Initial assessment of the psychometric properties of the Sexual Harassment Reporting Attitudes Scale

Abstract Although sexual harassment remains a common problem in the workplace, many victims and witnesses fail to report it to organizational authorities. Fear of retribution as well as perceptions of futility are widely cited as reasons for these reporting failures. Attitudes toward reporting sexual harassment in the workplace have only been marginally considered, and a scale to measure such attitudes does not currently exist. In the present study, we describe a newly developed scale to test this construct, as well as its initial psychometric properties. Our findings suggest that the Sexual Harassment Reporting Attitudes Scale (SHRAS) is reliable and valid, with three correlated subscales related to the risks of reporting, moral obligation to report, and its utility. The benefits and uses of the SHRAS for scholars and practitioners are discussed.

ABOUT THE AUTHORS Dr. Brian Cesario is an assistant professor of industrial-organizational psychology, and has research interests in workplace sexual harassment and other forms of interpersonal aggression in the workplace. He collaborated with Dr Elizabeth J. Parks-Stamm (an experimental social psychologist with research interests in self-regulation and the role of stereotypes in performance situations) and Ms Mujgan U. Turgut (a former graduate student of his) on this research. The goal of this research was to develop a new scale to measure attitudes about reporting workplace sexual harassment, which can be used as an organizational training aid as well as for future empirical investigation.

PUBLIC INTEREST STATEMENT
Although sexual harassment remains a common problem in the workplace, many victims and witnesses fail to report it to organizational authorities. Fear of retribution as well as perceptions of futility are widely cited as reasons for these reporting failures. Attitudes toward reporting sexual harassment in the workplace have only been marginally considered, and a scale to measure such attitudes does not currently exist. In the present study, we describe a newly developed scale to test this construct, as well as its initial psychometric properties. Our findings suggest that the Sexual Harassment Reporting Attitudes Scale (SHRAS) is reliable and valid, with three correlated subscales related to the risks of reporting, moral obligation to report, and its utility. The benefits and uses of the SHRAS for scholars and practitioners are discussed. about 25% of all US women and 10% of all US men have experienced some form of SH in their places of work, but these percentages vary from industry to industry (e.g., Stampler, 2014). Typically, harassment rates are higher in industries where gender-typed work is common (Chang, 2003). Although establishing precise prevalence rates has historically proven challenging methodologically (see Arvey & Cavanaugh, 1995), there is a general consensus that SH is a widespread phenomenon requiring continued attention and consideration from both the scientific and management communities.
With workplace SH victim counts numbering into the thousands, the pervasiveness of the effects borne by people who were sexually harassed has become an important area for research. The most common consequence to victims of workplace SH is a negative impact on mental health (Cortina & Leskinen, 2013;Dionisi, Barling, & Dupré, 2012;McDonald, 2012;Nielsen & Einarsen, 2012). Other serious consequences of SH victimization include cardiovascular disease (Tuckey, Dollard, Saebel, & Berry, 2010), gastrointestinal or digestive issues (Brown et al., 2011), and disordered eating (Buchanan, Bluestein, Nappa, Woods, & Depatie, 2013). Although much of the research regarding workplace SH consequences has focused on female victims, similar findings have been observed for male victims (Holland, Rabelo, Gustafson, Seabrook, & Cortina, 2016). Thus, it is evident that SH results in poor outcomes across a broad array of social groups.
In light of the negative effects that workplace SH victims may experience, it is important to examine what factors prevent these behaviors from being identified, reported, and addressed. In particular, attitudes about SH-held by victims, harassers, and organizational authority figures-may be useful in understanding why underreporting and failure to report persists. Research about workplace SH-related attitudes is not new, with some studies dating back to the 1980s (i.e., Konrad & Gutek, 1986;Tangri, Burt, & Johnson, 1982). The relationship between tolerance of sexually harassing behaviors and attitudes about SH has been long studied (McCabe & Hardman, 2005;Reilly, Lott, Caldwell, & DeLuca, 1992;Russell & Trigg, 2004;Sigal & Jacobsen, 1999). Findings in this realm have remained consistent over time: those who are less tolerant toward SH tend to appraise more behaviors as sexually harassing than those with greater tolerance toward SH (Bitton & Ben Shaul, 2013). Likewise, organizational climates or cultures moderate the relationship between organizational wrongdoingincluding SH-and reporting (Hertzog, Wright, & Beat, 2008;Miceli, Near, Rehg, & Van Scotter, 2012;Vijayasiri, 2008). Thus, workplace SH may persist due to high tolerance of sexually harassing behavior, altering how the behavior is interpreted, resulting in lower reporting rates.
Even when workplace SH is appraised as problematic, victims and bystanders do not always report these behaviors. Although reporting is considered an effective method of coping (Callahan, Dworkin, Fort, & Schipani, 2002), it is often underutilized for a variety of reasons. Fear of retaliation by offenders has been widely cited in the literature as a potential reason for under-reporting (Bergman, Langhout, Palmieri, Cortina, & Fitzgerald, 2002;Diekmann, Walker, Galinsky, & Tenbrunsel, 2013;Vodanovich & Piotrowski, 2014). There is also evidence to suggest that SH goes unreported because victims believe that reporting would be futile (Fitzgerald, Swan, & Fischer, 1995). Although there is a paucity of other potential explanations among SH victims, research pertaining to a broad array of crimes suggests that underreporting may be due to victims not appraising the situation as severe enough (Singer, 1988), beliefs that reporting will worsen a bad situation (Fleury, Sullivan, Bybee, & Davidson, 1998), emotion-based reasons (e.g., embarrassment and shame; Finkelhor, Wolak, & Berliner, 2001), social pressures (Feldman-Summers & Norris, 1984, or fear of further victimization from those in positions of authority (Kidd & Chayet, 1984). Although researchers have suggested multiple plausible reasons that could explain SH underreporting, attitudes toward reporting SH have yet to be empirically measured. Thus, an important next step is to develop a way to measure attitudes about reporting workplace SH, given the potential for such attitudes to influence actual reporting behaviors.
In light of the findings discussed here, a tool to measure SH reporting attitudes is needed for practitioners and scholars alike. Practitioners, particularly those who provide professional training and development in the areas of SH awareness and sensitivity, could benefit from gauging their trainees' attitudes and thus tailor their instruction accordingly. Understanding current attitudes before training is important, as some evidence in the literature suggests that training alone may not encourage SH victims to report, and, in fact, may discourage it (Goldberg, 2007). Thus, it may be necessary for training curriculum developers to emend or revise their materials to attenuate these potential effects. Researchers would also benefit from its use. Although research has suggested attitudes are important for predicting reporting of sexual harassment (e.g., Chen & Tang, 2006), no instrument exists to measure this construct. This paper will present a new measure, the SHRAS, as well as an initial assessment of internal consistency, concurrent validity, and discriminant validity.

Participants
The sample for this study consisted of 646 online panelists recruited via Amazon ® Mechanical Turk. The study was open to respondents worldwide; however, respondents were required to be fluent in English to complete the survey. After accounting for respondents who failed to complete the entire SHRAS, however, the final sample size was 586. After IRB approval was obtained, a link to the study was posted on the site. Those who completed the entire survey were given a nominal monetary incentive, consistent with typical research inducements for participants recruited via Mechanical Turk (Buhrmester, Kwang, & Gosling, 2011).
We restricted the sample to only persons who reported any current or former employment. The majority of our sample was currently employed in one full-time, or one part-time job, or one fulltime job and one part-time job (66%; n = 424). Employed respondents reported having worked at their current jobs an average of 54.68 months (SD = 61.73 months), and 158.81 months (SD = 130.28 months) overall. The largest proportion (45%; n = 292) of the sample was between the ages of 26 and 35 years, although a substantial portion of the sample was comprised of persons aged 18-25 years (21%; n = 135) and 36-45 years (19%; n = 125). The sample was predominantly female (55%; n = 357), Caucasian (55%; n = 354), and had completed a four-year college degree (38%; n = 242).

Materials and procedure
Participants completed the SHRAS, followed by the Sexual Harassment Attitudes Scale (SHAS; Mazer & Percival, 1989) in order to test discriminant validity. Brief descriptions of each of these instruments are below. Simple demographic data were captured at the end of the survey. All data were collected online using Google Forms.

SHRAS item development and expert review
After a thorough review of the literature surrounding SH reporting, 18 items reflecting the perceived utility and drawbacks of reporting workplace SH were drafted. The items were developed to reflect themes identified in the literature pertaining to reasons why people typically fail to report workplace SH and other offenses. Such themes included beliefs that reporting is inefficacious (e.g., Firestone & Harris, 2003;Fitzgerald et al., 1995;Near & Miceli, 1985;Peirce, Smolinski, & Rosen, 1998), that it results in retaliation from the offending party or from the organization (e.g., Kenny, Fotaki, & Scriver, 2018;Mesmer-Magnus & Viswesvaran, 2005;Pershing, 2003), and that reports will not be taken seriously (e.g., Kaptein, 2008;Williams, Fitzgerald, & Drasgow, 1999). We additionally included one item at the end of the scale to measure respondents' intentions to report workplace SH to establish concurrent validity as in Fitzgerald et al. (1988). The items were then subjected to expert review by members of the Society for Industrial-Organizational Psychology (SIOP) who had specific expertise in workplace sexual harassment or survey construction. Twenty-two SIOP members participated in the expert review, rating each survey item's relevance to the overall construct and its clarity on a scale of 1 ("not at all clear"/"not at all relevant") to 4 ("completely clear"/ "highly relevant"). Quantitative and qualitative data from this small-scale expert review study revealed that all 19 items were both relevant and clear (M Relevance > 3.0; M Clarity > 3.0; see Table 1), and that only minor revisions (e.g., wording modification for consistency) were needed prior to the current study. Based on the feedback received from the expert panel, we concluded that the initial set of items possessed sufficient content validity. We also added one new item (item # 8; see Appendix) to the scale. Twenty items were presented to participants for an initial test of the psychometric properties.

Sexual Harassment Reporting Attitudes Scale
The initial SHRAS consisted of 20 items; however, two were dropped after the factor analysis as their meaning was ambiguous and thus did not fit well in any of the subscales that emerged; these items are indicated in Table 1. The final SHRAS consists of 18 items designed to measure attitudes about reporting workplace SH (see Appendix). The items reflect attitudes regarding the perceived utility, importance, and benefits of reporting (e.g., "Reporting workplace sexual harassment is an effective way of stopping the problem") as well as the perceived drawbacks (e.g., "People who report workplace sexual harassment usually end up getting into trouble for it"). Response options ranged from 1 (strongly disagree) to 5 (strongly agree). One additional item was included at the end of the scale to measure intention to report workplace SH ("If I felt that I was being sexually harassed at my place of work, I would report it to a supervisor or other authority figure"); this was used to establish concurrent validity.

Sexual harassment attitudes scale
A modified version of the 19-item SHAS (Mazer & Percival, 1989) was used to measure overall attitudes about workplace sexual harassment in order to measure discriminant validity. Each item within this instrument reflects the respondent's attitudes regarding the nature (e.g., "Sexual assault and sexual harassment are two very different things") and severity (e.g., "Most of what people call sexual harassment is just innocent flirtation between two people") of SH. Items in the original version of the SHAS were written such that a female victim was assumed (e.g., "An attractive woman [emphasis added] has to expect sexual advances in the workplace and should learn how to handle them"). However, for the purposes of the present study, and considering the importance of male SH victimization, the items were reworded such that they were gender-neutral (e.g., "An attractive person [emphasis added] has to expect sexual advances in the workplace and should learn how to handle them"). Response options ranged from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating more favorable attitudes toward SH. An exploration of the internal consistency of the scale with the item revisions revealed that it was reliable (Cronbach's ɑ = .90).

Procedure
Once the link to the online survey was available on Mechanical Turk, interested participants were able to click on it and be taken directly to the survey page. Participants were required to indicate that they gave consent on an informed consent form before they were allowed to proceed to the survey instruments. Upon completion of the battery, participants submitted their data to the researchers and then entered a unique code to ensure that they received payment.

Factor structure
All data analyses were performed using SPSS version 24 software. A factor analysis was conducted using a principal components extraction method with an oblique rotation, because we expected the different factors to correlate with each other. Using a Promax rotation, three factors emerged with eigenvalues over 1.00, confirmed by an examination of the scree plot (see Figure 1) that accounted for 54.38% of the variability in responses.
The three factors included items related to risks of reporting, prescriptions about one's moral duty to report, and the utility of reporting. All items loaded primarily on one factor except one item ("If someone is being sexually harassed in his or her place of work, then s/he should report it to a supervisor"), which loaded equally on both the Moral Duty subscale and the Utility subscale. We included this item in the Moral Duty subscale based on its face validity although it does reflect both of these underlying constructs. See Table 2 for the factor loadings of each item on each factor.
The three subscales (risks, moral duty, and utility) were analyzed separately (see Table 3). All three subscales were significantly correlated with each other (all rs ≥ .32, p < .001) and strongly correlated with the full scale (all rs > = .68, p < .001). *Item omitted from final scale. **The wording in this item was changed from "creates tension" to "leads to animosity" in the final scale per feedback received from the expert panel.

Reliability and validity
Results of our analyses revealed that the internal consistency of the SHRAS was high (Cronbach's ɑ = .87). We observed a strong and significant negative correlation between the SHRAS total score and the SHAS total score (r = -.56, p < .001), which suggests that the SHRAS possesses discriminant validity. We observed a significant, moderately sized positive correlation between the SHRAS total score and the intention to report item (r = .42, p < .001), suggesting that the scale possesses good concurrent validity. All three subscales significantly correlated with the behavioral intention item (r > = .28, p < .001), with the moral duty subscale most strongly correlated with the behavioral intention to report (r = .47, p < .001). In conjunction with the findings of our pilot study (expert review), we conclude that the initial psychometric properties of this scale are sound, although further research is warranted to further explore the qualities of the instrument.

Discussion
The results of the present study provide evidence that the SHRAS is psychometrically sound, with mostly good reliability, a reasonable initial factor structure, and satisfactory concurrent validity with an intention to report item. Although these initial findings are promising, additional research is needed to further establish the psychometric properties of this newly developed scale. Given the known challenges associated with SH reporting, the authors recommend that the instrument's test-retest reliability be determined in order to verify that responses remain stable over time.
Future research should also test the predictive validity of the scale for actual reporting behavior.
We also recommend that discriminant validity be re-evaluated by comparing the SHRAS total score with measures of different constructs, namely those related to reporting behaviors. Given the three distinct underlying factors, future research could also examine the differential prediction of behavior based on attitudes about the risks, moral duty, or utility of reporting sexual harassment. In this initial study, attitudes about one's moral duty to report was most highly correlated with one's intention to report, although this factor has not been frequently discussed as a motivation in Figure 1. Scree plot for exploratory factor analysis the literature (see Hay, 2005). Future research should explore which factors most strongly predict actual reporting behavior. Finally, confirmatory factor analyses should be conducted to ensure that the three factors that emerged in this initial study are consistent with the nature of the model we developed. If someone is being sexually harassed in his or her place of work, then s/he should report it to a supervisor.
.56* .58 Reporting workplace sexual harassment is an effective way of stopping the problem. .75 A person who reports workplace sexual harassment is just a tattletale. (reversed)

−.41
Reporting workplace sexual harassment creates new problems for everyone. (reversed) .66 People should not be afraid to report sexual harassment in their places of work. .69 Supervisors have better things to do with their time than deal with reports of sexual harassment. (reversed) .50 Workplace sexual harassment problems will persist, even if people report them. (reversed) .69 People who witness workplace sexual harassment, but are not harassed themselves, should report it. .65 Supervisors need to take reports of workplace sexual harassment very seriously. .80 A person who reports workplace sexual harassment should not be afraid of losing his or her job because of it. .72 In general, reporting workplace sexual harassment does no good. (reversed) .77 Reporting workplace sexual harassment only makes the problem worse. (reversed) .77 Reporting sexual harassment leads to animosity in the workplace. (reversed) .79 An employee has the right to report workplace sexual harassment to his or her supervisor. .79 All things considered, reporting workplace sexual harassment is a waste of time. (reversed) .67 People who report workplace sexual harassment risk being looked upon badly by their coworkers. (reversed) .75 People who report workplace sexual harassment usually end up getting into trouble for it. (reversed) .74 Note. Loadings < .4 are omitted. * denotes the subscale this item is included within. Note. All correlations significant at p < .001.

Practical implications and applications of the SHRAS
Although the SHRAS is designed for use by both scientists and practitioners, we feel that the scale will be very helpful among applied psychologists, human resource managers, professional trainers, organizational consultants, and other members of the professional community. Practitioners, especially those who develop, provide, or supervise the implementation of workplace training curricula focusing on sensitivity and diversity, could use the scale to measure trainees' attitudes toward SH reporting. The results could then be used to inform or tailor specific training focal areas and content to address unfavorable or undesirable attitudes. This is an important consideration because attitudinal change may have long-term implications for the nature and quality of the relationship between employers and employees (Kickul & Lester, 2001;Saari & Judge, 2004;Van Dyne & Pierce, 2004).
We also believe that practitioners who provide consulting services in the realm of organizational change will find this scale useful. Employee attitudes are often indicative of some aspect of organizational climate (Patterson, Warr, & West, 2004); therefore, this scale may be a useful diagnostic and measurement aid in organizations in which sexual harassment is present. Specifically, organizational tolerance for sexual harassment manifests itself in several ways (see Fitzgerald, Drasgow, Hulin, Gelfand, & Magley, 1997), and it is plausible that unfavorable attitudes about reporting may indicate passive tolerance for SH.

Limitations and recommendations
Along with the generalizability afforded by a large, online sample of over 400 working adults, the diverse international sample also creates more variability in the present data. Our sample was comprised of individuals from a variety of cultural backgrounds, and culture has frequently been shown to affect the interpretation of and response to workplace SH (see Fasting, Brackenridge, & Sundgot-Borgen, 2003;Kennedy & Gorzalka, 2002;Shupe, Cortina, Ramos, Fitzgerald, & Salisbury, 2002). Second, the age range of the sample was very wide; thus, age-related or generational effects may have moderated response patterns as age has been shown to affect perceptions of SH as well as appropriateness of responses to such (O'Connor, Gutek, Stockdale, Geer, & Melançon, 2004). Finally, the nature of individual respondents' employment (i.e., industry, work settings) was not captured in our survey. This is an important consideration because reporting SH in certain industries and occupations has historically been frowned upon, regardless of individual employee's perceptions (Ilies, Hauserman, Schwochau, & Stibal, 2003). We recommend that all of these factors-cultural background, participant age, and nature of employment-be examined in future psychometric evaluation studies of this instrument. More importantly, we look forward to the use of this new instrument in research examining the role of attitudes in predicting sexual harassment reporting in the workplace.

Funding
The authors received no direct funding for this research.