Introduction

The dual purpose of this study was first, to examine gender differences in the labeling and subjective appraisal of teens experiencing sexual harassment; and second, to investigate the differential consequences for female and male teens distressed by harassing behavior at work. Evidence from the United States (U.S.) establishes the widespread prevalence of sexual harassment targeting more frequently, though not exclusively, women (Wasti and Cortina 2002). Cross-cultural research suggests that organizations worldwide face the costs and productivity declines that result from high incidence of such harassment (Cortina and Wasti 2005). A theory, known as the “wimpy male” hypothesis (and its counterpart, referred to as the “weak female” hypothesis), was tested with the use of survey methodology in the current study to predict detriments in the psychosocial maturation experienced by adolescent men appraising sexual harassment as threatening and relying upon behavioral coping, and by adolescent women making a similar appraisal of harassment and relying upon cognitive coping strategies (Stockdale 1998; Stockdale et al. 1999).

Adolescent Workforce Participation

In 1940, the U.S. government estimated that only 3% of 16-year-olds had formal employment while attending school (Steinberg and Cauffman 1995). More recent statistics indicate a dramatically different trend among today’s youth in that employment often begins in middle school and increases with age (Entwisle et al. 2000). Among today’s high school students, over 75% of 12th graders work in paid jobs while school is in session (Staff et al. 2009; Staff and Schulenberg 2010). The workplace has come to be an influential context for shaping the development of teenage women and men (Frone 1999). One unfortunate result of this trend is adolescents’ risk of exposure to the stressors too often found in work settings, including sexual harassment. Indeed research indicates that work experiences influence personality development during this formative time of life (Roberts et al. 2003). Moreover, the clinical psychology literature depicts employment as a central adult role, and suggests that sexual harassment carries the potential for triggering questions about identity and life purpose (Frone 1999; Koss 1990). Adolescent victims of harassment would be vulnerable to sexual harassment as a threat to their views of both personal competency and work as a safe environment (Taylor 1983). Unless explicitly stated to the contrary, all articles reviewed herein draw data from U.S. samples.

The adolescent development literature has established a considerable knowledge base about workforce participation among U.S. teenagers. Steinberg et al. (1982) found that aspects of the workplace, such as excessive hours and repetitive tasks, interfere with adolescent development. Yet, despite this body of scholarship on work stressors, that literature has virtually overlooked the ripe conditions for the sexualized exploitation of adolescents in their early work experience. One exception to this paucity of attention was a study by Fineran (2002), which offered some descriptive information about patterns among adolescent encounters with sexual harassment. Surveying U.S. high-school-age boys and girls who worked part-time, she reported that 63% of the girls had experienced some form of sexual harassment at work, and 37% of the boys reported the same.

Sexual Harassment

Because few studies have reported on sexual harassment directed at teens in the workplace, we turned to another well-developed body of literature to inform our model; that is, the literature on sexual harassment among adults. Till (1980) mapped the conceptual domain of sexual harassment at a time when scholars had not settled on one accepted definition of sexual harassment and measurement of the behavior was fairly haphazard. Drawing from descriptions of a sample of U.S. college women, Till conceptualized the psychological experience of sexual harassment by establishing a set of five distinct categories: Gender harassment, seductive behavior, sexual bribery, sexual coercion, and sexual assault. These categories were arranged in a hierarchical fashion, ranging from least to most egregious. Till’s system stands apart among the early work by its influence on scholars’ thinking about sexual harassment and, in particular, the work of Fitzgerald and colleagues in operationalizing the sexual harassment construct with the Sexual Experiences Questionnaire, or SEQ. One of the more recent SEQ versions (Fitzgerald et al. 1999), evolving from past empirical refinements, defines sexual harassment as a multidimensional construct and is comprised of sample behaviors representing each of the following components: sexist hostility (discriminatory behaviors of a non-sexual behavior); sexual hostility (offensive behaviors that are of a more explicit sexual nature); unwanted sexual attention (sexually-oriented verbal and nonverbal behaviors that are offensive and unwelcome); and sexual coercion (pressure to gain sexual cooperation in exchange for employment benefits or to avoid punishments).

Labeling of Sexual Harassment

A single item located at the end of the SEQ instrument, designed to address a cognitive judgment about the meaning of one’s exposure to sexually-oriented work behavior, directly asks the respondent whether he or she has been “sexually harassed” (Stark et al. 2002). The literature uses the term labeling to refer to this process of naming an event as “sexual harassment” (Munson et al. 2001). Use of this separate item reflects scholars’ acknowledgement that an individual may be the target of sexual harassment without necessarily labeling it as such.

The cognitive process of labeling is similar to that found in the sexual assault literature, wherein the majority of individuals experiencing assault fitting the legal definition of rape do not label their experience as “rape” (Koss et al. 1996; Littleton et al. 2006). A major theme from the rape literature suggests that rape is far more likely to go unacknowledged by the victim if the experience fails to match up with one’s “rape script,” or a person’s expectations of the typical rape scenario (Peterson and Muehlenhard 2004). For instance, sexual assault perpetrated by an acquaintance does not match up to the typical rape script which, for many people, typically views only violent strangers as capable of perpetrating rape (Peterson and Muehlenhard 2004). Men in the U.S., for whom a rape script does not extend so far as to include expectations that they may one day personally experience sexual assault, demonstrate very low reporting rates in the aftermath of rape due, in part, to a low likelihood of labeling their experience as rape (Weiss 2010). We applied the same rationale to predict gender differences in the labeling of sexual harassment. That is, sexual harassment is less inclined to be acknowledged, or labeled as such, if it deviates from one’s expectations of a typical sexual harassment script. Men in general, relative to women, may reasonably expect a lower probability of being targeted for sexual harassment (Berdahl et al. 1996). For example, men may typically expect to be insulated from lewd and derogatory remarks and behaviors directed at them in professional settings. As a result, the sexual harassment script, the pre-established notion of the goals, behaviors, and outcomes likely to happen in this particular kind of social setting, for a typical man is ill-defined and lacking in detail. Therefore a man, possessing a less well-defined harassment script arising from low expectations of personal victimization, is predicted to label a qualifying experience as harassment at a rate much lower than that of a woman under similar circumstances (McMullin and White 2006; Munson et al. 2001). A male victim of sexual harassment is more inclined to apply an alternative label to qualifying behavior, such as harmless banter. Indeed, prior research using the SEQ has routinely found that U.S. adult men label their experience as sexual harassment less often than do adult women, findings we seek to replicate among a U.S. teen sample (Street et al. 2007). Review of the aforementioned literatures suggested the following hypothesis regarding labeling of sexual harassment:

  • Hypothesis 1: Teen women, compared to teen men, will demonstrate a higher rate of labeling their experience as sexual harassment.

Appraisal of Sexually Harassing Events

Another important and distinct aspect of this study investigated the subjective appraisal of teens targeted by sexual harassment. Our analysis of this appraisal segment relies upon a general theoretical framework for work stress (Lazarus and Folkman 1984). Sexually harassing behavior is regarded here as a specific type of work stressor (Vaile Wright and Fitzgerald 2007). Organizational stress models assert that stressors give rise to the cognitive appraisal process. Appraisal is “the process of categorizing an encounter and its various facets, with respect to its significance for well-being” (Lazarus and Folkman 1984, p. 31). Appraisal is regarded by classic stress theorists as an evaluative process by which one determines the significance of a social event to one’s health (Vaile Wright and Fitzgerald 2007). According to Lazarus and Folkman (1984), the appraisal process can be more precisely described as comprising the primary appraisal, or the largely unconscious process of perceiving an event as threatening; followed by a secondary appraisal, or the more conscious cognitive determination of how to respond to, or cope with, the perceived stressor.

Aside from labeling, the first objective of the current study was to examine proposed gender differences in subjective appraisal. Fitzgerald et al. (1997) proposed a model of harm in which appraisal of sexual harassment was considered to be a function of multiple factors arising from the situation itself (e.g., frequency, duration, and intensity of the harassing situation), from the context (e.g., organizational climate, management norms, policies/procedures), and from the individual (e.g. victimization history, and degree of perceived control). The broad range of characteristics that make an individual more susceptible to sexual harassment comprises a construct frequently referred to as personal vulnerability (Bergman and Henning 2008; Fitzgerald et al. 1994). We contend that women are generally more vulnerable targets for the upsetting effects of sexual harassment, as a result of a combination of several of these risk factors. Research focusing upon factors arising from the situation has indeed established that women are frequently targeted for harassing behavior (Bergman and Henning 2008; Bergman et al. 2002), and experience intense forms of harassment over a substantial duration of time (Vaile Wright and Fitzgerald 2007). Past research examining contextual determinants has reported that women often perceive one’s organization as tolerant of sexual harassment and unlikely to take complaints seriously (Bergman and Henning 2008; Bergman et al. 2002; Hulin et al. 1996). Moreover, investigations directed at individual-level predictors have found that women frequently have a history of prior harassment victimization (Vaile Wright and Fitzgerald 2007) and report feeling unable to exert control over the situation (Berdahl 2007; Berdahl et al. 1996). In combination, these risk factors heighten women’s personal vulnerability to sexual harassment and its ill effects. Investigations among U.S. samples support this idea. For instance, Fineran (2002) offered evidence that female adolescents were more upset than their male counterparts by unwanted sexual attention at work, suggesting a gender difference in the appraisal of sexual harassment. We therefore developed the next hypothesis as follows:

  • Hypothesis 2: Teen women, compared to teen men, will demonstrate a higher rate of appraising sexually harassing behaviors at work as bothersome.

Psychosocial Maturity as an Outcome of Sexual Harassment Appraisal

This study’s second objective was to examine the developmental outcomes for adolescents experiencing bothersome sexual work behaviors. We contend that, in gauging the impact of sexual harassment on development, the key determinant is not labeling or degree of incidence, but rather the individual’s subjective appraisal of the behavior (Koss 1990). To meet the goal of relating subjective appraisal of harassment to developmental outcomes, the adolescent development literature was utilized to identify consequences specific to the adolescent population. Psychosocial maturity was adopted for use in the current model as a developmental outcome potentially susceptible to the stressful reactions arising from upsetting appraisal of sexual harassment.

Greenberger and colleagues (Greenberger 1984; Greenberger et al. 1975; Greenberger and Sorensen 1974) introduced the psychosocial maturity concept as a set of dispositions reflecting U.S. adolescents’ capacity to meet society’s demands. The theory of psychosocial maturity begins by compiling and organizing a set of ideal personal and social outcomes drawn from multiple disciplines, including biology, sociology, and psychology. Each of these disciplines, in addressing the aging and development of organisms (including people), identifies common threads of traits believed to contribute to the basic survival of the organism (Greenberger and Sorensen 1974). The theory of psychosocial maturity integrates these threads, drawn from multidisciplinary sources, to identify the fundamental maturation goals of human development (i.e., traits indicative of the optimal growth of the individual in his or her own right), as well as the goals of socialization (i.e., traits required of the individual to make a society function smoothly) as those attributes most necessary for individual and species survival. The model goes on to propose that psychosocial maturity is reflected in three basic capacities, which correspond to the general demands universally placed upon individuals by all societies: individual adequacy (the capacity to function effectively on one’s own), interpersonal adequacy (the capacity to interact effectively with others), and social adequacy (the capacity to contribute to group-level social cohesion; Greenberger et al. 1975; Greenberger and Sorensen 1974). Adolescence is the optimal period in the course of human development when the individual attains the physical, emotional, and cognitive abilities to begin meeting these societal demands (Dresner and Grolnick 1996).

Empirical testing via factor analytic procedures subsequently led psychologists to modify the model and conclude that adolescent psychosocial maturity can best be described by two higher-order and coherent categories of functioning corresponding to the two broad goals of intrapersonal functioning vs. interpersonal functioning. These two higher-order maturity skills were addressed in this study: autonomy and social responsibility (Greenberger et al. 1975; Lamborn and Groh 2009). Autonomy is the individual’s capacity to function on one’s own, displaying the specific characteristics of self-reliance, a positive work orientation, and a clear sense of personal identity. Social responsibility is the individual’s capacity to contribute to a cohesive society, displaying the specific characteristics of social commitment, tolerance of others’ differences, and openness to social and political change. Studies of construct validity and factor structure have established autonomy and social responsibility as two distinct and internally coherent dimensions of maturity (Greenberger 1984).

Coping and Gender as Moderators of the Appraisal-Autonomy and the Appraisal-Social Responsibility Relationships

Summarizing the full model thus far, we anticipated that the subjective appraisal of a sexually harassing behavior as bothersome will significantly predict the psychosocial maturity outcomes of autonomy and social responsibility. Theoretical frameworks from the organizational stress literature, however, suggest that there is substantial variability in people’s reactions to work stressors based upon (a) an individual’s chosen coping strategy, and (b) an individual’s gender.

First, a person’s choice of coping may play a role in determining whether distressing sexual harassment will result in deleterious outcomes (Magley 2002). Coping is the “cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person” (Lazarus and Folkman 1984, p. 141). Traditionally, sexual harassment coping was classified along a unidimensional scale, differentiated only by degree of assertiveness (Malamut and Offermann 2001). Drawing from the classic Lazarus and Folkman (1984) typology of problem-focused vs. emotion-focused coping strategies, Fitzgerald and her colleagues (Fitzgerald et al. 1995) proposed that coping with sexual harassment falls along the parallel dimensions of external strategies (designed to manage the situation) vs. internal strategies (designed to manage one’s own thoughts and emotions associated with the situation). Multidimensional conceptualizations of sexual harassment coping have evolved as the preferred approach (Knapp et al. 1997; Magley 2002). One notably well-articulated and empirically-supported classification system was offered by Magley (2002). Data supporting this classification system were drawn from a variety of samples, including U.S. and Italian employed women, U.S. students, and members of the U.S. armed forces. This framework positions coping along two dimensions, method of coping (behavior vs. cognition) and focus of coping (approach/engagement vs. avoidance/disengagement), and is arranged as a 2 × 2 typology of responses. Thus, four classes of coping serve as potential moderators in the relation between appraisal and autonomy, and between appraisal and social responsibility: behavioral engagement, behavioral disengagement, cognitive engagement, and cognitive disengagement (Magley 2002). The first class of coping, behavioral engagement, involves behaviors aimed at approaching the perpetrator or the situation to end the harassment, and includes specific tactics such as assertion and seeking organizational relief. A second class, behavioral disengagement, involves behaviorally-based attempts to seek relief by avoiding the situation, and includes specific tactics such as avoidance and seeking social support. Cognitive engagement involves a thought process aimed at accepting the reality of the situation. Cognitive engagement includes strategies such as relabeling an incident as something other than harassment, appeasement, and self-blame. Cognitive disengagement involves thought processes aimed at avoiding the harassment and accepting it as difficult to change. Cognitive disengagement includes tactics such as detachment, denial, and endurance.

Second, in addition to coping behavior, the anticipated appraisal-autonomy and appraisal-social responsibility relationships may be further refined by gender. Stockdale and colleagues (Stockdale 1998; Stockdale et al. 1999) promoted a theory of gender-specific normative expectations which are enforced within the work setting: norms that emphasize hyper-masculinity in men and hyper-femininity in women. Stockdale (1998) formulated what she referred to as a “wimpy male” hypothesis to explain her findings among U.S. employees that men suffered more negative consequences from sexual harassment when they relied upon more confrontational coping behaviors, rather than ignoring the abuse. The man actively resisting sexual harassment, according to the “wimpy male” hypothesis, is violating hyper-masculine workplace norms and, as a result, will be prone to a range of health and work consequences stemming from the social disapproval often directed at deviants (Stockdale et al. 1999). Adolescent men and women, as newcomers, are often the targets of socialization processes in the work setting and particularly prone to enforcement efforts for what is regarded as appropriate behavior (Kroger 2000). Drawing from the same theoretical explanation as that of Stockdale, we anticipated that adolescent men using visible, or behaviorally-based, coping will be regarded as openly and publicly resisting the hyper-masculine norm expectations that they should simply “tough it out” in the face of harassment, and consequently will experience significantly worse autonomy and social responsibility maturity outcomes.

We extended Stockdale’s (1998) rationale to formulate and test a “weak female” hypothesis counterpart. Some past research reports stronger associations between sexual harassment and depression, anxiety, and other psychological disorders among women, relative to men (Harned and Fitzgerald 2002; Murdoch et al. 2006). Moreover, women tend to rely upon relatively non-assertive tactics in trying to cope with sexual harassment (Fitzgerald et al. 1995). We anticipated that young women using non-visible, or cognitively-based, coping – in other words, rigidly conforming to hyper-feminine norm expectations – will be particularly at risk for developing perceptions of oneself as weak, powerless, and victimized and will consequently experience significantly worse autonomy and social responsibility outcomes than other groups relying on these tactics (Frydenberg 2008).

In summary, this line of reasoning suggests that both coping strategy and target gender are moderators in each of the appraisal-autonomy and appraisal-social responsibility relationships. To test the following hypotheses, a series of hierarchical regression analyses will be performed in which the effects common to all hypothesis tests will be accounted for in the initial steps of the analyses (i.e., all main effects and interactions associated with appraisal and gender), thereafter incrementally accounting for those effects unique to each of the following hypotheses (i.e., all main effects and interactions associated with the given coping behavior associated with each hypothesis, respectively). The data will then be examined for the presence of a significant three-way interaction among appraisal, gender, and coping strategy in the final step of the regression analyses, which would serve as evidence supporting the following hypotheses:

  • Hypothesis 3: High appraisal will be associated with lower autonomy and lower social responsibility. Furthermore, it is hypothesized that this relationship will be amplified for specific groups, as follows:

    • Hypothesis 3a: The wimpy male hypothesis predicts that the appraisal-autonomy and the appraisal-social responsibility relationships will be more strongly negative among men high in behavioral engagement coping strategy, relative to the following groups: men low in behavioral engagement, women high in behavioral engagement, women low in behavioral engagement.

    • Hypothesis 3b: The wimpy male hypothesis predicts that the appraisal-autonomy and appraisal-social responsibility relationships will be more strongly negative among men high in behavioral disengagement coping strategy, relative to the following groups: men low in behavioral disengagement, women high in behavioral disengagement, women low in behavioral disengagement.

    • Hypothesis 3c: The weak female hypothesis predicts that the appraisal-autonomy and appraisal-social responsibility relationships will be more strongly negative among women high in cognitive engagement coping strategy, relative to the following groups: women low in cognitive engagement, men high in cognitive engagement, men low in cognitive engagement.

    • Hypothesis 3d: The weak female hypothesis predicts that the appraisal-autonomy and appraisal-social responsibility relationships will be more strongly negative among women high in cognitive disengagement coping strategy, relative to the following groups: women low in cognitive disengagement, men high in cognitive disengagement, men low in cognitive disengagement.

Method

Participants

A total of 586 introductory psychology students at a Midwestern U.S. university served as participants for this study: 316 women and 270 men (representing a 43% response rate among those students offered the opportunity to participate). Participants’ ages ranged from 18 to 24. The average age of women was M = 18.77 (SD = 1.06) years, which was significantly less than the average age of men, M = 19.15 (SD = 1.35), t = 3.71, p < .05. The average age of women when they obtained their first job was M = 15.57 (SD = 1.04) years of age, which was significantly greater than the average age of men at time of first job, M = 15.27 (SD = 1.40), t = −2.99, p > .05. The majority of the sample comprised White/Caucasians (85% for both women and men), with a minority identifying as African American/Black (6% for men, 9% for women) and Hispanic (3% for men, 2% for women). The sample was composed of 73% freshmen (both women and men), and others were sophomores and upper-level students. They received extra credit in their introductory psychology courses in exchange for participation.

Materials

When visiting a class, a team recruiter described the set of surveys designed to examine their work attitudes and experiences. Potential recruits were also instructed that, to be eligible to participate, they were to have had some type of formal job before the age of 19, and to be currently under age 25. Repeated reminders appeared throughout the survey, instructing participants to respond to the questions relative to their “very first formal job.”

Labeling of Sexual Harassment

One instrument was an adapted version of the Sexual Experiences Questionnaire, the SEQ-DoD (Stark et al. 2002). The instructions were modified by directing participants to describe their experiences with other-gender co-workers and supervisors while working at their first job as a teenager. The contents of the male and female versions were identical with the exception of gender-specific pronouns. The SEQ first assessed frequency of experience with a set of sexually harassing behaviors (0 = “never,” 1 = “once or twice,” 2 = “sometimes,” 3 = “often,” and 4 = “many times”). Stark et al. (2002) reported that this 16-item instrument displayed a 4 4-item factor structure: sexist hostility, sexual hostility, unwanted sexual attention, and sexual coercion. The SEQ has also been validated for young adults reflecting upon their earlier experiences with sexual harassment during their first jobs as teenagers (Harris et al. 2005). A common stem for every question read, “While working at your first place of employment, were you in a situation where any of your male (female) supervisors or coworkers…” A sample item for sexist hostility completed the stem with, “treated you ‘differently’ because of your gender?”; a sample item for sexual hostility was, “repeatedly told sexual stories or jokes that were offensive to you?”; a sample item for unwanted sexual attention read, “made unwanted attempts to establish a romantic sexual relationship with you despite your efforts to discourage it?”; and an item from the sexual coercion scale stated, “made you feel threatened with some sort of retaliation for not being sexually cooperative?”

The single item of interest for the first hypothesis, the labeling item, read “Have you ever been sexually harassed?” It was located at the end of the SEQ, measured on the same scale as the other SEQ items, and represented the only mention of the phrase “sexual harassment” in the entire instrument. This single item was used for purpose of analysis in testing the first hypothesis.

Subjective Appraisal

The SEQ instrument further assessed the respondents’ subjective appraisal of the 16 harassing behaviors. Alongside the frequency scales, a second column was formed with the common stem, “If at least once, how much did this bother you?” Responses were measured on a Likert scale with options of 1 = “not at all,” 2 = “slightly,” 3 = “somewhat,” 4 = “very much,” and 5 = “extremely.” Respondents’ raw scores were then transformed to place all scores on the same scale, regardless of number of SEQ items selected. The transformation consisted of, first, converting 1 = “not at all” to zero, and making a similar subtraction for all scores such that 1 = 0, 2 = 1, 3 = 2, 4 = 3 and 5 = 4. Then, the summed “bothered” score for any given scale was divided by the number of SEQ items endorsed. This appraisal score then controls for the number of behaviors experienced. Coefficient alpha for the full 16-item appraisal scale was .92 and .94 for women and men, respectively. A high appraisal score represented greater upset, and was used for purpose of analysis in testing the second hypothesis.

Coping with Harassment

Coping strategy was assessed using the Coping with Harassment Questionnaire (Magley 2002). Four classes of coping were assessed: behavioral engagement (comprising the specific tactics of assertion and seeking organizational relief, represented by a total of ten items; coefficient alpha was .89 and .93 for women and men, respectively), behavioral disengagement (avoidance and seeking social support; ten items; coefficient alpha was .92 and .93 for women and men, respectively), cognitive engagement (relabeling, appeasement, and self-blame; 15 items; coefficient alpha was .85 and .89 for women and men, respectively), and cognitive disengagement (denial, detachment, and endurance; 15 items; coefficient alpha was .89 and .91 for women and men, respectively). Participants responded on a 5-point Likert scale the extent to which each item was descriptive of their reactions; high scores represented items that were highly descriptive. Scores were summed across items for each dimension of coping.

Autonomy and Social Responsibility

Two categories of psychosocial maturity, autonomy and social responsibility, served as the psychosocial maturity outcomes in this study (Greenberger 1984). Autonomy (the individual’s capacity to function on one’s own) was measured with 60 items tapping the characteristics of self-reliance, positive work orientation, and personal identity. Social responsibility (the individual’s capacity to contribute to a cohesive society) was measured with 67 items tapping the characteristics of social commitment, tolerance of others’ differences, and openness to social and political change. All items were assessed with 4 response options, with 1 = disagree strongly and 4 = agree strongly. Respondents were asked to rate their current levels of autonomy and social responsibility. Scores were summed across items for each dimension of psychosocial maturity. A high score represented high maturity. Coefficient alpha for autonomy was .93 for both women and men. Coefficient alpha for social responsibility was .88 and .89 for women and men, respectively.

Procedure

A team recruiter visited with introductory psychology courses in advance of survey administration, to inform students of an opportunity to gain extra credit in their classes. Students had other extra credit opportunities, aside from this one. During the class periods prior to a week’s break from school, the questionnaire packet was distributed to students indicating a willingness to participate. The cover page disclosed the purpose of the study, standard informed consent information, and an eligibility requirement (“You must have had some type of formal job before the age of 19 in order to participate in this survey”). The cover page disclosed that survey questions would be asking about work attitudes, behaviors, and experiences relevant to one’s first formal job. The cover page did not mention sexual harassment, to avoid priming participants’ memories of such events and perhaps influencing their survey responses. Students took the packet home during their break, and returned it when classes reconvened a week later. Surveys were collected in a hallway outside the classroom context, by one project team member, as students arrived to their classes. One project team representative was responsible for collecting the survey and ensuring that no name or other personally identifying information was on the document. Thereafter, a second team representative solicited the respondent’s name, and participation for extra credit purposes was recorded on a separate class roster.

Results

Descriptive Statistics

Preliminary analysis began with an examination of the means, standard deviations, intercorrelations, and coefficient alpha values among the major variables in this model. Table 1 provides these values for women and men separately. Visual inspection of values in Table 1 offers some insights into how people were responding to the scales. For instance, with few exceptions, women obtained higher average scores on the major variables, thus displaying higher levels of appraisal, higher labeling, higher behavioral engagement and disengagement, higher autonomy, and higher social responsibility scores. The exceptions, where men were displaying higher scores, included the variables of cognitive engagement and cognitive disengagement. Reliability estimates for both men and women were similar in value, thus suggesting that members of both genders were responding to the scales in a reasonably uniform fashion. Additionally, visual examination of Table 1 values yields no discernible difference in the intercorrelation values for men and women.

Table 1 Means, standard deviations, and intercorrelations for all variable for men (Lower Triangle) and women (Upper Triangle)
Full size table

Gender Differences on Labeling and Appraisal

Hypothesis 1 predicted greater frequency of labeling among women, relative to men. To test Hypothesis 1, a t-test was utilized comparing women and men on the single labeling item of the SEQ (which was located at the end of the SEQ instrument and read, “Have you ever been sexually harassed?”). A significant gender effect emerged, t(580) = −3.18, p < .01, Cohen’s d = .26, in support of the first hypothesis. Average labeling for women (M = .49, SD = .96) was greater than that for men (M = .27, SD = .75).

Hypothesis 2 predicted greater tendency to report upsetting appraisal among women, relative to men. Hypothesis 2 also utilized a t-test comparing women and men on the total appraisal score. An initial Levene’s test for equality of variance indicated that the distributions of the two groups deviated from the homogeneity of variance assumption, thus requiring appropriate adjustment of the statistical values. In support of the hypothesis, mean values indicated that women (M = 11.17, SD = 11.37) reported being significantly more bothered by incidents of sexual harassment, in comparison to men (M = 5.46, SD = 8.51), t(305.42) = −5.08, p < .01, Cohen’s d = .57.

The Relationship of Subjective Appraisal to both Autonomy and Social Responsibility

Hypotheses 3a through 3d predicted that appraisal would be related to lower autonomy and social responsibility, but this relation would be amplified for specific groups depending upon coping strategy and gender. A series of hierarchical regressions was conducted for each type of coping strategy and for each of autonomy and social responsibility, respectively, in which the effects that were common for all models were initially accounted for, and then the unique effects were thereafter added incrementally. Therefore, for each regression, the simple effects for appraisal and gender were added in the first step, and their interaction added in the second step. The simple effect for the relevant coping strategy was added in the third step. Then all relevant two-way interactions with coping, and the three-way interaction, were added in the fourth and fifth steps, respectively. Prior to creating the interaction terms, the predictor (appraisal) and moderator (coping strategy) were centered (by subtracting each value from its respective mean) to reduce multicollinearity and achieve better estimates of the interaction terms (Aiken and West 1991; Cohen et al. 2003). A significant regression weight on a two-way interaction term in Step 4 of the regression would signify that the relationship between appraisal and autonomy (or between appraisal and social responsibility) was moderated by either coping strategy or gender, respectively. Likewise, a significant regression weight on the three-way interaction term in Step 5 would serve as support for Hypotheses 3a through 3d by demonstrating that both the relevant coping strategy and gender serve as moderators in the association between appraisal and autonomy (or between appraisal and social responsibility). Because a total of seven significance tests were conducted for any set of regressions, the Bonferroni adjustment was applied, establishing a significance level of .007 (α of .05, divided by 7).

As a side note, an alternative set of analyses was conducted which included SEQ frequency as a control variable included in Step 1 of the regression. Diagnostic values indicated an exceedingly high degree of multicollinearity between the appraisal and SEQ frequency scores, even after the variable centering procedure, suggesting that appraisal and SEQ frequency contained highly redundant information. The variance inflation factors (VIF; An index of degree of multicollinearity) for SEQ frequency and appraisal, across both autonomy and social responsibility, and across all coping styles, were values ranging from 34.37 to 54.39. The common rule of thumb is that a VIF exceeding ten indicates significant multicollinearity problems. Therefore the choice was made to include only appraisal, the key theoretical predictor of our model, in the final analyses. After SEQ frequency was dropped from the analysis, the VIFs were re-calculated, and all were within the commonly accepted range of less than 10 (specifically, VIF values ranged from 1.06 to 6.03). One exception was found in the three-way interaction term appraisal x gender x behavioral disengagement, yielding a VIF of 12.07, which was retained in the equation due to its central role in the theoretical underpinnings of the study (Belsley et al. 1980).

The first set of analyses testing Hypothesis 3a was conducted to examine whether behavioral engagement and gender moderated each of the appraisal-autonomy and the appraisal-social responsibility relationships. Results did not support the three-way interaction in Hypothesis 3a. However, the behavioral engagement x gender two-way interaction emerged for social responsibility (see Table 2). For interpretation of the interaction, we used a graphing technique examining the simple slope for social responsibility regressed on appraisal for men and women separately (Aiken and West 1991). Figure 1 shows that behavioral engagement was related to lower levels of social responsibility among men, but that relation did not hold among women (Dawson and Richter 2006).

Table 2 Summary of hierarchical multiple regression results: behavioral engagement coping
Full size table
Fig. 1
figure 1

Relationship between behavioral engagement (BE) and social responsibility for women and men

Full size image

The next set of analyses testing Hypothesis 3b examined whether behavioral disengagement and gender moderated each of the appraisal-autonomy and the appraisal-social responsibility relationships. Both autonomy and social responsibility displayed significant interactions (in Step 5 of the regression analyses; See Table 3) involving behavioral disengagement, thus fully supporting the hypothesis. Three-way interactions were revealed for autonomy and social responsibility (see Fig. 2 for autonomy, which was similar to the pattern for social responsibility). In Fig. 2, we would expect that men relying on behavioral disengagement would show a pronounced appraisal-autonomy relationship, relative to the other three groups, if there is validity to the wimpy male hypothesis and, indeed, the trends point in that direction for both autonomy and social responsibility. Figure 2 demonstrates a strong negative relationship between appraisal and autonomy among the male/high behavioral disengagement group. Moreover, slope difference tests showed the male/high behavioral disengagement group as different from both of the female groups (female/high behavioral disengagement and female/low behavioral disengagement), and marginally different from the male/low behavioral disengagement group (Dawson and Richter 2006). A similar pattern of relationships was indicated in the three-way interaction for social responsibility.

Table 3 Summary of hierarchical multiple regression results: behavioral disengagement coping
Full size table
Fig. 2
figure 2

Relationship between appraisal and autonomy under varying levels of behavioral disengagement (BD) and gender

Full size image

The third and fourth sets of analyses testing Hypotheses 3c and 3d examined whether cognitive engagement and cognitive disengagement, respectively, and gender moderated each of the appraisal-autonomy and the appraisal-social responsibility relationships. The hypotheses were not supported in that no three-way interactions involving the cognitive coping strategies emerged as significant in Step 5 of the regression analyses. The full set of regression results for cognitive coping strategies is available from the corresponding author upon request.

In summary, the results for Hypothesis 3a and 3b lent support for the wimpy male hypothesis, whereas tests of Hypotheses 3c and 3d disconfirmed the weak female hypothesis (Stockdale 1998; Stockdale et al. 1999). Specifically, findings showed diminished autonomy and social responsibility among distressed men who reported using visible coping behaviors, which are incongruent with hyper-masculine normative expectations.

Discussion

Results suggest that U.S. families and institutions should be concerned about the exposure of adolescents to harassing behaviors upon introduction to the formal world of work. Given the pervasive nature of sexual harassment, it stands to reason that such concern for the welfare of working adolescents should extend beyond the border of the U.S., although data in the current study do not speak directly to the experiences of youth outside the U.S. The objectives for this study were to describe the labeling and appraisal of teens targeted by sexual harassment, and to examine the autonomy and social responsibility outcomes among teens who appraise such events as threatening to their welfare. Results replicated past findings indicating that women, relative to men, more often were upset by harassing behavior, and more frequently used the label of “sexual harassment.” The findings further revealed that men, and especially men appraising harassment as bothersome and using behaviorally-based strategies to cope, reported significant detriment to the development of autonomy and social responsibility.

Though not formally addressed as a hypothesis in the current study, our results did offer some information about the degree of teenager exposure to harassing behaviors at work. Results replicated prior research indicating that U.S. female adolescents are the more frequent targets of a range of sexualized behavior. In fact, 61% of women in this sample endorsed at least one scale item on the SEQ, a value not equal to – but approaching – the high endorsement rates reported among U.S. military women in other studies (77% reported in Fitzgerald et al. 1999; and 74% reported in Stark et al. 2002). Though cross-study comparisons must be interpreted cautiously, a similar examination among men revealed an intriguing trend wherein U.S. adolescent males in this study reported substantially higher exposure rates (26% for unwanted sexual attention; 46% for full SEQ scale) than previously identified levels of exposure to unwanted sexual behavior (8% for unwanted sexual attention; 37% for full SEQ scale) among U.S. male military samples (Fitzgerald et al. 1999). Moreover, the incidence rates obtained in the current study are likely attenuated because respondents were asked only about cross-gender harassment. Because same-gender harassment is common, especially among men, inclusion of that form of harassment would likely have inflated incidence values among adolescent males (Berdahl 2007). Fineran (2002) observed a similar pattern of substantially higher numbers of U.S. adolescent boys reporting sexual harassment, relative to other studies of adult male victimization. Though conclusions are tentative at this stage, this comparison among studies compels the speculation that both adolescent women and men may be subjected to relatively high levels of sexually harassing behaviors, and the risk of being a target decreases as people gain more workplace and life experience. Power-based theories in the sexual harassment literature offer a perspective that may aid in explaining the unusually high exposure rates among adolescent men. Though the particulars may vary among these gendered power theories, they have in common the suggestion that a primary motive for sexualized harassment toward members of any group is to protect one’s social status when it seems threatened (Berdahl 2007; Duncan 1999). Aside from qualities such as gender and race, age serves as yet another cue by which to stratify social status. By this reasoning, young men are vulnerable as targets of sexual harassment due to their lower status arising from younger age. Adolescent women are jeopardized by lower status arising from both gender and younger age (Richman et al. 1999).

The Subjective Experience of Sexual Harassment

Data collected from the current study speak to subjective experience in two ways: labeling and cognitive appraisal. First, exposure to potentially harassing behavior does not mean that a target will necessarily label such behavior as “sexual harassment” (Magley et al. 1999; Munson et al. 2001). Anecdotal evidence indicates that behaviors appearing on the SEQ are at times considered mere playful banter among young women and men, tacitly designed to maintain participants’ energy levels (Berdahl and Aquino 2009). In contrast to accounting of behavioral incidents, the scale’s final labeling item, “Have you ever been sexually harassed?” was designed to capture participants’ inclinations to label in the wake of a potentially harassing event (Bergman et al. 2002). Although a gender effect emerged, in absolute terms, the frequencies of labeling for both men and women were rather low (M = .49 for women and M = .27 for men, thereby falling between the response options of 0 = “never” and 1 = “once or twice,” on a 4-point scale). The generally low endorsement rates for our final labeling item indicate that adolescents do, indeed, sometimes perceive sexualized behavior as a form of social contact that does not rise to the level of what they would label as “sexual harassment,” but what might be considered as such by someone with greater life experience (Berdahl and Aquino 2009). Alternatively they may lack the knowledge and experience to be able to accurately interpret disconcerting social experiences as harassment due to an ill-defined sexual harassment script (Bremer et al. 1991), or fear related labels such as victim or whistle-blower (Munson et al. 2001). Despite the depressed labeling frequencies among men and women, the results did nonetheless reveal a significantly higher labeling rate for women, as might be expected among a group that is more often targeted for sexual behaviors. The support for Hypothesis 1, which predicted a gender difference in labeling, is consistent with our argument that men’s sexual harassment scripts, lacking in sufficient detail and personal relevance, increase their likelihood of failing to acknowledge their role as victims when experiencing behaviors that objectively qualify as sexual harassment. In contrast, women, even those in their adolescent years, possess more well-defined sexual harassment scripts and are therefore likely to label their experiences as sexual harassment at a greater rate, compared to their male counterparts. Future research may be designed to examine directly and compare the sexual harassment scripts of male and female adolescents.

Our second approach to capturing subjective experience was through the use of a separate scale rating degree to which respondents were “bothered” by a behavioral incident. These appraisals of discrete incidents reflected a generally low level of distress among all participants when responding to specific examples of sexualized behaviors. Apparently exposure to sexual behavior at work is not necessarily an upsetting experience, at least not among U.S. adolescents (Berdahl and Aquino 2009). Nonetheless further analysis revealed gender differences for the appraisal scale. Predictions for a gender difference in appraisal were drawn from a model of harm, developed by Fitzgerald and colleagues (Fitzgerald et al. 1997). This model suggests that level of appraisal for a target of sexual harassment is determined by a variety of factors arising from the situation, the organizational context, and characteristics of the target person. Because women are generally more vulnerable to several of the risk factors identified in this model, such as frequency of exposure and unfavorable organizational norms, our second hypothesis anticipated a gender difference in appraisal. As predicted, female respondents were more inclined to perceive sexualized behavior as a potential threat to their well-being than were male respondents, thus replicating a relation between gender and cognitive appraisal that has become established in the literature (Fineran 2002; Lacasse and Mendelson 2006).

The Role of Coping

Another area in the study of subjective experience is the role of coping by a target of harassment. A recent literature on coping with sexual harassment has focused upon predicting individuals’ coping behaviors (Cortina and Wasti 2005; Malamut and Offermann 2001). The current study contributes novel information aimed at the next question of coping strategy efficacy. The theme emerging from this study, labeled in prior research as the wimpy male concept, was the harm, in the form of reduced autonomy and social responsibility, befalling U.S. adolescent men who appraised sexual harassment as threatening and relied upon behavioral coping. These findings suggest that the high standards for masculine conduct continue to operate in U.S. work settings, and penalize men who are upset by and visibly oppose unwanted sexual behavior (Mosher and Tomkins 1988; Stockdale 1998; Stockdale et al. 1999). This pattern of results maps well onto the threat-to-status theory for sexual harassment perpetration (Berdahl 2007). According to the threat-to-status theory, sexual harassment is motivated by a desire to protect social status and, therefore, it is reasonable to expect that any man behaving contrary to masculine norms will be a prime target. It is worth noting that, ironically, behavioral coping in fact amounts simply to following the rules established by most modern organizations when one is victimized by sexual harassment. The weak female hypothesis counterpart, predicting disparate outcomes for men and women when bothered by sexual harassment and relying on cognitive coping, was not supported by the data. In hindsight, stronger support for the wimpy male hypothesis might have been anticipated because more social demands are placed upon men in the U.S. (relative to women) to meet gender-based ideals, thus the penalties for failure to fulfill traditionally masculine ideals would be heavier (Berdahl 2007).

A rich knowledge base from clinical psychology offers some explanation for the consequences for teenagers exploited in their earliest work experiences. That literature speaks of employment as a central adult role, and sexual harassment’s potential for triggering questions about identity and life purpose (Frone 1999; Koss 1990). The adolescent target would be vulnerable to the activation of negative self-images, views of oneself as weak and ineffectual. Moreover, this literature suggests that the at-risk teen may question previously established central beliefs of the world, and work, as meaningful and safe (Taylor 1983). The adolescent’s inability to adjust to sexual stressors at work may manifest in arrested development of adult capacities to operate autonomously and in a socially responsible manner. This research contributes to the literature by identifying young men as those most vulnerable to delayed development of autonomy and social responsibility when exposed to upsetting sexual harassment. These findings do not negate the widespread harmful effects of harassment for women, who are clearly the most frequent targets. Nonetheless the results for men parallel those in the sexual assault literature, which indicate that U.S. adult men experiencing sexual stressors, up to and including assault, may respond with questions about, and feelings of threat to, their own gender roles. Scholars from the clinical literature explain that men are more vulnerable to ill effects because adult sexual trauma is not considered normative in our society. Thus, a male target questions his own gender role and experiences increased feelings of powerlessness and self-blame (Platt and Busby 2009; Singer 1989; Street et al. 2007). Similarly, social power theory would suggest that men in general, possessing greater social power as conferred by society, would be more likely to perceive self-assertion as a viable option in response to harassment (Cortina and Wasti 2005; French and Raven 1959). Threat-to-status explanations would further indicate worsened outcomes, as a result of perceived threats to organizational, social, and physical status, for distressed men when their behaviorally-based responses fail to be effective (Berdahl 2007; Berdahl et al. 1996). This information from both the sexual assault and social power literatures may begin to offer explanations for the harsh outcomes realized by men in our sample who exercised behavioral opposition to distressing sexual attention at work.

Limitations and Future Directions

The current investigation is limited in generalizing beyond the types of jobs, gender of perpetrator, and the nationality and age groups from which the data were collected. Conclusions drawn from this data are applicable only to U.S. college-attending adults recalling their teen work experiences. Therefore interpretations reflect the self-reported responses of relatively inexperienced workers concentrated in select U.S. labor force sectors, for example, service industries, clerical, and low-skilled labor – who had advanced on to college. Further, participants were instructed to respond specifically about forms of harassment perpetrated by the other gender which, given that same-gender harassment (especially among men) is a somewhat common occurrence in select work contexts, restricts the degree to which all forms of sexual harassment are adequately captured by these data (Berdahl 2007).

Another limitation of this study was its use of retrospective recall. Respondents reflected back several years to their first formal jobs. The average age of respondents was about 19, no respondent was over age 24, and the average time from first job was about 3.5 years, so they were recalling events from the recent past. This methodology carries the challenges associated with memory limitations and reconstruction. We are nonetheless optimistic about the reasonable accuracy of the information derived from respondents. Past research in social cognition suggests that the recall and judgment of social events, including one’s thoughts, feelings, and behaviors associated with those events, are enhanced by the ease of retrieval of that information (Caruso 2008). Affect-laden past social events and consequences, like sexual harassment occurring during one’s first venture into adult work life, take on added informational value and would be easily retrieved from memory (Clark and Teasdale 1982; Schwarz and Clore 1983). In that vein, we have argued that cognitive appraisal, in particular, offers considerable value from a theoretical standpoint in that it reflects a person’s judgment of the intensity and significance of an affect-laden social event to his or her well-being (Koss 1990; Vaile Wright and Fitzgerald 2007). This kind of judgment carries the potential to have a long-lasting effect on a young adult in that the feelings associated with harassment are easily retrieved from memory years after the precipitating harassment event has taken place (Nisbett and Ross 1980).

A future research avenue is an elaboration upon the nature of coping. Scholars depict coping as a dynamic concept, which undergoes a process of trial-and-error (Gutek and Koss 2007; Malamut and Offermann 2001). In that vein, a target of harassing behavior faces a significant dilemma in choice of response. Assertive behavior may easily backfire, yet passivity is rarely effective at achieving an end to harassment. Knapp et al. (1997) offered a predictive framework in suggesting level of distress as a determinant of people’s movement over time toward more extreme forms of coping -- such as eventually leaving the organization in order to avoid the situation, or shifting across time toward whistle-blowing or legal action. Given the common finding that women show higher distress levels than do men, an examination of gender differences in movement toward extreme coping may be warranted. Moreover, we acknowledge an alternative explanation for our finding of a coping-outcome relationship, a reversal in causal sequencing in which individuals already experiencing undesirable health problems may be limited in their choice of coping behavior. Given the multitude of questions to untangle about choice and sequencing of coping, future research may benefit from a longitudinal examination of individuals’ coping efforts over the course of time.