رفتار تبعیض جنسیتی در محل کار: بررسی رابطه بین آزار جنسی و تبعیض جنسی

Abstract This paper examines survey-based reports of sexual harassment and sex discrimination in order to identify the stylized facts about the nature of the relationship between them. In particular, we are interested in assessing whether these concepts measure similar forms of gender-biased behavior and whether they have the same effect on workers’ job satisfaction and intentions to leave their jobs. Our results provide little support for the notion that survey-based measures of sexual harassment and sex discrimination capture the same underlying behavior. Respondents do appear to differentiate between incidents of sexual harassment and incidents of sex discrimination in the workplace. There are gender differences in the consequences, however. Both sex discrimination and sexual harassment are associated with a higher degree of job dissatisfaction. However, women’s intended job changes appear to be more sensitive to experiencing sex discrimination, while men’s are more sensitive to experiencing sexual harassment. Although exploratory, when taken together these results give us hope that in the future sufficiently detailed surveys could provide a useful foundation for quantifying the link between sexual harassment and sex discrimination. They also suggest that the best prospect for developing—and then testing—a conceptual framework of gender bias lies in adopting a multi-disciplinary approach incorporating the insights of disciplines such as sociology, psychology, and economics.

. Introduction Workplaces are rarely gender-neutral. Though gender differences in the terms and conditions of one’s employment are almost never codified in firms’ personnel policies or in employment law, women nonetheless frequently find that they are paid less, are promoted less often, and receive less training than their male colleagues (Blau, 1998 and Blau et al., 1998). Reports of sexual harassment are also common with many working women experiencing sexual harassment at some point in their careers (for example, Fitzgerald and Ormerod, 1993, Schneider et al., 1997 and Welsh, 1999). The complex—and often ill-defined—nature of workplace sex discrimination and sexual harassment poses significant challenges for researchers wishing to assess the extent of gender bias in employment relationships. Strong disciplinary roots have shaped the ways in which previous researchers have approached the issue of gender bias. The idiosyncrasies in conceptual frameworks, definitions, and research methodologies inherent in various academic disciplines have produced a dizzying array of results that, while individually enlightening, can be difficult to piece together to produce a comprehensive view of employment-related gender bias more generally. Economists, for example, typically define sex discrimination to be that portion of the gender gap in aggregate employment outcomes that is not attributable to productivity differentials and have largely been concerned with understanding how these disparities can best be measured (see Altonji & Blank, 1999). Until recently, however, economists have been almost silent on the issue of sexual harassment. A universally accepted definition of sexual harassment has not yet emerged for example (see Foulis & McCabe, 1997), though psychologists have made a great deal of progress in quantifying women’s experiences of sexual harassment (see Fitzgerald et al., 1997 and Schneider et al., 1997). Still, with the exception of a few evolutionary psychologists, psychologists have not been engaged in analyzing the psychological origins of sex discrimination as economists understand it.3 These somewhat artificial disciplinary boundaries have not been helpful in enriching our understanding of the causes and consequences of employment-related gender bias. More progress is almost certain to be made by developing a conceptual framework that does not focus exclusively on either sexual harassment or sex discrimination in isolation, but rather which explicitly views these as alternative forms of gender bias and considers the links between them. For economists, the successful strategy is likely to involve incorporating the insights from disciplines such as psychology and sociology into economic models of labor market behavior. There are a number of ways in which we might proceed. Sociologists, for example, often have an understanding of sexual harassment that is rooted in the power structure in society more generally (Skaine, 1996). They argue, for example, it is the social power structure that frequently puts male employers in positions of authority over female employees which underlies sexual harassment. Of course this is the same social power structure that is implicit in many “taste-based” theories of sex discrimination in economics (see for example, Becker, 1957). Similarly, legal scholars have spent the past two decades developing and then refining the argument that sexual harassment is sex discrimination (see MacKinnon, 1979, Siegel, 2004 and Skaine, 1996). In particular, MacKinnon (1979) argues: “Sexual harassment is discrimination ‘based on sex’ within the social meaning of sex, as the concept is socially incarnated in sex roles. Pervasive and ‘accepted’ as they are, these rigid roles have no place in the allocation of social and economic resources” (p. 178). Ideally, any conceptual model of employment-related gender bias would be multi-disciplinary, well grounded in the stylized facts, and take account of the complexities of workers’ experiences in the workplace. Sexual harassment and sex discrimination are surely related, but they are a long way from being the same thing despite the arguments of legal scholars.4 Yet we have very little empirical evidence on the nature of the relationship between them. There is evidence that the negative consequences of unwanted sexual behavior at work can be greater for women who believe themselves to be sexually harassed (Antecol & Cobb-Clark, 2006), though psychologists often conclude that labeling unwanted sexual experiences as sexual harassment is unrelated to many subsequent employment outcomes (Magley et al., 1999 and Munson et al., 2001). Moreover, previous research concludes that workers’ perceptions of harassment and discrimination are closely related to their labor market behavior. Women experiencing sex discrimination and older workers experiencing age discrimination are more likely to separate from their employers for example (Johnson and Neumark, 1997 and Neumark and McLennan, 1995), while women’s labor supply behavior appears to be particularly sensitive to sexual harassment (Goldsmith, Sedo, Darity, & Hamilton, 2004). Our objective is to add to this very limited empirical literature on the relationship between two forms of gender-biased behavior—sexual harassment and sex discrimination—by using data drawn from the 2002 General Social Survey (GSS). While far from perfect, these are the only data of which we are aware that separately identify incidents of sexual harassment from incidents of sex discrimination. Consequently, they allow us to begin to understand the extent to which these forms of gender bias might be related and to consider their respective consequences in terms of employment outcomes (i.e., job satisfaction and intentions to change jobs). Specifically, we address the following questions. Do survey-based measures of sexual harassment and sex discrimination capture separate forms of gender bias or are they simply reflections of the same underlying behavior? Do they have similar consequences for workers’ job satisfaction and intentions to remain in their current employment? Our results provide little support for the notion that survey-based measures of sexual harassment and sex discrimination capture the same underlying behavior. Respondents do appear to differentiate between incidents of sexual harassment and incidents of sex discrimination in the workplace. There are gender differences in the consequences, however. Both sex discrimination and sexual harassment are associated with a higher degree of job dissatisfaction. However, women’s intended job changes appear to be more sensitive to experiencing sex discrimination, while men’s are more sensitive to experiencing sexual harassment. Although exploratory, when taken together these results give us hope that in the future sufficiently detailed surveys could provide a useful foundation for quantifying the link between sexual harassment and sex discrimination. They also suggest that the best prospect for developing—and then testing—a conceptual framework of gender bias lies in adopting a multi-disciplinary approach incorporating the insights of disciplines such as sociology, psychology, and economics. In what follows, we discuss some of the issues involved in using surveys to measure the incidence of sexual harassment and sex discrimination. Details of the GSS data used in this analysis and the incidence of sexual harassment and sex discrimination as well as the link between them are provided in Section 3. Following that evidence on the consequences of these forms of gender-biased behavior for job satisfaction and intentions to leave ones current employment are presented. Finally, our conclusions and suggestions for future research are discussed in Section 5

Conclusions and suggested directions for future research Despite the decades that have passed since Equal Opportunity legislation was first passed, gender bias persists in many workplaces. The complex, ill-defined nature of sex discrimination and sexual harassment, however, have made it difficult to develop a fuller understanding of the ways in which gender bias might affect workers’ experiences, relationships and opportunities while on the job. This paper adds to the very limited empirical evidence on this issue by using General Social Survey (GSS) data to explore the links between workers’ perceptions of two gender-biased behaviors—sexual harassment and sex discrimination. Our objective has been to begin to understand the extent to which these forms of gender bias might be related and to consider the type of future data collections—and ultimately conceptual frameworks—that are likely to be the most useful in capturing the complexities of gender bias at work. While far from perfect, the GSS provides the only data of which we are aware that separately identify incidents of sexual harassment and incidents of sex discrimination. Still, our results give us reason to be optimistic about the potential for using surveys to enhance our understanding of the ways in which gender bias intrudes on men’s and women’s working lives. When asked directly, men and women in the GSS do appear to discriminate between incidents of sex discrimination and incidents of sexual harassment. This is important because these events appear to have distinct effects on men’s and women’s satisfaction with and intentions to remain in their current jobs. In particular, both are linked to a substantially higher degree of job dissatisfaction, especially amongst men. Women experiencing sex discrimination are somewhat more likely to intend to look for new work, though men’s future job changes are much more closely linked to incidents of sexual harassment. To the extent that reporting both sex discrimination and sexual harassment provides information about the intensity of gender bias a worker has experienced, it would seem that it is incidents—rather than the intensity—of gender bias that is important for understanding job dissatisfaction and the intention to quit. At the same time, there is a great deal that the GSS cannot tell us. Specifically, the GSS restricts us to analyzing what psychologists refer to as a “self-definition” of sexual harassment (see Fitzgerald et al., 1997). While it is certainly important to understand what drives workers’ beliefs about whether they have been harassed given that the legal system relies on a reasonable victim standard to make determinations in sexual harassment cases (Fitzgerald and Shullman, 1993 and Prior et al., 1997), this is by no means the only definition of sexual harassment that is relevant. Psychologists have made compelling arguments that sexual harassment as a legal concept should be distinguished from the psychological experience (see Fitzgerald et al., 1997)—the consequences of which do not necessarily depend on workers’ being prepared to label their experiences as sexual harassment (see Magley et al., 1999, Marin and Guadagno, 1999 and Munson et al., 2001). Thus, it is imperative to collect information about the totality of workers’ experiences of unwanted sex-related behavior at work.14 Moreover, sociological notions of sexual harassment often rest on the abuse of power or economic intimidation (Skaine, 1996). Consequently, evaluation of the sociology of sexual harassment requires detailed data on the social power structure within organizations. Developing richer surveys that allow for alternative notions of sexual harassment and sex discrimination to be identified is likely to be an important next step in deepening our understanding of gender bias in the workplace. Moreover, the GSS data are not sufficiently detailed to provide us with sensible exclusion restrictions or instruments that would allow us to account for the potential endogeneity of reported sex discrimination and sexual harassment. Thus, in this paper we have been left to speculate about the role that omitted variable bias might play in estimating the causal effect of gender bias on employment outcomes such as job dissatisfaction and the intention to quit ones job. Economists’ strong preference for the identification of causal relationships rather than associations is likely to continue to limit their engagement in the study of sexual harassment unless survey instruments are developed that would collect data suitable for econometric modeling. Panel data would be particularly useful in isolating the effects of individual heterogeneity and in understanding the consequences of gender bias on subsequent employment outcomes. Moreover, some economists have recently begun advocating the use of anchoring vignettes as a method of correcting for the possibility that different groups may systematically use different response scales when answering questions about subjective outcomes (Kapteyn et al., 2007, van Soest et al., 2006 and van Soest et al., 2007). By providing anchors (or benchmarks), this approach is likely to prove very useful in furthering our understanding of the ways in which both the scope and severity of employment-related gender bias vary across different groups of workers. There are certainly many challenges ahead of us in our efforts to quantify gender bias in its various guises. We believe, however, that such efforts will be important in paving the way for the development of a multi-disciplinary, conceptual framework that explicitly considers sexual harassment and sex discrimination as alternative forms of gender bias and considers the links between them. A modeling strategy that incorporates the insights from psychology and sociology into economic models of labor market behavior would seem to be particularly promising. We are of course not the first economists to call for a multi-disciplinary approach to understanding complex social behavior. A number of others have made similar arguments with respect to the importance of evolutionary psychology in understanding market behavior (Cosmides & Tooby, 1994) and social psychology in understanding implicit (or unintentional) discrimination (Bertrand, Chugh, & Mullainathan, 2005). However, the complex, ill-defined nature of gender bias makes it difficult to imagine proceeding in any other way.