Introduction

A substantial number of bachelor, master and doctoral students experience sexual harassment in academia (Feltes et al., 2012; National Union of Students, 2010; Lipinsky et al., 2022; Rudolfsson et al., 2022). A recent survey conducted in fifteen European countries shows that gender-based violence, including sexual harassment, is common in academia, with more than half of students reporting having experienced some form of gender-based violence, and almost one in three indicating they have experienced sexual harassment within their institution (Lipinsky et al., 2022). In line with previous research (Coulter et al., 2017; Wood et al., 2021), this large-scale study revealed an increased risk among minorities, with non-binary people, LGBTQ + people, ethnic minorities and people with disabilities or chronic illness reporting the highest rates of exposure (Lipinsky et al., 2022).

Cases of alleged sexual harassment in academia have been difficult to track and examine, particularly when no physical sexual aggression or violence is involved (Carstensen, 2016; Schaaf et al., 2019). The concern that most cases remain unreported (Lipinsky et al., 2022; Rudolfsson et al., 2022) may partially reflect obstacles that individuals encounter in evaluating whether a specific experience meets the criteria for sexual harassment. Previous research has in particular highlighted non-physical incidents as being difficult to categorize (Carstensen, 2016; Schaaf et al., 2019). Additionally, these difficulties might be compounded by social and organizational power dynamics (Carstensen, 2016; Coulter et al., 2017; Wood et al., 2021). At many European universities, sexual harassment is increasingly being recognized and relevant procedures established and advertised, often through Student Unions or offices devoted to the prevention of and response to sexual misconduct (for examples, see Dutch Association of Universities, 2022; University of Hamburg, 2022; University of Zagreb, 2022). Although such formal procedures and offices focus on helping students and employees to identify and report sexual harassment, their reach is often less than optimal.

One reason for the limited scope of institutional and other formal influences on students’ understanding of sexual harassment may be related to the fact that most of the cases that have been prosecuted, and sometimes widely publicized (National Union of Students, 2010), tend to involve the most serious examples of sexual misconduct—including those characterized by repeated attempts at physical sexual contact. Finally, previous and ongoing quantitative comparisons among European countries (Feltes et al., 2012; Lipinsky et al., 2022) suggest that there may be culture-specific discrepancies in the way specific experiences are evaluated and categorized.

Aim and Research Questions

The present study contributes to the field by exploring the perceived obstacles faced by students at the bachelor’s, master’s and doctoral levels in identifying potential incidents of sexual harassment. The main questions examined are the following:

  1. 1.

    How do the participants conceptualize sexual harassment, and how well do these conceptualizations align with their lived experiences?

  2. 2.

    What are the obstacles involved in distinguishing sexual harassment from other kinds of behaviour?

The authors would argue that this analysis is crucial to preventing sexual harassment in academia, since the capacity to identify such incidents is a prerequisite for reporting. As such, the results may be used to raise awareness about common pitfalls and challenges associated with the identification of sexual harassment incidents prior to the decision to report. The study’s participants include students at the bachelor’s, master’s and doctoral level. The study is based on individual interviews and focus groups conducted between 2020 and 2022 with a total of 85 students in five European countries (Belgium, Croatia, Germany, the Netherlands and Sweden).

Background

Legal Definitions and Frameworks

The regulation of sexual harassment varies among the participating countries. In most, sexual harassment is covered by legal frameworks that include anti-discrimination laws and provisions, as well as the countries’ penal codes and work-environment-related legislation.

In Croatia, sexual harassment is covered in the Anti-Discrimination Act (Croatian Anti-Discrimination Act. br., 85/2008) and defined as “every unwanted verbal, nonverbal and physical act that aims to violate or is a violation of, the dignity of a person, and which causes fear, or a hostile, humiliating, or offensive environment.” Similarly, anti-discrimination laws and frameworks in Germany, the Netherlands and Sweden also include a prohibition of sexual harassment based on very similar definitions as the Croatian example (German General Equal Treatment Act, 2006; Dutch Equal Treatment Act, 2020; Dutch Civil Code, 1992; Swedish Discrimination Act, SFS 2008:567). However, an important difference from Croatia is that in the other participating countries, sexual harassment must be committed with the intent to discriminate against the victim and must be committed by a representative of an organization or authority. Thus, these frameworks do not apply to incidents of sexual harassment that take place, for example, between students or by a student towards an employee. Instead, in Germany and Sweden, there is work environment legislation that prohibits both physical and verbal forms of sexual harassment. These regulations are intended to promote equal treatment among employees and to prohibit what in the Swedish case is defined as “derogatory and discriminatory mistreatment” (Swedish Work Environment Authority, 2015), and in the case of Germany is exemplified as sexual remarks and ambiguous jokes (German General Equal Treatment Act, 2006). Of relevance to the current study is the fact that these regulations apply to cases of sexual harassment that take place between students at the bachelor’s, master’s and doctoral levels, as well as sexual harassment directed against doctoral students by other staff.

In Belgium, legislation regarding sexual harassment was moved to the penal code in a recent legal reform, and placed among the provisions on violations of sexual integrity. This newly developed framework places a major emphasis on the presence of consent, defining such violations as ‘the performance of a sexual act on a person who does not consent to it’ (Belgian Penal Code, 1867, Art. 417/7). The provisions also emphasize that sexual consent can be withdrawn at any time, and that the presence of intoxication, illness, or disability may impede the capacity to freely provide consent (Belgian Penal Code, 1867, Art.417/5). Similarly, a new section on sexual harassment was included in the German Criminal Code (1998) after a series of sexual attacks on women in Cologne during the 2015/2016 New Year’s celebration (Eddy 2016). In a departure from other anti-discrimination laws, this framework only covers physical acts where the offender “physically touches another person in a sexually explicit manner and thereby harasses that person” (German Criminal Code, 1998).

University Guidelines and Policies

Four of the universities at which students were interviewed in the present study (KU Leuven, University of Amsterdam, University of Utrecht and University of Zagreb) have independent guidelines regarding the conceptualization of sexual harassment. The University of Utrecht has a code of conduct that prohibits sexual harassment between and among staff (Utrecht University, 2022), and at the University of Amsterdam, there is a policy prohibiting behaviour that is harmful to people and the institution, which includes sexual harassment (University of Amsterdam, 2020). The University of Zagreb has guidelines prohibiting sexual harassment, as exemplified further below (University of Zagreb, 2022). The definition employed by this university emphasizes incidents characterized by a lack of consent and incidents that are repeated over time. At KU Leuven, there is a code prohibiting ‘unacceptable sexual behaviour’ in the form of sexual intimidation, sexual abuse and sexual violence (KU Leuven, 2022).

At both the University of Hamburg and Malmö University, the university policies have guidelines prohibiting sexual harassment (Malmö University, 2023; University of Hamburg, 2022), which rely on definitions from the German General Equality Treatment Act (2006) and the Swedish Discrimination Act (SFS 2008:567), respectively. However, neither framework covers sexual harassment between students.

As regards sexual and/or romantic relationships between students and staff, none of the participating universities have policies in place that explicitly prohibit such contacts. However, this does not mean that such relationships are considered unproblematic. For example, most countries have policies that describe how such relationships can impair objectivity and impartiality on the part of the staff, and lead to the sexual exploitation of students as a result of power imbalances (Dutch Association of Universities, 2022; University of Hamburg, 2022; University of Zagreb, 2022). Malmö University in Sweden is the only university which has no code of conduct that problematizes sexual relationships between university staff and students.

Public Context

At the time of this study, the #MeToo movement had already impacted on all participating countries. Although the hashtag was first used on social media in 2006 by women’s activist advocate Tarana Burke to highlight the magnitude of sexual harassment and sexual abuse, its breakthrough was in 2017 following a large case in which several actors reported abuse by the Hollywood film producer Harvey Weinstein (Salmonsson, 2020).

In Sweden, the movement had a profound impact in late 2017. Of particular relevance to academia, an initiative to collect stories was conducted on social media, and a petition comprising 100 anonymized stories and 2648 signatures was subsequently submitted to the Swedish Minister of Higher Education (Salmonsson, 2020). In Germany, the #MeToo movement resulted in a recommendation, in 2018, to the 24th General Assembly of the German Rectors’ Conference, and subsequently the adaptation of the “Guidelines for Good Scientific Practice” by the German Research Foundation (German Research Foundation, 2022; Hochschulrektorenkonferenz, 2018). In Belgium, the movement has been active since 2017, but it took until 2020 for #MeToo-related cases of sexual harassment in academia to receive significant media attention (“#MeToo raast verder”, 2022; Paelinck, 2022; Struys, 2022). In Croatia, the #MeToo movement has been developing since 2018 (Šarić, 2022), but the main impact regarding sexual harassment in academia was felt in 2021, when over 60 cases of sexual harassment in academia were reported (Profesor optužen za seksualno uznemiravanje, 2021). Finally, in the Netherlands, the academic #MeToo initiative received media attention in 2021 following a report on sexual violence at universities (Driessen & Polet, 2021).

In the participating countries, the #MeToo movement brought the issue of sexual harassment in academia to the attention of the public via various media outlets. There have been recurrent media reports of sexual harassment in academia in all participating countries (“Göttinger Professor verurteilt”, 2022; Bahners, 2022), and often at the universities included in the present study. For example, two cases from Malmö University were recently prosecuted in the district court (Herkel, 2022a, b), while two cases at the University of Amsterdam (Het Parool, 2020; RTL Nieuws, 2019) and one case in Croatia received special media attention (“Profesor optužen za seksualno uznemiravanje”, 2021). Similarly, recurrent reports of sexual harassment by a professor at KU Leuven during the period 2010–2019 also received considerable media attention (Struys, 2022).

Previous Research on Sexual Harassment in Croatia, Belgium, Germany, the Netherlands and Sweden

To date, there has been a lack of research on obstacles to identifying potential incidents of sexual harassment among students in academia in the studied countries. Instead, the existing research has primarily taken the form of prevalence studies. These studies point to large variations in the prevalence of self-reported experiences of sexual harassment in student populations. There are no evident differences among those countries in which studies have been conducted; instead, different studies report both high and low rates for all countries, and in many cases, the variability in victimization rates appears to be attributable to methodological factors.

Studies with low victimization rates (e.g., 8–13%) tend to be based on studies using small- to medium-sized samples (approximately 100 to 500 participants) that measure sexual harassment using a single item (Gádyor et al., 2012; Jussen et al., 2019; Tydén et al., 2001). This is in line with previous methodological studies on the operationalization and measurement of sexual victimization (de Graaf & de Haas, 2018). In studies using multiple items, average rates of sexual harassment, even in mixed samples of men and women, tend to be higher (Geldolf et al., 2021; Larsson et al., 2003; Rademakers et al., 2008; Rudolfsson et al., 2022; Schoenfeld et al., 2021), although they vary considerably, with prevalences ranging between 20 and 59%. This variation appears to be largely due to differences across studies in how sexual victimization is operationalized in terms of the behaviours and experiences that are asked about, with some studies asking about lifetime prevalence or not distinguishing between victimization in or outside university settings.

Most prevalence studies point to substantial gender differences in the experience of sexual harassment, with a considerable over-representation of female victims (Gádyor et al., 2012; Geldolf et al., 2021; Larsson et al., 2003; Rudolfsson et al., 2022; Schoenfeld et al., 2021). The two largest studies to date conducted in Germany and Sweden, with samples of over 10,000 participants, found that between 40 and 55% of female students reported experiences of sexual harassment at university (Feltes et al., 2012; Rudolfsson et al., 2022). However, it would be inaccurate to present sexual harassment as a ‘women’s problem’. In most studies, between 12 and 35% of male students report experiences of sexual harassment (Geldolf et al., 2021; Larsson et al., 2003; Rudolfsson et al., 2022; Schoenfeld et al., 2021).

In addition to methodological differences, variability in prevalence estimates may in part also be due to actual differences in the frequency with which sexual harassment occurs. For example, medical training and education are highlighted in several studies as risk factors for students and doctoral students alike. This is believed to be the result of multiple factors, including a sexist working culture, meetings with patients who show a lack of respect for women, and strong hierarchies in which women have traditionally occupied lower-level positions (Gádyor et al., 2012; Jussen et al., 2019; Kisel et al., 2020; Kristoffersson et al., 2016; Larsson et al., 2003; Rademakers et al., 2008; Schoenfeld et al., 2021; Vanden Auweele et al., 2008).

Few of the studies conducted in the participating countries present information about offenders or the consequences of sexual harassment. When such information is included, victims have usually reported that the offender was a male co-student or colleague (Feltes et al., 2012; Geldolf et al., 2021; Rudolfsson et al., 2022). For those studying or working in a medical environment, patients were often highlighted as a significant offender group (Geldolf et al., 2021; Rademakers et al., 2008; Schoenfeld et al., 2021). However, some studies have also shown that here too, gender differences might play a role. For example, Schoenfield et al. (2021) found that the type of offender most often reported by male victims on medical training programs involved colleagues. At the same time, Larsson et al. (2003) found that women in medical training programs primarily reported having been victimized by lecturers/professors or physicians, while the most frequent types of offender reported by men in the same study were co-students or nurses. While media coverage tends to highlight cases in which professors or teachers have targeted students, these cases usually represent only a small proportion of all reported incidents (1–20%; Kury et al., 2004; Rudolfsson et al., 2022; Vanden Auweele et al., 2008). A quantitative Croatian study also pointed to gender differences in perceptions and the impact of sexual harassment (Balenović et al., 2000). For example, female participants favoured more severe punishment for sexual harassment than male participants. In terms of impact, the female participants generally reported feeling anger, fear and rage after being subjected to behaviours categorized as sexual harassment. Male participants, on the other hand, reported that they would feel indifferent or flattered if they were subjected to behaviours that are categorized as sexual harassment (Balenović et al., 2000). These gender differences may, at least in part, be attributed to the impact of heteronormative norms with regards to gender and sexuality, with the idea that men’s social status increases and women’s social status decreases with more sexual contacts (Habarth et al., 2020).

To date, only two qualitative studies have been published that involve one of the participating countries: a focus group study with doctoral students in medicine at a Swedish university (Kristoffersson et al., 2016) and a study that included focus groups with students at a German university (Feltes et al., 2012). Both studies primarily focused on the negative impact of sexual harassment. For example, Feltes et al. (2012) highlighted that experiences of sexual harassment made students feel unsafe on the university campus, while the results of the study by Kristoffersson et al. (2016) showed that women tended to make changes in both their appearance and their behaviour in order to avoid future sexual harassment.

Methods

The involved research groups conducted interviews and focus groups among students at the bachelor’s, master’s and doctoral levels in five European countries: Croatia, Belgium, Germany, the Netherlands and Sweden. The study was initiated by the Swedish research team, who invited colleagues from other European countries, all members of the European Sexual Medicine Network (ESMN), to participate. Across the participating countries, the proportion of the population with tertiary education varies from 24% to around 50%. In all countries, women represent 50–60% of the student population. In Belgium and the Netherlands, most students pay tuition fees for their university education. In Germany, tuition fees have been abolished, but students may still have to pay tuition fees for administrative costs. In Croatia, higher education used to be free, but universities are increasingly introducing fees. In Sweden, higher education is free for all EU citizens.

Participants

Participants were recruited between October 2020 and August 2022. As the Covid pandemic was accompanied by various restrictions, recruitment strategies varied at different times both between and within countries. During lockdown periods, recruitment was primarily carried out via digital means, such as social media, e-mail and digital learning platforms. During periods when campuses were open, recruitment strategies also included posters, flyers and presentations at lectures/meetings/student gatherings. For the same reasons, some interviews were conducted online using chat and video software.

To be eligible for inclusion in the study, students did not need to have personally experienced sexual harassment in academia. Instead, the researchers aimed to obtain rich and nuanced information from different perspectives with regard to both experiences and demographic characteristics.

In Sweden, participants were offered the choice between participating in focus groups or individual interviews. However, all students at the bachelor’s and master’s levels opted for individual interviews, while all doctoral students opted for focus groups. In Belgium, Croatia, the Netherlands and Germany, the research teams only recruited participants for individual interviews.

The sample comprises 85 participants, 24 of whom were bachelor’s students, 36 master’s students, 19 doctoral students and an additional 7 who were preparing for state exams in law or medicine (specific to the German higher education system). In total, 67 participated in individual interviews and 18 in focus group interviews. Of the participants, 58 identified as female, 26 as male and 1 as non-binary. A majority of the interviews (52) were conducted online. A summary of the participants’ study levels and gender identity, form of interview and data collection periods for each country are presented in Table 1.

Table 1 Overview of participants
Full size table

Interview Guide

The interview guide was semi-structured and contained questions about (a) the participant’s background, (b) perceptions about what constitutes sexual harassment, (c) perceptions about sexual harassment in academia, (d) perceptions about vulnerable groups and consequences of sexual harassment, (e) own experiences and (f) prevention of sexual harassment. All interviews were recorded and transcribed verbatim prior to analysis. The interview guide was initially developed by the Swedish research group, and subsequently translated by the research teams in Croatia, Belgium, Germany and the Netherlands.

In all participating countries, a formal legal term equivalent to sexual harassment was used when translating the interview guide. However, it should be noted that interviewers and study participants often also used other terms to denote sexual harassment. For example, in Belgium and the Netherlands, interviewers and participants often used the Dutch term ‘seksueel grensoverschrijdend gedrag’ (roughly translated as ‘sexually transgressive behaviour’). Similarly, Croatian participants used ‘sexual violence’, and German participants used the term ‘violation of sexual boundaries.’

Ethical Considerations

Ethical approval for the study was obtained from the Swedish Ethical Review Authority (DNR 2020–03393), the Research Ethics Committee of the Department of Sociology, Faculty of Humanities and Social Sciences at the University of Zagreb, Croatia (decision number 09–2020/21), the Local Psychological Ethics Committee at the Centre for Psychosocial Medicine in Germany, and the Ethics Committee Research UZ/UK at KU Leuven, Belgium. In the Netherlands, the Medical Research Ethics Committee NedMec informed the researchers that the study was exempt from formal medical approval under Dutch law (reference number 21/360). All participants received an information sheet containing information relevant to the informed consent procedure before participating. Interviewers also informed participants about available country-specific psychological health services. Participants in the Netherlands, Germany and Belgium all signed an informed consent form, while participants in Sweden were allowed to choose if they wanted to sign a consent form or consent verbally on the recording. The information letter was also read to Croatian interviewees before the start of the interview, and before the start of recording, they verbally consented to participate and to be recorded. To protect the identity and respect the privacy of the participants, quotes and details that might identify the participants have been altered or omitted from transcripts when required. In line with ethical regulations and guidelines, all participants have been given a pseudonym in the presentation of the results.

Analysis

The interview and focus group data were analysed using the thematic analysis approach (Braun & Clarke, 2022). This method allows researchers to identify, organize and obtain insights into patterns of meaning (themes) within and across data sets, systematically and actively. The method can be used to explore both shared collective experiences and group differences and individual nuances (Braun & Clarke, 2022).

The analysis for the present study is primarily based on the sections of the interview guide focused on perceptions of what constitutes sexual harassment and perceptions about sexual harassment in academia. Participants reflected on their definitions of and opinions on sexual harassment, often in relation to their own experiences, which ranged from being a victim of sexual harassment, to being a potential perpetrator or a bystander. Throughout the interviews, participants also often shared stories that they had heard from fellow students and friends or through the media. The data were analysed manually using qualitative research analysis software (NVivo or MaxQDA).

Thematic analysis consists of the following phases: familiarization with the data, generation of initial codes, development of candidate themes, reviewing the themes and defining and naming themes before writing the results (Braun & Clarke, 2022). While these phases might appear linear at the theoretical level, the actual process of analysis instead involves an iterative working strategy, whereby researchers develop codes and themes in parallel. For the purpose of the present study, all researchers coding the data attended a workshop in thematic analysis based on the work of Braun and Clarke (2022). Following this training, the researchers conducted an initial inductive coding of their interview material focusing in particular on obstacles to identifying potential incidents of sexual harassment.

Following the first phase of coding, researchers from all countries gathered for a writing retreat, during which they refined their initial codes and developed candidate themes and subthemes. Following the retreat, a second phase of coding took place, using the refined codes, candidate themes and subthemes.

Codes and themes were further reviewed and refined at a number of online follow-up meetings. This combination of meetings and continual dialogue among coders from the different research teams involved an iterative process and ensured that codes and themes matched and represented the findings from all countries.

In our final analysis, the first research question, which focuses on how the participants conceptualize sexual harassment and how well this conceptualization aligns with their lived experiences, has been formulated in the theme ‘Preconceived notions of sexual harassment’. With regard to the second question, which focuses on obstacles to distinguishing sexual harassment from other kinds of behaviour, we have formulated two themes, ‘Challenges to drawing the line’ and ‘Competing interpretations’. Each of these three themes includes 3 to 4 subthemes. An overview of the themes and subthemes by country is presented in Table 2 along with the number of interviews and focus groups each theme and subtheme were present in. The themes are used as headings, and subthemes italicized, in the “Results” section.

Table 2 Summary of themes and subthemes for each country
Full size table

In terms of disciplinary reflexivity (Braun & Clarke, 2022), the research group included researchers from the fields of sexology, sexual medicine, psychology, sociology, social work and criminology. Thus, all had professional and personal experiences of formal and informal sexual norms from their respective university contexts. Moreover, the research group varied in age from early 20 s to late 50 s. In sum, the research group can be described as relatively interdisciplinary and cross-generational.

Results

Preconceived Notions of Sexual Harassment

When asked to define sexual harassment, 83 of the participants indicated having preconceptions about what sexual harassment consists of. Firstly, the participants described a preconceived notion that sexual harassment is a top-down occurrence in the academic hierarchy, exercised by senior staff or teachers against students or doctoral students. This notion was an obstacle for those who had experienced sexual harassment that involved other students or doctoral students. Secondly, all participants endorsed the preconceived notion that sexual harassment tends to be associated with a gender hierarchy, performed by men against women. While consistent with most of the self-reported experiences, some participants also described cases in which women were offenders and/or men were victims. Thirdly, there was a preconceived notion that sexual harassment is strongly associated with age differences, with the offender being markedly older than the victim. This sub-theme was contrasted with the experience of sexual harassment between students, who are of roughly the same age, or instances in which doctoral students are sexually harassed by younger students in teaching settings. Finally, participants described a preconceived notion of sexual harassment as involving physical incidents that contain some form of physical sexual violence, while many had lived experiences of verbal harassment.

In the interviews and focus groups, these preconceived notions were often mentioned in connection with each other. The following examples illustrate how some participants constructed connections between the academic hierarchy, gender and age:

“I think that it [sexual harassment] exists at all levels of society, that somehow a person in a higher position, usually a man, abuses a female person in a lower position. So, female students and doctoral students are usually abused by male professors”. (Petar, male Croatian bachelor student)

“When I think about sexual harassment in academia, my first thought goes to younger females or doctoral students and older male senior professors. There are of course a lot more situations that can occur, but at the top of my mind, that is what I think about when I think of sexual harassment in academia”. (Kerstin, female Swedish doctoral student)

Some also discussed their own preconceived notions or described a process whereby they had gradually developed a more nuanced view of what constitutes sexual harassment, as in the following example from Germany:

Interviewer: You’ve said quite often “I have such stereotypical assumptions about it”. Can you maybe elaborate on that?

Interviewee: So, whether you have that from movies and stuff, because it’s also always portrayed in a movie that some man sexually harasses some woman. So, I don’t have any other examples of that at all. I’ve never heard of any other case of a woman sexually harassing a man, something like that. I’ve only ever heard of men who have somehow sexually harassed women. And somehow you only see it in the movies, that’s why it’s so stereotypical for me. (Lisa, female German bachelor student)

Recurrent in the first three sub-themes are broader notions of how sexual harassment is linked to structural power dynamics with regard to academic title, age and gender. At the same time, many described experiences that diverged from these preconceived notions. These primarily involved incidents with male victims, student-to-student harassment, student-to-staff harassment and non-physical incidents. In such cases, participants often described a decreased tendency to report. For example, one male participant who described having experienced unwanted sexual comments and sexual touching from women said:

“If a man gets sexually harassed then, well he might feel more reluctant to make a thing out of it than a woman would. I believe that women get encouraged to say no and take it to a legal instance or something. We [men] might be more prone to just saying ‘no thank you’ and just trying to forget the shit”. (Edvin, male Swedish master student)

As such, the theme expresses a gap between some of the participants’ lived experiences and their preconceived notions of what sexual harassment should look like. While there were many experiences that fit the participants’ preconceived notions about sexual harassment, as will be presented below, those who encountered incidents that diverged from preconceived notions described needing to revise their initial beliefs in order to be able to “see” other types of incidents of sexual harassment.

Challenges to Drawing the Line Between Sexual Harassment and Acceptable Behaviour

In the interviews and focus groups, participants repeatedly expressed facing challenges in drawing the line between acceptable social interactions and sexual harassment. This was a theme described by 81 of the participants. One of the first challenges the participants faced was connected to the degree of explicitness in each incident. There were often uncertainties around acts that involved either fleeting physical contact or gestures that might not be considered sexual by default, such as being looked at in specific ways. One participant said:

“So, this of course, in terms of comments, you have a grey zone of things that are not considered sexual harassment by some, while they are by others. What to do then is yet another question. I think there is room for discussion, where a very clear category is [harassment], a very clear category is not [harassment], and then [there is] a very large grey zone where it is difficult to know what it is”. (Thomas, male Belgian doctoral student)

One aspect that stood out was that the male participants often felt that a higher degree of explicitness was required for their experiences to be ‘counted’ as sexual harassment. Similarly, the setting in which an incident of sexual harassment occurs has a bearing on how easy or hard it is to draw the line. Here, it is of relevance whether or not the setting allows for displays of sexual interest or sexual comments. Broadly speaking, university facilities such as classroom lectures or seminars were largely perceived as nonsexual. In contrast, it becomes more complicated in informal university settings where sexual invitations are regarded as a potential part of the overall normative setting. Examples included student dorms, student bars or clubs and conferences and other more informal arenas for academic networking, as expressed by the following student from the Netherlands:

“Yes, if you are in a bar or something, that someone is touching you, that is actually quite normal. [...] Yeah, if you just went to the [Club], it happened. And yes, at house parties too. That I more often had to say no repeatedly, or walk away, or send someone away. And then indeed: Oh sorry, I was drunk”. (Sanne, female Dutch master student)

Interestingly, participants who were involved in courses that involved practical training outside the university often highlighted these as environments in which sexual harassment was frequent. For example, a medical student described it the following way:

“I think (about) this strong hierarchy in hospitals, the pressure that people are exposed to, and simply the very intensive cooperation, always in a team. That maybe people don’t think long enough before they say something, or that the power imbalance simply means that if a senior doctor is sexist towards me, I don’t dare say anything because I don’t know what the consequences will be for me. So, I talk about it with friends, but not with him”. (Marie, female German state exam student)

In these contexts, it was more common for the participants to talk about different forms of victim blaming, or their own fear of being blamed for having been sexually harassed. They also talked about how a broader societal normalization of sexual harassment impeded their ability to identify potential incidents of sexual harassment.

Aside from these more definitional types of challenges focused on where to draw the line, participants also described a cognitive challenge regarding evaluating something as sexual harassment—which was reflected in the time they needed to process and interpret their experiences. Since there were many layers of uncertainty involved in determining how to conceptualize or characterize a certain incident, many needed time to think and emotionally reflect on their experiences. For an incident to be perceived as sexual harassment, it needed both to fit an abstract concept and feel like an intrusion or violation of sexual integrity. Both require cognitive processing on the part of those who had been subjected to or witnessed potential incidents of sexual harassment:

“When I looked back at it a bit later, or now, or when I discussed it with my psychologist, I realised it was not okay. Then I do experience some general anger that people do that to other people. I am someone who tends to look at myself first and who only realizes in conversation with someone else that in the end it is not all my fault”. (Chloe, female Belgian bachelor student)

Since many participants with direct experiences of sexual harassment reported difficulties in identifying and responding to sexual harassment as it happens, the results point at the relevance of bystanders’ readiness to intervene as an important aspect of responding to instances of sexual harassment.

Competing Interpretations

In total, 80 of the participants described having to navigate between competing interpretations of sexual harassment and the challenges associated with this. One of these challenges was linked to conceptual confusion. Consistent with the many variations in how sexual harassment is defined and regulated legally, in society, and in university policy, participants described that they sometimes faced conceptual confusion when trying to match their lived experiences with the various concepts and definitions by which they were surrounded. Mostly, the competing interpretations that caused confusion were described as lay concepts and labels expressed in forms such as jokes, laypersons’ jargon and macho-culture.

“Well, I think a lot of such inappropriate remarks are meant as jokes. Then, when in doubt, the person who says it laughs too, and nobody else finds it funny. Or not everyone finds it funny. That would perhaps be the worse situation. Yes. Or like to hear yourself talk. I don’t know”. (Vanessa, female German master student)

Another aspect that led to competing interpretations involved incidents of sexual harassment that centred around the misuse of authority. For example, students talked about how some teachers would normalize sexist behaviours, for example, by commenting on how female students look.

“I think as soon as there is some kind of position of power that this forms a good basis for sexual harassment. I think we certainly shouldn’t be blind to that, and very often things are also seen as innocent, but when you look back at it you realise, ‘actually it wasn’t okay that that happened’”. (Amber, female Belgian master student)

Finally, there may be a disparity between the intention and the outcome of a social interaction. The participants described cases of competing stories in which one party might refer to an incident as sexual harassment, while the other party referred to it as normal flirting or a non-sexual interaction. The disparity between intention and impact was often linked to gender norms and the social scripting of acceptable behaviour for men and women:

“I think that it is actually this fine line that men often cross. The line between what is and is not acceptable, and usually what is ignored is the feeling - how a woman feels about it. They [men] can perceive it as something friendly [...], but if a woman does not perceive it the same way, it automatically creates discomfort in her”. (Lana, female Croatian master student)

“They [men] don’t experience that their one comment is not the only comment that a woman gets, so their piece of behaviour is not something in itself, it coincides [with more]. And, uh, that means they exhibit transgressive behaviour, and they are no longer aware of the impact of their behaviour. And the fault of this lies in, well, that lack of empathy”. (Bas, male Dutch master student)

The gap between intention and outcome was also associated with generational or cultural differences. For example, some described that problematic behaviour is often characteristic of senior faculty members, the generation influenced by the sexual liberation movement, whose sexual etiquette clashes with that of younger generations of students and doctoral students:

“I also think that this is a fine line, that’s why I think that… Yeah, the age difference can be a thing. If two 19-year-olds have this new or, or other culture in their life, rather than a 60-year-old”. (Elise, female Swedish bachelor student)

Discussion

Based on the findings of this joint research project, a host of obstacles to identifying sexual harassment reported by the interviewees can be assigned to one of three themes: (1) preconceived notions about what constitutes sexual harassment that do not necessarily concur with lived experiences, (2) navigating an often blurred or ambiguous line between acceptable and unacceptable behaviour and (3) the existence of competing interpretations about what actually happened. Based on the participants’ accounts, no single theme or subtheme occupied a central or a superordinate position in relation to the others. Although legal and academic definitions of sexual harassment may not always overlap perfectly, most of the obstacles that students encounter when faced with possible incidents of sexual harassment do not seem to be generated by such normative discrepancies. In daily life, most obstacles appear to be embedded in the complex and nuanced realities of social relationships and interactions, and are part of often messy and ambiguous forms of communication between individuals who do not necessarily share a common understanding of what constitutes (un)acceptable behaviour.

In relation to the study’s first research question, the identified obstacles included prevailing impressions that sexual harassment represents a consistent pattern of power dynamics (based on gender, age, or academic hierarchy) or that it must include some form of physical contact. Although often aware that this involved a stereotyped view, our interviewees’ cognitive representations of sexual harassment primarily pointed to female students as the victims and elderly male professors as the perpetrators. One practical contradiction of such a view is that sexual harassment seems to be far more prevalent among students or individuals of similar age and status (Kury et al., 2004; Rudolfsson et al., 2022; Vanden Auweele et al., 2008). And while it is indeed less common for women to be perpetrators or men to be victims, this normative perspective may create uncertainties in relation to the identification of such incidents as sexual harassment. The prevailing media presentation of sexual harassment in academia largely supports the existing stereotypes, which may serve to force academic institutions to implement rules and regulations focused on preventing status- and age-based sexual harassment, but which may not help in reducing the prevalence of cases of sexual harassment between students.

Similarly, the prevailing student and media accounts of sexual harassment in academia tend to revolve around physical aggression or transgressions that involve unwanted touching (see for example Bahners, 2022; Herkel 2022a, b; Het Parool, 2020; RTL Nieuws, 2019; “Profesor optužen za seksualno uznemiravanje”, 2021; Struys, 2022). At the very least, the common view of sexual harassment seems to be built upon the notion of systematic and repeated verbal sexual abuse. Such an emphasis on the most severe, and possibly most distressing, forms of sexual harassment entails several risks. For one, the stereotypical confounding of sexual harassment with physical aggression or persistent or repeated transgressions may result in the trivialization of perhaps less “severe” but more common forms of sexual harassment. This might also dissuade victims from reporting and responding to these forms of sexual harassment and reduce the likelihood that perpetrators will reflect on their behaviour.

With regard to the second research question, study participants frequently described challenges in drawing a clear line between sexual harassment and behaviours that fall outside the scope of sexual harassment, regardless of country. Apart from the most extreme and least common cases, interviewees often struggled to distinguish acceptable from unacceptable behaviour. The demarcation line was often dependent on context and obscured by multiple distinct or less distinct characteristics of the situation (e.g., the level of explicitness, the location of the incident, relationship history, etc.). This was further complicated by the potential ramifications of calling something sexual harassment, which led to reinterpretations and associated doubts and suspicion, especially in situations in which behaviour that potentially constituted sexual harassment was almost considered a ‘normal’ part of academic culture (e.g., in student bars and clubs). Many participants therefore needed time to process what had actually happened.

It should be noted that the uncertainties and difficulties associated with identifying sexual harassment, as observed in this study, were reported in the aftermath of the #MeToo movement and, in some countries, following several high-profile media reports about sexual abuse and harassment in academia. Thus, it appears that public discussions spurred by the #MeToo movement and local cases of academic sexual harassment have failed to substantially reduce the grey area of interpretation that surrounds sexual harassment, at least in our European student sample.

Finally, different interpretations of specific incidents were frequently mentioned during the interviews. These might be conflicting accounts or recollections about a specific situation on the part of a victim, a perpetrator, or individuals who had witnessed certain incidents. Unlike the second theme, which deals primarily with the individuals’ own uncertainties in defining a situation or a behaviour as sexual harassment, this last theme reflects the existence of external challenges to the recognition of sexual harassment. Competing perspectives not only complicated the inner process of distinguishing between acceptable and unacceptable behaviour as a result of direct expressions of disagreement from other participants involved in an event, but also had broader implications for defining, identifying and sanctioning sexual harassment. These contrasting perspectives were sometimes explained as the consequence of generational differences (e.g., professors not being in touch with, or accepting of, contemporary norms of social interaction) or gender-specific meanings associated with potential transgressions. When the gender of the involved sides was reversed (i.e., the victim male, the perpetrator female), the situation was substantially less likely to be interpreted as sexual harassment. It was also more difficult to characterize an incident as sexual harassment if it did not clearly involve a sexual intention—as judged by the individuals experiencing or witnessing the incident—or an intention to harass. This included situations that others in the participant’s social circle had interpreted as ‘innocent jokes’. To label such incidents as sexual harassment, rather than accepting them as ‘innocent jokes’, involves a risk of being labelled as someone who has no sense of humour or who is overly sensitive to benign personal remarks, which made it more difficult for participants who mentioned these situations to openly call them sexual harassment. The realization or recognition that power and social dynamics may transform a sexist remark into an ‘innocent joke’ can be a lonely, difficult and confusion-prone process, especially when those around you, who were not personally addressed by the joke, may not share this understanding.

There were a number of cross-cultural differences and similarities in the results that warrant some discussion. With regard to preconceived notions, none of the participants from Croatia or the Netherlands reported lived experiences of student-to-staff incidents, although the theoretical possibility of such situations occurring was mentioned in a few of the Croatian interviews. Moreover, none of the Croatian participants reported experiences of female offenders or male victims, which were mentioned in some of the interviews in Belgium, Germany, the Netherlands and Sweden. Furthermore, none of the participants in Belgium or the Netherlands challenged their beliefs that victims are usually younger than offenders, which is not consistent with the most common cases of student-to-student sexual harassment, in which victims and perpetrators are more often than not of a similar age. Additionally, with regards to the misuse of authority, participants from Croatia, Belgium, Germany and Sweden described having been bystanders to or hearing about cases of sexual relationships between staff and students in which sexual intimacy had been exchanged for improved grades. However, no such examples were given in the Dutch sample.

It is also important to note that male participants tended to apply higher thresholds with regard to what counts as sexual harassment than female participants. For example, male participants with experience of explicit sexual invitations, despite explicitly having said no, and/or who reported experiences of being touched in a sexual way, did not count such experiences as sexual harassment. In contrast, female participants with similar experiences did refer to them as sexual harassment. It is possible that this is related to gender differences in heteronormativity, as male students have been found to report more heteronormative values than female students (Habarth et al., 2020). As such, men might be less prone to self-advocate their physical and sexual integrity, which in turn can lead to negative social and health impacts.

There were no discernible differences in the obstacles that participants faced in identifying sexual harassment by the participants’ level of education. The main difference between bachelor’s, master’s and doctoral students was that students at master’s and doctoral levels tended to consider more types of behaviour as problematic and to be more analytical when reflecting over their own experiences. However, it is worth noting that this might also be the result of differences in interview format between students and doctoral students. It may be that the focus group format, rather than level of education, facilitated analytical reflection among participants in a way that individual interviews did not.

Policy Implications

As the three themes illustrate, the process of identifying and understanding sexual harassment in academia is complex, prone to cause confusion and self-doubt and susceptible to a host of challenges. To some extent, many of these challenges seem to be related to or a consequence of the complexities and nuances that are associated with, if not inherent to, social behaviour and interaction, which sometimes lead to miscommunication, misinterpretations of signals and misperceptions of intentions. These may be even more marked in hierarchical settings in which power dynamics may impact (even unintendedly) the meaning of what is said or the interpretations of certain behaviours.

As a direct or indirect effect of the #MeToo movement, which continues to impact both society in general and academia in particular, many universities and educational experts are directing resources towards the development of policies focused on tackling sexual harassment and abuse in academia. Efforts range from establishing clear and restrictive rules of conduct—particularly between students and teachers—to fostering a closer collaboration among university administrations, the police and judicial systems, setting up new academic offices and services for victims of sexual harassment and prioritizing prevention in the form of educating students and faculty about sexual harassment (Dutch Association of Universities, 2022; Malmö University, 2023; University of Hamburg, 2022; KU Leuven, 2022; University of Utrecht, 2022; University of Zagreb, 2022). However, this study’s findings suggest that these efforts, however important and in some cases long overdue, may not be enough.

Most of the challenges to identifying potential incidents of sexual harassment, as reflected in our interviews, seem to stem from a combination of stereotypes about sexual harassment, intricacies of internal processing and the existence of competing interpretations. While most of the educational, institutional and legal measures that have been adopted will hopefully improve the understanding of what constitutes sexual harassment in academia—and may facilitate more realistic perspectives on ‘typical’ instances of sexual harassment—and encourage its reporting, additional steps may be required to ensure that they have an impact in practice. Measures to prevent sexual harassment may still not pay sufficient attention to the nuances of social interaction in age-heterogeneous hierarchical settings. Thus, cognitive gaps created by often complex interpersonal dynamics, and other sociocultural obstacles that victims and bystanders experience when faced with potential sexual harassment in academia, are bound to produce dilemmas in the future.

To start filling the cognitive gaps that affect the way sexual harassment is perceived and interpreted, which appears to be of central importance to the prevention, reporting and sanctioning of sexual harassment in academia, efforts to reduce sexual harassment may need to adopt a more practical, case-based approach. When informed about the differences between beliefs and actual patterns of victimization, cases can be useful in addressing cross-cultural differences with regards to the relationships where sexual harassment occur, such as beliefs that the offender is always older than the victim, as well as gendered aspects of victimization, such as the presence of male victims, that was identified in the present study. Instead of focusing on the application and implementation of formal rules and institutional procedures, more attention should perhaps be paid to training based on the potential ambiguities, intended and unintended, that characterize more common types of sexual harassment. Such training would need to focus not only on problematic behaviours but also on the role of context and the possible impact of differing interpretations among bystanders, such as the gender, age, academic position, social status, race, and ethnicity of both victim and offender. Prevention education for both students and faculty should aim to improve their understanding of a range of different perceptions and interpretations of the same behaviour, based on principles of consent, respect, an awareness of consequences and a sense of safety and security.

Limitations

This study is one of the first to use qualitative methods to explore the crucial but often overlooked perceived uncertainties and challenges associated with identifying sexual harassment. One of the study’s strong points is its cross-cultural design, which has enabled us to conduct 85 interviews and focus groups in five different EU countries and to explore, in a preliminary manner, cross-cultural differences and similarities in experiences and perceptions of sexual harassment. However, some limitations also must be considered. First, although the total sample size of this qualitative study is considerable, the number of participants was relatively small in some of the participating countries. In addition, most researchers recruited students at their own universities, which limits the transferability of findings to other universities within the countries in question. Second, as a result of Covid pandemic restrictions, many interviews were conducted online. While this seemed to work quite well in general, some of the interview techniques that allow for the development of more depth in the data collected (e.g., picking up non-verbal cues, body language) are limited. Third, the cross-cultural component also produces some limitations. Although efforts were made to harmonize the methodology used in the different countries via frequent meetings and close collaboration, differences cannot be ruled out. These include the possible impact of conducting interviews in different languages, with different words being available and used to refer to ‘sexual harassment’ in the different countries. The translations of sexual harassment have slightly different legal meanings in the collaborating countries, irrespective of cross-cultural differences in their interpretation. Also, in some countries, a label other than ‘sexual harassment’ was used by participants, and by interviewers, especially when this label is more commonly used in the country in question. As a result, we have stayed as close as possible to the terminology used in each country, which introduces more variability and potentially makes the results less comparable across countries. Fourth, the participating countries were experiencing different stages of the #MeToo process, which may have impacted the participants’ understanding of sexual harassment. Fifth, only two countries included doctoral students. As a result, the doctoral perspective may be somewhat skewed towards Swedish perceptions, although five of the doctoral student participants were international students. Finally, while the first theme focused on structural power aspects such as academic seniority, gender and age, there are other power aspects relevant to understanding the vulnerability for or risk of sexual harassment, including being LGBTQ + and variables related to class, race, ethnicity and/or nationality, disability and neurotypicality (Coulter et al., 2017; Lipinsky et al., 2022; Wood et al., 2021). As these were framed as risk factors by the participants rather than something that comprises an obstacle for their ability to identify sexual harassment, we chose not to explore these as part of this study. However, since these are structural aspects that impact both victimization risk and individual ability to report and seek remedy, there is a need for further research exploring these connections (Coulter et al., 2017; Lipinsky et al., 2022; Wood et al., 2021).

Conclusion

The results of this study reveal that there are significant obstacles and challenges associated with identifying potential incidents of sexual harassment. These relate to characteristics of the incident, its context and where it takes place, the relationship between perpetrator and victim and also the power dynamics between them. While there may be a perception that sexual harassment is generally easy to identify, the results indicate that there are many obstacles associated with identifying different kinds of behaviour as sexual harassment. As such, the results point to a gap between on the one hand lived experiences and on the other individuals’ ability to match their experiences to theoretical understandings and legal or other definitions of sexual harassment. Measures to counteract and reduce the obstacles and challenges that victims and bystanders face in identifying sexual harassment in academia should include an awareness of such cognitive gaps and should address ambiguities and other variables and processes that may impede or otherwise negatively impact the identification and recognition of sexual harassment. In addition, such measures should address the many stereotypes and preconceived notions that exist about what constitutes sexual harassment, should combat sexism and should provide strategies for discerning whether an incident might fall within the definition of sexual harassment based on university policies and applicable legislation.