008 ** |
| Nontraditional undergraduate gender minority | −0.55 | 0.239 | −0.30 | 0.435 |
| Graduate/professional gender minority | 0.12 | 0.809 | −0.58 | 0.080 |
Table 6 below outlines the odd and rates ratios associated with the coefficients listed in Tables 4 and and5. 5 . Only the significant factor levels have been included in Table 6 . Each year of added exposure in the IHE setting increases the odds of experiencing faculty/staff perpetrated sexual harassment by 29% and the odds of experiencing of peer sexual harassment by 25%. While increased time of exposure does not significantly increase the expected rate at which faculty/staff behaviors are experienced, it does increase by 14% the expected rate at which peer sexual harassment behaviors are experienced. Being female increases the odds of experiencing both faculty/staff and peer sexual harassment by 86% and 147%, respectively. In addition, being female increases the expected rate of experiencing additional peer-perpetrated sexual harassment events by 56%. With odds ratios of 2.00 and 2.66, respectively, being a sexual minority substantially increases the risk of both faculty/staff (doubling) and peer (almost tripling) sexual harassment. While identifying as additional ethnicity (not nonLatinx White or Latinx) decreases the risk of experiencing both faculty/staff- and peer-perpetrated sexual harassment, being of an additional ethnicity increases the expected rate of additional behaviors by faculty or staff by 90% after an initial harassment experience.
Odds/Rates Ratios for Sexual Harassment.
| Ratios for Significant Predictors | ||||
|---|---|---|---|---|
| Faculty/Staff | Peer | |||
| Risk | Count | Risk | Count | |
| Baseline odds | 0.29 | 1.97 | 3.16 | 6.45 |
| Time | 1.29 | 1.25 | 1.14 | |
| Nontraditional undergraduate | 0.33 | |||
| Graduate/professional | 0.18 | |||
| Female | 1.86 | 2.47 | 1.56 | |
| Gender minority | 3.23 | 3.24 | ||
| Sexual minority | 2.00 | 2.66 | ||
| Latinx | 0.64 | 0.62 | ||
| Additional ethnicity | 0.69 | 1.90 | 0.58 | |
| Time × Nontraditional Undergraduate | 0.88 | |||
| Time × Graduate/Professional | 1.12 | 0.91 | ||
| Age0 × Traditional Undergraduate | 0.96 | 0.87 | ||
| Age0 × Nontraditional Undergraduate | 0.96 | 0.94 | 0.98 | |
| Age0 × Graduate/Professional | 0.97 | 0.95 | 0.99 | |
| Female sexual minority | 0.75 | |||
| Female Latinx | 0.77 | 0.82 | 0.72 | |
| Female additional ethnicity | 0.60 | |||
| Gender minority Additional ethnicity | 0.29 | |||
| Nontraditional undergraduate female | 00.72 | 0.72 | ||
| Graduate/Professional Female | 0.77 | |||
Exposure time does not differentially change the risk of faculty/staff perpetrated sexual harassment for nontraditional undergraduates or graduate students, but it does increase by 12% the expected rate of behaviors experienced from faculty or staff by graduate students. While time increases the odds of and expected rate of sexual harassment behaviors perpetrated by student among traditional undergraduates by 25% and 14%, respectively, per year, that rate is slightly lower for nontraditional undergraduates (12% lower) and graduate students (9% lower). Among those who experienced faculty/staff or peer harassment, increased age at enrollment differentially decreases the expected rate of harassment for all student statuses. It also decreases the expected rate of peer harassment among nontraditional undergraduates and graduate students compared to traditional undergraduates.
This study sought to identify predict risk and extent of sexual harassment by intersecting identity positions. Study hypotheses were partially supported, in particular for main effects. However, because main effects of independent identity positions were found to be impactful on risk overall as compared to interaction terms, or intersection of identity positions, this analysis is limited in its ability to meaningfully discuss intersectionality. The following discussion is, therefore, focused primarily on main effects for individual demographic factors; limitations and directions for future research are discussed below.
Female, gender minority, and sexual minority students are all at increased risk for both peer and faculty/staff harassment and face greater extent of peer harassment. These findings are similar to other research (see Cantor et al., 2015; Hill & Silva, 2005; McGinley et al., 2016). Time at institution increases risk for peer and faculty/staff harassment, extent of peer victimization, and among graduate/professional students, extent of faculty/staff harassment victimization. Age at enrollment is inversely related to risk of faculty/staff harassment and extent of peer harassment. Relative to nontraditional undergraduates, both traditional undergraduates and graduate/professional students are at higher risk for faculty/staff harassment, while traditional undergraduate students are at higher risk for peer victimization.
Gender and sexual minority predict significant increased risk for both peer and faculty/staff harassment. This is consistent with Hill and Silva (2005) and Silverschanz et al. (2007). Research on heterosexist harassment has linked it to depression, anxiety, problematic alcohol use, and lowered physical health (Meyer, 1995; Woodford, Kulick, & Atteberry, 2015). This study begins to contribute to a gap in the literature about the sexual harassment experiences of gender and sexual minority students, and indicates the urgent need for further research. Research needs include longitudinal and qualitative studies to increase understanding of impact and potential protective factors for gender and sexual minority students. This finding also indicates the need for focused outreach and support regarding sexual harassment to gender and sexual minority populations from faculty, staff, university health centers, and specialized centers at IHEs.
Findings regarding race and ethnicity were complex. While Latinx students and students identifying as an additional ethnicities were found to have decreased risk for both types of harassment, students identifying as additional ethnicities experienced greater extent of faculty/staff harassment. These findings are somewhat surprising given the literature that has suggested women of color experience higher rates of harassment (Moylan & Wood, 2016; Buchanan et al., 2009; Cortina et al., 1998), although other literature has found lower rates of harassment among Latinx women (Kearney & Gilbert, 2012; Shupe, Cortina, Ramos, Fitzgerald, & Salisbury, 2002). Regarding extent of victimization, the present study’s finding is consistent with Buchanan et al. (2009) and Yoon et al.’s (2010) findings around increased frequency of harassment experiences for women of color. Although these findings indicate students of color may not be at elevated risk for sexual harassment victimization at IHEs, they may remain vulnerable in other ways not addressed in this study. Societal and institutional racism may also lead to people of color experiencing dismissiveness, stereotyping, racialized harassment, an overall hostile campus climate, and elevated emotional discomfort and anger (Buchanan et al., 2009; Frederickson, 2002; Lundy-Wagner & Winkle-Wagner, 2013; Rankin & Reason, 2005). There continues to be an urgent need to design interpersonal violence and harassment screening and intervention protocols informed by culturally relevant theoretical frameworks that take into consideration the influence of multiple risk factors associated with interpersonal violence experiences against people of color (Klevens, 2007).
These findings also suggest potential measurement issues at the intersection of racial and sexual harassment. Lundy-Wagner and Winkle-Wagner (2013) and Calafell (2014) assert that artificially separating conversations of race and sexual harassment does a disservice to survivors and obscures the realities of intersectional oppression. Racial and sexual harassment cooccur, complicating measurement, as indicated by previous research. Woods et al. (2009) found when African American students face interracial sexual harassment, it may be more likely to include racialized harassment and be associated with increased posttraumatic stress. The SEQ-DoD may not capture the behavioral manifestations of sexual harassment intersecting with racism that affect Latinx students or other students of color, and may provide limited understanding of sexual harassment experiences for people of color (Cortina, 2001). This study may, therefore, underestimate the victimization experiences of women of color. Future sexual harassment studies with IHE students should develop and use new or adapted tools, such as Cortina’s (2001) modification of the SEQ for Latinas.
Unsurprisingly, traditional undergraduate students were at the highest risk for peer sexual harassment. Unlike previous findings, traditional undergraduate students were also at similar risk for faculty/staff harassment as graduate/professional students. This is inconsistent with literature indicating that graduate/professional students are more likely to experience harassment by faculty/staff, which may be due to increased interaction with and dependence on faculty (Cantalupo & Kidder, 2017; Rosenthal et al., 2016). Nontraditional undergraduates were at lower risk than both graduate/professional students and traditional undergraduates for both types of harassment. This may be an effect of exposure; nontraditional undergraduates often face competing demands for their time such as family responsibilities and off-campus employment, spending less time on campus with peers (Keith, 2007; Markle, 2015). While younger age students are at increased risk for harassment, time at the institution was also associated with increased risk and extent of both types of harassment. Students who have been at IHEs longer have had greater exposure to faculty and other students, and, therefore, more chances for victimization.
These findings have several implications for IHEs and future research. More research is needed about the interaction of identity positions in sexual harassment experiences to contribute to understanding of risk and extent. These findings have several implications for IHEs. IHEs should continue to recognize the reality of sexual harassment in their students’ lives, including the potential academic and mental health impacts of victimization, by addressing harassment through intervention and prevention. IHEs may unintentionally send messages to their students about appropriate gender, sexual, or racial norms. Yoon et al. (2010) recommend that IHEs pay particular attention to messaging, especially around athletics and masculinity, which may have unintended and harmful effect of endorsing sexism and aggression. Policies governing student and staff conduct must be culturally grounded and recognize that students’ identities affect potential harassment victimization (Calafell, 2014). Potential experiences of sexual harassment experienced by faculty and staff members from peers or from students is understudied area that merits additional research (Lampman, Phelps, Bancroft, & Beneke, 2009).
Violence prevention programs at IHEs can address sexual harassment through continued integration of harassment prevention strategies in existing programs, such as bystander intervention. Addressing gender bias and sexism as the underlying causes of harassment can provide a base to prevention work (Brinkman, Dean, Simpson, McGinley, & Rosen, 2015). Prevention and intervention strategies can use an intersectional approach to addressing sexual harassment by including racism and homophobia into programming. The recent National Academies of Sciences, Engineering and Medicine (NAS) report on sexual harassment in academia makes four recommendations toward culture change to end sexual harassment: (a) Integrate values of diversity and inclusion into policy and procedure, (b) Change power dynamics to diffuse advisor relationship dependencies, (c) Support sexual harassment survivors through services and reporting that minimizes risk for retaliation, and (d) Improve transparency and accountability (National Academies of Sciences, Engineering, and Medicine, 2018). These broad recommendations for science-based fields can easily apply in most academic settings.
This study has several limitations. In interpreting and contextualizing these findings, limitations due to measurement issues, at the intersection of sexual orientation and race, must be noted. The low survey response rate, which is typical of similar studies, can be improved but still represent student groups with minimal response bias concerns (Cook, Heath, & Thompson, 2000; Rosenthal & Freyd, 2018). Increased promotion and recruitment, especially in minority populations, would help with gaining significant statistical power to represent all groups in analysis. A major limitation of this study is the collapsing of all non-White and non-Latinx race and ethnicity categories into “additional ethnicities.” This approach, while utilized to provide sufficient statistical power for analysis, obscures interracial differences that may exist within this category, providing a simplified and potentially inaccurate representation for these students. Similarly, this study combines all gender identities other than male and female and all sexual orientations other than heterosexual. While underpowered for these groups, this study contributes to literature that points to heightened rates of sexual harassment for gender minority participants. Further research is required to fully understand these students’ experiences. Another limitation is the low average time at the institution for participants (1.2 years) indicating a younger sample and resulting in prevalence rates lower than those in similar studies. A sample with more students later in the undergraduate career likely would increase prevalence rates. There are many other aspects of students’ identities that may affect harassment risk that are not explored in this study. Future research should address factors such as disability and socioeconomic status, in particular. Finally, while the sample for this study was large, exploring intersections of gender, age, sexual orientation, race/ethnicity necessarily results in lower power for the corresponding intersectional estimates in the models used. Future studies would benefit from sampling plans to increase intersectional representation in the analysis. This will allow for more meaningful analysis of intersection effects of different identity variables, as described by Else-Quest and Hyde (2016).
Sexual harassment is a common problem for IHE students and is associated with significant negative outcomes for individuals and learning communities. Female and gender and sexual minority students, as well as traditional undergraduates are at increased risk for harassment. This study found that Latinx students and students with additional ethnicities were at decreased risk but faced greater extent of harassment. More research is needed to better understand population differences. The findings indicate IHEs must consider the impact of intersecting identities and forms bias and oppression while planning prevention and intervention response to sexual harassment. Broader culture change is needed to alter academic culture that allows sexual harassment to continue. Future research should address and explore the impact of sexual harassment on students with diverse and intersecting identities.
The authors wish to thank Dr. Lilia Cortina from University of Michigan for her review of this manuscript.
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Financial support for this research was provided by the University of Texas Board of Regents.
Leila Wood, PhD, MSW is a research assistant professor at the Institute on Domestic Violence & Sexual Assault (IDVSA) at The University of Texas at Austin School of Social Work. Her scholarship areas are intimate and interpersonal violence in the campus context, intimate partner violence and sexual assault service evaluation, and occupational stress related to interpersonal violence work. Wood’s research focuses on survivor-centered approaches and establishing evidence for community and campus-based practices in the intimate partner violence (IPV) and sexual assault (SA) field.
Sharon Hoefer, MSSW, is a research project manager at the Institute on Domestic Violence & Sexual Assault at The University of Texas at Austin Steve Hicks School of Social Work. She has worked with survivors of interpersonal violence in both research and community-based settings as a case manager and support group facilitator, and with children and young adults in several settings, including violence prevention education and social skills education for children with intellectual and developmental disabilities. Her professional interests include gender-based violence prevention and intervention, healthy relationship and sex education, program evaluation, and feminism in popular culture.
Matt Kammer-Kerwick, PhD, is a research scientist at the Bureau of Business Research (BBR) at The University of Texas at Austin where he is a coprincipal investigator on projects with the Institute on Domestic Violence & Sexual Assault. Current projects include the development of operations research methods and models toward the disruption of illicit networks and assessing the perceptions and prevalence of interpersonal violence at institutions of higher learning within the UT System. Prior to his role at BBR, he was President and Founder of Visionary Research, Inc. He has a PhD in Management Science and Information Systems.
José Rubén Parra-Cardona is an associate professor in the Steve Hicks School of Social Work (SHSSW) at The University of Texas at Austin. He is the coordinator for Initiatives in Mexico and Latin America and codirector of the Institute on Domestic Violence and Sexual Assault. His research focus is on the cultural adaptation of prevention parenting interventions for low-income immigrant Latino/a communities. He is currently funded by the National Institute on Drug Abuse (NIDA) for work with Latino/a immigrant families with adolescent children. An additional area of research is the cultural evaluation of programs for IPV survivors and offenders.
Noël Busch-Armendariz is the Director of the Institute on Domestic Violence and Sexual Assault, and a University Presidential Professor in the School of Social Work at The University of Texas at Austin. She is a nationally recognized expert in sexual assault and human trafficking. Busch-Armendariz has worked as a battered woman’s advocate, support group leader, program director, and registered lobbyist. Busch-Armendariz serves on the program committee of SafePlace, Inc. and the policy committee of the Texas Council on Family Violence.
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.
1. “Transgender” includes seven original response categories: transgender female, transgender male, genderqueer, gender nonconforming, intersex, two spirit, and I prefer to be called, please specify.
2. “Sexual minority” includes six original response categories: gay, lesbian, bisexual, asexual, queer, and a sexual orientation not listed.
3. “Additional races/ethnicities” includes nine original response categories: African American (5%), Asian (15.5%), American Indian/Alaskan Native, Native Hawaiian, Guamanian or Chamorro, Samoan, Other Pacific Islander (2%), biracial and multiracial (6.8%).
4. Participants who experienced faculty/staff-perpetrated sexual harassment were asked to pick the victimization that had the biggest impact on them and to provide additional information about that incident, including the gender identity and status of a member of faculty or staff as well as whether the incident occurred on or off campus. A similar set of questions were asked of victims of peer-perpetrated sexual harassment.