1. Introduction
Sexual harassment (SH) continues to be a significant problem at many workplaces, even though the general awareness of the problem increased after the #metoo movement in 2017 [
1]. Other forms of mistreatment also exist at workplaces and share some features in common with SH [
2]. As previous studies have mainly examined SH and other types of workplace harassment separately, research gaps exist concerning the relationship between various types of harassment, including SH [
3]. In this study we investigated whether experiences of other types of
harassment, defined as behaviours associated with one of the seven Swedish legal grounds for discrimination: sex, transgender identity or expression, ethnicity, religion or other belief, disability, sexual orientation, or age, and
derogatory treatment, defined as insulting acts such as withholding information, derogatory comments, or exclusion, were associated with SH at a large Swedish university. The overarching term for all three types of undesirable behaviours (harassment, derogatory treatment, and SH) that we use in this paper is workplace mistreatment.
SH leads to multiple negative health- and work-related outcomes [
3,
4,
5]. Individual and job-related consequences reported include anxiety, depression, decreased job satisfaction, absenteeism, and reduced productivity, not only affecting the individual but also causing considerable costs for organisations [
6,
7,
8,
9]. Negative consequences can be seen even in relatively mild cases of SH, such as sexist put-downs or offensive sexual remarks [
10,
11]. Typically, men perpetrate and women experience SH, although the reverse situation exists, as well as same-gender SH [
12]. Reported levels of SH vary widely between studies and between countries due to factors such as differing definitions used, recall periods and representativeness of the research, as well as variation in cultural values and norms in different societies [
4,
13]. It has, for example, been observed that EU countries that rank high in terms of gender equality, such as Sweden, also tend to report higher prevalence rates of SH [
14]. This could partly be due to greater awareness of SH in the society and of it being a fundamental rights abuse, leading to a greater willingness to report experiences of SH in a survey interview. A recent EU-wide survey on violence against women showed that every second woman (55%) in the EU has experienced sexual harassment at least once since the age of 15, among which 32% report that the perpetrator was someone from the workplace such as a colleague, a boss or a customer [
14]. Not only are women more often subjected to SH than men, research findings suggest that they are also more negatively impacted by it [
5,
15,
16]. For example, in one study, perceptions of workplace SH were associated with poor physical health among women, but not men [
17].
Besides SH, other types of mistreatment such as discrimination, general harassment, bullying, incivility or microaggressions also exist at workplaces and impact negatively on health- and work-related outcomes [
18,
19,
20,
21,
22,
23,
24]. As in the case of SH, the variety in definitions and terms used to capture these types of victimisation complicates comparisons across studies. In the European Working Conditions Survey (EWCS) from 2015, 12% of the participants reported verbal abuse and 6% reported humiliating behaviour in the month prior to the study [
25]. Further, 5% reported bullying/harassment and 7% reported discrimination on the basis of gender, age, race, religion, nationality, disability, or sexual orientation during the 12 months prior to the study. The above-mentioned adverse social behaviours were experienced by women to a greater extent than by men [
25].
There are different theories explaining why SH occurs [
3,
6]. A common understanding of SH is that it is one of the many expressions of gender-based violence [
26]. This perspective regards sexual harassment as part of a continuum of violent behaviours and attitudes against women on the basis of gender, rooted in gender inequality, the abuse of power and harmful norms [
26,
27]. It has been suggested that the primary motive underlying all forms of harassment is the desire to protect or enhance social status when it seems threatened [
28]. According to this view, SH would be motivated by a desire to protect or enhance a gender-based social status when it seems threatened, rather than by sexual desire. The defence of power in the form of identity or position has also been identified to be a shared commonality between SH and general harassment [
2]. According to McDonald [
3], “behaviours such as workplace bullying, mobbing, racial harassment and sex-based harassment, as well as SH, have hierarchical power relations at their core” (p. 12). Relevant in this context is the academic workplace setting, often filled with formal and informal power relations, which may influence the prevalence of SH and other types of mistreatment.
Besides the defence of power, McDonald points at other shared features of these workplace phenomena, such as the perception of the victim, the ambiguity of intent and the violation of organisational norms [
3]. It has been reported that women rarely experience SH in isolation, but in combination with general non-sexualised mistreatment or incivility [
29]. SH and gender discrimination are so closely linked that existing research often merges the two [
17]. Further, discrimination or harassment based on social status might be difficult to disentangle, as identities and social status based on gender, class, race, age, disability, or sexuality are not always easily separable [
17]. Clearly, there is an overlap between SH and various forms of other mistreatment at the workplace, yet few studies have examined these phenomena together [
2]. Increased knowledge about how SH relates to other forms of workplace mistreatment would give valuable insights as to how SH can be understood in the context of multiple destructive workplace behaviours undermining equal opportunities for all employees [
3]. This kind of insight can help improve preventive measures.
In the following section, we describe SH, harassment, and derogatory treatment according to Swedish legislation, which provides the definitions relevant to our study. In Sweden SH is mainly treated as a form of discrimination and falls under the Discrimination Act, defined as a conduct of a sexual nature that violates a person’s dignity [
30]. SH can also constitute criminal acts, such as unlawful invasion of privacy or sexual molestation and is then regulated in the Swedish Penal Code. Other types of harassment, defined as situations when a person’s dignity is violated, must be related to one of the seven legal grounds for discrimination, i.e., “sex, transgender identity or expression, ethnicity, religion or other belief, disability, sexual orientation, or age” to fall under the Discrimination Act. Harassment or SH are behaviours that are unwelcome, and it is the victim of harassment who determines what is unwelcome or offensive. This means that there are no specific behaviours that define harassment or SH mentioned in the Discrimination Act, but rather it is the victim’s perception of a behaviour that is of importance. The perpetrator, however, must understand that the behaviour is unwelcome for it to be classified as harassment or SH according to the law [
31].
Another form of mistreatment at the workplace regulated in Swedish law is defined as offensive or abusive actions towards one or more employees that may lead to ill health or to exclusion from the workplace community [
32]. In this study we refer to this phenomenon as derogatory treatment, which includes experiences of insulting acts such as withholding information, derogatory comments, or exclusion. Swedish employers are obliged to pursue active prevention work against SH, harassment, and derogatory treatment at the workplace. However, undesirable behaviours continue to be a problem at Swedish workplaces and preventive measures need to be improved [
33,
34,
35,
36,
37].
In summary, SH and various kinds of mistreatment exist at workplaces today with negative health- and job-related impacts on individuals and organisations. SH at the workplace seems to overlap with other forms of workplace mistreatment, and a better understanding of how these are related would be valuable when designing preventive measures. Workplace mistreatment is often rooted in hierarchical power relations. As the academic workplace is often characterised by formal and informal power relations, SH and other forms of mistreatment among the staff in this setting are of particular interest. Hence, the aim of this study was to examine whether experience of different forms of harassment and derogatory treatment was associated with SH among employees at a large public university in Sweden. Furthermore, a secondary aim was to investigate whether such associations differ by gender.
2. Methods
2.1. Data Collection
This study is part of the project “Tellus”, initiated at Lund University (LU), Sweden, 2018, with the aim of reinforcing preventive work against SH at the university [
38]. As part of the Tellus project, a cross-sectional survey targeting both staff/PhD students and students was conducted at LU in 2019. The survey included questions related to SH, but also harassment related to the seven Swedish legal grounds for discrimination (sex, transgender identity or expression, ethnicity, religion or other belief, disability, sexual orientation or age) and derogatory treatment. Questions about the gender of the perpetrator and the relationship between the victim and the perpetrator were also included. The survey instrument used, as well as the overall results, such as prevalence of SH while employed at LU, are presented in detail elsewhere [
39].
2.2. Study Population
In November 2019 a survey in both English and Swedish was sent by email to all staff at LU. PhD students are primarily employed by the university in Sweden and were therefore included as staff in our survey. The response rate was 33% (N = 2750). After exclusion of those with missing data on both sex and gender, those who did not answer any of the 10 questions on experiences of SH or those who did not specify when the SH took place, the final study population consisted of 2732 staff/PhD students.
A comparison between the participants and the target population by means of employment records found no major differences. However, a few observations were made; women were slightly over-represented among study participants and both men and women participants tended to be somewhat older compared to the target group. Further, a slight over-representation of staff with permanent employment (versus temporary) was seen [
39].
2.3. Outcome Variable: Sexual Harassment and Sexual Violence
As the outcome variable, we used experiences of SH during the past 12 months at LU. The variable was based on the following list of ten behaviours, including one behaviour representing sexual violence, to measure experiences of sexual harassment or sexual violence: unwelcome suggestive looks or gestures; unwelcome soliciting or pressuring for ‘dates’; unwelcome ‘inadvertent’ brushing or touching; unwelcome bodily contact such as grabbing or fondling; unwelcome gifts; unwelcome comments; unwelcome contact by post or telephone; unwelcome contact online, for example, by social media or email; stalking; and attempts to conduct or the conduct of oral, vaginal, or anal sex or other equivalent sexual activity in which one did not participate voluntarily [
39]. The study participants were asked if they had experienced any of these behaviours in connection with their employment at LU with the answer options: Yes, once; Yes, more than once; and No. All participants who answered yes to at least one of the above behaviours and stated that the behaviour/s occurred during the past 12 months were classified as exposed to SH during the past 12 months at LU.
2.4. Exposure Variables
2.4.1. Harassment
The following text introduced the section about harassment in the survey: “The simplified definition of harassment provided in the Discrimination Act is that harassment occurs when someone is subjected to an act that violates their dignity and that this violation is associated with one of the seven grounds for discrimination: sex, transgender identity or expression, ethnicity, religion or other belief, disability, sexual orientation or age. Harassment can be both individual and isolated events as well as subtle, almost imperceptible events that continue over time, so-called microaggressions. It can also be a process that is ongoing and permeates the entire working life.” The participants were asked if they had experienced harassment as described above in connection with their work at LU during the past 12 months. The answer options were: No; Yes, once; Yes, more than once; and Yes, in the form of microaggressions or ongoing process. It was possible to select more than one answer to indicate whether the harassment was in the form of isolated acts, microaggressions or both. All participants answering Yes, once, Yes, more than once, or Yes, in the form of microaggressions or a process that is ongoing, were categorised as exposed to harassment and all others as not exposed. To obtain information about the type of harassment that was experienced, the participant was asked to select, from a list of the seven legal grounds for discrimination, the particular grounds to which the harassment could be attributed. It was possible to select several grounds.
The Swedish language uses the same word for both sex and gender. The official translation in the Discrimination Act is “sex”. Unless directly citing the Discrimination Act, we use “gender” in the subsequent sections of this paper.
2.4.2. Derogatory Treatment
In the survey, the following information introduced the section with questions related to derogatory treatment: “This refers to derogatory or insulting acts directed at one or more employees. Examples of such acts include withholding information, derogatory comments and exclusion. The Swedish Work Environment Authority includes other examples such as the use of derogatory nicknames, shutting out, exclusion from meetings, unfair accusations, public personal attacks, and referring to someone in offensive terms in front of others. “The question was then put: Have you experienced derogatory treatment in conjunction with your work at Lund University during the past 12 months?” The answer options were: No; Yes, once; and Yes, more than once. All participants answering Yes, once or Yes, more than once were classified as exposed to derogatory treatment and all others as not exposed.
2.4.3. Multiple Forms of Harassment or Derogatory Treatment
Each study participant could report experiences of several forms of harassment associated with the legal grounds for discrimination (sex, transgender identity or expression, ethnicity, religion or other belief, disability, sexual orientation, or age) as well as derogatory treatment. To explore experiences of multiple forms of harassment or derogatory treatment, we created a new variable summarising all forms reported by each individual, including experiences of derogatory treatment, and thereafter categorised all participants into three categories: Not exposed to any form of harassment or derogatory treatment; Exposed to one form of harassment or derogatory treatment; and Exposed to two or more forms of harassment or derogatory treatment. One participant had reported experiences of harassment once and two other participants had reported experiences of microaggressions but not linked to any specific form of harassment. These three participants were classified as exposed to one form of harassment or derogatory treatment.
2.4.4. Sexual Harassment/Violence outside the University Setting
Survey participants were asked in two separate questions if they ever had been subjected to sexual harassment and/or sexual violence outside LU with the answer options: Yes, once; Yes, more than once; and No. All participants answering Yes, once or Yes, more than once to at least one of the two questions were categorised as exposed to SH/sexual violence outside LU and all others as not exposed.
2.5. Background Variables
Gender: Gender was based on two gender-related questions in the survey: ‘What gender were you assigned at birth?’ (female/male) and ‘what is your current gender identity?’ (Female/male/I do not identify as male or female). The answer to the second question was used to define participants as woman, man or non-binary; however, when the answer to this question was missing (N = 15), the answer to the first one was applied.
Age: Age was obtained by asking the participants what predefined age group they belonged to (≤30/31–40/41–49/50–59/≥60 years).
Form of employment: Participants were asked whether their form of employment was permanent or temporary.
Professional position: Professional position was specified according to nine types in the survey: “professor”, “senior lecturer”, “lecturer/teaching assistant”, “postdoc/associate senior lecturer”, “researcher/associate researcher”, “PhD student/research student”, “administrative staff/library staff”, “technical staff”, and “other”. For the purpose of analyses, some categories were grouped together. This resulted in the following six categories used in our analysis: “professors”, “senior lecturer”, “lecturer and researcher”, “PhD students”, “administrative and technical support staff”, and “others”.
Foreign background: Participants who were either born abroad or had two parents born abroad were categorised as individuals with a foreign background, in line with the definition used by Statistics Sweden [
40]. Participants lacking information about parents’ country of birth were presumed to have Swedish background if they were born in Sweden.
2.6. Gender, Function and Formal or Informal Power Relation of Perpetrators
Participants who reported experiences of harassment and derogatory treatment, respectively, were asked about the gender of the perpetrator/perpetrators (male, female, non-binary or don’t know) as well as their function (a person employed at LU, a PhD student/research student at LU, a student at LU, or/and another person that I met through my work at LU) and the power relation between the perpetrator and the participant. If the perpetrator was a university employee, the following options were offered: (1) a person upon whom I was reliant; (2) a person in position of formal seniority to me; (3) a person in a position of power (formal or informal) over me; (4) a person over whom I am/was in a position of (formal or informal) power; and (5) another person at Lund university. The alternatives (1), (2), and (3) were categorised together as “dominant/upper position” in the analyses, indicating that the perpetrator had a dominant or higher power position in relation to the victim. If the perpetrator was a PhD student, the above-described alternatives (3), (4), and (5) were offered and categorised as representing a perpetrator’s “dominant/upper position”, “dependent/lower position” or “another person/relationship”, respectively, in relation to the victim in the analysis. As a participant could have been subjected to harassment and derogatory treatment by several perpetrators, it was possible to mark several options regarding gender, function and power relation.
The same follow-up questions were asked about perpetrators of SH. However, as the questions about perpetrators referred to all experiences of SH, irrespective of time, we could not separate the information pertaining to perpetrators of SH during the last 12 months. However, the results regarding gender, function and power relation of perpetrators of SH (SH at any time) are published in a previous study presenting the overall results of the “Tellus” project [
39].
The question regarding the gender of the perpetrator was not put to the participants who reported that they had been exposed to harassment in the form of microaggressions only (81 participants). Therefore, those 81 participants were excluded from the total number of participants exposed to harassment in the top section of Table 3 showing the gender of perpetrators.
2.7. Statistical Analysis
We used descriptive statistics to analyse the sociodemographic characteristics of the sample, prevalence of SH, harassment and derogatory treatment, as well as gender, professional position and power relation of the perpetrators of harassment and derogatory treatment.
Association between background variables and SH was explored through a bivariate logistic regression analysis. A multivariable logistic regression analysis, including three different models adjusting for stepwise added background variables (age, foreign background and professional position) was performed to investigate associations between harassment, derogatory treatment or previous experience of SH outside LU and the outcome of SH. Interaction analyses were performed with dummy variables combining gender with exposure to harassment and gender with exposure to derogatory treatment, using men not exposed to harassment or derogatory treatment as the reference. When data allowed (not possible for all types of harassment due to small numbers), interaction analyses were performed by combining gender with specific types of harassment. The synergy indexes were calculated as proposed by Rothman, whereby a synergy index >1 indicates a synergistic effect, and a synergy index <1 an antagonistic effect [
41].
We thought it important to present data from non-binary participants in the descriptive statistics, as this is rarely carried out due to often small numbers. By including them we make our data available for future pooled studies. However, due to small numbers, we excluded non-binary participants (24 out of 2732) from further analysis.
Statistical analyses were performed in Stata version 13.
2.8. Ethics
The data collection and use were approved by the Regional Ethical Review Board in Lund (Dnr 2018/350) and was in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
4. Discussion
This study investigated the association between experience of different types of harassment or derogatory treatment and SH among employees at a large Swedish university. We found a three-times-higher risk of SH among women with experience of derogatory treatment and a sixfold-increased risk of SH among women with experience of harassment, where such harassment was mainly attributed to gender and age. The highest risk of experiencing SH among women (eight times higher risk) was found among those who had reported multiple forms of harassment or derogatory treatment. The increased risks remained statistically significant after adjusting for age, foreign background and professional position. A similar pattern was seen among men, although men reported lower prevalence of mistreatment in general. Further, we found a moderate synergistic effect of female gender on the association between harassment and the outcome of SH, indicating that women with experiences of harassment had a higher risk than men with the same experience of also experiencing SH. The non-binary individuals were not included in the above-mentioned analysis because of low statistical power, but they reported the highest prevalence of SH, harassment and derogatory treatment.
Our results indicate that experiences of SH co-occur with other forms of mistreatment. This is in line with results from a study by Lim and Cortina finding an association and co-occurrence between sexual harassment and general incivility [
29], as well as other research showing that individuals often report experiences of multiple forms of harassment and discriminatory behaviours at the workplace [
21]. Our results also indicate that power relations play a role in situations of harassment and derogatory treatment, as most of the victims reported that the perpetrators were in a dominant/higher position in relation to themselves. Similarly, previous results from the same survey showed that perpetrators of SH were in a dominant/higher position in relation to the victim [
39]. This result supports the theory that SH and other types of harassment may be driven by the need to defend or reinforce one’s power position [
2,
3]. We do not presuppose any causal direction between SH and other harassment or derogatory treatment. It is plausible that the defence of a social status or power position might involve a variety of behaviours directed towards the victim, thus generating experiences of multiple forms of mistreatment among the affected. The behaviours might vary as well as the perceptions of such behaviours, where victims’ perceptions might vary depending on previous experiences, social status or gender. For example, Berdahl and Aquino found a large gender difference regarding experiences of sexual behaviour at work; 46% of the men reported that they enjoyed it, compared to 10% of the women [
42]. Other researchers have identified individual factors, such as age, gender, gender role, past experiences of sexual harassment and perceptions of management’s tolerance of sexual harassment, to be related to attitudes toward sexual harassment [
43]. In this light, the identification of SH at the workplace can be an indicator of a larger problem, i.e., the existence of an unequal organisational culture where multiple types of abusive behaviours rooted in power imbalance and the defence of power contribute to unequal opportunities among employees. Our results have implications for designing preventive measures, suggesting that issues related to the defence of power and the various types of abusive behaviours including SH at a workplace both need to be addressed. On the other hand, it is possible that a positive spill-over effect could occur concerning other types of harassment behaviours if actions targeting SH behaviours are successful, given that the defence of power can be manifested in various ways by one individual.
In the university setting, there are various formal and informal hierarchical power relations that might trigger the defence of power irrespective of gender. However, underlying all other forms of power relations between staff is the informal and formal power imbalance between men and women. Men are generally attributed with higher status compared to the women in organisations, which gives men more informal and formal power [
44]. Moreover, a formal power imbalance is evident in that men dominate the higher positions in the Swedish academic sector [
45], which is also the case at Lund University, where, for example 72% of the professors were men in 2019 (personal communication human resources at LU) [
45]. This power imbalance between genders is the context within which our study results should be understood. A majority of the perpetrators of harassment and derogatory treatment in our study were men, and data from the same survey shows that this also applies to perpetrators of SH [
39]. The victims, on the other hand, were more often women; experiences of SH, harassment and derogatory treatment were about twice as common among women compared to men in our study. The formal and informal power imbalance between female and male university staff and its consequences should be systematically examined. Efforts to target and reduce workplace mistreatment at the academic workplace might mitigate the negative effects of gender inequality, although a true power balance is required to achieve real gender equality at academic workplaces.