Socioecological Risk and Protective Factors for Sexual Victimization in Post-Secondary Students in Ireland: A Brief Report

Abstract

Although numerous predictors of sexual violence (SV) have been identified, there is a need to further explore protective factors and examine the nature and strength of associations between predictors and SV outcomes using a hierarchical predictive model. Cross-sectional data from the My World Survey Post-Second Level (2019) in Ireland were used. The sample contained 8, 288 post-secondary students, 69% female, aged 18 to 25 years (M = 20.25, SD = 1.85). Univariate and multivariate binary logistic regression analyses were conducted to identify correlates of two SV outcomes (i.e., forced/pressured to have sex and being touched in a sexual manner without consent) across societal, community, relationship, and individual levels. Approximately 25% of females and 10% of males reported being forced/pressured to have sex, whereas 55% of females and 23% of males reported being touched in a sexual manner without consent. Key predictors of SV in multivariate analyses included being female and experiencing violence in a romantic relationship, whereas discrimination, living on campus or in rented accommodation (compared with living at home) were also significant predictors of both SV outcomes. Parental conflict, sexual initiation, depressive symptoms, older age, and an avoidant coping style significantly predicted being forced/pressured to have sex, whereas being bullied and engaging in problem drinking were unique predictors of being touched sexually without consent. Findings highlight the importance of considering salient risk and protective factors at the individual, relationship, community and societal level to predict SV and effectively guide the delivery of interventions and support services for post-secondary students affected by SV.

Introduction

Sexual violence (SV) is defined as “any sexual activity where consent is not obtained or freely given and ‘incorporates physically, verbally and/or psychologically coercive acts such as being touched in a sexual manner and/or being forced/pressured to have sex against one’s will’” (Tharp et al., 2013, p. 133). Despite growing awareness around issues of “consent” (i.e., when an individual freely and voluntarily agrees to engage in a sexual act; Department of Justice, Ireland, 2019) and gender equality among university/college campuses, SV remains a serious issue among post-secondary students (Adams-Curtis & Forbes, 2004; Anyadike-Danes et al., 2022). Studies have shown that one in three women and one in eight men report rape victimization at American universities (Koss et al., 2022). Similarly, in a national survey of 7,900 post-secondary students in Ireland, over one-third of female students (approx., 1,100 students) reported non-consensual vaginal penetration through coercion, incapacitation, force or threat of force, whereas 20% of males reported non-consensual touching and 8% reported being pressured to perform a sexual act (MacNeela et al., 2022).

SV can result in a host of adverse chronic physical, reproductive and psychological consequences for victims, both directly from assaults and indirectly via health problems and/or associated risky health behaviors (Jina & Thomas, 2013; Tharp et al., 2013). It is also associated with considerable psychological distress (Carey et al., 2014), poorer academic performance (Banyard et al., 2020; Jordan et al., 2014), and increased risk of dropout from post-secondary courses (Mengo & Black, 2016). Therefore, it is important that relevant risk factors, (i.e., factors that increase an individual’s vulnerability to experiencing SV), as well as protective factors (i.e., factors that can reduce the risk of experiencing SV) among post-secondary students are identified so as to reduce risk of SV and to inform prevention efforts and service provision (Ullman & Najdowski, 2010).

The socioecological model (Bronfenbrenner, 1979), which considers the complex interplay between individual, relationship, community, and societal factors can provide a holistic understanding of a range of risk/protective factors for SV. At an individual level, demographic risk factors for SV include being female, non-White, non-heterosexual, older in age or college year, and having low socioeconomic status (SES) or disability (Coulter et al., 2017; Eisenberg et al., 2016; MacNeela et al., 2022; Mellins et al., 2017). Alcohol use (binge drinking), drug use, and marijuana use tend to be strongly implicated in the risk for SV (Ellickson et al., 2004; Gidycz et al., 2007; Roudsari et al., 2009). Many studies also reveal associations between SV risk and engagement in risky sexual behaviors such as engaging in casual sex or the “hook up culture” (Bhochhibhoya et al., 2021; Siegel & Williams, 2003). Furthermore, increased time on the internet and engaging in risky online behaviors have been implicated in the risk for SV (Hines et al., 2012). Although less research has focused on psychological factors, the risk of SV is also thought to be elevated among individuals who have mental health difficulties, depression, anxiety, and use maladaptive coping (Aizpurua et al., 2021; Conley et al., 2017).

Peer and family risks include having a parent with a mental health difficulty or addiction, exposure to parental conflict, and/or domestic violence, and having low levels of parental warmth/support and low family support/concern (Bhochhibhoya et al., 2021; Fuller-Thompson & Agbeyaka, 2020; Messman-Moore & Brown, 2004). A previous history of sexual, physical, and/or emotional abuse is a primary predictor of SV in young adult students (Messman-Moore & Brown, 2004; Smith et al., 2003). Relationship conflict, violence, and/or emotional abuse in dating relationships also predicts heightened risk of SV in women (Duval et al., 2020; Testa et al., 2007).

Community factors including level of social disorganization (i.e., level of unemployment or financial inequality; Baron & Straus, 1989), community cohesion, safety and poverty (Yeo et al., 2022), post-secondary institution type (e.g., community colleges with more first-generation, lower income, female students conferring greater risk versus universities; Potter et al., 2020), and accommodation type (lower risk living at home with parents versus off-campus rented accommodation and sorority/fraternity houses; Tyler et al., 2017) are also thought to influence the risk for SV. Experiencing discrimination has been identified as a societal risk factor for SV (Bhochhibhoya et al., 2021; Sabidó et al., 2015).

Although many risk factors have been identified, protective factors, which can reduce an individual’s risk of being sexually assaulted, particularly in the presence of risk, have received less research attention (Ullman & Najdowski, 2010). In terms of individual-level factors, some studies have found that having higher levels of sexual assertiveness (Livingston et al., 2007) and to some extent, engaging in less risky sexual or health behaviors (e.g., lower drug use) can reduce the risk of experiencing some forms of SV (Ullman & Najdowski, 2010). High levels of relationship satisfaction and emotional support from a significant other (i.e., boyfriend/girlfriend/partner) and from parents have been identified as factors that can reduce the risk of SV (Banyard et al., 2009). However, less is known about whether adaptive coping strategies, personal resilience or help seeking intentions can protect against victimization by strengthening social support networks or reducing reliance on less adaptive coping strategies (e.g., alcohol use; Ullman & Najdowski, 2010). Further research on protective factors is required to inform more strengths-informed prevention efforts to reduce risk of SV (Ullman & Najdowski, 2010; Bonar et al., 2022).

Furthermore, few studies have concurrently examined multiple risk and protective factors for SV across several levels of the ecological model (Ullman & Najdowski, 2010). Most of the research has either focused on predictors at an individual level, with less consideration for contextual or ecological factors that contribute to SV risk, or have examined only one level of analysis at a time (Bonar et al., 2022). According to Ullman and Najdowski (2010), multilevel studies of risk and protective factors that account for micro, macro- and meso-level structures are necessary to better inform how individual factors (e.g., SV history) and behaviors (e.g., substance use) translate into risk for SV across various situational and structural contexts. One study by Herres et al. (2021) found that being female, involved in “Greek” societies and prior victimization predicted campus sexual assault in liberal arts students in the United States of America (USA). Bhochhibhoya et al. (2021) also evaluated risk of SV in college students using an ecological model and found that sexual victimization was associated with gender, sexual assertiveness, the frequency of hook-ups, peer deviance, parental involvement, and perceived discrimination. Although these studies did explore a range of risk factors, they did not incorporate many psychological characteristics, protective factors, or other relevant factors like risky online use (Jonsson et al., 2015). Furthermore, some studies have identified factors that combine in different ways to increase risk of SV (e.g., early sexual initiation, multiple sexual partners, and unprotected sex; Armstrong et al., 2016) whereas others have identified how certain protective factors, like empathy, may moderate the relationship between risk factors including peer approval of forced sex (Hudson et al., 2018). A recent systematic meta-review employed an ecological approach to understanding the impact of SV and found that the aftermath of SV involved a complex interplay of risk (e.g., chronosystem- time since the SV, individual level [avoidant coping, self-blame]) and protective factors (e.g., micro/meso system—social network, exosystem supportive reactions from formal support providers) in survivors’ recovery following SV (Stockman et al., 2023).

Additionally, most research on this topic has been conducted in the USA, often within relatively small convenience samples of college students, limiting generalizability to other contexts (Anyadike-Danes et al., 2024). Differences in university cultures and student demographics between the USA and other countries, like Ireland, may lead to varied risk for SV (Reynolds et al., 2023). For example, the American education system is not an open system and fees for attending university in the USA are considerably higher than Ireland, which may influence the profile of students that can attend universities and subsequently the demographic characteristics associated with SV risk in these contexts (Humphreys, 2015).

Additionally, in the USA, a large proportion of students travel far distances to attend university, and often, as a result, live in dormitories, campus accommodation, or shared apartments, whereas in Ireland, many students attend universities located close to where they live, and live in their childhood homes with their parents, which may structurally impact risk for SV across these contexts (Humphreys, 2015). Furthermore, fraternities/sororities are largely non-existent in Ireland, and we may be less likely to observe the “red zone” (the period of time from the beginning of fall semester to Thanksgiving break when sexual assaults on USA college campuses spike) due to the lack of Greek societies (Flack et al., 2008). Alternatively, this risk factor may manifest in the Irish context, as “lad culture” (i.e., excessive displays of macho behavior linked to sexist and homophobic abuse; Phipps & Young, 2013) and/or the large drinking culture among university students (Craig, 2016; Jordan et al., 2022).

Furthermore, according to work by Paul Donovan, students in Ireland are more likely to have lower levels of in-class engagement, school affiliation/connectedness, and less contact with lecturers compared with students the USA (as cited in Humphreys, 2015). This, coupled with Ireland’s history of being a Catholic country which may have stifled correct information on healthy sexual relationships (Sherlock, 2012), may impact protective community factors against SV risk.

In addition, the framework for institutional response to SV varies between the countries (Shannon, 2022a). The USA and Canada has a legalistic national framework for response, which contrasts with other countries, such as Ireland and the UK, where there is an individualized response focused on cultural change (MacNeela et al., 2022; Shannon, 2022b). As an example, in an Irish context, recommendations have been made with regard to implementing a campus climate approach to culture change on consent, SV, and harassment (MacNeela et al., 2022). These include campus wide awareness raising and education initiatives around SV and bystander trainings. Although these initiatives also exist in the USA, universities are neither legally bound to implement them in Ireland nor are they legally obliged to report on incidences of SV. Furthermore, due to “Title IX, the law prohibiting sex discrimination at federally funded institutions,” students in the USA have legal protections (e.g., process rights, rights to live hearings, allowances for students’ advisers to cross-examine parties, and witnesses involved). Although there is a national framework that hosts recommendations to tackle SV and change institutional culture in Ireland, there is no legally binding due process rights for students who report sexual misconduct and for those accused of it (Shannon, 2022b). This limited data on SV in countries outside of the USA, including Ireland and Europe, and reliance on USA data as a base of knowledge, is inappropriate given these differing cultural factors and reporting structures which may potentially influence risk/protective factors across the ecological model (e.g., perceived safety levels in an institution) (Reynolds et al., 2023). Therefore, there is a need for research from countries outside of the USA to allow for greater representation and to identify where potential gaps remain in our understanding, so that we can inform policies and practices in other contexts (e.g., Ireland, UK) who use similar reporting frameworks (Reynolds et al., 2023).

The present study extends the literature by exploring a comprehensive range of risk and protective factors, with a focus on psychological characteristics, in a large sample of post-secondary students in Ireland. Given broad descriptions of acts of SV, this study explored two SV outcomes; being forced/pressured to have sex and being touched against one’s will or without consent. Using an ecological model, this study aimed to identify prevalence of SV in post-secondary students in Ireland and to identify the key risk and protective factors for these SV outcomes in students so as to inform intervention efforts and service provision.

We expect based on predominantly USA-based research that individual-level factors (e.g., female gender, history of violence in a romantic relationship), micro-level factors (e.g., low peer/family support), and macro-level factors (e.g., neighborhood safety) would be associated with a higher likelihood of SV, whereas protective factors at individual levels (adaptive coping) and micro-levels (one supportive adult) might protect against the risk of SV.

Methods

Participants

Data from the My World Survey 2-Post Second Level (MWS2-PSL), a national sample of young adults in Ireland were used. Student data were collected from 12 post-secondary institutions across Ireland, including 5 out of 14 Institutes of Technologies (IoTs) and all seven Universities. Analyses were conducted on 8, 288 students; participants were excluded if they had greater than 70% missing data (n = 190). The sample contained 2, 364 (28.5%) males, and 5,727 (69.1%) females. Less than 2% of the sample identified as a gender minority: n = 77 (0.9%) were identified as non-binary, n = 42 (0.5%) as transgender, n = 78 (0.9%) were not sure/questioning or preferred not state their gender. Participants were aged 18 to 25 years (M = 20.25, SD = 1.85). Most n = 7,401 (91.1%) were undergraduates, and there were n = 723 (8.9%) postgraduates. Participants identified as White Irish (76.6%), White other (11.8%), Asian/Asian Irish (5.7%), mixed background (2.3%), Black/Black Irish (1.4%), other (2.0%), and Irish Traveler (.2%).

Measures

Variables included in analyses are outlined in Figure 1 and described in Table 1.

Figure 1.

Ecological model of predictors of sexual violence.

Table 1.

Constructs, Measures and Coding of Variables in the Predictive Model.

Construct and Measurement	Coding
Sexual violence outcome variables
Sexual violence (being forced/pressured to have sex) Indicated yes/no whether they had ever been pressured/forced to have sex.	“Yes” coded as 1, “No” coded as 0.
Sexual violence (being touched sexually without consent) Indicated yes/no whether they had been touched in a sexual manner against their will or without consent.	“Yes” coded as 1, “No” coded as 0.
Societal influences
Discrimination Indicated “yes/no” whether they had ever been treated unfairly because of their identity (i.e., sexual orientation, ethnicity, race, minority group status).	“Yes” coded as 1, “No” coded as 0.
Community influences
Neighborhood safety Indicated how safe they feel in their neighborhood. Response options: 1 “Very unsafe” to 5 “Very safe.”	Dichotomized responses 1–2 into “unsafe” (coded 1) and 3–5 “safe” (coded 0).
Place of residence Indicated whether they lived 1 “at home,” 2 “in rented accommodation” 3 “on campus” or 4 “other.”	Dichotomized responses 1 into “at home” (coded 0) and 2–4 into “other” (coded 1).
Institution Indicated whether they attended a University or Institute of Technology (IoT).	Dichotomized into University (coded 0) and IoT (coded 1).
Relationship influences
Parent mental health/addiction issues Indicated “yes/no” whether their parent/guardian ever had a long-term mental health problem (e.g., depression) and/or a long-term drug/alcohol addiction problem.	“Yes,” coded as 1, “No” coded as 0.
Parental conflict Indicated “yes/no” whether they experienced conflict between parents.	“Yes” coded as 1, “No” coded as 0.
Domestic violence Indicated “yes/no” whether they experienced violence at home (domestic violence).	“Yes” coded as 1, “No” coded as 0.
Social support (family, significant other, peer) Multidimensional Scale of Perceived Social Support (Zimet et al., 1988). Three four-item subscales “Family support” α = .92, “Significant other support” α = .96, and “Peer support” α = .94, subscales. Mean subscale score range (1–7), higher scores indicate higher levels of support.	Dichotomized scores one standard deviation (1 SD) below the mean as “low” (coded 1) and scores above this as “high” (coded 0) for each subscale.
Relationship status Indicated whether they were 1 “single,” 2 “in a relationship” 3 “married” 4 “cohabiting” or 5 “other.”	Dichotomized into 1 “single” (coded 1) and 2–5 “in a relationship” (coded 0).
Relationship violence Indicated “yes/no” whether they experienced violence in a romantic relationship.	“Yes” coded as 1, “No” coded as 0.
Distressing breakup If participants indicated “yes” to experiencing a breakup they were asked to indicate level of distress caused by breakup. Response options ranged from 1 “Very distressing” to 4 “Not at all distressing.”	Dichotomized responses 1–2 into “distressing” (coded 1) and 3–4 as “not distressing” (coded 0).
Emotional involvement in past relationship Indicated how emotionally involved they were with person at time of breakup, with response options ranging from 1 “Very involved” and 4 “Not at all involved.”	Dichotomized responses 1–2 into “high emotional involvement” (coded 1) and 3–4 into “low emotional involvement” (coded 0).
Bullying Indicated “yes/no” whether they had ever been bullied.	“Yes” coded as 1, “No” coded as 0.
Individual influences
Risky Behaviors Problem alcohol use The Alcohol Use Disorders Identification Test (WHO; Saunders et al., 1993) assesses frequency of alcohol consumption and engagement in risky alcohol use. Ten-items Cronbach’s α = .82. Recommended cut-offs: 1 “abstinent/low risk,” 2 “harmful” 3 “hazardous drinking,” and 4 “possible alcohol dependence.”	Dichotomized cut-offs scores of 1 into “no problem drinking” (coded 0) and 2–4 into “problem drinking” (coded 1).
Problem drug use The Drug Abuse Screen Test (Skinner, 1982) assesses drug use (excluding alcohol and tobacco) in the last 12 months. 10 items, total score range (0–10). Cronbach’s α = .74. Recommended cut offs: 1 “no problem” 2 “low” 3 “moderate” 4 “substantial,” and 5 “severe.”	Dichotomized cut-offs scores of 1 into “no drug problem” (coded 0) and 2–5 into “drug problem” (coded 1).
Cannabis use Indicated “yes/no” whether they had ever used cannabis	“Yes” coded as 1, “No” coded as 0.
Sexual initiation Indicated “yes/no” whether they had had sex. If “yes” participants indicated age at which they first had sex, with response options of “under 14 years,” “15–16 years,” and “17+ years.”	Dichotomized 1 into “early initiation” (coded 1) and 2–3 into “later initiation” (coded 0).
Sexual partners Indicated number of sexual partners in last 3 months.	6+ sexual partners classified as a “high” number of sexual partners (coded 1), 5 or fewer classified as “low” (coded 0).
Safe sex Indicated frequency with which they or their partner used condoms or another form of contraception when having sex, with response options of 1 “never” 2 “most times/sometimes,” and 3 “always.”	Dichotomized responses of 1 into “unsafe” (coded 1) and 2–3 into “safe” (coded 0).
Porn use Indicated how often they watched sexual content on internet over past month. Response options: 1 “Never” 2 “Less than once” 3 “2–3 times” 4 “once a week,” and 5 “more than once a week.”	Dichotomized responses 1–2 into “low frequency” (coded 0), 3–5 as “high frequency” (coded 1).
Meeting strangers online Indicated how often they met people online for social interaction that they hadn’t met in real life with response options of 0 “never” and 1 “at least once.”	Dichotomized responses of 0 into “safe online” (coded 0) and 1 into “risky online” (coded 1).
Personal characteristics Resilience Brief Resilience Scale (Smith et al., 2008) measures resilience. 6 items, mean score range (1–5). Cronbach’s αa = .88.	Scores 1 SD below the mean dichotomized into “low” (coded 1) and scores above this into “high” (coded 0).
Coping (avoidant and problem focused) The Adapted Coping Strategy Indicator (Amirkhan, 1990) assesses dimensions of coping strategies. Five-item problem-focused α = .86 and six-item avoidant coping subscales α = .82, were used. Mean score range (1–6).	Scores 1 SD below the mean were dichotomized into “low” and scores above this as “high.” For Problem-focused coping low (coded as 1), high (coded as 0). For Avoidant coping low (coded as 0) and high (coded as 1).
Comfort with sexuality Indicated on scale of 1–10 how comfortable they were with their sexuality, with higher scores indicating greater comfort.	Scores 1 SD below the mean dichotomized into “low” (coded 1) and the remainder “high” comfort with sexuality (coded 0).
Body esteem Body Esteem Scale for Adolescents and Adults (Mendelson et al., 2001), 10-item appearance subscale assesses self-evaluation/satisfaction with bodily appearance. Mean score range (0–4). Cronbach’s α = .85.	Scores 1 SD below the mean dichotomized into “low” (coded 1) and scores above this into “high” (coded 0).
Depression and anxiety The Depression and Anxiety Scale (Lovibond & Lovibond, 1995) assesses severity of negative emotions experienced by participants in past week. Depression α = .91, and anxiety α = .84, subscales were used; 21 items, mean item score range 0–4, recommended cut-offs: “normal,” “mild,” “moderate,” “severe,” and “very severe.”	Dichotomized cut-offs scores of 1-2 into “low” (coded 0) and 3–5 into “high” (coded 1).
Self-reported mental health condition Indicated yes/no whether they had long-term mental health condition.	“Yes” coded as 1, “No” coded as 0.
Self-reported physical health condition Indicated yes/no whether they had a long-term physical health condition/disability.	“Yes” coded as 1, “No” coded as 0.
Help-seeking intentions Indicated if experienced any serious emotional/personal/behavioral problems that caused considerable stress and would have benefited from professional help in previous year. Response options: (a) I have had few/no problems; (b) some problems, but did not feel professional help needed; (c) some problems but did not seek professional help when needed; And (d) some problems but sought professional help when needed.	Dichotomized responses 1 and 4 into “help not needed/got help” (coded as 0) and responses 2 and 3 into “Not get professional help when needed” (coded as 1).
Demographic characteristics Gender Indicted gender they identified with. Response options: Male, female [participants who identified as “other e.g., non-binary,” and “prefer not to say,” were deleted from sample due to low prevalence].	Dichotomized into “Female” (coded 1), “Male” (coded 0).
Ethnicity Indicated ethnicity they identified with. Response options: (a) White Irish, (b) White Other, (c) Black/Black Irish, (d) Asian/Asian Irish, (e) Mixed background, (f) Irish Traveler, and (g) Other.	Dichotomized response 1 into “White Irish” (coded 0) and responses 2–7 into “Other” (coded 1).
Sexual orientation Indicated sexual orientation they identified with. Response options: (a) Heterosexual, (b) Lesbian, (c) Gay, (d) Bisexual, (e) Pansexual, (f) Questioning, (g) Other, and (h) Prefer not say.	Dichotomized 1 into “Heterosexual” (coded 0), and 2–7 into “LGBQ+” (coded 1).
Socioeconomic status (SES) Mother’s highest level of education achieved served as proxy for SES. Participants who indicated that mother’s highest level of education attained was the Junior Certificate (i.e., the first stage of the education program for post-primary education within the Republic of Ireland) were coded as having “low SES.”	Dichotomized into “Low SES” (coded 1) and “High SES” (coded 0).
Student type Indicated whether they were an undergraduate or postgraduate student.	Dichotomized into “Undergraduate” (coded 1), and “Postgraduate” (coded 0).
Age Indicated their age.	Dichotomized ages 18–20 categorized into “younger” (coded 0) and ages 21–25 “older” (coded 1).

Procedure

Ethical approval was received from the host institution. Permission to host the study was sought from Registrars (or equivalent) of all post-secondary institutions. All students registered at participating institutions received an email inviting them to partake in a survey about mental health and well-being. Emails contained a weblink to an information sheet, consent form, and survey. Web-based surveys were administered using Qualtrics software and took approximately 25 min to complete. Participants provided informed consent prior to participation and were fully debriefed. Participation was voluntary and financial compensation was not provided.

Statistical Analyses

We calculated the prevalence of both SV outcomes across gender, sexual orientation, and student degree type. Chi-square tests of independence explored potential associations between gender, sexual orientation, and degree type with SV outcomes. Chi values of p < .01 and standardized residuals ±2 were reported as significant (Agresti, 2007). When a standardized residual for a category is greater than ±2, it can be concluded that this is a major contributor to the significant χ² value (Hinkle et al., 2003).

Hierarchical logistic regression analyses were conducted to examine associations between socioecological predictors and SV outcomes. Univariate logistic regressions (with unadjusted odds ratios) examined associations between SV outcomes and each predictor in the model. Multivariate logistic regressions, which included all predictor variables in the model, were also conducted to examine the relationships between risk and protective socioecological predictors and SV outcomes, while controlling for potential confounding variables. Separate models were run for SV outcomes of (a) being touched inappropriately without consent and (b) being forced/pressured to have sex without consent. Analyses were conducted using IMB SPSS Version 21.

Results

As seen in Table 2, approximately one-fifth of the sample reported that they had been forced or pressured to have sex, whereas a little under half of the sample reported that they had been touched in a sexual manner without consent. Females, postgraduates, bisexual, and pansexual individuals reported elevated rates of SV compared with males, undergraduates, and other sexual orientations. Table 3 presents the frequencies of predictor variables.

Table 2.

Prevalence and Chi-square Comparison of Sexual Violence by Gender, Sexual Orientation, and Student Degree Type.

Participant characteristics	Forced/Pressured to have Sex N (%)	χ²	Touched in A Sexual Manner Without Consent N (%)	χ²
Total % of sample	1,326 (20.3%)		3,039 (46.5%)
Gender		542.03*** Females > likely (9.1), Males < likely (−14.5)		153.47*** Female > likely (5.9) Male < likely (−9.4)
Male	182 (10.0%)		418 (23.1%)
Female	1,084 (23.8%)		2,522 (55.3%)
Sexual orientation		105.43*** Heterosexuals < likely (−3.4) Bisexuals (7.5), Pansexuals (5.0) >likely.		124.74*** Heterosexuals < likely (−2.9), Bisexual (6.2), Pansexuals (2.5) > likely
Heterosexual	847 (18.1%)		2,034 (43.4%)
Lesbian	32 (21.6%)		71 (48.3%)
Gay	49 (22.5%)		85 (39.0%)
Bisexual	253 (32.4%)		479 (61.4%)
Asexual	8 (11.8%)		33 (44.0%)
Pansexual	54 (39.4%)		83 (60.6%)
Questioning	58 (15.9%)		183 (49.3%)
Degree type		11.33*** PG (2.9) > likely		1.25
Undergraduate	1,148 (19.6%)		2,714 (46.4%)
Postgraduate	165 (26.7%)		306 (49.5%)

Note. Chi-square tests of independence explored potential associations between gender, sexual orientation, degree type, and SV outcomes. Chi-values of p < .01 and standardized residuals ±2 were reported as significant (Agresti, 2007). Only significant residuals (±2) are reported in the table, with > signifying that individuals were more likely and < were less likely to report the outcome.

^***p < .001.

Table 3.

Frequency Analysis for Socioecological Predictor Variables.

Societal	%	Individual	%
Discrimination	24.1	Risky behaviors
		Problem alcohol use	55.3
Community		Problem drug use	13.2
Neighborhood safety (unsafe)	5.1	Cannabis use	51.4
Where live (on campus)	49.7	Early sexual initiation	3
Institution (IoT)	15.5	Never contraceptive use	10.8
		Multiple sexual partners in 3 months	1.2
Relationship		Frequent porn use	65.9
Parental conflict	41	Meeting strangers online	30.2
Violence in the home	9.2	Psychological factors
Parent/guardian with mental health difficulty/addition	32.2	High anxiety	49
Family social support (low)	13.6	High depression	44.1
Peer social support (low)	11.8	Self-reported long-term mental health condition	24.1
Adult social support (low)	14.3	High avoidance coping	17.9
Relationship status (single)	64.5	Low problem focused coping	13.2
Violence in a romantic relationship	5.6	Low comfort with sexuality	12.6
Experienced distressing breakup	86.4	Low help-seeking intentions	25.2
High emotional involvement in previous relationship	75.8	Low resilience	18
Bullied	58	Low body esteem	18.4
		Demographic characteristics
		Gender (female)	70.8
		Ethnicity (Other than White)	23.4
		Sexual orientation (LGBQ+)	26
		SES (low)	8.7
		Age (old)	39.2
		Student type (undergrad)	91.1
		Self-reported physical health difficulty	6.9

Note. IoT = Institute of Technology; LGBQ+ = lesbian, gay, bisexual, questioning, and other; Undergrad = undergraduate; SES = socioeconomic status.

As seen in Table 4, unadjusted odds ratios show that most societal, community, relationship, and individual variables were significant predictors of SV. However, institution type, SES, comfort with sexuality, help seeking intentions, peer support, and having multiple sexual partners did not predict either SV outcome. Never using contraception and family support were not significant predictors of being forced/pressured to have sex, but they were significantly associated with being touched in a sexual manner without consent. Degree type did not significantly predict being touched in a sexual manner without consent.

Table 4.

Societal, Community, Relationship and Individual Risk and Protective Factors for Sexual Violence.

Potential risk/Protective factors	Being Forced/Pressured to have Sex Without Consent				Being Touched in A Sexual Manner Against will/without Consent
	Unadjusted Odds Ratio [95% CI]		Adjusted Odds Ratio [95% CI]		Unadjusted Odds Ratio [95% CI]		Adjusted Odds Ratio [95% CI]
Societal
Discrimination (yes)	1.63***	[1.43, 1.86]	0.98	[0.69, 1.41]	1.88***	[1.68, 2.11]	2.01***	[1.38, 2.93]
Community
Neighborhood safety (unsafe)	1.57***	[1.22, 2.01]	1.04	[0.55, 1.97]	1.29*	[1.03, 1.61]	0.93	[0.47, 1.82]
Where live (on campus)	1.32***	[1.17, 1.49]	1.55**	[1.14, 2.10]	1.14**	[1.03, 1.26]	1.38*	[1.02, 1.87]
Institution (university)	1.02	[0.86, 1.21]	0.96	[0.64, 1.44]	0.88	[0.76, 1.01]	0.65*	[0.43, 0.97]
Relationship
Parental conflict	1.71***	[1.52, 1.93]	1.61**	[1.18, 2.21]	1.49***	[1.35, 0.165]	1.10	[0.80, 1.51]
Violence in the home	2.04***	[1.71, 2.44]	0.83	[0.52, 1.31]	1.90***	[1.61, 2.25]	0.90	[0.55, 1.48]
Parent/guardian with mental health difficulty/addition	1.92***	[1.70, 2.17]	1.23	[0.90, 1.67]	1.79***	[1.62, 1.99]	1.18	[0.86, 1.63]
Family social support (high)	1.69***	[1.45, 1.96]	1.01	[0.67, 1.52]	1.29***	[1.12, 1.49]	1.16	[0.73, 1.86]
Peer social support (high)	1.13	[0.94, 1.36]	1.24	[0.73, 2.10]	0.92	[0.79, 1.07]	1.00	[0.57, 1.74]
Adult social support (high)	0.9	[0.76, 1.08]	1.14	[0.65, 1.98]	.81**	[0.70, .93]	0.95	[0.54, 1.66]
Relationship status (not in a relationship)	.56***	[0.50, 0.63]	1.29	[0.94, 1.77]	.67***	[0.61, 0.75]	0.74	[0.54, 1.02]
Relationship conflict/violence	8.75***	[7.00, 10.94]	4.98***	[3.08, 8.08]	4.89***	[3.77, 6.33]	3.01***	[1.72, 5.28]
Experienced distressing breakup	2.00***	[1.53, 2.61]	1.16	[0.62, 2.18]	1.38**	[1.11, 1.72]	1.69	[0.92, 3.09]
High emotional involvement in previous relationship	1.59***	[1.22, 1.89]	1.23	[0.70, 2.14]	1.26**	[1.08, 1.46]	0.75	[0.44, 1.28]
Bullied	1.93***	[1.69, 2.19]	1.18	[0.86, 1.63]	1.87***	[1.69, 2.06]	1.78***	[1.30, 2.44]
Individual
Risky behaviors
Problem alcohol use	1.77***	[1.56, 2.02]	1.06	[0.75, 1.49]	1.57***	[1.42, 1.75]	1.58**	[1.13, 2.22]
Problem drug use	2.35***	[2.01, 2.75]	1.30	[0.90, 1.88]	1.57***	[1.35, 1.81]	1.11	[0.77, 1.61]
Cannabis use	2.71***	[2.38, 3.08]	1.12	[0.77, 1.63]	2.04***	[1.85, 2.25]	1.39	[0.97, 1.99]
Early sexual initiation	2.83***	[1.98, 4.04]	2.19*	[1.08, 4.44]	1.86**	[1.27, 2.73]	1.43	[0.68, 3.01]
Never contraceptive use	1.56	[0.81, 3.00]	1.32	[0.38, 4.57]	3.06**	[1.40, 6.69]	2.50	[0.60, 10.48]
Multiple sexual partners in 3 months	0.837	[0.66, 1.06]	0.79	[0.51, 1.25]	0.91	[.74, 1.14]	0.69	[0.43, 1.12]
Frequent porn use (infrequent)	.60***	[0.52, 0.70]	1.01	[0.74, 1.39]	.53***	[0.47, 60]	1.09	[0.79, 1.51]
Meeting strangers online	1.61***	[1.42, 1.83]	1.21	[0.89, 1.66]	1.61***	[1.42, 1.83]	1.10	[0.81, 1.52]
Psychological factors
High anxiety	2.22***	[1.96, 2.52]	1.21	[0.86, 1.69]	1.82***	[1.66, 2.02]	1.19	[0.85, 1.65]
High depression	2.08***	[1.84, 2.35]	1.52**	[1.08, 2.14]	1.63***	[1.47, 1.80]	1.35	[0.96, 1.90]
Self-reported mental health condition	2.27***	[2.00, 2.58]	1.18	[0.84, 1.65]	1.76***	[1.57, 1.97]	1.28	[0.91, 1.81]
High avoidance coping	.58***	[0.48, 0.69]	1.38*	[0.90, 2.11]	.65***	[0.57, 0.74]	1.19	[0.79, 1.78]
Low problem focused coping	1.48***	[1.25, 1.75]	1.12	[0.74, 1.70]	1.26**	[1.09, 1.46]	0.83	[0.54, 1.27]
Low comfort with sexuality	1.08	[0.90, 1.30]	1.37	[0.81, 2.33]	1.12	[0.96, 1.30]	1.41	[0.77, 2.58]
Low help seeking intentions	1.08	[0.96, 1.23]	1.15	[0.82, 1.61]	1.02	[0.92, 1.12]	0.79	[0.57, 1.11]
Low resilience	1.75***	[1.51, 2.02]	1.03	[0.71, 1.49]	1.64***	[1.44, 1.86]	1.21	[0.82, 1.80]
Low body esteem	1.83***	[1.59, 2.11]	1.48*	[1.00, 2.17]	1.61***	[1.42, 1.83]	1.17	[0.77, 1.80]
Demographic characteristics
Gender (female)	2.79***	[2.36, 3.30]	4.07***	[2.76, 6.01]	4.13***	[3.65, 4.67]	6.09***	[4.26, 8.72]
Ethnicity (not White Irish)	.85*	[0.74, 0.99]	0.98	[0.67, 1.43]	.86*	[0.77, 0.97]	0.80	[0.54, 1.19]
Sexual orientation (heterosexual)	1.08***	[0.90, 1.30]	1.39	[0.97, 1.98]	1.56***	[1.4, 1.74]	1.35	[0.93, 1.96]
SES (low)	0.92	[0.73, 1.14]	0.75	[0.44, 1.27]	1.01	[0.84, 1.20]	0.83	[0.49, 1.43]
Age (old)	1.58***	[1.40, 1.78]	1.51*	[1.10, 2.07]	1.22***	[1.10, 1.34]	1.14	[0.83, 1.56]
Student type (undergraduate)	.71**	[0.58, 0.87]	0.82	[0.51, 1.33]	0.91	[0.76, 1.08]	0.80	[0.49, 1.32]
Experiencing a physical health condition	1.58***	[1.28, 1.95]	0.83	[0.48, 1.43]	1.77***	[1.46, 2.14]	1.63	[0.90, 2.94]
			Nagelkerke R2 = .308				Nagelkerke R2 = .332

Note. SES = Socioeconomic status.

The full regression model explained approximately 31% and 33% of the variance in SV outcomes. Adjusted odds ratios indicate that living on campus, experiencing parental conflict, violence in a romantic relationship, early sexual initiation, high levels of depressive symptoms, having high avoidant coping, being older in age, and being female were important significant predictors of being forced/pressured to have sex. Adjusted odds ratios also show that significant predictors of being touched in a sexual manner without consent include experiencing discrimination, living on campus/in rented accommodation, attending a university, being bullied, experiencing violence in a romantic relationship, engaging in problematic alcohol use, and being female. Adjusted odds ratios showed that risky online behavior did not remain significant in the model. Being of female gender and experiencing violence in a romantic relationship were the primary predictors of SV. Females were approximately four times more likely to be forced/pressured to have sex and six times more likely to be touched in a sexual manner without consent than males. Individuals who had experienced violence in a romantic relationship were five times more likely to report being forced/pressured to have sex and three times more likely to be touched in a sexual manner without consent.

Discussion

This study aimed to identify prevalence of SV and to examine risk and protective factors simultaneously using an ecological model in a national sample of post-secondary students in Ireland. In relation to the first aim, we found that approximately one-quarter of females and 10% of males reported being forced/pressured to have sex, whereas 55% of females and 23% of males reported being touched in a sexual manner without consent. These rates are similar to international data where one in three women and one in eight men at American universities report rape victimization (Koss et al., 2022) and recent national data in Ireland (MacNeela et al., 2022), where one-third of female students experience non-consensual vaginal penetration, 20% of men reporting non-consensual touching in a sexual manner, and 8% of men reporting being forced/pressured to have sex. This shows that despite recent efforts to raise awareness of consent within post-secondary institutions, SV is still commonly experienced by post-secondary students in Ireland.

Consistent with the literature in the USA, females and sexual minorities (MacNeela et al. 2022; Coulter et al., 2017; Mellins et al., 2017) and bisexual and pansexual individuals were at increased risk of SV (Eisenberg et al., 2016). Pansexuals, who hold attractions to individuals regardless of gender or sexual orientation, have not been extensively studied in relation to SV, therefore this finding, which suggests that pansexual individuals may represent a particularly “at risk” group for SV may be important for guiding and prioritizing supports and service provision toward these individuals (Eisenberg et al., 2016).

Univariate analyses showed that most of the theorized societal, community, relationship, and individual variables were significant predictors of SV, highlighting the importance of applying an ecological model to explore the confluence of risk and protective factors associated with SV risk (Tharp et al., 2013; Ullman & Najdowski, 2010). Consistent with a large body of evidence from literature in the USA, multivariate analyses revealed that being female in gender and experiencing violence in a romantic relationship were the primary predictors of both SV outcomes (Coulter et al., 2017; Duval et al., 2020; MacNeela et al., 2022; Mellins et al., 2017). Females were approximately four times more likely to be forced/pressured to have sex and six times more likely to be touched in a sexual manner without consent than males. Individuals who had experienced violence in a romantic relationship were five times more likely to report being forced/pressured to have sex and three times more likely to be touched in a sexual manner without consent. Interestingly although elevated levels of SV outcomes were reported by sexual minority groups, having a sexual minority identity status did not predict SV outcomes in multivariate analyses. Further contrasting with data from the USA (e.g., Coulter et al., 2017), ethnic minority status (non-white) neither did predict SV outcomes in multivariate analyses, nor did presence of a mental/physical health condition or the proxy we used to assess SES (i.e., mother’s highest level of education) indicating that demographic risk factors for SV may vary across contexts.

Consistent with the USA literature, developmental factors (experiencing parental conflict, early sexual initiation, older age; Bhochhibhoya et al., 2021), psychological factors (high levels of depressive symptoms and avoidant coping; Aizpurua et al., 2021; Conley et al., 2017), and health-risk behaviors (alcohol use; Gidycz et al., 2007; Roudsari et al., 2009) remained significant in multivariate models predicting SV outcomes and represent important risk factors for SV. However, protective factors, including relationship factors (e.g., peer support, relationship satisfaction; Banyard et al., 2009) and psychological factors (e.g., help seeking, comfort with sexuality), were not found to be significantly associated with SV outcomes, indicating a need to further explore protective factors, such as sexual assertiveness (recognizing warning signs of inappropriate sexual advances and having empowerment and skills to say no; Livingston et al., 2007), resistance/self-defense (using forceful resistant strategies as opposed to non-forceful strategies like pleading/begging; Gidycz et al., 2008; Ullman & Najdowski, 2010), school-connectedness, and anti-rape societal and peer attitudes and norms that could mitigate the risk of SV (Ullman & Najdowski, 2010).

Of note, living on campus/in rented accommodation versus living at home was a significant predictor of both sexual outcomes in multivariate analyses (Tyler et al., 2017), Although many students in Ireland travel out of county (state) to attend university and subsequently live on campus or in rental properties, as is common practice in the USA, many other students in Ireland live in their childhood homes with their parents and commute to college. Although MacNeela et al. (2022) found that a majority of students in Ireland felt safe in their accommodation on campus, they also observed that most students, particularly females and sexual minorities, did not feel safe from SV and harassment when socializing on or around campus or at night. These findings indicate the potential need for the development of policies and/or initiatives within post-secondary institutions to address threats of violence and harassment to students when socializing on campus.

Aligning with Bhochhibhoya et al. (2021), discrimination and bullying were significant predictors of SV outcomes. Indeed, high levels of harassment based on sex and sexuality have been reported in post-secondary campuses in Ireland (MacNeela et al., 2022), indicating that greater education and awareness around discrimination linked with sex and sexuality may be required across post-secondary institutions to reduce the risk of SV (Anyadike-Danes et al., 2022). Further research should also investigate how peer attitudes and influences related to “lad culture” potentially feed into narratives of discrimination and bullying in the Irish context, which may inform intervention efforts (Craig, 2016).

Study findings demonstrate how SV risk is influenced by factors operating across several levels of the ecological model (Bronfenbrenner, 1979) and highlight the need for prevention approaches to move beyond focus on individuals affected by SV, and also address the broader systems that may facilitate and maintain SV (Swift & Ryan-Finn, 1995). Indeed, reflecting international trends (Mujal et al., 2021; Muehlenhard et al., 2016), sexual consent and bystander training workshops have been recently introduced as part of college inductions for post-secondary students in several countries including Ireland, but given that empirical research on such programs is largely limited to the USA, these need to be systematically evaluated to ascertain whether they are successful in changing attitudes/perceptions toward SV (Muehlenhard et al., 2016). Additionally, supports and training around consent and SV need to be extended to families, carers, and broader peer and community networks so that there is a basic awareness and understanding of these issues within the wider community (Anyadike-Danes et al., 2022). Furthermore, there is a need to understand how systemic differences in reporting structures (i.e., the legalistic framework in the USA compared with individualistic framework in Ireland and Europe) influence help seeking for, reporting of, and perpetration of SV, and whether these differing frameworks and proxies, such as perceived safety levels in an institution, influence risk, and protective factors for SV across the ecological model.

Limitations

Although this study explored a wide range of risk and protective factors for SV using standardized variables, some relevant variables, such as sexual assertiveness, were not included. Additionally, the cross-sectional nature of the study makes it difficult to distinguish factors that increase the likelihood of experiencing SV versus the result of experiencing SV. Additionally, SV outcomes were measured via two single items, where experiences of pressure and force were combined in one item, and “sex” was not defined; therefore, there may be wide variability in the type of sexual activities participants were reporting on and there may be overlap in responses to the SV items, with participants potentially responding to both items while thinking about the same experience. In addition, independent variables were dichotomized to facilitate analysis and interpretation of this complex logistic model involving 33 variables; however, a level of nuance may have been lost in reducing multicategory variables to binaries. Compared with literature on SV, which typically relies on small convenience samples, this study contained a relatively large sample of post-secondary students; however, the data contained an overrepresentation of females and was a convenience sample of students who self-selected to participate, which may have introduced selection bias. Furthermore, these data reflect a response rate of 11% for University students and 2% for IoTs; therefore, findings might not be generalizable to the entire Irish or international student population. Due to the small sample size, we were precluded from conducting powered analyses on “non-binary” and “gender non-conforming” individuals and from exploring the intersectionality of identities; however, we recognize that gender and sexual orientations are often considered “fluid” constructs (Smalley et al., 2016), it would be beneficial to explore broader constructs of gender and their interactions with sexual orientation in future research. Furthermore, without data on perpetrators, efforts to prevent SV through identification of risk and protective factors among victims of SV may be limited in identifying root causes that inform primary prevention programs (Swift & Ryan-Finn, 1995; Ullman & Najdowski, 2010). MacNeela et al. (2022) observed that male post-secondary students tended to hold more lenient SV attitudes and demonstrated lower levels of bystander interference compared with women, suggesting that there may be a need for greater education and initiatives to highlight issues around consent among students in Ireland.

Future Directions

Further research is required to identify protective factors that could mitigate the risk of SV to inform strengths-focused approaches to reducing the risk of SV. Additionally future studies are required to understand the SV experiences of sexual minorities, including pansexual college students; qualitative studies could provide rich insights into sexual minority students’ lived experience which could inform prevention efforts. Additionally, more longitudinal research is needed to identify temporal relationships between a range of predictors and SV outcomes and how these operate independently or interact with one another to influence SV risk. Cross-cultural comparisons could inform potential gaps in our understanding of how risk and protective factors potentially vary across cultural contexts.

Conclusion

This study extends the literature as it simultaneously explored multiple levels of risk and protective factors for SV using a comprehensive ecological framework in a large sample of post-secondary students. Findings can inform multilevel intervention efforts to address risk factors for SV and the provision of support services to support post-secondary students affected by SV.

Author Biographies

Ciara Mahon, PhD, is a postdoctoral researcher in the Youth Mental Health (YMH) lab at University College Dublin (UCD). She analyzes data and communicates findings from My World Surveys 1 and 2 (MWS 1 & 2), the largest national surveys of youth mental health involving 30,000+ youths, across a range of topics including student mental health, sexuality, and sexual violence.

Barbara Dooley, PhD, is a Professor of Psychology, Dean of Graduate Studies, and Deputy Registrar at UCD. She is co-founder of the YMH lab and the primary investigator on MWS 1 & 2. She is a leading authority in youth mental health and has published extensively on risk and protective factors in youth mental health and supporting third-level student mental health.

Aileen O’Reilly, PhD, is a Research and Evaluation Manager at Jigsaw, The National Centre for Youth Mental Health and an Adjunct Assistant Professor in the School of Psychology, UCD. She is a lead researcher on the MWS-2, which was conducted in partnership with Jigsaw, and her research/evaluation work has played a critical role in the development of youth mental health services globally.

Amanda Fitzgerald, PhD, is an Associate Professor of Psychology and co-founder of the YMH lab at UCD. She is a principal researcher on the MWS-2 and leading expert in youth mental health. Her research focuses on understanding socioecological risk and protective factors of young people’s mental health and psychosocial interventions for youth mental health.