Abstract
Objectives. To describe the prevalence of and risk factors for workplace violence among Ugandan adolescents.
Methods. The analysis focused on adolescents recruited at primary schools who participated in the endline survey of a trial in 2014 (at ages 11–14 years) and were followed up in 2018–2019 (at ages 17–19 years). The analysis was restricted to those engaged in past-year paid work (n = 1406). We estimated the prevalence of past-year workplace violence and used mixed-effects multivariable logistic regression to explore associations with characteristics measured in early adolescence, current life circumstances, and work-related factors.
Methods. The analysis focused on adolescents recruited at primary schools who participated in a 2014 survey and were followed up in 2018–2019. The analysis was restricted to those engaged in past-year paid work (n = 1406). We estimated the prevalence of past-year workplace violence and used mixed-effects multivariable logistic regression to explore associations with characteristics measured in early adolescence, current life circumstances, and work-related factors.
Results. Overall, 40% (95% confidence interval [CI] = 37%, 43%) of adolescents in paid work experienced past-year workplace violence; odds were doubled among female domestic workers (vs retail/trade workers; adjusted odds ratio [AOR] = 2.07; 95% CI = 1.28, 3.35). Experiences measured in early adolescence, including eating less than 3 meals the previous day, experiencing severe physical violence (male adolescents: AOR = 1.48; 95% CI = 1.11, 1.98; female adolescents: AOR = 1.69; 95% CI = 1.13, 2.53) and bullying, and having poor mental health (male adolescents: AOR = 2.32 95% CI = 1.37, 3.92; female adolescents: AOR = 2.27; 95% CI = 1.05, 4.89), were associated with increased odds of workplace violence. Current life circumstances (fewer household assets, more moves, functional difficulties, poorer mental health) were also associated with workplace violence.
Conclusions. Interventions are needed to address the high prevalence of workplace violence across all sectors, with female domestic workers particularly vulnerable. Early prevention of violence and poor mental health may be promising. (Am J Public Health. 2022;112(11):1651–1661. https://doi.org/10.2105/AJPH.2022.306983)
Violence, including workplace violence, is an expression of power and dominance, and those who are in subordinate social positions because of age, gender, and poverty may be particularly vulnerable.1,2 Survey data, mainly from high-income countries, suggest that 30% to 60% of women experience sexual harassment at work and that younger women are especially at risk.3 Limited data from studies conducted in low- and middle-income countries suggest that the prevalence of sexual harassment is as high or higher.4 These studies have reported the prevalence of workplace violence within individual sectors such as domestic work, agriculture, and mining.4–6 National surveys in several countries have shown that 14% to 40% of children involved in work experience violence.7–9
The consequences of workplace violence include negative effects on mental and physical health and social outcomes.10,11 Building on the momentum of the #MeToo and #TimesUp movements, International Labour Organization (ILO) member states have adopted the Violence and Harassment Convention 2019,12 providing an impetus to reduce workplace violence and harassment.
More than 47.5 million adolescents 15 to 17 years of age face working conditions that expose them to environmental hazards, excessive hours, or violence, especially in low- and middle-income countries.13 Young workers are at higher risk of workplace violence than adults3 and are more likely to engage in low-status, insecure, and unstable work, further increasing their vulnerability to workplace violence.4 Most countries have legislation on workers’ safety, including implicit and explicit provisions covering workplace violence.4 Although these legal guarantees are important, their effectiveness may be reduced in countries such as Uganda, where the vast majority of workers (91.9%) remain in informal employment and out of the reach of labor inspections.14
Despite increased awareness of the pervasive nature of workplace violence, there are virtually no large-scale quantitative data documenting the prevalence of different forms of violence across all young workers and sectors in low-income settings. Existing data are difficult to compare because of differences in definitions,4 and violence is likely underreported. To our knowledge, no longitudinal studies have investigated risk factors for workplace violence in low-income settings.
In Uganda, almost 1 in 5 adolescents 15 to 17 years of age are out of school and working14 and 14% have experienced violence at work,7 despite the existence of laws guaranteeing fundamental freedoms and rights.15–17 Yet, very little is known about forms of violence and types of perpetrators in different work settings. Using data from a cohort study of male and female adolescents originally residing in Uganda’s Luwero District, we sought to (1) describe the prevalence of physical, sexual, and emotional workplace violence across work sectors; (2) determine what work-related factors are associated with workplace violence; (3) explore what characteristics measured in early adolescence, including childhood exposure to violence, are associated with workplace violence in later adolescence; and (4) explore how workplace violence is associated with current life circumstances.
METHODS
We used data from the Contexts of Violence in Adolescence Cohort Study,18 an ongoing closed cohort investigation of adolescents originally recruited as part of a trial of the Good School Toolkit, a whole-school intervention designed to reduce violence in schools.19 Trial recruitment involved a 2-stage process. A list of all primary schools in the Luwero District was obtained in 2012, and schools with more than 40 grade 5 students were eligible. The 151 eligible schools contained more than 80% of all students in the district; 42 schools were randomly selected, and all agreed to participate. Twenty-one received an intervention from 2012 to 2014, and 21 served as a wait-list control group that received an intervention from 2015 to 2018.
A simple random sample of up to 130 students per school in grades 5 through 7 (11–14 years of age) was selected, and students were approached to participate in a survey in 2014; if fewer than 130 students were present, the complete sample was invited to participate. In total, 93% of students agreed to participate and 90% of participants agreed to ongoing follow-up, forming the wave 1 sample (n = 3431). At wave 2 (2018–2019), we successfully traced 81% of the participants (n = 2773; Table C, available as a supplement to the online version of this article at http://www.ajph.org). Adolescents who had moved to districts beyond those neighboring Luwero (4% of participants) were not traced owing to study operational considerations. There was some evidence of differential attrition: adolescents not completing a wave 2 survey had eaten fewer meals the previous day at wave 1, were older, and were likely to be experiencing higher levels of adversity.
Procedures
Wave 1 and wave 2 data were captured by trained Ugandan interviewers, who administered face-to-face survey interviews using hand-held devices. At wave 2, interviews took place at school, home, work, or community settings. Adolescents 18 years or older and emancipated minors provided informed written consent before participation. In the case of adolescents younger than 18 years who were not emancipated minors, caregivers were first provided with information and could opt out from their child participating. Adolescents who were not opted out by their caregiver were approached to provide informed written assent before participation. Interviewers assessed participants’ ability to understand consent procedures in English or Luganda before proceeding with the informed consent process.
At the end of the interview, all adolescents were offered counseling, regardless of what they disclosed. Those who disclosed violence or a well-being concern in accordance with our predefined criteria were, on their agreement, referred to a child protection officer or counselor for appropriate action depending on severity and timing. Open Data Kit was used to capture, transmit, and store all data on a secure server.
Study Population
In our analysis, we used data from participants who reported at wave 2 that they had been involved in paid work in the preceding 12 months (934 male participants [70%] and 472 female participants [33%]).
Study Measures
Workplace violence was defined as self-reports of violent acts perpetrated by an employer or adult in a work-related position of authority or by peers at the workplace. Table A (available as a supplement to the online version of this article at http://www.ajph.org) lists workplace violence items captured at wave 2 and the composite measures modeled as binary variables in our analysis. Our main outcome was any workplace violence in the past year. This included physical, emotional, or sexual violence from an employer or peer at work. Other measures included any sexual workplace violence in the past year (from an employer or peer at work) and specific types of violence (emotional, physical, and sexual) from employers and peers at work in the past year.
Students were asked about lifetime, past-year, and past-week experiences of violence from different perpetrators. A full list of the violence questions asked and the violence exposure measure are provided in Table A. We constructed binary measures to describe lifetime experiences of caregiver violence, sexual violence, and severe physical violence. Two past-year school violence binary measures (multiple acts of physical or emotional violence from a teacher or school staff member and bullying) and 1 categorical variable (polyvictimization) were also constructed.
Other measures have been widely used and were adapted and piloted with this population before use. As in past analyses, we used number of meals eaten yesterday as a proxy for socioeconomic status at wave 1. A list of measures is provided in Table A.
We constructed work sector groupings (shown in Table B, available as a supplement to the online version of this article at http://www.ajph.org) using ILO definitions,13 Ugandan national surveys,14 and data response frequencies for male and female adolescents (groups including less than 5% of the sample were combined for analysis).
Data Analysis
Stata version 16 (StataCorp LLC, College Station, TX) was used in conducting all of our analyses. Few data were missing (less than 1% for any variable used in the analysis; less than 1% of participants were dropped in any multivariable model because of missing data). Descriptive statistics for continuous variables included numbers of observations, means and standard deviations or standard errors, and medians and interquartile ranges (IQRs) for skewed data. Categorical variables are presented as frequencies and percentages. When producing prevalence estimates with associated 95% confidence intervals, we accounted for clustering at the school level using the Stata svy command.
We initially determined the prevalence of different types of workplace violence by sex and work sector. We then used mixed-effects logistic regression, with school modeled as a random effect to account for clustering at the school level, to explore how work-related factors and individual characteristics (exposures) were associated with any workplace violence in the past year (the primary outcome). All analyses were conducted separately for male and female adolescents. Unadjusted and adjusted odds ratios (ORs) are presented. Adjusted models included original study arm as a covariate to adjust for any study arm effects. Age, number of meals eaten yesterday, and any functional difficulty were identified a priori as potential confounders of specific exposure–outcome relationships. The covariates included in each model are indicated in the relevant tables.
RESULTS
Table B shows that the nature of work in which young people engage differs by sex. Female adolescents most commonly reported engaging in domestic work (32%) and retail or trade (32%), whereas male adolescents worked in farming (33%) and manual labor (28%). A similar percentage of male and female adolescents reported working in street or recreational work sectors (12% male and 17% female). The median age of adolescents who had engaged in paid work within the past year was 18 years (IQR = 17–19; Table 1). A higher proportion of male than female adolescents were in school or in training (41% vs 26%) and lived with an adult family member (61% vs 49%). Just over half of the participants had eaten 3 or more meals on the day before the interview, and 13% had no household assets (no electricity, radio, television, or refrigerator).
TABLE 1—
Wave 2 Sociodemographics and Workplace Violence Among Adolescents Engaged in Paid Work in the Past Year, by Sex: Uganda, 2018–2019
| Total (n = 1406), Median (IQR) or No./Total No. (%; 95% CI) | Male (n = 934), Median (IQR) or No./Total No. (%; 95% CI) | Female (n = 472), Median (IQR) or No./Total No. (%; 95% CI) | |
| Sociodemographic factors | |||
| Age, y | 18 (17–19) | 18 (17–19) | 18 (17–19) |
| In school/training | 511/1406 (36; 31, 42) | 387/934 (41; 35, 48) | 124/472 (26; 20, 33) |
| Lives with adult family member(s)a | 803/1406 (57; 54, 60) | 571/934 (61; 57, 65) | 232/472 (49; 44, 54) |
| Three or more meals eaten yesterday | 765/1406 (54; 51, 58) | 515/934 (55; 52, 59) | 250/472 (53; 47, 59) |
| No household assetsb | 178/1406 (13; 11, 15) | 109/934 (12; 10, 14) | 69/472 (15; 11, 18) |
| Workplace violence, past year | |||
| Any workplace violencec | 565/1406 (40; 37, 43) | 371/934 (40; 37, 43) | 194/472 (41; 36, 46) |
| Any sexual workplace violenced | 103/1406 (7; 6, 7) | 40/934 (4; 3, 6) | 63/472 (13; 11, 17) |
| Any employer and peer violence | 159/1406 (11; 9, 14) | 104/934 (11; 9, 14) | 55/472 (12; 9, 15) |
| Any employer violence only | 332/1406 (24; 22, 26) | 213/934 (23; 21, 25) | 119/472 (25; 21, 29) |
| Any peer violence in workplace only | 74/1406 (5; 4, 8) | 54/934 (6; 4, 8) | 20/472 (4; 3, 7) |
| Employer violence, past year | |||
| Any violence | 491/1406 (35; 32, 38) | 317/934 (34; 31, 37) | 174/472 (37; 32, 42) |
| Emotional violence | 463/1406 (33; 30, 36) | 300/934 (32; 29, 36) | 163/472 (35; 30, 40) |
| Physical violence | 109/1406 (8; 7, 9) | 74/934 (8; 6, 10) | 35/472 (7; 5, 10) |
| Sexual violence | 78/1406 (6; 5, 7) | 31/934 (3; 2, 5) | 47/472 (10; 8, 13) |
| Physical and emotional violence | 94/1406 (7; 6, 8) | 63/934 (7; 5, 9) | 31/472 (7; 5, 9) |
| Any violent act many times | 156/1406 (11; 9, 13) | 88/934 (9; 7, 12) | 68/472 (14; 11, 18) |
| Peer workplace violence, past year | |||
| Any violence | 233/1406 (17; 14, 19) | 158/934 (17; 14, 20) | 75/472 (16; 13, 20) |
| Emotional violence | 215/1406 (15; 13, 18) | 150/934 (16; 13, 19) | 65/472 (14; 11, 17) |
| Physical violence | 39/1406 (3; 2, 4) | 34/934 (4; 3, 5) | 5/472 (1; 0, 3) |
| Sexual violence | 37/1406 (3; 2, 4) | 12/934 (1; 1, 2) | 25/472 (5; 4, 8) |
| Physical and emotional violence | 34/1406 (2; 2, 4) | 29/934 (3; 2, 5) | 5/472 (1; 0, 3) |
Note. CI = confidence interval; IQR = interquartile range.
One or more of the following: biological father, biological mother, stepfather, stepmother, grandfather, or grandmother.
No electricity, radio, television, or refrigerator.
Any violence from employers or peers in the workplace in the past year.
Any sexual violence from employers or peers in the workplace in the past year.
Prevalence of Workplace Violence
Overall, 40% of male and 41% of female adolescents reported experiencing any type of workplace violence from an employer or peer at work in the past year (Table 1), and 4% of male and 13% of female adolescents reported experiencing sexual violence at their workplace within the past year. Employers were the most common perpetrators of violence, with 34% of male and 37% of female adolescents reporting any type of violence from an employer. Emotional violence was the most common form of violence perpetrated by employers, but 8% of participants reported past-year physical violence from an employer. Peers were also common perpetrators of workplace violence; 17% of adolescents reported peer workplace violence, with emotional violence the most commonly perpetrated form.
Table 2 and Figure A (available as a supplement to the online version of this article at http://www.ajph.org) show how workplace violence differed across sectors. The prevalence of workplace violence among male adolescents ranged from 35% to 48% across sectors, with employer and peer violence lowest in the farming category. In the case of female adolescents, the prevalence of workplace violence ranged from 34% to 54% across sectors. Half of female domestic workers (accounting for a third of our sample) reported violence from their employer, and 18% reported violence from peers (i.e., young people associated with their employer’s household).
TABLE 2—
Past-Year Workplace Violence Victimization Across Work Sectors, by Perpetrator and Sex: Uganda, 2018–2019
| Retail/Trade, No./Total No. (%; 95% CI) | Street/Recreational, No./Total No. (%; 95% CI) | Farming or Farming+, No./Total No. (%; 95% CI) | Workshop, No./Total No. (%; 95% CI) | Manual Work, No./Total No. (%; 95% CI) | Domestic Work, No./Total No. (%; 95% CI) | |
| Male | ||||||
| Violence in the workplace | ||||||
| Any workplace violence, past year | 42/98 (43; 33, 53) | 52/111 (47; 36, 58) | 108/311 (35; 29, 39) | 43/89 (48; 39, 58) | 109/261 (42; 35, 47) | |
| Sexual workplace violence, past year | 1/98 (1; 0, 7) | 6/111 (5; 2, 11) | 9/311 (3; 2, 5) | 5/89 (6; 2, 15) | 15/261 (6; 3, 10) | |
| Violence from employers | ||||||
| Any violence, past year | 32/98 (33; 25, 42) | 40/111 (36; 27, 47) | 96/311 (31; 26, 36) | 33/89 (37; 28, 48) | 96/261 (37; 31, 43) | |
| Sexual violence, past year | 1/98 (1; 0, 7) | 4/111 (4; 1, 9) | 6/311 (2; 1, 4) | 3/89 (3; 1, 14) | 14/261 (5; 3, 9) | |
| Violence from peers at the workplace | ||||||
| Any violence, past year | 21/98 (21; 15, 30) | 30/111 (27; 19, 37) | 38/311 (12; 8, 16) | 27/89 (30; 23, 39) | 42/261 (16; 11, 21) | |
| Sexual violence, past year | 0/98 (0) | 5/111 (5; 2, 10) | 3/311 (1; 0, 4) | 2/89 (2; 1, 9) | 1/261 (0; 0, 3) | |
| Female | ||||||
| Violence in the workplace | ||||||
| Any workplace violence, past year | 51/149 (34; 27, 43) | 36/78 (46; 33, 58) | 22/64 (34; 22, 45) | 82/153 (54; 44, 60) | ||
| Sexual workplace violence, past year | 16/149 (11; 6, 15) | 14/78 (18; 10, 26) | 8/64 (13; 6, 19) | 25/153 (16; 11, 21) | ||
| Violence from employers | ||||||
| Any violence, past year | 42/149 (28; 21, 37) | 30/78 (38; 26, 52) | 20/64 (31; 21, 44) | 76/153 (50; 42, 57) | ||
| Sexual violence, past year | 10/149 (7; 3, 10) | 7/78 (9; 3, 15) | 8/64 (13; 6, 19) | 21/153 (14; 9, 18) | ||
| Violence from peers at the workplace | ||||||
| Any violence, past year | 29/149 (19; 14, 26) | 13/78 (17; 9, 24) | 6/64 (9; 4, 16) | 27/153 (18; 11, 24) | ||
| Sexual violence, past year | 10/149 (7; 3, 11) | 8/78 (10; 3, 17) | 3/64 (5; 0, 9) | 5/153 (3; 0, 5) | ||
Note. CI = confidence interval. For female adolescents, farming, workshop, and manual work were combined and named “Farming+.”
Table 2 and Figure A show that past-year workplace sexual violence was reported by fewer male (range = 1%–6%) than female (range = 11%–18%) adolescents across all sectors. Confidence intervals overlapped across all sectors; however, among female adolescents, sexual violence was most commonly reported by those working in street or recreational and domestic work, and sexual violence from peers was more prevalent among both male and female adolescents in street or recreational work.
Work-Related Factors and Workplace Violence
Female domestic workers had more than double the odds of violence victimization than female adolescents who worked in retail or trade (Table 3). Among both male and female adolescents, time spent at work was the strongest predictor of violence, with those spending 9 to 12 hours at work (vs 4 hours or less) on an average day having the highest odds of violence. Those working in more than 1 job also had higher odds of experiencing workplace violence.
TABLE 3—
Work-Related Factors Associated With Any Past-Year Workplace Violence Among Those in Paid Work Within the Past Year (Wave 2), by Sex: Uganda, 2018–2019
| Workplace Violence, Past-Year Male Victimization (n = 934) | Workplace Violence, Past-Year Female Victimization (n = 474) | |||||
| No. (%) | Crude OR (95% CI) | AOR (95% CI) | No. (%) | Crude OR (95% CI) | AOR (95% CI) | |
| Workplace violence | 371 (40) | 194 (41) | ||||
| Main workplace | ||||||
| Retail/trade | 98 (10) | 1 (Ref) | 1 (Ref) | 149 (32) | 1 (Ref) | 1 (Ref) |
| Street/recreational | 111 (12) | 1.18 (0.68, 2.03) | 1.11 (0.63, 1.93) | 78 (17) | 1.56 (0.89, 2.75) | 1.59 (0.90, 2.83) |
| Farming | 311 (33) | 0.68 (0.43, 1.08) | 0.72 (0.45, 1.16) | . . . | . . . | . . . |
| Workshop | 89 (10) | 1.25 (0.70, 22.22 | 1.29 (0.72, 2.31) | . . . | . . . | . . . |
| Manual work | 261 (28) | 0.93 (0.58, 1.48) | 1.02 (0.63, 1.64) | . . . | . . . | . . . |
| Farming+ | . . . | . . . | . . . | 64 (14) | 0.94 (0.50, 1.76) | 1.03 (0.53, 1.97) |
| Domestic work | . . . | . . . | . . . | 153 (32) | 2.10 (1.32, 3.35) | 2.07 (1.28, 3.35) |
| Other | 64 (7) | 0.70 (0.36, 1.34) | 0.72 (0.37, 1.39) | 28 (6) | 0.65 (0.26, 1.63) | 0.66 (0.26, 1.68) |
| Main employment | ||||||
| Seasonal | 139 (15) | 1 (Ref) | 1 (Ref) | 34 (7) | 1 (Ref) | 1 (Ref) |
| Occasional/casual | 249 (27) | 1.25 (0.82, 1.93) | 1.23 (0.80, 1.91) | 56 (12) | 0.76 (0.30, 1.94) | 0.81 (0.31, 2.08) |
| Full time | 340 (36) | 1.54 (1.02, 2.31) | 1.34 (0.88, 2.03) | 328 (69) | 1.77 (0.83, 3.78) | 1.69 (0.78, 3.68) |
| Part time | 102 (11) | 1.14 (0.67, 1.93) | 1.06 (0.62, 1.81) | 34 (7) | 1.28 (0.47, 3.52) | 1.29 (0.46, 3.60) |
| Weekends | 103 (11) | 0.65 (0.37, 1.14) | 0.71 (0.40,1.26) | 19 (4) | 0.56 (0.15, 2.12) | 0.68 (0.18, 2.65) |
| Time spent on an average day at main work | ||||||
| ≤4 hours | 163 (17) | 1 (Ref) | 1 (Ref) | 73 (15) | 1 (Ref) | 1 (Ref) |
| 5–8 hours | 275 (29) | 1.57 (1.01, 2.45) | 1.51 (0.97, 2.36) | 107 (23) | 2.00 (1.00, 3.98) | 1.84 (0.91, 3.73) |
| 9–12 hours | 372 (40) | 3.62 (2.39, 5.49) | 3.25 (2.12, 4.98) | 215 (46) | 3.60 (1.92, 6.72) | 3.35 (1.76, 6.40) |
| >12 hours | 123 (13) | 2.18 (1.31, 3.63) | 2.09 (1.24, 3.51) | 76 (16) | 2.50 (1.20, 5.20) | 2.21 (1.04, 4.69) |
| Working more than 1 job | 361 (39) | 1.44 (1.10, 1.89) | 1.45 (1.11, 1.91) | 90 (19) | 1.66 (1.04, 2.65) | 1.63 (1.01, 2.63) |
Note. AOR = adjusted odds ratio; CI = confidence interval; OR = odds ratio. All models were adjusted by age and meals and included original study arm as a covariate. Groups containing less than 5% were combined as follows: for female adolescents, farming, workshop, and manual work were combined and named “Farming+,” and for male adolescents domestic work was combined with other. See Table B for a detailed breakdown.
Characteristics Measured in Early Adolescence and Workplace Violence
Adolescents who had eaten 3 or more meals and those who had better mental health in early adolescence (wave 1; ages 11–14 years) were less likely to report past-year workplace violence at ages 17 to 19 years (wave 2; Table 4). Male adolescents who had been involved in paid work in early adolescence had higher odds of later workplace violence, whereas those who felt more connected to their family when they were younger were less likely to report workplace violence. Experiences of most types of childhood violence (wave 1; ages 11–14 years) were associated with increased odds of later workplace violence among male adolescents, including caregiver violence, multiple acts of physical or emotional violence from a teacher or school staff member, bullying, severe physical violence, and polyvictimization. In the case of female adolescents, being bullied in early adolescence and experiencing severe physical violence from any perpetrator were associated with later workplace violence.
TABLE 4—
Associations of Characteristics Measured in Early Adolescence, Childhood Violence Exposures (Wave 1), and Current Life Circumstances (Wave 2) With Any Workplace Violence in the Past Year (Wave 2), by Sex: Uganda, 2014–2019
| Workplace Violence, Past-Year Male Victimization (n = 934) | Workplace Violence, Past-Year Female Victimization (n = 472) | |||||
| No. (%) or Mean (SE) | Crude OR (95% CI) | AOR (95% CI) | No. (%) or Mean (SE) | Crude OR (95% CI) | AOR (95% CI) | |
| Workplace violence | 371 (40) | 194 (41) | ||||
| Characteristics in early adolescence, wave 1 | ||||||
| Urban schoola | 377 (40) | 1.12 (0.86, 1.47) | 1.21 (0.92, 1.58) | 154 (33) | 0.80 (0.52, 1.24) | 0.83 (0.53, 1.31) |
| Three or more mealsa | 444 (48) | 0.67 (0.52, 0.87) | 0.68 (0.52, 0.89) | 164 (35) | 0.53 (0.35, 0.80) | 0.55 (0.36, 0.83) |
| Lived with biological parent | 579 (62) | 0.81 (0.62, 1.06) | 0.87 (0.66, 1.14) | 293 (62) | 1.17 (0.79, 1.73) | 1.27 (0.85, 1.89) |
| Paid work, ever | 564 (60) | 1.61 (1.22, 2.11) | 1.44 (1.09, 1.91) | 101 (21) | 1.31 (0.84, 2.06) | 1.32 (0.83, 2.08) |
| Any functional difficulty | 191 (20) | 1.25 (0.90, 1.72) | 1.12 (0.81, 1.56) | 118 (25) | 1.40 (0.91, 2.15) | 1.41 (0.91, 2.18) |
| Mental health score | 0.44 (0.01) | 2.31 (1.37, 3.88) | 2.32 (1.37, 3.92) | 0.46 (0.01) | 2.29 (1.07, 4.90) | 2.27 (1.05, 4.89) |
| Peer support score | 3.44 (0.06) | 0.97 (0.90, 1.05) | 0.96 (0.89, 1.04) | 3.40 (0.09) | 1.10 (1.00, 1.22) | 1.09 (0.98, 1.20) |
| School connectedness | 9.72 (0.07) | 0.98 (0.93, 1.03) | 0.98 (0.93, 1.03) | 9.95 (0.10) | 1.01 (0.94, 1.09) | 1.01 (0.94, 1.09) |
| Family connectedness | 9.92 (0.07) | 0.94 (0.88, 0.99) | 0.93 (0.87, 0.99) | 10.01 (0.10) | 1.02 (0.93, 1.11) | 1.03 (0.95, 1.13) |
| Childhood violence exposures, wave 1b | ||||||
| Caregiver, lifetime | 152 (16) | 1.41 (0.99, 2.00) | 1.38 (0.97, 1.98) | 97 (21) | 0.91 (0.57, 1.46) | 0.86 (0.53, 1.40) |
| Sexual, lifetime | 28 (3) | 1.14 (0.53, 2.44) | 1.02 (0.47, 2.21) | 62 (13) | 1.64 (0.95, 2.84) | 1.39 (0.80, 2.44) |
| Multiple acts of physical/emotional violence from teacher/school staff member | 193 (21) | 1.47 (1.07, 2.03) | 1.50 (1.08, 2.09) | 72 (15) | 1.12 (0.67, 1.90) | 1.17 (0.68, 1.99) |
| Bullying | 97 (10) | 1.49 (0.98, 2.27) | 1.60 (1.04, 2.46) | 42 (9) | 2.80 (1.43, 5.47) | 2.64 (1.33, 5.22) |
| Severe physical, lifetime | 302 (32) | 1.47 (1.11, 1.94) | 1.48 (1.11, 1.98) | 159 (34) | 1.63 (1.10, 2.41) | 1.69 (1.13, 2.53) |
| Polyvictimization | ||||||
| None (of the 3) | 415 (44) | 1 (Ref) | 1 (Ref) | 176 (37) | 1 (Ref) | 1 (Ref) |
| One | 333 (36) | 1.54 (1.14, 2.08) | 1.68 (1.23, 2.28) | 184 (39) | 1.17 (0.76, 1.80) | 1.11 (0.71, 1.74) |
| Two or 3 | 186 (20) | 1.57 (1.10, 2.24) | 1.68 (1.16, 2.43) | 112 (24) | 1.54 (0.94, 2.53) | 1.48 (0.89, 2.47) |
| Current life circumstances, wave 2 | ||||||
| Lives with family members | 571 (61) | 0.58 (0.45, 0.76) | 0.66 (0.50, 0.88) | 232 (49) | 0.84 (0.58, 1.23) | 0.90 (0.61, 1.34) |
| In school or training | 387 (41) | 0.41 (0.31, 0.54) | 0.46 (0.34, 0.63) | 124 (26) | 0.47 (0.30, 0.73) | 0.54 (0.33, 0.89) |
| No household assets | 109 (12) | 1.78 (1.19, 2.66) | 1.76 (1.17, 2.65) | 69 (15) | 1.79 (1.06, 3.02) | 1.70 (1.00, 2.90) |
| Moves since 2014 | ||||||
| 0 | 396 (42) | 1 (Ref) | 1 (Ref) | 74 (16) | 1 (Ref) | 1 (Ref) |
| 1 | 283 (30) | 1.63 (1.19, 2.23) | 1.46 (1.06, 2.02) | 170 (36) | 1.36 (0.76, 2.45) | 1.23 (0.68, 2.25) |
| 2 | 187 (20) | 1.73 (1.21, 2.47) | 1.50 (1.04, 2.16) | 132 (28) | 1.53 (0.83, 2.81) | 1.38 (0.74, 2.57) |
| ≥3 | 68 (7) | 2.53 (1.50, 4.26) | 2.53 (1.49, 4.31) | 96 (20) | 2.59 (1.36, 4.92) | 2.30 (1.19, 4.45) |
| Any functional difficulty | 223 (24) | 1.65 (1.21, 2.25) | 1.65 (1.21, 2.26) | 164 (35) | 2.04 (1.38, 3.01) | 1.94 (1.31, 2.89) |
| Peer support score | 3.85 (0.05) | 0.84 (0.77, 0.91) | 0.84 (0.77, 0.91) | 3.54 (0.08) | 0.90 (0.81, 1.01) | 0.92 (0.82, 1.03) |
| Mental health score | 0.42 (0.01) | 21.96 (11.89, 40.56) | 21.24 (11.36, 39.70) | 0.53 (0.01) | 13.35 (6.19, 28.80) | 12.6 (5.75, 27.60) |
Note. AOR = adjusted odds ratio; CI = confidence interval; OR = odds ratio. Models were adjusted by age and meals and included original study arm as a covariate.
Adjusted by age only.
Also adjusted by wave 1 functional difficulty.
Current Life Circumstances and Workplace Violence
Life circumstances measured in later adolescence (at wave 2, when data on workplace violence were measured) were also associated with workplace violence (Table 4). Adolescents who were working while in school or training had half the odds of workplace violence. Economic vulnerability, a higher frequency of residential moves, and having functional difficulties were associated with increased odds of workplace violence. Currently living with adult family members and having more peer support were associated with less workplace violence among young men. Male and female adolescents who were victims of workplace violence had highly elevated odds of concurrent poor mental health.
DISCUSSION
Our study confirms the widespread nature of workplace violence across work sectors. Overall, 2 in 5 working Ugandan adolescents reported past-year physical, sexual, or emotional workplace violence; 13% of female adolescents reported past-year sexual violence in the workplace, as compared with 4% of male adolescents. Employers were the most common perpetrators. Violence was common across all sectors, with female domestic workers at particularly high risk. Early adolescent economic hardship, violence, and poor mental health were associated with an increased risk of workplace violence in later adolescence. Work intensity and a higher frequency of moves were strongly associated with past-year workplace violence, and those with poorer mental health had much higher odds of reporting past-year violence.
Our findings extend those of other workplace violence studies conducted mainly in population subgroups and high-income countries.20,21 Differences in definitions and measurements limit direct comparisons across studies.3 However, other research has identified domestic workers as being at high risk of violence,6 and qualitative research with adolescent domestic workers in Uganda has also revealed widespread workplace sexual harassment and violence.22,23 Workplace violence is an expression of power asymmetries between employers and workers, exacerbated by deeply rooted social norms that devalue and stigmatize paid domestic work.22,23 The often invisible, unprotected, and unregulated nature of this type of work perpetuates risks of violence.24 In Uganda, placing adolescents with wealthier friends or relatives may increase the likelihood of forced labor,22 which is associated with exploitation, abuse, and violence. This practice is common in Uganda because employers prefer to hire domestic workers whose families they know.22
Our study demonstrates the association between experiences in early adolescence and later experiences of workplace violence, highlighting the potential for compounding of disadvantage. Severe physical violence and bullying, in particular, increased adolescents’ risk for later workplace violence.
Our results accentuate that primary prevention of all childhood violence is paramount to protect and achieve a healthy society.25 Childhood violence is associated with a myriad of detrimental effects in terms of educational outcomes, confidence, self-worth, mental health, social bonding, and future experiences of violence.25–27 These effects may be compounded if adolescents leave school because of school violence, which leads to early entry into informal work and increased vulnerability to workplace violence.28 We found that those with poorer mental health in early adolescence were particularly likely to experience workplace violence in later adolescence. Also, as in other studies,29,30 we found that current poor mental health was strongly associated with recent workplace violence even when we adjusted for mental health in early adolescence.
According to the ILO, 3 in 5 young adolescents globally are able to find only informal jobs in which there are low wages, little stability, and no social protection3,14 and reporting mechanisms and enforcement of rights are unlikely. Young people from disadvantaged backgrounds may have substantial economic pressure to stay in any job, regardless of its quality, safety, and likelihood of violence.31 Engaging in work at a younger age, working long hours,14 and being forced to move to find work can cause disruptions in social networks and stress,31 and we found that these factors were associated with workplace violence.
Our study revealed some differences by sex, with experiences of violence in early adolescence more strongly and uniformly associated with workplace violence among male adolescents and family support protective for workplace violence among male but not female adolescents. Both violence and work are highly gendered, and further research is needed to explore these differences.
Implications
There is a clear need for interventions to address workplace violence against adolescents and to stop violence from figures of authority in public institutions and organizations. The ILO advocates for intensive occupational safety and health initiatives to reduce workplace violence with an emphasis on laws and regulations,4 and in Uganda employers with more than 25 workers are required to adopt a written policy designed to combat sexual harassment.4 However, the existence of this law and others intended to prevent violence in Uganda points to the gap between the availability of laws and their implementation.15,16,32 Furthermore, organizational-level policies are unlikely to benefit young people employed in the informal sector.
The diverse range of workplaces, both formal and informal, are likely to need a variety of tailored approaches to violence prevention informed by an understanding of local practices and societal and gender norms pertaining to entry into work. Informal jobs are least likely to be covered by labor inspections, collective bargaining agreements, or legislation.3 Rapid and sustainable change may hinge on engaging employers, adolescent workers, families, communities, and policy implementers in coproduction of interventions that can protect against violence in informal sectors. Initiatives in Uganda include collective bargaining agreements that challenge sexual harassment in the horticulture industry, support for women transport workers engaged in informal jobs on issues around violence and equality, and an initiative in which women working in Kampala markets are uniting against harassment.3,33
In the formal sector, in addition to strengthening policies, regulations, and reporting mechanisms, there is a need for institutional change interventions to improve work environments. These interventions might draw on successful cultural change initiatives in other types of institutions such as the Good School Toolkit, which is designed to change the operational culture of schools. Developed by a Ugandan nongovernmental organization, the toolkit has been shown to reduce physical violence committed by teachers against students by 42%.19
Schools and vocational skill training schools are important for reaching young people before and as they navigate into work spaces. The ILO has produced a toolkit to support institutions and organizations in raising young people’s awareness of their rights at work.34 However, adolescents’ awareness alone is unlikely to prevent violence by employers or colleagues in work contexts characterized by deep power imbalances and lack of regulation and oversight.
Strengths and Limitations
To our knowledge, this is the first longitudinal study presenting data on the prevalence of workplace violence from different perpetrators across a wide range of work sectors and exploring relationships with earlier life circumstances. Our cohort was broadly representative of adolescents in the Luwero District and was not selected on the basis of any characteristics related to violence or work. However, there was some evidence of differential attrition by wave 2 (Table D, available as a supplement to the online version of this article at http://www.ajph.org). Our cohort members were exposed to a successful violence prevention intervention during their primary school years, which may have reduced their likelihood of experiencing subsequent workplace violence. The prevalence of workplace violence in our sample may therefore have been underestimated.
We used questions regarding specific acts of violence to document experiences of violence, in line with gold-standard methods,35 and made every effort to support safe disclosures. However, as in all violence studies, it is likely that acts of violence were underreported as a result of the stigma and fear attached to reporting some forms of violence. We captured sexual, physical, and emotional violence in childhood and later at work but did not capture other forms of indirect, structural, or political violence. It would be interesting in future research to explore not only single types of exposures to violence but overall patterns of violence exposure early in adolescence (and their associations with workplace violence) as well as in the workplace. We did not capture past-year workplace violence perpetrated by customers or suppliers who were not peers.21 Number of meals eaten (captured at wave 1) might not be a perfect proxy for socioeconomic status, but this is challenging to measure in children who are not aware of household markers such as assets.
We treated male and female adolescents separately because of the gendered nature of both work and violence; however, our sample of female adolescents was limited in size (n = 475), and thus our study may have been underpowered with respect to detecting associations with workplace violence. We collected data on current life factors at the same time as data on workplace violence; therefore, associations may have been bidirectional, and life factors can potentially be interpreted as both a cause and a consequence of workplace violence. Finally, it may not be possible to extrapolate our findings to other populations and settings.
Conclusions
Workplace violence against adolescents is common across a range of sectors, and early adolescent economic disadvantage, violence, and poor mental health are associated with increased risk. Interventions to prevent and address such violence are urgently required.
ACKNOWLEDGMENTS
Funding for this study was provided by the UK Medical Research Council (grant MR/L004321/1); the UK Economic and Social Research Council (grant ES/S005196/1); the UK Medical Research Council, the Department for International Development, and the Wellcome Trust (grant MR/R002827/1); and the Hewlett Foundation.
Electronic data solutions were provided by MRC/UVRI & LSHTM Uganda Research Unit, Entebbe, Uganda. The authors would like to thank Ayoub Kakande and Michael Charles Mubiru. We also thank Jodie Pearlman for her help in finalizing the article for publication.
CONFLICTS OF INTEREST
The authors have no conflicts of interest to disclose.
HUMAN PARTICIPANT PROTECTION
This study was approved by the ethics committees of the London School of Hygiene and Tropical Medicine, the University of London, the Uganda Virus Research Institute, and the Uganda National Council of Science and Technology. All participants provided informed consent/assent to take part.
Footnotes
See also Amo-Adjei and Fry, p. 1535.
REFERENCES
- 1.Bograd M. Feminist perspectives on wife abuse: an introduction. In: Bograd M, editor. Feminist Perspectives on Wife Abuse. Newbury Park, CA: Sage Publications; 1988. pp. 11–26. [Google Scholar]
- 2.Kiss L, Yun K, Pocock N, Zimmerman C. Exploitation, violence, and suicide risk among child and adolescent survivors of human trafficking in the greater Mekong subregion. JAMA Pediatr. 2015;169(9):e152278. doi: 10.1001/jamapediatrics.2015.2278. [DOI] [PubMed] [Google Scholar]
- 3. Violence and Harassment Against Women and Men in the World of Work: Trade Union Perspectives and Action . 2017.
- 4. Safe and Healthy Working Environments Free from Violence and Harassment . 2020.
- 5.Jacobs S, Brahic B, Olaiya MM. Sexual harassment in an East African agribusiness supply chain. Econ Labour Relat Rev. 2015;26(3):393–410. doi: 10.1177/1035304615595604. [DOI] [Google Scholar]
- 6.Thi AM, Zimmerman C, Pocock NS, Chan CW, Ranganathan M. Child domestic work, violence, and health outcomes: a rapid systematic review. Int J Environ Res Public Health. 2022;19(1):427. doi: 10.3390/ijerph19010427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.The National Labour Force and Child Activities Survey 2011/2012: Child Labour Report. Kampala, Uganda: Uganda Bureau of Statistics; 2013
- 8.Ethiopia National Child Labour Survey 2015. Addis Ababa, Ethiopia: International Labour Organization; 2018
- 9.Bangladesh National Child Labour Survey 2013. Dhaka, Bangladesh: Bangladesh Bureau of Statistics; 2015
- 10.Tsuno K, Kawachi I, Kawakami N, Miyashita K. Workplace bullying and psychological distress: a longitudinal multilevel analysis among Japanese employees. J Occup Environ Med. 2018;60(12):1067–1072. doi: 10.1097/JOM.0000000000001433. [DOI] [PubMed] [Google Scholar]
- 11.Leach LS, Poyser C, Butterworth P. Workplace bullying and the association with suicidal ideation/thoughts and behaviour: a systematic review. Occup Environ Med. 2017;74(1):72–79. doi: 10.1136/oemed-2016-103726. [DOI] [PubMed] [Google Scholar]
- 12.International Labour Organization. 2022. https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:C190
- 13.World Report on Child Labour . 2015.
- 14. Labour Market Transition of Young People in Uganda: Highlights of the School-to-Work Transition Survey . 2015.
- 15.Parliament of Uganda. 2022. https://www.parliament.go.ug/documents/1240/constitution
- 16.Uganda Employment Act, 2006 (Act No. 6) 2022. https://www.ilo.org/dyn/natlex/natlex4.detail?p_lang=en&p_isn=74416
- 17.Government of Uganda. 2022. https://www.ilo.org/dyn/natlex/docs/SERIAL/92798/108265/F1420511694/UGA92798.pdf
- 18.Devries K, Parkes J, Knight L, et al. Context of Violence in Adolescence Cohort (CoVAC) study: protocol for a mixed methods longitudinal study in Uganda. BMC Public Health. 2020;20(1):43. doi: 10.1186/s12889-019-7654-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Devries KM, Knight L, Child JC, et al. The Good School Toolkit for reducing physical violence from school staff to primary school students: a cluster-randomised controlled trial in Uganda. Lancet Glob Health. 2015;3(7):e378–e386. doi: 10.1016/S2214-109X(15)00060-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.European Union Agency for Fundamental Rights. 2022. https://fra.europa.eu/en/publication/2014/violence-against-women-eu-wide-survey-main-results-report [PubMed]
- 21.Rauscher KJ. Workplace violence against adolescent workers in the US. Am J Ind Med. 2008;51(7):539–544. doi: 10.1002/ajim.20590. [DOI] [PubMed] [Google Scholar]
- 22.Namuggala V. Exploitation or empowerment? Adolescent female domestic workers in Uganda. Int J Child Youth Family Stud. 2015;6(4):561–580. doi: 10.18357/ijcyfs.64201514288. [DOI] [Google Scholar]
- 23.Kiss L, Zimmerman C. Human trafficking and labor exploitation: toward identifying, implementing, and evaluating effective responses. PLoS Med. 2019;16(1):e1002740. doi: 10.1371/journal.pmed.1002740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hatton E. Mechanisms of invisibility: rethinking the concept of invisible work. Work Employ Soc. 2017;31(2):336–351. doi: 10.1177/0950017016674894. [DOI] [Google Scholar]
- 25.Fry D, Blight S. How prevention of violence in childhood builds healthier economies and smarter children in the Asia and Pacific region. BMJ Glob Health. 2016;1(suppl 2):i3–i11. doi: 10.1136/bmjgh-2016-000188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.De Bellis MD, Hooper SR, Sapia JL. Early trauma exposure and the brain. In: Vasterling JJ, Brewin CR, editors. Neuropsychology of PTSD: Biological, Cognitive, and Clinical Perspectives. New York, NY: Guilford Press; 2005. pp. 153–177. [Google Scholar]
- 27.Devries K, Grundlingh H, Knight L. Cycles of violence in gendered social contexts: why does child maltreatment increase risk of intimate partner violence. In: Gideon J, editor. Gender and Health Handbook. Cheltenham, United Kingdom: Edward Elgar Publishing Inc; 2016. pp. 375–393. [Google Scholar]
- 28.Fry D, Fang X, Elliott S, et al. The relationships between violence in childhood and educational outcomes: a global systematic review and meta-analysis. Child Abuse Negl. 2018;75:6–28. doi: 10.1016/j.chiabu.2017.06.021. [DOI] [PubMed] [Google Scholar]
- 29.Houle JN, Staff J, Mortimer JT, Uggen C, Blackstone A. The impact of sexual harassment on depressive symptoms during the early occupational career. Soc Ment Health. 2011;1(2):89–105. doi: 10.1177/2156869311416827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Devries KM, Mak JY, Bacchus LJ, et al. Intimate partner violence and incident depressive symptoms and suicide attempts: a systematic review of longitudinal studies. PLoS Med. 2013;10(5):e1001439. doi: 10.1371/journal.pmed.1001439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. World Employment and Social Outlook: Trends 2020 . 2020.
- 32.Uganda National Council for Children. Children (Amendment) Act. 2016. https://www.unicef.org/uganda/media/5416/file/UGTheChildrenAmendmentActFinalWeb.pdf
- 33.McCool A.2022. https://www.theguardian.com/global-development/2019/aug/19/men-fear-us-kampalas-market-women-unite-against-harassment
- 34.Corbanese V, Rosas G. Rights@Work 4 Youth: Decent Work for Young People: Facilitators’ Guide and Toolkit. Geneva, Switzerland: International Labour Organization; 2016. [Google Scholar]
- 35.Garcia-Moreno C, Jansen HAFM, Ellsberg M, Heise L, Watts C. WHO Multi-Country Study on Women’s Health and Domestic Violence Against Women: Initial Results on Prevalence, Health Outcomes and Women’s Responses. Geneva, Switzerland: World Health Organization; 2005. [Google Scholar]