Networks of complicity: social networks and sex harassment

Abstract

Purpose

The purpose of this paper is to explore the question of why sex harassment persists in organizations for prolonged periods – often as an open secret.

Design/methodology/approach

In-depth interviews were conducted with 28 people in diverse organizations experiencing persistent sex harassment. Data were analyzed using standard qualitative methods.

Findings

The overarching finding was that perpetrators were embedded in networks of complicity that were central to explaining the persistence of sex harassment in organizations. By using power and manipulating information, perpetrators built networks that protected them from sanction and enabled their behavior to continue unchecked. Networks of complicity metastasized and caused lasting harm to victims, other employees and the organization as a whole.

Research limitations/implications

The authors used broad, open-ended questions and guided introspection to guard against the tendency to ask for information to confirm their assumptions, and the authors analyzed the data independently to mitigate subjectivity and establish reliability.

Practical implications

To stop persistent sex harassment, not only must perpetrators be removed, but formal and informal ties among network of complicity members must also be weakened or broken, and victims must be integrated into networks of support. Bystanders must be trained and activated to take positive action, and power must be diffused through egalitarian leadership.

Social implications

Understanding the power of networks in enabling perpetrators to persist in their destructive behavior is another step in countering sex harassment.

Originality/value

Social network theory has rarely been used to understand sex harassment or why it persists.

Citation

Cunningham, P., Drumwright, M.E. and Foster, K.W. (2021), "Networks of complicity: social networks and sex harassment", Equality, Diversity and Inclusion, Vol. 40 No. 4, pp. 392-409. https://doi.org/10.1108/EDI-04-2019-0117

Introduction

Sex harassment is a prevalent and persistent organizational phenomenon despite the many efforts of legislators, scholars and managers (Cortina and Berdahl, 2008; Tenbrunsel et al., 2019). This disheartening reality has been documented by widespread press coverage of sex harassment in Hollywood (e.g. Harvey Weinstein), newsrooms (e.g. Matt Lauer), government (e.g. Roy Moore), sports (e.g. Marshall Faulk) and the Catholic priesthood (e.g. Cardinal George Pell). Not only is sex harassment persistent and prevalent, it is often ignored as an open secret, tacitly supported and hidden (Almukhtar et al., 2018). It harms victims and organizations in deep and lasting ways (Cortina and Berdahl, 2008; Fitzgerald, 1993).

The #MeToo movement highlighted not only the prevalence of sex harassment, but also the difficulty of ending it (MacKinnon, 2019). Sex harassment research is extensive and rich, but its focus often has been on why sex harassment between a perpetrator and a victim occurs rather than on why it persists. Furthermore, while the relationship between the perpetrator and the victim has been studied extensively, less attention has been paid to the web of relationships surrounding the two central actors. The purpose of this research is to address these gaps in research.

Background

Sex harassment is a term that extends traditional definitions of sexual harassment (Berdahl, 2007). Sex harassment is defined as “harassment based on sex – behavior that derogates, demeans, or humiliates an individual based on that individual’s sex” (Berdahl, 2007, p. 641). It encompasses harassment driven by both sexual desire and gender-based norms.

Sex harassment is often intertwined with other types of inappropriate conduct such as incivility, microaggression, bullying, ridicule, work sabotage, physical threats, vandalism, slander and ostracism (Berdahl and Raver, 2011). It is typically embedded in a larger context of disrespect (Lim and Cortina, 2005). The same perpetrator may engage in multiple forms of bad behavior, or multiple perpetrators may behave inappropriately. As such, sex harassment may not only involve a dyadic relationship between a perpetrator and a victim, but it may also involve interactions among multiple individuals.

Considerable research and theoretical development has focused on the antecedents of sex harassment to explain why it occurs (O’Leary-Kelly et al., 2009), but it has not fully addressed why it persists. For example, sex harassment occurs because of the following phenomena: sex-role spillover (i.e. inappropriate carry over of sex-based behavior to the workplace; Gutek and Morasch, 1982); the contact hypothesis (i.e. problems arising from the contact of men and women in the workplace; Gutek et al., 1990); the power-dominance explanation (i.e. unequal power across men and women in the workplace and society; Cleveland and Kerst, 1993); organizational influences (i.e. gender domination by men and a toxic organizational culture; Fitzgerald et al., 1995) and aggressive work behavior (i.e. sex harassment as goal-directed behavior; Berdahl, 2007). Although the extant research is rich and varied, it does not fully explain how or why a toxic organizational climate that does not take sex harassment seriously develops, or why it persists when it is damaging to the organization. In addition, the existing theoretical approaches may not fully consider the role of third parties (e.g. bystanders) in either explicitly or implicitly condoning the behavior of the perpetrator. Social network theory may provide further insights into these phenomena.

Social network theory

Social network theory focuses on the role of social relationships in transmitting information and influencing attitudes and behavior (Granovetter, 1985). Social relations influence how individuals act in organizations, and sex harassment involves social relationships. While a perpetrator and a victim are central, they are embedded in an organization’s wider structure of relationships (Brass et al., 1998; Cortina et al., 2017).

Social network theory’s concept of “embeddedness” asserts that social influence – the desire to develop and maintain social relations – often dominates institutional and moral considerations to perpetuate the benefits of social connections. Personal relationships and the organization’s structure of social relations can generate trust and discourage or increase malfeasance depending on who is trusted (Granovetter, 1985). For example, the desire to continue past relationships may lead one to overlook another’s questionable behavior when it seems out of character. Embeddedness is a lens to look deeper into the organizational context to understand members of an organization who may – through action or inaction – be complicit, complacent or ignorant about sex harassment.

Organizational relationships involve different types of interpersonal connections or “social network ties” (Granovetter, 1985). These connections may be “strong” ties that create obligations, enforce organizational norms and restrict information flow, or “weak” ties that provide access to novel opportunities and information (Burt, 2000a; Granovetter, 1973). The ability to draw on the resources and support of others in the network is referred to as “tie activation” (Smith et al., 2012).

Social connections and ties are dynamic and may fluctuate with changes in organizational circumstances (Hurlbert et al., 2000). When there are few interpersonal ties, a “structural hole” exists in a network (Burt, 1992). Individuals who can create ties to span structural holes become brokers who gain an advantage relative to their peers by accessing resources, connecting others or combining information (Burt, 1992).

“Contagion” and “groupthink” are concepts related to embeddedness. Because people are embedded in social relationships, attitudes, behaviors and norms can spread rapidly in an “infectious” manner (Granovetter, 1978; Jackson, 2010). This diffusion of shared perspectives is referred to as “contagion.” Groupthink arises when members accept the group’s reasoning, avoid expressing individual opinions and interpret information in a positive way that supports the group’s assessments of events. It has been used to explain negative organizational outcomes in many contexts including the space shuttle Challenger disaster (Esser and Lindoerfer, 1989).

Social networks reflect and create power and status. The organizational hierarchy results in power differences (Brass and Burkhardt, 1993), and the organization’s formal structure enables and constrains individuals’ opportunities to connect or form ties with others (Kleinbaum et al., 2013). “Network centrality” is a concept related to power and status. It emerges from the number (density) and structure of ties. High (or low) centrality may result from the formal hierarchy or from the informal or emergent patterns of interactions among individuals (Gulati and Puranam, 2009). A central position in a network typically enables an individual to understand the network, which is referred to as “network cognition” (Krackhardt, 1987). It, along with the number of ties, can be sources of power.

Social networks provide members with access to resources and opportunities (Coleman, 1988) and expose them to social norms, interpersonal loyalties and expectations and moral obligations (Pescosolido, 1992; Wellman and Worley, 1990). Social network theory holds promise as a theory to address the question of why sex harassment persists in many organizations.

Methods

Qualitative methods such as in-depth interviews are useful when the research goal is to understand complex interactions and deeply embedded perceptions, beliefs, and values, and when researchers cannot be sure what interpretation, norm, affect or rule is guiding individuals (Braybrooke, 1965; McCracken, 1988). Persistent sex harassment is a phenomenon in which the complexities, subtleties and perceptions of various organizational members must be examined, so in-depth interviews were conducted. In-depth interviews are different from highly structured survey research interviews (McCracken, 1988). While the latter are useful for predicting behavior and generalizing to a population, in-depth interviews are designed to ascertain the informant’s understanding and definition of the situation. They encourage the informant to structure the account of the situation and allow the informant to reveal his or her view of what is relevant (McCracken, 1988; Strauss and Corbin, 2015).

Informants for in-depth interviews need to be selected based on the knowledge they have or the position they occupy that is relevant to the phenomenon being studied (Hochschild, 2009). Informants were recruited because they had worked in organizations where sex harassment had persisted. They were recruited using various methods. The researchers read reports of sex harassment (e.g. press coverage, legal investigations) and then contacted individuals who had been involved. The researchers also contacted people within their personal networks whom they suspected had experienced sex harassment or worked in organizations in which sex harassment had persisted. The researchers also used snowballing (Moriarty, 1983). At the end of each interview, they asked informants if they would refer them to others who had worked in organizations in which sex harassment had persisted.

Interviews were conducted with 28 informants – 18 women and 10 men. In total, 16 were US based, and 12 were Canada based. Some informants described multiple cases of persistent sex harassment. In all, 37 cases were described. Informants were victims in 20 of the cases, and bystanders in 17 cases. The cases were drawn from diverse sectors: business, government, higher education, health, journalism, the military and nonprofits. The organizations ranged in size from four employees to almost 200,000. Of the 37 cases, nine occurred in organizations with less than 500 employees, and eight occurred in organizations with between 500 and 5,000 employees. Nine cases occurred in organizations with between 5,001 and 10,000 employees, and 11 cases occurred in organizations with more than 10,000 employees.

Table I describes informants and perpetrators – gender, role, organizational position, power of victims and bystanders relative to the perpetrator and type of knowledge of sex harassment. First-hand knowledge indicates that the informants had observed or experienced sex harassment themselves; second-hand knowledge conveys that a victim or someone who witnessed sex harassment told them, and general knowledge indicates that the informants knew about sex harassment from hearsay. In Table I, the term “perpetrator” refers to primary perpetrators. They held a variety of organizational positions, were predominantly male and typically had higher formal power in terms of organizational position than the victims. Though the number of female perpetrators was small (3), there were no notable differences between the female and male perpetrators.

The interview protocol , which is in the Appendix, was composed of broad, open-ended questions designed to permit informants to define the situation and decide what was important. Informants were prompted to engage in what Wallendorf and Brucks (1993, p. 341) referred to as “guided introspection,” which elicits reports of specific experiences associated with a unique context in time and place. Sample questions were, “Would you describe the problematic behavior that you witnessed or heard about in your work organization?” and “What do you think enabled this problematic behavior to continue?” Interviews typically ranged in length from 60 to 90 minute, but several extended to 3 hour. Informants were assured of anonymity.

The interviews were recorded and transcribed. The data were analyzed systematically and intensively using standard qualitative methods (Corbin and Strauss, 2015). Themes that were common across multiple informants were identified and are reported in the next section.

The limitations of in-depth interviews are well known, and the authors attempted to mitigate them. Potential weaknesses include the tendency of researchers to ask for information to confirm their assumptions and the tendency of respondents to give “canned” or self-serving answers to the questions (Jensen and Jankowski, 1991). The interview protocol was designed to minimize these weaknesses with questions that were broad and open-ended. The authors used guided introspection to encourage informants to think aloud in an unrestrained manner. Informants reported encountering sex harassment in varied ways, and they represented a variety of positions and experience levels in different types of organizations. Portraying the range of the phenomenon is important to developing theory in qualitative research (Bonoma, 1985).

To guard against subjectivity and establish reliability (Corbin and Strauss, 2015), authors did initial data analysis independently. They reported only findings that they agreed upon. Many key findings were corroborated by secondary data including a report outlining an extensive investigation into sex harassment at a large national organization. When reporting the themes, quotations are used to add objectivity and depth of understanding. Unless otherwise noted, the quotations are representative of ideas expressed by multiple informants.

Findings

The overarching finding was that perpetrators were embedded in networks of complicity (social networks) that were central to explaining the persistence of sex harassment in organizations. These networks protected perpetrators and enabled them to continue sex harassment and other problematic behavior. Networks of complicity encompassed a wide range of people whose roles ranged from support staff to high-level executives and professionals. Members of the network can be classified as two types: active and passive. Active participants were drawn into the network by a perpetrator’s charisma, formal power and ability to control and shape information. They proactively protected perpetrators through actions that ranged from making excuses for perpetrators and shielding them from criticism, to sabotaging those trying to expose or counter perpetrators’ behavior. They are referred to as “active enablers.”

Passive enablers were recruited to the network of complicity by perpetrators or active enablers. They enabled perpetrators to continue sex harassment largely through continually ignoring perpetrators’ bad behavior, turning a blind eye, making light of it or rationalizing it to the point that it was not on their radar.

A third category of people who surrounded the network of complicity were bystanders. They had (or should have had) some knowledge, or at least suspicion, of bad behavior. Perhaps unwittingly, their inaction permitted the perpetrator and the network of complicity to continue. Passive enablers and bystanders were on a spectrum in terms of how much they knew about the perpetrator and the network’s activities as well as how much power they had to resist and how closely they mirrored the perpetrator’s values. However, one thing was clear: perpetrators could not have persisted in sex harassment without the support of a network of complicity. Four themes that emerged from the data are reported below.

Theme 1: networks of complicity are enabling mechanisms

Organizations harbored sex harassers because networks of complicity protected them from sanction in ways that would seem incredible to an objective observer. The enabling took many forms. Common types of active support involved “fixing” problems and rewarding people who complied with the norms of the network of complicity as Gary (journalism, bystander) described:

Okay, so how many people around [the perpetrator] know that he is a sexual predator? Well, he’s got to have his fixers, and he has to have the people who are doing the “honey pot” work […]. They all end up having to be burdened with the necessity of deceiving themselves and others as to who [the perpetrator] is and what he’s doing.

Members fell prey to “motivated blindness” – the tendency not to see the unethical behavior of others when it is not in one’s best interest to do so (Tenbrunsel et al., 2019) – and “moral myopia” – a short-sightedness in moral vision in which moral issues do not come clearly into focus (Drumwright and Murphy, 2004). Blindness to bad behavior and moral myopia resulted from the choice to preserve ongoing relations over moral obligations. Ignoring bad behavior within the network of complicity was common as Lillian (journalism, victim) observed:

They [network members] all know each other’s bad things because they enable each other’s bad behavior. And they let each other’s bad behavior continue saying, “Well, I’ll overlook that.”

Other network of complicity members said that they would do something about the perpetrator, but then they did not as Anita (journalism, victim) recounted:

I’ve heard from multiple people who worked at [the company] that there were rumors about him – that people knew that he attacked interns in the backs of cabs […]. Somebody told her manager that he had assaulted her in his office under the pretenses of talking to her about mentorship. She told her manager, and he said, “Oh, yeah. We know that there have been issues. I’ll report it.” And then nothing happened.

Some network of complicity members provided passive support by turning a blind eye and a deaf ear to rumors, jokes and innuendos. Others refused to dig too deeply into evidence of inappropriate behavior, as Brianna (higher education, bystander) noted:

I don’t know the nature of the conversation [behind closed doors]. I do know there were some loud voices in there. I knew people came out crying. I knew some of this was happening, but I didn’t know what that was. So that’s somewhat like turning, like seeing something – not really understanding what is, but also not wanting to dig any further.

Karen (nonprofit, bystander) attributed the inaction of some network of complicity members to the fact that they did not know what to do to stop the sex harassment:

They’re passive enablers in that they don’t do anything directly. They almost don’t want to know this [existence of sex harassment] because it challenges them to do something, and they have no clue what to do.

Other network of complicity members and some bystanders unwittingly provided passive support for perpetrators because they did not recognize the gravity of the perpetrators’ behavior. As Sarah (journalism, bystander) described:

And so, I look back, and I think I really only saw it [sex harassment] back then as immature guy humor that had to be tolerated to play on that ballfield […]. I wish that I had seen it as having the potential to be more damaging and had said something […] and maybe put the brakes on things earlier before the culture deepened and worsened.

Network theory has an explanation for such behavior. The convergence of norms in the network of complicity included an acceptance or tolerance of sex harassment, and as the members became socially embedded into the network, it also engendered groupthink. For example, when confronted with evidence of sex harassment by a member of the in-group (i.e. the network of complicity), other network members rationalized that a fellow member of the in-group could not be capable of such behavior (Janis, 1971). Acknowledging that a co-worker would sex harass others is difficult; doing so risks damage to favorable perceptions of the organization, the in-group and one’s relationships.

Clearly, perpetrators could not have persisted in sex harassment without the support of the network of complicity. This finding raises the question, “How are these networks of complicity built and maintained?”

Theme 2: perpetrators are network builders and power brokers

Perpetrators were talented boundary spanners who built broad networks both inside and outside their organizations. These networks helped perpetrators gain or increase their power and use power differentials to protect themselves from condemnation. Nancy (journalism, victim) described such a network:

They had a network of people in power that were actually all connected, and it was, to me, more of just an older, elite, privileged, white boys’ network.

Organizational social networks create and perpetuate social norms, interpersonal loyalties and moral obligations (Pescosolido, 1992; Wellman and Worley, 1990). Networks of complicity featured high levels of connectedness. Since network members had a number of connections with other network members, including the perpetrator, the network was dense (e.g. Otte and Rousseau, 2002). Because the networks were dense, norms, values and opinions coalesced creating protection for the perpetrator. The network provided power, but many perpetrators tapped other forms of power as well. Anita (journalism, victim) explained how a perpetrator had formed ties with powerful people outside their New York-based organization:

He spent a lot of time building relationships with critical stakeholders […] investing in his relationships […]. [The perpetrator] was connected and an insider in Washington […]. He had tremendous influence behind the scenes.

Perpetrators had the power of weak ties as well as strong ties (Burt, 2000a; Granovetter, 1973). Their weak ties outside of their organizations enabled them to span structural holes in the networks (Burt, 1992), which gave them access to unique resources and served as another source of power. Their strong ties within their organizations gave them internal clout to make or break careers, and they were more than willing to use their power to break careers as Andrew (business, victim) noted:

They exert power […]. They hold things over you like your position. They hold things over you like your job. They hold things over you like your promotion. And then they have a connected network of, I believe, other people in power.

Perpetrators also used their legitimate power resulting from their positions in the organizational hierarchy (e.g. a senior manager) in various ways to build their networks. They doled out rewards to members and potential members of their networks, who owed their positions, bonuses or promotions to perpetrators. Informants noted that people receiving these benefits often did not receive them because of merit, but because they were willing to support the perpetrator. These rewards increased loyalty to the perpetrator. Michael (higher education, bystander) described the loyalty these rewards engendered:

There are people put in roles or given roles that were way above their pay grade. They had no competency to be successful in those roles, but they were there. And when you put a mediocre person in a role that is way above their pay grade, they’re going to be beholden to whoever put them in that role, and they’re going to overlook anything else because now they are grateful for being in a role that they could never hope to earn.

Perpetrators included or excluded others based on their support of the network. They frequently used fear, along with power and authority, to exclude, dominate or intimidate others. Mary (journalism, victim) noted:

[The perpetrator] does focus on maintaining power using really harsh management tactics, and it’s very manipulative. All those things really fit into the definition of Machiavellian behaviors or attitudes.

Perpetrators used verbal and physical intimidation techniques. When people – often older men and women – opposed perpetrators, the perpetrators disparaged or threatened them. One perpetrator frequently formed his hand into a gun-like shape and pointed it at people he disagreed with or disliked. Even when perpetrators did not actually harm others, people assumed that they would be willing to do so as Anita (journalism, victim) explained:

Somebody bold enough to do this [unwanted attention and unwelcomed sex proposition] to a young woman behind the scenes was probably a narcissist who would hurt her if she rejected him.

Perpetrators and active enablers increased and reinforced their power through bullying and discrimination. They often preyed upon and took advantage of the most vulnerable organizational members, whom more powerful people were unlikely to know well or be concerned about. However, informants noted that perpetrators and active enablers also attacked talented high performers if they viewed them as a threat to their agendas. Some informants such as Sophia (higher education, victim) described the bullying and incivility by active enablers whom she complained to as more difficult to tolerate than the threat of rape that prompted her to complain. The bullying and disparagement caused her to have significant self-doubt and to feel dismissed and undervalued. She also developed a major medical condition, which she attributed to the bullying and incivility. Like Sophia, other informants described the harm from bullying and incivility as more insidious than overt sexual affronts because these behaviors were incessant, unrelenting and impossible to avoid, while the more overt sexual affronts were episodic and sometimes avoidable with careful planning.

As the above findings illustrate, perpetrators wielded power arising from various formal and informal sources. Perpetrators had central positions (network centrality) resulting from their connections with others (their strong and weak ties) (Freeman, 1979; Granovetter, 1973). Because of their centrality, perpetrators understood the networks around them; that is, they had a high degree of network cognition (Burt, 1982; Krackhardt, 1987). Having the ties and understanding the network enabled perpetrators to activate their ties when they needed them for support (i.e. tie activation, Shea et al., 2015). In contrast, victims were either on the periphery of their organizations or were pushed into marginal, isolated positions by the perpetrator and network members. As such, they had far fewer connections, weaker ties and a much lower level of network cognition than perpetrators. Victims, therefore, were often unable to activate their ties for support.

Social network theory illuminates the ways that perpetrators built and maintained their networks of complicity. Two additional questions arise: “What are the characteristics of the perpetrators,” and “How were they so successful in building networks and wielding power?”

Theme 3: perpetrators are myth builders and information manipulators. Network members perpetuate the myths and misinformation

Perpetrators were able to build and maintain their networks because they were master myth builders and information manipulators, and network members perpetuated the myths and misinformation. Our informants described three types of perpetrators: stars (or rising stars), incompetents and the presumed harmless.

Stars (or rising stars)

Some perpetrators had rare skills or scarce talents that greatly benefitted their organizations – e.g. a renowned neurosurgeon or a television news anchor. They were perceived as indispensable to the organization’s ability to accomplish its goals. Mary (journalism, victim) described a star:

When one company relies so much on one man – when that man is seen as the brains of the operation – that is a recipe for disaster […]. His lieutenants, they all bestowed him with magical powers and belonged to his cult of personality. They were all very loyal to him.

While stars had considerable talent, they also were highly adept at promoting themselves and elevating their indispensability to mythic proportions. Nancy (journalism, victim) noted that a “star” perpetrator was never disciplined because the organization and its identity are “wrapped up in those key journalists […] so they have power because they embody the organization.”

Stars relished having devout followers. These individuals, in turn, perpetuated the image of the perpetrator as having superior capabilities essential to the organization. Gary (journalism, bystander) noted:

My experience has been that they [perpetrators] have “groupies.” So, they have people on the staff who like to be in the aura of the powerful, and the people in power like to have the “groupies.”

Despite any façades to the contrary, stars were highly egotistic and cared much more about their own success and continuing their illicit behavior than they did about the success or well-being of their organizations.

Incompetents

Some perpetrators were surprisingly incompetent, but they were extraordinarily adept at building myths about their talent and indispensability. They constantly spoke in private meetings, in public forums and to the media about their talents and supposed successes.

Active enablers appeared to be blind to signs of the perpetrators’ incompetence, embraced the myths and perceived the perpetrators to be stars. Bystanders and others who could not directly observe the work of the perpetrators also believed the myths and saw these perpetrators as gifted leaders and managers. More astute observers outside the network of complicity described perpetrators’ weak leadership skills, their lack of technical competency and their inability to manage resources. Frances (higher education, bystander) noted:

But incompetence was my major concern […] incompetence in, really in most aspects of the role – the leadership skills, the communications. It was the whole thing […] like saying that he had a plan when clearly there was no plan.

Perpetrators were adept at recruiting others with talent to their networks, so that they could conceal their own deficits. They were quite willing to defy organizational policies to put their chosen lieutenants in power. Francis described serving on a search committee forced to hire an unqualified candidate who had been preselected by the perpetrator:

We had a huge thing where we had to hire a new [senior manager]. And that was not a good search […]. [The perpetrator] had no regard for policy or procedure. And it was very tainted right from the beginning because the decision had already been made. So, that’s a very helpless feeling, especially when you’re on a committee, and it’s not the direction that the committee had decided to go on. But of course, we were not the final say.

Presumed harmless

Some perpetrators disguised the harm they wreaked behind a veil of humor. They built a myth that they were completely benign, funny and fun-loving people. Active enablers and even bystanders perpetuated that myth. For example, Tammy (journalism, bystander) said, “I just saw them as immature goof-offs.” Others rationalized the conduct as “doing no harm,” “as just being his way”; he was an “old-school insider” who was merely “quirky,” a “character” and fun. Debra (journalism, bystander) described one such perpetrator:

We had an older male […] and numerous times he would say inappropriate things toward women in general, toward co-workers, toward his boss […]. Many people excused the behavior as, “Oh that’s just him. Hahaha” because he was a character […]. He was a funny guy who crossed the line all the time, all the time. And it was well known among everybody who worked there, and we all let it happen.

The humor and light-heartedness not only protected the perpetrator, it also sent a message. Observers and victims alike assumed that people who laughed at the inappropriate comments of presumed-harmless perpetrators condoned, accepted and affirmed the behavior instead of objecting to it.

All three types of perpetrators manipulated information to their advantage decreasing the probability that their bad behavior would be reported. Active enablers were quick to disseminate misinformation for perpetrators. For example, perpetrators would start rumors about their victims’ lack of skills, the insignificance of their accomplishments or their lack of integrity. Network members perpetuated these slurs. This behavior marginalized victims and created self-doubt. The way perpetrators shaped and framed information, often in deceptive ways, also kept organizational members off balance. Jayson (higher education, victim) noted:

And my personal experience really had to do more with the mismatch in what he tells me versus what he says in public. So a lot of it’s really deception […] and what I refer to as “selective incivility.” […] We call this person “Trump-Lite.”

Others noted that perpetrators were clever in how they framed their personas to screen and manipulate information depending on the situation. Gary (journalism, bystander), who had a particularly close working relationship with a perpetrator, testified that the perpetrator’s behavior was completely unassailable in his presence:

[The perpetrator] was a completely different person with me […]. Never once in my presence in any way, shape, or form, did [he] ever say anything that could even be misconstrued as locker room talk or bad behavior. I think he knew that that wouldn’t fly with me […]. I think [perpetrators] just have a very fine-tuned antenna as to read people.

Importantly, the people who had hired the perpetrators appeared to be willing – if not eager – to buy into the perpetrators’ myths and misinformation as were the active enablers.

As the above examples illustrate, perpetrators had positive as well as negative characteristics. They were described as being social, energetic, charismatic, passionate and charming but also cunning, dishonest, opportunistic, two-faced, unscrupulous, egotistical, manipulative and Machiavellian – characteristics others have documented (e.g. Ziegler-Hill et al., 2016). Perpetrators’ ability to build myths and manipulate, shape and control information shrouded their pathologies and negative characteristics so they could harass with impunity. Perpetrators’ centrality in their networks increased their ability to control and manipulate information, build myths and use their status and power to encourage other network members – especially active enablers – to embrace and perpetuate the myths and misinformation. These findings converge with extant literature (e.g. McLaughlin et al., 2012) and illustrate how perpetrators increase their power and span of influence, which became self-reinforcing as suggested by Magee and Galinsky (2017). That is, the higher the status, power and influence that perpetrators’ attained, the more network connections they established, the greater their network centrality became and the greater their ability to create myths and manipulate information to camouflage their bad behavior.

The previous findings raise the question, “What are the effects of networks of complicity?”

Theme 4: networks of complicity metastasize

Like a cancer, networks of complicity spread bad behavior causing lasting harm to victims, other organizational members and the organization as a whole. A vicious organizational cycle was set in motion. Shielded by the network of complicity, perpetrators got away with sex harassment despite organizational policies forbidding it. They felt immune from punishment, and their sense of entitlement grew. As the perpetrator became more empowered and the network became more embedded, the bad behavior increased, spread and often became more severe. In speaking of a doctor who harassed a nurse and eventually threatened her life, Andrew (health care, bystander)[1] described the consequences of a perpetrator’s immunity from sanction:

It created a bit of a pandemic [of bad behavior] and the belief that it really doesn’t matter what a physician does; we’re not going to get rid of them.

The bad behavior hurt victims who reported physical symptoms ranging from sleeplessness to autoimmune diseases and psychological symptoms ranging from inability to concentrate to severe anxiety. Incivility among network members created, contributed to and perpetuated a toxic organizational culture. Anita (journalism, victim) described the behavior of an active enabler:

[…] she [active enabler] was a toxic leader who would gossip about her subordinates. Nobody knew where they stood. Your value was based on your pedigree […]. She also favored men, and she would gossip about their sexuality and their personal lives […]. The culture just didn’t allow us to do our best work.

Even incivility that was less severe took its toll and kept workers from feeling respected and safe.

Bystanders suffered as well. Hugh (journalism, bystander) expressed angst and regret that he felt even years later:

And I like to think of myself as progressive and caring about people, and I was even a manager for some of that time [when sex harassment persisted], and I think I just let it slide. And it’s the regret of my career – that I let that slide.

A number of informants highlighted unfairness in hiring and promotion as a form of organizational harm. Perpetrators often put unqualified people whom they knew would be loyal to them in power. Another informant, Kayla (higher education, victim), reported that she and her co-workers became so demoralized that they could no longer perform at the level they had previously:

The impact that they [perpetrators] have on their direct reports is to create such dysfunction and such angst – or people feeling not valued or respected or appreciated or not even able to get their job done – because they’re being so micromanaged. And then that in turn can result in illness because people who are used to accomplishing things are now being blocked.

Informants described their organization as being transformed by the network of complicity from being a “great place to work” with a collaborative climate to one that developed a toxic climate devoid of trust and cooperation. Destructive competition developed as people and departments were pitted against each other in fights for recognition, favor or resources. Informants spoke of a breakdown in organizational trust as Andrew (higher education, victim) did:

I had to ask myself, “Do you trust your organization? Do you trust that they will do the right thing? Do you trust they’ll make the right decisions if you got yourself into trouble or if somebody else got in trouble?” […] And I can honestly say at this point, I would say no. So, I’ve lost trust in [the organization] […] to do the right thing.

When organizational trust broke down, people were even less likely to report wrongdoing.

The toxic culture, the breakdown in trust and the widespread disillusionment had devastating effects on the organization. It prompted capable people to find jobs elsewhere robbing the organization of much needed talent. As Andrew noted, “You just run; you just leave.” Anita (journalism, victim) spoke of how some people even left the professions they loved as she had done:

If you put into power people who sexually harass their subordinates […] or their colleagues behind the scenes, and those who scream at and don’t acknowledge young people just starting out in the profession, it creates a culture in which you think, “This is what it takes to get ahead – these are the types of people that get ahead.” And that wasn’t a place that I wanted to be despite my love for the profession.

As employees who remained became disillusioned and fearful, they stopped innovating or developing new initiatives. The organization’s future was jeopardized as Kayla (higher education, victim) explained:

And I said to [a senior leader], “With people being so afraid and abused, innovation has come to a halt.” And I perceived that as a huge risk, and I told him that I suspect that shortly we would see a decline in [key success measures] and other things because we just weren’t able to move forward any longer.

Theme 4 illustrates the network property of contagion – the infectious spread of attitudes, behavior and norms from “infected” to “susceptible” people (Granovetter, 1978; Jackson, 2010). While the perpetrator was the initial sex harasser, other network members began to exhibit other harassing behaviors such as bullying and incivility. Improper behavior became the norm, and it was exacerbated and spread. A toxic culture developed, and such cultures have long been recognized as lowering organizational productivity (e.g. Goldman, 2008).

In summary, the data robustly documented that networks of complicity formed to protect, insulate and enable perpetrators to persist in sex harassment. The findings are summarized in the following list:

• Overarching theme: perpetrators were embedded in networks of complicity that were central to explaining the persistence of sex harassment in organizations.

  • − The network of complicity supported, shielded and enabled perpetrators to continue their inappropriate behavior without sanction for prolonged periods of time.

• Theme 1: networks of complicity are enabling mechanisms.

  • − Members of the network of complicity enabled perpetrators to continue their bad behavior by providing active support (e.g. covering for perpetrators) and passive support (e.g. turning a blind eye).

• Theme 2: perpetrators are network builders and power brokers.

  • − Perpetrators used formal and informal power to build networks, and they then derived more power from their central positions in the networks. Their ability to activate network ties and span structural holes in networks inside and outside of their organizations also heightened their power.

• Theme 3: perpetrators are myth builders and information manipulators.

  • − The three types of perpetrators – stars (or rising stars), incompetents and presumed harmless – built and perpetuated myths about themselves. They manipulated information to enhance their status and increase their power and influence. Network of complicity members perpetuated these myths and spread misinformation.

• Theme 4: networks of complicity metastasize.

  • − Networks of complicity created toxic organizational cultures that spread like a cancer and reduced the performance of individuals and the organization as a whole.

Discussion

The findings demonstrated the usefulness of social network theory in explaining the effects of networks of complicity, how and why they are built and maintained and why sex harassment persists. They also documented how social network theory can supplement and provide insights regarding other theoretical approaches used to study sex harassment. For example, in organizations in which networks of complicity allowed sex harassment to persist, social network theory highlighted other sources of power-dominance (e.g. network centrality) and other types of organizational influences (e.g. social influence through embeddedness and contagion) (Granovetter, 1978).

Importantly, network theory provides guidance to minimize the influence of perpetrators and networks of complicity and reverse the toxic organizational cultures they created. For example, the strong, reinforcing ties among perpetrators and network members must be broken or weakened (Burt, 2000b; Feld, 1981). Perpetrators must be removed from organizations, and when they are removed, formal work ties with them are broken; however, the connections among other network members typically remain and enable the network to persist. In fact, when members of a dominant group – e.g. a network of complicity – feel threatened by another group (people who report a perpetrator), backlash can occur, and network members are likely to strive to retain their power and status (Dobbin and Kalev, 2019).

Connections and ties have a decay function, and infrequent interaction can result in the loss of a connection (Kleinbaum, 2012). As such, leaders must understand both formal ties (e.g. work assignments) and informal ties (e.g. repeated social interactions) within the organization (Feld, 1981), and then they must work to reduce interaction opportunities of network members (Burt, 2000b). For example, to weaken formal ties, work units can be reorganized, and work responsibilities can be reassigned to reduce task interdependence (Feld, 1981). Informal ties typically are based on interaction choices that individuals make (Kleinbaum et al., 2013). Leaders must communicate that there are consequences to choosing to be connected to perpetrators and their supporters (Kleinbaum, 2012). All should understand that there is no value in remaining in a network of complicity and that those connections should be pruned (Hernandez et al., 2015). Internal communication vehicles or social media channels can be used to correct misinformation generated by perpetrators and networks of complicity and unpack the myths they created (Mergel, 2011).

Network theory provides guidance for supporting victims of sex harassment. As demonstrated by the findings, perpetrators used their power to isolate their victims and prevent them from accessing support. Victims must be connected and integrated into networks of support. Many victims are hesitant to report sex harassment, so organizations can provide victims access to people who can provide them support services, such as counseling and confidential legal and professional consultation, without making a formal report (NASEM, 2018). If victims launch formal reports, ombudspersons can support them while advocating for a fair process. Organizations can connect victims with people who will keep them informed about what to expect throughout the process and let them know about developments in the investigation and any disciplinary actions taken (NASEM, 2018). Mentoring programs can help victims form relationships with higher-ranked employees who have more power and influence (Hezlett and Gibson, 2007).

Victims and their supporters can also receive validation and support from professional networks, professional societies and others outside of their organizations (NASEM, 2018). For example, the #MeToo movement highlighted the power of networking outside of a given organization. Professional networks – e.g., Press Forward, a network of current and former journalists (Boucher, 2017), and CongressToo, a network of former Congressional staffers (Akin, 2018) – have offered many victims support.

Bystanders can play important roles in mobilizing positive networks that counteract networks of complicity. Evil leaders cannot lead in evil ways without bystanders who are aware of the leader’s behavior but do nothing to stop it (Kellerman, 2004). Bystanders can benefit from intervention training based on the “Confronting Prejudiced Responses” (CPR) model, and they can use confrontation as a lever to change organizational culture (Ashburn-Nardo et al., 2008). The CPR model helps bystanders learn intervention techniques that increase their ability to detect discriminatory behavior, help them understand harm and teach them how to confront discriminatory behavior. However, sex harassment training can backfire and increase victim blaming and the likelihood of harassing if it is framed in a way that puts trainees in the defensive position of potential perpetrators (Dobbin and Kalev, 2019). Such “forbidden behavior” training reviews harassment law, specifies forbidden behavior and outlines complaint processes and punishments (e.g. Dobbin and Kalev, 2019). In contrast, bystander intervention training that is not accusatory prompts trainees to identify and empathize with victims, to intervene and to offer help (Katz and Moore, 2013). Sex harassment training has been most successful when “training lasted more than four hours, was conducted face-to-face, included active participation with other trainees on interdependent tasks, was customized for the audience, and was conducted by a supervisor or external expert” (NASEM, 2018, p. 152).

In organizations suffering from persistent sex harassment, network theory highlights ways of preventing networks of complicity. As noted in the findings, perpetrators derive power from their centrality in the network and their ability to use power differentials to their advantage. Likewise, sex harassment thrives in environments with significant power differences – e.g., the star network anchor, the star academic researcher – in which one party has a great deal of influence over the lives and careers of others (NASEM, 2018). Feeling powerless increases people’s sensitivity to the obstacles to confronting perpetrators and inhibits their intentions to confront (Ashburn-Nardo et al., 2014). As such, the power structure within an organization must be altered in a manner that diffuses power and reduces the isolation of followers (NASEM, 2018). To diffuse power, leaders should be educated on the value and effectiveness of egalitarian leadership styles based on shared power and trained to use them (NASEM, 2018). For example, transformational leadership, a form of egalitarian leadership in which leaders motivate followers through inspiration, is positively associated with team effectiveness (Flood et al., 2000). Leaders with a socialized perspective identify and solve problems based on collective interests and instill self-responsibility, self-initiative and autonomy in their followers (House and Howell, 1992). A family-like culture of caring, looking out for others and accommodating differences empowers employees throughout an organization (Cunningham, 2018). These leadership approaches reduce the risk of sex harassment because those in power are more approachable, tasks are shared and subordinates are treated respectfully as people of value who can further the organization’s interests (Nelson et al., 2017). In addition, leaders must understand that others view them as most responsible for confronting problematic behavior because they are authority figures, and they must assume personal responsibility for doing so (Ashburn-Nardo et al., 2019).

In sum, the approaches discussed above can be used to reverse the toxic culture created by networks of complicity and transform organizations into places in which careers flourish and sex harassment does not.

Directions for future research

Opportunities abound to investigate how to detect networks of complicity, neutralize their negative effects and prevent them from forming. This study demonstrated that networks of complicity thrive in many different sectors, but little is yet known about the commonalities and differences associated with these networks across sectors, which may be key to understanding their dynamics more fully. This study highlighted the importance of identifying perpetrators and network members early before they cause extreme harm. Research is needed to uncover the best methods of early detection. Once networks of complicity are identified, researchers can explore optimal ways to disband these networks without interfering with the organization’s performance as well as how network members can be rehabilitated and reintegrated into the organization in positive, productive ways.

The findings demonstrated the power of network theory in understanding the dark side of networks, but questions arise about the bright side of networks that empower and support victims, bystanders and managers. Understanding the characteristics of positive networks – how they activate, grow, evolve and dissolve and what leads to their success or failure – is fertile ground for future research. The #MeToo movement has demonstrated that victims of sex harassment can mobilize within and across organizations to provide solutions. #MeToo and similar movements provide unique opportunities to understand the role of social media in activating and shaping network evolution.

Network theory can increase understanding of how and why sex harassment persists and how efforts aimed at making positive change within organizations and in society are mobilized. However, much remains to be learned about how to best leverage network theory to the advantage of organizations and society.

Table I

Characteristics of Informants and Perpetrators

	Case characteristics
	Informant	Perpetrator
Gender
Male	15	34
Female	22	3
Role
Victim	20
Bystander	17
Position
Professional	15	2
Senior leader	10	18
Middle manager	4	8
Manager	3	1
Rank and file	5	1
Power relative to perpetrator/victim
Higher	8	34
Equal (peers)	3	2
Lower	26	1
Knowledge of sex harassment
First-hand	25
Second-hand	9
General	3

Note

1. This is the second case described by Andrew.

Appendix. Interview protocol

1. Would you describe the problematic behavior that you witnessed or heard about in your work organization?

   • Tell me about the perpetrator.

   • Tell me about the victim.

   • Tell me about any others who were involved.

2. How did you react?

3. How did others react?

4. Was anything done to try to stop the problematic behavior?

5. What was the organization like when the problematic behavior occurred?

6. What do you think enabled this problematic behavior to continue?

7. In retrospect, are there things that you would have done differently?

8. What did you learn from this experience?

9. Is there anything else that I should know or should ask about?