Direct and indirect effects of workplace sexual harassment on the productivity of victims and witnesses: The preventive role of equitable management

Abstract

This research demonstrates the impact of equitable management as a protective factor against workplace sexual harassment (WSH) and its consequences on labor productivity. It also shows that there are invisible costs for colleagues who witness WSH, through counterproductive behaviors, such as sabotage or production deviance, with an indirect decrease in labor productivity. We used a structured questionnaire that was answered by 827 women from 37 small, medium, and large private companies in the Lima Metropolitan Area, Peru. We designed a conceptual model and tested it using structural covariance equations. The results indicate that 33.5 % of women have been sexually harassed over the last 12 months, an average of 6.6 times, while 18.9 % of women have supported co-workers who were victims of WSH. Being sexually harassed at work decreases labor productivity by 43.1 % and increases the intention to desert the company by 15.2 %. Witnessing WSH increases the intention to drop out by 11.3 % and increases counterproductive behaviors by 39.6 %. We found that equitable management is a preventative factor for WSH. Equitable management not only decreases the probability of the occurrence of WSH by 2.2 times but also—if it exists—reduces its pernicious impact on productivity through various indirect effects. Equitable management can reduce the labor productivity costs caused by WSH by 4.6 times.

1. Introduction

Workplace sexual harassment (WSH) is one of the most prevalent forms of gender violence at work. A meta-analysis based on 55 probability samples finds that—on average—58 % of women report having experienced potentially harassing behavior, and 24 % report having experienced WSH at work [1]. WSH is not an isolated incident but rather a pattern of behaviors. WSH has three essential dimensions: sexual coercion, which consists of implicit or explicit promises or threats conditioned on “sexual favors”; unwanted sexual attention, which consists of unwanted verbal or physical sexual advances that can be offensive and intimidating; and gender harassment, which involves verbal and non-verbal behaviors that convey hostility, objectification, exclusion, or subordinate status based on gender [2].

Research shows that women suffer more WSH than men [3,4], with co-workers and bosses being the main perpetrators. WSH creates a hostile, intimidating, and offensive environment, interfering with job performance. A meta-analysis by Willness et al. [5], based on 41 studies, found that WSH decreases job satisfaction and organizational commitment, increases job abandonment and poor physical and mental health, and even causes stress disorder symptoms, such as post-traumatic stress disorder in victims. Indeed, women who are sexually harassed at work rarely report it [6], yet they quit their job more often due to the harassment [7]. WSH diminishes self-confidence and increases post-traumatic stress, fatigue, sleep problems, depersonalization, anxiety, burn-out, psychotropic self-medication, depression, and suicidal ideation [[8], [9], [10], [11], [12], [13], [14], [15]].

WSH generates significant business costs, including increased absenteeism, turnover, use of health services, legal costs, decreased motivation, productivity, and organizational reputation [5,16]. Although the research is still exploratory, so conceptual models to explain its mechanisms are lacking, the International Center for Research on Women [17] proposed a promising model to explain the business costs of WSH, which shows that these costs come from both the harassed and the harassers and tend to increase as WSH continues or goes unpunished. According to the model, victims may experience damage in various physical and emotional aspects, including increased stress, distraction, physical ailments, reduced job satisfaction, and increased intention to desert. This can decrease individual productivity and team performance and generate transfer costs to move the harassed person to another team or department and retention costs, due to the harassed person resigning. The model also shows that perpetrators generate business costs since their productivity is also diminished, there are transfer costs to transfer the harasser to another team or department, and, in some cases, there are turnover costs due to the final dismissal of the aggressor. Added to these are legal expenses, including court judgments for damages. Moreover, an organization may experience reputational costs, and the perceived value of its brand may decrease [18], which makes talent acquisition more difficult.

This model is parsimonious and uses the mechanism of damage or morbidity to explain the decrease in productivity in the case of the victims. However, some other mechanisms and agents could explain the additional productive costs: 1. Regarding the mechanisms, many counterproductive behaviors can also be affected by WSH. For example, the desire to leave the company or the will to sabotage or reduce the quality of work can emerge because of WSH, which do not require morbidity as an explanatory mechanism. 2. As for the agents, the business costs can also originate from the colleagues who attest to WSH. Indeed, when harassed women do not formally report WSH but ask their colleagues for help [19], likely, these colleagues are also affected by the time they dedicate to assisting victims or as a negative consequence of this vicarious experience. Despite not being direct victims, the vicarious PTSD of WSH witnesses could create invisible business costs. However, the conceptual model and empirical proof of these hypotheses are nonexistent, so this constitutes a knowledge gap that needs to be addressed.

Another important line of research on WSH is its prevention. Therefore, many studies have tried to identify the risk factors that increase the probability of its occurrence. Regarding companies, a majority male-dominated population and an organizational climate that communicates tolerance of sexual harassment (for example, by having leadership that does not take complaints seriously, does not punish perpetrators, or does not protect against whistleblower retaliation), are the two environmental characteristics most associated with high rates of WSH [20]. For this reason, strategies have focused primarily on the promotion of gender equality, the establishment of protocols and cognitive training, and a cultural change against sexual harassment. However, few studies have analyzed the influence of managerial style on the prevalence of WSH, and no studies have focused on its business costs. To what extent does management style play a preventive or risk role in WSH? We hypothesize that a type of leadership focused on fairness, respect, inclusion, and equality, which we call “equitable management”, has a significant preventive effect on the prevalence of WSH and its costs.

2. Proposed model

Next, we present a conceptual model that explains the preventive role of the equitable management of workplace sexual harassment and its business costs and a new mediating mechanism that explains how WSH translates into lost labor productivity. This model applies to sexually harassed women and their colleagues who witness it.

The model is based on three assumptions: mechanism, agent, and predictor.

• •Mechanism. WSH increases tardiness, absenteeism, and presenteeism at work. In addition, it increases labor incidents. One mechanism that explains this effect is physical or emotional morbidity, to the extent that women's health decreases, affecting their work capacities. However, many types of sexual harassment are too low an intensity to cause significant physical or emotional damage, so it can be erroneously assumed that they do not significantly impact work productivity since work capacities are unaffected. Previous studies found that WSH also increases other types of work behaviors, not necessarily linked to health in the strictest sense, but rather to well-being, such as counterproductive behaviors or those related to work incivility [21].

  Counterproductive behaviors are voluntary behaviors that violate organizational norms, affecting an organization's and its personnel's well-being [22,23]. This concept is quite broad and usually also includes harassment of others, such as sexual harassment, production deviation (decreasing performance or spoiling quality on purpose), sabotage (deliberately damaging company property or its reputation), theft, and withdrawal (absenteeism, tardiness, or leaving early). For the present research, we focus on two of these dimensions: sabotage and production deviation.

  We hypothesize that WSH can activate a series of responses in female workers. The most immediate is flight, expressed through the desire to desert. However, another is confrontation, made impossible by the difference in gender power with the harassers, which can translate into counterproductive behaviors such as sabotage or silent resignation (voluntarily lowering performance). Both flight and confrontation are frustrated by inequitable labor conditions: women cannot lose or risk their jobs by directly confronting the harasser. This impossibility can later manifest itself in indicators of decreased productivity, such as tardiness, absenteeism, presenteeism, and labor incidents related to the quality of their work.

  • •
    Agent. The costs of WSH likely come not only from the harassed women but also from their co-workers who witness WSH. In this regard [[24], [25], [26]], have found that colleagues who witness WSH are likelier to suffer emotional and psychological consequences than their non-witness counterparts. These consequences can affect productivity, and they can also trigger counterproductive behaviors. Indeed, as in victims, the causal hypothesis in witnesses can be extended beyond the explanatory mechanism of the damage. Counterproductive behaviors can also be used to explain the decrease in labor productivity. Thus, this vicarious circuit can be activated in women who are not victims but witness it when assisting their harassed co-workers. As is often the case, the main perpetrators are other co-workers or bosses; therefore, witnessing colleagues also know the perpetrators and can be part of the same work team. This may trigger flight behavior (increasing the desire to desert) to reduce the risk of potential WSH. Alternatively, it may promote indirect confrontational behavior through sabotage or productivity deviance to restore fairness and solidarity with the harassed colleague. Research on labor incivility proves this hypothesis is possible [27,28].
  • •
    Predictor. Organizational justice can reduce counterproductive behaviors [29] and the prevalence of WSH by discouraging perpetrators [30]. Similarly, Rubino et al. [31] found that sexual harassment is less frequent and, when it occurs, has a lesser impact when greater organizational justice is perceived. Robotham and Cortina [32], Banner et al. [33], and Lim and Cortina [34] found that a climate of respect is a protective factor against workplace sexual harassment. Perry et al. [35] proposed that inclusive leadership can promote an environment of respect, making WSH less likely. Inclusive leadership fosters employee uniqueness (for example, by promoting diversity), strengthens membership within a team, and shows appreciation by recognizing efforts and contributions [36,37]. Conversely, Lee [38] found that passive leadership increases the probability of WSH by not addressing the issue of social tolerance. Research on toxic leadership (abusive, despotic, narcissistic, and tyrannical) indicated that it decreases motivation, dedication, and job satisfaction, and increases desertion, the perception of injustice, and counterproductive behaviors [39]. Liang et al. [40] found that abusive leadership tends to activate feelings of injustice in staff and encourage vengeful behavior to restore injustice. A meta-analysis by Park et al. [41] found that abusive supervision can increase productive deviations, understood as those behaviors that violate organizational norms, and threaten organizations' integrity and that of their members. Conversely, leaders perceived as constructive, ethical, fair, and participatory have an inhibitory influence on uncivil behavior at work [[42], [43], [44]].

All these indicators account for the enormous predictive role, which we will call “equitable”, that fair leadership has in preventing WSH and mitigating its consequences. Indeed, equitable relationships in organizations develop in a context where equality and respect for human rights are guaranteed and demanded without discrimination or abuse. Equitable management is the frequent behavior of a manager toward their staff, characterized by rationality, impartiality, inclusion, and respect. Vara-Horna [45] postulated three essential dimensions of the concept: 1. openness (when managers consider the opinions and ideas of their team and are open to new visions of the problems, facilitating more rational decision-making with more information); 2. inclusion (when managers manage their team fairly, with appropriate rewards for ability and performance); and 3. respect (when managers treat and promote dignity and equality without abuse or aggressive behavior within their workgroup).

Thus, based on these three assumptions, we propose a model that states that both victims and witnesses can lose labor productivity directly or indirectly through increased counterproductive behaviors, such as the intention to leave, sabotage, and productive deviation (Fig. 1). The model also suggests that WSH is less prevalent within inequitable management contexts (direct effect), and, in existing cases, there will be a decreased loss of labor productivity (indirect effect).

Fig. 1.

Proposed conceptual model of how equitable management can reduce workplace sexual harassment and its labor productivity costs, for both victims and witnesses.

3. Materials and methods

We have adopted an observational study methodology grounded in surveys to probe our conceptual model, relying on theoretical frameworks to infer causality. Although experimental design studies are often recognized as the standard in determining causal relationships, their use in the social sciences is only sometimes feasible or ethically defensible. Under certain circumstances, controlled experiments are hindered by pragmatic and ethical constraints. Consequently, while observational studies cannot unequivocally establish causality, they are capable of evidencing robust and consistent correlations between variables. While inconclusive, these correlations can strongly indicate a causal relationship. This indication is further strengthened when combined with well-founded theories and other types of evidence. In the context of our research on the direct and indirect effects of workplace sexual harassment on the productivity of victims and witnesses, an observational study is particularly appropriate. This approach lets us collect data realistically and assess correlations within an authentic work environment. Moreover, it allows us to respect ethical considerations, as it would be inappropriate, not to mention unfeasible, to design a controlled experiment involving workplace sexual harassment.

3.1. Participants

Our study was conducted within the private sector context in Lima, Peru, where an estimated 73,000 large, medium, and small private companies operate [46]. We chose the companies from Lima, for various reasons. Few investigations have measured the prevalence of workplace sexual harassment in Peruvian companies. GenderLab reported in 2021 that 34 % of female workers had been sexually harassed [47]. As per the data from the National Superintendency of Labor Inspection (Sunafil), a subsidiary body of the Ministry of Labor and Employment Promotion (MTPE), there were 5059 inspections pertinent to workplace sexual harassment carried out from 2020 to 2023. A significant escalation was observed in 2022, with the number of investigations swelling by 26 % compared to the preceding year. Predominantly, women employees in the private sector came forward, contributing to 91.29 % of the total complaints lodged. Regarding geographical distribution, the region of Metropolitan Lima was at the forefront, leading the charge in sexual harassment investigations from 2020 to 2023, accounting for 2827 cases.

From a directory of 1720 accessible companies, we extended participation invitations via email and telephone outreach. A total of 37 companies consented to engage in our research. Subsequently, we administered our survey exclusively to female employees within these corporations, yielding 913 responses. Following a rigorous data cleaning process, eliminating 86 surveys due to incomplete or invalid responses, our final sample constituted 827 women.

We applied questionnaires after coordinating with the authorities of each company. The applications were virtual or in-person, depending on the conditions warranted. Some decision criteria were the type of work they carry out, the number of company offices or their geographical dispersion, and the staff's level of education. Each company previously publicized the surveys, which reported that the Spanish Agency for International Development Cooperation and the Lima Chamber of Commerce were conducting independent research on the quality of working life. Then, they decided whether to participate. After the socialization campaign, the surveys were applied—in person or virtually—emphasizing their confidentiality and anonymity, guaranteeing total secrecy, and the company's authorities would only access a global report. Indeed, this present study follows a design that complies with the principles of ethics established in the Belmont Report (1979): respect, beneficence, and justice. Regarding the first principle, all participants were informed of the objective and nature of the study and guaranteed informed consent. Regarding non-maleficence, the research sought to preserve the personal safety of the respondents, giving it priority over information, through anonymous and confidential surveys and without individual access for the participating companies. The companies accessed a global report without identifying who responded in each case. Regarding the third principle of justice, the research results - globally – were returned to the organizations with specific recommendations and dissemination events.

The participants have an average age of 36.50 (SD = 9.81), ranging from 19 to 74 years. Regarding marital status, 52.9 % are single, 38.1 % are married or cohabiting, 8.3 % are divorced or separated, and 0.6 % are widows. In addition, 47.9 % of respondents have children. Regarding the educational level, the majority had completed higher education (54.7 %), while others had completed postgraduate studies (12.8 %), technical studies (21 %), or primary education (2.3 %) or had incomplete university studies (9.2 %).

The majority work as administrative personnel (46 %) or in sales, customer service, or maintenance services (37.6 %), 12.8 % are command personnel (manager, supervisor, or boss), and 3.5 % carry out manufacturing activities (elaboration of products, storage, or distribution). As for the hiring modality, 56.2 % have an indefinite contract, 37.2 % have a temporary contract, 2.1 % have employment by location of services, 1.8 % have a labor training agreement (pre-professional practices), and 2.7 % have other employment circumstances. Moreover, 90.8 % have a full-time contract (40 h or more per week). Regarding seniority, 22 % have worked for less than a year, 18.4 % have worked between one and two years, 19.8 % have worked between three and five years, and 39.8 % have worked more than five years.

Regarding the characteristics of their direct boss, 46.4 % report that they are a man, and 53.6 % report that they are a woman. In addition, 38 % report that this person has been their boss for less than a year, 28 % report that this person has been their boss between 1 and 2 years, and 34 % report that this person has been their boss for more than a year. This immediate boss oversees an average of 13.87 (SD = 25.2) people, ranging from 1 to 400 employees.

Regarding work location, 50.8 % work exclusively at the company's facilities, 42.8 % alternate between the office and teleworking for a few days, and 7.4 % exclusively telework. In addition, 64 % have had this situation for over two years, 12.9 % have had it for a year, and 23.1 % have had it for less than a year.

3.2. Measure

The research was carried out by applying an anonymous and confidential self-report questionnaire. The questionnaire design attempted to balance as many questions as possible to obtain as much information as possible with the women's time and willingness to complete the questionnaire. Thus, a questionnaire of 67 questions was achieved with the following variables:

Demographic and employment information. The questionnaire inquired about work activity, type of contract, hours worked per week, job seniority and remote work conditions, the gender of the person in charge, and the number of personnel under their command. It also recorded demographic variables such as age, number of children, educational level, and marital status.

Equitable management. The reflective scale of 10 items measured the frequency of management behaviors based on justice, respect, and inclusion. The women surveyed responded by evaluating their immediate boss or the commanding staff to whom they report. Each item had six response options ranging from never to always. The global scale consisted of the average of the ten items. Three classification groups are created based on the average score: daily GE (5.0 and higher), intermittent GE (from 4.0 to 4.99), and absent GE (3.99 and lower). This scale is based on the Inequitable Management Patterns Scale developed by Vara-Horna [45] and validated in Bolivia [48]. Initially, the scale had good internal consistency (Alpha = .906) and construct validity (AVE = 0.552). In this research, the values are equally satisfactory (Alpha = 0.847 and AVE = 0.614).

Workplace sexual harassment. The reflective scale of 10 items was based on the Sexual Experiences Questionnaire-Workplace (SEQ-W) by Fitzgerald et al. [49], which measures the level of gender harassment, unwanted sexual contact, and sexual coercion in women perpetrated by colleagues, clients, or superiors within organizations. The questions consider the last 12 months, with seven response options (never; 1 time; 2 times; between 3 and 5 times; 6 and 10 times; between 11 and 19 times, and more than 20 times). This brief scale has been adapted by Vara-Horna [48] in Bolivian companies, showing good reliability indicators (Alpha = .871) and construct validity (AVE = 0.542). In this research, the values are equally satisfactory (Alpha = 0.762, and AVE = 0.672).

Sabotage and deviance of production. The formative scale of six items was based on Spector et al.‘s scale of counterproductive behaviors [49]. The original scale evaluated five dimensions: sabotage, withdrawal, deviance of production, theft, and abuse toward the rest of the staff. Only two were chosen: sabotage (3 items) and production deviance (3 items). Initially, these scales had a low level of internal consistency (Alpha = 0.55 and 0.63, respectively) [50], which we suspect is not a construct but a compound. In this research, the values are similar (Alpha = 0.545 and AVE = 0.547).

Intention of labor leave. The reflective scale of three items registered the desires and behaviors related to leaving work. The scale was based on the proposals of Nielsen, Bergheim, and Eid [51] and Mobley, Horner, and Hollingsworth [52]. The original scale showed acceptable reliability (Alpha = 0.70). In this research, the values were more satisfactory (Alpha = 0.862 and AVE = 0.790).

Presenteeism. The reflective scale of six items recorded the level of distraction and job exhaustion for multiple reasons. Response options ranged from never, one day, two days, 3–5 days, 6–10 days, 11–20 days, and more than 20 days. For the calculation of days lost, the lower limit of each option was taken as a base. This scale was based on the one developed by Vara-Horna [53] and validated by Duvvury et al. [54] to measure work exhaustion and distraction due to domestic violence against women. It showed good reliability and validity indicators in business samples from Peru, Bolivia, Ecuador, and Paraguay [48,53,55]. The most recent use showed good internal consistency (Alpha = .908) and construct validity (AVE = 0.686) [47]. In this research, the values were equally satisfactory (Alpha = 0.939 and AVE = 0.766).

Absenteeism and tardiness. The formative scale of 8 items recorded the number of absences from work according to various reasons related to personal or family health (2 items), legal aspects to attend to (1 item), or aversive labor relations (2 items). Tardiness recorded the times they have been late for work based on the delay time (less than 1 h, between 1 and 2 h, or more than 2 h). Response options ranged from never, one day, two days, 3–5 days, 6–10 days, and more than ten days. For the calculation of days lost, the lower limit of each option was taken as a base. These scales were based on those developed by Vara-Horna [53] and validated by Duvvury et al. [54] to measure absenteeism and tardiness due to domestic violence against women. As a formative scale, its internal consistency was also low (Alpha = 0.533).

Labor incidents. The reflective scale of 4 items recorded the number of labor problems related to the quality of work and the labor relations affected by that quality. Response options varied from never, one time, two times, 3 to 5 times, 6 to 10 times, and more than 10 times. For the calculation of incidents, the lower limit of each option was taken as a base. This scale was initially developed by Vara-Horna [53] and showed good indicators of reliability (Alpha = .824) and construct validity (AVE = 0.658) [48]. In this research, the values were equally satisfactory (Alpha = 0.793 and AVE = 0.62).

Witness workplace sexual harassment. This formative scale (five items) recorded the amount of time allocated to assist colleagues who are victims of WSH and the number of days lost due to worry, tardiness, or absenteeism directly caused by having witnessed the harassment. As a formative scale, its internal consistency was also low (Alpha = 0.533).

Lost labor productivity. This second-order reflective scale was based on the linear combination of the scales for presenteeism, absenteeism, tardiness, and work incidents. The factorial structure of this second-order scale had an acceptable fit (X2 = 343.162, df = 111, p < .001; CFI = 0.96; TLI = 0.951; RMSEA = 0.053, p = .182; SRMR = 0.04). The overall reliability of the scale was good (Alpha = 0.901).

As shown in Table 1, the scales are related parsimoniously and with a theoretical sense between them. As expected, equitable management maintains an inverse relationship with the other variables; furthermore, the moderate relationships between them show no discriminant validity problems. However, the different scales of measurement and distribution of the scales will require some adjustments and transformations in the analyses.

Table 1.

Descriptive statistics, reliability, validity, and correlation matrix between scales.

	1	2	3	4	5	6	7	8	9
1. Equitable management	1
2. Workplace sexual harassment	−0.289	1
3. Intention to leave	−0.376	0.318	1
4. Presenteeism	−0.123	0.234	0.385	1
5. Work incidents	−0.269	0.266	0.403	0.577	1
6. Lost labor productivity (2 order)	−0.172	0.264	0.401	0.805	0.580	1
7. Witness WSH (c)	−0.174	0.166	0.161	0.091	0.189	0.118	1
8. Sabotage and deviance production (c)	−0.217	0.165	0.301	0.285	0.305	0.303	0.098	1
9. Absenteeism and tardiness (c)	−0.110	0.167	0.309	0.412	0.360	0.774	0.073	0.224	1
Number of items	10	10	3	6	4	18	5	6	8
Dimensions	3	3	1	1	2	4	1	2	2
Average variance extracted (AVE)	0.614	0.672	0.790	0.766	0.620	0.673	0.707	0.547	0.610
Reliability (Cronbach's Alpha)	0.847	0.762	0.862	0.939	0.793	0.901	0.533	0.545	0.533
Means	5.081	11.585	6.242	16.658	6.326	5.491	5.705	6.481	12.639
Standard deviation	0.81	3.544	3.254	8.471	3.155	5.459	1.828	1.383	4.095
Kurtosis	1.758	13.621	0.352	0.045	5.674	5.47	11.411	21.379	2.061
Skewness	−1.229	3.37	0.974	0.777	2.062	2.024	3.039	4.149	1.197

Notes: Spearman's Rho. All are statistically significant relationships, except for 7 and 9. The lost labor productivity scale is the linear combination of presenteeism, absenteeism, tardiness, and work incidents. (c): Compound or Formative variables. In reflective models, items are expected to be interchangeable and highly correlated, reflecting the same underlying construct. In such cases, a high Cronbach's Alpha is anticipated to indicate internal consistency. On the other hand, in formative models, items are seen as causes of the construct and may not be correlated with each other, as they might represent unique facets of the construct. Hence, a high Cronbach's Alpha is neither necessary nor expected in formative models and could even suggest redundancy among the items [56].

3.3. Analysis

The data were tabulated and processed using the statistical packages SPSS, Lavaan (R), and Stata 17.

In addition to descriptive statistics, categorical associations (X2) and comparisons of means (F Anova), linear modeling equations were used. We used generalized least squares with robust error estimation to control for heteroskedasticity problems, distribution differences, and skewness [57]. Natural logarithms of base ex were used to eliminate the effect of the units of each scale on the coefficients. On the other hand, the effect of some potentially confounding variables was controlled in the equations. These were chosen based on their relationship with the study scales. Table 2 shows the correlation matrix between them. Having a male boss or being younger increases the risk of WSH. Low income, being younger, or not having children are associated with a higher intention to leave the job. On the other hand, higher income and education are associated with fewer counterproductive behaviors such as sabotage or production deviance.

Table 2.

Correlations between demographic and occupational variables and covariates.

	Witness WSH	WSH	Intention to Leave	Sabotage and D. Produc.	Absenteeism	Presenteeism	Work Incidents	Equitable Management
Income	−.035	.003	−.130***	−.129***	.007	.055	−.039	.129***
Boss malea	.081	.111**	−.004	.002	−.021	.013	−.014	−.047
Have childrena	.009	−.081	−.172***	.003	.069	−.153***	−.016	−.048
Age	.012	−.121**	−.158***	−.073	.019	−.110**	−.042	−.054
Education level	−.002	−.001	−.080*	−.080*	−.037	.114***	.021	.032
Time working	.001	.020	.007	.007	.147***	.037	.041	−.083*
Work hours per week	.061	.003	−.007	−.007	−.001	−.005	.070	−.128***
Indefinite-term contracta	.016	−.070	−.069	−.039	.092*	−.002	−.004	−.016

Notes: Rho Spearman. ***p < .001, **p < .01, and * p < .05.

^aDummy variable. In these cases, biserial point correlations were used.

Through the GSEM package of Stata 17 [58], structural equations of covariance (SEM) were used to test the existence of significant, direct, and indirect relationships between the variables proposed in the conceptual model. SEM is useful for determining how independent variables influence dependent variables through mediating variables. Thus, it is assumed that the relationship between the independent and dependent variables is direct and indirect [59,60]. In this case, we used maximum likelihood estimators to identify the standardized Beta coefficients' precision and determine the statistical significance of the hypothesis test; we used robust errors instead of standard errors. This technique estimates the standard error by correcting for heteroscedasticity, which makes it possible to calculate the Z distribution and the p-values of the path coefficients. These are considered significant in cases where p < .05 and the Z score is greater than the critical value (1.96 at the 5 % significance level). Potentially confounding covariates are included in the model as controls.

Calculation of lost days. We used the estimation algorithms for annual costs of presenteeism and work incidents designed by Vara-Horna and internationally validated by Duvvury et al. [54] to calculate the days lost due to sexual harassment at work. For this research, the number of days lost in work productivity during the last 12 months due to any cause was computed using the following formula: days of work productivity lost = [absenteeism ((average of (a1 + a2)) + a3 + (average of (a4 + a5)))] + [tardiness (ta1 * 0.125 + ta2 * 0.25 + ta3 * 0.375)] + [work incidents ((average of (il1 + il2 + il3 + il4)) * 0.5)] + [presenteeism (average of (p1 + p2 + p3 + p4 + p5 + p6))].

The per capita estimate of days lost due to WSH alone was obtained by the difference between the women who were sexually harassed and those who did not report harassment. A regression analysis was performed with demographic and occupational covariates to adjust for the differences between the two groups. The same procedure was carried out to determine the days lost for witnessing WSH.

The formula to obtain the total number of days lost by the company consisted of multiplying—separately—the number of days lost due to WSH by the number of women affected in each case (corresponding to the percentage of prevalence). Productivity costs were calculated by converting the number of lost days to lost labor (total lost days/288 days = labor force).

4. Results

4.1. WSH prevalence

Table 3 shows the prevalence data of sexual harassment in the last 12 months. In total, 33.5 % of women report being sexually harassed at work during the last 12 months, with an average of 6.65 incidents (SD = 10.88). According to Dimension, 29 % suffered gender harassment, 17.6 % suffered unwanted sexual attention, and 1.9 % suffered sexual coercion. Considering only the group of harassed women (33.5 %), 86.5 % suffered gender harassment, 52.6 % suffered unwanted sexual attention, and 5.7 % suffered sexual coercion.

Table 3.

Workplace Sexual Harassment Prevalence in the past 12 months.

	Percentage
Gender harassment Dimension (mean incidents = 5.48, S.D. = 9.42)	29.0
At work, they tell stories with sexual content or make offensive jokes toward women.	18.5
Women are treated differently at work just because they are women (e.g., they are mistreated, looked down on, or ignored).	12.1
At work, sexist comments are made against women (e.g., suggesting that women are too emotional to be bosses or that they are incapable of doing a good job).	21.1
Unwanted sexual attention Dimension (mean incidents = 3.39, S.D. = 5.14)	17.6
Someone at work has made uncomfortable comments to you about your physical appearance or has pointed out their desire for you.	15.7
Someone at work has insisted on you by asking you out on dates or to go out for a drink or to eat, even though you said no.	7.9
Someone from work attempted to fondle or touch your private body parts (e.g., stroking your leg or neck or touching your chest, buttocks, etc.).	2.3
Sexual coercion Dimension (mean incidents = 2.00, S.D. = 1.22)	1.9
Someone at work has suggested that you receive some kind of reward (e.g., preferential treatment or job improvements) in exchange for accepting invitations or sexual encounters.	1.0
You have felt threatened with some kind of retaliation if you did not accept invitations or sexual encounters at work (e.g., hurting you in your evaluation, not renewing your contract, losing some benefit, etc.).	1.3
Someone at work has threatened you into accepting sexual invitations or encounters.	0.4
Someone at work tried to force you to have sex against your will.	0.3
Total (mean incidents = 6.65, S.D. = 10.88)	33.5

Overall, 76.1 % of sexually harassed women do not recognize themselves. Women aware of the harassment identify the perpetrator as their co-worker (12.2 %), boss or supervisor (4.8 %), client (3.5 %), security personnel (0.9 %), cleaning personnel (0.4 %), and other personnel (2.6 %). Of the women who acknowledged sexual harassment, only 16.9 % reported it to company authorities.

Regarding witnesses, 18.9 % supported a co-worker being sexually harassed during the last 12 months. Of this group, 74.4 % dedicated an average of 24.5 (S.D. = 129.18) minutes to assist. This support affected their productivity: 55 % were worried about what happened, 28.6 % were late or delayed due to assisting, and 15.8 % missed work to do so. On average, colleagues who witnessed WSH report losing 0.635 (S.D. = 0.89) days of work productivity as a direct consequence of witnessing and attending these events.

4.2. WSH effects

Indicators of absenteeism, presenteeism, and work incidents are more prevalent in women who have been sexually harassed. As shown in Table 4, the difference is statistically significant in 15 of 18 indicators. All aspects of presenteeism, including difficulty concentrating, distraction, concerns affecting work, slower work pace, tiredness or exhaustion, and lack of energy, were reported more frequently by those who had experienced WSH. The chi-square tests were significant at the p < .001 level, and the odds ratios indicate that those who have experienced WSH were more than twice as likely to report these issues. In the same way, those who had experienced WSH were more likely to report issues such as difficulty doing their job, problems with work quality, and being admonished for performance by both bosses and colleagues. All chi-square tests were significant at the p < .001 level. The odds ratios indicate that those who have experienced WSH are around twice as likely to experience these labor incidents. Finally, those who had experienced WSH were generally more likely to report absenteeism for health-related reasons, to avoid someone at work, and because they no longer felt well at work. These experiences had significant chi-square tests, except for failure to attend due to legal issues. Those who have experienced WSH were also more likely to report tardiness, with the most significant differences in being more than 2 h late.

Table 4.

Presenteeism, work incidents, absenteeism, tardiness, and presenteeism, according to the experience of sexual harassment at work (numbers shown are percentages).

	Total	WSH		X2	O.R.
		No	Yes
Presenteeism
Have you had difficulty concentrating at work?	71.2	66.3	81.3	16.837***	2.210
Have you been very distracted at work?	63.1	56.5	75.2	23.038***	2.341
Have you had concerns that have affected your work?	69.1	65.4	79.1	13.676***	2.004
Have you worked slower than usual?	68.9	63.2	80.0	20.205***	2.333
Have you been tired or exhausted at work?	80.8	77.0	88.6	13.338***	2.329
Have you been without energy to work?	74.4	68.9	82.6	14.705***	2.141
Labor incidents
Have you had difficulty doing your job?	46.8	41.4	60.0	21.320***	2.127
Have you had problems with the quality of your work?	40.9	34.6	55.0	26.294***	2.325
Has your boss admonished/claimed you for your performance?	28.2	24.6	39.1	15.620***	1.974
Have your colleagues admonished/claimed you for your performance?	15.5	11.3	24.0	18.241***	2.486
Absenteeism and Tardiness
You were absent because you were ill or had some ailment or health indisposition.	61.2	59.1	69.4	6.977**	1.537
You were absent to attend to your physical or mental health.	48.7	45.6	56.6	7.132**	1.554
You failed to attend to legal issues.	15.5	14.7	19.8	2.910	1.436
Missed work to avoid running into someone from work.	2.3	1.5	4.0	3.930*	2.654
Missed work because you no longer feel well at work.	9.2	6.1	16.2	17.986***	2.961
You did show up but less than 1 h late.	50.8	48.0	55.7	3.582	1.361
You did show up but 1–2 h late.	17.5	15.2	20.3	2.777	1.422
You did show up but were more than 2 h late.	8.9	6.4	14.4	11.766***	2.474

Notes: Statistically significant differences are ***p < .001, **p < .01, and * p < .05.

The same trend occurs for the indicators of the intention to leave work and the counterproductive behaviors of sabotage and deviation of production, which are more prevalent in sexually harassed women, (see Table 5). Respondents who had experienced WSH showed significantly higher percentages of intention to leave compared to those who had not, with highly significant Chi-square test results (p < .001). The Odds Ratios imply that those who have experienced WSH are about 3 times more likely to have intentions to leave their job. In concordance, the percentage of respondents who intentionally wasted company materials/supplies was significantly higher among those who had experienced WSH. However, there was no significant difference in the percentages of those who intentionally damaged company property or littered at the workplace. Finally, respondents who had experienced WSH were more likely to report doing their job incorrectly, working slowly, and not following work instructions. These differences were significant, particularly for working slowly and not following instructions.

Table 5.

Intention to leave work and counterproductive behaviors of sabotage and deviance of production, according to the experience of sexual harassment at work.

	Total (%)	WSH (%)		X2	O.R.
		No	Yes
Intention to leave
Have you thought about quitting your job?	58.7	49.5	74.8	40.279***	3.031
Have you been looking for a new job?	50.2	40.9	69.1	48.71***	3.233
If you could, would you leave your job?	59.0	49.2	77.4	50.021***	3.53
Sabotage
Have you intentionally wasted company materials/supplies?	4.9	3.3	7.8	6.908**	2.502
Have you intentionally damaged a piece of company equipment or property?	1.8	2.2	1.3	0.644	0.591
Have you littered or littered on purpose at your workplace?	1.0	0.9	1.3	0.279	1.497
Production deviance
Have you done your job wrong?	2.0	1.3	3.5	3.593	2.709
Have you worked slowly?	13.0	10.3	19.6	11.285***	2.117
Have you not followed the instructions given at work?	7.4	5.5	11.7	8.594**	2.298

Notes: Statistically significant differences are ***p < .001, **p < .01, and * p < .05.

The results showed a robust direct association between workplace sexual harassment and decreased labor productivity. These results are better visualized in Fig. 2, with the difference in standardized means between both groups of variables (at the scale level), showing that all of them were statistically significant, as their confidence intervals did not cross.

Fig. 2.

Differences in standardized averages of lost productivity scales, according to whether they are victims of sexual harassment at work. Notes: Lost days of work productivity is the linear combination between absenteeism, presenteeism, and work incidents. The bars show 95 % confidence intervals.

As observed in Table 6, being sexually harassed at work decreased labor productivity by 43.1 % and increased the intention to desert the company by 15.2 %. Despite the trend, no significant effect on counterproductive behaviors was found. On the other hand, witnessing WSH increased the intention to leave by 11.3 % and increased counterproductive behaviors by 39.6 %. Despite the trend, no significant effect was found on the decrease in labor productivity.

Table 6.

Effects of workplace sexual harassment on labor productivity, intention to leave, and counterproductive behaviors of sabotage and production deviation.

	Lost labor productivity			Intention to leave			Sabotage and deviance of productivity
	B	X2 Wald	Exp(B)	B	X2 Wald	Exp(B)	B	X2 Wald	Exp(B)
Intersection	0.846	7.500***	2.330	0.906	38.372***	2.475	0.754	2.259	2.125
WSH	0.413	24.797***	1.511	0.152	17.616***	1.165	0.225	3.231	1.253
WSH witnesses	0.152	2.353	1.164	0.113	7.527**	1.120	0.396	6.330**	1.486
Labor seniority	0.150	21.753***	1.161	0.025	2.277	1.025	−0.106	4.537*	0.899
Age	−0.010	3.129	0.990	−0.010	12.360***	0.990	0.007	0.659	1.007
Have children	−0.090	0.870	0.914	−0.002	0.002	0.998	0.187	1.879	1.206
Education level	0.036	0.656	1.037	0.026	1.414	1.027	−0.123	2.936	0.884
Income	0.004	0.015	1.004	−0.025	3.048	0.976	0.155	12.55***	1.167

Notes: Dependent variables in natural logarithms. Estimation by generalized least squares using robust errors. Statistically significant differences are ***p < .001, **p < .01, and * p < .05.

These main effects demonstrated the enormous influence that WSH has on organizational and productivity variables. However, the theoretical link between them may be hiding other indirect effects. In Table 7, this hypothesis is tested, and it was verified that counterproductive behaviors are essential mechanisms that explain the effects of WSH on labor productivity. This was particularly important in the case of female WSH witnesses, who saw their labor productivity affected mainly through full mediation via intention to leave and sabotage and deviance of production behaviors.

Table 7.

The mediating effect of counterproductive behaviors between WSH and lost work productivity.

	B	Z	Sig.	95 % Confidence Interval
				Lower	Upper
Direct effects
WSH → Productivity lost	0.092	2.020	0.043*	0.003	0.181
WSH witnesses → Productivity lost	0.056	1.273	0.203	−0.030	0.142
Indirect effects
WSH → Intention to leave → Productivity lost	0.079	4.402	<.001***	0.044	0.114
WSH → Sabotage and Deviance P. → Productivity lost	0.043	2.338	0.019*	0.007	0.080
WSH witnesses → Intention to leave → Productivity lost	0.028	2.198	0.028*	0.003	0.052
WSH witnesses → Sabotage and Deviance P. → Productivity lost	0.042	2.292	0.022*	0.006	0.078
Total effects
WSH → Productivity lost	0.214	4.342	<.001***	0.117	0.310
WSH witnesses → Productivity lost	0.125	2.801	0.005**	0.038	0.213

Notes: Robust standard errors and robust confidence intervals. Maximum likelihood estimator. Control covariates: age, male boss, educational level, have children, and job seniority. In parentheses (standard error).

4.3. Preventive role of equitable management

In general, as observed in Table 8, the indicators of equitable management in command personnel were widespread. Although fair and respectful treatment (84 %) and inclusive treatment (86.8 %) were very prevalent, empathic flexibility was less common (50 %). Each of the equitable management indicators was significantly correlated with workplace sexual harassment.

Table 8.

Equitable management indicators (%) and their relationship with workplace sexual harassment.

	N	AN	R	O	AA	A	Rho with WSHa
Openness (opposite of rigidity)
Does he/she listen to reasons, ideas, or advice even though they contradict his/her point of view?	7.4	8.9	14.4	21.0	19.2	29.0	−.163
Does he/she admit to making mistakes to his/her staff without blaming others?	12.0	15.2	18.8	13.7	16.1	24.2	−.137
Does he/she care about his/her staff's problems (try to understand and support them)?	4.7	6.2	11.4	16.1	17.9	43.8	−.224
Abuse of power * (opposite of fair and respectful treatment)
Does he/she reward his/her staff arbitrarily and unfairly? Does he/she have favorites?	55.0	16.6	12.8	6.3	4.4	4.9	.227
Is he/she selfish? Does he/she only think about himself/herself and not about the organization's benefit? Does he/she take credit from others?	69.8	16.2	6.7	4.4	1.5	1.4	.319
Does he/she treat his/her employees as if they were his/her servants or inferior to him/her?	79.3	10.6	5.5	2.8	1.2	0.6	.237
Does he/she communicate aggressively with his/her staff? Does he/she raise his/her voice in an intimidating, scolding, or yelling manner or ridicule them by making humiliating comments?	77.1	11.5	6.3	1.8	2.1	1.2	.227
Exclusion * (opposite of inclusive treatment)
Does he/she allow conflict between his/her collaborators, since he/she does not attend to complaints about insults or harassment between them?	80.1	10.9	5.4	1.7	0.7	1.2	.306
When he/she gets angry with someone on his/her team, does he/she push them aside or look down on them, ignore them, or exclude them?	73.3	14.7	7.3	2.2	1.4	1.1	.299
Does he/she limit the creativity and innovative ideas of his/her team?	63.9	17.6	12.3	4.0	1.4	0.8	.270

Notes: N = never, AN = almost never, R = rarely, O = often, AA = almost always, and A = always. * Reversible items to form the global scale. ^a Spearman's correlation. All relationships are significant at p < .001.

Openness: Managers who listen to differing ideas, admit to mistakes, and care about their staff's problems are considered open. A negative correlation with WSH suggests that these positive behaviors are associated with less sexual harassment at the workplace. The correlation is strongest for “Does he/she care about his/her staff's problems (try to understand and support them)?", indicating that emotional support and understanding from managers might be particularly impactful in preventing WSH.

Abuse of Power: This category covers negative behaviors such as favoritism, selfishness, treating employees as inferiors, and aggressive communication. These behaviors positively correlate with WSH, suggesting that they contribute to an environment where sexual harassment is more likely to occur. The strongest correlation is with “Is he/she selfish? Does he/she only think about himself/herself and not about the organization's benefit? Does he/she take credit from others?”, suggesting that a manager's self-centered behavior might be a key factor in creating an unhealthy work environment.

Exclusion: This category includes allowing conflicts, excluding team members, and limiting creativity. All these behaviors show positive correlations with WSH, indicating that they might foster conditions where sexual harassment can occur. The correlation is strongest for “Does he/she allow conflict between his/her collaborators, since he/she does not attend to complaints about insults or harassment between them?”, suggesting that a manager's negligence in handling conflicts can significantly contribute to an environment conducive to sexual harassment.

Overall, 65.4 % of respondents reported that their bosses practiced equitable daily management; 25 % reported intermittent equitable management, and 9.6 % reported the absence of equitable management. These three groups had important significant differences concerning the prevalence of workplace sexual harassment and labor productivity. As shown in Table 9, there was an inverse correlation between WSH and equitable management. The less frequent equitable management was the more prevalent workplace sexual harassment was in its three dimensions. The most substantial change was sexual coercion, which dropped from 7.5 % to 0.04 % when equitable management became routine.

Table 9.

Workplace sexual harassment, work productivity, intention to leave, and counterproductive behaviors according to the practice of equitable management in the organization.

	Equitable management			X2 or F
	Absent (9.6 %)	Intermittent (25 %)	Daily (65.4 %)
Percentages
Workplace sexual harassment	58.2	44.4	25.6	40.108***
WSH: gender harassment	53.7	39.8	21.2	42.968***
WSH: unwanted sexual attention	29.9	28.7	11.6	32.541***
WSH: sexual coercion	7.5	3.5	0.04	18.668***
WSH witnesses	29.0	21.5	16.4	7.272*
Intention to leave	85.7	82.7	63.5	31.667***
Sabotage and productivity deviance	30.0	32.4	12.8	37.535***
Means (S.D.)a
WSH incidents	9.87 (11.35)	7.43 (12.06)	5.05 (9.63)	14.938***
Lost labor productivity	8.82 (7.17)	5.94 (5.08)	4.94 (5.26)	16.13***
Sabotage and productivity deviance	1.27 (2.70)	0.782 (1.63)	0.328 (1.37)	13.191***
Intention to leave	2.32 (1.09)	2.03 (1.15)	1.44 (1.27)	25.906***

Notes: Statistically significant differences are ***p < .001, **p < .01, and * p < .05. ^a Between parentheses is the standard deviation.

When controlling the prevalence of WSH according to equitable management, labor productivity was also associated with the management style of the command personnel. Whether WSH existed or not, equitable management reduced labor productivity losses, the intention of job desertion, and the counterproductive behaviors of sabotage and productive deviation. However, when WSH already existed, equitable management decreased its intensity and labor productivity costs. Equitable management, directly and indirectly, affected labor productivity through the decrease in WSH. Finally, the impact of equitable management on productivity was exclusively indirect, through the reduction in WSH, counterproductive behaviors, and the intention to drop out.

Indeed, more equitable management is associated with less WSH (B = −0.303, Z = −5.51) and less intention to leave (B = −0.379, Z = −8.71) both directly and overall, (see Table 10). It also has a direct negative effect on sabotage and productivity deviance (B = −0.166, Z = −3.71), and an indirect negative effect on the same (B = −0.058, Z = −2.40), leading to a larger total negative effect (B = −0.225, Z = −4.39). However, it has an insignificant direct effect on lost labor productivity (B = −0.027, Z = 0.63) but a significant indirect negative effect (B = −0.289, Z = −7.19), leading to a significant total negative effect (B = −0.317, Z = −4.89). In the case of Workplace Sexual Harassment, its directly increases sabotage and productivity deviance (B = 0.117, Z = 2.28), intention to leave (B = 0.141, Z = 4.07), and lost labor productivity (B = 0.140, Z = 2.27). It also indirectly increases lost labor productivity (B = 0.090, Z = 3.50), leading to a larger total effect (B = 0.231, Z = 3.70). Finally, Witnessing WSH directly reduces the likelihood of WSH (B = −0.127, Z = −2.75) and increases sabotage and productivity deviance (B = 0.182, Z = 2.49), but these effects are relatively small. However, it has a non-significant direct effect on the intention to leave (B = 0.051, Z = 1.63), but indirectly increases the likelihood of lost labor productivity (B = 0.077, Z = 2.70), leading to a significant total effect (B = 0.187, Z = 1.97).

Table 10.

Direct, indirect, and total effects of equitable management on WSH and labor productivity.

	Beta standardized coefficients
	Direct effects	Z	Indirect effects	Z	Total effects	Z
Workplace sexual harassment (R2 = .114)
Equitable management	−0.303***	−5.51	–		−0.303***	−5.51
WSH witnesses (R2 = .016)
Equitable management	−0.127**	−2.75	–		−0.127**	−2.75
Sabotage and productivity deviance (R2 = .100)
Workplace sexual harassment	0.117*	2.28	–		0.117*	2.28
WSH witnesses	0.182*	2.49	–		0.182*	2.49
Equitable management	−0.166***	−3.71	−0.058*	−2.40	−0.225***	−4.39
Intention to leave (R2 = .239)
Workplace sexual harassment	0.141***	4.07	–		0.141***	4.07
WSH witnesses	0.051	1.63	–		0.051	1.63
Equitable management	−0.379***	−8.71	−0.049***	−3.65	−0.429***	−10.76
Lost labor productivity (R2 = .391)
Workplace sexual harassment	0.140*	2.27	0.090**	3.50	0.231***	3.70
Sabotage and productivity deviance	0.322***	3.85	–		0.322***	3.85
Intention to leave	0.372***	6.31	–		0.372***	6.31
WSH witnesses	0.109	0.72	0.077**	2.70	0.187**	1.97
Equitable management	−0.027	0.63	−0.289***	−7.19	−0.317***	−4.89

Notes: Structural equations of covariance, controlling for age, educational level, male boss, job seniority, and monthly income. Robust error correction for heterogeneous variances. Maximum likelihood estimation. SRMR = 0.042; R² global model = 0.294; CFI = 0.952. Statistically significant differences are ***p < .001, **p < .01, *p < .05.

As an example, which is based on the previous calculations, it can be affirmed that promoting equitable management in organizations means a reduction of 4.6 times the labor productivity costs caused by WSH, a reduction of 2.2 times the prevalence of WSH, and a reduction of 1.7 times the number of women witnessing WSH (see Table 11).

Table 11.

Hypothetical comparison of labor productivity costs caused by WSH, assuming three companies of 1000 workers each.

	Equitable Management Company
	Absent	Intermittent	Daily
Sexually harassed women (last 12 months)	582	444	256
Days lost per victim	4.70	1.80	2.20
Total days lost per year	2735.4	799.2	563.2
WSH-attesting women (past 12 months)	290	215	164
Days lost per witness	3.07	1.62	1.40
Total days lost per year	889.7	348.6	230.3
Total days lost per year	3625.1	1147.8	793.5
Workforce lost per yeara	12.6	4.0	2.8

^aEquivalence with full-time staff (288 days each), who would have stopped producing for 12 months.

5. Discussion

This is the first study demonstrating the impact of equitable management as a protective factor against workplace sexual harassment (WSH) and its consequences on labor productivity. Equitable management not only decreases the probability of the occurrence of WSH but also—if it exists—reduces its pernicious impact on productivity through various indirect effects. The research also shows, for the first time, the invisible costs of witnessing WSH in colleagues through hitherto undocumented indirect costs: the increase in counterproductive behaviors, such as sabotage or production deviance, with indirect repercussions on the reduction in labor productivity.

Businesses are organizations designed to maintain high levels of productivity. Variables such as absenteeism and tardiness are easy to identify, but other variables, such as presenteeism (exhaustion and distraction) or performance incidents, are more challenging to identify. To these must now be added the indirect costs of counterproductive behaviors, which are very significant. New empirical research in other samples and societies may provide further evidence of these relationships.

It was found that 1 in 3 female workers report WSH. These data agree with the prevalence of 34 % obtained in Peruvian workers [47] and 35 % obtained in Bolivian workers [48] and with previous studies carried out in various countries [61]. It was also shown that most harassed women do not recognize the abuse, consistent with previous studies [47]. This is worrying because it reflects sociocultural patterns that normalize gender-based harassment behaviors that facilitate their justification and discourage reporting.

It was also found that the client was the third-most-reported agent. This is worrying given that the rights of thousands of women who have demanding jobs in customer service and hotel service may be violated [[62], [63], [64], [65]], in addition to the invisible costs that can affect productivity. This is particularly sensitive in the case of female direct sales workers. Many sales performance organizations require demanding indicators, which increase the risk of exposure to workplace sexual harassment by clients, who, seeing themselves as having more bargaining power, can abuse their position. In addition, the same sales jobs require soft skills and charisma that, in macho contexts, can be misinterpreted as sexual interest. In the case of the hotel sector, the service culture oriented toward “guest sovereignty” can expose employees to different types of harassment [66]. Adding to this is managers' tolerance, who consider these behaviors “part of the job”, ignoring complaints and discouraging confrontations [63]. Companies must become aware of these facts and assume prevention mechanisms against harassment.

Here, we analyzed the impacts of WSH on productivity, but other aspects of the victims' personal lives may be affected. One of them is her personal and family relationships. We are particularly concerned about to what extent being a victim of WSH may be associated with being a victim of another type of gender violence. WSH is prevalent; however, it is not the most prevalent type of gender violence. Violence carried out by a partner or ex-partner is one of the most prevalent in the world, and in South America, Bolivia, Ecuador, and Peru have the highest rates. Thus, many women can be sexually harassed at work and assaulted by their partners or ex-partners at home. In this regard, very few investigations analyze the simultaneous frequency of both types of violence and, even more so, their joint impact on organizations. For example, in Bolivia, it was recently found that both types of violence are strongly associated and that their combined effects on productivity are much greater than their isolated effects [48]. Future research is pending to include the added costs in the productivity of simultaneous gender violence.

The age and gender of the boss are factors that increase the risk of WSH, as previously corroborated by other research [3,6,67]. However, independent of these demographic factors, the most significant risk factor is that command staff have inequitable management patterns. In other words, bosses with exclusive, rigid, and disrespectful management increase the risk of workplace sexual harassment. Indeed, labor productivity can be affected by various organizational factors; one of the main ones is the leadership style of the hierarchy at all levels, which mainly highlights fair and respectful labor relations without abuse of power, rigidity, or exclusion. By chance, we know that organizations that tolerate rigid, abusive, and exclusionary management patterns tend to affect the productivity of their staff, as they affect their staff's emotional health. However, a recent discovery in Bolivia found that this inequitable management can also increase the probability that high workplace sexual harassment and violence against women in intimate relationships exist or are tolerated [48]. With the present research, we verify that this hypothesis is correct.

These results—broadly—are familiar within the organizational sciences. Similar results come from research on leadership, including toxic leadership, abusive supervision, organizational justice, and workplace incivility. Our study is based on a prescriptive conception of what equity is in boss–staff relationships, which we define as equitable management, that tries to harmonize all these advances and apply them to the problem of workplace sexual harassment. Indeed, organizational research on sexual harassment in the workplace has been dormant in recent decades. However, the #MeToo and #TimesUp movements show that these experiences remain prevalent in organizations, making it necessary to reactivate academic interest [68]. In this regard, the test of the proposed conceptual model attempts to provide new evidence that explains how WSH causes business costs and suggests new predictors that contribute to prevention. In this sense, equitable management, as a concept, can promote strategies focused on the profitability of prevention, to the extent that its exercise has been shown to reduce WSH by 2.2 times, reduce the number of women who testify to WSH by 1.7 times, and reduce labor productivity costs caused by WSH by 4.6 times.

Many companies believe that complying with national legislation or internal regulations is enough to deal with complaints of sexual harassment. However, as found in this investigation and many others, very few harassed women report sexual harassment, so these actions become ineffective. Faced with this, organizations need alternative mechanisms. One should focus on leadership [69], specifically on effectively managing the climate of justice and respect that would discourage sexual harassment [31]. Previous studies investigated the effects of this equitable climate on the prevalence of sexual harassment through different models and theories, demonstrating its predictive value [30,70,71]. With the present investigation, this effect was confirmed by the leadership. It also extends to the reduction in productivity costs of the harassed and of the women who witness these harassments.

Finally, concerning the impacts on WSH witnesses, witnessing WSH causes costs and damages that translate into organizational losses. These results are according to “second-order sexual harassment”, where individuals assisting GBV survivors may also be assaulted [72]. Second-order harassment includes reprisals, injuries, slander, rumor spreading, mobbing, bullying, stigmatization, isolation, discrimination in a person's professional life or social life, or stereotyping a person, their families, friends, or colleagues. Although we have not directly measured this critical variable, the results show the need to expand its measurement for future research because cost categories that are not visible to organizations may be hidden.

5.1. Limitations

A paramount limitation of this study lies in using an observational design to infer causal relationships. It is vital to bear in mind the inherent constraints of correlational studies when interpreting the results. The most recognized constraint is that correlation does not imply causality, as unmeasured factors could potentially influence the variables under study, creating possible confounding effects. Likewise, correlational studies do not allow for controlling and manipulating independent variables, limiting the ability to establish causal relationships. This consideration is particularly pertinent in the study of workplace sexual harassment, a multifaceted and complex phenomenon influenced by many interconnected factors. However, observational studies are essential when experimental research is impractical due to ethical or practical limitations.

Furthermore, applying theory-guided analyses and using structural equation models to decompose direct and indirect effects while controlling for potential confounding variables lend robustness to the findings. This allows for understanding the causal relationship between workplace sexual harassment and productivity, identifying the pathways through which harassment may influence labor productivity. In this context, the results of the present investigation are parsimonious and consistent with theory. Although experimentation may be limited at the diagnostic stage, it could be the best approach during the implementation of solutions. It is important to note that experiment could be feasible when implementing solutions, such as evaluating changes in productivity and sexual harassment after companies have trained their management in Equitable management. This research approach would allow for determining whether interventions aimed at preventing workplace sexual harassment effectively result in productivity improvements, as theory predicts.

Although our research has addressed various sizes of companies and sectors, it is impossible to generalize the results for the case of microenterprises. The management and governance models of microenterprises require independent research. In addition, women who work in microenterprises have more precarious contracts and less safe working conditions, so the probability of women suffering sexual harassment and then not reporting it is much higher [73].

While our research comprehensively addresses workplace sexual harassment in private-sector companies, caution should be exercised when extrapolating these findings to public-sector organizations. The divergent organizational cultures, policies, structures, workforce compositions, and accountability measures inherent in public sector entities necessitate distinct investigation. 1. Public organizations typically demonstrate unique organizational cultures which may significantly influence employee attitudes and behaviors, including the prevalence and perceptions of WSH. 2. Public enterprises’ typically more bureaucratic nature, with more extensive human resource policies and procedures for handling WSH, may affect how such incidents are reported and managed, thus potentially altering perceived prevalence. 3. Structural and size differences between private and public organizations could also impact the dynamics of WSH. Public sector entities often exhibit more complex organizational structures and larger workforce sizes, factors which may necessitate individual investigation. Furthermore, the demographic composition of staff in public sector organizations may differ substantially in aspects such as gender, age, and educational background, potentially influencing the prevalence and perceptions of WSH. 4. The heightened scrutiny, transparency requirements, and accountability measures that public sector organizations are subject to may influence how WSH incidents are reported and managed. Given these considerations, generalizing our findings from private to public sector organizations could overlook significant contextual factors. Therefore, dedicated research is imperative to understand WSH in public sector organizations accurately.

This study acknowledges the limitation presented by a high non-response rate, which may potentially bias the outcomes. A substantial number of companies exhibited reluctance to participate in the study, and among those who did participate, a significant proportion did not complete the survey, rendering their data invalid for subsequent analysis. This high degree of non-response is not atypical of observational studies of this nature, yet it raises pertinent concerns regarding the potential implications of the research's findings. It is plausible that the non-response rate, in this case, could be interlinked with an aversion or resistance to addressing the issue of workplace sexual harassment [74]. If true, this could mean that the actual prevalence of sexual harassment and its economic impacts are underrepresented in our results. Given the sensitivity and the potential stigmatization associated with the subject matter, it is conceivable that those experiencing harassment or those within organizations where harassment is prevalent might be less likely to respond or complete the survey. Therefore, while our findings provide valuable insights into the costs of workplace sexual harassment, they may be conservative estimates. This underlines the necessity for future research to employ strategies that might encourage greater participation and completion rates, such as ensuring confidentiality, providing assurances about the use of data, and communicating the importance and societal benefit of such research. With more prominent and more representative samples, future studies might be able to provide a more comprehensive reflection of the true prevalence of sexual harassment and its impact on costs within organizations.

Finally, the study focused on WSH against women. However, the results could be extended to all victims of gender violence, provided that the conditions of each case are analyzed. For the LGBTQIA + population, it is necessary to consider broader aspects of harassment, such as gender hostility. Future research is pending to replicate these results in the LGBTQIA + population.

5.2. Implications

This study carries crucial implications from a practical and a policy standpoint. Providing pivotal evidence, it underscores the detrimental effects of workplace sexual harassment (WSH) not only from a humanitarian perspective but also from an economic viewpoint. The costs associated with WSH permeate beyond the direct victims, impacting the entire organization, including bystanders, and, as can be reasonably inferred, even the perpetrators when they are employees.

Our research unveils the high social cost resulting from the suffering imposed on female workers by WSH, a significant portion of which could be mitigated through effective prevention policies and practices. Simultaneously, the study illuminates the economic burden that WSH imposes on organizations. This suggests that the prevention of WSH benefits female workers by providing a safer and more equitable working environment and may also positively impact organizations by reducing the costs associated with harassment.

The pivotal role of equitable management as a preventive factor of WSH and its consequent business implications stands out as one of this study's most significant practical implications. Like any other management form, equitable management must be learned and integrated into the fabric of an organization. To this end, it becomes necessary to incorporate it as part of the curriculum at business schools, both at the undergraduate level and in executive training through extension courses. Furthermore, guided by the recommendations of the International Labor Organization and UN Women's “Addressing Violence and Harassment Against Women in the World of Work”, organizational culture and leadership can be steered towards equality through adequate training, the introduction of zero-tolerance management policies for violence, care mechanisms, and the prevention of violence and gender discrimination.

These revelations should catalyze organizations to adopt more robust preventative measures and ensure that work environments are respectful and free from harassment. Moreover, these findings offer a solid foundation for lawmakers and policymakers to promote labor regulations fostering equity and safety in the workplace.

5.3. Conclusions

In conclusion, this study illustrates the dual impacts of Workplace Sexual Harassment (WSH) on individual well-being and organizational productivity. The findings underscore the critical role of equitable management in mitigating the occurrence of WSH, highlighting its potential to not only prevent harassment but also alleviate its detrimental effects on productivity. A noteworthy revelation is the “invisible” costs incurred by bystanders of WSH, manifested in counterproductive behaviors and diminished productivity.

The study also draws attention to the insidious yet pervasive nature of WSH, with one in three female workers reporting harassment, a statistic echoed in multiple national contexts. The data also indicate the disturbing reality that most victims do not recognize the abuse they endure. This reflects deep-seated sociocultural patterns normalizing gender-based harassment, discouraging reporting, and potentially perpetuating the issue. Furthermore, the study reveals that clients, rather than colleagues or superiors, are often perpetrators of WSH, putting women in customer-facing roles at significant risk. This highlights the urgent need for organizations to recognize and address this issue.

Moreover, the study delves into the broader implications of WSH, including its potential association with other forms of gender violence and the cumulative effect on victims' productivity. The findings reinforce the importance of addressing WSH and violence in intimate relationships, which can compound the harm inflicted on women at work and at home.

The study further identifies the risk factors contributing to WSH, emphasizing the significance of leadership style and management patterns. While the age and gender of supervisors were found to be potential risk factors, the most potent predictor of WSH was inequitable management patterns, a factor also linked to decreased productivity. As such, the study provides robust empirical evidence advocating for the prioritization of equitable management and organizational justice in the fight against WSH. While compliance with legislation and internal regulations is essential, the findings suggest that organizational culture and leadership play a significant role in preventing harassment and mitigating its impacts.

The study concludes by highlighting the need for future research to address its limitations and expand its scope, particularly concerning the representation of different populations, such as those within the LGBTQIA + community, and different organizational structures, like microenterprises and public sector entities. This will further enrich our understanding of WSH and contribute to developing effective prevention strategies, fostering safer and more equitable workplaces.”

Funding

This research was financed by the European Union and the Spanish Agency for International Development Cooperation (AECID), through the management of the Lima Chamber of Commerce (CCL), within the framework of the project “Gender Violence against Women: Strengthen Prevention in the Private Sector”.

The content of this publication should not necessarily be interpreted as the opinions of the European Union or AECID.

Institutional review board statement

All The Belmont Report (1979) recommendations were followed, ensuring informed consent, non-maleficence, and fairness.

Informed consent statement

Informed consent was obtained from all subjects involved in the study.

Data availability statement

Data from the current study is available upon request to the corresponding author.

CRediT authorship contribution statement

Arístides A. Vara-Horna: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Software, Validation, Visualization, Writing – original draft. Alberto Díaz-Rosillo: Investigation, Methodology, Project administration, Writing – review & editing. Zaida Asencios-Gonzalez: Data curation, Investigation, Methodology, Software, Supervision, Writing – review & editing. Liliana Quipuzco-Chicata: Investigation, Methodology, Project administration, Resources, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: