Exploring the Validity of a Modified Version of the SES-SFV with Students Attending Northern Irish Universities

ABSTRACT

Compared to US university students, far less is known about the scale of unwanted and non-consensual sexual experiences [USEs] faced by UK university students, particularly those in Northern Ireland [NI]. The Sexual Experiences Survey (Short Form [SEF-SFV]) is considered a popular tool for measuring USEs but has not been updated since 2007; there is some indication that additional perpetrator tactics may be more inclusive of students’ experiences and that certain scoring strategies may impact our understanding of data. This paper examines the USEs reported by 1033 students attending either of NI’s traditional universities. Participants completed a modified version of the SES-SFV that included two additional perpetration tactics: “ignorance of refusal” and “taken by surprise.” Sixty-three percent (n = 650) reported experiencing at least one USE, but this reduced to 53% (n = 546) without the new perpetrator tactics. Female and non-heterosexual students reported significantly more USEs than male and heterosexual students, respectively. “Taken by surprise” was highly endorsed (81%, n = 525) and the most commonly endorsed tactic. Whilst dichotomous scoring is the most straightforward, continuous scoring affords greater analytical opportunities whilst still retaining frequency of USEs. “Taken by surprise” may be a relevant addition but further mixed-methodological research is required to assess its validity among larger and more diverse samples. SES-SFV scoring options should be also validated using male and mixed-gender samples, particularly categorical scoring to ensure current construction is reflective of the wider student experience.

In the last 30 years, rape victimization at American universities has risen significantly (Koss et al., 2022): the 1 in 4 women and 1 in 19 men reporting rape is now 1 in 3 women and 1 in 8 men. Further evidenced in Fedina et al.’s (2016) review, between 2000 and 2015, an estimated 18–19% of female, and 1–3% of male students in the US were subjected to an unwanted and non-consensual sexual experience [USE] whilst at university (Fedina et al., 2016). The scale and impact of USEs at American universities has resulted in sexual victimization (or, USEs) being considered a major public health concern (Basile et al., 2011; Koss et al., 2022). It is widely believed that students attending UK campuses are subjected to USEs on a similar scale, but currently there is a dearth of peer-reviewed research publicly available (Dworkin et al., 2021; Holloway & Bennett, 2018), with an overreliance on bespoke reporting within specific localities. For example, a recent report provided by Revolt Sexual Assault (2018) indicated that 62% of UK students surveyed (n = 4500) reported some form of sexual violence at university; however, lack of demographic information makes it difficult to determine how representative these figures are across all four UK nations (i.e., England, Northern Ireland [NI], Scotland and Wales) and it was unclear how the different types of sexual violence (e.g., sexual harassment, sexual assault) were operationalized. The most recent national research from the UK was published in 2011 by the National Union of Students; their results suggested that more than half of female students had experienced sexual harassment and approximately 14% had experienced serious physical or sexual assault. There is no comparative male or non-binary student data available. A taskforce developed by Universities UK in 2016 reconfirmed that there was a lack of comprehensive data available in a UK context. Lack of generalizability across studies examining the prevalence of USEs in the university context has been consistently highlighted, particularly with respect to how USEs are defined and measured (Cook et al., 2011; Krause et al., 2019; Krebs et al., 2016; Moylan et al., 2021).

Defining USEs in the Context of Higher Education

The scope of behaviors defined under the umbrella term of USE differ significantly depending on the focus of the research (Krebs et al., 2021; Rennison & Addington, 2014); for example, researchers seeking to identify the prevalence of sexual offenses may operationalize legal definitions (Koss, 1996) whilst those seeking to explore students’ USEs that do not meet the standard of criminal behavior may include broader definitions (Krebs et al., 2021). These choices have a knock-on effect upon prevalence estimates and comparability across studies (Krause et al., 2019; Krebs et al., 2021), particularly between studies conducted in different countries where legal definitions may differ. Higher education institutions [HEIs] may choose to source their definition of USEs from the law or elsewhere (Cowan & Munro, 2021; Krause et al., 2019) so students’ conceptualization (and operationalization) of USEs may vary depending on, for example, their awareness of university policy and the law. In this paper, USE is used as a collective term (similar to sexual violence) to group together a range of unwanted and non-consensual sexual behaviors, including unwanted and non-consensual non-penetrative, attempted penetrative, and completed penetrative sexual acts. The avoidance of terms like “sexual assault” and “rape” and the adoption of behaviorally specific language (e.g., non-penetration, penetration) is considered best practice in surveying USEs because it reduces the risk of participants operationalizing their own definitions (Johnson et al., 2017; Krause et al., 2019). The underreporting of sexual victimization is well-recognized and one such issue relates to individuals and their lack of acknowledgment of their experiences as sexual victimization, even when these experiences clearly meet the definition of a sexual crime (Haugen et al., 2018; Lamarche & James-Hawkins, 2022; Peterson & Muehlenhard, 2004). Further, it allows for greater ease in comparability between studies and countries because the language is not limited to a specific university’s (or, country’s) definition.

Alongside the use of behaviorally specific language, researchers should be mindful of how they direct participants to report on non-consensual acts. For example, research suggests that university students may conflate consent and wantedness, such that the determination of whether an act is non-consensual (even acts they have been subjected to) depends on the absence of indicators of sexual desire (Artime & Peterson, 2015; Hills et al., 2020; Peterson & Muehlenhard, 2007). Peterson and Muehlenhard (2007) argued that consent and wantedness are distinct components; feelings of wantedness and sexual desire may contribute to an individual’s decision to consent but should not be mistaken for actual consent. As such, it is important to highlight that the current study focuses on USEs that are both unwanted and non-consensual.

Measuring USEs in the Context of HEI

The Sexual Experiences Survey [SES] is one of the most commonly used tools to assess the incidence of USEs (Anderson et al., 2018; Canan et al., 2020; Johnson et al., 2017; Littleton et al., 2019). Several modifications have been made to improve its reliability and validity with the current version [SES-R] developed in 2006 (Koss et al., 2007). These changes include the phrase “when you didn’t want to” (Koss & Gidycz, 1985, p. 422) changed to “without my consent” (Koss et al., 2007, p. 359) because it was suggested that the previous wording did not adequately convey non-consent. Similarly, the verbal coercion category was expanded to include two new items focusing on negative coercion (e.g., using lies or false promises, showing displeasure) to reflect growing knowledge about the different types of sexual coercion (Koss et al., 2007; Livingston et al., 2004).

The SES-R is a collective name for four different versions of prevalence surveys that were developed during the revision process: a short and long form for assessing sexual victimization (SES-SFV and SES-LFV, respectively) and perpetration (SES-SFP and SES-LFP, respectively; Koss et al., 2007). Both versions include a set of behaviorally specific questions for participants to report the number of times a given perpetration tactic was used to achieve a certain behavior. The short form includes seven behaviorally specific items (each associated with five tactics). During data analysis, these behavioral items are combined to create five USE types: unwanted sexual contact [USC], attempted coercion, coercion, attempted rape, and rape. USC is the only non-penetrative sexual act within the SES-SFV and includes behavior such as sexualized touching, kissing, and groping. It is also the only USE type that depends exclusively on the act rather than a combination of a specific set of tactics. The remaining categories all reference different types of penetration (see Table 1). The current study focused specifically on the SES-SFV.

Table 1. Different behaviors measured within the SES-R.

USE behavior type	Description
Unwanted sexual contact	Sexualized touching, kissing, and groping; irrespective of the type of tactic used
Attempted coercion	Lies, showing displeasure; either of these tactics in combination with attempted anal, oral, or vaginal penetration
Coercion	Lies, showing displeasure; either of these tactics in combination with completed anal, oral, or vaginal penetration
Attempted rape	Incapacitation, threats, or force; any of these tactics in combination with attempted anal, oral, or vaginal penetration
Rape	Incapacitation, threats, or force; any of these tactics in combination with completed anal, oral, or vaginal penetration.

Two notable outliers of the USE behavioral categorizations are attempted coercion and coercion – the only two categories that are labeled by the perpetration tactic resulting in the USE. Koss et al. (2007) appeared to indicate two reasons for this choice. First, unlike rape or USC, coerced unwanted sexual acts that did not involve threats or similar were not previously considered a crime (despite acknowledgment that they should be). Therefore, it would be inappropriate to group these experiences with recognized criminal acts. Moreover, previous research (Koss et al., 1987; Livingston et al., 2004) indicated that coercive tactics reflected female experiences; therefore, the coercion category was retained. The extent to which coercive tactics reflect the male or gender minority ([GM]; e.g., students who identify as neither man nor woman) experience appears not to have been evaluated during the modification process.

Secondly, a community sample of women did not rate sexual coercion as seriously as rape (Abbey et al., 2004); therefore, it was argued that attempted and completed penetrative sexual acts should be differentiated between those by coercion and, those by incapacitation, threat, or force. The severity of coercion (in reference to the other USEs) was only evaluated by women whose mean age was older than the typical university sample at the time and were not university students themselves. In sum, the SES-SFV categories may only represent the experiences of women because their construction appears to have been explicitly based on women’s experiences. As such, it is necessary to assess the validity of the SES-SFV to determine whether the categories are representative of all participants’ experiences. The extent to which the SES-SFV is a valid and reliable measure, or its accuracy in measuring a number of USE types and how consistently it is able to do this, can be evaluated in several different ways. The current paper explores two such methods – the perpetration tactics included in the SES-SFV and how it is scored.

Validity of the SES-SFV

Research evidence indicates that the SES-SFV is a valid and reliable measure for female community members (Davis et al., 2014) and female university students (Johnson et al., 2017). Less evidence exists, however, as to its validity for use with male (Anderson et al., 2018), GM and sexual minority ([SM]; e.g., lesbian, gay, bisexual) students (female SM: Canan et al., 2020). Given that the SES was developed and modified using female samples, it is not surprising that some research suggests that the SES-SFV is more valid for female students (heterosexual and SM) than male students (Anderson et al., 2021). It might be possible to improve the validity of the SES-SFV by examining whether the current response options (perpetration tactics) are representative of current students. For example, previous research has indicated that two additional perpetration tactics (“making me feel refusing was useless” and “just doing the behaviour without giving me a chance to say ‘no’”; Canan et al., 2020, pp. 1071–72) might be more inclusive of SM students’ experiences. Similarly, Anderson et al. (2020) modified the SES-SFV to include “forced-to-penetrate” (the victim is forced to penetrate the perpetrator) and the results suggested that male students, in particular, reported this as a common experience. This suggests that current scoring options may require modification to ensure that all students’ experiences are captured.

Scoring the SES-SFV

In addition to the content of the SES-SFV, a recent area of discussion has concerned the scoring of SES-SFV data. Koss et al. (2007) outlined three options for scoring the SES-SFV: (1) calculating the frequency of individual items; (2) creating redundant scores to report prevalence by USE categories (e.g., non-victimization, sexual contact); or (3) creating non-redundant (or, mutually exclusive) scores to report prevalence according to participants’ most severe USE. Redundant scoring may be most reflective of participants’ experiences because percentages represent all USEs reported; however, participants can be included in multiple USE categories depending on their experiences. Non-redundant scoring may be useful for analytical purposes because participants are placed into the category of their most “severe” experience; however, these scores do not account for the frequency of less severe experiences. Research suggests that severity ratings of USEs are subjective, and that the assessment of additional characteristics is required to ensure that this type of scoring is reliable (Davis et al., 2014; Testa et al., 2004). Further, the rationale for the severity ordering, as discussed above, appears to primarily rely upon female experiences of non-consensual heterosexual sexual activity, so it is unclear whether men, GM or SM would order the severity of USEs differently. Dichotomous scoring of categories (i.e., yes/no), however, has good test–retest reliability among female university students (Johnson et al., 2017) and adequate reliability among male students (Anderson et al., 2018) indicating that, until further testing occurs, it may be the most reliable scoring method. To our knowledge, only one study (Davis et al., 2014) has comprehensively examined the different scoring protocols for the SES-SFV, but this study was limited to a female-only community-based sample.

Current Study

The present study focused on an analysis of data from a larger project assessing USEs of university students attending two NI universities. The aims of the paper were to (1) examine the prevalence of USEs reported by these students since attending university; (2) evaluate the validity of two new perpetrator tactics by assessing their unique contribution; and (3) determine the impact of different scoring strategies on how data is presented.

Method

Procedure

Data was collected concurrently but, respectively, at two NI universities using an online survey between November 2020 until May 2021. Having had a USE was not an inclusion criterion of survey participation, as future analyses would involve a comparison between participants reporting USEs and participants who did not. Participants were recruited through an invitation sent to their university e-mail address; reminders to participate in the survey were sent four times within both universities. This invitation briefly described the research (including the focus on USEs), stated the criteria (over 18 years old and registered at the university), the length of time required to complete the survey (30–40 minutes), information about the prize draw incentive (six £50 Amazon gift cards with winners chosen using a random number generator) and a link to the survey hosted on the Qualtrics online software platform. Potential participants were also invited to complete the survey via social media platforms.

After clicking the link, participants received more detail about the study and a consent form to complete. The informed consent process described participants’ contribution as voluntary, data management procedures, confidentiality, and anonymity and the data withdrawal process. Participants were not asked to provide any identifying information. All research procedures were approved by each university’s ethics committee prior to commencing data collection. Further, to improve the face validity of our survey, it was reviewed by internal and external support organizations, academics, and student representatives local to NI to ensure that the language, level, and experiences included were appropriately captured based on the experiences of support providers in NI.

Participants

Participants were 830 women, 188 men and 15 GM¹ (e.g., did not describe their gender as man or woman) attending two universities in NI. Participants were aged between 18 and 59 years (M = 23.77, SD = 6.14) and most identified as either Irish (37%) or Northern Irish (31%). The majority (90%) of participants were White/European. Most participants identified as heterosexual (75%), with 23% identifying as an SM (e.g., lesbian, gay, bisexual). Data from both universities concerning the demographics of the student body during 2019/20 suggested that the student body was approximately 48,882 students; most students were women (57%), aged over 18 (91%) and identified as White ((86%); minority and sexual orientation data was not available).

Measures

Demographics

Participants reported their age, gender, sexual orientation, ethnicity, nationality, year, and mode of study (e.g., full time).

Sexual Experiences Survey (Short Form-Victimization) ([SES-SFV]; Koss et al., 2007)

A modified version of the SES-SFV assessed participants’ USEs since starting university. Participants responded to seven behaviorally specific questions about (1) sexualized touching; (2) attempted oral, vaginal, or anal penetration; and (3) completed oral, vaginal, or anal penetration, and all were described using gender neutral language. Male participants (or those with a penis) did not complete questions concerning vaginal penetration. Each question (see Table 2) was accompanied by seven subquestions focusing on the tactics used to elicit the activity including two forms of verbal coercion (tactics a and b), taken advantage whilst incapacitated (tactic c) and two forms of physical force (tactics d and e). We included two additional tactics (tactics f and g) as research indicated that they might uniquely capture participants’ experience (Canan et al., 2018, 2020; Jozkowski et al., 2017). Participants were asked how often each experience had occurred since starting university and could report using response options ranging from 0 (0 times) to 3 (3 or more times). Participants could also decline to answer by selecting “prefer not to say” (4) but these selections were excluded from analyses. Cronbach’s alpha was not computed for this measure because the SES-SFV does not measure a latent construct that causes sexual victimization, rather the occurrence of it (Diamantopoulos et al., 2008; Hulme, 2007; Koss et al., 2007).

Table 2. Tactics included in the modified SES-SFV.

Tactic	Description
A	Telling lies, threatening to end the relationship, threatening to spread rumors about me, making promises I knew were untrue, or continually verbally pressuring me after I said I didn’t want to.
B	Showing displeasure, criticizing my sexuality or attractiveness, getting angry but not using physical force, after I said I didn’t want to.
C	Taking advantage of me when I was too drunk or out of it to stop what was happening.
D	Threatening to physically harm me or someone close to me.
E	Using force, for example, holding me down with their body weight, pinning my arms, or having a weapon.
F	Making me feel as though refusing was useless.
G	Just doing the behavior without giving me a chance to say “no” (e.g., surprising me with the behavior).

Analytic Plan

Both university-specific datasets were cleaned and screened prior to data merging for the purposes of analyses. Participants were removed for starting but not responding to any items, failing to provide consent, not completing the demographics measure or completing 20% or less of the SES-SFV. Whilst 1760 participants began the survey, only 1412 completed the demographic questions (completion rate = 80%) and, of those participants, 1033 completed more than 80% of the SES-SFV. Data from the latter group were included in all subsequent analyses. Data was merged in an SPSS version 26 data file with corresponding variables recoded for compatibility.

Following Davis et al. (2014) and Anderson et al. (2018), descriptive statistics and frequency scores were calculated using dichotomous, continuous (or, frequency) and severity scores. Gender and sexual orientation differences were assessed using the continuous scores with Mann–Whitney U tests because the data were non-normally distributed. The two new tactics included in the current study have only been tested on a mixed-sexual orientation female sample; as such, descriptive statistics are reported for all tactics by gender and sexual orientation. The calculation of these scores is described in full below and reflects the scoring procedures used by Koss et al. (2007).

Scoring Procedures

Dichotomous Scores

To calculate dichotomous scores (USE: yes/no), any SES item endorsed as “1” (once) or greater were coded as having been subjected to at least one USE. Participants endorsing all items with a response of “0” were coded as having experienced no USEs. Dichotomous scores were also calculated by behavioral outcome (see mutually exclusive categories column in Table 3) but these scores are not mutually exclusive (i.e., a participant could be included in two categories).

Table 3. Process of creating mutually exclusive categories.

Corresponding question	Tactics combined	Mutually exclusive category
All questions	Reported “0” across all tactics	No experiences
Question 1	All tactics attached to sexualized touching	Unwanted sexual contact
Question 5, 6 or 7	Tactics a or b attached to attempted oral, vaginal, or anal penetration	Attempted coercion
Question 2, 3 or 4	Tactics a or b attached to completed oral, vaginal, or anal penetration	Coercion
Question 5, 6, or 7	Tactics c, d, e, f or g attached to attempted oral, vaginal, or anal penetration	Attempted rape
Question 2, 3 or 4	Tactics c, d, e, f or g attached to completed oral, vaginal, or anal penetration	Rape

Continuous Scores

To calculate continuous scores, dichotomous scores (as calculated above) were summed. For participants completing all items (including tactics f and g), scores ranged from 0 to 72; participants who did not view the vaginal items had scores ranging from 0 to 59. With tactics (f) and (g) excluded, scores ranged from 0 to 35 (or 0–28 for participants who did not view the vaginal items). These scores represent a summed dichotomous score of each type of incident per tactic (e.g., USC by lies, USC by force etc.) so cannot differentiate between, for example, multiple incidents involving the same perpetrator and a single incident involving several tactics.

Severity Scores

Severity scores were calculated using the procedure outlined by Koss et al. (2007) to create mutually exclusive categories visualized in Table 3. Participants were then scored according to their most “severe” experience. For example, participants reporting a rape were scored as “5” irrespective of whether they had also reported USC.

Validation Efforts

Frequency endorsements of the two response options (tactics f and g) were determined by calculating the percentage of participants endorsing each tactic at least once. Then, unique endorsements of the new response options were calculated and compared to the percentage of participants uniquely endorsing the existing SES-SFV response options. Scoring procedures described above were performed twice, with existing response options and the addition of tactics (f) and (g).

Results

Dichotomous Scores

Sixty-three percent (n = 650) of the sample reported experiencing at least one USE whilst attending university. If tactics (f) and (g) were excluded from the survey this reduced to 53% (n = 546) of participants recorded as experiencing a USE. Prevalence estimates were also analyzed against participants’ gender and sexual orientation with the results presented in Tables 4 and 5. Due to low cell counts and to prevent class disclosure (also referred to as jigsawing),² participants who chose not to report their gender (unidentified) or self-described their gender [SD] were grouped together; similarly, participants who self-described their sexual orientation or chose not to report it were also grouped together. Sixty-seven percent of women (n = 559), 44% (n = 83) of men and 53% (n = 8) of SD/unidentified participants reported at least one USE. If tactics (f) and (g) were excluded from the results all these figures decreased (women: n = 475, 57%; men: n = 66, 35%; SD/unidentified participants: n = 5, 33%).

Table 4 Dichotomous scoring across reported USEs by gender.

			Women (n=830)	Men (n=188)	SD/unidentified (n=15)	Total (n=1033)
All reported USEs	Original items only	At least one experience	475 (57%)	66 (35%)	5 (33%)	546 (53%)
		No experiences	355 (43%)	122 (65%)	10 (67%)	487 (47%)
	Additional items^2	At least one experience	559 (67%)	83 (44%)	8 (53%)	650 (63%)
		No experiences	271 (33%)	105 (56%)	7 (47%)	383 (37%)
Unwanted sexual contact	Original items only	At least one experience	424 (51%)	51 (27%)	5 (33%)	480 (47%)
		No experiences	406 (49%)	137 (73%)	10 (67%)	553 (54%)
	Additional items^2	At least one experience	507 (61%)	69 (37%)	7 (47%)	583 (56%)
		No experiences	323 (39%)	119 (63%)	8 (53%)	450 (44%)
Attempted coercion	Original items only^1	At least one experience	187 (23%)	21 (11%)	2 (13%)	210 (20%)
		No experiences	643 (78%)	167 (89%)	13 (87%)	823 (80%)
Coercion	Original items only^1	At least one experience	223 (27%)	24 (13%)	0 (0%)	247 (24%)
		No experiences	607 (73%)	164 (87%)	15 (100%)	786 (76%)
Attempted rape	Original items only	At least one experience	210 (25%)	26 (14%)	2 (13%)	238 (23%)
		No experiences	620 (75%)	162 (86%)	13 (87%)	795 (77%)
	Additional items^2	At least one experience	291 (35%)	36 (19%)	2 (13%)	329 (32%)
		No experiences	539 (65%)	152 (81%)	13 (87%)	704 (68%)
Rape	Original items only	At least one experience	285 (34%)	32 (17%)	2 (13%)	319 (31%)
		No experiences	545 (66%)	156 (83%)	13 (87%)	714 (69%)
	Additional items^2	At least one experience	385 (46%)	40 (21%)	5 (33%)	430 (42%)
		No experiences	445 (54%)	148 (79%)	10 (67%)	603 (58%)

¹Attempted coercion and coercion are not associated with items (f) and (g); ²Additional items include all tactics (a-g)

Table 5. Dichotomous scoring across reported USEs by sexual orientation.

			Heterosexual (n = 775)	Gay/Lesbian (n = 39)	Bisexual (n = 174)	SD/unidentified (n = 45)	Total (n = 1033)
All reported USEs	Original items only	At least one experience	396 (51%)	22 (56%)	111 (64%)	17 (38%)	546 (53%)
		No experiences	379 (49%)	17 (44%)	63 (36%)	28 (62%)	487 (47%)
	Additional items^b	At least one experience	478 (62%)	24 (62%)	126 (72%)	22 (49%)	650 (63%)
		No experiences	297 (38%)	15 (39%)	48 (28%)	23 (51%)	383 (37%)
Unwanted sexual contact	Original items only	At least one experience	344 (44%)	17 (44%)	104 (60%)	15 (33%)	480 (47%)
		No experiences	431 (56%)	22 (56%)	70 (40%)	30 (67%)	553 (54%)
	Additional items^b	At least one experience	425 (55%)	19 (49%)	119 (68%)	20 (44%)	583 (56%)
		No experiences	350 (45%)	20 (51%)	55 (32%)	25 (56%)	450 (44%)
Attempted coercion	Original items only^a	At least one experience	152 (20%)	11 (28%)	40 (23%)	7 (16%)	210 (20%)
		No experiences	623 (80%)	28 (72%)	134 (77%)	38 (84%)	823 (80%)
Coercion	Original items only^a	At least one experience	180 (23%)	11 (28%)	49 (28%)	7 (16%)	247 (24%)
		No experiences	595 (77%)	28 (72%)	125 (72%)	38 (84%)	786 (76%)
Attempted rape	Original items only	At least one experience	168 (22%)	12 (31%)	47 (27%)	11 (24%)	238 (23%)
		No experiences	607 (78%)	27 (69%)	127 (73%)	34 (76%)	795 (77%)
	Additional items^b	At least one experience	235 (30%)	15 (39%)	64 (37%)	15 (33%)	329 (32%)
		No experiences	540 (70%)	24 (62%)	110 (63%)	30 (67%)	704 (68%)
Rape	Original items only	At least one experience	224 (29%)	18 (46%)	69 (40%)	8 (18%)	319 (31%)
		No experiences	551 (71%)	21 (54%)	105 (60%)	37 (82%)	714 (69%)
	Additional items^b	At least one experience	304 (39%)	21 (54%)	92 (53%)	13 (29%)	430 (42%)
		No experiences	471 (61%)	18 (46%)	82 (47%)	32 (31%)	603 (58%)

^aAttempted coercion and coercion are not associated with items (f) and (g).

^bAdditional items include all tactics (a-g).

Sexual Orientation

Sixty two percent of heterosexual participants (n = 478), 62% (n = 24) of gay/lesbian participants, 72% (n = 126) of bisexual participants and 49% (n = 22) of SD/unidentified participants reported at least one USE. If tactics (f) and (g) were excluded from the results all these figures decreased (heterosexuals: n = 396, 51%; gay/lesbians: n = 111, 64%; bisexuals: n = 22, 56%; SD/unidentified: n = 17, 38%).

Continuous Scores

Using total continuous scores, the mean SES-SFV score for the sample was 7.84 (SD = 11.46) with a range from 0 to 72. Fifty-five percent of participants (n = 571) reported more than one incident or that a perpetrator used more than one tactic. When tactics (f) and (g) were excluded from the survey results, the mean score was 3.29 (SD = 5.27) with a range from 0 to 35 and 44% of participants (n = 451) reporting multiple USEs.

Gender

The mean SES-SFV score for women was 8.64 (SD = 11.54), 4.69 (SD = 10.91) for men and 3.33 (SD = 4.55) for SD/unidentified participants. When tactics (f) and (g) were excluded from the results, the mean scores (and standard deviations) were 3.62 (SD = 5.30), 2.04 (SD = 5.11) and 1.00 (SD = 1.60), respectively. Statistically significant differences were found between men and women whether tactics (f) and (g) were included (U = 54,485.50, p < .001) or excluded (U = 57,575.00, p < .001), with male participants scoring lower.

Sexual Orientation

The mean SES-SFV score for heterosexual participants was 7.28 (SD = 11.08), 11.59 (SD = 15.86) for gay/lesbian participants, 10.13 (SD = 12.13) for bisexual participants and 5.47 (SD = 9.15) for SD/unidentified participants. When tactics (f) and (g) were excluded from the results, the mean scores (and standard deviations) were 3.07 (SD = 5.10), 5.21 (SD = 7.41), 4.20 (SD = 5.64) and 2.02 (SD = 3.51), respectively. Statistically significant differences were found between heterosexual and non-heterosexual³ participants whether tactics (f) and (g) were included (U = 83,565.00, p = .007) or excluded (U = 84,177.00, p = .008), with heterosexual participants scoring lower.

Severity Scores

Participants’ reported USEs were divided into six mutually exclusive categories ranging from “no experiences” to “rape” (most severe). The results of this categorization, with tactics (f) and (g) included and excluded, are summarized in Table 6.

Table 6. Severity categorical scoring across reported USEs against gender.

		Women (n = 830)	Men (n = 188)	SD/unidentified (n = 15)	Total (n = 1033)
No experiences	Original items only	355 (43%)	122 (65%)	10 (67%)	487 (47%)
	Additional items^b	271 (33%)	105 (56%)	7 (47%)	383 (37%)
Unwanted sexual contact	Original items only	69 (8%)	16 (9%)	2 (13%)	87 (8%)
	Additional items^b	90 (11%)	28 (15%)	3 (20%)	121 (12%)
Attempted coercion	Original items only	18 (2%)	2 (1%)	0 (0%)	20 (2%)
	Additional items^b	10 (1%)	0 (0%)	0 (0%)	10 (1%)
Coercion	Original items only^a	55 (7%)	10 (5%)	0 (0%)	65 (6%)
	Additional items^b	21 (3%)	7 (4%)	0 (0%)	28 (3%)
Attempted rape	Original items only^a	48 (6%)	6 (3%)	1 (7%)	55 (5%)
	Additional items^b	53 (6%)	8 (4%)	0 (0%)	61 (6%)
Rape	Original items only	285 (34%)	32 (17%)	2 (13%)	319 (31%)
	Additional items^b	385 (46%)	40 (21%)	5 (33%)	430 (42%)

^aAttempted coercion and coercion are not associated with items (f) and (g).

^bAdditional items include all tactics (a-g).

Gender

The results of severity categorization against participants’ gender are reported in Table 6. When comparing the severity categories across genders, the majority of women (n = 385, 46%) reported the most severe experience (rape) whilst the majority of men (n = 105, 56%) and SD/unidentified (n = 7, 47%) participants were categorized as reporting no experiences. The exclusion of tactics (f) and (g) resulted in an increase in participants being categorized as reporting no experiences across all genders (women: n = 355, 43%; men: n = 122, 65%; SD/unidentified: n = 10, 67%) because they had only reported on USEs tied to these tactics.

Sexual Orientation

The results of severity categorization against participants’ sexual orientation are reported in Table 7. When comparing the severity categories across sexual orientation, the majority of gay/lesbian (n = 21, 54%), bisexual (n = 92, 53%) and heterosexual (n = 304, 39%) participants were categorized with the most severe experience and the majority of SD/unidentified participants (n = 23, 51%) were categorized as reporting no experiences. When tactics (f) and (g) were excluded, the majority of gay/lesbian participants (n = 18, 46%) and bisexual participants (n = 69, 40%) were still categorized with the most severe experience (rape) whilst the majority of heterosexual (n = 379, 49%) and SD/unidentified (n = 28, 62%) participants were re-categorized as reporting no experiences.

Table 7. Severity categorical scoring across reported USEs against sexual orientation.

		Heterosexual (n = 775)	Gay/Lesbian (n = 39)	Bisexual (n = 174)	SD/unidentified (n = 45)	Total (n = 1033)
No experiences	Original items only	379 (49%)	17 (44%)	63 (36%)	28 (62%)	487 (47%)
	Additional items^b	297 (38%)	15 (39%)	48 (28%)	23 (51%)	383 (37%)
Unwanted sexual contact	Original items only	67 (9%)	0 (0%)	18 (10%)	2 (4%)	87 (8%)
	Additional items^b	97 (13%)	1 (3%)	20 (12%)	3 (7%)	121 (12%)
Attempted coercion	Original items only^a	15 (2%)	1 (3%)	3 (2%)	1 (2%)	20 (2%)
	Additional items^b	8 (1%)	0 (0%)	1 (1%)	1 (2%)	10 (1%)
Coercion	Original items only^a	50 (7%)	2 (5%)	11 (6%)	2 (4%)	65 (6%)
	Additional items^b	24 (3%)	1 (3%)	3 (2%)	0 (0%)	28 (3%)
Attempted rape	Original items only	40 (5%)	1 (3%)	10 (6%)	4 (9%)	55 (5%)
	Additional items^b	45 (6%)	1 (3%)	10 (6%)	5 (11%)	61 (6%)
Rape	Original items only	224 (29%)	18 (46%)	69 (40%)	8 (18%)	319 (31%)
	Additional items^b	304 (39%)	21 (54%)	92 (53%)	13 (29%)	430 (42%)

^aAttempted coercion and coercion are not associated with items (f) and (g).

^bAdditional items include all tactics (a-g).

Validation Efforts

Of those reporting a USE, 53% (n = 342) were subjected to a USE because the perpetrator “[made them] feel as though refusing was useless” (tactic f). Comparatively, 81% (n = 525) reported that they were subjected to a USE because the perpetrator “just [did] the behaviour without giving [them] a chance to say ‘no’” (tactic g). By comparison, “showing displeasure …, getting angry but not using physical force, after I said I didn’t want to” (tactic b) and “taking advantage … when I was too drunk or out of it to stop what was happening” (tactic c) were endorsed by 53% (n = 342) and 60% (n = 392) of participants, respectively. In actuality, tactic (g) was the most commonly endorsed perpetration tactic across the whole sample.

Each response item’s unique contribution was also analyzed. Of those who reported at least one experience, 149 participants only responded to one item. Tactic (f) uniquely captured 13 (9%) participants’ responses that would not have been captured by any of the original SES-SFV items. Tactic (g) was uniquely endorsed by 75 (50%) participants. By comparison, tactic (b) and tactic (c) were uniquely endorsed by 17 (11%) and 34 (23%) participants, respectively.

Gender

Within the different gender groups, 53% (n = 299) of women, 45% (n = 37) of men and 75% (n = 6) of SD/unidentified participants reported that the perpetrator used tactic (f). By comparison, 82% (n = 458) of women, 72% (n = 60) of men and 88% (n = 7) of SD/unidentified participants reported that the perpetrator used tactic (g). When examining the gender dimensions of unique endorsements, tactic (f) was uniquely endorsed by 10 (8%) women, 2 (7%) men and 1 (55%) SD/unidentified person. Tactic (g), however, was uniquely endorsed by 61 (51%) women, 13 (48%) men and 1 (50%) SD/unidentified person.

Sexual Orientation

Across sexual orientation, 50% (n = 240) of heterosexual participants, 54% (n = 13) of gay/lesbian participants, 61% (n = 77) of bisexual participants and 55% (n = 12) of SD/unidentified participants reported that the perpetrator used tactic (f). By comparison, 79% (n = 376) of heterosexual, 83% (n = 20) of gay/lesbian, 89% (n = 112) of bisexual and 77% (n = 17) of SD/unidentified participants reported that the perpetrator used tactic (g). When examining unique endorsements against participants’ sexual orientation, tactic (f) was uniquely endorsed by 10 (8%) heterosexuals, 2 (13%) bisexuals and 1 (14%) SD/unidentified person. Tactic (g), however, was uniquely endorsed by 60 (54%) heterosexual, 2 (50%) gay/lesbian, 9 (56%) bisexual and 4 (57%) SD/unidentified people.

Discussion

The current study is the first in NI to assess the prevalence of USEs using a large, diverse sample of students attending two different universities in NI. The main findings can be summarized succinctly. First, using only the original SES-SFV response options, 47% of students reported at least one incident of USC and 31% were subjected to at least one incident of rape. Comparatively, including the two new response options, 56% reported at least one USC incident and 42% reported at least one incident of rape. Second, tactic g (“taken by surprise”) uniquely captured 50% of participants’ experiences and was the most commonly endorsed tactic across all USE types. Third, dichotomous scoring appears to be the most straightforward scoring strategy but there needs to be further consideration regarding individual tactics rather than outcomes.

Concerning the first aim, our findings were consistent with recent national research from the Republic of Ireland where students were asked to report on their experiences over the last 4 years at their HEI (MacNeela et al., 2022): 45% of students reported non-consensual sexual touching, 28% reported rape and women reported more USEs than men (overall USE percentage were not reported). Without conducting this kind of research, prevalence can only be estimated by reviewing individual universities’ internal reporting figures and/or national crime reports. Generally, sexual offenses are often underreported to university authorities and to the police, and research suggests that students are often unaware that their experience may meet the criminal threshold (Phipps & Smith, 2012). Moreover, they may blame themselves for their experience and feel a significant amount of shame (Ahrens, 2006; Fisher et al., 2003; Khan et al., 2018; Phipps & Smith, 2012). Specifically, regarding the former, reporting frameworks are not consistent across universities and may not be easy to find, thereby reducing the likelihood that students will report their experience (Axinn et al., 2021; Fonteñez et al., 2020).

New SES-SFV Response Options

The second aim was to assess the unique contribution of two new perpetrator tactics (tactic f: “ignorance of refusal” and tactic g: “taken by surprise”) following research by Canan (2017) and others (Canan et al., 2018, 2020; Jozkowski et al., 2017). Research would suggest that these additional response options may capture when people are made to feel that their refusal is useless (tactic f), or the perpetrator’s use of surprise to engage in sexual activity (tactic g). Results indicated that they did offer a unique contribution (particularly, “taken by surprise”); 13 (9%) and 75 (50%) students were provided the opportunity to report at least one USE, respectively. When included in the analysis, the prevalence of students reporting at least one USE increased to 63%. Similarly, Canan et al. (2020) found that “taken by surprise” uniquely captured 9% (n = 86) of participants’ experiences but “ignorance of refusal” was less effective (2%, n = 17). They suggested that “ignorance of refusal” might be a redundant item that is actually captured in the original SES-SFV set but also that it focuses on how the victim felt rather than the actual perpetrator tactics used.

Nonetheless, the SES-SFV has not been updated by its authors since 2007 and indication that “taken by surprise” might be a useful conclusion begs the question whether there are other tactics that might be more in keeping with different types of students’ experiences. Among a national sample of lesbian, bisexual, and heterosexual women, an additional 12 perpetration tactics were identified using an open-ended narrative (Canan et al., 2020) and these are not currently reflected in the SES-SFV. Examples of these tactics include intentionally seeking to incapacitate the victim (though this is in the SES-LFV) or taking the victim to a less public space to prevent others from intervening. There is some suggestion that mixed methodology may be useful in evaluating the SES-SFV, although there are not many examples of this. Canan (2017), for example, surveyed students with a modified version of the SES-SFV; participants were given the opportunity to add more detail to any victimization experience and this data was thematically analyzed. This process provided the opportunity to better understand how participants understood each tactic and the parallels they drew to their own experience. Recently, Littleton et al. (2019) sought to evaluate the reliability and validity of a modified SES-SFV (referred to as SES-R) using a mixed-gender sample but only assessed USC, attempted rape, and rape by way of threats, force and/or incapacitation. When comparing participants’ narratives of USEs to the reported USEs, they found moderate consistency but a false-positive rate of 20%. Notably, they highlighted that some USEs reported in the written narratives were inconsistent with those captured by the SES-SFV; they suggested that this could be because some of the written experiences would not meet the legal threshold for sexual assault. They concluded that it might be more appropriate to use the SES-SFV in conjunction with another tool to evaluate prevalence of sexual victimization and this recommendation was also made by Koss et al. (2007).

SES-SFV Scoring Options

The final aim was to determine the impact of different scoring options within the SES-SFV. Similar to Davis et al. (2014), the results of this study would suggest that the choice of scoring option largely depends on the study aim(s). Dichotomous scoring appears to be a popular option and the most straightforward (Williams et al., 2020). In our study, study we chose to calculate frequency using the dichotomous scores; therefore, our continuous scores represent the number of different types of events reported by participants rather than the total number of events reported. If researchers are interested in calculating the frequency of USEs (e.g., total number of events reported) or detecting relationships relevant to repeated USEs. In these instances, continuous scoring should be the preferred option. In our study, for example, 67% women reported at least one USE, but their mean score was 8.64 so, on average, women were subjected to almost 9 different types of incidents (but not necessarily 9 different incidents). If examining tactic endorsement rather than the outcomes – 78% of participants endorsed more than one tactic and this has been found in previous research (Canan, 2017). Unique tactic endorsement may be more informative than multi-tactic endorsement because, in its current iteration, the SES-SFV cannot identify whether continuous scores or multi-tactic endorsement refer to multiple USEs across one single incident involving one perpetrator or multiple USEs by different perpetrators across multiple incidents.

An additional consideration would be the composition of the categories themselves. In the SES-SFV, participants subjected to attempted or completed penetrative USEs as a result of incapacitation, threat or physical force are categorized as having been subjected to attempted rape or rape, respectively (Koss et al., 2007). This may reflect legal definitions of certain sexual offenses (jurisdiction dependent), but it ignores key differences across tactics. Research suggests that the psychological outcomes differ depending on the type of USE perpetration tactic used on the individual (Brown et al., 2009; Littleton et al., 2009; O’Callaghan & Ullman, 2021; Zinzow et al., 2012). For example, Brown et al. (2009) reported that the traumatic impact of incapacitated rape was less than that of forcible rape but greater than that of coerced rape. Their recommendation was to consider examining forcible and incapacitated rape separately, particularly with respect to psychological consequences. This returns to the point of scoring choice reflecting this study’s aims. Cook et al. (2011), for example, acknowledged that presenting USE data in a way that conforms to legal statutes is not always necessary, and that empirical data that does not run parallel to the law could be used as a method of highlighting where the law does not match experience. To this point, it is also not clear whether severity scoring benefits all participants. The determination of the increasing severity afforded to each USE type was initially made using an entirely female sample (Koss, 1985) and the validation of these severity categories also involved an all-female community sample (Testa et al., 2004) but used an earlier iteration of the SES (Koss et al., 1987). It was during this validation process that coerced penetrative sex was deemed to be less severe than attempted penetrative sex. Our team is not aware of any studies that have validated the SES-SFV severity categories using a mixed or all-male sample.

Limitations

This study was not without limitations. Whilst it is lacking in non-female and LGBTQIA+ participants, both universities report a majority female student body (57%). Though we achieved a relatively large sample by comparison to similar studies (see Fedina et al. (2016) for review), our sample only represented 2% of both universities’ combined student population; therefore, it would not be appropriate to generalize these figures and further research is required to validate prevalence. Similarly, whilst women outnumber men in the student population, we recognize that our sample is still disproportionately female. The estimated victimization recorded in this study is high and it is possible that the study’s focus attracted a greater number of students who had been subjected to USEs. Though we were clear that participation was not dependent on a history of USEs, it is possible that students took advantage of the opportunity to report their experiences as this type of university-based study has not occurred since 2016 (Haughey et al., 2016). Whilst participants with vaginas answered a specific set of questions concerning vaginal penetration, we did not include items concerning participants (typically, male) being forced to penetrate someone else (Bates, 2020; Hines & Douglas, 2016; Weare, 2018). We would suggest that future researchers consider this as a possible modification to the SES-SFV. Lastly, we focused on physical USEs so cannot speak to, for example, online sexual victimization that may have increased as a result of the pandemic (Eaton et al., 2022).

Implications

Information is power. Inasmuch as we lack UK-based research, students attending NI universities have been disproportionately underrepresented in the available research. In 2016, students attending one NI university (Haughey et al., 2016) could report their experiences and results suggested that 33% experienced sexual harassment, 8% experienced attempted penetrative sex and 6% experienced penetrative sex. The explicit focus on one university aside, the manner in which the data was presented makes it difficult to compare results reliably. Nonetheless, 5 years later it is clear that the problem is no less serious. Our study provides an indication of the scale of the problem; further research is always a necessity, but we would advocate for more support from institutions and the relevant government organizations to implement trauma informed support services, reporting frameworks and awareness campaigns.

Our findings indicate that a mixed methodological research approach (e.g., SES-SFV and open narrative responses) is necessary to evaluate the validity of the SES-SFV among mixed-gender and sexual orientation populations to determine the current validity of the SES-SFV and, if necessary, how it could be improved. Whilst the brevity of the SES-SFV will likely reduce fatigue, if there are items in the SES-LFV that are highly relevant to the student experience, these should be considered as possible modifications. Davis et al.’s (2014) work laid a foundation for evaluating the validity and reliability of the scoring protocol yet there is no examination of the suitability of scoring options among non-female student samples. More specifically, the time has come for the perceived severity ordering to be evaluated by non-female samples to ensure that it is wholly representative rather than female-centric. Lastly, it is recognized that different types of tactics result in differing psychological impact, yet, the current recommendation of data presentation in the SES-SFV can only differentiate between USE type rather than tactic; an evaluation of the different types of data ordering might highlight differences that were previously hidden. These differences strengthen our contextual knowledge of university-based USEs and can result in better informed awareness campaigns, consent knowledge workshops and so on. Without these data, these figures only supplement what the police (and others) might be missing rather than highlight the nuanced issues.

Conclusion

Much debate has centered around how USEs are defined, particularly within the context of HEI (Cook et al., 2011; Krebs, 2014; Moylan et al., 2018). At least three USE definitions are at play within the HEI focused research: the research team’s definition, the participant’s definition, and the HEI’s definition. The legal definition of a USE may also play a role depending on whether this is operationalized by the research team and/or the HEI. Though it is possible that students may also operationalize the legal definition, research suggests that this is uncommon (Artime & Peterson, 2015; Baldwin-White & Bazemore, 2020; Franiuk, 2007; Hills et al., 2020; Peterson & Muehlenhard, 2007). With this understanding, the necessity of reporting prevalence of USE by criminal classification appears to be less about the student’s experience and more to compare against official reports where the legal framework is more common. Relying on a legal framework to report USEs may have been more appropriate in the past yet this does not seem in keeping with the behaviorally specific phrasing of the SES-SFV items – it would seem more fitting to report USEs by the different tactics reported, sectioned by USE type if this meets the aims of the study. Our region-specific knowledge in this area is growing as UK-based research becomes more available but, to match this growth, methodological tools must also evolve to ensure that they accurately capture experiences.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ Five participants did not identify their gender, nor did they identify within the gender binary.

² “Class disclosure” refers to situations where all or none of the observations fall into one category, therefore, running the risk that an individual (or set of individuals) would be identified by nature of being the only ones who belong to that “class” (or group).

³ Participants choosing “prefer not to say” were not included.