Why sexual violence and harassment are not addressed: comparative insights from Cambodia's construction sector

The prevailing policy approach to addressing sexual violence and harassment (SVH) in the workplace rests on the assumption that formal regulatory and institutional frameworks create the conditions for reporting and redress. Recent evidence complicates this assumption, with researchers extensively documenting why victims do not report and what happens when they do (Cortina and Areguin, 2021), and examining how organisational contexts, structures and workplace conditions shape the likelihood that victims will report (Adikaram and Kailasapathy, 2022; Knox and Bohle, 2025). A study of eight Sub-Saharan African countries, for example, found no significant differences in reporting between organisations with and without formal policies, unless training had been provided to staff (Blumell and Mulupi, 2025), raising questions about what mediates between the existence of formal mechanisms and their operation in practice. While salient to all workplaces, these observations are especially pressing in the Global South, where enforcement of regulatory protections is generally weakest and reliance on the discretion of workplace actors greatest.

Far less attention has, however, been paid to the interpretive work performed by those who receive and process those complaints, or to the impact of this interpretative work on victims' decisions about whether or not to report. In any workplace, formal mechanisms are operationalised by the intermediaries through which complaints must first pass. Where there are no formal mechanisms, those same intermediaries constitute the sole arbiters of SVH complaints. In comparing how intermediaries interpret and respond to SVH in cement factories and construction sites in Cambodia, we address three questions. First, do these intermediaries recognise reported behaviours as SVH, and if so, do they accept that they constitute a legitimate workplace grievance? Second, where complaints are pursued, does the presence or absence of a formal grievance mechanism determine the course of action taken and the outcomes achieved? Third, how do intermediaries' responses, and the mechanisms available to them, shape workers' decisions about whether to report SVH in the first place?

The comparative design of this study provides a natural test of these propositions, allowing us to examine more and less formal workplaces in the same regulatory and cultural context. This comparison serves two purposes. By tracing intermediaries' interpretive work across the three decision points identified above, it renders visible a gatekeeping process that existing scholarship has largely overlooked, in the process drawing into question the widespread assumption that formal structures necessarily produce better reporting rates and outcomes. As its findings reveal, intermediaries in the cement factories were more likely to have access to at least some formal workplace mechanisms and more likely to have some rudimentary training on SVH. However, overall, the greater level of formality of the cement factories had little effect on intermediaries' responses to workplace SVH, which were strikingly similar to those on the construction sites.

Workplace SVH is typically defined as unwelcome sexual conduct that serves to intimidate or offend (Lewis and Henderson, 1994). It involves acts ranging from sexual comments, jokes and advances through to coercion and physical assault. They are interconnected and often escalate, with seemingly minor acts contributing to workplace cultures in which more serious violations are enabled and tolerated. SVH is not fundamentally about sex, but about gendered forms of power, both formal and informal (Acker, 1990; Rospenda et al., 1998). Power is also the fundamental element in determining how harmful conduct is interpreted and addressed (Martin, 2003). Although men experience SVH at work, workplace power structures tend to concentrate authority in the hands of men and license the abuse of women (Edelman and Cabrera, 2020).

In contexts characterised by weak accountability and limited enforcement, such conditions enable misconduct to persist (Bongiorno et al., 2020; Lucero et al., 2006; McDonald and Charlesworth, 2016). The prevailing approach to addressing workplace SVH rests on the assumption that regulatory protections produce workplace policy, which generates awareness and increases rates of reporting and redress (Bruschini et al., 2025). However, a growing body of evidence demonstrates that each link in this chain is weaker than commonly assumed. Even where legal frameworks are robust and workplaces highly formalised, legislation has no direct influence on the structural power relations that enable workplace violence (Chen, 2025). Survey data from Belgian accounting and law firms show that women reported elevated experiences of harassment despite firms' stated commitments to equity (Hardies, 2023). Sommerlad (2016) demonstrates that, despite extensive regulatory coverage, profit-driven law firms in England and Wales tolerate SVH as an expression of the demands of professional performance. In military contexts, rigid command structures and cultures valorising physical dominance generate intense pressure for victims to remain silent despite the availability of dedicated complaint channels (Lade et al., 2026). Since these diverse settings are all characterised by strong legal frameworks and, in most cases, mandatory workplace policies, it is clear that failure to address workplace SVH cannot be understood simply as a product of the absence of regulation, policy or training (cf. Giovannone and Lamberti, 2026).

If regulation, policy and training do not guarantee outcomes even in contexts with a strong record of enforcement, what determines whether formal mechanisms actually function? Existing research has documented failures of recognition where SVH is narrowly defined (Leskinen et al., 2011) or where it is understood to be private or interpersonal (Adikaram and Kailasapathy, 2022). The literature also suggests that response failures are ultimately shaped by organisational culture rather than by the availability of formal procedures (Knox and Bohle, 2025; McDonald et al., 2015). We integrate these insights into a three-stage gatekeeping model that captures the sequential process through which intermediaries – in our case, team leaders, union officials, subcontractors and security guards – engage with complaints (Figure 1). At the first decision point in our model, intermediaries must decide whether the reported behaviour constitutes SVH. At the second, they must decide whether the behaviour they recognise as problematic constitutes a workplace matter requiring a response. At the third, they must determine what course of action, if any, to take. This disaggregation is analytically important because it reveals that the failure of complaint systems is a cumulative process which not only affects the outcomes of specific complaints but also reduces the likelihood that complaints are made at all.

Scholars identify two broad sets of factors that shape intermediaries' decision-making at each of these points. The first of these consists of the structural conditions reflected in workplace power relations (Cortina and Areguin, 2021; Minnotte and Legerski, 2019). Hierarchical authority, discretionary decision-making and organisational climate influence not only who perpetrates SVH, and against whom, but also whether perpetrators go unpunished (Knapp, 2016). A review of 29 studies on the agricultural sector across the Global North and South identifies the informality of employment relationships, traditional concepts of masculinity and weak regulatory enforcement as key drivers of workplace SVH (Rocha et al., 2024). A systematic review of SVH in informal sector employment in low- and middle-income countries identifies consistent patterns whereby informal working conditions heighten women's vulnerability to SVH and foreclose the possibility of formal redress (Hardt et al., 2023). Critically, informality also operates through the discretionary space available to intermediaries within formal workplaces, as evidenced by the discussion below.

The second set of factors consists of cultural norms, which operate independently of the level of workplace formality and even structural power (Oosterom and Nazneen, 2023; Zulfiqar et al., 2025). McDonald et al. (2015) have argued that the framing of SVH as an individual problem rather than one with systemic causes and consequences has constrained the development of effective organisational responses. The same individualising logic operates at a lower level in the everyday interpretive work of the intermediaries. Norms that accept SVH or attach shame to those who experience it (cf. Ruebottom and Toubiana, 2024) influence targets' willingness to report. Nahar (2024), for example, identifies fear of retaliation, economic insecurity and cultural stigma as key deterrents to reporting in Bangladesh's garment sector. Equally, these factors influence intermediaries' recognition of inappropriate behaviour, their classification of it as private or institutional, and their willingness to act, as demonstrated by work in Sri Lanka, which has shown that intermediaries with access to formal complaint mechanisms still frequently classify SVH as a private or interpersonal matter falling outside their institutional domain (Adikaram and Kailasapathy, 2022).

The situation is demonstrably worse in hypermasculine workplace cultures where these norms become embedded through workplace habitus (Powell and Sang, 2015), and especially where hypermasculine workplace cultures with high levels of employment informality (French and Strachan, 2017; Galea et al., 2020). Survey evidence from South Korean construction sites confirms that women experience elevated levels of workplace violence (including SVH), compounded by their isolation within a male-dominated workforce (Park et al., 2022). Women in Canada's construction sector frequently leave their employer without reporting harassment, largely because it is normalised to the point where challenging it appears futile or dangerous (Hamel-Roy and Dumont-Lagacé, 2021). The transient nature of construction work and complex subcontracting arrangements exacerbate the situation (Pun and Lu, 2010), limiting opportunities to build the kind of trust with supervisors that might otherwise support reporting. Under these conditions, intermediaries occupy an especially consequential position since they are the point at which complaints (fail to) enter whatever infrastructure is in place.

At the international level, the International Labour Organization's Violence and Harassment Convention (C190), adopted in 2019, recognises the right of all workers to a world of work free from violence and harassment, including gender-based violence and harassment (ILO, 2019). The Convention takes a structural approach, acknowledging that violence and harassment are facilitated not only by individual behaviour but by discrimination, abuse of power relations and gendered social norms, and requires ratifying states to adopt comprehensive prevention strategies, enforce prohibitions and ensure access to remedy. Its scope is deliberately broad, covering the formal and informal economy, and extending to employer-provided accommodation and work-related travel. Critically, it also extends obligations beyond employers, requiring states to address violence and harassment involving third parties and to account for the role of public authorities and employers (and in some cases unions) in preventing workplace violence and harassment. This is particularly relevant in sectors characterised by complex subcontracting arrangements, where lines of employer responsibility are blurred.

Cambodia has not ratified C190, and its legal framework falls well short of the standards that C190 sets out. Its legal framework nevertheless does address SVH. SVH is referred to in Article 250 of the Criminal Code, which defines it as abuse of authority involving repeated pressure for sexual favours, thus effectively excluding peer-to-peer harassment and single incidents regardless of their severity. Situated within a section on the employment of women and children, Article 172 of the 1997 Labour Law proscribes all forms of “sexual violation (harassment)” in establishments employing minors or women, but does not provide any definition of what constitutes SVH, no standard of proof and no specification of employer obligations. Responsibility for monitoring and enforcing these provisions rests primarily with the Ministry of Labour and Vocational Training. In practice, however, inspections are rare and, when they occur, focus on physical hazards (MGML1-1). Indeed, none of the intermediaries interviewed were aware of any labour inspection that had targeted SVH.

It is also important to note that references to SVH in the legislative framework do not apply to all workers, nor to all workplace relationships. Article 3 of the 1997 Labour Law defines workers as those who have signed an employment contract in return for remuneration, systematically excluding many of those most exposed to SVH, even in formal workplaces. This provision has different consequences for the settings examined in this study. In the cement factories, workers directly hired by the company fall firmly within the scope of the Labour Law and are, in principle, covered by its provisions on SVH. Subcontracted workers in the same factories occupy a more ambiguous position: while they work within the physical premises of the factory, their employment relationship is with the subcontracting company, which may or may not have a written contract with individual workers. On construction sites, the situation is far worse. Most workers are engaged through verbal agreements with subcontractors and have no formal employment contract, placing them at even higher risk. The Criminal Code's definition of SVH as abuse of authority further disadvantages workers without contracts, since the informality of their employment relationships makes it difficult to establish the hierarchical relationship on which prosecution depends. Meanwhile, the absence of meaningful enforcement means that even the limited protections that exist on paper have little practical effect.

In Cambodia's cement factories, team leaders and subcontractors serve as workers' first point of contact, human resources departments and safety officers manage formal processes in the factories and enterprise unions attempt to represent workers' interests. These intermediaries exercise considerable discretion in deciding what constitutes a legitimate grievance but also whether complaints are addressed formally or informally or are dismissed out of hand, as their responses, in turn, shape victims' decisions about whether to report. Lead contractors are largely absent from Cambodia's construction sites. Team leaders and subcontractors provide daily oversight and, while unions have something of a presence, it is much more sporadic than in the cement factories. Various other intermediaries also play a role, including the security guards who police workers in some of the accommodation provided to construction workers by their employers.

In investigating the role of these intermediaries, this study adopts a comparative case study approach (Bartlett and Vavrus, 2017). Cambodia's cement factories and construction sites operate within the same national legal and cultural context but differ sharply in levels of workplace formality. As such, they provide a strong basis for determining whether formality determines how SVH complaints are handled. This research design enables analysis not only of similarities and differences in these settings but also of how the interpretive frameworks that shape complaint-handling operate across them. Data collection proceeded in overlapping phases between January 2022 and June 2025, a period during which the way that Cambodia's legal framework dealt with workplace SVH did not change and the sectoral conditions documented in this study remained largely constant. Before it began, interviewers and survey enumerators underwent specialised preparation on workplace violence and harassment, ethical protocols for research on sensitive topics, and procedures for maintaining participant safety and privacy. A specialist local counselling service was commissioned to support any respondents or researchers who needed it. Initial interviews with enterprise union officials and organisers from the Building and Wood Workers Trade Union Federation of Cambodia (BWTUC), the key national federation in the construction sector, were conducted to build contextual understanding and establish initial access, after which the team developed its own networks.

Qualitative and quantitative components served complementary analytical purposes in the study's mixed-methods design (Venkatesh et al., 2024). Quantitative data establish the prevalence and distribution of SVH experiences and reporting behaviour, while qualitative data illuminate the motivations, barriers and institutional dynamics that shape those patterns. As scholars of gender-based violence research have argued, studies addressing sensitive topics must employ methods that enable participants to recount their experiences on their own terms and that foster a sense of autonomy and of being believed (Romero Gutierrez et al., 2024). Our semi-structured interview format was designed with this principle in mind, using open-ended prompts that allowed participants to narrate their experiences at length while also enabling systematic comparison across respondents. The field design also incorporated field observations around workplaces, in employer-provided accommodation, and in social spaces frequented by workers, which helped contextualise formal interview and survey data within the everyday dynamics of the work environment (cf. Loosemore et al., 2015). The resulting dataset represents the largest body of primary data on SVH in Cambodia's construction sector ^[1].

The survey instrument was developed to encompass both sexual and non-sexual forms of violence and harassment in the workplace but also in employer-provided accommodation and during work-related travel, in line with the scope of C190. In the process of its development, the team referred to existing surveys on workplace violence and harassment and to insights from preliminary qualitative fieldwork ^[2]. The draft instrument was then refined through two rounds of feedback from Australian unionists and academics working on related topics, and two further rounds from local unionists, then translated into Khmer and piloted to ensure questions were culturally appropriate and clearly understood. The survey was administered using Qualtrics. Given that many respondents had limited functional literacy, and none had access to personal computing devices, enumerators read each question aloud and recorded responses on their phones or tablets.

Two hundred workers from each of three subsectors (cement factories, commercial construction sites and residential construction sites) were surveyed. The construction subsectors were merged for this particular study. Women were sampled in equal numbers to men in each setting, despite constituting a minority of the construction workforce, because they are disproportionately affected by SVH and their experiences were central to the study's research questions (Table 1). Gender-matching was used to ensure respondents felt comfortable discussing sensitive topics. Female workers were surveyed between March 2022 and March 2023 by a team of female Cambodian enumerators and male workers were surveyed between March and July 2023 by male Cambodian final-year university students. The sample is not statistically representative of Cambodia's construction workforce, as access was shaped by the dispersed nature of employment and the sector's heavy reliance on subcontracting ^[3].

In-depth semi-structured interviews averaging between 45 and 90 minutes were then conducted with 40 female and 20 male cement workers and 40 female and 40 male construction workers. The interview pool was drawn from survey respondents who expressed interest in sharing their perspectives in greater detail or who reported direct or indirect experiences of violence and harassment. Female workers were interviewed by the second author (a Khmer woman). Male workers were interviewed by a Khmer male researcher hired and trained specifically for this component, who was trained and supervised by the second author. Interviews explored workers' experiences of workplace violence and harassment, their decisions about whether and how to report, and their assessments of workplace grievance mechanisms. We were guided by the existing literature on workplace SVH in our identification of these elements (Cortina and Areguin, 2021). Particular attention was paid to how workers understand the role of intermediaries in these processes, and how intermediary responses influence workers' decisions to report or remain silent. All workers interviewed are referred to in the article using pseudonyms.

A total of 41 workplace-level intermediaries were interviewed (Table 2). These included eight company representatives (human resources staff, engineers, administrators and safety officers), six subcontractors, four security guards, 14 team leaders, six enterprise union officials and four BWTUC organisers. Although women held some of these intermediary roles, men accounted for the overwhelming majority.

The intermediary interviews were designed to capture the perspectives of actors who exercise institutional and relational power over workers, and who therefore play a central role in shaping how SVH is interpreted, addressed and reported, an approach informed by the literature, which highlights the importance of organisational actors, particularly human resource personnel and supervisors, in shaping reporting behaviour and organisational responses to harassment (McDonald et al., 2015; Cortina and Areguin, 2021). The interviews focused on these intermediaries' understandings of workplace violence and harassment, how cases are handled in practice, and how their responses influence workers' decisions to report or not to report ^[4].

For this particular study, survey data were analysed to capture the prevalence of different forms of SVH, reporting rates and reasons for non-reporting across the two settings, using Qualtrics. Interview data were transcribed in Khmer and then translated into English and then checked by the second author. The data were then analysed using iterative thematic coding techniques in Excel, which was chosen because of its ease of use and its accessibility in Cambodia. Initial codes were derived from the research questions and interview guides, then refined and supplemented as additional themes emerged through close reading of the transcripts. To ensure analytical consistency, sample data were coded independently by two researchers and the results were compared before the full coding commenced; the first author then spot-checked coded data against the original transcripts at several points throughout the coding process. For this study, the analysis focused on three areas: how intermediaries classified and categorised SVH incidents; the outcomes of complaints that were pursued; and workers' reasons for reporting or not reporting.

The data reveal a fundamental disconnect between how intermediaries and workers understood SVH. Across both cement factories and construction sites, intermediaries largely failed to recognise SVH as a legitimate workplace grievance and, even where they acknowledged that behaviours were problematic, they framed them as private matters to be managed by the individuals involved. The workers they supervised saw things differently: while many normalised less severe forms of SVH, those who experienced more serious behaviours understood them as harmful, yet rarely reported them. The few who did pursue complaints encountered dismissive responses, victim-blaming, or informal resolutions that left perpetrators in place. Those who chose not to report either cited feeling shame or fear – or had made a pragmatic assessment that reporting would create more problems than it would solve.

In both settings, SVH fell into an accountability gap. In the cement factories, management used subcontractors to insulate themselves from direct responsibility for workers' well-being. General grievance procedures were used to handle SVH cases but were not designed for them, and the factories' zero-conflict policies actively discouraged complaints. On the construction sites, the gap was even wider. Lead contractors eschewed responsibility entirely, leaving all responsibility for workers to their subcontractors, many of them small and unregistered. Within this institutional vacuum, intermediaries exercised even more discretion over how SVH complaints were handled than in the cement factories. As the evidence presented below demonstrates, however, intermediaries' interpretive frameworks were nevertheless relatively consistent, regardless of the level of workplace formality.

Across both cement factories and construction sites, company management had no SVH-specific policies or grievance procedures. In the cement factories, however, general grievance procedures were in place. When asked how SVH complaints would be handled, intermediaries in the factories consistently pointed to these existing channels as an appropriate means to deal with an SVH complaint. One HR officer confirmed that workers could “contact the administration office or their shop steward” if they experience SVH (MACHR1-1). A female team leader confirmed this process: “Both parties would be called in for an investigation. It depends on the victim, whether they want to resolve it quietly or escalate it” (FCGL2-1). On the construction sites, by contrast, no such channels existed. Subcontractors managed grievances on an entirely ad hoc basis, some confirming that they had received no guidance from lead contractors on how to handle any workplace dispute, let alone SVH (MXSU2-1, MXGL1-1).

Across all three cement factories, there was an assumption that, since a grievance mechanism existed, the absence of reported cases meant that SVH did not occur. According to a safety and health officer in one factory, there had been no cases of SVH at all (MASO1-1). A human resources officer from the same factory said, “Since I've been working here, no serious cases have been brought to our attention, and most cases we have seen have been resolved” (MAHR2-1). At a second factory, an administrator stated: “I don't think we have had any cases. We have security guards everywhere … so no one can do anything” (MBHR1-1). Notably, all three factories had policies under which all workers involved in any workplace conflict would be terminated. These policies did not distinguish between perpetrator and complainant, which both discouraged workers from approaching intermediaries in the first instance and signalled to intermediaries that SVH was not a matter they were expected to address. This contrasted starkly with the reporting of physical safety hazards, where workers could raise concerns without being treated as parties to a dispute and where at least one factory actively incentivised reporting through a team rewards program for injury-free periods.

The cement companies' heavy reliance on subcontractors compounded these problems. Only a small proportion of the workforce (typically those with technical skills or individuals employed during the initial stages of factory operations) was directly hired by the companies. One factory relied on around 15 subcontracting companies to supply both skilled and unskilled workers. This company claimed to manage factory-wide safety standards. As an administration officer noted, “Even though we use subcontractors, we still need to monitor and train the workers, especially on safety issues” (MAHR1-1). It was clear, however, that company staff in all three factories had little knowledge of how subcontracted workers were supervised or how grievances were handled within subcontracting firms. A subcontracting company in another factory confirmed that this assessment held true: “The factory doesn't want to be involved in anything related to workers. That's why they use subcontracting companies” (MBHR1-1). On the construction sites, lead contractors distanced themselves even more firmly from responsibility for workers' wellbeing. In one case, an administrator acknowledged that rapes had occurred on-site (FXMA1-1). She went on to explain that responsibility for addressing SVH sat with the subcontractor unless the victim was employed directly by the company. Subcontractors confirmed this assessment. One said, “The company doesn't really care about this issue” (MXSU2-1). Another described how outcomes depended entirely on subcontractors, saying “The company won't take responsibility for anything” (MXSU4-1). This applied even to physical injuries, where evidence and eyewitnesses are generally readily available. The situation was worse for SVH, which is generally far harder to substantiate.

While these accountability gaps operated through different mechanisms, they produced functionally equivalent outcomes whereby intermediary discretion went unchecked. In the cement factories, the absence of SVH-specific policies, combined with zero-conflict policies that treated complainants and perpetrators identically, sent an unambiguous signal that SVH was not a matter of great concern to management. On the construction sites, lead contractors' wholesale abrogation of responsibility made subcontractors the de facto authority over all workplace grievances, including those they had no capacity to handle. In both settings, the institutional vacuum not only left intermediaries to their own devices but actively shaped their understanding of what fell within their remit.

In the cement factories, workers were expected to raise concerns with their section leader, supervisor or group leader who would “help them resolve the problem” (MAHR1-1). Although these intermediaries had access to formal structures and, in principle, to more institutional guidance, many could not identify SVH at all. The leader of an all-male team, for example, dismissed the possibility of SVH in his section entirely: “No women work in our section … and it's impossible that men harass men!” (FXGL2-1). Subcontractors who directly managed workers on the construction sites also had little to no grasp of the concept. A security guard spoke at length about physical altercations between workers but struggled to articulate a response when asked what SVH meant (MXSG2-1). A union organiser was consulted about an attempted rape in employer-provided accommodation but did not accompany the victim to the police station or approach the company. According to a worker who knew the victim, he reasoned that the perpetrator had “stopped short of actual rape” (F02CPP).

Where intermediaries did recognise that a behaviour was problematic, they most often framed it as a private matter rather than a legitimate workplace grievance. On the construction sites, a subcontractor admitted advising a female worker who reported harassment simply to avoid the perpetrator, telling her: “If a man flirts with you, and you reply, it will keep on happening. If you ignore them, they'll stop doing it” (MXSU3-1). This flippant response placed responsibility for managing the behaviour squarely on the victim. Reflecting on other kinds of SVH cases, his reasoning became more explicit: “A woman has to let a man rape her or have sex with her … If she doesn't let him, he can't.” Sophea, a female team leader on a construction site, demonstrated a better understanding of SVH. Yet even though she was willing to intervene when sufficient evidence was available, she also framed SVH largely in interpersonal terms rather than as an organisational responsibility (FXGL1-1). In each of these cases, moreover, intermediaries' responses functioned to attach shame to the act of complaining. By identifying the victim's behaviour as the problem, they signalled that raising SVH as a workplace matter would expose the complainant to further humiliation rather than produce accountability for the perpetrator.

Even male and female union officials – who might be expected to be willing to challenge this framing – frequently reproduced it. A female union leader said SVH would not occur in her workplace because of occupational gender segregation and the marital status of her members (FEUL3-1). Another believed, like many of her male colleagues, that SVH could only occur with the victim's consent, adding that all the women in her union were middle-aged and could deal with perpetrators themselves (FEUL2-1). A male union official at the second cement factory described advising a female worker who reported being repeatedly touched by her male team leader to stay away from the perpetrator. He also explained that she had left herself open to unwelcome advances by being late for work. The same official recounted coaching a Chinese supervisor on how to proposition female workers without triggering a formal complaint: “I tell them, ‘If you're interested in a girl, you can ask them whether they have husbands or boyfriends. If you use those words, they can't file a complaint’” (MEUL1-1). The consistency of these responses across the two settings is significant. Security guards, team leaders and union officials in the cement factories demonstrated the same narrow conceptions of SVH, the same tendency to frame it as a private matter, and the same reliance on victim-blaming as their counterparts on the construction sites.

Even in the minority of cases where intermediaries recognised SVH and accepted that something needed to be done, complaints were almost always addressed through informal channels. On the construction sites, subcontractors were the most common point of contact for victims. One subcontractor said that, if there were a case, he would call the perpetrator and the victim in to discuss the matter, then assign them to different teams (MXSU2-1). However, he had never encountered such a case. Other subcontractors said that they would adopt a similar approach (MXSU1-1, MXSU3-1). In more serious cases, subcontractors might call the police, but only as a last resort. As another subcontractor explained, “If we ask police to come, we need to pay them … it wastes a lot of time and money” (MXSU4-1). Team leaders had even less capacity to act. Sophana reported a case of voyeurism to the head engineer, but no action was taken (FXGL2-1). In employer-provided accommodation, meanwhile, security guards are the key intermediaries. In cases involving women filmed while showering, one guard did nothing (M09RPP). In a case of attempted rape, the perpetrator was dismissed but no compensation was provided to the victim (F07CPP). In the most extreme case reported to us, a female worker was raped and killed, and her body was found in the toilet (F10CPP). The case was closed without any kind of institutional response.

Outcomes in the cement factories were not substantially different, despite the existence of formal grievance handling mechanisms. Champa described an incident in which a female worker employed in a karaoke lounge at one of the factories was propositioned for sex and grabbed when she declined. The victim reported the incident through the formal grievance process, but no action was taken. She felt she had no choice but to resign (F11BKP). When Srey Nara discovered that her locker had been forced open and her towel smeared with semen and pubic hair, she reported it to her male team leader, who responded: “If you want to know who did it, you need to check every man's special part [penis] to see whose has lost hairs” (F06AKP). Although she proposed installing lighting around the locker area, her suggestion was ignored. In a case where a man spied on a woman using the toilet, the team leader conducted what he described as a “practical investigation,” asking a worker to climb up to see what was visible from the adjacent cubicle. After concluding that only the victim's head could be seen, the perpetrator was told to pay the victim just USD 50 in compensation (F23AKP, F04AKP).

Where more serious cases reached upper management, perpetrators were most often left in place, especially where they were expatriates rather than local workers. A Chinese staff member who groped a female cleaner received a formal warning but remained employed (M01CKP). In another case, a Chinese supervisor offered money to a female cleaner for sex and attempted to grab her when she refused. She beat him off with her broom and called a union official, who called in the military police stationed at the factory. Management gave the victim USD 2,000 in compensation, but the perpetrator did not lose his job (MEUL1-1). As these examples suggest, the cement factories' formal grievance channels and union structures provided virtually no additional protection for victims.

The cases described above reveal how intermediaries interpreted and handled SVH complaints when they were made, but most incidents were never reported at all. Survey data provide insight into the scale of this under-reporting and the reasons behind it. In total, 62 of the 100 female cement workers surveyed had experienced one or more forms of SVH in their current workplace, of whom just nine (14.5%) lodged a report with their team leader or with their subcontractor (Table 3). None said they reported their experience to union officials, local authorities or the police – although as evidenced above, unions were sometimes involved in the management of SVH cases. Of the nine who had reported to their team leader or subcontractor, three said there was no impact, while six said the behaviour had stopped; further action was taken in only one case, where the perpetrator received a formal warning. Of the 200 female construction workers surveyed, 116 had experienced at least one of these behaviours, of whom nine (7.8%) lodged a report. Of those nine, four said nothing changed, two said the harassment had stopped, and in three cases the perpetrator was given a formal warning, moved, or dismissed.

Reporting rates were even lower among men. Of the 100 male cement workers surveyed, 37 had experienced at least one form of SVH, but just two lodged a report; both approached their subcontractor, who intervened and made the harassment stop. Of the 200 male construction workers surveyed, 91 had experienced one or more forms of SVH, but again only two had reported, one with no impact and one in which the harassment ceased. The key finding here is not that reporting rates were low, but that they were comparably low across both settings. Despite the cement factories' formal grievance procedures, human resources departments and union representation, reporting rates among women were still under 15%. Among men, the numbers were too small for meaningful comparison but followed the same pattern. This convergence confirms that the availability of formal mechanisms did not translate into substantially higher rates of reporting or more effective outcomes.

As Table 4 reveals, the top two reasons given by workers across all four categories (male and female, cement and construction) who had experienced SVH but not reported it were that the behaviour was considered normal or not serious. Over one-quarter of female construction workers and a sizeable proportion of male cement workers believed that reporting would exacerbate the situation. A substantial minority of men and women on construction sites, and a smaller number in the cement factories, were too scared or ashamed to report. Reflecting the fact that most SVH was perpetrated by coworkers rather than supervisors, only a small number said they had not reported because the perpetrator was a powerful figure.

These survey findings align with the qualitative data but also reveal additional nuances. Workers normalised less severe forms of SVH, such as comments on physical appearance, sexually suggestive jokes or being pinched or patted. This pattern is in line with research showing that many forms of SVH are widely perceived to be inconsequential by those who experience it, as well as by those who might otherwise be expected to address it (Leskinen et al., 2011). In more serious cases, shame and fear operated as powerful deterrents, as suggested by research showing that the primary driver of non-reporting is fear of disbelief, career damage and retaliation, compounded by perceptions that complaints will be trivialised or that offenders will not be sanctioned (Knapp, 2016). In the case described above, where a woman was groped by her supervisor, the union official had been prepared to take the case further, but the victim refused because she feared that news of it would spread and further damage her reputation (MEUL3-1). Phava, a laboratory worker, was sent a video of a Chinese supervisor having sex with a Cambodian woman, which she deleted without reporting because she was afraid of the consequences (F01BKP). A union official reported that a victim of attempted rape in a cement factory elevator decided not to pursue a complaint even though the union had CCTV evidence (FEUL1-1).

Critically, these decisions not to report did not reflect passivity but rather a pragmatic assessment of likely outcomes, informed by direct or indirect experience of how previous complaints were handled. Rothara, a cleaner on a construction site, spoke to her female team leader when she heard men speaking disparagingly about a co-worker's body (F19RPP). The team leader called the ringleader in, who brushed off the accusation, saying he was merely talking to his mates. The team leader took no further action. Some intermediaries, who had themselves experienced SVH, responded similarly. A female team leader in a cement factory had been subjected to insinuations that she was romantically involved with multiple men and to jokes about unclear paternity by a group of Chinese supervisors. Rather than make a complaint, she chose to disengage: “I told them that it was an inappropriate comment, then I stopped talking to them and just focused on my work” (FBGL2-1). Her response was not a failure to recognise SVH as a workplace issue but rather a rational response to her institutional environment.

The conceptual framework proposed above identifies three points at which intermediaries exercise discretion: (1) whether they recognise reported behaviour as SVH; (2) whether they classify it as a legitimate workplace matter; and (3) what course of action, if any, they take in response. As noted above, if legal frameworks exist, are effective in driving workplace policy, and that policy produces awareness and prompts action, intermediaries in more formalised settings should both recognise SVH more readily and demonstrate a greater willingness to classify it as a workplace matter, escalating complaints even if available grievance handling mechanisms are imperfect. What our data shows is that formal processes and mechanisms do not displace the intermediaries' discretion as they assess, classify and respond to SVH complaints.

Where intermediaries could not identify SVH, it precluded any likelihood of redress. Even among intermediaries who had received SVH training, the majority treated it as a private matter to be managed by the individuals involved. This dynamic transcended the structural differences between formal and informal workplaces, confirming that the gatekeeping process is driven primarily by normative rather than structural factors. In theoretical terms, this is in line with Powell and Sang’s (2015) concept of workplace habitus, but extends it by demonstrating that habitus shapes not only how workers experience SVH but also how those who receive complaints interpret and categorise it at each stage of the complaint process. Ruebottom and Toubiana (2024) theorise shame as an embodied disciplinary mechanism through which those who experience SVH come to internalise blame rather than challenge misconduct. Our data confirm this dynamic among workers, who cited shame as a reason for not reporting more serious incidents. When intermediaries located the cause of SVH in victims' own conduct and responded to complaints in ways that sexualised or ridiculed the complainant, their responses did not merely fail to address the reported behaviour; they also made the process of complaint itself a source of stigma, thereby reinforcing the likelihood of non-reporting.

Where intermediaries accepted that some response was needed, they most often sought to address complaints informally. On the construction sites, this was partly a function of the absence of any structured grievance mechanisms. However, cement factory intermediaries had access to general grievance mechanisms that were recognised as accommodating SVH complaints. Knox and Bohle (2025) argue that structural conditions determine the likelihood of reporting, and our findings partially support this, since institutional limitations in both settings constrained what intermediaries could do. However, our data also demonstrate that intermediaries' interpretive frameworks effectively neutralised whatever additional capacity the cement factories' (imperfect) formal structures provided. This suggests that structural conditions are necessary but not sufficient given the pivotal role of intermediaries in handling complaints.

Workers' own responses to SVH provide a final insight into the consequences of intermediary gatekeeping. Many of the workers surveyed shared the view that unwanted behaviours were simply normal. This perspective is itself a product of workplace cultures in which SVH goes unchallenged. Significantly, however, victims of more serious forms of SVH understood that reporting would not produce a meaningful outcome or could make matters worse. The decision not to report was thus a calculated response, shaped by institutional signals that reporting would not produce accountability and could in fact expose the complainant to further harm ^[5], producing low reporting rates that were taken by management as confirmation that SVH was not occurring. As McDonald (2012) observes, fewer complaints may indicate either a lower incidence of SVH or a workplace climate in which workers are reluctant to come forward. Our findings extend this insight by identifying the specific mechanism through which this ambiguity is produced.

These findings carry practical as well as scholarly implications. Cambodia's weak and poorly enforced legislative framework is a key reason why neither the cement factories nor the construction sites had developed SVH-specific policies: in the absence of specified employer obligations or credible enforcement, employers had little reason to treat SVH as distinct from other forms of workplace misconduct. Despite this evident source of weakness, there are nevertheless observable differences in the mechanisms available in the cement factories and on construction sites. While cement factories' general grievance mechanisms were understood by both management and intermediaries to cover SVH, these mechanisms produced no better outcomes than the complete absence of any mechanism on the construction sites, which suggests that outcomes cannot be explained by inadequate regulation alone.

It is clear, moreover, from these cases that the fragmentation of employer responsibility across subcontracting tiers allowed both cement companies and lead contractors to avoid taking accountability for workplace SVH, a gap that must be closed by establishing stronger workplace mechanisms for dealing with SVH and establishing joint liability for SVH across all tiers of subcontracting. In addition, reporting channels that do not position intermediaries as the sole or primary point of entry are essential, as are complaint systems that recognise the different positionality of parties to a dispute. Policies like the conflict-termination policies, which function as de facto deterrents to reporting, should be replaced with frameworks that distinguish between victims and perpetrators, and that offer institutional protection to victims wishing to report. Finally, training must be robust enough to shift the practices through which intermediaries consign complaints to the private sphere since – as the findings presented here confirm – procedural and regulatory reform, while necessary, is insufficient unless it confronts the cultural norms that shape how intermediaries interpret SVH complaints.

This study has assessed whether workplace formality determines how SVH complaints are handled in cement factories and on construction sites in Cambodia, a country which has poorly defined and poorly implemented regulatory restraints on SVH. In the case presented here, the availability of at least some formal structures for dealing with SVH in more formal workplaces did not translate into more systematic or more effective responses, with reporting rates low in both settings and workers in both citing similar reasons for remaining silent. Where complaints were pursued, outcomes also converged around outright dismissal of the incident or its informal resolution, with perpetrators most often left in place. In short, as this article has demonstrated, these outcomes were shaped less by institutional design than by the normative frameworks that intermediaries brought to the complaints they encountered.

This study has limitations that should be acknowledged. The findings are drawn from a single national context, and the survey sample was not statistically representative of Cambodia's construction workforce. These limitations notwithstanding, it makes several contributions to scholarship on workplace SVH, especially as it pertains to the Global South. In addition to providing a robust comparison of reporting in formal and informal workplaces in Cambodia, it speaks to the broader literature in three ways. First, by disaggregating intermediaries' decision-making, it reveals that gatekeeping is a sequential process in which complaints can be deflected at multiple junctures – a finding that has salience in better regulated contexts as well as for other contexts in the Global South. Second, it identifies a self-reinforcing dynamic in which intermediary filtering produces low reporting rates, which organisations then treat as evidence that intervention is not required, thereby perpetuating the conditions under which SVH goes unaddressed. Third, it provides further evidence that the causal link between legislation and workplace practice is tenuous at best, especially in contexts like Cambodia, where legislation is poorly enforced.

We acknowledge and thank all our survey participants and interview respondents, but especially the workers who so generously shared their stories with us. We also extend our thanks to our partner organizations (Building and Woodworkers’ International, the Solidarity Center and Union Aid Abroad–APHEDA), and the Building and Wood Workers Trade Union of Cambodia (BWTUC), which helped facilitate access to research participants, and acknowledge the contribution of our research assistants and the mental health support provided by Transcultural Psychosocial Organization.