Risk factors for workplace sexual harassment and violence among a national cohort of women in Iceland: a cross-sectional study.

BACKGROUND: Sexual harassment and violence in the workplace are a serious public health concern for women worldwide with substantial costs due to sick leave and personnel turnover. Yet little is known about the prevalence of sexual harassment and violence at a population level, especially across work sectors. The aim of this study was to investigate the prevalence of workplace sexual harassment and violence by demographic factors and work sectors among Icelandic women.
METHODS: For this cross-sectional study we analysed nationally representative, de-identified individual-level data from women who had responded to an online survey item about self-labelled current and lifetime workplace sexual harassment or violence as part of the Stress and Gene Analysis (SAGA) study, a cross-sectional nationally representative study done from March 1, 2018, to July 1, 2019. Eligible participants were women who resided in Iceland, were aged between 18 and 69 years, spoke Icelandic, and had a registered address from the Icelandic Population Register or a telephone number from the online 1819 service. We used binomial and Poisson regression analysis to study the cross-sectional association between workplace sexual harassment and violence and demographic factors (eg, age, sexual orientation, and education) and factors relating to the workplace (eg, work schedule), across works sectors.
FINDINGS: Of 113 814 eligible women, 104 197 were invited to complete the online survey, of whom 30 403 women responded and were included in the SAGA cohort. 15 799 women answered the item about exposure to workplace sexual harassment or violence. 11 286 [71·4%] of 15 799 women answered the question about sexual orientation that were included in the survey from June, 2018. 5291 (33·5%) of 15 799 of participants reported having experienced workplace sexual harassment or violence during their lifetime, and 1178 (7·5%) in their current workplace. Such exposure in the current workplace was most common among women who were young (age 18-24 years: prevalence ratio [PR] 3·89 [95% CI 2·66-5·71]; age 25-34 years: 3·66 [2·53-5·31]), single (1·27 [1·12-1·43]), and worked shifts (2·32 [2·02-2·67]), with the highest prevalence rates observed among women in work sectors of public figures (15·67 [9·34-25·12]), tourism (15·01 [11·01-20·13]), and the legal system and security (13·56 [7·00-24·66]). Lifetime exposure to workplace sexual harassment or violence was more common among women who belonged to sexual minorities than among heterosexual women (PR 1·35 [1·24-1·46]).
INTERPRETATION: Lifetime exposure to workplace sexual harassment or violence seems common among women in a Nordic welfare state. These findings provide nuanced targets for prevention and for public policies aimed at promoting women's safety in the work environment. FUNDING: Icelandic Gender Equality Fund, European Research Council, and Icelandic Centre for Research.

Introduction

WHO defines sexual violence as any sexual act, attempt to obtain a sexual act, or an act directed against a person's sexuality using coercion, including rape, attempted rape, or sexual harassment. Workplace sexual harassment includes requests for sexual favours, and other verbal or physical harassment of a sexual nature that creates an intimidating, hostile, or offensive work environment, interferes with work performance, or affects an individual's employment. Women who are subjected to sexual harassment or sexual violence (sexual harassment or violence hereafter) at work seldom file complaints or report their experiences to their supervisors. However, they are more likely than other women to quit their job owing to the harassment. Being subjected to workplace sexual harassment has been associated with decreased self-confidence and increased stress, fatigue, sleeping problems, anxiety, depression, and suicidal ideation.5, 6, 7, 8, 9 Despite these reported negative effects, workplace sexual harassment and violence has been understudied worldwide.

The reported global prevalence of workplace sexual harassment and violence among women varies greatly between studies, ranging from 11–72%,3, 7, 11, 12, 13 and accumulated evidence suggests that rates might differ by work sector.10, 14 High prevalence has been documented among female social and health-care workers (51–69%), journalists (57%), and women working in the science, technology, engineering, and mathematics (STEM) sector (46%). Sexual harassment and violence in the academic environment is also common; 22–47% of female graduate students report that they were sexually harassed by a faculty member and 34–58% by another student.18, 19, 20, 21 The great variability in prevalence across studies is largely owing to varying methods used to assess sexual harassment or violence.10, 14

Evidence before this study

We searched Web of Science, PubMed, the SAGE Journals website, ScienceDirect, and Google Scholar for articles published in English from Jan 1, 1990 to Feb 13, 2022, using the search terms “workplace sexual harassment”, “workplace sexual harassment or violence”, “occupational violence”, and “workplace violence”, and also sometimes “among women”, “prevalence”, “risk factors”, “national study”, and “population-based study” mainly within Science Direct and Google Scholar. Additionally, we searched for reports on sexual harassment on the websites of international public sector bodies and organisations (eg, US Centre for Gender Equity and Health). We excluded studies that did not report prevalence separately for men and women. Previous research indicates that workplace sexual harassment and violence is common among women worldwide, although across studies, the prevalence (11–72%) varies considerably due to the study populations selected (eg, women in academia) and variations in the measurement of sexual harassment and violence. Previous studies have assessed the prevalence of sexual harassment, or sexual violence, or both, either by asking participants directly for a subjective rating of experience of these behaviours (self-labelled), or by asking them to respond to a list of such behaviours (predefined); the predefined method usually yields higher prevalence rates than the self-labelled one. Epidemiological reports indicate that workplace sexual harassment is more often aimed at women than at men and poses a serious threat to public health. For example, being the target of workplace sexual harassment and violence has been associated with negative health outcomes such as anxiety, depression, and suicidal ideation. Previous studies have shown that perpetrators include colleagues, subordinates, and supervisors, but also third-party individuals in the work environment, such as customers, clients, or patients. Despite these reported negative effects, historically sexual harassment and violence has been understudied worldwide.

Added value of this study

Increased knowledge and understanding about the epidemiology of sexual harassment and violence is crucial for the development of effective workplace policies and preventive measures. Using a nationally representative cohort of 15 799 women aged 18–69 years old, this study is, to our knowledge, the first to assess the prevalence of sexual harassment and violence in a nationally representative cohort of women employed in various work sectors. We found the prevalence of workplace sexual harassment and violence to be highest among women who were defined as public figures (eg, politicians and actors), those who worked in the legal system and security sector or in tourist services, and among women who were young, highly educated, had a minority sexual identity, had a high-income, and for whom shift work or long or irregular work hours were common. The findings of this study provide valuable information to improve public policies aimed at promoting women's safety in the work environment.

Implications of all the available evidence

The current study provides evidence for how widespread workplace sexual harassment and violence is among women residing in a Nordic welfare state, particularly among young women and those who work shifts. The current study adds nuanced information on the prevalence of sexual harassment across various work sectors, by identifying sectors in which women are at high risk of such exposure. Taken together, these findings provide guidance as to where interventions might be needed, for example, in work sectors with a third-party presence (eg, patients, clients, or customers) and a general high male-to-female ratio.

Few studies have assessed what factors might place women at increased risk of workplace sexual harassment and violence. The few existing studies on this topic indicate that workplace sexual harassment and violence is more commonly experienced by younger women (ie, 18–39 years old)22, 23 and women with a university degree. An Australian report published in 2018 found that individuals belonging to sexual minorities (eg, gay, lesbian, bisexual, or pansexual) were more likely to report being exposed to such behaviour at work in the past 5 years than heterosexuals were (52% vs 31%).

Despite being informative, previous studies10, 14 had limited generalisability because of small sample sizes and a focus on a narrow range of work sectors. To our best knowledge, no previous study has examined the rates of sexual harassment and violence across work sectors at the population level. The present study aimed to expand on the existing literature by using a nationally representative cohort of 15 799 women to assess the prevalence of workplace sexual harassment and violence by demographic factors (eg, age, sexual orientation, and education), factors relating to the workplace (eg, work schedule), and across 18 works sectors.

Methods

Study design and population

This cross-sectional cohort study used de-identified data from women in the Icelandic Stress and Gene Analysis (SAGA) cohort, a nationally representative population-based study that assessed the effect of trauma on women's health. The SAGA cohort invited all women who were residing in Iceland, aged 18–69 years, who spoke Icelandic, and had a registered address or telephone number by which they could be contacted to complete an online survey to collect information on demographics, work-related factors, and self-labelled current and lifetime workplace sexual harassment or violence. Our study included data for those women who answered a SAGA cohort survey item about workplace sexual harassment or violence that was added to the survey on May 4, 2018.

Participants used their unique Icelandic national identity number to log into the SAGA survey website, provided informed consent electronically, then answered an extensive online survey of 374 questions before Jan 3, 2018, and 389 questions thereafter. For further details about the SAGA cohort study procedure, see the appendix (p 1). The question on sexual orientation was not added to the survey until June 22, 2018; values were coded as unknown for participants who were not asked the question.

Measurement of exposure

Workplace was defined as the work environment or the academic environment. In accordance with previous research, workplace sexual harassment and violence was assessed using the following question: “Have you experienced sexual harassment or violence in your work or academic environment?” The response options were: no; yes, in my current work or academic environment; yes, in my former work or academic environment; and yes, in both my current and former work or academic environment. Answers were coded according to two exposures of interest: exposure to sexual harassment and violence in a current workplace (irrespective of previous exposure), and exposure to sexual harassment or violence over one's lifetime (in the current workplace, or a previous workplace, or both).

Demographics and work factors

Demographic characteristics included: age (18–24, 25–34, 35–44, 45–54, or 55–69 years); country of birth (Iceland or other); sexual orientation (heterosexual, sexual minority [lesbian, bisexual, not defined, or other], relationship status (married or in a relationship, or single); highest level of education (elementary, high school or vocational, bachelor's degree or equivalent, master's degree, doctoral degree, or other); current monthly income (≤€2040, €2041–3400, €3401–4761, or >€4761); and geographical location (capital, north, east, west or Westfjords, south, or south peninsula). The response options for current employment status were full-time work, part-time work, self-employed, student, homemaker or on parental leave, retired, on disability benefits, on sick leave for more than 2 months, or unemployed. If respondents selected retired, not working, homemaker, or on parental leave, in addition to stating that they were active in the labour market (ie, working, studying, or an employer), respondents were classified as being active in the labour market. We also assessed the number of working hours per week (<20, 20–39, 40–59, 60–79, or >80 h), and work schedule (day job, shift work, or irregular hours; appendix pp 6–7). Participants were also able to select cannot or will not answer in response to all questions except those on age, weekly hours worked, and work schedule.

Current work sector was assessed with the open-ended question: how would you describe your job title? The answers were manually coded into 43 groups that represented similar occupations or work environments. The groups of occupations were then grouped into 18 different work sectors on the basis of the US Bureau of Labor Statistics categorisation system, as adapted to the Icelandic work environment (eg, a tourist services sector was added; and the architecture and engineering and the computer and mathematical sectors were combined into the work sector technology and software; appendix p 7).

Statistical analysis

We used R (version 4.1.3) for the statistical analysis. We used the χ2 test to assess differences between different demographic and work-related factors in exposure to workplace sexual harassment and violence, first, in the current workplace (current vs previous only) and second, over the participant's lifetime (exposed vs never exposed). We used Poisson log-linear models with robust error variance to report the association of demographic and work-related factors with prevalence ratios (PRs) and 95% CIs, because the relative risk interpretation given to odds ratios can be misleading for outcomes with a high prevalence. The log-transformed effect estimates from the modified Poisson model can be interpreted as prevalence ratios.28, 29 Results are shown with and without adjustment for age in years as a continuous variable. Effect modification by age was statistically significant for some of the demographics (eg, education, income, and relationship status) for current workplace exposure and exposure over the lifetime. Results were also stratified by age (<45 years vs ≥45 years). Binomial logistic regression models were fitted to the data to assess the consistency between the models, evaluate goodness of fit, and obtain prevalence estimates of harassment and violence by work sector. Adjusted marginal estimates of the prevalence of harassment and violence in the current workplace, with 95% CIs, were obtained using the R emmeans package (version 1.7.2), including adjustment for age, income, and work schedule. The appendix (p 4) shows the unadjusted prevalence. To assess the effect of missingness on our results, we completed two seperate case analyses: one that included both the cannot or will not answer and unknown values; and a second that excluded these values. Overall, our findings remained the same. We used multiple binomial logistic regression models with adjustment for all demographic characteristics to estimate a predicted prevalence of workplace sexual harassment and violence. These estimates were then split into four equal quartile groups from low risk to high risk (for further details see the appendix pp 10–11). The study protocol is in the appendix (pp 12–13). Our study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology checklist (appendix p 14).

The SAGA cohort study was approved by the National Bioethics Committee of Iceland (grant VSNb2017110046/03.01) and the Icelandic Data Protection Authority.

Role of the funding source

The funders of the study had no role in study design, data collection, data analysis, data interpretation, writing of the manuscript, or the decision to submit.

Results

Between March 1, 2018, and July 1, 2019, 104 197 women were invited to complete the SAGA cohort online survey, and 30 403 women responded and were included in the SAGA cohort, of whom 15 799 answered the item about exposure to workplace sexual harassment or violence (figure 1). 7557 (47·8%) of 15 799 women answered the question about their occupation.

Figure 1

Study profile

Characteristics of the study population are shown in table 1. We compared the distribution of age, education level, monthly wages, and geographical region of participants with those for the total female population of Iceland (according to available data from Statistics Iceland30, 31, 32) and found that participants in this study were representative of the adult female population in Iceland (appendix pp 2–3).

Table 1

Baseline characteristics

	Participants (n=15 799)
Age, years
18–24	1585 (10·0%)
25–34	2762 (17·5%)
35–44	3241 (20·5%)
45–54	7039 (44·6%)
55–69	1172 (7·4%)
Country of birth
Iceland	15 302 (96·9%)
Other	488 (3·1%)
Cannot or will not answer	9 (<0·1%)
Sexual orientation
Heterosexual	10 544 (66·7%)
Sexual minority	742 (4·7%)
Cannot or will not answer	38 (0·2%)
No information*	4475 (28·3%)
Relationship status
Married or in a relationship	12 099 (76·6%)
Single	3625 (22·9%)
Cannot or will not answer	75 (0·5%)
Highest level of education
Elementary	2623 (16·6%)
High school or vocational	4180 (26·5%)
Bachelor's degree or equivalent	4766 (30·2%)
Master's degree	2957 (18·7%)
Doctorate degree	183 (1·2%)
Other	1020 (6·5%)
Cannot or will not answer	70 (0·4%)
Monthly income
≤€2040	4817 (30·5%)
€2041–3400	4647 (29·4%)
€3401–4761	3573 (22·6%)
>€4761	2083 (13·2%)
Cannot or will not answer	679 (4·3%)
Time worked per week, h
<20	1025 (6·5%)
20–39	4267 (27·0%)
40–59	6665 (42·4%)
60–79	507 (3·2%)
≥80	140 (0·9%)
Not relevant†	3195 (20·2%)
Work schedule
Day job	8379 (72·0%)
Shift work	1555 (13·4%)
Irregular	1697 (14·6%)
Not relevant†	4168 (26·0%)
Geographical location in Iceland
Capital	10 555 (66·8%)
North or east	2187 (13·8%)
West or Westfjords	1034 (6·5%)
South	1206 (7·6%)
South peninsula	793 (5·0%)
Cannot or will not answer	24 (0·5%)

Data are n (%).

11 286 (71%) of 15 999 participants answered questions about sexual orientation, which was not added to the survey until June 22, 2018.

Participants who were unemployed, exclusively on disability benefits, or non-working students were not asked this question.

Overall, 5291 (33·5%) of 15 799 of women had been exposed to workplace sexual harassment or violence, with 1178 (7·5%) of 15 799 women having experienced such treatment in their current workplace.

Table 2, Table 3 show the percentage of women exposed to sexual harassment or violence in the workplace, by demographic and work-related factors. Differences in the prevalence of sexual harassment or violence by demographic characteristics and work-related factors are shown in table 4 expressed as PRs and 95% CIs. The mean age for women in this study was 44·5 years (SD 14·0). We found that, using Poisson analysis, younger women (<55 years) were at increased risk of sexual harassment or violence in their current workplace compared with women aged 55–69 years (age 18–24 years PR 3·89 [95% CI 2·66–5·71]; and 25–34 years 3·66 [2·53–5·31]). Results stratified by age (younger than 45 years [n=7588] vs 45 years and older [n=8211]) are shown in the appendix (pp 5, 8, 9). Adjusting for age, single women (PR 1·27; 95% CI 1·12–1·43), women with a high school degree (1·27 [1·05–1·53]), bachelor's degree (1·42 [1·18–1·70]), or master's degree (1·48 [1·21–1·80]); and women with a moderate monthly income (€2041–3400; 1·42 [1·22- 1·65]) or high monthly income (1·92 [1·65–2·24] for €3401–4761 and 1·93 [1·61–2·30] for >€4761) were at increased risk of sexual harassment and violence in the current workplace compared with women in a relationship or with other levels of education and income. Women residing in the capital city area (PR 1·21 [95% CI 1·02–1·44]) and women working long hours each week (40–59 h 1·36 [1·09–1·70]; 60–79 h 1·97 [1·45–2·67]; >80 h 1·74 [1·05–2·88]), working shifts (2·32 [2·02–2·67]), and working irregular hours (1·42 [1·20–1·67]) also had an elevated risk of exposure. Women with a high income (>€3401 per month) and university-level education (bachelor, master, and doctorate) were more likely to work at least 40 h per week than women with lower education (64·8% vs 46·6%, p<0·001) or lower income (56·8% vs 25·6%, p<0·001).

Table 2

Exposure to workplace sexual harassment or sexual violence by demographic characteristic

	Total participants (n=15799)	Exposure over lifetime		Setting of exposure
		Exposed (n=5291)	Never exposed (n=10 508)	Current workplace, irrespective of previous exposure (n=1178)	Previous workplace only (n=4113)
Age, years*
18–24	1585/15 799 (10·0%)	543/1585 (34·3%)	1042/1585 (65·7%)	158/543 (29·1%)	385/543 (70·9%)
25–34	2762/15 799 (17·5%)	1068/2762 (38·7%)	1694/2762 (61·3%)	259/1068 (24·3%)	809/1068 (75·7%)
35–44	3241/15 799 (20·5%)	1228/3241 (37·9%)	2013/3241 (62·1%)	284/1228 (23·1%)	944/1228 (76·9%)
45–54	7039/15 799 (44·6%)	2172/7039 (30·9%)	4867/7039 (69·1%)	447/2172 (20·6%)	1725/2172 (79·4%)
55–69	1172/15 799 (7·4%)	280/1172 (23·9%)	892/1172 (76·1%)	30/280 (10·7%)	250/280 (89·3%)
Country of birth†
Iceland	15 302/15 799 (96·9%)	5139/15 302 (33·6%)	10 163/15 302 (66·4%)	1142/5139 (22·2%)	3997/5139 (77·8%)
Other	488/15 799 (3·1%)	148/488 (30·3%)	340/488 (69·7%)	35/148 (23·6%)	113/148 (76·4%)
Cannot or will not answer	9 (<0·1%)	4/9 (44·4%)	5/9 (55·6%)	1/9 (11·1%)	3/9 (33·3%)
Sexual orientation*
Heterosexual	10 544/15 799 (66·7%)	3548/10 544 (33·6%)	6996/10 544 (66·4%)	810/3548 (22·8%)	2738/3548 (77·2%)
Sexual minority	742/15 799 (4·7%)	361/742 (48·7%)	103/742 (13·9%)	71/361 (19·7%)	290/361 (80·3%)
Cannot or will not answer	38/15 799 (0·3%)	10/38 (26·3%)	28/38 (73·7%)	5/38 (13·2%)	5/38 (13·2%)
Unknown‡	4475/15 799 (28·3%)	1372/4475 (30·7%)	3103/4475 (69·3%)	292/1372 (21·3%)	1080/1372 (78·7%)
Relationship status*
Married or in a relationship	12 099/15 799 (76·6%)	3848/12 099 (31·8%)	8251/12 099 (68·2%)	837/3848 (21·8%)	3011/3848 (78·2%)
Single	3625/15 799 (22·9%)	1419/3625 (39·1%)	2206/3625 (60·9%)	337/1419 (23·7%)	1082/1419 (76·3%)
Cannot or will not answer	75/15 799 (0·5%)	24/75 (32·0%)	51/75 (68·0%)	4/24 (16·7%)	20/24 (83·3%)
Highest level of education*
Elementary	2623/15 799 (16·6%)	780/2623 (29·7%)	1843/2623 (70·3%)	149/780 (19·1%)	631/780 (80·9%)
High school or vocational	4180/15 799 (26·5%)	1410/4180 (33·7%)	2770/4180 (66·3%)	324/1410 (23·0%)	1086/1410 (77·0%)
Bachelor's degree or equivalent	4766/15 799 (30·2%)	1568/4766 (32·9%)	3198/4766 (67·1%)	376/1568 (24·0%)	1192/1568 (76·0%)
Master's degree	2957/15 799 (18·7%)	1078/2957 (36·5%)	1879/2957 (63·5%)	243/1078 (22·5%)	835/1078 (77·5%)
Doctorate degree	183/15 799 (1·1%)	72/183 (39·3%)	111/183 (60·7%)	13/72 (18·1%)	59/72 (81·9%)
Other	1020/15 799 (6·5%)	369/1020 (36·2%)	651/1020 (63·8%)	73/369 (19·8%)	296/369 (80·2%)
Cannot or will not answer	70 (0·4%)	14/70 (20·0%)	56/70 (80·0%)	0	14/14 (100·0%)
Monthly income*
≤€2040	4817/15 799 (30·5%)	1716/4817 (35·6%)	3101/4817 (64·4%)	287/1716 (16·7%)	1429/1716 (83·3%)
€2041–3400	4647/15 799 (29·4%)	1455/4647 (31·3%)	3192/4647 (68·7%)	336/1455 (23·1%)	1119/1455 (76·9%)
€3401–4761	3573/15 799 (22·6%)	1172/3573 (32·8%)	2401/3573 (67·2%)	332/1172 (28·3%)	840/1172 (71·7%)
>€4761	2083/15 799 (13·2%)	745/2083 (35·8%)	1338/2083 (64·2%)	192/745 (25·8%)	553/745 (74·2%)
Cannot or will not answer	679/15 799 (4·3%)	203/679 (29·9%)	476/679 (70·1%)	31/203 (15·3%)	172/203 (84·7%)
Time worked per week, h*
<20	1025/15 799 (6·5%)	351/1025 (34·2%)	674/1025 (65·8%)	83/351 (23·6%)	268/351 (76·4%)
20–39	4267/15 799 (27·0%)	1290/4267 (30·2%)	2977/4267 (69·8%)	311/1290 (24·1%)	979/1290 (75·9%)
40–59	6665/15 799 (42·2%)	2192/6665 (32·9%)	4473/6665 (67·1%)	584/2192 (26·6%)	1608/2192 (73·4%)
60–79	507/15 799 (3·2%)	221/507 (43·6%)	286/507 (56·4%)	67/221 (30·3%)	154/221 (69·7%)
≥80	140/15 799 (0·9%)	57/140 (40·7%)	83/140 (59·3%)	16/57 (28·1%)	41/57 (71·9%)
Unknown§	3195/15 799 (20·2%)	1180/3195 (36·9%)	2015/3195 (63·1%)	117/1180 (9·9%)	1063/1180 (90·1%)
Work schedule*
Day job	8379/15 799 (53·0%)	2456/8379 (29·3%)	5923/8379 (70·7%)	563/2456 (22·9%)	1893/2456 (77·1%)
Shift work	1555/15 799 (9·8%)	642/1555 (41·3%)	913/1555 (58·7%)	275/642 (42·8%)	367/642 (57·2%)
Irregular	1697/15 799 (10·7%)	650/1697 (38·3%)	1047/1697 (61·7%)	166/650 (25·5%)	484/650 (74·5%)
Unknown§	4168/15 799 (26·4%)	1543/4168 (37·0%)	2625/4168 (63·0%)	174/1543 (11·3%)	1369/1543 (88·7%)
Geographical location in Iceland*
Capital	10 555/15 799 (66·8%)	3780/10 555 (35·8%)	6775/10 555 (64·2%)	850/3780 (22·5%)	2930/3780 (77·5%)
North or east	2187/15 799 (13·8%)	628/2187 (28·7%)	1559/2187 (71·3%)	143/628 (22·8%)	485/628 (77·2%)
West or Westfjords	1034/15 799 (6·5%)	303/1034 (29·3%)	731/1034 (70·7%)	62/303 (20·5%)	241/303 (79·5%)
South	1206/15 799 (7·6%)	352/1206 (29·2%)	854/1206 (70·8%)	73/352 (20·7%)	279/352 (79·3%)
South peninsula	793/15 799 (5·0%)	220/793 (27·7%)	573/793 (72·3%)	47/220 (21·4%)	173/220 (78·6%)
Cannot or will not answer	24/15 799 (0·2%)	8/24 (33·3%)	16/24 (66·7%)	3/24 (12·5%)	5/24 (20·8%)

Data are n/N (%). p values calculated using a χ2 test on workplace sexual harassment and violence (never, current workplace, previous workplace) and demographic characteristics. Respondents were not given a “cannot or will not answer” option for the questions on age, weekly hours worked, and work schedule. NA=not reported.

p<0·001.

p=0·573.

Did not receive the question; 11 286 (71%) of 15 999 participants answered the question about sexual orientation, which was not added to the questionnaire until June 22, 2018.

Did not receive question as they were not employed or only students.

Table 3

Exposure to workplace sexual harassment or sexual violence by current employment status

	Total participants (n=15 799)	Age, years	Exposure during lifetime		Setting of exposure
			Exposed (n=5291)	Never exposed (n=10 508)	Current workplace, irrespective of previous exposure (n=1178)	Previous workplace only (n=4113)
Full-time work	7002/15799 (44·3%)	46 (36–55)	2205/7002 (31·5%)	4797/7002 (68·5%)	610/2205 (27·7%)	1595/2205 (72·3%)
Part-time work	2186/15799 (13·8%)	51 (40–60)	632/2186 (28·9%)	1554/2186 (711%)	175/632 (27·7%)	457/632 (72·3%)
Employer	955/15799 (6·0%)	49 (41–57)	333/955 (34·9%)	622/955 (65·1%)	47/333 (14·1%)	286/333 (85·9%)
Student—not working	973/15799 (6·2%)	25 (21–30)	363/973 (37·3%)	610/973 (62·7%)	57/363 (15·7%)	306/363 (84·3%)
Student—working	1272/15799 (8·1%)	26 (22–37)	501/1272 (39·4%)	771/1272 (60·6%)	156/501 (31·1%)	345/501 (68·9%)
Receiving disability benefits—not working	1228/15799 (7·8%)	54 (43–61)	485/1228 (39·5%)	743/1228 (60·5%)	NA	454/485 (93·6%)
Receiving disability benefits—working	216/15799 (1·4%)	49 (42–57)	77/216 (35·6%)	139/216 (64·4%)	16/77 (20·8%)	61/77 (79·2%)
On sick leave for >2 months	399/15799 (2·5%)	47 (36–56)	172/399 (43·1%)	227/399 (56·9%)	NA*	140/172 (81·4%)
Homemaker or on parental leave	601/15799 (3·8%)	33 (29–41)	220/601 (36·6%)	381/601 (63·4%)	NA†	188/220 (85·5%)
Retired	559/15799 (3·5%)	68 (67–69)	126/559 (22·5%)	433/559 (77·5%)	NA	119/126 (94·4%)
Unemployed	290/15799 (1·8%)	41 (29–54)	128/290 (44·1%)	162/290 (55·9%)	NA	120/128 (93·8%)
Cannot or will not answer	118/15799 (0·8%)	49 (34–58)	49/118 (41·5%)	69/118 (58·5%)	7/49 (14·3%)	42/49 (85·7%)

Data are n (%) or median (IQR). Workplace was defined as the work environment, or the academic environment, or both. NA=not reported.

Women on sick leave were not asked about their current workplace from which they were off on sick leave.

Women who either had a formal job to go back to after parental leave or were homemakers without a job.

Table 4

Prevalence of exposure to workplace sexual harassment or sexual violence by demographic characteristic

	All participants	Exposed			Exposed in current workplace
		Yes (%)	Prevalence ratio (95% CI)	Prevalence ratio adjusted for age (95% CI)	Yes (%)	Prevalence ratio (95% CI)	Prevalence ratio adjusted for age (95% CI)
Age, years
55–69	1172/15 799 (7·4%)	280/1172 (23·9%)	1·00 (Ref)	NA	30/1172 (2·6%)	1·00 (Ref)	NA
45–54	7039/15 799 (44·6%)	2172/7039 (30·9%)	1·29 (1·16–1·44)	NA	447/7039 (6·4%)	2·48 (1·72–3·57)	NA
35–44	3241/15 799 (20·5%)	1228/3241 (37·9%)	1·59 (1·42–1·77)	NA	284/3241 (8·8%)	3·42 (2·36–4·96)	NA
25–34	2762/15 799 (17·5%)	1068/2762 (38·7%)	1·62 (1·45–1·81)	NA	259/2762 (9·4%)	3·66 (2·53–5·31)	NA
18–24	1585/15 799 (10·0%)	543/1585 (34·3%)	1·43 (1·27–1·62)	NA	158/1585 (10·0%)	3·89 (2·66–5·71)	NA
Sexual orientation
Heterosexual	10 544/11 286 (93·4%)	3548/10 544 (33·6%)	1·00 (Ref)	1·00 (Ref)	810/10 544 (7·7%)	1·00 (Ref)	1·00 (Ref)
Sexual minority	742/11 286 (6·6%)	361/742 (48·7%)	1·45 (1·34–1·56)	1·35 (1·24–1·46)	71/742 (9·6%)	1·25 (0·99–1·57)	1·03 (0·82–1·31)
Relationship status
Married or in a relationship	12099/15 724 (76·9%)	3848/12 099 (31·8%)	1·00 (Ref)	1·00 (Ref)	837/12 099 (6·9%)	1·00 (Ref)	1·00 (Ref)
Single	3625/15 724 (23·1%)	1419/3625 (39·1%)	1·23 (1·17–1·29)	1·20 (1·15–1·26)	337/3625 (9·3%)	1·34 (1·19–1·52)	1·27 (1·12–1·43)
Highest level of education
Elementary	2623/15 729 (16·7%)	780/2623 (29·7%)	1·00 (Ref)	1·00 (Ref)	149/2623 (5·7%)	1·00 (Ref)	1·00 (Ref)
High school or vocational	4180/15 729 (26·6%)	1410/4180 (33·7%)	1·13 (1·06 −1·22)	1·10 (1·02–1·18)	324/4180 (7·8%)	1·36 (1·13–1·65)	1·27 (1·05–1·53)
Bachelor's degree or equivalent	4766/15 729 (30·3%)	1568/4766 (32·9%)	1·11 (1·03–1·19)	1·11 (1·03–1·19)	376/4766 (7·9%)	1·39 (1·16–1·67)	1·42 (1·18–1·70)
Master's degree	2957/15 729 (18·8%)	1078/2957 (36·5%)	1·23 (1·14–1·32)	1·38 (1·14–1·67)	243/2957 (8·2%)	1·45 (1·19–1·76)	1·48 (1·21–1·80)
Doctorate degree	183/15 729 (1·1%)	72/183 (39·3%)	1·32 (1·09–1·59)	1.16 (1·09–1·24)	13/183 (7·1%)	1·25 (0·72–2·16)	1·41 (0·82–2·42)
Other	1020/15 729 (6·5%)	369/1020 (36·2%)	1·22 (1·10–1·35)	1·28 (1·16–1·42)	73/1020 (7·2%)	1·26 (0·96–1·65)	1·44 (1·10–1·89)
Monthly income
€≤2040	4817/15 120 (31·9%)	1716/4817 (35·6%)	1·00 (Ref)	1·00 (Ref)	287/4817 (5·9%)	1·00 (Ref)	1·00 (Ref)
€2041–3400	4647/15 120 (30·7%)	1455/4647 (31·3%)	0·88 (0·83–0·93)	0·92 (0·87–0·98)	336/4647 (7·2%)	1·21 (1·04–1·41)	1·42 (1·22–1·65)
€3401–4761	3573/15 120 (23·6%)	1172/3573 (32·8%)	0·92 (0·87–0·98)	0·98 (0·92–1·04)	332/3573 (9·3%)	1·56 (1·34–1·82)	1·92 (1·65–2·24)
€>4761	2083/15 120 (13·8%)	745/2083 (35·8%)	1·00 (0·94–1·08)	1·07 (1·00–1·15)	192/2083 (9·2%)	1·55 (1·29–1·84)	1·93 (1·61–2·30)
Time worked per week, h
<20 hours	1025/12 604 (8·1%)	351/1025 (34·2%)	1·00 (Ref)	1·00 (Ref)	83/1025 (8·1%)	1·00 (Ref)	1·00 (Ref)
20–39 hours	4267/12 604 (33·9%)	1290/4267 (30·2%)	0·88 (0·80–0·97)	0·99 (0·90–1·10)	311/4267 (7·3%)	0·90 (0·71–1·14)	1·13 (0·89–1·43)
40–59 hours	6665/12 604 (52·9%)	2192/6665 (32·9%)	0·96 (0·88–1·05)	1·08 (0·98–1·19)	584/6665 (8·8%)	1·08 (0·87–1·35)	1·36 (1·09–1·70)
60–79 hours	507/12 604 (4·0%)	221/507 (43·6%)	1·27 (1·12–1·45)	1·40 (1·23–1·60)	67/507 (13·2%)	1·63 (1·20–2·21)	1·97 (1·45–2·67)
>80 hours	140/12 604 (1·1%)	57/140 (40·7%)	1·19 (0·96–1·48)	1·32 (1·06–1·64)	16/140 (11·4%)	1·41 (0·85–2·34)	1·74 (1·05–2·88)
Work schedule
Day job	8379/11 631 (72·0%)	2456/8379 (29·3%)	1·00 (Ref)	1·00 (Ref)	563/8379 (6·7%)	1·00 (Ref)	1·00 (Ref)
Shift work	1555/11 631 (13·4%)	642/1555 (41·3%)	1·41 (1·32–1·50)	1·33 (1·24–1·43)	275/1555 (17·7%)	2·63 (2·30–3·01)	2·32 (2·02–2·67)
Irregular	1697/11 631 (14·6%)	650/1697 (38·3%)	1·31 (1·22–1·40)	1·29 (1·21–1·39)	166/1697 (9·8%)	1·46 (1·23–1·72)	1·42 (1·20–1·67)
Geographical location in Iceland
North or east	2187/15 775 (13·9%)	628/2187 (28·7%)	1·00 (Ref)	1·00 (Ref)	143/2187 (6·5%)	1·00 (Ref)	1·00 (Ref)
Capital area	10 555/15 775 (66·9%)	3780/10 555 (35·8%)	1·25 (1·16–1·34)	1·24 (1·16–1·33)	850/10 555 (8·1%)	1·23 (1·04–1·46)	1·21 (1·02–1·44)
West or Westfjords	1034/15 775 (6·6%)	303/1034 (29·3%)	1·02 (0·91–1·15)	1·03 (0·91–1·15)	62/1034 (6·0%)	0·92 (0·69–1·22)	0·93 (0·69–1·24)
South	1206/15 775 (7·6%)	352/1206 (29·2%)	1·02 (0·91–1·13)	1·03 (0·92–1·14)	73/1206 (6·1%)	0·93 (0·70–1·22)	0·94 (0·72–1·24)
South peninsula	793/15 775 (5·0%)	220/793 (27·7%)	0·97 (0·85–1·10)	0·96 (0·85–1·09)	47/793 (5·9%)	0·91 (0·66 1·25)	0·90 (0·66–1·24)

Data are n/N (%) or prevalence ratio (95% CI). Workplace was defined as the work environment, or the academic environment, or both. NA=not applicable.

Figure 2 shows the prevalence of current workplace sexual harassment and violence by work sector, adjusted for age, work schedule, and income. The highest prevalence of sexual harassment and violence in the current workplace was among women working as public figures (15·7% [95% CI 9·3–25·1]), in tourist services (15·0% [11·0–20·1]), and in the legal system and security (13·6% [7·0–24·7]). The lowest prevalences were among women working in office and administration jobs (5·8% [95% CI 4·2–8·0]), and restaurant and catering (6·4% [3·8–10·6]).

Figure 2

Prevalence of sexual harassment and violence in the current workplace by work sector

Prevalence rates and 95% CIs adjusted for age, work schedule, and income.

The percentages of women exposed to workplace sexual harassment or violence at some point during their life, by demographic and work-related factors, are shown in Table 2, Table 3. Results from the Poisson analysis indicated that younger women (<55 years) appear to be at elevated risk of exposure during their lifetime compared with women aged 55–69 years (18–24 years PR 1·43 [95% CI 1·27–1·62], 25–34 years 1·62 [1·45–1·81], for age; and 35–44 years 1·59 [1·42–1·77], 45–54 years 1·29 [1·16–1·44]). Adjusting for age, women belonging to a sexual minority (PR 1·35 [95% CI 1·24–1·46]), and single women (1·20 [1·15–1·26]) were also at elevated risk. Compared with other groups of women, those with high school or vocational education (PR 1·10 [95% CI 1·02–1·18]), university-level education (bachelor's degree 1·11 [1·03–1·19], master's degree 1·38 [1·14–1·67], and doctorate degree 1·16 [1·09–1·24]), and those residing in the capital city area (1·24 [1·16–1·33]) were at elevated risk of lifetime exposure to sexual harassment or violence. Women working long hours per week (60–79 h PR 1·40 [95% CI 1·23–1·60], >80 h 1·32 [1·06–1·64]), working shifts (1·33 [1·24–1·43]), and working irregular hours (1·29 [1·21–1·39]) were also at increased risk (table 4).

Discussion

To our knowledge, this large-scale cohort study is the first to assess the prevalence of workplace sexual harassment and violence among women across multiple work sectors at the population-level. Our findings indicate that approximately a third of women report exposure to sexual harassment or violence during their lifetime, and 8% report exposure in their current workplace. Building and expanding on previous findings, we identified several factors associated with increased risk of workplace sexual harassment and violence among women. Specifically, sexual harassment and violence within the current workplace was most frequently reported by women who were young, highly educated, working shifts or irregular hours, or working within tourism, the legal system and security, or as public figures. Furthermore, lifetime exposure to workplace harassment and violence was more common among women in sexual minorities than among heterosexual women.

Two population-based studies done in Iceland (n=3219) and Sweden (n=85 205), found that the prevalence of current workplace sexual harassment, for both men and women, was 4% in Iceland and 5% in Sweden. Previous reports17, 22 suggest that women are more commonly the targets of sexual harassment or violence than men are, which probably explains the higher prevalence in our study than in the other two population-based studies. Also, our study was done in the wake of the global #MeToo movement, which received substantial public and media attention in Iceland, which might partly explain higher prevalence ratings of sexual harassment and violence in the current study than the previous two population-based studies. Importantly, our study and the other two population-based studies9, 33 were done in Nordic welfare states, which share many cultural aspects and repeatedly score among the highest countries in the world on measures of gender equality. Future studies are needed to map the extent of sexual harassment and violence at population level in diverse cultures.

The prevalence of sexual harassment or violence in the current workplace differed by demographic factors. We found that the prevalence was higher for women who were highly educated and had a high income, which was consistent with a study done in 28 EU member states. The women in our study who were university-educated and had a high income tended to work long hours (≥40 h per week), which might increase their risk of exposure, given the increased amount of time spent in the workplace. As documented in previous reports,22, 23 we also found that sexual harassment and violence were most commonly reported by young women (aged <35 years), despite the fact that older women have had a higher number of years in the labour market. This difference between age groups might result from generational differences in perceptions of what behaviours constitute workplace sexual harassment and violence; the #MeToo movement led to open discussions about violence against women, especially among young adults, which might have influenced how young women in the study conceptualised sexual harassment and violence and perhaps made them increasingly willing to acknowledge that they had been exposed to such behaviour in their workplace. Another possibility is that recall bias might have contributed to older women under-reporting lifetime exposure to workplace sexual harassment or violence.

In line with reports published in 2018 and 2019, our results indicated that sexual harassment and violence over the lifetime was more likely to happen to women who belonged to a sexual minority (lesbian, bisexual, other, or not defined) than to heterosexual women. This imbalance might be linked to the sexualisation of the identities of women belonging to sexual minorities, and the misconception that they focus heavily on sexual activity. For example, women belonging to sexual minorities might be subjected to inappropriate comments at work regarding their sex lives. Up to now, few studies have assessed workplace sexual harassment and violence among sexual minorities and further research is needed to assess the manifestation of sexual harassment and violence that these populations are subjected to.

Higher rates of sexual harassment and violence in a previous workplace were reported by unemployed women than by women who were active in the labour market. Unemployed women might have quit their jobs because of being sexually harassed in their workplace, which would be in line with results from a national sample of American workers (n=364), which found that 40% of women who reported experiencing workplace sexual harassment left their job because of the occurrence.

We found that current workplace sexual harassment and violence were more common among women who worked shifts and irregular hours than among women who worked during the day. This might be because women who work evenings and nights often work alone (due to understaffing or lower workloads outside of office hours) and might be in contact with third-party individuals (eg, patients, clients, or customers). Also consistent with previous studies,18, 19, 20, 21 our findings suggest that sexual harassment and violence are common among students (PR 6%). Similar factors to those given for women who work shifts and irregular hours might contribute to this elevated risk for students. For example, students might keep irregular schedules while studying (perhaps alone during evenings or weekends). Students also interact to varying degrees with a large number of people (eg, teachers, fellow students, and people or patients at internship sites). One report suggested that contact with third-party individuals is linked to increased risk of workplace sexual harassment and violence. Unfortunately, we do not have information about either the circumstances in which the exposure occurred or about the perpetrators. Future studies on the specific circumstances in which sexual harassment or violence occurs might provide important insights into why prevalence is high in some work environments.

We found that some work sectors had higher rates of sexual harassment and violence than others, namely public figures (eg, politicians and actors), tourist services, legal system and security, manufacturing and repair (eg, factories and construction work), and health care. Our findings add to previous evidence15, 16, 17, 37, 38 based on smaller sample sizes than ours, which indicates that women are at greater risk of sexual harassment or violence in some work environments than others. There are several explanations for higher rates in those work sectors. First, the sectors all involve the presence of a third-party, which has been found to increase the risk of exposure. Second, these sectors tend to be male-dominated work environments with higher male-to-female ratios than in many other work sectors. Research has shown that there tends to be an increased prevalence of sexual harassment and violence in male-dominated environments, potentially because of increased interaction with men, who are more likely than women to be perpetrators.17, 22

Most workplaces in Iceland and elsewhere in the western Europe and North America have implemented sexual harassment grievance procedures and anti-harassment training for their employees. However, our findings suggest that more refined and targeted preventive and intervention efforts are needed within specific work environments to eliminate sexual harassment and violence, despite there being little empirical support for such efforts. Interventions include teaching women specific strategies to reduce the risk of sexual harassment and violence and interventions led by professionals that focus on gender role socialisation (eg, the consideration of gender role stereotyping). Given the high prevalence of sexual harassment or violence, and the increased risk in workplaces in which women are in contact with third-party individuals, community-based interventions are needed to change public discourse, practices, and norms regarding this problem, both within and outside of the workplace. It is crucial to monitor the effectiveness of interventions that target sexual harassment or violence within the workplace, as such interventions might not have the intended effect, and might even contribute to a backlash against women (eg, victim blaming and retaliation) or increase the likelihood of harassment.

The strengths of our study include a large sample size (n=15 799) that was representative of the adult female population in Iceland with respect to demographic characteristics. Because few sexual harassment and violence cases come to the attention of the authorities, self-reported data provide a more accurate picture of the prevalence of workplace sexual harassment and violence than official records, such as legal reports, do. In terms of study limitations, the generalisability of our results is limited to women who have an Icelandic national identity number and could complete the survey in Icelandic. We therefore do not have data on residents who do not speak the native language, who might be particularly vulnerable to workplace sexual harassment and violence. Young women (age <35 years) and those with a high school or vocational education as their highest degree were slightly under-represented in our sample compared with the general female population in Iceland. Additionally, participants were asked to rate subjectively whether they had been exposed to sexual harassment or violence at work, as opposed to being asked for objective ratings of predefined behaviour. Subjective ratings have been found to lead to under-reporting compared with objective ratings. Furthermore, we do not have information about the type (ie, verbal or physical) or severity of the sexual harassment and violence involved, and therefore cannot comment on how these might differ depending on the characteristics of the women or the workplace. Because of the framing of the question (ie, have you experienced sexual harassment or violence in the workplace?), we cannot exclude the possibility that some women reported physical, rather than sexual, violence within the workplace. We also do not have information about who the perpetrators of workplace sexual harassment or violence were in the present study. Given the nature of our cross-sectional analysis, we are unable to reflect on whether rates of workplace sexual harassment or violence increase or decrease over time. Furthermore, despite being a population-based study, the sample size of women who worked in the legal system and security or as public figures, was small. Thus, results from those work sectors need to be interpreted with caution. Finally, although we adjusted for age, we cannot exclude the possibility of unmeasured or residual confounding.

In conclusion, our findings support that there is a high prevalence of sexual harassment and violence aimed at women in the workplace, with approximately a third of women in Iceland reporting such exposure over their lifetime. Expanding upon previous work, our results identified several potential areas in which to target prevention and intervention efforts. Namely, we found that sexual harassment and violence within the workplace was most common for women who were young, single, highly educated, who belonged to a sexual minority, had high incomes, and who worked long or irregular hours, in shifts, or in the sectors of tourist services, legal system and security, manufacturing and repair, health care, or as a public figure.

Data sharing

The data used in this study were compiled in the Stress-And-Gene-Analysis (SAGA) cohort. A detailed description of the study protocol and self-report measures used in this study and R codes can be found on the SAGA cohort's website (https://afallasaga.is/english/) and in the appendix (pp 6–7,12–13,15–38). These data cannot be made publicly available, according to Icelandic data protection laws and the terms of approval for the current study that were stipulated by the National Bioethics Committee of Iceland. The SAGA cohort contains extremely sensitive data, and all use of data is subject to NBC approval (email: vsn@vsn.is). Interested researchers can obtain access to de-identified data by submitting a proposal to the SAGA cohort data management board (email: afallasaga@hi.is), which will assist with submitting a request to the NBC to obtain data from the study. The corresponding author of the present study submitted a research proposal to the SAGA cohort data management board and the NBC and was given access only to de-identified data, which the corresponding author is not permitted to share in any way.

Declaration of interests

We declare no competing interests.