Introduction

Sexual aggression is a public health problem that can affect people at any moment in life, it produces serious consequences at physical, psychological and occupational levels, for this reason it is necessary to know its magnitude and scope ^1,3.

Sexual aggression is a subtype of sexual violence, it is defined as the use of certain mechanisms like physical force, intimidation and other coercive actions to force men or women to any kind of sexual encounter, against their will, thus infringing their sexual freedom through coercion, manipulation and deceit ^4,7.

Based on the severity degree of the sexual aggression it is possible to identify four subtypes of sexual aggression: (a) sexual contact, defined as the union of sexual nature with another person, without penetration, and against her/his will be means of verbal pressure, by taking advantage of the victim's inability to resist the use, or the threat of use, of physical force; (b) sexual coercion, understood as the use of practices such as subtle psychological and verbal pressure in order to have a sexual encounter, thus achieving oral, vaginal or anal penetration; (c) attempted rape, defined as the attempt to penetrate orally, vaginally or anally by means of exploiting the victim's inability to oppose resistance to the use, or the threat of use, of physical force; and (d) rape, understood as the violent carnal access with the penis or other objects, through the vagina, anus, and/or the mouth, including the use of physical force and/or taking advantage of the victim's inability to oppose resistance ^7,10.

Determining the prevalence of sexual aggression and its subtypes in the environment of higher education institutions is a task that has been done for over 30 years ^9,11. Data collection on this problematic has been carried out for a long time through researches that use self-administered tools in physical format and now also with the help of technology, in the digital format ^6,7. Current figures of sexual aggression in the university environment can be considered an estimation of the magnitude of the problem, since it is not possible to assure that all persons that have been victims or offenders of this situation report this in the investigations, and even less in the case of the victims who had denounced the occurrence before any institution or responsible entity while they were studying at the university ¹².

At present, there is no updated evidence that at a global level synthesizes prevalence studies on sexual aggression in students from day one of their university life. It is relevant to know the prevalence of sexual aggression in university campuses since there is the contrasting fact that academic environments have been considered for years safe spaces for students. It is therefore necessary to have updated figures that allow synthesizing the reports on this global problematic in order to implement strategies that permit creating guidelines for helping victims and for establishing protocols for the prevention of sexual aggression. Considering the above, the research question for the present review is which is the prevalence of victimization and perpetration of sexual aggression in undergraduate students from day one at the university?

This is the first proposal developed in Latin America concerning studies conducted with samples of Latin America and other regions, whose objective was to systematically review and synthetize the findings of the studies on prevalence of sexual aggression -perpetration and victimization- in undergraduate students of both genders, from the beginning of their university life on, in the last ten years (2008-2018).

Materials and Methods

The methods applied in this study are based on the PRISMA Declaration and the recommendations given by Montero and Leon on the research methodologies in psychology ^13,14. Two authors carried out the systematic and independent search in the Scopus®, Web of Science®, Scielo®, Redalyc® and Eric® databases of the researches published between January 1^st 2008 and February 20^th 2018, in English as well as in Spanish. The descriptors used were University/College/Universidad, Sexual Violence/Sexual Assault/ Sexual Aggression/Agresión Sexual/Violencia Sexual/Violación, and Prevalence/Prevalencia, all of them terms included in the Biblioteca Virtual en Salud (http://bvsalud.org/es/). One of the search strategies used in the data base Web of Science® was subject: University AND college AND "sexual assault" AND prevalence; time span: 2008-2018; indices: SCI-EXPANDED, SSCI, A&HCI, ESCI.

Independently, two authors extracted data of the studies using an Excel® matrix with the information of the selected articles including ID, author, publication year, journal, title, abstract, accepted (yes/no), and motive. On completion of the extraction, data bases were unified, verifying that both matrixes contained the same information.

Subsequently, two evaluators, HH and EC, carried out the reading of the titles and abstracts (figure 1), considering as inclusion criteria studies that, regardless of the method used, reported prevalence of victimization and/or perpetration of sexual aggression, understood as a subtype of sexual aggression including the use of some actions such as physical force, intimidation or other coercive acts to force men or women to have some kind of sexual encounter, despite his/her lack of will, attacking their sexual freedom through coercion, manipulation and deceit and subtypes thereof in university students from the moment they entered to higher education institutes, including the studies that reported perpetration and/or victimization by partners, former partners, friends, acquaintances or unknown people in university life ^4,7.

Figure 1 Flow chart - Systematic review process 

Victimization of the sexual aggression is understood as the self-report of having suffered any encounter of sexual nature against one's own will, as well as the self-report of having used any mechanism to force another person to have an encounter of sexual nature. Prevalence for this study was conceived as the percentage of victimization and perpetration of the sexual aggression reported by students since they entered in the university. Longitudinal studies that reported victimization and perpetration of sexual aggression in their base line as frequency measurement were included.

Articles that measured prevalence of sexual aggression in the lives of the students that did not discriminated possible victimization or perpetration from the moment of university entrance were excluded, as well as studies on reported sexual aggression only in dating or in dates that did not connected the occurrence to university life of the participants; studies of reports carried out by a third party like friends or colleagues of the victims or of the perpetrators, and reports from secondary studies.

In cases of doubt concerning the initial selection of some articles by the evaluators, these articles were included in the list of accepted for an in-depth review in order to ensure consensus; when the discrepancy persisted the opinion of a third evaluator (PR) was consulted.

The reasons for article exclusion in the selection stage are reported in the Figure below. Articles mentioning victimization/perpetration outside the university and before university life, violence in the workplace, partner violence, violence in dates or domestic violence predominated, all of them psychometric properties that did not provide data on prevalence of victimization/perpetration of sexual aggression, and secondary studies that segmented the sample for new analysis, among others. Then, in the selection stage, a comprehensive reading was undertaken to the remaining articles, setting aside articles related to harassment/stalking; victimization/perpetration in the pre-university time, as well as articles lacking a report on findings of prevalence of sexual aggression, among others.

Those articles eventually included in the review were evaluated independently by two researchers in order to establish the risk of bias according to the STROBE checklist.

Results

The search produced a total of 1238 articles. After passing through the different screening phases, eligibility, selection and inclusion (see figure 1), 35 articles met the established inclusion criteria. A new Excel® matrix was created indicating authors (year), design, sample, measuring instrument, findings / types of sexual aggression and relation perpetrator-victim (table 1).

Table 1 Studies included in the systematic review 

ID = Identification; S. A. = Sexual Aggression; n = Sample; Rel = Relation; P. = Perpetrator; V. = Victim; SES = Sexual Experiences Survey; sabs= Sexually Aggressive Behavior Scale; sav-s = Sexual Aggression and Victimization Scale; N/I = No information; nmupm = Non-Medical Use of Prescription Medicine.

Prevalence of victimization and of perpetration of the sexual aggression identified in the studies ranged between 0.4 %-85 % and 0.6 %-46.2 %, respectively, these reports varied according to the subtype of sexual aggression and to the victim's/perpetrator's gender (table 1).

The most used design in the selected studies was the transversal design (n = 28), followed by longitudinal studies (n = 7). Most analyzed studies assessed victimization (n = 27); followed by studies on victimization and perpetration (n = 6), and only on perpetration (n = 2). Considering the characteristics of the studies' samples, 51.4 % (n = 18) included men and women; 45.7 % (n = 16) only women, and 2.9 % only men.

22.9 % (n = 8) of the studies reported figures on victimization and/or perpetration in general, not specifying subtype nor seriousness of the event. The remaining studies (n = 27) reported that the most frequent aggression type is the unwanted sexual contact, both in persons who have been aggression victims and in those who have perpetrated the aggression. Regarding the relation perpetrator-victim, 48.6 % (n = 17) of the studies concluded that the most frequent perpetrator was the one who was within the most immediate sphere of the victim, being this a friend, partner or acquaintance; the other studies did not report this information (n = 18).

A total of 17 studies used only one measuring tool for sexual aggression: ad hoc questionnaires or an adjustment of questions of instruments of previous studies. The most used instrument (n = 13) was the Sexual Experiences Survey in its different versions ^10,11,15.

Out of the 35 articles included for the review, 5.7 °% (n = 2) were studied with a trans-cultural sample of Chile, Bolivia, Mexico and the US (16, 17). The remaining 33 studies were reviewed with samples from the US (n = 23), Nigeria (n = 4), Portugal (n = 1), Ethiopia (n = 1), Turkey (n = 1), the UK (n = 1), China (n = 1), and Chile (n = 1).

Although the review period covered the last ten years (2008-2018), 45.7 °% of the articles included in this review (n = 16) were published between the years 2014 and 2016.

Discussion

The currently used figures of sexual aggression in the university context range between 85.5 % and 0.4 % for sexual aggression victimization, and between 46.2 % to 0.5 % for perpetration of sexual aggression ^7,18,21,31. Nevertheless, these figures in different studies are confusing since they use diverse classifications of sexual aggression that can be understood as rape or sexual coercion, an example of this is when the presented figure relates to an unwanted sexual relation, or a sexual relation under threat, or unwanted anal sex, among others ^18,31,56.

This can be considered as an estimate of the scope of the problem because there are factors that can influence the prevalence ranges of victimization and perpetration of sexual aggression reported by the studies included in the review. Among these, the assessed definitions and sub-categories of sexual aggression stand out, as well as the proposed research designs (transversal/longitudinal); the representativeness the selected sample could have of the study's target population; the gender of the respondents, and the used measurement instruments. In spite of these differences, and in light of the numbers collected in this review, it is clear that sexual aggression in university students is a frequent phenomenon that affects both women and men.

In all researches, including the present review, the used collection method for data on sexual aggression was the self-report, mainly through digital means, a fact that provided security for the students in terms of confidentiality in their answers. However, the reason why this is the most frequently used method is because in most cases the education institutes have neither instruments nor resources for assisting victims in the institutions, that is why the reports on possible accusations are limited ^12,58.

Although most studies measured the forms of sexual aggression separately (sexual contact, sexual coercion, attempted rape and rape), some researches used for these measurements only one question. For example, the study carried out by Forke, Myers, Catallozzi and Schwarz ²⁹ was aimed at establishing the prevalence of victimization and perpetration of sexual aggression before and during university life, according to the type of victim-perpetrator relation, using a broader definition of sexual aggression and asking the respondent if he/she was "pressured, coerced or forced to have sexual relations with someone", which produced lower figures of prevalence of sexual aggression victimization and perpetration (12 % and 1.6 %, respectively), compared with the study carried out by Geidam, Njoku and Bako ³⁰, in which the figures of sexual aggression and their reported subtypes exceeded 30 percentage points.

Likewise, there is no evidence of consensus on the studied types of sexual aggression, a fact that could indicate an underestimation of the numbers reported by the students, and with it a generalization of the results in relation to the sexual aggression type is not possible; this could in turn be considered a risk of bias in the given results. For example, in the study carried out by Philipps and Smith the aggression types considered where sexual relations against the victim's will, attempted rape and rape, disregarding the subtype unwanted sexual contact ⁴⁷. Something similar happened with the study by Adejimi, Sabageh and Adedokun, where the subcategories "unwanted sexual contact", "unwanted sexual relations", "unwanted oral sex", "unwanted anal sex", and "penetration with finger or object" were included; these categories do not allow an adequate assessment of whether the case was coercion and/or rape ¹⁸.

Another factor to be considered is the design of the studies. For instance, in the study conducted by Flack, Kimble, Campbell, Hopper, with a transversal design, prevalence of any type of sexual aggression was 18.8 % ²⁷. In contrast, in the study of Turchik, Probst, Irvin, Chau and Gidycz, with a longitudinal design, prevalence of any sexual aggression type was 31.5 %, with a difference of over ten percentage points with respect to the transversal study, something that can be related to an underreporting of the event, to social desirability or to recall bias ⁵⁴.

Likewise, in the majority of the studies included in this review, the female gender was the prevailing sample. In some studies the prevalence of victimization of sexual aggression is higher among women than among men, with substantial percentage differences ^3,7,9. However, in other studies, such as the study conducted in Chile by Schuster, Krahe, Ilabaca and Muñoz-Reyes, this difference tends to disappear, thus evidencing that both men and women are victims of sexual aggression and that, depending on its subtype, men reported a higher victimization than women in the last twelve months of university life ⁶ (table 1).

Comparing the present review with previous studies, it was established that prevalence of victimization of sexual aggression and its subtype were similar to the reported in our study. A review conducted in 2016 in the US, for example, synthetized the studies on prevalence of victimization of sexual aggression in US university campuses published between 2000 and 2015, establishing a total of 34 studies, out of which 22 had a transversal design, 11 longitudinal and one was a randomized controlled trial; the sample of most studies was composed of only women (n = 22); the results on prevalence of victimization of sexual aggression varied in each study ⁵⁹. Unwanted contact ranged between 1.8 % and 34 %; reported sexual coercion oscillated between 1.7 % and 32 %; attempted rape varied between 1.1 % and 3.8 %; rape ranged between 0.5 % and 8.4 %, and finally any form of sexual aggression measured in a broad manner as just one type oscillated between 6 % and 44.2 % ⁵⁹.

Another research that showed findings similar to this study was conducted by Krahé, Tomaszewska, Kuyper and Vanwesenbeeck, it documented available evidence on perpetration and victimization of sexual aggression in young people between 12 and 25 years old in 27 countries of the European Union ⁶⁰. Results registered a total of 113 studies, in which Germany (n = 17), followed by Spain (n = 14) and the Netherlands (n = 11), have developed a higher number of studies on this problem ⁶⁰. Data on prevalence of victimization and perpetration of sexual aggression were divided according to gender of the sample: in the case of women, victimization ranged between 0.3 % and 58.9 %, and in men, it oscillated between 0.3 % and 59.5 % for the last year ⁶⁰. In the case of perpetration of sexual aggression, women reported figures ranging between 0.1 % and 48.2 %, and men figures between 0.1 % and 62.2 % ⁶⁰.

Even having these reviews, the study conducted by Fedina, Holmes and Backes considered only studies carried out in the United States; on the other hand, the study of Krahé, Tomaszewska, Kuyper and Vanwesenbeeck considered a broader sample than only the universities that, although it included studies conducted only with university students, does not allow assessing prevalence of victimization and perpetration of sexual aggression in this population specifically during their university life ^59,61.

The present systematic review allows identifying and synthetizing evidence that provides support to the hypothesis of the existence of sexual aggression in the university environment, proving that, regardless of the environment or the culture, this problematic exists and therefore it is necessary to take actions for creating protocols for the prevention of sexual aggression in the university environment against students when they start their university life. At the same time, it allows knowing the limitations of the different studies, such as not generalizing the results since the samples are not representative of the studied population, a fact that implies a selection bias; the designs used (transversal/ longitudinal); the measurement instruments; and the sample specific characteristics.

These findings allow opening a new field of study in order to produce new evidence on the prevalence of victimization and perpetration of sexual aggression against students when they start their university life, around the world and more specifically in Latin America.

Limitations

Although the present review is updated for the last ten years, one of the main limitations this study had was the exclusion of other sources of information, such as grey literature, that would have provided further proof. Other limitations that were evidenced in this review are related to the criterion of the time in which the event took place, for this reason it is possible that some studies do not have prevalence data of sexual aggression from the moment the students started their university life on; therefore it is necessary to specify in future studies if the event takes place along university life or from the moment the university life started. As to the studies including students from their day one in college and that assessed the sexual aggression during the last twelve months, a wrong classification phenomenon took place and, therefore, these reports could relate to situations previous to university life.

Further research could investigate in a greater number of data bases and consult grey literature that provide further information on prevalence of sexual aggression in university students (e.g. national surveys, documents of non-profit organizations, foundations).

Disclaimer

The authors declare no conflicts of interests for publishing this article.