Legal Consciousness and Responses to Sexual Harassment

Abstract

Studies of legal mobilization often focus on people who have perceived some wrong, but rarely consider the process that selects them into the pool of potential “mobilizers.” Similarly, studies of victimization or targeting rarely go on to consider what people do about the wrong, or why some targets come forward and others remain silent. We here integrate sociolegal, feminist, and criminological theories in a conceptual model that treats experiencing sexual harassment and mobilizing in response to it as interrelated processes. We then link these two processes by modeling them as jointly determined outcomes and examine their connections using interviews with a subset of our survey respondents. Our results suggest that targets of harassment are selected, in part, because they are least likely to tell others about the experience. Strategies that workers employ to cope with and confront harassment are also discussed. We find that traditional formal/informal dichotomies of mobilization responses may not fully account for the range of ways individuals respond to harassment, and we propose a preliminary typology of responses.

How do individuals respond when they feel their rights have been violated? Do those who perceive a wrong simply tell the wrong-doer, do they tell others, or do they ignore it? Following Ewick and Silbey’s (1998) groundbreaking work on the common place of law, a growing body of literature has taken up these and related questions of legal consciousness and mobilization (e.g., Albiston 2005; Connolly 2002; Fleury-Steiner 2003; Hoffman 2005; Litowitz 2000; Lovell 2006; Marshall 2005; Marshall and Barclay 2003; Richman 2001). Central to these studies are the interrelated processes by which individuals first come to experience or perceive some wrong and then do something about it. In this study of sexual harassment, we attempt to link these two processes – in this case, subjectively experiencing sexual harassment and then going on to tell a supervisor or government agency that it has occurred. We also examine how individuals respond to harassment and why they employ particular response strategies. We find multidimensional responses to harassment, such that traditional formal/informal dichotomies may not adequately account for diversity of strategies that individuals employ. We therefore propose a preliminary multidimensional typology of mobilization responses.

We first describe legal consciousness, targeting, and mobilization within the context of sexual harassment. Next we outline a general model of sexual harassment and legal mobilization based on the sociology of law, feminist scholarship, and criminology. Using survey data, we then test predictions from the model and consider harassment and mobilization as jointly determined outcomes. In the selection portion of the model, we ask who experiences a behavior or behaviors they define as sexual harassment. In the substantive portion, we model the likelihood that individuals will notify a supervisor or government agency about these experiences, before and after adjusting for the non-random selection of targets using a bivariate probit technique. We contextualize quantitative findings using data from interviews with a subset of the survey respondents. The interview data show the multiple, varied forms mobilization takes. Finally, we consider our results in the context of mobilization and legal consciousness more generally.

TARGETING, MOBILIZATION, AND THE CONTEXT OF SEXUAL HARASSMENT

Legal Consciousness and Mobilization

Individuals’ perceptions of legality affect whether and how they mobilize the law. Ewick and Silbey’s (1998) work on legal consciousness has been a catalyst for a line of research on the relationship between formal institutions and everyday perceptions of law (see also Merry 1990). Legal consciousness concerns individuals’ awareness and understanding of the law and legal rights, a process whereby people experience and understand “meanings, sources of authority, and cultural practices that are commonly recognized as legal” (1998:22). Legal consciousness thus drives how individuals understand, experience, and respond to the law and legality.

Cultural schemas as well as individual resources shape understandings of law. Theories of legal consciousness therefore identify cultural and social-psychological factors in shaping perceptions of justice and responses to perceived injustices (Bumiller 1988; Ewick and Silbey 1998; Felstiner, Abel, and Sarat, 1980–81; Gallagher 2006; Kritzer et al. 1991). Together, socioeconomic and political power affect individuals’ purposes, stakes, and constraints, which in turn affect legal consciousness. At the individual level, these theories emphasize qualities such as self-efficacy or self-esteem in understanding the subjective processes through which experiences become grievances and grievances become disputes (Felstiner et al. 1980–81). A more recent study adds to these theories, suggesting that perceptions and expressions of grievances are linked not only to individuals’ self-esteem but to their emotions, such as anger or sadness, about their experience (Cormier 2007). While most studies of legal consciousness focus on those in less privileged social positions, such as the working class, people of color, and women (Munkres 2008; for exceptions see Nielson 2000; Quinn 2000), our study will include women and men from a variety of social backgrounds and whose relative positions of power therefore vary.

Importantly, the concept of legal consciousness allows for the expression and examination of legality in ways not officially recognized by the law. Apart from formal mobilization such as lawsuits, law can also be invoked informally, creating a conceptual distinction between “the law” and “legality” (Hull 2003). The law represents a set of formal institutions, whereas legality refers to how those formal institutions shape and are shaped by everyday beliefs and actions. We can thus observe the “imprint of law” (Ewick and Silbey 1998:20) in social interactions and in individuals’ perceptions of and responses to those interactions. Of course, prior to mobilization some precipitating event must occur, such as being burglarized, experiencing discrimination, or disputing with one’s neighbor. Legal consciousness is thus a dynamic process wherein people experience events in their lives, make sense of those events, and respond to them. In doing so, individuals not only express a perspective on legality, they shape its meaning and boundaries. The same forces are likely to drive both a mobilizing response and the perception that an event or dispute warrants such a response.

Legal consciousness refers to how people perceive the law and how such perceptions are translated into action. One study distinguishes two categories of such actions: either they qualify as “formal” actions, such as filing complaints, or as “informal” actions, such as directly confronting those with whom a worker has a complaint (Hoffman 2003). Actions falling outside of either category include tolerating grievances and leaving a position. In this paper, we too distinguish the multiple ways individuals invoke the law in response to grievances at work. We examine workers’ experiences with sexually harassing behaviors, asking how legal consciousness shapes mobilization in response to those behaviors. In our statistical models, we measure “formal” responses, predicting the probability of reporting the harassment to a supervisor or government authority. In our qualitative analysis, we consider forms of mobilization that may fall outside traditional formal/informal dichotomies. Before proceeding, however, we describe sexual harassment as a site for examining consciousness and mobilization.

The Context of Sexual Harassment

Sexual harassment became recognized as a form of sex discrimination in the 1970s, with scholarly research on the topic flourishing since publication of Catherine MacKinnon’s (1979) influential Sexual Harassment of Working Women. As legal and scholarly understandings of sexual harassment developed, so too did informal public understandings. In particular, public awareness increased dramatically during the 1991 televised hearings of Supreme Court nominee Clarence Thomas. Although working adults may have difficulty reciting formal legal definitions of sexual harassment, many have a strong abstract comprehension of the concept (Quinn 2000). The abundance of popular films (e.g., Working Girl in 1988, Disclosure in 1994; and North Country in 2005) and advice publications (e.g., Bravo and Cassedy 1999; Rutter 1997; Webb 1999) addressing sexual harassment further suggest its pervasiveness in public consciousness. The strong public association of sexual harassment with legality make it an apt site for analyzing the relationships between targeting, legal consciousness, and mobilization.

Bringing together sociolegal and social movements literatures, Marshall (2003) employs the social movements concept of “framing” to describe how women workers construct sexual harassment from within an injustice, management, or sexual freedom frame (see also Pedriana 2006). In other work focused on legal consciousness and harassment policies, Marshall (2001) suggests that even the slightest amount of “oppositional consciousness” on the part of targets can generate policy changes. Other sexual harassment research suggests that consciousness about harassment varies with social and demographic characteristics such as gender and race (Nielsen 2000; Quinn 2000; Welsh et al. 2006). Most typically, however, studies linking legal consciousness and sexual harassment do not go on to consider why some targets come forward while others remain silent. Those that do consider mobilization are often limited to samples of people who have experienced harassment, not considering the process of targeting or victimization that selects these individuals into the pool of potential “mobilizers.” Because legal consciousness is linked to multiple processes -- experiencing a wrong, framing it for oneself and others, and formulating a response -- a more comprehensive analysis would consider both the predictors of sexual harassment and individuals’ understandings of and responses to harassment.

Exclusive focus on the predictors of sexual harassment targeting has been controversial, in part because of an understandable reluctance to “blame the victims.” Further, as Quinn rightly argues, victim-focused research can lead to “an incomplete, and potentially distorted, understanding of sexual harassment perpetration” (2005). Without denying the need for more research on perpetrators and the cultural contexts in which harassment thrives, information about who is targeted remains critical to assessing risks, crafting policy interventions, and understanding why some targets come forward while others remain silent.

Gender, age, and organizational culture are among the few well-established correlates of sexual harassment (Morgan 1999; Welsh 1999). Harassment is especially salient in certain occupations and industry sectors, particularly sales (Collinson and Collinson 1996; Morgan and Martin 2006) and restaurant work (Giuffre and Williams 1994; Lerum 2004). Moreover, supportive work cultures, such as those with high coworker solidarity and sympathetic supervisors, have been linked to lower incidences of harassment (Chamberlain et al. 2008; DeCoster, Estes, and Mueller 1999). The effect of work friendships on reactions to harassment, however, may vary by gender. Morrison (2009) reports that women are more likely to describe the social and emotional support gained from work friends while men describe the career-related benefits of friends, suggesting that close work friends may facilitate mobilization for women to a greater extent than for men. The detailed accounts of harassment experiences provided in qualitative studies also suggest that work friendships affect mobilization choices (Morgan 1999; Stambaugh 1997). For example, while close friendships with coworkers may prevent sexual harassment from occurring at all, individuals in supportive work settings may seek support and advice from work friends if and when harassment does occur.

TOWARD A GENERAL MODEL OF HARASSMENT AND MOBILIZATION

Despite recent advances, there is “no unified theoretical framework” of sexual harassment (Welsh 1999:186). As one study concludes, “data have far outstripped theory, in both quantity and quality” in this area of research (Gelfand, Fitzgerald, and Drasgow 1995) and most extant theories have focused only on the occurrence of sexual harassment. What we do know about outcomes remains unconnected to the structural and social-psychological contexts of disputants’ lives (Felstiner et al. 1980–81). We view harassment and mobilization as interrelated processes, drawing on sociocultural, routine activities, and sociolegal theories. We do not attempt a critical test of competing hypotheses because, in general, these frameworks are logically consistent and mutually compatible.

Sociocultural models stress gender-based power differentials as the primary predictor of targeting (Barr 1993) and suggest that factors such as workplace authority determine who will experience harassment. Some argue that women supervisors are more likely to be harassed, others that power makes them less likely to experience harassment (see, e.g., Gruber 1998; Rospenda at al. 1998; Simpson and Ellis 1996). Sociocultural models emphasize macro-level cultural understandings of gender and sexuality that determine how individual characteristics shape harassment experiences, the meaning of these experiences to targets, and the choices targets make about mobilization. In light of existing gender and power relations, and prior research (Baker, Terpstra, and Larntz 1990), we expect women to perceive more harassment than men and to be more likely to mobilize in response to it. When presented with a series of sexual harassment scenarios, women are more likely than men to react – either physically or verbally, or by reporting the behavior – to more serious scenarios such as fondling or sexual propositions (Baker, Terpstra, and Larntz 1990). At the same time, however, women were more likely than men to ignore less threatening behaviors, such as directed gestures or wolf-whistling. Gender may thus interact with situational factors to affect mobilization choices. Baker and colleagues also found religiosity significantly affected mobilization, as those with high levels of religiosity were more likely to report particular behaviors than to directly confront harassers.

Routine activities theory attributes targeting to the presence of motivated offenders, the proximity of vulnerable targets, and the absence of solidaristic work groups (DeCoster et al. 1999). Based on this theory, we would expect those who work more hours to be more vulnerable to harassment because of greater opportunity or exposure to risk. Close friendships at work could also increase targeting opportunity, though such friendships may also provide guardianship that reduces harassment. Further, based on routine activities theory we expect a supportive supervisor – one who is willing to listen to problems – to provide guardianship that prevents harassment and encouragement for those who come forward with complaints.

Theories of legal consciousness (Ewick and Silbey 1998) and dispute transformation (Felstiner et al. 1980–81) underscore social position and social psychological factors, suggesting that efficacy, socioeconomic resources, and workplace relationships shape perceptions of having been wronged and subsequent responses. For example, Baker et al. (1990) find that self-efficacy predicts mobilization in a rape scenario, such that individuals with an internal locus of control were more likely to react physically or verbally. We expect self-efficacy to be negatively related to harassment and positively related to mobilization. Similarly, we expect education to influence mobilization, as more educated individuals may be more knowledgeable about their rights and the process of filing a formal complaint, and therefore more capable of taking formal action in response to harassment. Furthermore, workers of lower status may be disproportionately targeted, in part because their financial vulnerability and unwillingness to jeopardize their positions may have a chilling effect on mobilization options. Workplace training and knowledge of harassment policies also affect how individuals define sexual harassment and make sense of their own work experiences (Tinkler 2008). Workers with extensive on-the-job training may be more aware of sexual harassment policies and therefore more likely to perceive it and mobilize in response to it. Because legal consciousness theories emphasize that experience with the law shapes consciousness, we would also expect prior experience with discrimination to alter perceptions of and responses to sexual harassment.

Drawing from the above perspectives, we propose a statistical and theoretical model of sexual harassment that considers both targeting and mobilization. Organization-level characteristics and routine work activities, such as work hours and interactions with supervisors and coworkers, structure the opportunities to experience and to report harassment. Individual-level resources, such as self-efficacy, education, and religiosity (Terpstra and Baker 1986) also shape a target’s willingness to name and protest harassment when it occurs.

As the preceding discussion makes clear, similar factors are likely to affect the probability of perceiving harassment and the probability of notifying others in response to the experience. We conceptualize and measure these processes using a two-step sample selection approach. In the selection portion of the model, we ask who experiences sexual harassment or becomes the target of harassing experiences. In the substantive portion, we model the likelihood that individuals will notify others about these experiences, before and after adjusting for the non-random selection of targets. We then draw upon in-depth interview evidence to contextualize and extend the findings from our statistical models. Our qualitative interviews help answer questions about how and why harassment and mobilization occur.

DATA, MEASURES, AND METHODOLOGY

We analyze data from the Youth Development Study (YDS), a prospective longitudinal investigation of adolescents in St. Paul, Minnesota. Beginning with a random sample of 1,010 ninth graders in St. Paul, Minnesota in 1988, the YDS has since been administered annually. With about 74 percent of the panel identifying as white, 10 percent African-American, 5 percent Hispanic, and 4 percent Asian, the sample well represents the St. Paul community (Finch et al. 1991; Mortimer 2003). Sexual harassment items were placed on the survey in 1999, when respondents were 25–26 years old. Intensive interviews were conducted with a subset of the survey respondents in 2002 and 2003. In previous research using these data, we found significant levels of sexually harassing behaviors reported by both males and females. We therefore examine the entire sample, but also test for interactions between gender and key independent variables. To our knowledge, no other longitudinal data set contains a set of behavioral sexual harassment items to indicate targeting and broad measures of mobilization.

Measurement

To measure sexual harassment we use a facet-free global item, asking whether respondents considered any of the behaviors they experienced to be sexual harassment.¹ To measure mobilization, we asked respondents whether they notified supervisors or government agencies about these behaviors. To predict this form of formal mobilization, we include a measure of informal mobilization (reporting a harassing behavior to a coworker) and a lag for past formal mobilization (reporting harassing workplace behavior experienced during high school to a supervisor or government authority). Descriptive statistics for these and other measures are provided in Table 1.

Table 1.

Descriptive Statistics

Variable	Description	Coding	Mean	SD
Subjective Harassment
Perceived Harassment	% that consider their experiences to be sexual harassment	0 = No 1 = Yes	24.69%
Objective Measures of Sexual Harassment
Offensive Joking	% experiencing offensive jokes/remarks about themselves during adulthood	0 = No 1 = Yes	36.12%
Personal Questions	% experiencing questions about their private life during adulthood	0 = No 1 = Yes	41.78%
Invasion of Space	% experiencing staring or invasion of space during adulthood	0 = No 1 = Yes	23.89%
Unwanted Touching	% experiencing unwanted touching during adulthood	0 = No 1 = Yes	9.83%
Offensive Materials	% experiencing exposure to offensive material during adulthood	0 = No 1 = Yes	7.55%
Physical Assault	% experiencing physical assault during adulthood	0 = No 1 = Yes	2.84%
Total Adulthood Harassment	Number of objective measures experienced during adulthood	0 = Low 6 = High	1.20	1.32
Total High School Harassment	Number of objective measures experienced during high school	0 = Low 6 = High	.75	1.22
Mobilization
Formal Mobilization	% who reported harassment to	0 = No	35.00% of
During Young Adulthood	supervisor or government authority	1 = Yes	harassed
Informal Mobilization in	% who reported harassment to a	0 = No	35.36% of
Young Adulthood	coworker during young adulthood	1 = Yes	harassed
Formal Mobilization	% who reported harassment to	0 = No	14.36% of
During High School	supervisor or government authority	1 = Yes	harassed
Individual-level
Male	Self-reported gender of respondent	0 = Female 1 = Male	48.02%
Self-efficacy	Index measuring respondents’ beliefs that the future depends on them, that they can do anything, and that they control their own destiny	7.73 = Low 21.29 = High	16.26	2.61
Education	Highest level of education completed	1 = elementary or junior high 6 = bachelor’s degree (4 yrs)	3.49	1.50
Financial Problems	any financial difficulties in the past 5 years	0 = No 1 = Yes	43.79%
Religiosity	How important is religion in your life?	1 = not at all 5 = extremely	2.16	.92
Discrimination	% reporting discrimination in education or housing	0 = No 1 = Yes	16.53%
Work Characteristics
Work Hours	Number of weekly hours worked	0 = Low 80 = High	24.86	17.37
Work Friends	How close do you feel to your best friend at work?	0 = Not at all 4 =Extremely	1.68	1.41
Workplace Training	Received on-the-job training	0 = No 1 = Yes	79.81%
Supervisor Listens	How often is your supervisor willing to listen to your problems and help find solutions?	1 = Never 5 = Almost always	4.20	1.00
Supervise Others	Supervise others in primary workplace	0 = No 1 = Yes	23.74%
Restaurant Industry	% employed in the restaurant industry during high school or young adulthood	0 = No 1 = Yes	34.18%	.47
Sales Industry	% employed in the sales industry during high school or young adulthood	0 = No 1 = Yes	39.23%

We include six individual-level measures in our models predicting harassment and mobilization: gender, self-efficacy, education, financial problems, religiosity, and discrimination (see Table 1 for coding and basic descriptive statistics). Gender was self-reported by respondents. Self-efficacy is measured with a 3-item scale based on respondents’ beliefs that the future depends on them, that they can do anything, and that they control their own destiny. Higher scores indicate greater self-efficacy. Education is an ordered categorical variable representing respondents’ highest level of education completed. Financial problems is a dummy variable for whether individuals reported experiencing any financial difficulties in the past five years (spanning the same period of young adulthood as the harassment and mobilization measures), and religiosity measures the importance of religion in respondents’ lives. Lastly, the discrimination measure indexes experiences of discrimination in either schooling or housing. Because we expect women to experience more sexual harassment and discrimination in other realms, we will test for a gender interaction term for this measure.

We also employ seven measures of work characteristics, described in detail in Table 1: work hours, work friends, workplace training, the extent to which supervisors listen, supervisory status, and restaurant and sales industry occupations. Work hours is a continuous measure of weekly self-reported work hours, ranging from zero to eighty. Work friends measures how close individuals feel to their best friend at work. Because work friendships may function differently for men and women, the effect of work friends on mobilization may vary substantially by gender; therefore, we also include a gender interaction in our models measuring mobilization. Workplace training is a dummy variable indicating whether workers received on-the-job training at their current job. Supervisor listens assesses the extent to which respondents’ supervisors are willing to listen to their problems and help find solutions. Supervisory status indicates whether the respondent supervises others at work. Because our harassment and mobilization measures span five years during young adulthood, we are unable to connect these measures to a specific workplace or occupational industry. For this reason, our restaurant and sales industry measures are not mutually exclusive, reflecting only whether respondents held these positions at any point in young adulthood. Because of the gendered nature of restaurant and sales work, we include gender interaction terms for these measures.²

To understand mobilization from the research participants’ perspectives, we needed to hear their stories in their own words. We selected 33 survey respondents for intensive interviews, based on whether they had told others about their harassment or remained silent. Fourteen men and nineteen women participated in the interviews, responding to open-ended questions about rapport with co-workers, experiences with offensive behaviors in the workplace, and their responses to those behaviors. Some of the interview excerpts presented in this paper underscore patterns that emerged from the quantitative analysis, while others demonstrate the complexity of participants’ experiences in a way that cannot be easily quantified.

Harassment and Mobilization as Interrelated Processes

We first present simple univariate probit models to identify the predictors of perceiving harassment in young adulthood. We then estimate a separate model, in which we use individual and workplace characteristics, past harassment, and mobilization experiences to predict legal mobilization among those who are harassed. To examine harassment and mobilization simultaneously, we then use a generalization of James Heckman’s (1976) technique for modeling selective samples. Our sample is selective because we only observe mobilization among those subject to harassment, yet our model suggests that factors related to targeting will also be tied to the propensity to notify others about the harassment.

The most direct, and typical, way to assess mobilization of harassment complaints is to simply exclude respondents who did not experience harassment from the analysis. Because mobilization information is only available for those who experience some form of harassment, however, there is no assurance that the effects found among targets would hold among those who were not targeted. Moreover, if targets are characterized by some unmeasured trait associated with both harassment and mobilization, such as legal consciousness, any estimates based on this subsample could be seriously biased. We therefore estimate two-equation probit models that allow for interdependence in the processes leading to harassment and mobilization.

The first equation models the selection process predicting harassment, and hence entry into the pool of targets or potential mobilizers, while the second equation predicts the likelihood of mobilization. To the extent that the unmeasured factors driving mobilization also affect harassment, the error terms for the mobilization and harassment equations will be correlated. For example, if targets have low legal consciousness and legal consciousness is positively associated with mobilization, this would be reflected in a negative correlation (denoted by the Greek rho (ρ)) between the error terms of the harassment and mobilization equations. By introducing unmeasured characteristics associated with harassment into the mobilization equation, rho adjusts the estimates observed among the subsample of harassed persons for the unmeasured factors affecting the propensity to be harassed.

In this bivariate probit selectivity model, Y_1i, representing the harassment experience of person i, determines whether the effects of our predictors on Y_2i (mobilization), are observed (Winship and Mare 1992; Liao 1995). This model takes the form:

1. Harassment:	Y*1i = β1 X1i + ε1i
	Y1i = 1 if Y*1i > 0; 0 otherwise
2. Mobilization:	Y*2i = β2 X2i+ ε2i
	Y2i = 1 if Y*2i > 0; 0 otherwise

where, for the i^th observation, Y*₁ is the latent propensity to become a target of harassment (an unobserved continuous latent variable) and Y₁ is a dichotomous self-reported indicator of harassment; Y*₂ represents the propensity to mobilize the law and Y₂ represents observed self-reported mobilization; X₁ is a vector of values on the independent variables in the harassment equation and X₂ represents values on the independent variables that are observed when Y₁ = 1; β₁ and β₂ are vectors of coefficients, and ε₁ and ε₂ are error terms with correlation (ε₁,ε₂) = ρ (see Liao 1995: 91). This correlation represents the interrelatedness of the two processes, or the association between the probabilities of harassment and mobilization net of the effects of the independent variables in the model. We use the same basic set of individual and workplace predictors in each equation, with two primary identifying restrictions. We include a past measure of high school harassment experiences in the adult harassment (but not mobilization) equation and a past measure of high school mobilization experiences in the mobilization (but not harassment) equation. We also include a different set of gender interactions in each equation, such as the hypothesized gender interaction with work friendships in predicting mobilization, corresponding to extant research on gender, harassment, and legal consciousness.

We further explore the interrelation between processes of targeting and mobilization in the accounts shared with us by a subset of the survey respondents selected for interviews. Our interview guide was designed to elicit information from participants about their harassment and mobilization experiences. All questions were open-ended and presented to participants as related topics arose during each interview. Thus, as is characteristic of intensive interviewing, the same topics were covered with all participants but not presented in the same order (Lofland and Lofland 1995). Each 60–90 minute interview was recorded then transcribed and analyzed using NVivo, a qualitative data analysis software program. To maintain interview participants’ confidentiality, we use pseudonyms and, in some cases, have changed minor but potentially identifying details (e.g., company names and locations).

RESULTS

Results of the Selection or Targeting Model

The first step in the quantitative analysis is to estimate the harassment targeting equation. The probit estimates in Table 2 consider the effects of individual and workplace factors, along with specific behavioral measures of harassment. As shown in Model A, four of the six behavioral indicators are significant predictors of self-reporting that one was sexually harassed at work. Those who experienced offensive joking, invasion of personal space, unwanted touching, or physical assault were more likely to label their experience as sexual harassment, net of other indicators. Note that this model does not take into account the severity of each indicator, so we are unable to determine whether the women in our sample experience more severe or intentionally derogatory behaviors (though prior analysis of these data show that women are more likely to experience such “classic” indicators of sexual harassment as invasion of personal space and unwanted touching). Model A also includes a measure of the number of objective harassment indicators experienced during high school to net out person-specific differences in the propensity to experience targeting. Those experiencing a higher number of prior indicators are also more likely to self-report sexual harassment.

Table 2.

Univariate Probit Estimation of Perceived Harassment

Variable	Model A	Model B
Objective Harassment
Offensive Joking	.352*	.348*
	(.138)	(.138)
Personal Questions	−.061	−.063
	(.141)	(.141)
Invasion of Space	.443**	.453**
	(.147)	(.147)
Unwanted Touching	.625**	.620**
	(.197)	(.198)
Offensive Materials	.140	.140
	(.214)	(.215)
Physical Assault	.751*	.793*
	(.369)	(.380)
Total High School	.127*	.131*
Harassment	(.053)	(.053)
Individual-level
Male	−.577**	−.654**
	(.127)	(.195)
Self-efficacy	.021	.018
	(.026)	(.026)
Education	.081#	.074#
	(.043)	(.044)
Religiosity	.098	.091
	(.067)	(.067)
Financial Problems	−.028	−.037
	(.128)	(.129)
Other Discrimination	.425**	.488*
	(.156)	(.198)
Other Discrimination*Male		−.136
		(.316)
Work Characteristics
Work Hours	.004	.004
	(.005)	(.005)
Work Friends	−.082	−.085
	(.057)	(.057)
Workplace Training	.302#	.315#
	(.182)	(.183)
Supervisor Listens	.073	.075
	(.072)	(.073)
Supervise Others	−.132	−.133
	(.171)	(.171)
Restaurant Industry	−.126	−.058
	(.124)	(.151)
Restaurant*Male		−.196
		(.262)
Sales Industry	−.207#	−.346*
	(.120)	(.145)
Sales*Male		.426#
		(.251)
Intercept	−2.169**	−2.077**
	(.593)	(.580)
Number of cases	705	705
Log-Likelihood	−311.770	−309.869
Chi-Squared	153.02**	156.82**

^#p < .10

^*p < .05

^**p < .01

Note: Standard errors in parentheses

We also include individual and work characteristics as predictors of subjective harassment. First, gender is a strong predictor, as men were significantly less likely than women to understand a workplace experience as sexual harassment. Because our model includes behavioral indicators of harassment, this effect indicates that even when men and women are exposed to the same behaviors, such as offensive joking or unwanted touching, women are more likely to consider their experiences to be sexual harassment. Individuals who reported past discrimination in education and/or housing were also more likely to report workplace sexual harassment. Several other estimates are suggestive, though they do not exceed standard significance levels: respondents with more education and on-the-job training, and those lacking sales experience all appear more likely to report harassment (p < .10).

We anticipate that gender will interact with a number of workplace characteristics in structuring harassment. In Model B, we include gender interactions for three variables: discrimination in education or housing, and restaurant and sales occupations. Although the inclusion of these three product terms does not significantly improve the overall fit of the model, the sales interaction term is statistically significant. Under the interaction coding, the main effect of discrimination, restaurant, and sales represents their effects among women (effects for men are computed using both the main effects and the interaction terms). While we find no significant gender interactions for discrimination in education or housing or for restaurant workers, women in sales positions are significantly less likely to be harassed than women in non-sales positions. For men, however, sales work may increase the likelihood of harassment (at p < .10). In general, the results in Model B parallel those in Model A.

As previous researchers (Quinn 2005) suggest, focusing on targets to the exclusion of the harassers provides only a partial understanding of harassment. When asked in our interview why she thinks she was targeted, Pam said, “I just don’t understand it. Maybe they thought I wouldn’t care. But I don’t believe that either, you know? I think it has to do more with the person that is giving it more than the person receiving it.” Although our data do not allow us to analyze harassers’ motivations, we can ask whether a harasser’s choice of target is linked to the propensity to mobilize by modeling harassment and mobilization as interrelated processes.

Mobilization Models

Turning from harassment to mobilization, we first present univariate probit estimations in Table 3, and then include selectivity corrections in Table 4. As shown in Table 3, Model A, few individual characteristics and work activities are significant predictors of mobilization. While women were more likely to experience harassment in Table 2, Table 3 shows that men and women are equally likely to formally mobilize in response. The presence of close work friends, however, is a strong and consistent predictor of mobilization. The closer one feels to friends at work, the more likely s/he is to report sexual harassment to a supervisor or government agency.

Table 3.

Univariate Probit Estimation of Formal Mobilization

Variable	Model A	Model B	Model C
Objective Harassment
Offensive Joking		−.234	−.295
		(.323)	(.337)
Personal Questions		−.015	−.040
		(.317)	(.329)
Invasion of Space		.662*	.644*
		(.283)	(.292)
Unwanted Touching		.376	.355
		(.327)	(.334)
Offensive Materials		.489	.422
		(.385)	(.392)
Physical Assault		.811#	.796
		(.480)	(.512)
Individual-level
Male	−.431	−.339	−1.209
	(.275)	(.344)	(.781)
Self-efficacy	−.065	−.042	−.069
	(.052)	(.062)	(.065)
Education	.131	.150	.180#
	(.086)	(.105)	(.109)
Religiosity	−.162	−.207	−.240
	(.125)	(.148)	(.153)
Financial Problems	.481#	.375	.488
	(.252)	(.297)	(.305)
Other Discrimination	.278	.314	.237
	(.264)	(.303)	(.315)
Work Characteristics
Work Hours	.004	−.005	−.005
	(.010)	(.012)	(.012)
Work Friends	.240*	.284*	.315*
	(.111)	(.133)	(.144)
Work Friends*Male			.088
			(.291)
Workplace Training	.757#	.910#	.807
	(.438)	(.549)	(.555)
Supervisor Listens	.034	−.028	.043
	(.135)	(.174)	(.188)
Supervise Others	−.443	−.281	−.437
	(.339)	(.390)	(.413)
Restaurant Industry	−.192	−.574*	−.597#
	(.228)	(.288)	(.316)
Restaurant*Male			.170
			(.795)
Sales Industry	.121	.179	.069
	(.228)	(.277)	(.287)
Mobilization
High School Formal Mobilization		1.208**	.994**
		(.351)	(.361)
Young Adult Informal Mobilization		.576*	.325*
		(.292)	(.317)
Adult Informal Mobilization*Male			1.713*
			(.828)
Intercept	−1.000	−1.824	−1.421
	(1.156)	(1.422)	(1.491)
Number of cases	171	171	171
Log-Likelihood	−92.159	−71.261	−68.561
Chi-Squared	33.37*	75.17**	80.56**

^#p < .10

^*p < .05

^**p < .01

Note: Standard errors in parentheses

Table 4.

Probit Estimation of Mobilization with and without Selectivity Correction

	Probit	Bivariate Probit with Selectivity
Variable	A	B	C	D	E
Objective Harassment
Invasion of Space				.652	.724*
				(.415)	(.363)
Physical Assault				.672	.746
				(.579)	(.578)
Individual-level
Male			.196	−.313	−1.223
			(.250)	(.421)	(.784)
Self-efficacy			−.028	−.020	−.045
			(.038)	(.059)	(.062)
Education			.038	.131	.168
			(.070)	(.105)	(.104)
Religiosity			−.179#	−.235#	−.274#
			(.094)	(.143)	(.152)
Financial Problems			.133	.369	.507#
			(.226)	(.301)	(.306)
Other Discrimination			−.132	.246	.222
			(.211)	(.267)	(.373)
Work Characteristics
Work Hours			−.009	−.013	−.015
			(.008)	(.012)	(.012)
Work Friends			.271**	.261*	.293#
			(.086)	(.132)	(.152)
Work Friends*Male					.047
					(.284)
Workplace Training			.483	.835	.749
			(.351)	(.526)	(.529)
Supervisor Listens			−.049	.021	.107
			(.106)	(.168)	(.179)
Supervise Others			−.157	−.301	−.481
			(.258)	(.366)	(.390)
Restaurant Industry			−.220	−.527#	−.566#
			(.182)	(.271)	(.303)
Restaurant*Male					.158
					(.793)
Sales Industry			.198	.226	.074
			(.174)	(.290)	(.323)
Mobilization
High School Formal				.999**	.819*
Mobilization				(.362)	(.355)
Young Adult Informal				.572#	.337
Mobilization				(.301)	(.298)
Adult Informal					1.788*
Mobilization*Male					(.820)
Intercept	−.415**	.513**	.772	−1.626	−1.671
	(.099)	(.151)	(.877)	(1.873)	(1.810)
Rho (p)		−.841**	−.929**	−.176	.075
		(.089)	(.093)	(.633)	(.648)
Objective Harassment
Offensive Joking		.338**	.293*	.342**	.345**
		(.118)	(.116)	(.132)	(.132)
Invasion of Space		.509**	.525**	.450**	.448**
		(.127)	(.121)	(.144)	(.144)
Unwanted Touching		.517**	.483**	.614**	.608**
		(.183)	(.184)	(.197)	(.120)
Physical Assault		.957**	.937**	.806*	.791*
		(.350)	(.341)	(.384)	(.382)
Total High School		.111*	.117*	.126*	.128*
Harassment		(.047)	(.046)	(.052)	(.051)
Individual-level
Male		−.689**	−.675**	−.670**	−.669**
		(.179)	(.181)	(.195)	(.196)
Self-efficacy		.012	.018	.018	.017
		(.023)	(.025)	(.026)	(.026)
Education		.072#	.067	.070	.071
		(.040)	(.043)	(.044)	(.044)
Religiosity		.054	.096	.091	.091
		(.061)	(.066)	(.067)	(.067)
Financial Problems		−.056	−.090	−.052	−.047
		(.118)	(.128)	(.129)	(.129)
Other Discrimination		.467**	.537**	.506**	.498*
		(.176)	(.190)	(.197)	(.198)
Other		−.047	−.140	−.126	−.123
Discrimination*Male		(.284)	(.279)	(.314)	(.315)
Work Characteristics
Work Hours		.002	.005	.004	.004
		(.005)	(.005)	(.005)	(.005)
Work Friends		−.024	−.081	−.087	−.087
		(.054)	(.056)	(.057)	(.057)
Workplace Training		.365*	.335#	.324#	.318#
		(.173)	(.181)	(.183)	(.183)
Supervisor Listens		.080	.090	.077	.077
		(.066)	(.071)	(.072)	(.072)
Supervise Others		−.133	−.087	−.126	−.129
		(.154)	(.169)	(.170)	(.169)
Restaurant Industry		−.194	−.101	−.068	−.061
		(.136)	(.143)	(.152)	(.151)
Restaurant*Male		−.035	−.030	−.189	−.196
		(.241)	(.235)	(.263)	(.261)
Sales Industry		−.290*	−.332*	−.359*	−.358*
		(.134)	(.143)	(.145)	(.145)
Sales*Male		.380#	.343	.434#	.437#
		(.233)	(.242)	(.250)	(.252)
Intercept		−1.973**	−2.117**	−2.043**	−2.034**
		(.554)	(.594)	(.598)	(.599)
Number of cases	171	171, 536	171, 536	171, 536	171, 536
Log-Likelihood	−109.524	−408.195	−394.080	−385.503	−382.548

^#p < .10

^*p < .05

^**p < .01

Note: Standard errors in parentheses

In Model B, we also control for six objective sexual harassment indicators experienced in young adulthood, individuals’ past mobilization in response to sexual harassment, and their informal mobilization in young adulthood. Of the six measures of harassment, those who experience invasion of space are significantly more likely to report the harassment to a supervisor or government authority. Moreover, informal mobilization (i.e., reporting the harassment to a coworker) and past formal mobilization are also strong and significant predictors of formal mobilization. The inclusion of these independent variables in Model B significantly improves the overall fit of the model, yet the effect of close work friends remains significant.

Model C of Table 3 includes gender interactions for closeness to work friends, restaurant work, and informal mobilization. As in the harassment equation in Table 2, the inclusion of these interactions did not significantly improve the overall model fit, yet one important interaction is worth noting. As shown in Model C, informal mobilization is a predictor of formal mobilization for both men and women, yet men who report the harassment to a coworker are significantly more likely to also report the harassment to a supervisor or government authority. This suggests that women may talk about the harassment with others, but not file a formal complaint, whereas men are more likely to keep the harassment to themselves unless they are prepared to mobilize more formally. This finding does not reflect women’s greater subjective closeness to work friends, as the gender interaction term for work friends is not significant.

Table 4 reports estimates for models predicting mobilization among respondents who experienced harassment, and two-equation models that adjust these estimates using a selectivity correction. Probit coefficients can be interpreted as the effects of a unit change in the independent variable on the cumulative normal probability of the dependent variable, or as effects on z-scores. Model A fits only a constant, indicating that the probability of mobilization relative to non-mobilization is .34, which is the probability associated with a z-score of -.415 (see, e.g., Hardy 1989). Therefore, only 34 percent of individuals who reported they were sexually harassed elected to notify a supervisor or government agency of their mistreatment.

The bivariate probit selection models, B, C, D, and E, test whether unmeasured factors that increase the probability of harassment also affect the likelihood of mobilization. These “substantive” models were jointly estimated with a “selection” equation predicting targeting (based on estimates from Model B of Table 2). If the unmeasured propensity to mobilize is lower among harassment targets than non-targets, for example, this would be manifest in a negative correlation between the error terms of the harassment and mobilization equations. Alternatively, if targets had higher unmeasured propensity to mobilize, then the correlation between the error terms would be positive. Model B, which simply fits a constant in the substantive equation, gauges the effects of unmeasured traits that increase the likelihood of harassment on the probability of mobilization. Here, the correlation is negative in direction and statistically significant (rho = -.841). This means that unmeasured factors associated with sexual harassment decreased the probability of mobilization. Note that the intercept term in Model B is no longer negative as in the uncorrected model (Model A). The probability of mobilization net of unmeasured factors related to harassment is approximately .70, the probability associated with a z-score of .513. Because the correlation is negative, the overall probability of mobilization would thus have been higher had harassment targets represented a random draw from the sample, rather than a selected subgroup. This suggests that perpetrators may select potential targets, in part, because they are those least likely to mobilize the law or notify others of the harassment.

Model C includes individual-level and work characteristics in the mobilization model, along with the selection equation. Closeness to work friends is the only significant predictor of formal mobilization, which parallels the univariate probit estimation of formal mobilization in Table 3 (see Model A). Selection again plays a significant role in this model, however, as the correlation remains strong and significant (rho = -.929).

We next include the two behavioral indicators of harassment that were significant in the univariate probit estimation (see Table 3), informal mobilization, and past formal mobilization. As shown in Model D, measures of invasion of space and physical assault are no longer associated with mobilization, net of selection. In contrast, past formal mobilization and informal mobilization in young adulthood are statistically significant. Those who reported sexual harassment to a supervisor or government authority in the past, and those who reported harassment to a coworker in young adulthood, were more likely to formally mobilize in response to harassment. The effect of work friends on mobilization remained significant despite the inclusion of these additional measures, which significantly improved the overall model fit. The addition of these controls, however, dramatically reduces rho’s magnitude and significance. This reduction suggests that the effects of unmeasured factors related to both harassment and mobilization are captured in large part by these strongly predictive independent variables.

Lastly, we include gender interaction terms for work friends, restaurant workers, and informal mobilization in Model E. While the inclusion of these interactions does not significantly improve the overall model fit, we report three noteworthy findings. First, one of the behavioral measures of harassment, invasion of space, re-emerges as an important predictor of mobilization. Net of other independent variables, individuals who experience invasion of space are more likely to formally mobilize than those who did not experience this measure of harassment. Second, the effect of close work friends remains strong in magnitude, though the standard error increases slightly in size so that the statistical significance drops slightly below the 95 percent confidence interval (p = .054). Lastly, the inclusion of a gender interaction for informal mobilization revealed a significant difference between men and women. Men, but not women, who report harassment to a coworker are more likely to also report the harassment to a supervisor or government authority. As in Model D, rho remains small and is not statistically significant.

We found one other important difference between the selection-corrected models and the univariate probit presented in Table 3. In Models C, D, and E, religiosity emerges as a marginal negative predictor of mobilization (p < .10). In all three models, the likelihood of reporting harassment to a supervisor or government authority decreases as religiosity increases. Relying solely on uncorrected models could cause researchers to overlook the potential role of religiosity in suppressing formal mobilization.

In some ways, our interview results affirm these findings. In particular, having close relationships with co-workers was important for several interview participants. Laurie, a white woman who worked as a waitress in high school, said she and her coworkers warned one another about customers known to be harassers. Describing the content of such discussions, Laurie said:

[We’d say] like, “Oh, careful, he's a jerk,” or, “He's slimy,” or “He's a creep,” whatever. … So then you track records… It would go back and forth, and people would say, “Watch yourself.” So you just want to stand further away, or whatever, to safeguard [against being harassed].

Laurie’s experience shows how coworkers may collaborate to protect one another from harassment and it might help explain the low rates of formal mobilization we observed for women in restaurant work. Dan, on the other hand, did not have the sort of supportive workplace relationships Laurie described. A white man who worked for the postal service as a young adult, Dan said he felt discomfort after being “hit on” by several older female coworkers:

A few older women said stuff like, I don’t know if it would be considered harassment but asking, “Hey would you go out” type thing. Stuff like that. I don’t know if that would be considered harassment. It did cause me to have kind of an uncomfortable feeling working with them in the future.

When asked whether he talked about those experiences with anyone, Dan said:

No, no. You know, it wasn’t anything that I wanted to confide with anyone else. I just kind of kept it underground.

Our quantitative results suggest that without confiding in workplace friends, men such as Dan may have been even less likely to report his experience to a supervisor.

Finally, consistent with our quantitative results, Lisa’s story shows how relationships with co-workers might shape participants’ legal consciousness. Upon hearing that her co-workers believed she did not object to a vendor’s offensive remarks, her own consciousness changed:

At the convenience store where I worked in high school we had a vendor who was pretty sexist and would make comments on the customers walking in. About their boobs. It didn’t occur to me that he shouldn’t be saying that stuff. I just naturally expected him, because he was such a sleazy guy, to be saying stuff like that. Then some of the other employees filed a complaint about him and he got fired. Later I asked them why they hadn’t asked me to join the complaint and there was a comment like, “We thought you didn’t have a problem with it.” And, you know, I can see why they would have thought that just ‘cuz I wasn’t complaining about him to anybody else. [The experience] changed my feelings since I can speak up if something like this happens, “Oh wait a minute, there’s something I can do about this!” I can say, “Hey! This is not right or appropriate and knock it off!”

Our quantitative analysis suggests that workers such as Lisa may be targeted for harassment, in part, because they are perceived as unlikely or unwilling to mobilize. Lisa’s story suggests a more dynamic process, in which legal consciousness develops in response to social relationships.

Our interview data may also help explain why workplace training, at least sexual harassment training, was non-significant in the quantitative models that adjust for selectivity. Several interview participants noted that an increasing number of companies are beginning to require their employees to attend some form of sexual harassment training. Yet few participants put much stock in the training. Angela recalls how she and her coworkers rarely connected their everyday retail work to the sexual harassment training they had received:

There were times when [men] coworkers would comment on a woman coming into the department, make really explicit comments about her breasts or body parts. That always made me feel uncomfortable. And they had posters in the warehouse, actual posters of women in bikinis and stuff. … There were a few women that I spoke with [about it]. I think we all just kind of dismissed it as ‘well, that’s just guy behavior,’ not really ever acknowledging that maybe we can have a different expectation, or that if we’re uncomfortable with it that we could really do anything about it. It was funny – we had the little sexual harassment lectures and we’d go through the sexual harassment training and we’d just all laugh and then we’d walk away.

While our interview data complement the quantitative findings in some ways, they also go beyond our quantitative models, revealing additional forms of mobilization and factors that may play a role in individuals’ mobilization decisions.

THE COMPLEXITY OF MOBILIZATION: EVIDENCE FROM OUR INTERVIEWS

Our interviews also raise new questions about the complexities of legal consciousness and mobilization. We first discuss factors not considered in the statistical models that affect participants’ consciousness regarding harassment, such as the nature or quality of personal relationships. We then offer a preliminary multidimensional typology of mobilization to describe the range of possible responses to harassment.

Consciousness and Sexual Harassment

Consciousness about sexual harassment involves how people understand harassment, how they define it, and how their own views develop over time. In describing the processes by which they come to understand and respond to sexual harassment, we observe “the imprint of law” (Ewick and Silbey 1998) reflected in and reflective of those processes. In many ways, our interview data echo the quantitative results, suggesting that participants’ relationships with coworkers shape their conceptions of and responses to harassment. Beyond our quantitative results, interview participants indicated that their understandings of harassment changed as they gained life experience. Additionally, they described how factors such as their family’s attitude toward harassment, which was not included in our statistical models, may impact consciousness.

Our interview data show that relationships outside of the workplace may be just as important in shaping consciousness as those within the workplace. For example, Pam’s survey responses indicate that she mobilized by telling her coworkers and bosses about offensive questions, jokes, and invasion of her personal space. In our interview, however, Pam revealed an additional form of mobilization that she perceived as even more important. Discussions with friends and family, according to Pam, transformed how she perceived the “constant sex jokes” at one of her first workplaces. In thinking about the joking, Pam said “This was sort of before sexual harassment really came into the limelight so a lot of those things that happened were just considered acceptable. Looking back on it now, there’s just no way it would have flown today. [If the diner were still in business] it would have had to have a major overhaul.” Pam still remembers some of the jokes, recalling the following in her interview:

One time one of the waitresses was kneeling on the floor. She was trying to get something that had fallen under the jukebox and the manager said, “Hey Sally, while you’re down there…” [speaking in a suggestive tone]. And he just started laughing.

Pam told some of the jokes to her friends and family because she thought they were funny. But the reactions of people she respected caused her to reconsider her perception of the joking:

I told my girlfriends about some of the stuff that was said and they didn’t like it. And I remember telling a family member and she especially didn’t like it. I don’t remember exactly what was said, but I remember that they didn’t see it in the same light that I did.

Pam’s consciousness about her own experiences thus shifted after discussing them with others outside of work. This important aspect of how Pam’s consciousness developed could not have been captured using our survey methodology.

Pam’s job at the diner was one of her first jobs and she was relatively young at the time. But more experienced workers also relied on personal relationships to understand and cope with harassment. Like Pam, Marie told supportive people in her life about harassing workplace experiences. Marie, one of few women managers in the construction industry, said she discusses her workplace experiences with her husband, who works in the same industry:

If something is borderline, I talk to him about it. I don’t think it’s good to keep things hidden. Just being a female in management is difficult, and guys don’t like it.

Marie went on to describe how discussions with her husband helped her face men in the industry who were resistant, sometimes even hostile, to a woman manager. Thus it was in the hours after work and in the company of her spouse rather than colleagues that Marie’s consciousness about harassment developed. In addition to relationships outside of work, our interview data suggest one more way that relationships with others may impact consciousness.

While our survey results indicate that having close work friends increases mobilization, the interview data show that co-workers may sometimes inhibit mobilization decisions. In fact, co-workers led some interviewees to believe they may have overreacted to a situation. Though she reported no mobilization in the survey, Angela later described her response to a customer’s proposition while at her high school waitressing job.

I was sitting at a table wrapping the silverware in cloth napkins and he came and sat down and was talking to me for a long time. I really didn’t want to be rude and ask him to leave but I was becoming more and more uncomfortable and he was becoming more and more intimate and forward with his questions. Finally he said “I’m staying at this hotel across the highway” and he invited me to join him. At that point I asked him to leave.

After “loitering for a bit” the man did leave the restaurant. Angela described the experience as “a turning point in my view of that job” and said “that was really the first time where I experienced that blatant sexual propositioning. I was pretty innocent and a naive Catholic girl, so I really felt unsafe for a while after that.” Angela then told an older waitress about the experience:

I told another waitress. She was a lot older. Basically I just shared it with her to find out if she knew him -- was he someone in the community or was he just traveling through. She had seen him before but he wasn’t in the community. She basically just said that it was a pattern he had, that when he came in he would talk to other waitresses that way. She was kind of nonchalant. She had worked there for probably fifteen years. I think it was just part of daily life for her in a sense.

As Angela’s remarks indicate, co-workers might sometimes discourage mobilization. It is also worth noting that Angela’s background as a “naive Catholic girl” shaped her perception of the experience. Angela’s own observation about the possible impact of her religious background on her perception of appropriate and inappropriate behavior echoes the finding from our survey data that religiosity seems to play a role in responses to sexual harassment.

Relationships not only impacted how participants’ perceived their experiences but also how they responded to them. In Pam’s interview, she described not wanting to disrupt the family-like atmosphere at her workplace. She did not like the sexual joking, remarks about her body, and intrusive questions by her male coworkers at a radio station, but she also felt some responsibility for wanting to maintain the existing atmosphere. In Pam’s words, “The atmosphere there was so much like a family that it was really hard to [speak up]. Most people just got along there. That may be one of the reasons why I never, you know, put up a fuss about it.” For Pam having close friends at work may have actually discouraged her from mobilizing. Without the interview data, we would not have observed this aspect of participants’ experiences.

Across our interviews, routine activities at work, social relationships, and individual characteristics seem to work together in shaping consciousness about and responses to sexual harassment. These data also show how responses to harassment are intertwined with individuals’ unique understandings of the concept. In the following section, we describe participants’ responses in greater depth, analyzing the variety of mobilization strategies employed.

Responding to Harassment: A Typology of Mobilization?

Participants employed many strategies to deal with workplace sexual harassment, from ignoring it to reporting it up the chain of command. Our interview results suggest that a multidimensional typology best represents participants’ diverse mobilization strategies . Moving beyond traditional formal/informal mobilization dichotomies allows us to more closely examine the complexities of mobilization. In general, we find multiple forms of “informal” mobilization. Many of those who were classified as “non-mobilizers” from the survey data spent time during their interviews describing response strategies that were empowering and effective for themselves -- strategies that should not be overlooked. Our preliminary typology of mobilization is represented by the figure below, followed by interview data bearing on the proposed typology.

Mobilization Continuum: Ignore→Avoid→Self help→Tell friends/family→Tell equals→Tell superiors→Tell attorney/agency

Some participants ignored harassment as a strategy for coping with it. Both women and men described ignoring harassment, though women’s descriptions more clearly identified this as a conscious strategy. Like Dan, the postal worker described previously, Cam, an Asian American man who worked in a factory during high school, tried to keep “uncomfortable” experiences with female coworkers under wraps. In one of the few instances of female-on-male harassment in our sample, Cam says he felt uncomfortable when female coworkers came on to him:

There were girls who approached me that I didn’t know. They tended to grab on you, like on the elbows or shoulders. They tended to stand really close to you, and they smiled a lot, and they offered help, and they also asked for help. Those [kinds of] things. I didn’t know how to react to the situation. So it was hard for me and it was uncomfortable. I usually pretended there was nothing happening.

Like Dan and Cam, several women participants also talked about attempting to ignore harassment, though they did so less out of embarrassment that it had occurred and more as a strategy to avoid further harassment. Women seemed to hold out hope that if they avoided directly confronting harassment, it would eventually stop. Hannah did not formally mobilize because she observed her harassers’ responses to another woman whom they teased regularly:

Basically I watched this one woman get really upset all the time, which would always really goad the guys on more to tease her. So I would just let them tease me, and I wouldn't react, or try not to. I was hoping they would have gotten tired of it.

Other women did not ignore harassment per se, but employed “passive” strategies to deal with it. At the retail store where she worked during high school, Marie faced coworkers who made unwanted passes at her: “Most of the time my response was [to tell them] that I had a boyfriend, even if I didn’t. Sometimes I [really] did [have a boyfriend] and sometimes I didn’t.” Pam used a similar strategy to avoid harassment from customers while working at a bakery:

I remember actually buying a fake wedding ring and putting it on my ring finger so men wouldn’t bother me. [I started doing it] after men started bothering me. Because of it. I just didn’t want to mess with it. I just wanted to protect myself.

Angela avoided areas at work where she would encounter certain groups of male coworkers:

I remember a manager who made a lot of disparaging comments about sexual frustration like “I’m not serviced by my wife at home.” And he would make those comments around me and that really made me uncomfortable. If I saw this manager standing around talking to guys I definitely wouldn’t walk past and I wouldn’t be around it. I would avoid [certain] places around the store, and sometimes alter my work habits to avoid it.

Finally, April describes making strategic wardrobe choices to avoid harassment by male coworkers in an automotive repair shop:

I would have to watch what I wore. I wouldn’t wear anything too tight. If I did, well, I have, this is so stupid, but my nipples stick out, they always stick out. So I would have to put band-aids over them because I felt like they’d be staring at me if I didn’t. A couple times [they said things to me about it]. They would just say, “You cold?” or, “Your headlights are sticking out!” or something like that. It’s embarrassing. It embarrasses me.

Marie, Angela, and April did not ignore the harassment, nor did they take any steps toward formal mobilization. Instead, they each came up with creative strategies to deal with or avoid it. When ignoring or avoiding harassment proved ineffective, some participants actually quit their jobs or transferred to new locations. Marie tried to avoid harassment but eventually moved to a different store when her strategy of pretending to have a boyfriend did not work. Like Marie, Liz removed herself from a harassing situation by literally leaving her workplace. Liz said, “I took personal leave for the last two months of my pregnancy, and I truly took personal leave because I was so emotionally distressed.” Marie and Liz’s actions might be viewed as avoidance but it the act of leaving a job is a form of action. Traditional dichotomies of mobilization may omit such actions, overlooking a response that some participants perceived as both satisfying and effective.

When not avoiding harassment, participants sometimes used self-help by taking matters into their own hands, especially in cases where the harassment was isolated or rare. While working as a waitress, Laurie employed self-help in the following way:

One [experience] that sticks out the best is when a male customer put up two fingers and motioned for me to come over, and he asked me if I always came with two fingers. And then I did pour coffee in his lap, and then went down and said to my manager, “I have to tell you, I just did something that you might be angry, but this is what I did.” My manager said the customer had it coming.

In addition to the above experience, one might also understand Laurie’s previously described experience (in which co-workers warned each other about potential harassers) as a form of self-help. In that instance, rather than reporting harassment up the chain of command or demanding an improved workplace culture, these restaurant workers relied upon each other for protection and upon a strategy of standing “further away” from known harassers.

As Laurie did with her coffee, Jerry engaged in self-help when he directly confronted his harasser, a male coworker who groped him while on the job as a prison guard. Jerry immediately told the man to stop but did not report any mobilization in our survey. He explained, “I never did anything about it. You know, I felt like I dealt with it right there. You know [I said] ‘Don’t ever touch me again,’ basically.” Jerry’s harasser was eventually fired after word got around about his experience. Even though Jerry did not formally mobilize, several coworkers witnessed the incident and he also discussed it with a few coworker friends. Additionally, several others with whom Jerry worked later reported their own harassment experiences with the same man.

Having coworkers know about his experience may have influenced Jerry’s mobilization decisions. Interview participants also described the importance of relationships outside of work in shaping their consciousness and responses to harassment (see, e.g., Marie’s discussions with her husband, described in the previous section). Relationships with coworkers and friends, and discussions with them about their experiences, mattered even for adolescents in informal jobs. Janice’s friends helped her deal with the father in a family for whom she and her friends babysat.

We called him ‘horn-man’ because he was just icky. He was really icky. He would come out and just have shorts and a t-shirt on but he wouldn’t have anything under his shorts and he’d sit across from us and have his legs up and be talking to us. We’d be like, “Okay we have to go,” and he’d say “No, you can’t leave yet.” And if his wife wasn’t home he’d say “No, you can’t leave until you give me a hug.”

When asked whether she or her friends ever hugged the man, Janice said:

Yeah. We wanted to leave. We were afraid of him. I never went there alone because I was seriously afraid of him. [I always brought someone with me.] Always. Because he scared me too much. He was just so icky. I just can’t [pauses] I mean, he was just icky. Trying to be close to us, [saying things] like “Let me just hug you. Let me just squish you. You’re a really good girl.” Icky. And no. We didn’t [tell any adults at the time].

Although Janice and her friends did not report the “icky” man to any adults, and she reported no mobilization in her survey responses, their reliance on one another to cope with the situation could in fact be viewed as a form of informal mobilization.

For most of the interviewees, mobilization in response to harassment was not one-dimensional. That is, participants typically tried a response represented early on the continuum, perhaps ignoring harassment or more actively avoiding it, before talking with others about the experience or reporting it up the chain of command. Erin’s story, below, describes how her own mobilization progressed from avoidance to self-help to discussions with friends outside of and at work and finally to reporting the harassment to her boss. Erin, a school custodian, first tried to deflect the sexually suggestive remarks made by a male coworker. For example:

He said something about how he’d been looking in the Victoria’s Secret catalogue and was wondering what I’d look like in a pair of thong underwear and I said, “Why don’t you buy your girlfriend a pair of thong underwear. I don’t wear thong underwear.” You know, just trying to avoid the subject. I always tried to keep it at a joking manner.

Eventually the coworker’s remarks went too far, and he began grabbing and touching her without her consent, so Erin confronted him more directly:

I was like, “Don’t do that. You can’t be doing that at work,” and, “There’s kids around,” or, “Just don’t touch me. I don’t want you to touch me.” That kind of thing. He’d [stop] for a day or two and then go back to trying to jokingly touch me and all that crap again.

Despite her repeated requests for him to stop, and then enlisting the help of her coworkers who tried to get him to stop, Erin’s harasser continued to touch her and make sexually suggestive remarks, so she eventually mobilized more formally by reporting the harassment to her boss:

I told my boss about it and my boss told him the same day, within minutes, not to ever touch me again at work. That was uncomfortable but he came up and apologized to me the first time and said, “Oh I’m sorry. I won’t do it again.” But he did. So I told [my boss] again. He said he was sorry and all that and then a couple weeks later he touched me again and I was like, all right. So I told [my boss] again. There was two occasions where he actually unsnapped my bra while we were at work. I was like, all right, that’s it, no more! He would laugh and say, “Ha ha, lookit, I can unsnap your bra with one finger.” I said, “I don’t care how many fingers it takes, don’t do it. Don’t ever do it again.”

Although Erin did report the harassment to her supervisor, she initially pursued several other strategies to deal with it. Simply labeling Erin as a formal mobilizer because she reported the harassment to her supervisor overlooks these previous attempts and her multidimensional mobilization strategy.

Hannah’s experience in an office setting provides another example of how individuals might progress along the mobilization continuum. Like Erin, Hannah first attempted to respond to harassment by taking care of it on her own. Later, she reported the incident to her human resources department:

This sales guy [walked up behind me]. I was standing in the doorway, and I think I probably had my arms up on my [gestures, standing with hands on hips], and one of the slimiest sales guys came up, and went like this [slides hands up and down ribs], and tickled me, and I was like “Ew!” I was just so wigged out. I turned around, and I think I said pretty loud, that that was not an okay touch, which made everyone come out of their cubes. And then I did talk to [Human Resources] about it.

As described with Hannah and Erin above, other participants also recalled first attempting self-help techniques in order to stop harassment but then moving on to more formal mobilization.

Holly, the only woman manager at her company, formally mobilized after being harassed by a client at a company event but only after trying to deal with the situation on her own.

It was the VP of their company, and he would get really touchy-feely. And I’m like, uh, you know he’s kind of from that old boys’ school and him and [the owner of my company] have known each other for a lot of years and so I’m like whatever. He’d put his arm around me and say, “Oh I love her. She’s beautiful.” And I’m just sitting next to him and I’m the only girl at the table and I’m just like, oh ok this is fun [sarcastically].

The harassment continued, even escalated, in spite of Holly’s attempts to get him to stop.

And he’d put his arm around me and pull me towards him and, kind of uncomfortable, and I’d push away. And he just kept going on and touching me and put his hand on my leg very forcefully and then he was playing the game of trying to unhook my bra with two fingers which he did after I tried to get up and get away once. I went to get up and he put his arm on my shoulders and said, “Stay.” And the whole nine yards, and I was really shaking and he kept being excessively persistent with his hand on the leg thing and I sat with my leg crossed away from him, my arms across my body kind of protecting me.

Holly had attended the event alone, was the only woman present at the table, and did not have any close work friends. Holly thus lacked the coworker support that other respondents described. Eventually, others at the table began to notice what was happening and the client’s harassment finally ended, but only after he was physically separated from her.

It was just really awkward and people started to notice the different things that were going on and finally the regional guy just kind of [made a gesture indicating someone calling someone over]. So I got up and then I said, “You need to move, physically move your butt into my chair because I’m not sitting there.” So, we sat down at about six and he was touchy-feely for maybe an hour or so and then he started the leg and everything else and I’d push his hand off and it just progressed. I think I left about nine.

Holly formally mobilized by contacting her attorney the next morning. At the time of our interview, she and her attorney were negotiating a satisfactory response to the incident with the two companies involved. Holly found the experience of reporting the incident and contacting an attorney empowering. She described feeling emotional turmoil and difficulty at work following the incident, but she was also proud of her swift response. When asked to describe her feelings about how things had proceeded to that point, and whether she had any regrets about having reported it, she said simply, in a proud and matter-of-fact tone, “Nope, none.”

However targets respond, experiencing harassment alters their perceptions of the workplace and of themselves. As our interviews show, the process of mobilizing in response to harassment is dynamic, with every incident and response shaping how subsequent incidents are managed and how previous incidents are perceived.

WHAT IS MOBILIZATION? DISCUSSION AND CONCLUSION

Together, our quantitative and qualitative results show how the personal and social contexts that structure sexual harassment also shape mobilization choices. Although our sample is limited to younger workers in a single community, we can offer generalizations based on our longitudinal survey and intensive interview data. Our quantitative results suggest that having close work friends may provide guardianship that promotes mobilization. While much of the academic and policy literature on sexual harassment concerns hierarchical relationships, our findings suggest that horizontal worker-to-worker relationships are crucial for understanding both harassment and mobilization. Though close work friends may facilitate mobilization, others at work, such as mentors or those with whom participants are not personally close, may have a different influence. Angela’s experience, in which an older colleague caused her to doubt her interpretation of a customer’s proposition as harassing, shows how coworkers may sometimes squelch mobilization. The process by which Angela’s experience transformed into a grievance was not linear. Instead, her understanding of her experience was shaped by her initial gut response, by the fact that she felt “unsafe for a while” following the experience, and by the discussion with her “unimpressed” coworker. The dynamic nature of Angela’s account reflects Ewick and Silbey’s (1998) conceptualization of legal consciousness as a process and as a sort of cultural practice. For Angela as well as our other interview participants, mobilization was not simply a question of doing it or not, nor of formal versus informal mobilization. Instead, consciousness and mobilization unfolded dynamically, developing in response to conversations with others and their own varied attempts to cope with and confront harassment, from active avoidance to talking it over with coworkers.

The effects of self-efficacy in our quantitative models were non-significant, though the interview data suggest that efficacy may be worth investigating further. Jerry, the prison guard, and Laurie, the waitress, whose survey responses indicated higher than average scores on self-efficacy, both took matters into their own hands rather than pursuing formal remedies. To a greater extent than the survey analysis, the interviews highlight the dynamic nature of the relationships between targeting, mobilization, and perceptions. Participants’ legal consciousness is shaped by their harassment experiences and by other social interactions, and these in turn shape their feelings about subsequent experiences. We believe this paper underscores the need for multi-method work that allows for both breadth and depth in terms of research methodology.

Consistent with previous work (Hoffman 2003, 2006; Jacobs 2007) we find that legal consciousness and mobilization take several forms. Filing lawsuits and consulting attorneys represent formal responses, but even the seemingly simple act of telling others about one’s experiences represents a form of mobilization. Our interview data reveal great diversity in mobilization strategies, beyond the classic formal/informal dichotomy. Some individuals opt to confront their harassers directly, others engage in strategies to avoid or deflect harassment, and still others seek out alternative employment. While each approach has benefits and drawbacks, individuals facing workplace harassment clearly employ a great range of strategies in response. Future researchers might examine the extent to which our proposed mobilization typology applies not only to sexual harassment responses but to other mobilization processes as well. Does the sexual harassment mobilization continuum apply to workers who have greater experience than the adolescent and young adult workers in our study? Do individual mobilization responses to other forms of discrimination resemble the responses outlined in our preliminary typology?

Taken as a whole, our findings suggest that sexual harassment targeting and mobilization are best understood as interrelated processes. Targets appear to be selected, in part, because they are less likely to mobilize the law in response. While results from our selectivity-corrected models were generally consistent with those from uncorrected models, we were fortunate to have strong longitudinal data on past harassment and mobilization experiences in this study. Before adding these variables to our mobilization models, the harassment and mobilization equations were quite strongly correlated. Our interview data, including Lisa’s story about not being asked by coworkers to participate in a joint complaint about a vendor, further support this finding. Moreover, those who are targeted respond in part based on their prior harassment experiences and their suppositions about future experiences, such that reflections on past events shape subsequent mobilization strategies. Just as Lisa or others may respond differently should they face offensive workplace behavior in the future, Janice, who told us about the “icky” man who harassed her while working as a young babysitter, carries these harassment experiences with her into adulthood. Harassment targeting and mobilization are thus intertwined at any given time and individual responses and strategies evolve over time.