Introduction

Sexual harassment is widespread. Up to half of working women report experiencing sexual harassment at some point in their careers [2, 20, 32, 39, 50]. These widespread unwanted sexual behaviors take a heavy toll on those experiencing them. As demonstrated in several studies, sexual harassment can result in poor psychological, health-related, and work-related outcomes for the victim [6, 14, 15, 18, 19, 24, 32, 36]. Sexually harassed women are more likely to report psychological symptoms like depression, nervousness and anger, and physical symptoms like weight loss and sleeplessness. They also report higher levels of absenteeism and stronger turnover intentions. Sexual harassment also incurs financial costs for organizations. The federal government, for instance, estimated a loss of $189 million in a 2-year period owing to sexual harassment. These losses resulted from job turnover, medical insurance claims, absenteeism and reduced productivity [49].

Surprisingly, given this widespread occurrence and serious consequences, formal reporting of workplace harassment is relatively rare. The objective of this research is to discuss why victims of unwanted sexual behavior are reluctant to use the internal grievance process to address these discriminatory behaviors. I extend beyond the more often studied circumstances of the incident, such as the seriousness of the behavior experienced, and demographic and personal makeup of the victim, such as age and tenure in the organization, to the less studied organizational context [28]. Organizational conditions in which a victim finds herself may influence victim response to harassment. Certain structural conditions may dampen assertive responses even for victims who are conscious of and eager to exercise their legal rights. The current study examines such contextual and organizational factors, including the victim’s trust in the system and fear of reprisal.

Literature Review

Public recognition and debate about workplace sexual harassment led to positive legal developments witnessed during the last three decades. The courts presently recognize unwelcome sexual behavior at work as a form of sex discrimination violating Title VII of the Civil Rights Act, and the legal developments, coupled with the willingness on the part of the courts to award substantial damages to victims, have compelled employers to adopt policies and procedures aimed at eliminating sexual harassment in the workplace.

Today most companies/corporations have sexual harassment policies and grievance procedures in place.Footnote 1 The typical sexual harassment policy informs employees that sexual harassment is prohibited and attempts to prevent or reduce its occurrence. An internal grievance procedure enforces a company’s policy against sexual harassment by investigating harassment complaints and sanctioning harassers. These policies and procedures allow victims to call on the formal power of the organization in order to address and resolve a sexual harassment situation. Additionally, bringing an incident of sexual harassment to the attention of the employer is vitally important since notification of harassment to an employer could be crucial for the success of a later legal claim against the employer.Footnote 2

Despite the potential benefits of filing a complaint, organizational grievance policies are rarely used by sexual harassment victims [2, 13, 23, 32, 40]. In one study, only 8% of nearly 2,000 women formally reported the harassment they had experienced [23]. In another study of university students and staff regarding responses to unwanted sexual attention, only 2% reported making a formal complaint [13]. The federal workplace shows similarly low rates of reporting. Practically all government agencies surveyed by the Merit Systems Protection Board in 1994 had sexual harassment policies and complaint mechanisms in place, but only 6% of harassment victims surveyed had taken formal action against their harassers [50]. Data reported by organizations themselves show that filing of formal complaints is rare. Out of a sample of US based organizations, no formal complaint of sexual harassment was filed in 70% of the organizations within a 1-year period [42].

In recent years scholars have begun to examine victim reluctance to file sexual harassment complaints. Studies generally show that factors relating to the incident, such as the type of behavior, its severity, offensiveness, and job-related consequences associated with the harassment have an effect on reporting rates [4, 10, 13, 23, 43–45]. Studies examining the relationship between the personal characteristics of the victim—such as age, marital status, ethnicity, sex role attitudes, and sexual orientation—and reporting behavior have produced mixed results [10, 22, 33, 38, 44, 45]. For example, Rudman and colleagues [38] found that older women are more likely to report sexual harassment, while a study by Terpstra and Cook [45] suggests that younger women may be more likely to file complaints than older women. Furthermore, Schneider [39] found that lesbian women were more likely to label unwanted sexual attention as “sexual harassment” than heterosexual women while Swan and colleagues [44] found no relationship between sexual orientation and likelihood of reporting. Organizational context may help explain the failure of these individual traits to explain variation in behavior.

Several studies have also looked at how organizational politics affect the complaint process [16, 17, 27, 34, 37, 51, 52]. These studies generally point to a conflict of interest for organizations between victim empowerment and concerns over litigation. These studies show how the implementation of the complaint process by complaint handlers creates obstacles to filing of complaints. Kihnely [27] found that administrators of the formal complaint process tended to align themselves with the alleged harasser because of liability issues. Marshall [34] found that complaint handlers tended to interpret sexual harassment policies narrowly, thereby discouraging victim complaints. In their study, complaint investigators engaged in various practices that discouraged women from complaining: they dismissed complaints on the ground that the incidents were not serious even though the incidents met the definition of sexual harassment under the written policy. Complaint handlers also added steps that made complaining a less attractive option—for example, they made confronting the harasser a prerequisite to complaining.

Research also suggests that in addition to these biases inherent in the complaint handling process, the current social definition of discrimination and the operation of anti-discrimination law discourages the filing of complaints [11]. Bumiller argues that anti-discrimination law requires persons subjected to mistreatment to assume the role of the victim before they can file a claim. For the women in her sample, filing a claim became a contest between the need to maintain a positive self-image and having to assume the label of victim in order to claim a right. These women also feared that by filing a grievance they will not gain power but would lose control over a hostile situation. Bumiller’s findings suggest that, apart from the rational costs of filing a grievance, the psychological needs of victims may operate as barriers.

The literature on trust and legal compliance is of particular use for understanding how the organizational culture around sexual harassment might influence victim decisions to file sexual harassment complaints. Research looking at the association between trust of legal authorities and institutions and decisions to obey the law suggests that legitimacy and trustworthiness promote compliance [47, 48]. According to Tylor [47], persons’ decisions about legitimacy and trustworthiness depend on whether legal authorities adhere to procedural justice concerns (that is, if they follow fair procedures) rather than achieving a personally favorable outcome. The following aspects of a procedure are considered by individuals when they evaluate the fairness of legal authorities or procedures: neutrality, interpersonal aspects of the decision making process, having an opportunity to present one’s case, perceptions about the authority’s effort to be fair, and the fairness of the outcome.

There is also some evidence that trust of procedures might determine the type of dispute resolution method that employees use. Thus, in one study, women employees perceived greater opportunities for procedural justice in the formal dispute resolution process than in the informal grievance process. Accordingly, these women expressed a greater willingness to use formal grievances than settling grievances informally [25].

Based on this literature, a potential determinant of a woman’s decision to file a formal sexual harassment complaint may be the level of trust she places in the internal grievance process. However, depending on the sexual harassment climate of the specific organization, the level of trust of those who have actually encountered the grievance process and those who have not might differ. We might expect that in organizations which are not committed to EEO mandates and do not take the issue of sexual harassment seriously, women might run into obstacles in their attempt to use the internal complaint process and thus might express lower trust in such processes than women who have not encountered the actual functioning of the complaining process.

It is plausible to expect that lack of trust in the reporting process will dampen the willingness of harassment victims to file complaints. Hence, in cross sectional studies we might expect to find a positive association between lack of trust and non-reporting. If, on the other hand, the internal grievance process is dysfunctional, it is likely that we might observe a negative association in cross-sectional designs between trust and reporting—that is low trust being associated with a high rate of reporting. The few studies that have looked at the association between the perceived efficacy of complaint channels and reporting assume that the causal order is from lack of trust to lower rates of reporting [3, 38]. But, as pointed out above, it is also likely that, at least for reporters, the outcome of the sexual harassment complaint will influence perceptions of efficacy of the complaint procedures (i.e., the causal order is from reporting to trust). Figure A1 (See Appendix A) presents the possible causal pathways with regard to the association between trust in complaint channels and reporting. Whether we observe a positive or negative association between trust in complaint channels and reporting will depend, in large part, on the sexual harassment climate of the specific organization. Once a significant association between trust and reporting is observed, researchers need to do further analysis to determine the underlying organizational dynamics that in fact drive the observed association. The current study examines this potential, dynamic association between trust in grievance processes and reporting of sexual harassment.

Determining the exact causal pathway of the association between trust in reporting channels and filing of complaints is important from a policy point of view. If the avoidance of using formal channels is mainly due to unfounded fears about the fairness and efficacy of complaint channels, the employers need to focus on reassuring and building trust among employees. If, on the other hand, employees’ mistrust of the complaint channels is in response to actual experiences of retaliation or poor handling of complaints, overcoming these organizational failures might prove to be more challenging, given that employers must balance their interest in protecting the organization and avoiding legal liability with the need for victim empowerment.

A woman may also fail to formally complain about sexual harassment because she fears negative consequences. Fear of not being believed, fear of being blamed for the incident, and fear of adverse career consequences are common reasons for refraining from taking formal action [50]. Fear of retaliation by the harasser as well as the employer may also discourage the use of internal grievance processes [2, 43]. A survey of adult residents of the state of Connecticut showed that victims who pursued active strategies in response to their harassment suffered the most negative consequences—lowered evaluations, denial of promotions, and terminations [32]. Though workplace sexual harassment policies generally reassure employees against retaliation by the organization or the harasser, the above studies suggest that fear of retaliation is well grounded for many harassment victims. However, one study found that reporters of sexual harassment expressed greater concern about retaliation than non-reporters [38], again suggesting that a more complex causal cycle, such as that displayed in Fig. A1, applies because reporters may learn through experience. While greater scholarly attention has been paid to retaliation by harassers or the organization, less attention has been paid to another type of fallout that may follow a complaint: this is reprisal that does not involve the formal organizational power structure, like being ostracized by fellow coworkers or gossiped about following a complaint. The current study looks at the impact of this type of non-formal, organizational reprisal on a victim’s decision to use the internal grievance process.

The current study also attempts to replicate whether the characteristics of the harassment experience itself influence responses to sexual harassment. Research findings on the effect of harassment severity on reporting rates have been generally consistent [13, 23, 31, 32, 38, 43]. All of these studies suggest that factors associated with the harassment incident, such as its level of offensiveness and severity, explain a major portion of reporting behavior. For instance, in its study of the federal workforce, the U.S. Merit Systems Protection Board [50] found that 50% of sexual harassment victims said they failed to take formal action because they did not consider the harassing behavior to be serious enough. Brooks and Perot’s [10] study of a sample of university women found that the frequency of the behavior had an indirect effect on reporting by influencing the perceived offensiveness of the behavior. In my sample, I also expect to observe that several indicators of severity—namely the offensiveness and threat associated with the behavior and the frequency and duration of the behavior—have a positive effect on reporting.

Summary of Research Questions

The current study examines whether victim trust in the internal grievance process is associated with decisions to file sexual harassment complaints. Literature on trust and organizational justice suggests that individuals, in their evaluations of the trustworthiness of authorities and institutions, focus on both the fairness of the procedures and the fairness of outcomes [46]. Based on this literature, this study examines the effect of two indicators of trust: (1) whether the organization takes sexual harassment complaints seriously, and (2) whether the organization appropriately sanctions harassers subsequent to a finding of harassment. The study also investigates (within the limitations of current cross-sectional data) the complex causal cycle that may operate between trust and reporting. This could reveal at what stage of the sexual harassment prevention process organizations fail most or face most challenges. The current study also focuses on an organizational aspect that has not been seriously investigated but may affect victim response—fears of reprisal that do not involve the formal organizational power structure. The study looks at the impact of non-formal reprisal by examining the association between fears of coworker mistreatment following a complaint and likelihood of reporting. The study also examines how various attributes of the harassing behavior experienced affect filing of complaints. Often studies focus on one or two attributes of the behavior, but the simultaneous use four attributes—frequency, duration, offensiveness, and threat—provide a more comprehensive measure of severity of the harassing incident.

This study examines these research questions using sexual harassment data from the military. The military, with its long history of sexual harassment, its segregated and gendered job context, and the clear cut, formalized reporting processes provides an especially useful context for examining the problem of sexual harassment and victim response. Though women have served in the U.S. military throughout history, it is one of the very few organizations with explicit legal restrictions on the employment of women. Currently women are barred from 20% of all military positions—those mainly involving direct ground combat. While the percentage of women serving in the military has dramatically increased since the establishment of the All Volunteer Force in 1973, by the end of 1998 the percentage of women serving in the military still stood at 14%. The legal restrictions on the employment of women as well as the definition of the military job and the culture that exalts heterosexual masculinity promote an image of women service members as second-class members. This may in turn encourage ‘gate-keeping’ behavior such as sexual harassment [9]. The predominantly male, masculinized work environment of the military is a suitable medium to examine how gender segregated work cultures and their associated pressures and dynamics affect women employees’ responses to harassment.

Distrust of formal grievance procedures and fears of reprisal are problems common to all sexual harassment victims irrespective of the type of organization they serve. Thus, a survey of U.S. federal workers suggests that problems associated with the formal complaint procedure are among the reasons for not filing a formal complaint. In this 1995 survey, 29% of the sexual harassment victims said that filing a formal complaint would make the work situation unpleasant, 20% said that nothing would be done, and 19% felt that their confidentiality would not be protected [50]. A study of university students also shows that, retaliation, being blamed for the incident, and the perpetrator not being disciplined are among reasons for not filing a formal report [2]. While this study is most helpful in shedding light on how trust in grievance processes might affect victim response in male dominated, gendered work contexts, such as the military, firefighting, and law enforcement, the findings on this possible association and the underlying causal mechanisms will be of relevance to all organizations where sexual harassment is a problem.

Data

To examine the research questions, the current study uses data from the “2002 Status of Armed Forces: Workplace and Gender Relations Survey (2002 WGRS)” conduced by the Defense Manpower Data Center (DMDC). The 2002 WGRS sample consists of 60,416 active-duty members of the Army, Navy, Marine Corps, Air Force, and Coast Guard.Footnote 3 The overall purpose of the 2002 WGRS was to document the extent to which Service members reported experiencing unwanted sexual attention, the details surrounding those events, and Service members’ perceptions of the effectiveness of sexual harassment policies and programs [30]. A total of 19,960 eligible members returned usable surveys yielding a response rate of 36%, of which number 10,309 were females. My study is limited to the sub-sample of women respondents from the survey because the female sample provides a good context for examining responses to sexual harassment, given that sexual harassment is predominantly experienced by women [20, 50].

Measures

The 2002 WGRS measures sexual harassment experiences of military women by asking respondents about “sex/gender related talk or behavior” they experienced during the 12 months prior to the survey.Footnote 4 The survey questions on victim perceptions of, and responses to, sexual harassment focus on the harassing situation that occurred during this period and “had the greatest effect” on the respondent.

Dependent Variable

Reporting

The dependant variable ‘reporting’ looks at whether the respondent reported the incident of unwanted sexual behavior ‘that had the greatest effect on her’ to at least one of the following persons/offices: respondent’s immediate supervisor, someone else in the respondent’s chain of command, special office responsible for handling harassment complaints or, other office or person responsible for following up such complaints. The coding category “yes” indicates that the respondent reported the behavior to at least one such person/office, while coding category “no” indicates that the respondent did not report the behavior.

Independent Variables

Attributes of the Sexually Harassing Behavior

I code four characteristics that victims were asked to report about the harassing behavior. Two of the attributes of interest are the offensiveness and the level of threat associated with the harassing behavior. Both of these variables are coded using five categories: “not at all,” “slightly,” “moderately,” “very,” and “extremely.” Frequency of harassing behavior is measured by asking respondents how often the alleged behaviors took place, and the categories for this variable are coded “once,” “occasionally,” “frequently,” “almost every day,” and “more than once a day.” The duration of the behavior is measured with seven categories—“less than 1 week,” “1 week to less than 1 month,” “1 month to less than 3 months,” “3 months to less than 6 months,” “6 months to less than 9 months,” “9 months to less than 12 months,” and “12 months or more.”

Trust in Reporting Channels and Fear of Non-Formal Reprisal or Retaliation

The study measures employee trust in complaint channels using two indicators: whether the organization treats complaints of sexual harassment seriously and whether the organization takes corrective action in response to a complaint. The other indicator of the sexual harassment climate of the organization—fear of non-formal reprisal following a complaint—is measured by examining whether victims of harassment entertained fears of co-worker mistreatment were they to file a formal complaint.

The measures of trust in complaint channels and fear of co-worker mistreatment were constructed in the following manner. In the 2002 WGRS, women were asked about three hypothetical harassment situations. These three situations were, (1) attempts by a coworker to draw others into discussion of sexual matters, (2) a coworker who continues to ask for dates despite being refused, and (3) a suggestion by a supervisor that the way to get good assignments is to be sexually cooperative with him/her. For each situation, respondents were asked whether a complaint by a coworker about the harassing behavior would be taken seriously by the organization, whether corrective action would be taken, and whether the complainant would be treated badly by coworkers. For each, the response categories were, “strongly agree,” “agree,” “neither agree nor disagree,” “disagree,” and “strongly disagree.” I created summary items by summing responses across the three situations, resulting in measures of “complaint taken seriously,” “corrective action taken,” and “complainant treated badly.” Each of the three summary measures for the three predictors ranges from 3 to 15 with higher scores indicating stronger agreement with the statement. To simplify the data analysis and interpretation, response choices were collapsed such that 3–6 was coded as “1 = weakly agree,” 7–11 was coded as “2 = moderately agree,” and 12–15 was coded as “3 = strongly agree.”

Control Variables

Race, marital status, education, pay grade, years of military service, and sexual harassment training were used as control variables. Dummy variables were created for race, marital status, and education. The four race dummies are “Hispanic,” “Non-Hispanic white,” “Non-Hispanic Black or African American,” and “Non-Hispanics reporting some other race or Non-Hispanics reporting more than one race.” The two dummy variables for marital status are “married or separated” and “never married, divorced, or widowed,” while dummies for education are “high school diploma or less,” “some college but no 4-year college degree,” and “4-year college degree or more.”

The study measures pay grade with five categories, “E1-E4,” “E5-E9,” “W1-W5,” “O1-O3,” and “O4-O6” coded “1” through “5.” The variable “years of service” measures the years of active-duty service completed with four items—“less than 6 years,” “6 years to less than 10 years,” “10 years to less than 20 years,” and “20 years or more,” coded “1” through “4.” The study also measures whether the respondent received sexual harassment training during the 12 months preceding the survey.

Findings

The demographic distribution of the female WGRS sample was as follows. Fifty five percent of the sample were white, 24% were Black, 11% were Hispanic, and 10% were some “other” race. Forty eight percent of the sample had some college education and another 34% had a 4-year college degree or more education. Fifty three percent of these respondents were either married or separated, while 47% were either never married, divorced, or widowed.

Findings indicate that despite the considerable public attention paid to the problem of sexual harassment in the military including congressional hearings on the issue and wide publicity given to recent incidents of harassment in the Service branches and military academies, sexual harassment still remains a fairly widespread problem in the military. Sixty one percent of women in the WGRS sample reported experiencing at least one type of sexually harassing behavior within a 12-month period. The most common types of sexually harassing behaviors reported by military women were offensive gender terms (39%), sexual stories or jokes (32%), and being treated differently based on gender (34%). Five percent of the women said they were bribed with offers of special treatment for engaging in sexual behavior, while 3% of women reported being threatened with retaliation for refusing sexual cooperation. Two percent of the women also reported attempted sex against their will. An overwhelming proportion of these harassing behaviors was carried out by men. Thus, 85% of women reported that the person(s) involved in the harassment was male while 12% said the persons involved were both males and females. Unlike in the civilian workplace where coworker harassment is more common, in the military supervisory harassment appears to be somewhat more widespread than co-worker harassment. Seventy percent of the military women said the harassment involved a military superior. Sixty four percent of women reported that the harassment involved a military coworker of equal rank.

As for the characteristics of the harassing behavior, most women found the behaviors they experienced to be offensive. Forty eight percent of the women reported that the behavior was slightly to moderately offensive, and 35% said the behavior was very or extremely offensive. A third of the military women also reported that they felt threatened by the behavior. Sixty nine percent of the women said the harassing behaviors happened occasionally or frequently, while for another 8% these behaviors were a daily occurrence. As for the duration of the harassing behavior, 35% said the behavior lasted less than a week, 36% said the harassment continued for a week to 6 months, while 29% said the behavior lasted in excess of 6 months.

Most Service women trust the internal grievance processes provided by their Service or unit. Thus, out of the overall sample, 70% strongly agreed that the military takes sexual harassment complaints seriously, while 30% of the sample expressed moderate or weak agreement on the issue. On the issue of whether the military takes remedial action in response to a complaint, 64% strongly agreed that the organization takes action, while 35% weakly or moderately agreed with the statement. However, military women who had filed a sexual harassment complaint expressed lower trust than women who had not used the complaining process. This fact is especially significant, given the focus of the current study. More than half (53%) of the women in the WGRS also expressed fears of non-formal reprisal by coworkers due to a complaint.

Compared to studies of reporting rates in private and federal organizations that generally found complaint rates below 10%, the complaint rate in the military is significantly higher. Twenty nine percent of military women who experience sexually harassing behavior complained to a superior or other office designated to handle such complaints. A considerable proportion of women (45%) who experience sexually harassing behaviors also labeled such behaviors as sexual harassment. There may be several possible explanations for this comparatively high rate of identification and reporting in the military. For one, the formal hierarchy of the military may invoke a greater sense of risk/vulnerability among female victims of sexual harassment. Self-help measures of confrontation or avoidance also may be viewed by victims as less effective in this context, encouraging them to seek organizational intervention. It is also likely that highly publicized incidents of sexual harassment inside the military such as the Tailhook and Aberdeen Proving Grounds incidents have highlighted the problem of sexual harassment as illegal and remediable, encouraging the identification of the behaviors and filing of complaints. Feminist-led activism inside the military in recent decades also may have changed the way the military operates. For example, feminist led protest has been instrumental in the integration of women into male only occupational specialties and in the repeal of the combat exclusion laws [26]. It is likely that this feminist activism and consequent efforts to reduce gender inequality in the military had a positive impact on eliciting more assertive responses on the part of military women.

Most military women have undergone sexual harassment training. Of the survey respondents, 76% had received sexual harassment training at some point during the year preceding the survey. This indicates that the military takes steps to deter sexual harassment at least at the front end of the prevention process by increasing employee awareness of sexual harassment, the part of the prevention process that is arguably less impacted by organizational politics.

In order to examine the net relationships between the predictor variables and reporting, multivariate analysis was conducted using logistic regressions. Table A1 presents the effects of attributes of the harassing behavior on the log-odds (i.e., likelihood) of reporting sexual harassment, the effects of trust in complaint procedures on the log-odds of reporting harassment, and the effects of fear of informal reprisal on the log-odds of reporting harassment. The full model, Model 4, in Table A1 shows the net associations between these predictors and reporting, controlling for race, marital status, education, pay grade, period of service, and sexual harassment training. According to Model 4, all four attributes of the harassing behavior are significantly associated with reporting. Women who experience more offensive, more threatening, and more frequent behavior are more likely to report; women who experience harassment that continues for longer periods of time are also more likely to report such harassment. My study also shows that severity of the harassing behavior experienced (defined as the combined effect of offensiveness, frequency, duration of the behavior, and the threat accompanying the behavior) also influences reporting rates. These results are consistent with previous research that generally finds that reporting rates increase with severity of harassment [13, 23, 38, 43].

According to the same model, the first indicator of trust in complaint channels—the belief that a complaint would be taken seriously by authorities—is negatively associated with reporting. However, the second indicator of trust in complaint procedures—the belief that corrective action would be taken in response to a complaint—is not significantly related to reporting. I hypothesized that fears of informal retaliation will discourage complaints. Model 4, which points to a negative association between fears of coworker mistreatment and the likelihood of filing a complaint appears to support this hypothesis. This finding is consistent with most previous research indicating that fear of negative outcomes, such as fear of adverse career consequences, or being blamed for the incident are a major deterrent to reporting [2, 43, 50].

Model 4 also shows that women who received sexual harassment training during the 12 months preceding the survey are more likely to make a harassment complaint than those who did not receive such training. A few of the other control variables are also associated with reporting. Thus, results show that married or separated women are more likely to report harassment than never married, divorced, or widowed women. According to results, race and education level, however, have no significant impact on reporting. Results also show that women who have had a longer career with the military are more likely to report harassment. While one may expect that formal occupational power of a victim would encourage more assertive responses, the data show that this is not the case in the military. According to data, higher occupational status appears to discourage filing of complaints. Again, this may not be surprising, given that in male dominated institutions such as the military high status women will have few allies who would support their actions against harassers. Also they might be more invested in their careers and have more to lose if their complaints upset the military leadership.

Discussion

The main objective of this study was to understand the ways in which the organizational climate influences victim decisions to file formal sexual harassment complaints. The findings indicate that certain aspects of the organizational context—namely employee perceptions of the grievance process, actual functioning of that process, and the fall-out associated with utilizing this organizational remedy—influence victim response.

Sexual harassment incurs substantial costs both to its victims and the organizations. Filing a grievance is the last recourse available within the organizational boundaries to a victim when other attempts to stop the harassment have failed. Organizational intervention to stop harassment can also discourage other desperate attempts at avoidance such as quitting a job or absenteeism, outcomes that are costly to the employee as well as the organization. My study reveals that participation in sexual harassment training programs encourages the seeking of organizational intervention. Thus, it would appear that harassment training empowers victims to come forward, and the fact that most military women have received sexual harassment training may be a possible explanation for the relatively high rate of reporting in the military compared to other workplaces. My study, however, points to ways in which current training programs might be improved further. My findings suggest that fears of coworker disapproval keep women from seeking organizational intervention. This indicates that training programs should broaden their focus, so that they are aimed not only at addressing victim fears of formal or obvious methods of retaliation, such as loss of promotions, bad evaluations, being labeled a “troublemaker” by the organization, or retaliation by the perpetrator, but at negative outcomes that may not be readily anticipated or not related to the formal workplace authority structure.

Rudman and colleagues [38] found that, compared to reporters of sexual harassment, non-reporters were more likely to express concerns about the efficacy of filing a grievance, possibly indicating that lack of trust causes victims to avoid the internal grievance process. My results on the other hand showed a positive association between doubts about the efficacy of complaint channels and likelihood of filing a complaint. One interpretation of this finding could be that women who mistrust the complaint process are more likely to take formal action, but this interpretation seems implausible on face value. A more plausible interpretation is that the observed positive association is due to most reporters becoming disenchanted with the grievance process due to their own experiences with its operation.

In fact, in the current sample, a considerable number of military women who filed formal complaints expressed dissatisfaction with the complaint handling process. Out of the 1,341 military women who filed a complaint of sexual harassment, 35% said they were dissatisfied with the outcome of the complaint, while 33% said they were “neither satisfied nor dissatisfied.” Fifty percent of the reporters said nothing was done about the complaint, 45% said no disciplinary action was taken against the perpetrator. Another 36% said that the harassing situation was not corrected. Seven percent of the women who filed a complaint also said that the organization took action against them. Chi-square analyses I conducted also showed that these unsatisfying outcomes were significantly associated with women’s perceptions of the internal grievance process: women who experienced these unsatisfying outcomes were more likely to express doubts about the complaint mechanisms in general than women who expressed satisfaction with the handling of their complaints. While the cross sectional nature of the current data set cannot conclusively determine the issue, these results appear to provide some support for the causal cycle discussed earlier (Fig. A1). It would appear that while the military through the high visibility given to the issue of sexual harassment and their sexual harassment training programs encourages Service women to use the complaint procedure, it often fails them in its implementation. This might be due to a lack of leadership commitment in the military to equal employment opportunities for women. Also, the gendered job context of the military, with its code of silence and pressures to cover up transgressions, might be a root cause for these breakdowns. These deficiencies in the functioning of the internal grievance process in the military possibly explain the observed finding of a negative association between trust in complaint channels and reporting in the military.

The findings in the study bear on more general difficulties in resolving sexual harassment problems. The legal system penalizes victims who fail to report harassment. Thus, often in cases of hostile environment harassment carried out by coworkers, the plaintiff has to prove that she reported the harassment to the employer [29]. In the case of hostile environment harassment carried out by supervisors, courts, in their implementation of the affirmative defense laid out by the Supreme Court, generally tend to dismiss cases when the victim failed to report the harassment, provided the employer had a complaint procedure in place [5, 21, 35]. However, the courts should not presume that the victim unreasonably failed to report the harassment, if such failure is common and arguably rational, given the sexual harassment climate of the defending organization [21]. In the current sample, a large proportion of women said they were unsatisfied with the actual implementation of the complaint procedure. Additionally, almost half of the women believed that a person filing a sexual harassment complaint would face coworker mistreatment. Such fears deter filing of formal complaints. The findings in the current study highlight why courts should, in their adjudication of sexual harassment lawsuits, look deeper into the sexual harassment climate of the organization and the specific reasons as to why the victim failed to use the internal grievance process and not assume that the mere presence of a sexual harassment policy is sufficient to encourage reporting.

The current study also underlines the difficulties of determining the dynamic relationship between efficacy of internal grievance processes and their use, using cross sectional data and why sexual harassment research should advance from current cross-sectional research designs to longitudinal designs. As pointed out in Fig. A1, two causal pathways are possible in the association between trust in complaint channels and reporting, and which of these two dynamics in fact goes on in an organization depends on the sexual harassment climate of the specific organization. While cross sectional data can tell us whether an association exists between the predictor of interest and reporting, and can provide supportive evidence for the presence (or absence) of the reverse causal path in an organization, such data cannot conclusively determine the causal direction of this association.

Current research on the reporting of sexual harassment, when assessing victim perceptions of in-house grievance processes and potential negative outcomes suffers from problems of measurement validity. Surveys either ask respondents to recall their feelings prior to filing of the complaint (for example, the study by Rudman and colleagues) or surveys try to assess how victims feel about complaint channels at the time of the survey (as done in the 2002 WGRS survey). Both suffer from the limitation that the outcome of the complaint—whether the victim was satisfied with the way the investigation was conducted or with the final outcome—is likely to color responses to these survey questions. Because of this, the causal order with regard to the association between trust in complaint channels and probability of filing a complaint cannot be conclusively proved using the cross sectional data used in these studies.

The way to overcome this problem in the future is to conduct longitudinal studies that could produce valid measures of victim sentiments prior to filing a complaint. Data collected at two points in time from the same organization should lend itself to a determination of the causal pathway with regard to the association between efficacy of complaint channels and reporting rates. Consider surveying respondents about (1) their current level of trust in the sexual harassment complaint channels, (2) their experience with harassing behavior, filing of complaints, and complaint outcomes during the 12 months preceding the survey, and (3) their level of satisfaction with complaint outcome at both time points, T1 and T2, separated by 1 year. In order to determine how trust affects reporting (the first leg of Fig. A1), trust at T1 can be regressed on the likelihood of the respondent filing a complaint between T1 and T2, which is measured at T2. This predicted association takes the following form: reportT2 = α + β * trustT1. To examine whether the reverse causal path (the second leg of Fig. A1) operates in an organization, the outcomes of a complaint that was filed between T1 and T2, and complainant satisfaction with the investigation process and final outcome can be regressed on trust in complaint channels at T2. This predicted association takes the following form: trustT2 = α + β * complaint outcome T2. In this model, it is not possible to know for certain whether the change in trust occurred before or after the complaint outcome—it is possible that a decline in trust occurred soon after T1 and stayed low until T2, and that the poor complaint outcome happened midway through. To overcome this limitation, a qualitative component could be added to the research design to probe the timing of and the reason for the change in trust. At T2, personal interview data can be gathered from respondents who filed a complaint and reported a change in trust between T1 and T2 about the outcome of the complaint, timing of the complaint outcome, outcome satisfaction, and the timing of the change in trust. This design would allow us to test whether only the first leg of Fig. A1 operates in an organization or whether both processes in Fig. A1 occur in the organization.

The current study suffers from certain limitations that should be kept in mind when interpreting and generalizing the findings. The 2002 WGRS survey was limited to personnel currently serving in the military and does not include those who may have left the military because of sexual harassment. Additionally, only 36% of the survey participants completed the survey. Perhaps women who have experienced sexual harassment or have suffered negative outcomes as a result of a complaint are more likely to take part in a harassment survey than women who had no such adverse experiences. These sample biases should be kept in mind when interpreting the findings. Additionally, the legal environment and the organizational culture of the sample organization and the civilian workplace might differ on various dimensions. In the military, the job has traditionally been defined as masculine, the complete functioning of the organization is based on obeying orders, and there is strong pressure to conform to the existing institutional culture [53]. Also, while the sexual harassment law that applies to the civilian workplace and the military share many commonalities, they are not completely identical.Footnote 5 For these reasons, caution should be exercised when generalizing the findings to non-military work settings.

While improved research designs in the future will broaden our understanding of obstacles to using the internal grievance process, the current study reveals that the organizational climate with regard to sexual harassment is a vital determinant in women’s decisions to file formal complaints. Work environments where employees fear the possibility of coworker backlash discourage formal complaining. Similarly, the study suggests that poor handling of sexual harassment complaints by an organization can erode employee trust in these grievance procedures. This mistrust may in turn influence victim response to unwanted sexual behavior. The study also highlights the limitations of current research designs and suggests ways in which these limitations may be overcome in the future so that the complex dynamics involved in accessing the internal grievance procedures can be more clearly understood.