thesis about gender bias

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

Research: How Bias Against Women Persists in Female-Dominated Workplaces

• Amber L. Stephenson,
• Leanne M. Dzubinski

A look inside the ongoing barriers women face in law, health care, faith-based nonprofits, and higher education.

New research examines gender bias within four industries with more female than male workers — law, higher education, faith-based nonprofits, and health care. Having balanced or even greater numbers of women in an organization is not, by itself, changing women’s experiences of bias. Bias is built into the system and continues to operate even when more women than men are present. Leaders can use these findings to create gender-equitable practices and environments which reduce bias. First, replace competition with cooperation. Second, measure success by goals, not by time spent in the office or online. Third, implement equitable reward structures, and provide remote and flexible work with autonomy. Finally, increase transparency in decision making.

It’s been thought that once industries achieve gender balance, bias will decrease and gender gaps will close. Sometimes called the “ add women and stir ” approach, people tend to think that having more women present is all that’s needed to promote change. But simply adding women into a workplace does not change the organizational structures and systems that benefit men more than women . Our new research (to be published in a forthcoming issue of Personnel Review ) shows gender bias is still prevalent in gender-balanced and female-dominated industries.

• Amy Diehl , PhD is chief information officer at Wilson College and a gender equity researcher and speaker. She is coauthor of Glass Walls: Shattering the Six Gender Bias Barriers Still Holding Women Back at Work (Rowman & Littlefield). Find her on LinkedIn at Amy-Diehl , X/Twitter @amydiehl , and visit her website at amy-diehl.com .
• AS Amber L. Stephenson , PhD is an associate professor of management and director of healthcare management programs in the David D. Reh School of Business at Clarkson University. Her research focuses on the healthcare workforce, how professional identity influences attitudes and behaviors, and how women leaders experience gender bias.
• LD Leanne M. Dzubinski , PhD is professor of leadership and director of the Beeson International Center at Asbury Seminary, and a prominent researcher on women in leadership. She is coauthor of Glass Walls: Shattering the Six Gender Bias Barriers Still Holding Women Back at Work (Rowman & Littlefield).

Partner Center

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Front Psychol

Gender inequalities in the workplace: the effects of organizational structures, processes, practices, and decision makers’ sexism

Gender inequality in organizations is a complex phenomenon that can be seen in organizational structures, processes, and practices. For women, some of the most harmful gender inequalities are enacted within human resources (HRs) practices. This is because HR practices (i.e., policies, decision-making, and their enactment) affect the hiring, training, pay, and promotion of women. We propose a model of gender discrimination in HR that emphasizes the reciprocal nature of gender inequalities within organizations. We suggest that gender discrimination in HR-related decision-making and in the enactment of HR practices stems from gender inequalities in broader organizational structures, processes, and practices. This includes leadership, structure, strategy, culture, organizational climate, as well as HR policies. In addition, organizational decision makers’ levels of sexism can affect their likelihood of making gender biased HR-related decisions and/or behaving in a sexist manner while enacting HR practices. Importantly, institutional discrimination in organizational structures, processes, and practices play a pre-eminent role because not only do they affect HR practices, they also provide a socializing context for organizational decision makers’ levels of hostile and benevolent sexism. Although we portray gender inequality as a self-reinforcing system that can perpetuate discrimination, important levers for reducing discrimination are identified.

Introduction

The workplace has sometimes been referred to as an inhospitable place for women due to the multiple forms of gender inequalities present (e.g., Abrams, 1991 ). Some examples of how workplace discrimination negatively affects women’s earnings and opportunities are the gender wage gap (e.g., Peterson and Morgan, 1995 ), the dearth of women in leadership ( Eagly and Carli, 2007 ), and the longer time required for women (vs. men) to advance in their careers ( Blau and DeVaro, 2007 ). In other words, workplace discrimination contributes to women’s lower socio-economic status. Importantly, such discrimination against women largely can be attributed to human resources (HR) policies and HR-related decision-making. Furthermore, when employees interact with organizational decision makers during HR practices, or when they are told the outcomes of HR-related decisions, they may experience personal discrimination in the form of sexist comments. Both the objective disadvantages of lower pay, status, and opportunities at work, and the subjective experiences of being stigmatized, affect women’s psychological and physical stress, mental and physical health ( Goldenhar et al., 1998 ; Adler et al., 2000 ; Schmader et al., 2008 ; Borrel et al., 2010 ),job satisfaction and organizational commitment ( Hicks-Clarke and Iles, 2000 ), and ultimately, their performance ( Cohen-Charash and Spector, 2001 ).

Within this paper, we delineate the nature of discrimination within HR policies, decisions, and their enactment, as well as explore the causes of such discrimination in the workplace. Our model is shown in Figure ​ Figure1 1 . In the Section “Discrimination in HR Related Practices: HR Policy, Decisions, and their Enactment,” we explain the distinction between HR policy, HR-related decision-making, and HR enactment and their relations to each other. Gender inequalities in HR policy are a form of institutional discrimination. We review evidence of institutional discrimination against women within HR policies set out to determine employee selection, performance evaluations, and promotions. In contrast, discrimination in HR-related decisions and their enactment can result from organizational decision makers’ biased responses: it is a form of personal discrimination. Finally, we provide evidence of personal discrimination against women by organizational decision makers in HR-related decision-making and in the enactment of HR policies.

A model of the root causes of gender discrimination in HR policies, decision-making, and enactment .

In the Section “The Effect of Organizational Structures, Processes, and Practices on HR Practices,” we focus on the link between institutional discrimination in organizational structures, processes, and practices that can lead to personal discrimination in HR practices (see Figure ​ Figure1 1 ). Inspired by the work of Gelfand et al. (2007) , we propose that organizational structures, processes, and practices (i.e., leadership, structure, strategy, culture, climate, and HR policy) are interrelated and may contribute to discrimination. Accordingly, gender inequalities in each element can affect the others, creating a self-reinforcing system that can perpetuate institutional discrimination throughout the organization and that can lead to discrimination in HR policies, decision-making, and enactment. We also propose that these relations between gender inequalities in the organizational structures, processes, and practices and discrimination in HR practices can be bidirectional (see Figure ​ Figure1 1 ). Thus, we also review how HR practices can contribute to gender inequalities in organizational structures, processes, and practices.

In the Section “The Effect of Hostile and Benevolent Sexism on How Organizational Decision Makers’ Conduct HR Practices,” we delineate the link between organizational decision makers’ levels of sexism and their likelihood of making gender-biased HR-related decisions and/or behaving in a sexist manner when enacting HR policies (e.g., engaging in gender harassment). We focus on two forms of sexist attitudes: hostile and benevolent sexism ( Glick and Fiske, 1996 ). Hostile sexism involves antipathy toward, and negative stereotypes about, agentic women. In contrast, benevolent sexism involves positive but paternalistic views of women as highly communal. Whereas previous research on workplace discrimination has focused on forms of sexism that are hostile in nature, we extend this work by explaining how benevolent sexism, which is more subtle, can also contribute in meaningful yet distinct ways to gender discrimination in HR practices.

In the Section “The Effect of Organizational Structures, Processes, and Practices on Organizational Decision Makers’ Levels of Hostile and Benevolent Sexism,” we describe how institutional discrimination in organizational structures, processes, and practices play a critical role in our model because not only do they affect HR-related decisions and the enactment of HR policies, they also provide a socializing context for organizational decision makers’ levels of hostile and benevolent sexism. In other words, where more institutional discrimination is present, we can expect higher levels of sexism—a third link in our model—which leads to gender bias in HR practices.

In the Section “How to Reduce Gender Discrimination in Organizations,” we discuss how organizations can reduce gender discrimination. We suggest that, to reduce discrimination, organizations should focus on: HR practices, other closely related organizational structures, processes, and practices, and the reduction of organizational decision makers’ level of sexism. Organizations should take such a multifaceted approach because, consistent with our model, gender discrimination is a result of a complex interplay between these factors. Therefore, a focus on only one factor may not be as effective if all the other elements in the model continue to promote gender inequality.

The model we propose for understanding gender inequalities at work is, of course, limited and not intended to be exhaustive. First, we only focus on women’s experience of discrimination. Although men also face discrimination, the focus of this paper is on women because they are more often targets ( Branscombe, 1998 ; Schmitt et al., 2002 ; McLaughlin et al., 2012 ) and discrimination is more psychologically damaging for women than for men ( Barling et al., 1996 ; Schmitt et al., 2002 ). Furthermore, we draw on research from Western, individualistic countries conducted between the mid-1980s to the mid-2010s that might not generalize to other countries or time frames. In addition, this model derives from research that has been conducted primarily in sectors dominated by men. This is because gender discrimination ( Mansfield et al., 1991 ; Welle and Heilman, 2005 ) and harassment ( Mansfield et al., 1991 ; Berdhal, 2007 ) against women occur more in environments dominated by men. Now that we have outlined the sections of the paper and our model, we now turn to delineating how gender discrimination in the workplace can be largely attributed to HR practices.

Discrimination in HR Related Practices: HR Policy, Decisions, and their Enactment

In this section, we explore the nature of gender discrimination in HR practices, which involves HR policies, HR-related decision-making, and their enactment by organizational decision makers. HR is a system of organizational practices aimed at managing employees and ensuring that they are accomplishing organizational goals ( Wright et al., 1994 ). HR functions include: selection, performance evaluation, leadership succession, and training. Depending on the size and history of the organization, HR systems can range from those that are well structured and supported by an entire department, led by HR specialists, to haphazard sets of policies and procedures enacted by managers and supervisors without formal training. HR practices are critically important because they determine the access employees have to valued reward and outcomes within an organization, and can also influence their treatment within an organization ( Levitin et al., 1971 ).

Human resource practices can be broken down into formal HR policy, HR-related decision-making, and the enactment of HR policies and decisions. HR policy codifies practices for personnel functions, performance evaluations, employee relations, and resource planning ( Wright et al., 1994 ). HR-related decision-making occurs when organizational decision makers (i.e., managers, supervisors, or HR personnel) employ HR policy to determine how it will be applied to a particular situation and individual. The enactment of HR involves the personal interactions between organizational decision makers and job candidates or employees when HR policies are applied. Whereas HR policy can reflect institutional discrimination, HR-related decision-making and enactment can reflect personal discrimination by organizational decision makers.

Institutional Discrimination in HR Policy

Human resource policies that are inherently biased against a group of people, regardless of their job-related knowledge, skills, abilities, and performance can be termed institutional discrimination. Institutional discrimination against women can occur in each type of HR policy from the recruitment and selection of an individual into an organization, through his/her role assignments, training, pay, performance evaluations, promotion, and termination. For instance, if women are under-represented in a particular educational program or a particular job type and those credentials or previous job experience are required to be considered for selection, women are being systematically, albeit perhaps not intentionally, discriminated against. In another example, there is gender discrimination if a test is used in the selection battery for which greater gender differences emerge, than those that emerge for job performance ratings ( Hough et al., 2001 ). Thus, institutional discrimination can be present within various aspects of HR selection policy, and can negatively affect women’s work outcomes.

Institutional discrimination against women also occurs in performance evaluations that are used to determine organizational rewards (e.g., compensation), opportunities (e.g., promotion, role assignments), and punishments (e.g., termination). Gender discrimination can be formalized into HR policy if criteria used by organizational decision makers to evaluate job performance systematically favor men over women. For instance, “face time” is a key performance metric that rewards employees who are at the office more than those who are not. Given that women are still the primary caregivers ( Acker, 1990 ; Fuegen et al., 2004 ), women use flexible work arrangements more often than men and, consequently, face career penalties because they score lower on face time ( Glass, 2004 ). Thus, biased criteria in performance evaluation policies can contribute to gender discrimination.

Human resource policies surrounding promotions and opportunities for advancement are another area of concern. In organizations with more formal job ladders that are used to dictate and constrain workers’ promotion opportunities, women are less likely to advance ( Perry et al., 1994 ). This occurs because job ladders tend to be divided by gender, and as such, gender job segregation that is seen at entry-level positions will be strengthened as employees move up their specific ladder with no opportunity to cross into other lines of advancement. Thus, women will lack particular job experiences that are not available within their specific job ladders, making them unqualified for advancement ( De Pater et al., 2010 ).

In sum, institutional discrimination can be present within HR policies set out to determine employee selection, performance evaluations, and promotions. These policies can have significant effects on women’s careers. However, HR policy can only be used to guide HR-related decision-making. In reality, it is organizational decision-makers, that is, managers, supervisors, HR personnel who, guided by policy, must evaluate job candidates or employees and decide how policy will be applied to individuals.

Personal Discrimination in HR-Related Decision-Making

The practice of HR-related decision-making involves social cognition in which others’ competence, potential, and deservingness are assessed by organizational decision makers. Thus, like all forms of social cognition, HR-related decision-making is open to personal biases. HR-related decisions are critically important because they determine women’s pay and opportunities at work (e.g., promotions, training opportunities). Personal discrimination against women by organizational decision makers can occur in each stage of HR-related decision-making regarding recruitment and selection, role assignments, training opportunities, pay, performance evaluation, promotion, and termination.

Studies with varying methodologies show that women face personal discrimination when going through the selection process (e.g., Goldberg, 1968 ; Rosen and Jerdee, 1974 ). Meta-analyses reveal that, when being considered for male-typed (i.e., male dominated, believed-to-be-for-men) jobs, female candidates are evaluated more negatively and recommended for employment less often by study participants, compared with matched male candidates (e.g., Hunter et al., 1982 ; Tosi and Einbender, 1985 ; Olian et al., 1988 ; Davison and Burke, 2000 ). For example, in audit studies, which involve sending ostensibly real applications for job openings while varying the gender of the applicant, female applicants are less likely to be interviewed or called back, compared with male applicants (e.g., McIntyre et al., 1980 ; Firth, 1982 ). In a recent study, male and female biology, chemistry, and physics professors rated an undergraduate science student for a laboratory manager position ( Moss-Racusin et al., 2012 ). The male applicant was rated as significantly more competent and hireable, offered a higher starting salary (about $4000), and offered more career mentoring than the female applicant was. In summary, women face a distinct disadvantage when being considered for male-typed jobs.

There is ample evidence that women experience biased performance evaluations on male-typed tasks. A meta-analysis of experimental studies reveals that women in leadership positions receive lower performance evaluations than matched men; this is amplified when women act in a stereotypically masculine, that is, agentic fashion ( Eagly et al., 1992 ). Further, in masculine domains, women are held to a higher standard of performance than men are. For example, in a study of military cadets, men and women gave their peers lower ratings if they were women, despite having objectively equal qualifications to men ( Boldry et al., 2001 ). Finally, women are evaluated more poorly in situations that involve complex problem solving; in these situations, people are skeptical regarding women’s expertise and discredit expert women’s opinions but give expert men the benefit of the doubt ( Thomas-Hunt and Phillips, 2004 ).

Sometimes particular types of women are more likely to be discriminated against in selection and performance evaluation decisions. Specifically, agentic women, that is, those who behave in an assertive, task-oriented fashion, are rated as less likeable and less hireable than comparable agentic male applicants ( Heilman and Okimoto, 2007 ; Rudman and Phelan, 2008 ; Rudman et al., 2012 ). In addition, there is evidence of discrimination against pregnant women when they apply for jobs ( Hebl et al., 2007 ; Morgan et al., 2013 ). Further, women who are mothers are recommended for promotion less than women who are not mothers or men with or without children ( Heilman and Okimoto, 2008 ). Why might people discriminate specifically against agentic women and pregnant women or mothers, who are seemingly very different? The stereotype content model, accounts for how agentic women, who are perceived to be high in competence and low in warmth, will be discriminated against because of feelings of competition; whereas, pregnant women and mothers, who are seen as low in competence, but high in warmth, will be discriminated against because of a perceived lack of deservingness ( Fiske et al., 1999 , 2002 ; Cuddy et al., 2004 ). Taken together, research has uncovered that different forms of bias toward specific subtypes of women have the same overall effect—bias in selection and performance evaluation decisions.

Women are also likely to receive fewer opportunities at work, compared with men, resulting in their under-representation at higher levels of management and leadership within organizations ( Martell et al., 1996 ; Eagly and Carli, 2007 ). Managers give women fewer challenging roles and fewer training opportunities, compared with men ( King et al., 2012 ; Glick, 2013 ). For instance, female managers ( Lyness and Thompson, 1997 ) and midlevel workers ( De Pater et al., 2010 ) have less access to high-level responsibilities and challenges that are precursors to promotion. Further, men are more likely to be given key leadership assignments in male-dominated fields and in female-dominated fields (e.g., Maume, 1999 ; De Pater et al., 2010 ). This is detrimental given that challenging roles, especially developmental ones, help employees gain important skills needed to excel in their careers ( Spreitzer et al., 1997 ).

Furthermore, managers rate women as having less promotion potential than men ( Roth et al., 2012 ). Given the same level of qualifications, managers are less likely to grant promotions to women, compared with men ( Lazear and Rosen, 1990 ). Thus, men have a faster ascent in organizational hierarchies than women ( Cox and Harquail, 1991 ; Stroh et al., 1992 ; Blau and DeVaro, 2007 ). Even minimal amounts of gender discrimination in promotion decisions for a particular job or level can have large, cumulative effects given the pyramid structure of most hierarchical organizations ( Martell et al., 1996 ; Baxter and Wright, 2000 ). Therefore, discrimination by organizational decision makers results in the under-promotion of women.

Finally, women are underpaid, compared with men. In a comprehensive US study using data from 1983 to 2000, after controlling for human capital factors that could affect wages (e.g., education level, work experience), the researchers found that women were paid 22% less than men ( U.S. Government Accountability Office, 2003 ). Further, within any given occupation, men typically have higher wages than women; this “within-occupation” wage gap is especially prominent in more highly paid occupations ( U.S. Census Bureau, 2007 ). In a study of over 2000 managers, women were compensated less than men were, even after controlling for a number of human capital factors ( Ostroff and Atwater, 2003 ). Experimental work suggests that personal biases by organizational decision makers contribute to the gender wage gap. When participants are asked to determine starting salaries for matched candidates that differ by gender, they pay men more (e.g., Steinpreis et al., 1999 ; Moss-Racusin et al., 2012 ). Such biases are consequential because starting salaries determine life-time earnings ( Gerhart and Rynes, 1991 ). In experimental studies, when participants evaluate a man vs. a woman who is matched on job performance, they choose to compensate men more ( Marini, 1989 ; Durden and Gaynor, 1998 ; Lips, 2003 ). Therefore, discrimination in HR-related decision-making by organizational decision makers can contribute to women being paid less than men are.

Taken together, we have shown that there is discrimination against women in decision-making related to HR. These biases from organizational decision makers can occur in each stage of HR-related decision-making and these biased HR decisions have been shown to negatively affect women’s pay and opportunities at work. In the next section, we review how biased HR practices are enacted, which can involve gender harassment.

Personal Discrimination in HR Enactment

By HR enactment, we refer to those situations where current or prospective employees go through HR processes or when they receive news of their outcomes from organizational decision makers regarding HR-related issues. Personal gender discrimination can occur when employees are given sexist messages, by organizational decision makers, related to HR enactment. More specifically, this type of personal gender discrimination is termed gender harassment, and consists of a range of verbal and non-verbal behaviors that convey sexist, insulting, or hostile attitudes about women ( Fitzgerald et al., 1995a , b ). Gender harassment is the most common form of sex-based discrimination ( Fitzgerald et al., 1988 ; Schneider et al., 1997 ). For example, across the military in the United States, 52% of the 9,725 women surveyed reported that they had experienced gender harassment in the last year ( Leskinen et al., 2011 , Study 1). In a random sample of attorneys from a large federal judicial circuit, 32% of the 1,425 women attorneys surveyed had experienced gender harassment in the last 5 years ( Leskinen et al., 2011 , Study 2). When examining women’s experiences of gender harassment, 60% of instances were perpetrated by their supervisor/manager or a person in a leadership role (cf. Crocker and Kalemba, 1999 ; McDonald et al., 2008 ). Thus, personal discrimination in the form of gender harassment is a common behavior; however, is it one that organizational decision makers engage in when enacting HR processes and outcomes?

Although it might seem implausible that organizational decision makers would convey sexist sentiments to women when giving them the news of HR-related decisions, there have been high-profile examples from discrimination lawsuits where this has happened. For example, in a class action lawsuit against Walmart, female workers claimed they were receiving fewer promotions than men despite superior qualifications and records of service. In that case, the district manager was accused of confiding to some of the women who were overlooked for promotions that they were passed over because he was not in favor of women being in upper management positions ( Wal-Mart Stores, Inc. v. Dukes, 2004/2011 ). In addition, audit studies, wherein matched men and women apply to real jobs, have revealed that alongside discrimination ( McIntyre et al., 1980 ; Firth, 1982 ; Moss-Racusin et al., 2012 ), women experience verbal gender harassment when applying for sex atypical jobs, such as sexist comments as well as skeptical or discouraging responses from hiring staff ( Neumark, 1996 ). Finally, gender harassment toward women when HR policies are enacted can also take the form of offensive comments and denying women promotions due to pregnancy or the chance of pregnancy. For example, in Moore v. Alabama , an employee was 8 months pregnant and the woman’s supervisor allegedly looked at her belly and said “I was going to make you head of the office, but look at you now” ( Moore v. Alabama State University, 1996 , p. 431; Williams, 2003 ). Thus, organizational decision makers will at times convey sexist sentiments to women when giving them the news of HR-related decisions.

Interestingly, whereas discrimination in HR policy and in HR-related decision-making is extremely difficult to detect ( Crosby et al., 1986 ; Major, 1994 ), gender harassment in HR enactment provides direct cues to recipients that discrimination is occurring. In other words, although women’s lives are negatively affected in concrete ways by discrimination in HR policy and decisions (e.g., not receiving a job, being underpaid), they may not perceive their negative outcomes as due to gender discrimination. Indeed, there is a multitude of evidence that women and other stigmatized group members are loath to make attributions to discrimination ( Crosby, 1984 ; Vorauer and Kumhyr, 2001 ; Stangor et al., 2003 ) and instead are likely to make internal attributions for negative evaluations unless they are certain the evaluator is biased against their group ( Ruggiero and Taylor, 1995 ; Major et al., 2003 ). However, when organizational decision makers engage in gender harassment during HR enactment women should be more likely to interpret HR policy and HR-related decisions as discriminatory.

Now that we have specified the nature of institutional gender discrimination in HR policy and personal discrimination in HR-related decision-making and in HR enactment, we turn to the issue of understanding the causes of such discrimination: gender discrimination in organizational structures, processes, and practices, and personal biases of organizational decision makers.

The Effect of Organizational Structures, Processes, and Practices on HR Practices

The first contextual factor within which gender inequalities can be institutionalized is leadership. Leadership is a process wherein an individual (e.g., CEOs, managers) influences others in an effort to reach organizational goals ( Chemers, 1997 ; House and Aditya, 1997 ). Leaders determine and communicate what the organization’s priorities are to all members of the organization. Leaders are important as they affect the other organizational structures, processes, and practices. Specifically, leaders set culture, set policy, set strategy, and are role models for socialization. We suggest that one important way institutional gender inequality in leadership exists is when women are under-represented, compared with men—particularly when women are well-represented at lower levels within an organization.

An underrepresentation of women in leadership can be perpetuated easily because the gender of organizational leaders affects the degree to which there is gender discrimination, gender supportive policies, and a gender diversity supportive climate within an organization ( Ostroff et al., 2012 ). Organizational members are likely to perceive that the climate for women is positive when women hold key positions in the organization ( Konrad et al., 2010 ). Specifically, the presence of women in key positions acts as a vivid symbol indicating that the organization supports gender diversity. Consistent with this, industries that have fewer female high status managers have a greater gender wage gap ( Cohen and Huffman, 2007 ). Further, women who work with a male supervisor perceive less organizational support, compared with those who work with a female supervisor ( Konrad et al., 2010 ). In addition, women who work in departments that are headed by a man report experiencing more gender discrimination, compared with their counterparts in departments headed by women ( Konrad et al., 2010 ). Some of these effects may be mediated by a similar-to-me bias ( Tsui and O’Reilly, 1989 ), where leaders set up systems that reward and promote individuals like themselves, which can lead to discrimination toward women when leaders are predominantly male ( Davison and Burke, 2000 ; Roth et al., 2012 ). Thus, gender inequalities in leadership affect women’s experiences in the workplace and their likelihood of facing discrimination.

The second contextual factor to consider is organizational structure. The formal structure of an organization is how an organization arranges itself and it consists of employee hierarchies, departments, etc. ( Grant, 2010 ). An example of institutional discrimination in the formal structure of an organization are job ladders, which are typically segregated by gender ( Perry et al., 1994 ). Such gender-segregated job ladders typically exist within different departments of the organization. Women belonging to gender-segregated networks within organizations ( Brass, 1985 ) have less access to information about jobs, less status, and less upward mobility within the organization ( Ragins and Sundstrom, 1989 ; McDonald et al., 2009 ). This is likely because in gender-segregated networks, women have less visibility and lack access to individuals with power ( Ragins and Sundstrom, 1989 ). In gender-segregated networks, it is also difficult for women to find female mentors because there is a lack of women in high-ranking positions ( Noe, 1988 ; Linehan and Scullion, 2008 ). Consequently, the organizational structure can be marked by gender inequalities that reduce women’s chances of reaching top-level positions in an organization.

Gender inequalities can be inherent in the structure of an organization when there are gender segregated departments, job ladders, and networks, which are intimately tied to gender discrimination in HR practices. For instance, if HR policies are designed such that pay is determined based on comparisons between individuals only within a department (e.g., department-wide reporting structure, job descriptions, performance evaluations), then this can lead to a devaluation of departments dominated by women. The overrepresentation of women in certain jobs leads to the lower status of those jobs; consequently, the pay brackets for these jobs decrease over time as the number of women in these jobs increase (e.g., Huffman and Velasco, 1997 ; Reilly and Wirjanto, 1999 ). Similarly, networks led by women are also devalued for pay. For example, in a study of over 2,000 managers, after controlling for performance, the type of job, and the functional area (e.g., marketing, sales, accounting), those who worked with female mangers had lower wages than those who worked with male managers ( Ostroff and Atwater, 2003 ). Thus, gender inequalities in an organization’s structure in terms of gender segregation have reciprocal effects with gender discrimination in HR policy and decision-making.

Another contextual factor in our model is organizational strategy and how institutional discrimination within strategy is related to discrimination in HR practices. Strategy is a plan, method, or process by which an organization attempts to achieve its objectives, such as being profitable, maintaining and expanding its consumer base, marketing strategy, etc. ( Grant, 2010 ). Strategy can influence the level of inequality within an organization ( Morrison and Von Glinow, 1990 ; Hunter et al., 2001 ). For example, Hooters, a restaurant chain, has a marketing strategy to sexually attract heterosexual males, which has led to discrimination in HR policy, decisions, and enactment because only young, good-looking women are considered qualified ( Schneyer, 1998 ). When faced with appearance-based discrimination lawsuits regarding their hiring policies, Hooters has responded by claiming that such appearance requirements are bona fide job qualifications given their marketing strategy (for reviews, see Schneyer, 1998 ; Adamitis, 2000 ). Hooters is not alone, as many other establishments attempt to attract male cliental by requiring their female servers to meet a dress code involving a high level of grooming (make-up, hair), a high heels requirement, and a revealing uniform ( McGinley, 2007 ). Thus, sexist HR policies and practices in which differential standards are applied to male and female employees can stem from a specific organizational strategy ( Westall, 2015 ).

We now consider institutional gender bias within organizational culture and how it relates to discrimination in HR policies. Organizational culture refers to collectively held beliefs, assumptions, and values held by organizational members ( Trice and Beyer, 1993 ; Schein, 2010 ). Cultures arise from the values of the founders of the organization and assumptions about the right way of doing things, which are learned from dealing with challenges over time ( Ostroff et al., 2012 ). The founders and leaders of an organization are the most influential in forming, maintaining, and changing culture over time (e.g., Trice and Beyer, 1993 ; Jung et al., 2008 ; Hartnell and Walumbwa, 2011 ). Organizational culture can contribute to gender inequalities because culture constrains people’s ideas of what is possible: their strategies of action ( Swidler, 1986 ). In other words, when people encounter a problem in their workplace, the organizational culture—who we are, how we act, what is right—will provide only a certain realm of behavioral responses. For instance, in organizational cultures marked by greater gender inequality, women may have lower hopes and expectations for promotion, and when they are discriminated against, may be less likely to imagine that they can appeal their outcomes ( Kanter, 1977 ; Cassirer and Reskin, 2000 ). Furthermore, in organizational cultures marked by gender inequality, organizational decision makers should hold stronger descriptive and proscriptive gender stereotypes: they should more strongly believe that women have less ability to lead, less career commitment, and less emotional stability, compared with men ( Eagly et al., 1992 ; Heilman, 2001 ). We expand upon this point later.

Other aspects of organizational culture that are less obviously related to gender can also lead to discrimination in HR practices. For instance, an organizational culture that emphasizes concerns with meritocracy, can lead organizational members to oppose HR efforts to increase gender equality. This is because when people believe that outcomes ought to go only to those who are most deserving, it is easy for them to fall into the trap of believing that outcomes currently do go to those who are most deserving ( Son Hing et al., 2011 ). Therefore, people will believe that men deserve their elevated status and women deserve their subordinated status at work ( Castilla and Benard, 2010 ). Furthermore, the more people care about merit-based outcomes, the more they oppose affirmative action and diversity initiatives for women ( Bobocel et al., 1998 ; Son Hing et al., 2011 ), particularly when they do not recognize that discrimination occurs against women in the absence of such policies ( Son Hing et al., 2002 ). Thus, a particular organizational culture can influence the level of discrimination against women in HR and prevent the adoption of HR policies that would mitigate gender discrimination.

Finally, gender inequalities can be seen in organizational climates. An organizational climate consists of organizational members’ shared perceptions of the formal and informal organizational practices, procedures, and routines ( Schneider et al., 2011 ) that arise from direct experiences of the organization’s culture ( Ostroff et al., 2012 ). Organizational climates tend to be conceptualized and studied as “climates for” an organizational strategy ( Schneider, 1975 ; Ostroff et al., 2012 ). Gender inequalities are most clearly reflected in two forms of climate: climates for diversity and climates for sexual harassment.

A positive climate for diversity exists when organizational members perceive that diverse groups are included, empowered, and treated fairly. When employees perceive a less supportive diversity climate, they perceive greater workplace discrimination ( Cox, 1994 ; Ragins and Cornwall, 2001 ; Triana and García, 2009 ), and experience lower organizational commitment and job satisfaction ( Hicks-Clarke and Iles, 2000 ), and higher turnover intentions ( Triana et al., 2010 ). Thus, in organizations with a less supportive diversity climate, women are more likely to leave the organization, which contributes to the underrepresentation of women in already male-dominated arenas ( Miner-Rubino and Cortina, 2004 ).

A climate for sexual harassment involves perceptions that the organization is permissive of sexual harassment. In organizational climates that are permissive of harassment, victims are reluctant to come forward because they believe that their complaints will not be taken seriously ( Hulin et al., 1996 ) and will result in negative personal consequences (e.g., Offermann and Malamut, 2002 ). Furthermore, men with a proclivity for harassment are more likely to act out these behaviors when permissive factors are present ( Pryor et al., 1993 ). Therefore, a permissive climate for sexual harassment can result in more harassing behaviors, which can lead women to disengage from their work and ultimately leave the organization ( Kath et al., 2009 ).

Organizational climates for diversity and for sexual harassment are inextricably linked to HR practices. For instance, a factor that leads to perceptions of diversity climates is whether the HR department has diversity training (seminars, workshops) and how much time and money is devoted to diversity efforts ( Triana and García, 2009 ). Similarly, a climate for sexual harassment depends on organizational members’ perceptions of how strict the workplace’s sexual harassment policy is, and how likely offenders are to be punished ( Fitzgerald et al., 1995b ; Hulin et al., 1996 ). Thus, HR policies, decision-making, and their enactment strongly affect gender inequalities in organizational climates and gender inequalities throughout an organization.

In summary, gender inequalities can exist within organizational structures, processes, and practices. However, organizational leadership, structure, strategy, culture, and climate do not inherently need to be sexist. It could be possible for these organizational structures, processes, and practices to promote gender equality. We return to this issue in the conclusion section.

The Effect of Hostile and Benevolent Sexism on How Organizational Decision Makers’ Conduct HR Practices

In this section, we explore how personal biases can affect personal discrimination in HR-related decisions and their enactment. Others have focused on how negative or hostile attitudes toward women predict discrimination in the workplace. However, we extend this analysis by drawing on ambivalent sexism theory, which involves hostile sexism (i.e., antagonistic attitudes toward women) and benevolent sexism (i.e., paternalistic attitudes toward women; see also Glick, 2013 ), both of which lead to discrimination against women.

Stereotyping processes are one possible explanation of how discrimination against women in male-typed jobs occurs and how women are relegated to the “pink ghetto” ( Heilman, 1983 ; Eagly and Karau, 2002 ; Rudman et al., 2012 ). Gender stereotypes, that is, expectations of what women and men are like, and what they should be like, are one of the most powerful schemas activated when people encounter others ( Fiske et al., 1991 ; Stangor et al., 1992 ). According to status characteristics theory, people’s group memberships convey important information about their status and their competence on specific tasks ( Berger et al., 1974 ; Berger et al., 1998 ; Correll and Ridgeway, 2003 ). Organizational decision makers will, for many jobs, have different expectations for men’s and women’s competence and job performance. Expectations of stereotyped-group members’ success can affect gender discrimination that occurs in HR-related decisions and enactment ( Roberson et al., 2007 ). For example, men are preferred over women for masculine jobs and women are preferred over men for feminine jobs ( Davison and Burke, 2000 ). Thus, the more that a workplace role is inconsistent with the attributes ascribed to women, the more a particular woman might be seen as lacking “fit” with that role, resulting in decreased performance expectations ( Heilman, 1983 ; Eagly and Karau, 2002 ).

Furthermore, because women are associated with lower status, and men with higher status, women experience backlash for pursuing high status roles (e.g., leadership) in the workplace ( Rudman et al., 2012 ). In other words, agentic women who act competitively and confidently in a leadership role, are rated as more socially deficient, less likeable and less hireable, compared with men who act the same way ( Rudman, 1998 ; Rudman et al., 2012 ). Interestingly though, if women pursue roles in the workplace that are congruent with traditional gender expectations, they will elicit positive reactions ( Eagly and Karau, 2002 ).

Thus, cultural, widely known, gender stereotypes can affect HR-related decisions. However, such an account does not take into consideration individual differences among organizational decision makers (e.g., managers, supervisors, or HR personnel) who may vary in the extent to which they endorse sexist attitudes or stereotypes. Individual differences in various forms of sexism (e.g., modern sexism, neosexism) have been demonstrated to lead to personal discrimination in the workplace ( Hagen and Kahn, 1975 ; Beaton et al., 1996 ; Hitlan et al., 2009 ). Ambivalent sexism theory builds on earlier theories of sexism by including attitudes toward women that, while sexist, are often experienced as positive in valence by perceivers and targets ( Glick and Fiske, 1996 ). Therefore, we draw on ambivalent sexism theory, which conceptualizes sexism as a multidimensional construct that encompasses both hostile and benevolent attitudes toward women ( Glick and Fiske, 1996 , 2001 ).

Hostile sexism involves antipathy and negative stereotypes about women, such as beliefs that women are incompetent, overly emotional, and sexually manipulative. Hostile sexism also involves beliefs that men should be more powerful than women and fears that women will try to take power from men ( Glick and Fiske, 1996 ; Cikara et al., 2008 ). In contrast, benevolent sexism involves overall positive views of women, as long as they occupy traditionally feminine roles. Individuals with benevolently sexist beliefs characterize women as weak and needing protection, support, and adoration. Importantly, hostile and benevolent sexism tend to go hand-in-hand (with a typical correlation of 0.40; Glick et al., 2000 ). This is because ambivalent sexists, people who are high in benevolent and hostile sexism, believe that women should occupy restricted domestic roles and that women are weaker than men are ( Glick and Fiske, 1996 ). Ambivalent sexists reconcile their potentially contradictory attitudes about women by acting hostile toward women whom they believe are trying to steal men’s power (e.g., feminists, professionals who show competence) and by acting benevolently toward traditional women (e.g., homemakers) who reinforce conventional gender relations and who serve men ( Glick et al., 1997 ). An individual difference approach allows us to build on the earlier models ( Heilman, 1983 ; Eagly and Karau, 2002 ; Rudman et al., 2012 ), by specifying who is more likely to discriminate against women and why.

Organizational decision makers who are higher (vs. lower) in hostile sexism should discriminate more against women in HR-related decisions ( Glick et al., 1997 ; Masser and Abrams, 2004 ). For instance, people high in hostile sexism have been found to evaluate candidates, who are believed to be women, more negatively and give lower employment recommendations for a management position, compared with matched candidates believed to be men ( Salvaggio et al., 2009 ) 1 . In another study, among participants who evaluated a female candidate for a managerial position, those higher in hostile sexism were less likely to recommend her for hire, compared with those lower in hostile sexism ( Masser and Abrams, 2004 ). Interestingly, among those evaluating a matched man for the same position, those higher (vs. lower) in hostile sexism were more likely to recommend him for hire ( Masser and Abrams, 2004 ). According to ambivalent sexism theorists ( Glick et al., 1997 ), because people high in hostile sexism see women as a threat to men’s status, they act as gatekeepers denying women access to more prestigious or masculine jobs.

Furthermore, when enacting HR policies and decisions, organizational decision makers who are higher (vs. lower) in hostile sexism should discriminate more against women in the form of gender harassment. Gender harassment can involve hostile terms of address, negative comments regarding women in management, sexist jokes, and sexist behavior ( Fitzgerald et al., 1995a , b ). It has been found that people higher (vs. lower) in hostile sexism have more lenient attitudes toward the sexual harassment of women, which involves gender harassment, in the workplace ( Begany and Milburn, 2002 ; Russell and Trigg, 2004 ). Furthermore, men who more strongly believe that women are men’s adversaries tell more sexist jokes to a woman ( Mitchell et al., 2004 ). Women also report experiencing more incivility (i.e., low level, rude behavior) in the workplace than men ( Björkqvist et al., 1994 ; Cortina et al., 2001 , 2002 ), which could be due to hostile attitudes toward women. In summary, the evidence is consistent with the idea that organizational decision makers’ hostile sexism should predict their gender harassing behavior during HR enactment; however, more research is needed for such a conclusion.

In addition, organizational decision makers who are higher (vs. lower) in benevolent sexism should discriminate more against women when making HR-related decisions. It has been found that people higher (vs. lower) in benevolent sexism are more likely to automatically associate men with high-authority and women with low-authority roles and to implicitly stereotype men as agentic and women as communal ( Rudman and Kilianski, 2000 ). Thus, organizational decision makers who are higher (vs. lower) in benevolent sexism should more strongly believe that women are unfit for organizational roles that are demanding, challenging, and requiring agentic behavior. Indeed, in studies of male MBA students those higher (vs. lower) in benevolent sexism assigned a fictional woman less challenging tasks than a matched man ( King et al., 2012 ). The researchers reasoned that this occurred because men are attempting to “protect” women from the struggles of challenging work. Although there has been little research conducted that has looked at benevolent sexism and gender discrimination in HR-related decisions, the findings are consistent with our model.

Finally, organizational decision makers who are higher (vs. lower) in benevolent sexism should engage in a complex form of gender discrimination when enacting HR policy and decisions that involves mixed messages: women are more likely to receive messages of positive verbal feedback (e.g., “stellar work,” “excellent work”) but lower numeric ratings on performance appraisals, compared with men ( Biernat et al., 2012 ). It is proposed that this pattern of giving women positive messages about their performance while rating them poorly reflects benevolent sexists’ desire to protect women from harsh criticism. However, given that performance appraisals are used for promotion decisions and that constructive feedback is needed for learning, managers’ unwillingness to give women negative verbal criticisms can lead to skill plateau and career stagnation.

Furthermore, exposure to benevolent sexism can harm women’s motivation, goals and performance. Adolescent girls whose mothers are high in benevolent (but not hostile) sexism display lower academic goals and academic performance ( Montañés et al., 2012 ). Of greater relevance to the workplace, when role-playing a job candidate, women who interacted with a hiring manager scripted to make benevolently sexist statements became preoccupied with thoughts about their incompetence, and consequently performed worse in the interview, compared with those in a control condition ( Dardenne et al., 2007 ). These findings suggest that benevolent sexism during the enactment of HR practices can harm women’s work-related motivation and goals, as well as their performance, which can result in a self-fulfilling prophecy ( Word et al., 1974 ). In other words, the low expectations benevolent sexists have of women can be confirmed by women as they are undermined by paternalistic messages.

Ambivalent sexism can operate to harm women’s access to jobs, opportunities for development, ratings of performance, and lead to stigmatization. However, hostile and benevolent sexism operate in different ways. Hostile sexism has direct negative consequences for women’s access to high status, male-typed jobs ( Masser and Abrams, 2004 ; Salvaggio et al., 2009 ), and it is related to higher rates of sexual harassment ( Fitzgerald et al., 1995b ; Mitchell et al., 2004 ; Russell and Trigg, 2004 ), which negatively affect women’s health, well-being, and workplace withdrawal behaviors ( Willness et al., 2007 ). In contrast, benevolent sexism has indirect negative consequences for women’s careers, for instance, in preventing access to challenging tasks ( King et al., 2012 ) and critical developmental feedback ( Vescio et al., 2005 ). Interestingly, exposure to benevolent sexism results in worsened motivation and cognitive performance, compared with exposure to hostile sexism ( Dardenne et al., 2007 ; Montañés et al., 2012 ). This is because women more easily recognize hostile sexism as a form of discrimination and inequality, compared with benevolent sexism, which can be more subtle in nature ( Dardenne et al., 2007 ). Thus, women can externalize hostile sexism and mobilize against it, but the subtle nature of benevolent sexism prevents these processes ( Kay et al., 2005 ; Becker and Wright, 2011 ). Therefore, hostile and benevolent sexism lead to different but harmful forms of HR discrimination. Future research should more closely examine their potentially different consequences.

Thus far, we have articulated how gender inequalities in organizational structures, processes, and practices can affect discrimination in HR policy and in HR-related decision-making and enactment. Furthermore, we have argued that organizational decision makers’ levels of hostile and benevolent sexism are critical factors leading to personal discrimination in HR-related decision-making and enactment, albeit in different forms. We now turn to an integration of these two phenomena.

The Effect of Organizational Structures, Processes, and Practices on Organizational Decision Makers’ Levels of Hostile and Benevolent Sexism

Organizational decision makers’ beliefs about men and women should be affected by the work environments in which they are embedded. Thus, when there are more gender inequalities within organizational structures, processes, and practices, organizational decision makers should have higher levels of hostile sexism and benevolent sexism. Two inter-related processes can account for this proposition: the establishment of who becomes and remains an organizational member, and the socialization of organizational members.

First, as organizations develop over time, forces work to attract, select, and retain an increasingly homogenous set of employees in terms of their hostile and benevolent sexism ( Schneider, 1983 , 1987 ). In support of this perspective, an individual’s values tend to be congruent with the values in his or her work environment (e.g., Holland, 1996 ; Kristof-Brown et al., 2005 ). People are attracted to and choose to work for organizations that have characteristics similar to their own, and organizations select individuals who are likely to fit with the organization. Thus, more sexist individuals are more likely to be attracted to organizations with greater gender inequality in leadership, structure, strategy, culture, climate, and HR policy; and they will be seen as a better fit during recruitment and selection. Finally, individuals who do not fit with the organization tend to leave voluntarily through the process of attrition. Thus, less (vs. more) sexist individuals would be more likely to leave a workplace with marked gender inequalities in organizational structures, processes, and practices. The opposite should be true for organizations with high gender equality. Through attraction, selection, and attrition processes it is likely that organizational members will become more sexist in a highly gender unequal organization and less sexist in a highly gender equal organization.

Second, socialization processes can change organizational members’ personal attributes, goals, and values to match those of the organization ( Ostroff and Rothausen, 1997 ). Organizational members’ receive both formal and informal messages about gender inequality—or equality—within an organization through their orientation and training, reading of organizational policy, perceptions of who rises in the ranks, how women (vs. men) are treated within the organization, as well as their perception of climates for diversity and sexual harassment. Socialization of organizational members over time has been shown to result in organizational members’ values and personalities changing to better match the values of the organization ( Kohn and Schooler, 1982 ; Cable and Parsons, 2001 ).

These socialization processes can operate to change organizational members’ levels of sexism. It is likely that within more sexist workplaces, people’s levels of hostile and benevolent sexism increase because their normative beliefs shift due to exposure to institutional discrimination against women, others’ sexist attitudes and behavior, and gender bias in culture and climate ( Schwartz and DeKeseredy, 2000 ; Ford et al., 2008 ; Banyard et al., 2009 ). These processes can also lead organizational decision makers to adopt less sexist attitudes in a workplace context marked by greater gender equality. Thus, organizational members’ levels of hostile and benevolent sexism can be shaped by the degree of gender inequalities in organizational structures, processes, and practices and by the sexism levels of their work colleagues.

In addition, organizational decision makers can be socialized to act in discriminatory ways without personally becoming more sexist. If organizational decision makers witness others acting in a discriminatory manner with positive consequences, or acting in an egalitarian way with negative consequences, they can learn to become more discriminatory in their HR practices through observational learning ( Bandura, 1977 , 1986 ). So, organizational decision makers could engage in personal discrimination without being sexist if they perceive that the fair treatment of women in HR would encounter resistance given the broader organizational structures, processes, and practices promoting gender inequality. Yet over time, given cognitive dissonance ( Festinger, 1962 ), it is likely that discriminatory behavior could induce attitude change among organizational decision makers to become more sexist.

Thus far we have argued that gender inequalities in organizational structures, processes, and practices, organizational decision makers’ sexist attitudes, and gender discrimination in HR practices can have reciprocal, reinforcing relationships. Thus, it may appear that we have created a model that is closed and determinate in nature; however, this would be a misinterpretation. In the following section, we outline how organizations marked by gender inequalities can reduce discrimination against women.

How to Reduce Gender Discrimination in Organizations

The model we present for understanding gender discrimination in HR practices is complex. We believe that such complexity is necessary to accurately reflect the realities of organizational life. The model demonstrates that many sources of gender inequality are inter-related and have reciprocal effects. By implication, there are no simple or direct solutions to reduce gender discrimination in organizations. Rather, this complex problem requires multiple solutions. In fact, as discussed by Gelfand et al. (2007) , if an organization attempts to correct discrimination in only one aspect of organizational structure, process, or practice, and not others, such change attempts will be ineffective due to mixed messages. Therefore, we outline below how organizations can reduce gender discrimination by focusing on (a) HR policies (i.e., diversity initiatives and family friendly policies) and closely related organizational structures, processes, and practices; (b) HR-related decision-making and enactment; as well as, (c) the organizational decision makers who engage in such actions.

Reducing Gender Discrimination in HR Policy and Associated Organizational Structures, Processes, and Practices

Organizations can take steps to mitigate discrimination in HR policies. As a first example, let us consider how an organization can develop, within its HR systems, diversity initiatives aimed at changing the composition of the workforce that includes policies to recruit, retain, and develop employees from underrepresented groups ( Jayne and Dipboye, 2004 ). Diversity initiatives can operate like affirmative action programs in that organizations track and monitor (a) the number of qualified candidates from different groups (e.g., women vs. men) in a pool, and (b) the number of candidates from each group hired or promoted. When the proportion of candidates from a group successfully selected varies significantly from their proportion in the qualified pool then action, such as targeted recruitment efforts, needs to be taken.

Importantly, such efforts to increase diversity can be strengthened by other HR policies that reward managers, who select more diverse personnel, with bonuses ( Jayne and Dipboye, 2004 ). Organizations that incorporate diversity-based criteria into their performance and promotion policies and offer meaningful incentives to managers to identify and develop successful female candidates for promotion are more likely to succeed in retaining and promoting diverse talent ( Murphy and Cleveland, 1995 ; Cleveland et al., 2000 ). However, focusing on short-term narrowly defined criteria, such as increasing the number of women hired, without also focusing on candidates’ merit and providing an adequate climate or support for women are unlikely to bring about any long-term change in diversity, and can have detrimental consequences for its intended beneficiaries ( Heilman et al., 1992 , 1997 ). Rather, to be successful, HR policies for diversity need to be supported by the other organizational structures, processes, and practices, such as strategy, leadership, and climate.

For instance, diversity initiatives should be linked to strategies to create a business case for diversity ( Jayne and Dipboye, 2004 ). An organization with a strategy to market to more diverse populations can justify that a more diverse workforce can better serve potential clientele ( Jayne and Dipboye, 2004 ). Alternatively, an organization that is attempting to innovate and grow might justify a corporate strategy to increase diversity on the grounds that diverse groups have multiple perspectives on a problem with the potential to generate more novel, creative solutions ( van Knippenberg et al., 2004 ). Furthermore, organizational leaders must convey strong support for the HR policies for them to be successful ( Rynes and Rosen, 1995 ). Given the same HR policy within an organization, leaders’ personal attitudes toward the policy affects the discrimination levels found within their unit ( Pryor, 1995 ; Pryor et al., 1995 ). Finally, diversity programs are more likely to succeed in multicultural organizations with strong climates for diversity ( Elsass and Graves, 1997 ; Jayne and Dipboye, 2004 ). An organization’s climate for diversity consists of employees’ shared perceptions that the organization’s structures, processes, and practices are committed to maintaining diversity and eliminating discrimination ( Nishii and Raver, 2003 ; Gelfand et al., 2007 ). In organizations where employees perceive a strong climate for diversity, diversity programs result in greater employee attraction and retention among women and minorities, at all levels of the organization ( Cox and Blake, 1991 ; Martins and Parsons, 2007 ).

As a second example of how HR policies can mitigate gender inequalities, we discuss HR policies to lessen employees’ experience of work-family conflict. Work-family conflict is a type of role conflict that workers experience when the demands (e.g., emotional, cognitive, time) of their work role interfere with the demands of their family role or vice versa ( Greenhaus and Beutell, 1985 ). Work-family conflict has the negative consequences of increasing employee stress, illness-related absence, and desire to turnover ( Grandey and Cropanzano, 1999 ). Importantly, women are more adversely affected by work-family conflict than men ( Martins et al., 2002 ). Work-family conflict can be exacerbated by HR policies that evaluate employees based on face time (i.e., number of hours present at the office), as a proxy for organizational commitment ( Perlow, 1995 ; Elsbach et al., 2010 ).

Formal family friendly HR policies can be adopted to relieve work-family conflict directly, which differentially assists women in the workplace. For instance, to reduce work-family conflict, organizations can implement HR policies such as flexible work arrangements, which involve flexible schedules, telecommuting, compressed work weeks, job-shares, and part-time work ( Galinsky et al., 2008 ). In conjunction with other family friendly policies, such as the provision of childcare, elderly care, and paid maternity leave, organizations can work to reduce stress and improve the retention of working mothers ( Burke, 2002 ).

Unfortunately, it has been found that the enactment of flexible work policies can still lead to discrimination. Organizational decision makers’ sexism can lead them to grant more flexible work arrangements to white men than to women and other minorities because white men are seen as more valuable ( Kelly and Kalev, 2006 ). To circumvent this, organizations need to formalize HR policies relating to flexible work arrangements ( Kelly and Kalev, 2006 ). For instance, formal, written policies should articulate who can adopt flexible work arrangements (e.g., employees in specific divisions or with specific job roles) and what such arrangements look like (e.g., core work from 10 am to 3 pm with flexible work hours from 7 to 10 am or from 3 to 6 pm). When the details of such policies are formally laid out, organizational decision makers have less latitude and therefore less opportunity for discrimination in granting access to these arrangements.

To be successful, family friendly HR policies should be tied to other organizational structures, processes, and practices such as organizational strategy, leadership, culture, and climate. A business case for flexible work arrangements can be made because they attract and retain top-talent, which includes women ( Baltes et al., 1999 ). Furthermore, organizational leaders must convey strong support for family friendly programs ( Jayne and Dipboye, 2004 ). Leaders can help bolster the acceptance of family friendly policies through successive interactions, communications, visibility, and role modeling with employees. For instance, a leader who sends emails at 2 o’clock in the morning is setting a different expectation of constant availability than a leader who never sends emails after 7:00 pm. Family friendly HR policies must also be supported by simultaneously changing the underlying organizational culture that promotes face time. Although it is difficult to change the culture of an organization, the leaders’ of the organization play an influential role in instilling such change because the behaviors of leaders are antecedents and triggers of organizational culture ( Kozlowski and Doherty, 1989 ; Ostroff et al., 2012 ). In summary, HR policies must be supported by other organizational structures, processes, and practices in order for these policies to be effective.

Adopting HR diversity initiative policies and family friendly policies can reduce gender discrimination and reshape the other organizational structures, processes, and practices and increase gender equality in them. Specifically, such policies, if successful, should increase the number of women in all departments and at all levels of an organization. Further, having more women in leadership positions signals to organizational members that the organization takes diversity seriously, affecting the diversity climate of the organization, and ultimately its culture ( Konrad et al., 2010 ). Thus, particular HR policies can reduce gender inequalities in all of the other organizational structures, processes, and practices.

Reducing Gender Discrimination in HR-Related Decision-Making and Enactment

A wealth of research demonstrates that an effective means of reducing personal bias by organizational decision makers in HR practices is to develop HR policies that standardize and objectify performance data (e.g., Konrad and Linnehan, 1995 ; Reskin and McBrier, 2000 ). To reduce discrimination in personnel decisions (i.e., employee hiring and promotion decisions) a job analysis should be performed to determine the appropriate knowledge skills and abilities needed for specific positions ( Fine and Cronshaw, 1999 ). This ensures that expectations about characteristics of the ideal employee for that position are based on accurate knowledge of the job and not gender stereotypes about the job ( Welle and Heilman, 2005 ). To reduce discrimination in performance evaluations, HR policies should necessitate the use of reliable measures based on explicit objective performance expectations and apply these practices consistently across all worker evaluations ( Bernardin et al., 1998 ; Ittner et al., 2003 ). Employees’ performance should be evaluated using behaviorally anchored rating scales ( Smith and Kendall, 1963 ) that allow supervisors to rate subordinates on examples of actual work behaviors. These evaluations should be done regularly, given that delays require retrieving memories of work performance and this process can be biased by gender stereotypes ( Sanchez and De La Torre, 1996 ). Finally, if greater gender differences are found on selection tests than on performance evaluations, then the use of such biased selection tests needs to be revisited ( Chung-Yan and Cronshaw, 2002 ). In summary, developing HR policies that standardize and objectify the process of employee/candidate evaluations can reduce personal bias in HR practices.

Importantly, the level of personal discrimination enacted by organizational decision makers can be reduced by formalizing HR policies, and by controlling the situations under which HR-related decisions are made. We have articulated how HR-related decisions involve social cognition and are therefore susceptible to biases introduced by the use of gender stereotypes. This can occur unwittingly by those who perceive themselves to be unprejudiced but who are affected by stereotypes or negative automatic associations nonetheless ( Chugh, 2004 ; Son Hing et al., 2008 ). For instance, when HR policies do not rely on objective criteria, and the context for evaluation is ambiguous, organizational decision makers will draw on gender (and other) stereotypes to fill in the blanks when evaluating candidates ( Heilman, 1995 , 2001 ). Importantly, the context can be constructed in such a way as to reduce these biases. For instance, organizational decision makers will make less biased judgments of others if they have more time available to evaluate others, are less cognitively busy ( Martell, 1991 ), have higher quality of information available about candidates, and are accountable for justifying their ratings and decisions ( Kulik and Bainbridge, 2005 ; Roberson et al., 2007 ). Thus, if they have the time, motivation, and opportunity to make well-informed, more accurate judgments, then discrimination in performance ratings can be reduced.

Reducing Organizational Decision Makers’ Sexism

Another means to reduce gender discrimination in HR-related decision-making and enactment is to focus directly on reducing the hostile and benevolent sexist beliefs of organizational decision makers. Interventions aimed at reducing these beliefs typically involve diversity training, such as a seminar, course, or workshop. Such training involves one or more sessions that involve interactive discussions, lectures, and practical assignments. During the training men and women are taught about sexism and how gender roles in society are socially constructed. Investigations have shown these workshop-based interventions are effective at reducing levels of hostile sexism but have inconsistent effects on benevolent sexism ( Case, 2007 ; de Lemus et al., 2014 ). The subtle, and in some ways positive nature of benevolent sexism makes it difficult to confront and reduce using such interventions. However, levels of benevolent sexism are reduced when individuals are explicitly informed about the harmful implications of benevolent sexism ( Becker and Swim, 2012 ). Unfortunately, these interventions have not been tested in organizational settings. So their efficacy in the field is unknown.

Gender inequality in organizations is a complex phenomenon that can be seen in HR practices (i.e., policies, decision-making, and their enactment) that affects the hiring, training, pay, and promotion of women. We propose that gender discrimination in HR-related decision-making and the enactment of HR practices stems from gender inequalities in broader organizational structures, processes, and practices, including HR policy but also leadership, structure, strategy, culture, and organizational climate. Moreover, reciprocal effects should occur, such that discriminatory HR practices can perpetuate gender inequalities in organizational leadership, structure, strategy, culture, and climate. Organizational decision makers also play an important role in gender discrimination. We propose that personal discrimination in HR-related decisions and enactment arises from organizational decision makers’ levels of hostile and benevolent sexism. While hostile sexism can lead to discrimination against women because of a desire to keep them from positions of power, benevolent sexism can lead to discrimination against women because of a desire to protect them. Finally, we propose that gender inequalities in organizational structures, processes, and practices affect organizational decision makers’ sexism through attraction, selection, socialization, and attrition processes. Thus, a focus on organizational structure, processes, and practices is critical.

The model we have developed extends previous work by Gelfand et al. (2007) in a number of substantive ways. Gelfand et al. (2007) proposed that aspects of the organization, that is, structure, organizational culture, leadership, strategy, HR systems, and organizational climates, are all interrelated and may contribute to or attenuate discrimination (e.g., racism, sexism, ableism, homophobia). First, we differ from their work by emphasizing that workplace discrimination is most directly attributable to HR practices. Consequently, we emphasize how inequalities in other organizational structures, processes, and practices affect institutional discrimination in HR policy. Second, our model differs from that of Gelfand et al. (2007) in that we focus on the role of organizational decision makers in the enactment of HR policy. The attitudes of these decision makers toward specific groups of employees are critical. However, the nature of prejudice differs depending on the target group ( Son Hing and Zanna, 2010 ). Therefore, we focus on one form of bias—sexism—in the workplace. Doing so, allows us to draw on more nuanced theories of prejudice, namely ambivalent sexism theory ( Glick and Fiske, 1996 ). Thus, third, our model differs from the work of Gelfand et al. (2007) by considering how dual beliefs about women (i.e., hostile and benevolent beliefs) can contribute to different forms of gender discrimination in HR practices. Fourth, we differ from Gelfand et al. (2007) by reviewing how organizational decision makers’ level of sexism within an organization is affected by organizational structures, processes, and practices via selection-attraction-attrition processes and through socialization processes.

However, the model we have developed is not meant to be exhaustive. There are multiple issues that we have not addressed but should be considered: what external factors feed into our model? What other links within the model might arise? What are the limits to its generalizability? What consequences derive from our model? How can change occur given a model that is largely recursive in nature? We focus on these issues throughout our conclusion.

In this paper, we have illustrated what we consider to be the dominant links in our model; however, additional links are possible. First, we do not lay out the factors that feed into our model, such as government regulations, the economy, their competitors, and societal culture. In future work, one could analyze the broader context that organizations operate in, which influences its structures, processes, and practices, as well as its members. For instance, in societies marked by greater gender inequalities, the levels of hostile and benevolent sexism of organizational decision makers will be higher ( Glick et al., 2000 ). Second, there is no link demonstrating how organizational decision makers who are more sexist have the capacity, even if they sit lower in the organizational hierarchy, to influence the amount of gender inequality in organizational structures, processes, and practices. It is possible for low-level managers or HR personnel who express more sexist sentiments to—through their own behavior—affect others’ perceptions of the tolerance for discrimination in the workplace ( Ford et al., 2001 ) and others’ perceptions of the competence and hireability of female job candidates ( Good and Rudman, 2010 ). Thus, organizational decision makers’ levels of hostile and benevolent sexism can affect organizational climates, and potentially other organizational structures, processes, and practices. Third, it is possible that organizational structures, processes, and practices could moderate the link between organizational decision makers’ sexist attitudes and their discriminatory behavior in HR practices. The ability of people to act in line with their attitudes depends on the strength of the constraints in the social situation and the broader context ( Lewin, 1935 , 1951 ). Thus, if organizational structures, processes, and practices clearly communicate the importance of gender equality then the discriminatory behavior of sexist organizational decision makers should be constrained. Accordingly, organizations should take steps to mitigate institutional discrimination by focusing on organizational structures, processes, and practices rather than focusing solely on reducing sexism in individual employees.

Our model does not consider how women’s occupational status is affected by their preferences for gender-role-consistent careers and their childcare and family responsibilities, which perhaps should not be underestimated (e.g., Manne, 2001 ; Hakim, 2006 ; Ceci et al., 2009 ). In other words, lifestyle preferences could contribute to gender differences in the workplace. However, it is important to consider how women’s agency in choosing occupations and managing work-life demands is constrained. Gender imbalances (e.g., in pay) in the workplace (e.g., Moss-Racusin et al., 2012 ; Sheltzer and Smith, 2014 ) and gender imbalances in the home (e.g., in domestic labor, childcare; Bianchi, 2000 ; Bianchi et al., 2000 ) shape the decisions that couples (when they consist of a woman and a man) make about how to manage dual careers. For instance, research has uncovered that women with professional degrees leave the labor force at roughly three times the rate of men ( Baker, 2002 ). Women’s decisions to interrupt their careers were difficult and were based on factors, such as workplace inflexibility, and their husbands’ lack of domestic responsibilities, rather than a preference to stay at home with their children ( Stone and Lovejoy, 2004 ). Thus, both factors inside and outside the workplace constrain and shape women’s career decisions.

Our model is derived largely from research that has been conducted in male-dominated organizations; however, we speculate that it should hold for female-dominated organizations. There is evidence that tokenism does not work against men in terms of their promotion potential in female-dominated environments. Rather, there is some evidence for a glass-escalator effect for men in female-dominated fields, such as nursing, and social work ( Williams, 1992 ). In addition, regardless of the gender composition of the workplace, men are advantaged, compared with women in terms of earnings and wage growth ( Budig, 2002 ). Finally, even in female-dominated professions, segregation along gender lines occurs in organizational structure ( Snyder and Green, 2008 ). Thus, the literature suggests that our model should hold for female-dominated environments.

Some might question if our model assumes that organizational decision makers enacting HR practices are men. It does not. There is evidence that decision makers who are women also discriminate against women (e.g., the Queen Bee phenomenon; Ellemers et al., 2004 ). Further, although men are higher in hostile sexism, compared with women ( Glick et al., 1997 , 2000 ), they are not necessarily higher in benevolent sexism ( Glick et al., 2000 ). More importantly, the effects of hostile and benevolent sexism are not moderated by participant gender ( Masser and Abrams, 2004 ; Salvaggio et al., 2009 ; Good and Rudman, 2010 ). Thus, those who are higher in hostile or benevolent sexism respond in a more discriminatory manner, regardless of whether they are men or women. Thus, organizational decision makers, regardless of their sex, should discriminate more against women in HR practices when they are higher in hostile or benevolent sexism.

In future work, the consequences of our model for women discriminated against in HR practices should be considered. The negative ramifications of sexism and discrimination on women are well known: physical and psychological stress, worse physical health (e.g., high blood pressure, ulcers, anxiety, depression; Goldenhar et al., 1998 ); lower job satisfaction, organizational commitment, and attachment to work ( Murrell et al., 1995 ; Hicks-Clarke and Iles, 2000 ); lower feelings of power and prestige ( Gutek et al., 1996 ); and performance decrements through stereotype threat ( Spencer et al., 1999 ). However, how might these processes differ depending on the proximal cause of the discrimination?

Our model lays out two potential paths by which women might be discriminated against in HR practices: institutional discrimination stemming from organizational structures, processes, and practices and personal discrimination stemming from organizational decision makers’ levels of sexism. In order for the potential stressor of stigmatization to lead to psychological and physical stress it must be seen as harmful and self-relevant ( Son Hing, 2012 ). Thus, if institutional discrimination in organizational structures, processes, and practices are completely hidden then discrimination might not cause stress reactions associated with stigmatization because it may be too difficult for women to detect ( Crosby et al., 1986 ; Major, 1994 ), and label as discrimination ( Crosby, 1984 ; Stangor et al., 2003 ). In contrast, women should be adversely affected by stigmatization in instances where gender discrimination in organizational structures, processes, and practices is more evident. For instance, greater perceptions of discrimination are associated with lower self-esteem in longitudinal studies ( Schmitt et al., 2014 ).

It might appear that we have created a model, which is a closed system, with no opportunities to change an organization’s trajectory: more unequal organizations will become more hierarchical, and more equal organizations will become more egalitarian. We do not believe this to be true. One potential impetus for organizations to become more egalitarian may be some great shock such as sex-based discrimination lawsuits that the organization either faces directly or sees its competitors suffer. Large corporations have been forced to settle claims of gender harassment and gender discrimination with payouts upward of $21 million ( Gilbert v. DaimlerChrysler Corp., 2004 ; LexisNexis, 2010 ; Velez, et al. v. Novartis Pharmaceuticals Crop, et al., 2010 ). Discrimination lawsuits are time consuming and costly ( James and Wooten, 2006 ), resulting in lower shares, lower public perceptions, higher absenteeism, and higher turnover ( Wright et al., 1995 ). Expensive lawsuits experienced either directly or indirectly should act as a big driver in the need for change.

Furthermore, individual women can work to avoid stigmatization. Women in the workplace are not simply passive targets of stereotyping processes. People belonging to stigmatized groups can engage in a variety of anti-stigmatization techniques, but their response options are constrained by the cultural repertoires available to them ( Lamont and Mizrachi, 2012 ). In other words, an organization’s culture will provide its members with a collective imaginary for how to behave. For instance, it might be unimaginable for a woman to file a complaint of sexual harassment if she knows that complaints are never taken seriously. Individuals do negotiate stigmatization processes; however, this is more likely when stigmatization is perceived as illegitimate and when they have the resources to do so ( Major and Schmader, 2001 ). Thus, at an individual level, people engage in strategies to fight being discriminated against but these strategies are likely more constrained for those who are most stigmatized.

Finally, possibly the most efficacious way for organizational members (men and women) to challenge group-based inequality and to improve the status of women as a whole is to engage in collective action (e.g., participate in unions, sign petitions, organize social movements, recruit others to join a movement; Klandermans, 1997 ; Wright and Lubensky, 2009 ). People are most likely to engage in collective action when they perceive group differences as underserved or illegitimate ( Wright, 2001 ). Such a sense of relative deprivation involves feelings of injustice and anger that prompt a desire for wide scale change ( van Zomeren et al., 2008 ). Interestingly, people are more likely to experience relative deprivation when inequalities have begun to be lessened, and thus their legitimacy questioned ( Crosby, 1984 ; Kawakami and Dion, 1993 ; Stangor et al., 2003 ). If organizational leaders respond to such demands for change by altering previously gender oppressive organizational structures, processes, and practices, this can, in people’s minds, open the door for additional changes. Therefore, changes to mitigate gender inequalities within any organizational structure, policy, or practice could start a cascade of transformations leading to a more equal organization for men and women.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

This research was supported by funding from the Canadian Institute for Advanced Research (CIFAR) awarded to Leanne S. Son Hing.

1 In this study, candidates were identified with initials and participants were asked to indicate the presumed gender of the candidate after evaluating them.

• Abrams K. (1991). Social construction, roving biologism, and reasonable women: a response to Professor Epstein. DePaul Law Rev. 41 1021–1040. [ Google Scholar ]
• Acker J. (1990). Hierarchies, jobs, bodies: a theory of gendered organizations. Gend. Soc. 4 139–158. 10.1177/089124390004002002 [ CrossRef ] [ Google Scholar ]
• Adamitis E. M. (2000). Appearance matters: a proposal to prohibit appearance discrimination in employment. Wash. Law Rev. 75 195–223. [ Google Scholar ]
• Adler N. E., Epel E. S., Castellazzo G., Ickovics J. R. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy White women. Health Psychol. 19 586–592. 10.1037/0278-6133.19.6.586 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Baker J. G. (2002). The influx of women into legal professions: an economic analysis. Mon. Labor Rev. 125 12–24. [ Google Scholar ]
• Baltes B. B., Briggs T. E., Huff J. W., Wright J. A., Neuman G. A. (1999). Flexible and compressed workweek schedules: a meta-analysis of their effects on work-related criteria. J. Appl. Psychol. 84 496–513. 10.1037/0021-9010.84.4.496 [ CrossRef ] [ Google Scholar ]
• Bandura A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychol. Rev. 84 191–215. 10.1037/0033-295X.84.2.191 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Bandura A. (1986). The explanatory and predictive scope of self-efficacy theory. J. Soc. Clin. Psychol. 4 359–373. 10.1521/jscp.1986.4.3.359 [ CrossRef ] [ Google Scholar ]
• Banyard V. L., Moynihan M. M., Crossman M. T. (2009). Reducing sexual violence on campus: the role of student leaders as empowered bystanders. J. Coll. Stud. Dev. 50 446–457. 10.1353/csd.0.0083 [ CrossRef ] [ Google Scholar ]
• Barling J., Dekker I., Loughlin C. A., Kelloway E. K., Fullagar C., Johnson D. (1996). Prediction and replication of the organizational and personal consequences of workplace sexual harassment. J. Manag. Psychol. 11 4–25. 10.1108/02683949610124771 [ CrossRef ] [ Google Scholar ]
• Baxter J., Wright E. O. (2000). The glass ceiling hypothesis: a comparative study of the United States, Sweden, and Australia. Gend. Soc. 14 275–294. 10.1177/089124300014002004 [ CrossRef ] [ Google Scholar ]
• Beaton A. M., Tougas F., Joly S. (1996). Neosexism among male managers: is it a matter of numbers? J. Appl. Soc. Psychol. 26 2189–2203. 10.1111/j.1559-1816.1996.tb01795.x [ CrossRef ] [ Google Scholar ]
• Becker J. C., Swim J. K. (2012). Reducing endorsement of benevolent and modern sexist beliefs: differential effects of addressing harm versus pervasiveness of benevolent sexism. Soc. Psychol. 43 127–137. 10.1027/1864-9335/a000091 [ CrossRef ] [ Google Scholar ]
• Becker J. C., Wright S. C. (2011). Yet another dark side of chivalry: benevolent sexism undermines and hostile sexism motivates collective action for social change. J. Pers. Soc. Psychol. 101 62–77. 10.1037/a0022615 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Begany J. J., Milburn M. A. (2002). Psychological predictors of sexual harassment: authoritarianism, hostile sexism, and rape myths. Psychol. Men Masc. 3 119–126. 10.1037/1524-9220.3.2.119 [ CrossRef ] [ Google Scholar ]
• Berdhal J. L. (2007). The sexual harassment of uppity women. J. Appl. Psychol. 92 425–437. 10.1037/0021-9010.92.2.425 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Berger J., Conner T. L., Fisek M. H. (eds). (1974). Expectation States Theory: A Theoretical Research Program. Cambridge, MA: Winthrop Publishers. [ Google Scholar ]
• Berger J., Fisek M. H., Norman R. Z., Wagner D. G. (1998). “Formation of reward expectations in status situations,” in Status, Power, and Legitimacy eds Berger J., Zelditch M., Jr. (New Brunswick, NJ: Transaction Publishers; ) 121–153. [ Google Scholar ]
• Bernardin H. J., Hagan C. M., Kane J. S., Villanova P. (1998). “Effective performance management: a focus on precision, customers, and situational constraints,” in Performance Appraisal: State of the Art in Practice ed. Smither J. W. (San Francisco, CA: Jossey-Bass; ) 3–48. [ Google Scholar ]
• Bianchi S. M. (2000). Maternal employment and time with children: dramatic change or surprising continuity? Demography 47 401–414. 10.1353/dem.2000.0001 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Bianchi S. M., Milkie M. A., Sayer L. C., Robinson J. P. (2000). Is anyone doing the housework? Trends in the gender division of household labor. Soc. Forces 79 191–228. 10.1093/sf/79.1.191 [ CrossRef ] [ Google Scholar ]
• Biernat M., Tocci M. J., Williams J. C. (2012). The language of performance evaluations: gender-based shifts in content and consistency of judgment. Soc. Psychol. Personal. Sci. 3 186–192. 10.1177/1948550611415693 [ CrossRef ] [ Google Scholar ]
• Björkqvist K., Österman K., Hjelt-Bäck M. (1994). Aggression among university employees. Aggress. Behav. 20 173–184. 10.1002/1098-2337(1994)20:3<173::AID-AB2480200304>3.0.CO;2-D [ CrossRef ] [ Google Scholar ]
• Blau F. D., DeVaro J. (2007). New evidence on gender differences in promotion rates: an empirical analysis of a sample of new hires. Ind. Relat. 46 511–550. 10.1111/j.1468-232X.2007.00479.x [ CrossRef ] [ Google Scholar ]
• Bobocel R. D., Son Hing L. S., Davey L. M., Stanley D. J., Zanna M. P. (1998). Justice-based opposition to social policies: is it genuine? J. Pers. Soc. Psychol. 75 653–669. 10.1037/0022-3514.75.3.653 [ CrossRef ] [ Google Scholar ]
• Boldry J., Wood W., Kashy D. A. (2001). Gender stereotypes and the evaluation of men and women in military training. J. Soc. Issues 57 689–705. 10.1111/0022-4537.00236 [ CrossRef ] [ Google Scholar ]
• Borrel C., Artazcoz L., Gil-González D., Pérez G., Rohlfs I., Pérez K. (2010). Perceived sexism as a health determinant in Spain. J. Womens Health 19 741–750. 10.1089/jwh.2009.1594 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Branscombe N. R. (1998). Thinking about one’s gender group’s privileges or disadvantages: consequences for well-being in women and men. Br. J. Soc. Psychol. 37 167–184. 10.1111/j.2044-8309.1998.tb01163.x [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Brass D. J. (1985). Men’s and women’s networks: a study of interaction patterns and influence in an organization. Acad. Manag. J. 28 327–343. 10.2307/256204 [ CrossRef ] [ Google Scholar ]
• Budig M. J. (2002). Male advantage and the gender composition of jobs: who rides the glass escalator? Soc. Probl. 49 258–277. 10.1525/sp.2002.49.2.258 [ CrossRef ] [ Google Scholar ]
• Burke R. J. (2002). Organizational values, job experiences and satisfactions among managerial and professional women and men: advantage men? Women Manag. Rev. 17 228–236. 10.1108/09649420210433184 [ CrossRef ] [ Google Scholar ]
• Cable D. M., Parsons C. K. (2001). Socialization tactics and person-organization fit. Pers. Psychol. 54 1–23. 10.1111/j.1744-6570.2001.tb00083.x [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Case K. A. (2007). Raising male privilege awareness and reducing sexism: an evaluation of diversity courses. Psychol. Women Q. 31 426–435. 10.1111/j.1471-6402.2007.00391.x [ CrossRef ] [ Google Scholar ]
• Cassirer N., Reskin B. (2000). High Hopes: organizational position, employment experiences, and women and men’s promotion aspirations. Work Occup. 27 438–463. 10.1177/0730888400027004002 [ CrossRef ] [ Google Scholar ]
• Castilla E. J., Benard S. (2010). The paradox of meritocracy in organizations. Admin. Sci. Q. 55 543–576. 10.2189/asqu.2010.55.4.543 [ CrossRef ] [ Google Scholar ]
• Ceci S. J., Williams W. M., Barnett S. M. (2009). Women’s underrepresentation in science: sociocultural and biological considerations. Psychol. Bull. 135 218–261. 10.1037/a0014412 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Chemers M. M. (1997). An Integrative Theory of Leadership. Mahwah, NJ: Lawrence Erlbaum Associates. [ Google Scholar ]
• Chugh D. (2004). Societal and managerial implications of implicit social cognition: why milliseconds matter. Soc. Justice Res. 17 203–222. 10.1023/B:SORE.0000027410.26010.40 [ CrossRef ] [ Google Scholar ]
• Chung-Yan G. A., Cronshaw S. F. (2002). A critical re-examination and analysis of cognitive ability tests using the Thorndike model of fairness. J. Occup. Organ. Psychol. 75 489–509. 10.1348/096317902321119709 [ CrossRef ] [ Google Scholar ]
• Cikara M., Lee T. L., Fiske S. T., Glick P. (2008). “Ambivalent sexism at home and at work: how attitudes toward women in relationships foster exclusion in the public sphere,” in Social and Psychological Bases of Ideology and System Justification eds Jost J. T., Kay A. C., Thorisdottir H. (New York: Oxford University Press; ) 444–462. [ Google Scholar ]
• Cleveland J. N., Stockdale M., Murphy K. R., Gutek B. A. (2000). Women and Men in Organizations: Sex and Gender Issues at Work. Mahwah, NJ: Lawrence Erlbaum Associates. [ Google Scholar ]
• Cohen P. N., Huffman M. L. (2007). Working for the woman? Female managers and the gender wage gap. Am. Sociol. Rev. 72 681–704. 10.1177/000312240707200502 [ CrossRef ] [ Google Scholar ]
• Cohen-Charash Y., Spector P. E. (2001). The role of justice in organizations: a meta-analysis. Organ. Behav. Hum. Decis. Process. 86 278–321. 10.1006/obhd.2001.2958 [ CrossRef ] [ Google Scholar ]
• Correll S. J., Ridgeway C. L. (2003). “Expectation states theory,” in Handbook of Social Psychology ed. Delamater J. (New York, NY: Kluwer Academic Press; ) 29–51. [ Google Scholar ]
• Cortina L. M., Lonsway K. A., Magley V. J., Freeman L. V., Collinsworth L. L., Hunter M., et al. (2002). What’s gender got to do with it? Incivility in the federal courts. Law Soc. Inq. 27 235–270. 10.1111/j.1747-4469.2002.tb00804.x [ CrossRef ] [ Google Scholar ]
• Cortina L. M., Magley V. J., Williams J. H., Langhout R. D. (2001). Incivility in the workplace: incidence and impact. J. Occup. Health Psychol. 6 64–80. 10.1037/1076-8998.6.1.64 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Cox T. (1994). Cultural Diversity in Organizations: Theory, Research, and Practice. San Francisco, CA: Berrett-Koehler Publishers. [ Google Scholar ]
• Cox T. H., Blake S. (1991). Managing cultural diversity: implications for organizational competitiveness. Executive 5 45–56. 10.5465/AME.1991.4274465 [ CrossRef ] [ Google Scholar ]
• Cox T. H., Harquail C. V. (1991). Career paths and career success in the early career stages of male and female MBAs. J. Vocat. Behav. 39 54–75. 10.1016/0001-8791(91)90004-6 [ CrossRef ] [ Google Scholar ]
• Crocker D., Kalemba V. (1999). The incidence and impact of women’s experiences of sexual harassment in Canadian workplaces. Can. Rev. Sociol. 36 541–558. 10.1111/j.1755-618X.1999.tb00963.x [ CrossRef ] [ Google Scholar ]
• Crosby F. (1984). The denial of personal discrimination. Am. Behav. Sci. 27 371–386. 10.1177/000276484027003008 [ CrossRef ] [ Google Scholar ]
• Crosby F., Clayton S., Alksnis O., Hemker K. (1986). Cognitive biases in the perception of discrimination: the importance of format. Sex Roles 14 637–646. 10.1007/BF00287694 [ CrossRef ] [ Google Scholar ]
• Cuddy A. J. C., Fiske S. T., Glick P. (2004). When professionals become mothers, warmth doesn’t cut the ice. J. Soc. Issues 60 701–718. 10.1111/j.0022-4537.2004.00381.x [ CrossRef ] [ Google Scholar ]
• Dardenne B., Dumont M., Bollier T. (2007). Insidious dangers of benevolent sexism: consequences for women’s performance. J. Pers. Soc. Psychol. 93 764–779. 10.1037/0022-3514.93.5.764 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Davison H. K., Burke M. J. (2000). Sex discrimination in simulated employment contexts: a meta-analytic investigation. J. Vocat. Behav. 56 225–248. 10.1006/jvbe.1999.1711 [ CrossRef ] [ Google Scholar ]
• de Lemus S., Navarro L., Velázquez M. J., Ryan E., Megías J. L. (2014). From sex to gender: a university intervention to reduce sexism in Argentina, Spain, and El Salvador. J. Soc. Issues 70 741–762. 10.1111/josi.12089 [ CrossRef ] [ Google Scholar ]
• De Pater I. E., Van Vianen A. E. M., Bechtoldt M. N. (2010). Gender differences in job challenge: a matter of task allocation. Gend. Work Organ. 17 433–453. 10.1111/j.1468-0432.2009.00477.x [ CrossRef ] [ Google Scholar ]
• Durden G. C., Gaynor P. E. (1998). More on the cost of being other than white and male. Am. J. Econ. Sociol. 57 95–103. 10.1111/j.1536-7150.1998.tb03259.x [ CrossRef ] [ Google Scholar ]
• Eagly A. H., Carli L. L. (2007). Through the Labyrinth: The Truth about How Women become Leaders. Boston, MA: Harvard Business School Publishing. [ Google Scholar ]
• Eagly A. H., Karau S. J. (2002). Role congruity theory of prejudice toward female leaders. Psychol. Rev. 109 573–598. 10.1037/0033-295X.109.3.573 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Eagly A. H., Makhijani M. G., Klonsky B. G. (1992). Gender and the evaluation of leaders: a meta-analysis. Psychol. Bull. 111 3–22. 10.1037/0033-2909.111.1.3 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ellemers N., Heuvel H., Gilder D., Maass A., Bonvini A. (2004). The underrepresentation of women in science: differential commitment or the queen bee syndrome? Br. J. Soc. Psychol. 43 315–338. 10.1348/0144666042037999 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Elsass P. M., Graves L. M. (1997). Demographic diversity in decision-making groups: the experiences of women and people of color. Acad. Manag. Rev. 22 946–973. 10.5465/AMR.1997.9711022111 [ CrossRef ] [ Google Scholar ]
• Elsbach K. D., Cable D. M., Sherman J. W. (2010). How passive ‘face time’ affects perceptions of employees: evidence of spontaneous trait inference. Hum. Relat. 63 735–760. 10.1177/0018726709353139 [ CrossRef ] [ Google Scholar ]
• Festinger L. (1962). A Theory of Cognitive Dissonance. Stanford, CA: Stanford University Press. [ Google Scholar ]
• Fine S. A., Cronshaw S. F. (1999). Functional Job Analysis: A Foundation for Human Resources Management. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. [ Google Scholar ]
• Firth M. (1982). Sex discrimination in job opportunities for women. Sex Roles 8 891–901. 10.1007/BF00287858 [ CrossRef ] [ Google Scholar ]
• Fiske A. P., Haslam N., Fiske S. T. (1991). Confusing one person with another: what errors reveal about the elementary forms of social relations. J. Pers. Soc. Psychol. 60 656–674. 10.1037/0022-3514.60.5.656 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Fiske S. T., Cuddy A. J. C., Glick P., Xu J. (2002). A model of (often mixed) stereotype content: competence and warmth respectively follow from perceived status and competition. J. Pers. Soc. Psychol. 82 878–902. 10.1037//0022-3514.82.6.878 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Fiske S. T., Xu J., Cuddy A. J. C., Glick P. (1999). (Dis)respecting versus (dis)liking: status and interdependence predict ambivalent stereotypes of competence and warmth. J. Soc. Issues 55 473–489. 10.1111/0022-4537.00128 [ CrossRef ] [ Google Scholar ]
• Fitzgerald L. F., Gelfand M. J., Drasgow F. (1995a). Measuring sexual harassment: theoretical and psychometric advances. Basic Appl. Soc. Psychol. 17 425–445. 10.1207/s15324834basp1704_2 [ CrossRef ] [ Google Scholar ]
• Fitzgerald L. F., Hulin C. L., Drasgow F. (1995b). “The antecedents and consequences of sexual harassment in organizations: an integrated model,” in Job Stress in a Changing Workforce: Investigating Gender, Diversity, and Family Issues eds Keita G., Hurrell J., Jr. (Washington, DC: American Psychological Association; ) 55–73. [ Google Scholar ]
• Fitzgerald L. F., Shullman S. L., Bailey N., Richards M., Swecker J., Gold Y., et al. (1988). The incidence and dimensions of sexual harassment in academia and the workplace. J. Vocat. Behav. 32 152–175. 10.1016/0001-8791(88)90012-7 [ CrossRef ] [ Google Scholar ]
• Ford T. E., Boxer C. F., Armstrong J., Edel J. R. (2008). More than “just a joke”: the prejudice-releasing function of sexist humor. Pers. Soc. Psychol. Bull. 34 159–170. 10.1177/0146167207310022 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ford T. E., Wentzel E. R., Lorion J. (2001). Effects of exposure to sexist humor on perceptions of normative tolerance of sexism. Eur. J. Soc. Psychol. 31 677–691. 10.1002/ejsp.56 [ CrossRef ] [ Google Scholar ]
• Fuegen K., Biernat M., Haines E., Deaux K. (2004). Mothers and fathers in the workplace: how gender and parental status influence judgements of job-related competence. J. Soc. Issues 60 737–754. 10.1111/j.0022-4537.2004.00383.x [ CrossRef ] [ Google Scholar ]
• Galinsky E., Bond J., Sakai K. (2008). 2008 National Study of Employers. Available at: http://familiesandwork.org/site/research/reports/2008nse.pdf [ Google Scholar ]
• Gelfand M. J., Nishii L. H., Raver J. L., Schneider B. (2007). Discrimination in Organizations: An Organizational-Level Systems Perspective (CAHRS Working Paper #07-08). Ithaca, NY: Cornell University. Retrieved from Cornell University, School of Industrial and Labor Relations, Center for Advanced Human Resource Studies. [ Google Scholar ]
• Gerhart B., Rynes S. (1991). Determinants and consequences of salary negotiations by male and female MBA graduates. J. Appl. Psychol. 76 256–262. 10.1037/0021-9010.76.2.256 [ CrossRef ] [ Google Scholar ]
• Gilbert v. DaimlerChrysler Corp. (2004). 470 Mich. 749, 685 N.W.2d 391 2004 Lansing, MI: Supreme Court of Michigan. [ Google Scholar ]
• Glass J. (2004). Blessing or curse? Work-family policies and mother’s wage growth over time. Work Occup. 31 367–394. 10.1177/0730888404266364 [ CrossRef ] [ Google Scholar ]
• Glick P. (2013). “BS at work: how benevolent sexism undermines women and justifies backlash,” in Paper Presented at the Harvard Business School symposium Gender & Work: Challenging Conventional Wisdom Boston, MA. [ Google Scholar ]
• Glick P., Diebold J., Bailey-Werner B., Zhu L. (1997). The two faces of Adam: ambivalent sexism and polarized attitudes toward women. Pers. Soc. Psychol. Bull. 23 1323–1334. 10.1177/01461672972312009 [ CrossRef ] [ Google Scholar ]
• Glick P., Fiske S. T. (1996). The ambivalent sexism inventory: differentiating hostile and benevolent sexism. J. Pers. Soc. Psychol. 70 491–512. 10.1037/0022-3514.70.3.491 [ CrossRef ] [ Google Scholar ]
• Glick P., Fiske S. T. (2001). An ambivalent alliance: hostile and benevolent sexism as complementary justifications for gender inequality. Am. Psychol. 56 109–118. 10.1037/0003-066X.56.2.109 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Glick P., Fiske S. T., Mladinic A., Saiz J. L., Abrams D., Masser B., et al. (2000). Beyond prejudice as simple antipathy: hostile and benevolent sexism across culture. J. Pers. Soc. Psychol. 79 763–775. 10.1037/0022-3514.79.5.763 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Goldberg P. (1968). Are women prejudiced against women? Transaction 5 316–322. [ Google Scholar ]
• Goldenhar L. M., Swanson N. G., Hurrell J. J., Jr., Ruder A., Deddens J. (1998). Stressors and adverse outcomes for female construction workers. J. Occup. Health Psychol. 3 19–32. 10.1037/1076-8998.3.1.19 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Good J. J., Rudman L. A. (2010). When female applicants meet sexist interviewers: the costs of being a target of benevolent sexism. Sex Roles 62 481–493. 10.1007/s11199-009-9685-6 [ CrossRef ] [ Google Scholar ]
• Grandey A. A., Cropanzano R. (1999). The conservation of resources model applied to work-family conflict and strain. J. Vocat. Behav. 54 350–370. 10.1006/jvbe.1998.1666 [ CrossRef ] [ Google Scholar ]
• Grant R. M. (2010). Contemporary Strategy Analysis, Seventh Edition. New York, NY: Wiley. [ Google Scholar ]
• Greenhaus J. H., Beutell N. J. (1985). Sources of conflict between work and family roles. Acad. Manag. Rev. 10 76–88. 10.5465/AMR.1985.4277352 [ CrossRef ] [ Google Scholar ]
• Gutek B. A., Cohen A. G., Tsui A. (1996). Reactions to perceived sex discrimination. Hum. Relat. 49 791–813. 10.1177/001872679604900604 [ CrossRef ] [ Google Scholar ]
• Hagen R. L., Kahn A. (1975). Discrimination against competent women. J. Appl. Soc. Psychol. 5 362–376. 10.1111/j.1559-1816.1975.tb00688.x [ CrossRef ] [ Google Scholar ]
• Hakim C. (2006). Women, careers, and work-life preferences. Br. J. Guid. Counc. 34 279–294. 10.1080/03069880600769118 [ CrossRef ] [ Google Scholar ]
• Hartnell C. A., Walumbwa F. O. (2011). “Transformational leadership and organizational culture,” in The Handbook of Organizational Culture and Climate eds Ashkanasy N. M., Wilderom C. P. M., Peterson M. F. (Thousand Oaks, CA: Sage Publications; ) 225–248. [ Google Scholar ]
• Hebl M. R., King E. B., Glick P., Singletary S. L., Kazama S. (2007). Hostile and benevolent reactions toward pregnant women: complementary interpersonal punishments and rewards that maintain traditional roles. J. Appl. Psychol. 92 1499–1511. 10.1037/0021-9010.92.6.1499 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Heilman M. E. (1983). “Sex bias in work settings: the lack of fit model,” in Research in Organizational Behavior Vol. 5 eds Staw B., Cummings L. (Greenwich, CT: JAI press; ) 269–298. [ Google Scholar ]
• Heilman M. E. (1995). Sex stereotypes and their effects in the workplace: what we know and what we don’t know. J. Soc. Behav. Pers. 10 3–26. [ Google Scholar ]
• Heilman M. E. (2001). Description and prescription: how gender stereotypes prevent women’s ascent up the organizational ladder. J. Soc. Issues 57 657–674. 10.1111/0022-4537.00234 [ CrossRef ] [ Google Scholar ]
• Heilman M. E., Block C. J., Lucas J. A. (1992). Presumed incompetent? Stigmatization and affirmative action efforts. J. Appl. Psychol. 77 536–544. 10.1037/0021-9010.77.4.536 [ CrossRef ] [ Google Scholar ]
• Heilman M. E., Block C. J., Stathatos P. (1997). The affirmative action stigma of incompetence: effects of performance information ambiguity. Acad. Manag. J. 40 603–625. 10.2307/257055 [ CrossRef ] [ Google Scholar ]
• Heilman M., Okimoto T. G. (2007). Why are women penalized for success at male tasks? The implied communality deficit. J. Appl. Psychol. 92 81–92. 10.1037/0021-9010.92.1.81 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Heilman M., Okimoto T. G. (2008). Motherhood: a potential source of bias in employment decisions. J. Appl. Psychol. 93 189–198. 10.1037/0021-9010.93.1.189 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Hicks-Clarke D., Iles P. (2000). Climate for diversity and its effects on career and organisational attitudes and perceptions. Person. Rev. 29 324–345. 10.1108/00483480010324689 [ CrossRef ] [ Google Scholar ]
• Hitlan R. T., Pryor J. B., Hesson-McInnis M. S., Olson M. (2009). Antecedents of gender harassment: an analysis of person and situation factors. Sex Roles 61 794–807. 10.1007/s11199-009-9689-2 [ CrossRef ] [ Google Scholar ]
• Holland J. L. (1996). Exploring careers with a typology: what we have learned and some new directions. Am. Psychol. 51 397–406. 10.1037/0003-066X.51.4.397 [ CrossRef ] [ Google Scholar ]
• Hough L. M., Oswald F. L., Ployhart R. E. (2001). Determinants, detection, and amelioration of adverse impact in personnel selection procedures: issues, evidence, and lessons learned. Int. J. Sel. Assess. 9 152–194. 10.1111/1468-2389.00171 [ CrossRef ] [ Google Scholar ]
• House R. J., Aditya R. N. (1997). The social scientific study of leadership: quo vadis? J. Manag. 23 409–473. 10.1177/014920639702300306 [ CrossRef ] [ Google Scholar ]
• Huffman M. L., Velasco S. C. (1997). When more is less: sex composition, organizations, and earnings in U.S. firms . Work Occup. 24 214–244. 10.1177/0730888497024002005 [ CrossRef ] [ Google Scholar ]
• Hulin C. L., Fitzgerald L. F., Drasgow F. (1996). “Organizational influences on sexual harassment,” in Sexual Harassment in the Workplace: Perspectives, Frontiers, and Response Strategies. Women and Work: A Research and Policy Series Vol. 5 ed. Stockdale M. S. (Thousand Oaks, CA: Sage Publications; ) 127–150. [ Google Scholar ]
• Hunter J. E., Schmidt F. L., Jackson G. B. (1982). Meta-Analysis: Cumulating Research Findings Across Studies. Thousand Oaks, CA: Sage Publications. [ Google Scholar ]
• Hunter L. W., Bernhardt A., Hughes K. L., Skuratowicz E. (2001). It’s not just the ATMs: technology, firm strategies, jobs, and earnings in retail banking. Ind. Labor Relat. Rev. 54 402–424. 10.1177/001979390105400222 [ CrossRef ] [ Google Scholar ]
• Ittner C. D., Larcker D. F., Meyer M. W. (2003). Subjectivity and the weighting of performance measures: evidence from a balanced scorecard. Account. Rev. 78 725–758. 10.2308/accr.2003.78.3.725 [ CrossRef ] [ Google Scholar ]
• James E. H., Wooten L. P. (2006). Diversity crises: how firms manage discrimination lawsuits. Acad. Manag. J. 49 1103–1118. 10.5465/AMJ.2006.23478091 [ CrossRef ] [ Google Scholar ]
• Jayne M. E., Dipboye R. L. (2004). Leveraging diversity to improve business performance: research findings and recommendations for organizations. Hum. Resour. Manag. 43 409–424. 10.1002/hrm.20033 [ CrossRef ] [ Google Scholar ]
• Jung D., Wu A., Chow C. W. (2008). Towards understanding the direct and indirect effects of CEOs’ transformational leadership on firm innovation. Leadersh. Q. 19 582–594. 10.1016/j.leaqua.2008.07.007 [ CrossRef ] [ Google Scholar ]
• Kanter R. M. (1977). Men and Women of the Corporation. New York, NY: Basic Books. [ Google Scholar ]
• Kath L. M., Swody C. A., Magley V. J., Bunk J. A., Gallus J. A. (2009). Cross-level, three-way interactions among work-group climate, gender, and frequency of harassment on morale and withdrawal outcomes of sexual harassment. J. Occup. Organ. Psychol. 82 159–182. 10.1348/096317908X299764 [ CrossRef ] [ Google Scholar ]
• Kawakami K., Dion K. L. (1993). The impact of salient self-identities on relative deprivation and action intentions. Eur. J. Soc. Psychol. 23 525–540. 10.1002/ejsp.2420230509 [ CrossRef ] [ Google Scholar ]
• Kay A. C., Jost J. T., Young S. (2005). Victim derogation and victim enhancement as alternate routes to system justification. Psychol. Sci. 16 240–246. 10.1111/j.0956-7976.2005.00810.x [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Kelly E. L., Kalev A. (2006). Managing flexible work arrangements in US organizations: formalized discretion or a ‘right to ask’. Soc. Econ. Rev. 4 379–416. 10.1093/ser/mwl001 [ CrossRef ] [ Google Scholar ]
• King E. B., Botsford W., Hebl M. R., Kazama S., Dawson J. F., Perkins A. (2012). Benevolent sexism at work: gender differences in the distribution of challenging developmental experiences. J. Manag. 38 1835–1866. 10.1177/0149206310365902 [ CrossRef ] [ Google Scholar ]
• Klandermans B. (1997). The Social Psychology of Protest. Oxford: Basic Blackwell. [ Google Scholar ]
• Kohn M. L., Schooler C. (1982). Job conditions and personality: a longitudinal assessment of their reciprocal effects. Am. J. Sociol. 87 1257–1286. 10.1086/227593 [ CrossRef ] [ Google Scholar ]
• Konrad A. M., Cannings K., Goldberg C. B. (2010). Asymmetrical demography effects on psychological climate for gender diversity: differential effects of leader gender and work unit gender composition among Swedish doctors. Hum. Relat. 63 1661–1685. 10.1177/0018726710369397 [ CrossRef ] [ Google Scholar ]
• Konrad A. M., Linnehan F. (1995). Formalized HRM structures: coordinating equal employment opportunity or concealing organizational practices? Acad. Manag. J. 38 787–820. 10.2307/256746 [ CrossRef ] [ Google Scholar ]
• Kozlowski S. W., Doherty M. L. (1989). Integration of climate and leadership: examination of a neglected issue. J. Appl. Psychol. 74 546–553. 10.1037/0021-9010.74.4.546 [ CrossRef ] [ Google Scholar ]
• Kristof-Brown A. L., Zimmerman R. D., Johnson E. C. (2005). Consequences of individuals’ fit at work: a meta-analysis of person-job, person-organization, person-group, and person-supervisor fit. Pers. Psychol. 58 281–342. 10.1111/j.1744-6570.2005.00672.x [ CrossRef ] [ Google Scholar ]
• Kulik C. T., Bainbridge H. T. J. (2005). “Psychological perspectives on workplace diversity,” in Handbook of Workplace Diversity eds Konrad A. M., Prasad P., Pringle J. K. (Thousand Oaks, CA: Sage Publications; ) 25–52. [ Google Scholar ]
• Lamont M., Mizrachi N. (2012). Ordinary people doing extraordinary things: responses to stigmatization in comparative perspective. Ethn. Racial Stud. 35 365–381. 10.1080/01419870.2011.589528 [ CrossRef ] [ Google Scholar ]
• Lazear E. P., Rosen S. (1990). Male-female wage differentials in job ladders. J. Labor Econ. 8 106–123. 10.1016/j.sapharm.2011.06.003 [ CrossRef ] [ Google Scholar ]
• Leskinen E. A., Cortina L. M., Kabat D. B. (2011). Gender harassment: broadening our understanding of sex-based harassment at work. Law Hum. Behav. 35 25–39. 10.1007/s10979-010-9241-5 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Levitin T., Quinn R. P., Staines G. L. (1971). Sex discrimination against the American working woman. Am. Behav. Sci. 15 237–254. 10.1177/000276427101500207 [ CrossRef ] [ Google Scholar ]
• Lewin K. (1935). A Dynamic Theory of Personality. New York, NY: McGraw-Hill. [ Google Scholar ]
• Lewin K. (1951). Field Theory in Social Science: Selected Theoretical Papers. Oxford: Harpers. [ Google Scholar ]
• LexisNexis. (2010). Mealy’s Daily News Update: Pharmaceutical Firm Settles Gender Bias Class Claims for $ 175 Million. Available at: http://www.lexis.com [accessed July 15, 2010] [ Google Scholar ]
• Linehan M., Scullion H. (2008). The development of female global managers: the role of mentoring and networking. J. Bus. Ethics 83 29–40. 10.1007/s10551-007-9657-0 [ CrossRef ] [ Google Scholar ]
• Lips H. M. (2003). The gender pay gap: concrete indicator of women’s progress toward equality. Anal. Soc. Issues Public Policy 3 87–109. 10.1111/j.1530-2415.2003.00016.x [ CrossRef ] [ Google Scholar ]
• Lyness K. S., Thompson D. E. (1997). Above the glass ceiling? A comparison of matched samples of female and male executives. J. Appl. Psychol. 82 359–375. 10.1037/0021-9010.82.3.359 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Major B. (1994). From social inequality to personal entitlement: the role of social comparisons, legitimacy appraisals, and group membership. Adv. Exp. Soc. Psychol. 26 293–355. 10.1016/S0065-2601(08)60156-2 [ CrossRef ] [ Google Scholar ]
• Major B., Quinton W. J., Schmader T. (2003). Attributions to discrimination and self-esteem: impact of group identification and situational ambiguity. J. Exp. Soc. Psychol. 39 220–231. 10.1016/S0022-1031(02)00547-4 [ CrossRef ] [ Google Scholar ]
• Major B., Schmader T. (2001). “Legitimacy and the construal of social disadvantage,” in The Psychology of Legitimacy: Emerging Perspectives on Ideology, Justice, and Intergroup Relations eds Jost J. T., Major B. (Cambridge: Cambridge University Press; ) 176–200. [ Google Scholar ]
• Manne A. (2001). Women’s preferences, fertility and family policy: the case for diversity. People Place 9 6–25. [ Google Scholar ]
• Mansfield P., Koch P., Henderson J., Vicary J., Cohn M., Young E. (1991). The job climate for women in traditionally male blue-collar occupations. Sex Roles 25 63–79. 10.1007/BF00289317 [ CrossRef ] [ Google Scholar ]
• Marini M. M. (1989). Sex differences in earnings in the United States. Annu. Rev. Sociol. 15 343–380. 10.1146/annurev.so.15.080189.002015 [ CrossRef ] [ Google Scholar ]
• Martell R. F. (1991). Sex bias at work: the effects of attentional and memory demands on performance ratings of men and women. J. Appl. Soc. Psychol. 21 1939–1960. 10.1111/j.1559-1816.1991.tb00515.x [ CrossRef ] [ Google Scholar ]
• Martell R. F., Lane D. M., Emrich C. (1996). Male-female differences: a computer simulation. Am. Psychol. 51 157–158. 10.1037/0003-066X.51.2.157 [ CrossRef ] [ Google Scholar ]
• Martins L. L., Eddleston K. A., Veiga J. F. (2002). Moderators of the relationship between work-family conflict and career satisfaction. Acad. Manag. J. 45 399–409. 10.2307/3069354 [ CrossRef ] [ Google Scholar ]
• Martins L. L., Parsons C. K. (2007). Effects of gender diversity management on perceptions of organizational attractiveness: the role of individual differences in attitudes and beliefs. J. Appl. Psychol. 92 865–875. 10.1037/0021-9010.92.3.865 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Masser B. M., Abrams D. (2004). Reinforcing the glass ceiling: the consequences of hostile sexism for female managerial candidates. Sex Roles 51 609–615. 10.1007/s11199-004-5470-8 [ CrossRef ] [ Google Scholar ]
• Maume D. J. (1999). Glass ceilings and glass escalators: occupational segregation and race and sex differences in managerial promotions. Work Occup. 26 483–509. 10.1177/0730888499026004005 [ CrossRef ] [ Google Scholar ]
• McDonald P., Backstrom S., Dear K. (2008). Reporting sexual harassment: claims and remedies. Asia Pac. J. Hum. Resour. 46 173–195. 10.1177/1038411108091757 [ CrossRef ] [ Google Scholar ]
• McDonald S., Lin N., Ao D. (2009). Networks of Opportunity: gender, race, and job leads. Soc. Probl. 56 385–402. 10.1525/sp.2009.56.3.385 [ CrossRef ] [ Google Scholar ]
• McGinley A. (2007). Babes and beefcake: exclusive hiring arrangements and sexy dress codes. Duke J. Gend. Law Policy 14 257–283. [ Google Scholar ]
• McIntyre S., Moberg D. J., Posner B. Z. (1980). Preferential treatment in preselection decisions according to sex and race. Acad. Manag. J. 23 738–749. 10.2307/255560 [ CrossRef ] [ Google Scholar ]
• McLaughlin H., Uggen C., Blackstone A. (2012). Sexual harassment, workplace authority, and the paradox of power. Am. Sociol. Rev. 77 625–647. 10.1177/0003122412451728 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Miner-Rubino K., Cortina L. M. (2004). Working in a context of hostility toward women: implications for employees’ well-being. J. Occup. Health Psychol. 9 107–122. 10.1037/1076-8998.9.2.107 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Mitchell D., Hirschman R., Angelone D. J., Lilly R. S. (2004). A laboratory analogue for the study of peer sexual harassment. Psychol. Women Q. 28 194–203. 10.1111/j.1471-6402.2004.00136.x [ CrossRef ] [ Google Scholar ]
• Montañés P., de Lemus S., Bohner G., Megías J. L., Moya M., Garcia-Retamero R. (2012). Intergenerational transmission of benevolent sexism from mothers to daughters and its relation to daughters’ academic performance and goals. Sex Roles 66 468–478. 10.1007/s11199-011-0116-0 [ CrossRef ] [ Google Scholar ]
• Moore v. Alabama State University. (1996). 980 F. Supp. 426 (M.D. Ala. 1996) M.D. Alabama: United States District Court. [ Google Scholar ]
• Morgan W. B., Walker S. S., Hebl M. R., King E. B. (2013). A field experiment: reducing interpersonal discrimination toward pregnant job applicants. J. Appl. Psychol. 98 799–809. 10.1037/a0034040 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Morrison A. M., Von Glinow M. A. (1990). Women and minorities in management. Am. Psychol. 45 200–208. 10.1037/0003-066X.45.2.200 [ CrossRef ] [ Google Scholar ]
• Moss-Racusin C. A., Dovidio J. F., Brescoll V. L., Graham M. J., Handelsman J. (2012). Science faculty’s subtle gender biases favor male students. PNAS 109 16474–16479. 10.1073/pnas.1211286109 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Murphy K. R., Cleveland J. (1995). Understanding Performance Appraisal: Social, Organizational, and Goal-Based Perspectives. Thousand Oaks, CA: Sage Publications. [ Google Scholar ]
• Murrell A. J., Olson J. E., Frieze I. H. (1995). Sexual harassment and gender discrimination: a longitudinal study of women managers. J. Soc. Issues 51 139–149. 10.1111/j.1540-4560.1995.tb01313.x [ CrossRef ] [ Google Scholar ]
• Neumark D. (1996). Sex discrimination in restaurant hiring: an audit study. Q. J. Econ. 111 915–942. 10.2307/2946676 [ CrossRef ] [ Google Scholar ]
• Nishii L. H., Raver J. L. (2003). “Collective climates for diversity: evidence from a field study,” in Paper Presented at the Annual Conference of the Society for Industrial and Organizational Psychology Orlando, FL. [ Google Scholar ]
• Noe R. A. (1988). Women and mentoring: a review and research agenda. Acad. Manag. Rev. 13 65–78. 10.5465/AMR.1988.4306784 [ CrossRef ] [ Google Scholar ]
• Offermann L. R., Malamut A. B. (2002). When leaders harass: the impact of target perceptions of organizational leadership and climate on harassment reporting and outcomes. J. Appl. Psychol. 87 885–893. 10.1037/0021-9010.87.5.885 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Olian J. D., Schwab D. P., Haberfeld Y. (1988). The impact of applicant gender compared to qualifications on hiring recommendations: a meta-analysis of experimental studies. Organ. Behav. Hum. Decis. Process. 41 180–195. 10.1016/0749-5978(88)90025-8 [ CrossRef ] [ Google Scholar ]
• Ostroff C., Atwater L. E. (2003). Does whom you work with matter? Effects of referent group gender and age composition on managers’ compensation. J. Appl. Psychol. 88 725–740. 10.1037/0021-9010.88.4.725 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ostroff C., Kinicki A. J., Muhammad R. S. (2012). “Organizational culture and climate,” in Handbook of Psychology, Industrial and Organizational Psychology 2nd Edn Vol. 12 eds Schmitt N. W., Highhouse S. (New York, NY: Wiley and Sons; ) 643–676. [ Google Scholar ]
• Ostroff C., Rothausen T. J. (1997). The moderating effect of tenure in person—environment fit: a field study in educational organizations. J. Occup. Organ. Psychol. 70 173–188. 10.1111/j.2044-8325.1997.tb00641.x [ CrossRef ] [ Google Scholar ]
• Perlow L. A. (1995). Putting the work back into work/family. Group Organ. Manag. 20 227–239. 10.1177/1059601195202009 [ CrossRef ] [ Google Scholar ]
• Perry E. L., Davis-Blake A., Kulik C. T. (1994). Explaining gender-based selection decisions: a synthesis of contextual and cognitive approaches. Acad. Manag. Rev. 19 786–820. 10.5465/AMR.1994.9412190219 [ CrossRef ] [ Google Scholar ]
• Peterson T., Morgan L. A. (1995). Separate and unequal: occupation-establishment sex segregation and gender wage gap. Am. J. Sociol. 101 329–365. 10.1086/230727 [ CrossRef ] [ Google Scholar ]
• Pryor J. B. (1995). The psychosocial impact of sexual harassment on women in the US Military. Basic Appl. Soc. Psychol. 17 581–603. 10.1207/s15324834basp1704_9 [ CrossRef ] [ Google Scholar ]
• Pryor J. B., Giedd J. L., Williams K. B. (1995). A social psychological model for predicting sexual harassment. J. Soc. Issues 51 69–84. 10.1111/j.1540-4560.1995.tb01309.x [ CrossRef ] [ Google Scholar ]
• Pryor J. B., LaVite C. M., Stoller L. M. (1993). A social and psychological analysis of sexual harassment: the person/situation interaction. J. Vocat. Behav. 42 68–83. 10.1006/jvbe.1993.1005 [ CrossRef ] [ Google Scholar ]
• Ragins B. R., Cornwall J. M. (2001). Pink triangles: antecedents and consequences of perceived workplace discrimination against gay and lesbian employees. J. Appl. Psychol. 86 1244–1261. 10.1037/0021-9010.86.6.1244 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ragins B. R., Sundstrom E. (1989). Gender and power in organizations: a longitudinal perspective. Psychol. Bull. 105 51–88. 10.1037/0033-2909.105.1.51 [ CrossRef ] [ Google Scholar ]
• Reilly K. T., Wirjanto T. S. (1999). Does more mean less? The male/female wage gap and the proportion of females at the establishment level. Can. J. Econ. 32 906–929. 10.2307/136410 [ CrossRef ] [ Google Scholar ]
• Reskin B. F., McBrier D. B. (2000). Why not ascription? Organizations’ employment of male and female managers. Am. Sociol. Rev. 65 210–233. [ Google Scholar ]
• Roberson L., Galvin B. M., Charles A. C. (2007). When group identities matter: bias in performance appraisal. Acad. Manag. Ann. 1 617–650. 10.1080/078559818 [ CrossRef ] [ Google Scholar ]
• Rosen B., Jerdee T. H. (1974). Effects of applicant’s sex and difficulty of job on evaluations of candidates for management positions. J. Appl. Psychol. 59 511–512. 10.1037/h0037323 [ CrossRef ] [ Google Scholar ]
• Roth P. L., Purvis K. L., Bobko P. (2012). A meta-analysis of gender group differences for measures of job performance in field studies. J. Manag. 38 719–739. 10.1177/0149206310374774 [ CrossRef ] [ Google Scholar ]
• Rudman L. A. (1998). Self-promotion as a risk factor for women: the costs and benefits of counterstereotypical impression management. J. Pers. Soc. Psychol. 74 629–645. 10.1037/0022-3514.74.3.629 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Rudman L. A., Kilianski S. E. (2000). Implicit and explicit attitudes toward female authority. Pers. Soc. Psychol. Bull. 26 1315–1328. 10.1177/0146167200263001 [ CrossRef ] [ Google Scholar ]
• Rudman L. A., Phelan J. E. (2008). Backlash effects for disconfirming gender stereotypes in organizations. Res. Organ. Behav. 28 61–79. 10.1016/j.riob.2008.04.003 [ CrossRef ] [ Google Scholar ]
• Rudman L. A., Moss-Racusin C. A., Phelan J. E., Nauts S. (2012). Status incongruity and backlash effects: defending the gender hierarchy motivates prejudice against female leaders. J. Exp. Soc. Psychol. 48 165–179. 10.1016/j.jesp.2011.10.008 [ CrossRef ] [ Google Scholar ]
• Ruggiero K. M., Taylor D. M. (1995). Coping with discrimination: how disadvantaged group members perceive the discrimination that confronts them. J. Pers. Soc. Psychol. 68 826–838. 10.1037/0022-3514.68.5.826 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Russell B. L., Trigg K. Y. (2004). Tolerance of sexual harassment: an examination of gender differences, ambivalent sexism, social dominance, and gender roles. Sex Roles 50 565–573. 10.1023/B:SERS.0000023075.32252.fd [ CrossRef ] [ Google Scholar ]
• Rynes S., Rosen B. (1995). A field survey of factors affecting the adoption and perceived success of diversity training. Pers. Psychol. 48 247–270. 10.1111/j.1744-6570.1995.tb01756.x [ CrossRef ] [ Google Scholar ]
• Salvaggio A. N., Streich M., Hopper J. E. (2009). Ambivalent sexism and applicant evaluations: effects on ambiguous applicants. Sex Roles 61 621–633. 10.1007/s11199-009-9640-6 [ CrossRef ] [ Google Scholar ]
• Sanchez J. I., De La Torre P. (1996). A second look at the relationship between rating and behavioral accuracy in performance appraisal. J. Appl. Psychol. 81 3–10. 10.1037/0021-9010.81.1.3 [ CrossRef ] [ Google Scholar ]
• Schein E. H. (2010). Organizational Culture and Leadership , Vol. 2 New York, NY: Jossey-Bass [ Google Scholar ]
• Schmader T., Johns M., Forbes C. (2008). An integrated process model of stereotype threat effects on performance. Psychol. Rev. 115 336–356. 10.1037/0033-295X.115.2.336 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Schmitt M. T., Branscombe N. R., Kobrynowicz D., Owen S. (2002). Perceiving discrimination against one’s gender group has different implications for well-being in women and men. Pers. Soc. Psychol. Bull. 28 197–210. 10.1177/0146167202282006 [ CrossRef ] [ Google Scholar ]
• Schmitt M. T., Branscombe N. R., Postmes T., Garcia A. (2014). The consequences of perceived discrimination for psychological well-being: a meta-analytic review. Psychol. Bull. 140 921–948. 10.1037/a0035754 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Schneider B. (1975). Organizational climates: an essay. Pers. Psychol. 28 447–479. 10.1111/j.1744-6570.1975.tb01386.x [ CrossRef ] [ Google Scholar ]
• Schneider B. (1983). “The attraction–selection–attrition framework,” in Organizational Effectiveness: A Comparison of Multiple Models eds Cameron K. S., Whetten D. A. (New York, NY: Academic Press; ). [ Google Scholar ]
• Schneider B. (1987). The people make the place. Pers. Psychol. 40 437–453. 10.1111/j.1744-6570.1987.tb00609.x [ CrossRef ] [ Google Scholar ]
• Schneider B., Ehrhart M. G., Macey W. H. (2011). “Organizational climate research: achievements and the road ahead,” in The Handbook of Organizational Culture and Climate 2nd Edn eds Ashkanasy N. M., Wilderom C. P. M., Peterson M. F. (Thousand Oaks, CA: Sage; ) 29–49. [ Google Scholar ]
• Schneider K. T., Swan S., Fitzgerald L. F. (1997). Job-related and psychological effects of sexual harassment in the workplace: empirical evidence from two organizations. J. Appl. Psychol. 82 401–415. 10.1037/0021-9010.82.3.401 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Schneyer K. L. (1998). Hooting: public and popular discourse about sex discrimination. Univ. Mich. J. Law Reform 31 551–636. [ Google Scholar ]
• Schwartz M. D., DeKeseredy W. S. (2000). Aggregation bias and woman abuse: variations by male peer support, region, language, and school type. J. Interpersh. Violence 15 555–565. 10.1177/088626000015006001 [ CrossRef ] [ Google Scholar ]
• Sheltzer J. M., Smith J. C. (2014). Elite male faculty in the life sciences employ fewer women. PNAS 111 10107–10112. 10.1073/pnas.1403334111 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Smith P. C., Kendall L. M. (1963). Retranslation of expectations: an approach to the construction of unambiguous anchors for rating scales. J. Appl. Psychol. 47 149–155. 10.1037/h0047060 [ CrossRef ] [ Google Scholar ]
• Snyder K. A., Green A. I. (2008). Revisiting the glass escalator: the case of gender segregation in a female dominated occupation. Soc. Probl. 55 271–299. 10.1525/sp.2008.55.2.271 [ CrossRef ] [ Google Scholar ]
• Son Hing L. S. (2012). Responses to stigmatization: the moderating roles of primary and secondary appraisals. Du Bois Rev. 9 149–168. 10.10170/S1742058X11000592 [ CrossRef ] [ Google Scholar ]
• Son Hing L. S., Bobocel D. R., Zanna M. P. (2002). Meritocracy and opposition to affirmative action: making concessions in the face of discrimination. J. Pers. Soc. Psychol. 83 493–509. 10.1037/0022-3514.83.3.493 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Son Hing L. S., Bobocel D. R., Zanna M. P., Garcia D. M., Gee S. S., Orazietti K. (2011). The merit of meritocracy. J. Pers. Soc. Psychol. 101 433–450. 10.1037/a0024618 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Son Hing L. S., Chung-Yan G. A., Hamilton L. K., Zanna M. P. (2008). A two-dimensional model that employs explicit and implicit attitudes to characterize prejudice. J. Pers. Soc. Psychol. 94 971–987. 10.1037/0022-3514.94.6.971 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Son Hing L. S., Zanna M. P. (2010). “Individual differences,” in The Sage Handbook of Prejudice, Stereotyping and Discrimination eds Dovidio J. F., Hewstone M., Glick P., Esses V. (London: SAGE Publications Ltd.) 163–179. [ Google Scholar ]
• Spencer S. J., Steele C. M., Quinn D. M. (1999). Stereotype threat and women’s math performance. J. Exp. Soc. Psychol. 35 4–28. 10.1006/jesp.1998.1373 [ CrossRef ] [ Google Scholar ]
• Spreitzer G. M., McCall M. W., Mahoney J. D. (1997). Early identification of international executive potential. J. Appl. Psychol. 82 6–29. 10.1037/0021-9010.82.1.6 [ CrossRef ] [ Google Scholar ]
• Stangor C., Lynch L., Duan C., Glass B. (1992). Categorization of individuals on the basis of multiple social features. J. Pers. Soc. Psychol. 62 207–218. 10.1037/0022-3514.62.2.207 [ CrossRef ] [ Google Scholar ]
• Stangor C., Swim J. K., Sechrist G. B., DeCoster J., Van Allen K. L., Ottenbreit A. (2003). Ask, answer, and announce: three stages in perceiving and responding to discrimination. Eur. Rev. Soc. Psychol. 14 277–311. 10.1080/10463280340000090 [ CrossRef ] [ Google Scholar ]
• Steinpreis R. E., Anders K. A., Ritzke D. (1999). The impact of gender on the review of the curricula vitae of job applicants and tenure candidates: a national empirical study. Sex Roles 41 509–528. 10.1023/A:1018839203698 [ CrossRef ] [ Google Scholar ]
• Stone P., Lovejoy M. (2004). Fast-track women and the “choice” to stay home. Ann. Am. Acad. Polit. Soc. Sci. 596 62–83. 10.1177/0002716204268552 [ CrossRef ] [ Google Scholar ]
• Stroh L. K., Brett J. M., Reilly A. H. (1992). All the right stuff: a comparison of female and male managers’ career progression. J. Appl. Psychol. 77 251–260. 10.1037/0021-9010.77.3.251 [ CrossRef ] [ Google Scholar ]
• Swidler A. (1986). Culture in action: symbols and strategies. Am. Sociol. Rev. 51 273–286. 10.2307/2095521 [ CrossRef ] [ Google Scholar ]
• Thomas-Hunt M. C., Phillips K. W. (2004). When what you know is not enough: expertise and gender dynamics in task groups. Pers. Soc. Psychol. Bull. 30 1585–1598. 10.1177/0146167204271186 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Tosi H. L., Einbender S. W. (1985). The effects of the type and amount of information in sex discrimination research: a meta-analysis. Acad. Manag. J. 28 712–723. 10.2307/256127 [ CrossRef ] [ Google Scholar ]
• Triana M. D. C., García M. F. (2009). Valuing diversity: a group-value approach to understanding the importance of organizational efforts to support diversity. J. Organ. Behav. 30 941–962. 10.1002/job.598 [ CrossRef ] [ Google Scholar ]
• Triana M. D. C., García M. F., Colella A. (2010). Managing diversity: how organizational efforts to support diversity moderate the effects of perceived racial discrimination on affective commitment. Pers. Psychol. 63 817–843. 10.1111/j.1744-6570.2010.01189.x [ CrossRef ] [ Google Scholar ]
• Trice H. M., Beyer J. M. (1993). The Cultures of Work Organizations. Englewood Cliffs, NJ: Prentice-Hall. [ Google Scholar ]
• Tsui A. S., O’Reilly C. A. (1989). Beyond simple demographic effects: the importance of relational demography in superior-subordinate dyads. Acad. Manag. J. 32 402–423. 10.2307/256368 [ CrossRef ] [ Google Scholar ]
• U.S. Census Bureau. (2007). Statistical Abstract of the United States: 2007. Washington, DC. Available at: http://www.census.gov/prod/www/statistical_abstract.html [ Google Scholar ]
• U.S. Government Accountability Office. (2003). Women’s Earnings: Work Patterns Partially Explain Difference Between Men’s and Women’s Earnings (GAO-04-35). Available at: http://www.gao.gov [ Google Scholar ]
• van Knippenberg D., De Dreu C. K., Homan A. C. (2004). Work group diversity and group performance: an integrative model and research agenda. J. Appl. Psychol. 89 1008–1022. 10.1037/0021-9010.89.6.1008 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• van Zomeren M., Postmes T., Spears R. (2008). Toward an integrative social identity model of collective action: a qualitative research synthesis of three socio-psychological perspectives. Psychol. Bull. 134 505–535. 10.1037/0033-2909.134.4.504 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Velez et al. v. Novartis Pharmaceuticals Crop et al. (2010). No. 04 Civ. 9194 (S.D. N.Y., May 19, 2010). [ Google Scholar ]
• Vescio T. K., Gervais S. J., Snyder M., Hoover A. (2005). Power and the creation of patronizing environments: the stereotype-based behaviors of the powerful and their effects on female performance in masculine domains. J. Pers. Soc. Psychol. 88 658–672. 10.1037/0022-3514.88.4.658 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Vorauer J. D., Kumhyr S. M. (2001). Is this about you or me? Self-versus other-directed judgments and feelings in response to intergroup interaction. Pers. Soc. Psychol. Bull. 27 706–719. 10.1177/0146167201276006 [ CrossRef ] [ Google Scholar ]
• Wal-Mart Stores Inc. v. Dukes. (2004/2011). 222 F.R.D. 137 (N.D. Cal. 2001) aff’d 509 F.3d. 1168 (9C 2007) aff’d 603 F.3d 571 (9C 2010) rev’d 564 U.S. ___ (2011), Docket No. 10–277. [ Google Scholar ]
• Welle B., Heilman M. E. (2005). “Formal and informal discrimination against women at work: the role of gender stereotypes,” in Research in Social Issues in Management eds Steiner D., Gilliland S. W., Skarlicki D. (Westport, CT: Information Age Publishers; ) 23–40. [ Google Scholar ]
• Westall R. (2015). Restaurant Dress Codes Open to Sexual Discrimination Complaints. CBC. Available at: http://www.cbc.ca/news/canada/restaurant-dress-codes-open-to-sexual-discrimination-complaints-1.3012522 [accessed March 31, 2015] [ Google Scholar ]
• Williams C. L. (1992). The glass escalator: hidden advantages for men in the “female” professions. Soc. Probl. 39 253–267. 10.2307/3096961 [ CrossRef ] [ Google Scholar ]
• Williams J. C. (2003). Beyond the glass ceiling: the maternal wall as a barrier to gender equality. T. Jefferson L. Rev. 26 1–14. [ Google Scholar ]
• Willness C. R., Steel P., Lee K. (2007). A meta-analysis of the antecedents and consequences of workplace sexual harassment. Pers. Psychol. 60 127–162. 10.1111/j.1744-6570.2007.00067.x [ CrossRef ] [ Google Scholar ]
• Word C. O., Zanna M. P., Cooper J. (1974). The nonverbal mediation of self-fulfilling prophecies in interracial interaction. J. Exp. Soc. Psychol. 10 109–120. 10.1016/0022-1031(74)90059-6 [ CrossRef ] [ Google Scholar ]
• Wright P., Ferris S. P., Hiller J. S., Kroll M. (1995). Competitiveness through management of diversity: effects on stock price valuation. Acad. Manag. J. 38 272–287. 10.2307/256736 [ CrossRef ] [ Google Scholar ]
• Wright P. M., McMahan G. C., McWilliams A. (1994). Human resources and sustained competitive advantage: a resource-based perspective. Int. J. Hum. Resour. Manag. 5 301–326. 10.1080/09585199400000020 [ CrossRef ] [ Google Scholar ]
• Wright S. C. (2001). “Strategic collective action: social psychology and social change,” in Blackwell Handbook of Social Psychology: Intergroup Processes eds Brown R., Gaertner S. (Oxford: Blackwell Publishing; ) 409–430. 10.1002/9780470693421.ch20 [ CrossRef ] [ Google Scholar ]
• Wright S. C., Lubensky M. E. (2009). “The struggle for social equality: collective action versus prejudice reduction,” in Intergroup Misunderstandings: Impact of Divergent Social Realities eds Demoulin S., Leyens J. P., Dovidio J. F. (New York, NY: Psychology Press; ) 291–310. [ Google Scholar ]

• USF Research
• USF Libraries

Digital Commons @ USF > College of Arts and Sciences > Women's and Gender Studies > Theses and Dissertations

Women's and Gender Studies Theses and Dissertations

Theses/dissertations from 2023 2023.

Social Media and Women Empowerment in Nigeria: A Study of the #BreakTheBias Campaign on Facebook , Deborah Osaro Omontese

Theses/Dissertations from 2022 2022

Going Flat: Challenging Gender, Stigma, and Cure through Lesbian Breast Cancer Experience , Beth Gaines

Incorrect Athlete, Incorrect Woman: IOC Gender Regulations and the Boundaries of Womanhood in Professional Sports , Sabeehah Ravat

Transnational Perspectives on the #MeToo and Anti-Base Movements in Japan , Alisha Romano

Theses/Dissertations from 2021 2021

Criminalizing LGBTQ+ Jamaicans: Social, Legal, and Colonial Influences on Homophobic Policy , Zoe C. Knowles

Dismantling Hegemony through Inclusive Sexual Health Education , Lauren Wright

Theses/Dissertations from 2020 2020

Transfat Representation , Jessica "Fyn" Asay

Theses/Dissertations from 2019 2019

Ain't I a Woman, Too? Depictions of Toxic Femininity, Transmisogynoir, and Violence on STAR , Sunahtah D. Jones

“The Most Muscular Woman I Have Ever Seen”: Bev FrancisPerformance of Gender in Pumping Iron II: The Women , Cera R. Shain

"Roll" Models: Fat Sexuality and Its Representations in Pornographic Imagery , Leah Marie Turner

Theses/Dissertations from 2018 2018

Reproducing Intersex Trouble: An Analysis of the M.C. Case in the Media , Jamie M. Lane

Race and Gender in (Re)integration of Victim-Survivors of CSEC in a Community Advocacy Context , Joshlyn Lawhorn

Penalizing Pregnancy: A Feminist Legal Studies Analysis of Purvi Patel's Criminalization , Abby Schneller

A Queer and Crip Grotesque: Katherine Dunn's , Megan Wiedeman

Theses/Dissertations from 2017 2017

"Mothers like Us Think Differently": Mothers' Negotiations of Virginity in Contemporary Turkey , Asli Aygunes

Surveilling Hate/Obscuring Racism?: Hate Group Surveillance and the Southern Poverty Law Center's "Hate Map" , Mary McKelvie

“Ya I have a disability, but that’s only one part of me”: Formative Experiences of Young Women with Physical Disabilities , Victoria Peer

Resistance from Within: Domestic violence and rape crisis centers that serve Black/African American populations , Jessica Marie Pinto

(Dis)Enchanted: (Re)constructing Love and Creating Community in the , Shannon A. Suddeth

Theses/Dissertations from 2016 2016

"The Afro that Ate Kentucky": Appalachian Racial Formation, Lived Experience, and Intersectional Feminist Interventions , Sandra Louise Carpenter

“Even Five Years Ago this Would Have Been Impossible:” Health Care Providers’ Perspectives on Trans* Health Care , Richard S. Henry

Tough Guy, Sensitive Vas: Analyzing Masculinity, Male Contraceptives & the Sexual Division of Labor , Kaeleen Kosmo

Theses/Dissertations from 2015 2015

Let’s Move! Biocitizens and the Fat Kids on the Block , Mary Catherine Dickman

Interpretations of Educational Experiences of Women in Chitral, Pakistan , Rakshinda Shah

Theses/Dissertations from 2014 2014

Incredi-bull-ly Inclusive?: Assessing the Climate on a College Campus , Aubrey Lynne Hall

Her-Storicizing Baldness: Situating Women's Experiences with Baldness from Skin and Hair Disorders , Kasie Holmes

In the (Radical) Pursuit of Self-Care: Feminist Participatory Action Research with Victim Advocates , Robyn L. Homer

Theses/Dissertations from 2013 2013

Significance is Bliss: A Global Feminist Analysis of the Liberian Truth and Reconciliation Commission and its Privileging of Americo-Liberian over Indigenous Liberian Women's Voices , Morgan Lea Eubank

Monsters Under the Bed: An Analysis of Torture Scenes in Three Pixar Films , Heidi Tilney Kramer

Theses/Dissertations from 2012 2012

Can You Believe She Did THAT?!:Breaking the Codes of "Good" Mothering in 1970s Horror Films , Jessica Michelle Collard

Don't Blame It on My Ovaries: Exploring the Lived Experience of Women with Polycystic Ovarian Syndrome and the Creation of Discourse , Jennifer Lynn Ellerman

Valanced Voices: Student Experiences with Learning Disabilities & Differences , Zoe DuPree Fine

An Interactive Guide to Self-Discovery for Women , Elaine J. Taylor

Selling the Third Wave: The Commodification and Consumption of the Flat Track Roller Girl , Mary Catherine Whitlock

Theses/Dissertations from 2010 2010

Beyond Survival: An Exploration of Narrative Healing and Forgiveness in Healing from Rape , Heather Curry

Theses/Dissertations from 2009 2009

Gender Trouble In Northern Ireland: An Examination Of Gender And Bodies Within The 1970s And 1980s Provisional Irish Republican Army In Northern Ireland , Jennifer Earles

"You're going to Hollywood"!: Gender and race surveillance and accountability in American Idol contestant's performances , Amanda LeBlanc

From the academy to the streets: Documenting the healing power of black feminist creative expression , Tunisia L. Riley

Developing Feminist Activist Pedagogy: A Case Study Approach in the Women's Studies Department at the University of South Florida , Stacy Tessier

Women in Wargasm: The Politics of Womenís Liberation in the Weather Underground Organization , Cyrana B. Wyker

Theses/Dissertations from 2008 2008

Opportunities for Spiritual Awakening and Growth in Mothering , Melissa J. Albee

A Constant Struggle: Renegotiating Identity in the Aftermath of Rape , Jo Aine Clarke

I am Warrior Woman, Hear Me Roar: The Challenge and Reproduction of Heteronormativity in Speculative Television Programs , Leisa Anne Clark

Theses/Dissertations from 2007 2007

Reforming Dance Pedagogy: A Feminist Perspective on the Art of Performance and Dance Education , Jennifer Clement

Narratives of lesbian transformation: Coming out stories of women who transition from heterosexual marriage to lesbian identity , Clare F. Walsh

The Conundrum of Women’s Studies as Institutional: New Niches, Undergraduate Concerns, and the Move Towards Contemporary Feminist Theory and Action , Rebecca K. Willman

Theses/Dissertations from 2006 2006

A Feminist Perspective on the Precautionary Principle and the Problem of Endocrine Disruptors under Neoliberal Globalization Policies , Erica Hesch Anstey

Asymptotes and metaphors: Teaching feminist theory , Michael Eugene Gipson

Postcolonial Herstory: The Novels of Assia Djebar (Algeria) and Oksana Zabuzhko (Ukraine): A Comparative Analysis , Oksana Lutsyshyna

Theses/Dissertations from 2005 2005

Loving Loving? Problematizing Pedagogies of Care and Chéla Sandoval’s Love as a Hermeneutic , Allison Brimmer

Exploring Women’s Complex Relationship with Political Violence: A Study of the Weathermen, Radical Feminism and the New Left , Lindsey Blake Churchill

The Voices of Sex Workers (prostitutes?) and the Dilemma of Feminist Discourse , Justine L. Kessler

Reconstructing Women's Identities: The Phenomenon Of Cosmetic Surgery In The United States , Cara L. Okopny

Fantastic Visions: On the Necessity of Feminist Utopian Narrative , Tracie Anne Welser

Theses/Dissertations from 2004 2004

The Politics of Being an Egg “Donor” and Shifting Notions of Reproductive Freedom , Elizabeth A. Dedrick

Women, Domestic Abuse, And Dreams: Analyzing Dreams To Uncover Hidden Traumas And Unacknowledged Strengths , Mindy Stokes

Theses/Dissertations from 2001 2001

Safe at Home: Agoraphobia and the Discourse on Women’s Place , Suzie Siegel

Theses/Dissertations from 2000 2000

Women, Environment and Development: Sub-Saharan Africa and Latin America , Evaline Tiondi

Advanced Search

• Email Notifications and RSS
• All Collections
• USF Faculty Publications
• Open Access Journals
• Conferences and Events
• Theses and Dissertations
• Textbooks Collection

Useful Links

• Women's and Gender Studies Department Homepage
• Rights Information
• SelectedWorks
• Submit Research

Home | About | Help | My Account | Accessibility Statement | Language and Diversity Statements

Privacy Copyright

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

An analysis of unconscious gender bias in academic texts by means of a decision algorithm

Contributed equally to this work with: Pedro Orgeira-Crespo, Carla Míguez-Álvarez, Miguel Cuevas-Alonso, Elena Rivo-López

Roles Conceptualization, Data curation, Formal analysis, Software, Validation, Writing – original draft

* E-mail: [email protected]

Affiliation Department of Mechanical Engineering, Heat Engines and Machines, and Fluids, Aerospace Area, Aerospace Engineering School, University of Vigo, Vigo, Spain

Roles Investigation, Methodology, Resources, Validation, Writing – review & editing

Affiliation Language Variation and Textual Categorization (LVTC), Philology and Translation School, University of Vigo, Vigo, Spain

Roles Investigation, Methodology, Validation, Writing – review & editing

Roles Project administration, Resources, Writing – review & editing

Affiliation Faculty of Economic Sciences, Business Organization and Marketing Department, University of Vigo, Vigo, Spain

• Pedro Orgeira-Crespo, 
• Carla Míguez-Álvarez, 
• Miguel Cuevas-Alonso, 
• Elena Rivo-López

• Published: September 30, 2021
• https://doi.org/10.1371/journal.pone.0257903
• Reader Comments

Inclusive language focuses on using the vocabulary to avoid exclusion or discrimination, specially referred to gender. The task of finding gender bias in written documents must be performed manually, and it is a time-consuming process. Consequently, studying the usage of non-inclusive language on a document, and the impact of different document properties (such as author gender, date of presentation, etc.) on how many non-inclusive instances are found, is quite difficult or even impossible for big datasets. This research analyzes the gender bias in academic texts by analyzing a study corpus of more than 12,000 million words obtained from more than one hundred thousand doctoral theses from Spanish universities. For this purpose, an automated algorithm was developed to evaluate the different characteristics of the document and look for interactions between age, year of publication, gender or the field of knowledge in which the doctoral thesis is framed. The algorithm identified information patterns using a CNN (convolutional neural network) by the creation of a vector representation of the sentences. The results showed evidence that there was a greater bias as the age of the authors increased, who were more likely to use non-inclusive terms; it was concluded that there is a greater awareness of inclusiveness in women than in men, and also that this awareness grows as the candidate is younger. The results showed evidence that the age of the authors increased discrimination, with men being more likely to use non-inclusive terms (up to an index of 23.12), showing that there is a greater awareness of inclusiveness in women than in men in all age ranges (with an average of 14.99), and also that this awareness grows as the candidate is younger (falling down to 13.07). In terms of field of knowledge, the humanities are the most biased (20.97), discarding the subgroup of Linguistics, which has the least bias at all levels (9.90), and the field of science and engineering, which also have the least influence (13.46). Those results support the assumption that the bias in academic texts (doctoral theses) is due to unconscious issues: otherwise, it would not depend on the field, age, gender, and would occur in any field in the same proportion. The innovation provided by this research lies mainly in the ability to detect, within a textual document in Spanish, whether the use of language can be considered non-inclusive, based on a CNN that has been trained in the context of the doctoral thesis. A significant number of documents have been used, using all accessible doctoral theses from Spanish universities of the last 40 years; this dataset is only manageable by data mining systems, so that the training allows identifying the terms within the context effectively and compiling them in a novel dictionary of non-inclusive terms.

Citation: Orgeira-Crespo P, Míguez-Álvarez C, Cuevas-Alonso M, Rivo-López E (2021) An analysis of unconscious gender bias in academic texts by means of a decision algorithm. PLoS ONE 16(9): e0257903. https://doi.org/10.1371/journal.pone.0257903

Editor: Jie Zhang, Newcastle University, UNITED KINGDOM

Received: May 24, 2021; Accepted: September 13, 2021; Published: September 30, 2021

Copyright: © 2021 Orgeira-Crespo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data are available at https://github.com/pdrinio/inclusive_CNN .

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

The frequent problem of gender bias has generated a considerable amount of research (as will be seen in later paragraphs). However, there is little research on the phenomenon of second-generation gender bias. This bias occurs when a person continues to unconsciously make stereotype-based evaluations but consciously rejects "gender stereotypes" ([ 1 , 2 ]). This is possible because this indeterminate number of people unconsciously continue to make evaluations based on stereotypes derived from traditions, life experience, norms and/or culture. Academic texts and in particular doctoral theses, we argue, are overwhelmingly in this situation. The very object of the thesis itself, its development, and its expression in a document (in the case of having gender biases), can be considered to be due to involuntary and automatic mental associations based on gender.

Currently, research on unconscious bias has focused mainly on labor issues or workplace organization [ 3 – 5 ] or on barriers to women’s leadership ([ 2 , 6 – 10 ]).

Unconscious bias can be revealed by analyzing data patterns and identifying gender gaps, but doing so in large volumes of data can be a herculean or unmanageable task. With the development of artificial intelligence, it has been possible to process large volumes of data that can address various research domains. There are still certain difficulties present: for example, if that data is loaded with gender-stereotyped concepts, the resulting application of the technology will maintain this bias. Currently, the use of inclusive language is one of the initiatives that try to guarantee equal treatment for all people. This work develops an algorithm for the detection of the non-inclusive use of the Spanish language in the field of academic texts, using machine-learning techniques. Language is one of the areas where discrimination is observed due to the lack of equal treatment of women and men. Many of the debates in artificial intelligence on the subject of gender bias reflect those related to gender equality in society since the 1970s.

Different studies and organizations have studied or even developed user-friendly toolkits or guides with recommendations on language bias [ 11 – 13 ]. Other studies have focused on different aspects of gender and language, mainly on the differences in the way women and men speak as well as the structural and functional aspects of gender-related variation and change in individual languages [ 14 – 16 ]. Different comparative analyses of gender in different languages [ 17 ] have been developed, and a sociolinguistic analysis of several languages of different morphological characteristics [ 18 ].

Text mining (that describes procedures for obtaining key information from a text by detecting hidden patterns in unstructured documents), has been under research over the last years for many different areas [ 19 – 22 ]. The studies of the use of inclusive language in a document might be considered similar to automatic sentimental classification: opinion mining tries to extract opinion, which are subjective from expressions, and the detection of non-inclusive language is about finding whether the use of the same terms is considered as inclusive or not [ 23 – 25 ]. Most of the investigation lines for text classification were based on extracting algorithms from samples set to train algorithms, including support vector machine (SVM) [ 26 – 28 ], k-nearest neighbor (kNN) [ 29 , 30 ], Naïve Bayes (NB) [ 31 – 34 ], CNN [ 35 ] decision tree [ 36 – 39 ], inside the classification paradigm. Big data analysis using different flavors of artificial intelligence has also shown interesting results in many other areas ranging DNA studies [ 40 , 41 ], vehicle theft identification [ 42 ], intrusion detection [ 43 , 44 ] and more. This research analyzes the possible unconscious gender bias in a study corpus consisting of more than one hundred thousand doctoral theses, written in the Spanish universities. For this purpose, an automated algorithm was developed to evaluate the different characteristics of the document and to search for interactions according to age, year of publication, gender and field of knowledge in which the doctoral thesis is framed. The main contributions of this work are: a) The algorithm that detects non-inclusive usages of the language within an Spanish text (given a generic text, every word in every sentence is evaluated and qualified as non-inclusive usage or not); b) The study along over one hundred thousand doctoral theses (the whole Spanish production in electronic format) to find out, according to the developed algorithm, how different properties of the document (author’s age, gender, year of publication) affect the non-inclusiveness index of the doctoral thesis. The paper is structured as follows: section 2 presents the background of previous work provided by the literature, and how the problem was approached; section 3 explains the proposed methodology followed during this research, including the automated algorithm; section 4 display and analyzes the obtained experimental results; finally, section 5 details the conclusions and future work.

Background and related work

Constraints to the description of gender have been explored across languages or countries [ 45 ], as in the case of English [ 46 ], German [ 47 ], Swedish [ 48 ], Chinese [ 49 – 51 ], Polish [ 52 ], Italian [ 53 , 54 ] or French [ 55 ]. In this line and with respect of the Spanish language, there have also been several research studies [ 56 – 58 ]. For the case of academic texts, [ 59 ] analyzed documents from a university to detect the presence of a neutral and sex-linked language as a measure of the impact of regulations on language change. They studied documents between 1969–1972 and 1978–1979. The newer documents were found to contain significantly fewer sex-linked language elements and markedly more neutral elements than the older documents. Most of the changes involved language structures that had received considerable public debate.

When working with large amounts of data, it is necessary to resort to machine learning techniques. For the general classification problem in text mining, a number of classification techniques are explored: [ 60 – 63 ] combined several to obtain an adequate categorization. The support vector machine has historically been tested as a solution for binary classification in text environments under different approaches [ 64 ].

An algorithm can also be biased by gender, ethnicity, nationality, wealth or other social characteristic if it performs differently or with different levels of efficiency for different groups. A common issue found through literature review is the problem of the bias that word embeddings suffer from, according to the text corpora the different solutions are built with. The subject has been studied and some research suggest that gender bias has not been solved yet [ 65 ]. Automated detection of a use of specific language has been already covered in previous research: [ 66 ] deduce the meaning of words by associating them with other words that tend to occur in the same document. Using this approach can lead to conservative implications, such as that “homemaker” is part of the common meanings of the word “woman”, and that “programmer” is part of the meaning of the term “man” [ 67 ].

Several attempts to eliminate that favoritism in the context of gender are present in the literature, including those that approach the problem with a similar point of view as our research [ 68 ]. In [ 69 ], unintended bias is addressed from the perspective of detecting misogyny; in [ 67 ], authors claim to remove gender stereotypes while keeping reasonable and natural embeddings; sexist messages are also studied [ 70 ] or any offensive language detection [ 71 ].

This discrimination is increasingly manifested on social media: [ 72 ] examines the problem of gender discrimination and attempts to move beyond the typical surface-level text classification approach by identifying differences between genders in the way they use the same words. [ 73 ] presented several configurations of a language-independent classifier for predicting the gender of Twitter users. [ 74 ] proposed a set of criteria that a tweet should exhibit in order to be classified as offensive. [ 75 ] covered detecting hate speech and proposed novel methods that take advantage of knowledge-based generalizations for bias-free learning, using Wikipedia and Twitter as the dataset. [ 76 ] finds that racist and homophobic tweets are more likely to be classified as hate speech, but that sexist tweets are generally classified as offensive. Tweets without explicit hate keywords are also harder to rank for. [ 77 ] also studies hate speech on Facebook. [ 78 ] proposes the application of text mining techniques to online gender discourse through the analysis of shared reviews in electronic word-of-mouth communities (eWOM).

[ 67 ] shows that even word embeddings trained on Google News articles exhibit female/male gender stereotypes to a disturbing extent. They quantitatively demonstrated that word-embeddings contain biases in their geometry that reflect gender stereotypes present in broader society. Due to their wide-spread usage as basic features, word embeddings not only reflect such stereotypes but also can amplify them. This poses a significant risk and challenge for machine learning and its applications. All these approaches consider only English embeddings. A few recent studies focus on measuring and reducing gender bias in contextualized word embeddings [ 79 – 82 ]. [ 83 ] mitigates bias by saving one dimension of the word vector for gender. [ 84 ] proposes a regularization loss term for word level language models. Moreover, [ 65 ] shows that mitigation methods based on gender directions are not sufficient, since the embeddings of socially biased words still cluster together.

[ 85 ] work lends itself to a computational approach to identify gender bias and could be used to remove it from training data for a machine learning algorithm. Similarly, Facebook uses algorithms to select content for users’ newsfeeds [ 86 ], and combines lexical-based and machine-learning techniques. The results obtained through this approach show that it is feasible to perform sentiment analysis in Facebook with high accuracy (83.27%) [ 87 ].

Albeit an important part of the previous research is focused on the English language, there are as well some minor attempts to address the issue in different languages such as French and Spanish [ 55 , 56 ]. That research explains how word embeddings experiment bias when it comes to gender analysis (and also when looking for hate in speech); they as well show certain methods to tackle that problem.

Table 1 summarizes the most relevant previous works in this area, shown in this section:

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0257903.t001

The research hypotheses are as follows:

• H1: The development of an automatic text extraction algorithm would help to analyze gender bias.
• H2. The use of inclusive terms is directly related to age.
• H3: The use of inclusive terms is related to the field of knowledge.
• H4: The use of inclusive terms is directly related to gender, being lower in women.
• H5: Unconscious bias is a characteristic of academic texts, particularly doctoral theses.

Proposed method

The context for this research is the corpora of doctoral theses wrote in Spain since 1974, when the first digital treatable documents are found. Table 2 summarizes the key indicators for this process as regards the amount of text to be treated:

https://doi.org/10.1371/journal.pone.0257903.t002

Although the total number of theses found were over 257,000, only 100,450 could be used (the other ones were either not public domain or not in a legible file).

The first area of interest was the age of the author, and how it could affect the use of inclusive language within his/her doctoral thesis. The age and gender of the authors are show in Fig 1 .

https://doi.org/10.1371/journal.pone.0257903.g001

The documents were created along 46 years at the Spanish universities. Only those that were public domain and obtained as PDF files were treated. In this data corpus, over 12 billion words were found using over 9000 different elements, being, therefore, a very complete representation of the Spanish language.

Another important aspect is that every document in the data set belongs to one of the UNESCO categories for doctoral thesis, as displayed on Table 3 :

https://doi.org/10.1371/journal.pone.0257903.t003

A deep learning model was the option chosen due to the need to provide the algorithm with information on a small set of training documents (what should or should not be inclusive for a term in context), and subsequently label terms in different contexts to validate the model. Metadata were also obtained that described the document: the date of publication, the language in which it was written, the name of the author, the title of the research, and the UNESCO code that characterizes the area of knowledge; that information was obtained, structured, and stored in a database, as depicted in Fig 2 :

https://doi.org/10.1371/journal.pone.0257903.g002

Algorithm design

Convolutional Neural Networks (CNN) are well known in the area of text classification [ 88 ]. Their capacity to learn from sentences having the same or similar words in different parts of the phrase allows us to capture relevant information. Anyway, the architecture of the CNN needs to be adapted to the specific problem to be solved, and also a future work should be to analyze whether the generated embeddings depend on the domain of the corpus or not. Much research has been explored in text mining and NLP, as well as the world of text classification. The whole process is summarized in Fig 3 :

https://doi.org/10.1371/journal.pone.0257903.g003

Document collection.

The documents that make up the corpus were obtained from the universities that make up the Spanish university system, hosted on their official servers. In the first place, the raw documents were obtained, to later process them and extract from them the information related to authorship, date, subject, etc.

An ecosystem of virtual machines has provided the IT infrastructure capabilities: hosted in Azure, it consisted of Debian MVs on which Python scripts were executed, and hosted the information in SQL-Server and MongoDB databases. Fig 4 describes that infrastructure:

https://doi.org/10.1371/journal.pone.0257903.g004

Text extraction, curation and storage.

Once the raw text of each document has been extracted, it has been processed to prepare it to be analyzed by an automatic algorithm. Special emphasis has been placed on eliminating those words that, belonging to areas of the document where studying its inclusiveness does not really apply, could negatively impact the process without providing information: mathematical formulae, portions of texts in other languages within the text in the Spanish language, page numbers, section names, table headers, etc.

This process has been carried out using different libraries oriented to NLP (natural language processing), both on Python and in C # using.Net Core.

Data obtention.

Once the information was available as extracted text, properly structured, and stored, the next step was to provide a numerical representation to the text. The first focus was to locate all those adjectives and common nouns that, present in the text in their masculine form, also have a valid feminine form in Spanish (for example: “profesor” -teacher-, which has its feminine form “profesora”-female teacher-, valid). Common nouns and adjectives are those that, being able to exist in their masculine, feminine, or neutral form, are key to determine whether their use might be considered non-inclusive. Accordingly, the Freeling library (specialized in morphologic analysis) was used to provide noun and adjective detection; in this case, the hmm_tagger (Hidden Markov Model PoS Tagger) was employed to provide PoS (part of speech) on the text, implementing a classical trigram Markovian tagger. Tables 4 and 5 display the characteristics of the pairs of attribute/value for common nouns and adjectives:

https://doi.org/10.1371/journal.pone.0257903.t004

https://doi.org/10.1371/journal.pone.0257903.t005

Under this methodology, the whole corpus could be analyzed putting special emphasis on the search of parts of sentences where common nouns and adjectives could be found under their masculine form, ruling out those that do not have a feminine form in the Spanish dictionary, and creating parts of sentences to be evaluated. Since sentences need to have a context to be properly analyzed, the algorithm looked for their beginning, using a punctuation mark (comma, period, semi-colon, etc.); that punctuation mark became the beginning of a new sub-sentence to be isolated and stored. In that new sub-sentence, the first instance of a common noun or adjective in their masculine form was searched for, until the next punctuation mark (if during this process no element could be found, the sub-sentence was discarded). Once an element was found, the next punctuation mark was looked for; after several iterations, it was determined that the subsentence should not contain more than 7 other words before the next punctuation mark was found, counting from the element that was discovered: that provided meaningful sub-sentences that could be properly analyzed and characterized, in the next step.

Vector representation.

The goal of this step was to generate a sequence of word embeddings using the input acquired in the previous step, to finally obtain a numerical representation of the relevant sub-sentences that contained potential non-inclusive usages of the elements found within the text. The idea behind this approach is that it is the context where a masculine common noun or adjective is used, that give rise to its non-inclusiveness (not the term on its own). Therefore, every sub-sentence was characterized with a sequence of vectors, where each vector was calculated as shown in Table 6 :

https://doi.org/10.1371/journal.pone.0257903.t006

Architectural model and algorithm training.

https://doi.org/10.1371/journal.pone.0257903.g005

The architecture is based on a CNN model, to provide classification to determine when each sub-sentence can be considered as non-inclusive use of the language or not. After experimenting with the data iterating through different methods, we empirically determined (after a tenfold cross validation) that every convolution operation to be applied would be implemented as 128 filters; we decided to use a pooling layer based on max pooling, and a final perceptron with two layers. Fig 5 displays the architecture:

Data labelling was performed by a group of experts on the field, which were fed with 8,200 sub-sentence samples and their extra context (validating, as well, the size of the sub-sentence for relevant non-inclusive identification). Every sub-sentence was reviewed by two different persons, providing a binary classification (“Inclusive”/”Non inclusive” usage). The process was accomplished in two rounds, where the second was to categorize only the sub-sentences that had obtained different classification; finally, that third vote provided the final categorization for the sub-sentence.

Out of the labelled samples, 6,550 were used to train the algorithm, and 1,640 were kept for validation purposes, in a standard 80/20 ratio.

Validation.

A confusion matrix is a very common description of the performance of categorization models. The matrix (indeed, a table) displays the number of accurately and erroneously classified samples, compared to the actual outcomes (target value) in the test data. In the field of artificial intelligence and especially in the problem of statistical classification, a confusion matrix is a tool that allows the visualization of the performance of an algorithm that is used in supervised learning [ 89 ].

As well, certain indicators were also added to cope with the common issue binary classification comes with: false negatives (any erroneously identification by the algorithm of a sub-sentence as non-inclusive usage, when the truth is that the usage was inclusive), are more harmful than false positives. When the panel of experts reviewed an undetected non-inclusive, they felt more permissive than when the algorithm claims a non-inclusive usage for a proper language utilization.

Table 7 shows the notation that was used:

https://doi.org/10.1371/journal.pone.0257903.t007

Results and analysis

The analysis was divided in two parts: first, testing the reliability of the algorithm to determine the inclusive usage of language in a part of speech (subphrase); second, testing on a large number of documents, in this case Spanish theses, to draw conclusions about bias.

To evaluate the performance of the algorithm, several data from the labelled dataset was kept for validation (a 20% split). The results calculated by the algorithm were compared to those obtained by the panel of linguistics, and compared using well-known standard indicators; Table 8 displays the scores obtained:

https://doi.org/10.1371/journal.pone.0257903.t008

The results can be considered as quite reasonable. The accuracy is over 94%, where a certain level of misclassification was expected. The false negative ratio is over 30%, which may be reasoned as cases where a human determines a truly non-inclusive usage and the algorithm could not tell; the false positive ratio stays well below 1%, matching the expectations at the beginning of the project. Moreover, the false positive ratio kept well under the false negative ratio: in a non-inclusive classification problem, the fact of finding a misclassification by an algorithm is not positive at all, but much worse is the case where a false positive occurs, than a false negative: a false positive is assumed as something that should have been detected, but that the algorithm did not perceive; a false negative is perceived as a less acceptable failure, since it generates greater mistrust by having considered a text that was not inclusive as non-inclusive. In our case, as the FPR (false positive ratio) remains below the FNR (false negative ratio), and both are relatively low, the final result is considered positive.

We may find in Fig 6 the confusion matrix for the obtained values, showing quite reasonable results as for the performance of the algorithm:

https://doi.org/10.1371/journal.pone.0257903.g006

Every document was characterized by its index and, consequently, different conclusions arise studying the evolution of the value of the index according to different origin regions, age of the author, and evolution throughout time.

Referring to Fig 1 , the distribution of the average number of non-inclusive terms per document according to the age of the author and its gender is shown in Table 9 :

https://doi.org/10.1371/journal.pone.0257903.t009

Considering the average for each age group, it can be seen that the highest mean score corresponds to the oldest group (> 55 years) while the lowest mean score would be that of the youngest group (24–29 years) regardless of gender. Based on these results, it can be said that there is a direct relationship between age and the use of inclusive terms in the texts. The most important gap happens after the age of 50, both in the average divergence with the previous bands and in gender discrimination.

It is also important to note that the older the age group, the more marked the gender difference between men and women is, also decreasing with the age of the group. It also turns out that the youngest have the lowest incidence. It is true in all cases that the male gender continues to use more non-inclusive terms than the female gender, regardless of the age group that we consider. In the same vein, the results would indicate a significantly higher support for inclusive language on the side of the female participants compared to the male ones.

Fig 7 displays the non-inclusiveness occurrences according to the age of the author. As expected, the youngest people are the ones who had the least use of non-inclusive terms, which means that current social conditions are changing the perception of their use.

https://doi.org/10.1371/journal.pone.0257903.g007

These results are in line with those obtained in other investigations that point to the existence of a direct relationship between age and the use of inclusive terms [ 90 – 93 ]. It was not found in the literature that at a younger age, opinions towards inclusive language were more polarized according to gender [ 94 ]. This also has its explanation in the type of research carried out focused on doctoral theses, which results in that in the case of bias, it should be unconsciously [ 85 – 88 , 95 ].

Fig 8 shows a visual representation of the non-inclusive indicator according to the UNESCO subject group displayed in Table 10 , where the doctoral theses in Humanities achieve the greatest score, and other sciences as logic, mathematics, physics, chemistry or engineering obtain the best inclusive results. This fact makes sense with the typical topics under these subjects, commonly involving facts, numbers, formulae, etc.; social sciences and humanities might be prone to the use of words or opinions, where the inclusiveness might be affected. Although in the Humanities group, the field of Linguistics (9.90) obtains the lowest value of all (showing that in this field special attention is paid to the use of non-inclusive terms), the opposite occurs with the Law group, which obtains the highest value of all with 31.7:

https://doi.org/10.1371/journal.pone.0257903.g008

https://doi.org/10.1371/journal.pone.0257903.t010

Fig 9 shows the evolution of the use of non-inclusive words by doctoral thesis over the years. It can be seen that, until 2010, doctoral theses maintained a steady level in the use of non-inclusive words. Since the publication in Spain of Royal Decree 99/2011, which regulates official doctoral education, there has been a very significant and continuous increase in the preparation and reading of theses. This increase also occurred in the use of non-inclusive terms, which could have a sociological behavioral explanation not considered in this study. In the last years (2018, 2019), there is a sharp decrease in the use of non-inclusive terms, which could be related to the current trend of avoiding gender discrimination. With the current values, the ratios prior to 2010 have not yet been reached.

https://doi.org/10.1371/journal.pone.0257903.g009

Recalling the initial hypothesis enunciated in the introduction:

• H2: The use of no-inclusive terms is directly related to age.
• H3: The use of no-inclusive terms is related to the field of knowledge.
• H4: Gender-bias is directly related to gender, being lower in women.

R1: The development of an automatic text extraction algorithm makes it possible to analyze gender bias in the case of doctoral theses written in Spanish. We understand that the system is applicable to documents outside the academic field, in the same language. The possibility of analyzing hundreds of thousands of texts with millions of words would become unapproachable otherwise.

R2: As can be seen in Table 9 , the use of non-inclusive terms is directly related to age, increasing their use very deeply, assuming about 5.85 points above the average in the case of men and 4.25 in the case of women.

R3: Gender bias is grouped by areas with a behavior that could be considered logical. Where it appears least is in those documents with a use of scientific or mathematical terms. The areas where gender bias occurs most are those of Humanities and Social Sciences: with two extremes, the largest being the Laws subarea and the smallest being the Linguistics subarea.

R4: It has been shown that in the case of doctoral theses written in Spanish this is true in all cases and in all age ranges, also accentuating this male/female difference with age.

R5: The unconscious bias is characteristic of academic texts and, particularly, of doctoral theses. This has been shown to be true in the case of doctoral theses written in Spanish. That explains why the fields of engineering and basic sciences have the lowest values. If it was intentional, it would not depend on the scope and it would occur in any of them without a behavior pattern.

The innovation provided by this research lies mainly in the ability to detect, within a digital text document in Spanish, whether the use of the language can be considered non-inclusive, based on a CNN that has been trained in the context of the doctoral thesis. A significant number of documents have been used, by means of all accessible doctoral theses from Spanish universities of the last 40 years, only manageable by data mining systems, so that the training allows identifying the terms within the context and effectively and compiling them in a novel dictionary of non-inclusive terms.

This research addresses the problem of providing an algorithm that can detect non-inclusive uses of the Spanish language without human supervision. It also shows that, although CNN usually requires a significant set of labeled data, when properly trained, it can show reasonable results in terms of accuracy, error, false positive and false negative rates.

As can be seen, there is a slight increasing trend in the use of non-inclusive terms as the doctoral student gets older, going from an average value of 13.68 to 21.18. Furthermore, for each age group, it is also observed that the number of non-inclusive occurrences is always higher in men than in women, concluding that there is a greater awareness of inclusiveness in women than in men, and also that this awareness grows as the candidate is younger. In terms of field of knowledge, the humanities are the most biased, discarding the subgroup of Linguistics, which has the least bias at all levels, and the field of science and engineering, which also have the least influence. This supports the assumption that the bias in academic texts (doctoral theses) is due to unconscious issues: otherwise, it would not depend on the field, age, gender, and it would occur in any field in the same ratio.

The results also allow us to study the evolution of inclusivity according to the author’s age, and the evolution over time. It is important to mention that new forms of inclusive language that included "wildcard" characters to abstract the word from being masculine or feminine (such as "todes", "tod@s", "tod*s" or "todxs") were explicitly searched for, but were not found in any document as part of the dissertation.

As for the limitations of this work, we can mention the following: this study is limited to a specific type of academic text, namely doctoral dissertations. However, the method of data collection and analysis could be extended to other types or to non-academic texts, and checked whether the most common non-inclusive terms match those obtained in this research. A proper training of the CNN on a different context of documents could provide the algorithm with adequate customization to address documents from different environments.

• 1. Hill , C.M., Kevin; Benson, Kathleen; Handley, Grace, Barriers and bias: The status of women in leadership. 2016, American Association of University Women AAUW. p. 66.
• View Article
• Google Scholar
• 3. (act/emp), B.f.E.A., Breaking barriers: unconscious gender bias in the workplace. 2017, International Labour Office.
• PubMed/NCBI
• 5. McKinsey_Company, Women in the workplace. 2020.
• 6. (ACT/EMP), I.-B.f.E.A., Women in Business and Management: Gaining momentum. 2015.
• 7. Bierema L.L, No women left behind: Critical leadership development to build gender consciousness and transform organizations, in Handbook of research on gender and leadership, Madsen S.R, Editor. 2017, Edward Elgar Publishing.: Cheltenham, England. p. 145–162.
• 11. EIGE, E.I.f.G.E., Toolkit on Gender-sensitive Communication. 2018, Vilnius, Lithuania: Publications Office of the European Union,.
• 12. Association A., Publication manual of the American Psychological Association: the official guide to APA style. Seventh edition ed. 2020, Washington, D.C.: American Psychological Association.
• 14. Pauwels A., English. Spreading the feminist word: The case of the new courtesy title Ms in Australian English, in Gender Across Languages. 2001, John Benjamins.
• 15. Romaine S., English. A corpus-based view of gender in British and American English, in Gender Across Languages. 2001, John Benjamins.
• 17. Hellinger M. and Bußmann H., Gender Across Languages: The linguistic representation of women and men. Volume 3. 2003: John Benjamins.
• 18. Teso E., A comparative study of gender-based linguistic reform across four European countries. 2010, Liverpool John Moores University.: England.
• 24. Singh, V.K., et al. Sentiment analysis of movie reviews: A new feature-based heuristic for aspect-level sentiment classification. in 2013 International Mutli-Conference on Automation, Computing, Communication, Control and Compressed Sensing (iMac4s). 2013.
• 28. Pratama, B.Y. and R. Sarno. Personality classification based on Twitter text using Naive Bayes, KNN and SVM. in 2015 International Conference on Data and Software Engineering (ICoDSE). 2015.
• 33. Zhang, H. and D. Li. Naïve Bayes Text Classifier. in 2007 IEEE International Conference on Granular Computing (GRC 2007). 2007.
• 35. Gadekallu T. R., Alazab M., Kaluri R., Maddikunta P. K. R., Bhattacharya S., Lakshmanna K., & Parimala M. (2021). Hand gesture classification using a novel CNN-crow search algorithm. Complex & Intelligent Systems, 1–14.
• 39. Zhang, X., J. Zhao, and Y. LeCun, Character-level convolutional networks for text classification, in Proceedings of the 28th International Conference on Neural Information Processing Systems—Volume 1. 2015, MIT Press: Montreal, Canada. p. 649–657.
• 42. Kumar, M. E., Reddy, G. T., Sudheer, K., Reddy, M. P. K., Kaluri, R., Rajput, D. S., et al. (2017, November). Vehicle theft identification and intimation using gsm & iot. In IOP Conference Series: Materials Science and Engineering (Vol. 263, No. 4, p. 042062). IOP Publishing.
• 56. Calero Fernández M., Sexismo lingüístico. Análisis y propuestas ante la discriminación sexual en el lenguaje. 1999, Madrid: Editorial Narcea. 225.
• 58. Bosque, I. Sexismo lingüístico y visibilidad de la mujer. 2012; Available from: http://www.rae.es/sites/default/files/Sexismo_linguistico_y_visibilidad_de_la_mujer_0.pdf .
• 59. García Meseguer Á., ¿Es sexista la lengua española?. Una investigación sobre el género gramatical. 1994, Barcelona (Spain): Editorial Paidós.
• 61. Orgeira P., et al., Decision Algorithm for the Automatic Determination of the Use of Non-Inclusive Terms in Academic Texts. Publications, 2020. 8: p. 41. https://doi.org/10.1016/j.scitotenv.2018.01.219 pmid:29426134
• 65. Gonen H. and Goldberg Y. Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them. in NAACL. 2019.
• 67. Bolukbasi, T., et al., Man is to computer programmer as woman is to homemaker? debiasing word embeddings, in Proceedings of the 30th International Conference on Neural Information Processing Systems. 2016, Curran Associates Inc.: Barcelona, Spain. p. 4356–4364.
• 68. Dixon, L., et al., Measuring and Mitigating Unintended Bias in Text Classification, in Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society. 2018, Association for Computing Machinery: New Orleans, LA, USA. p. 67–73.
• 69. Nozza, D., C. Volpetti, and E. Fersini. Unintended Bias in Misogyny Detection. in 2019 IEEE/WIC/ACM International Conference on Web Intelligence (WI). 2019.
• 73. Burger, J.D., et al., Discriminating gender on Twitter, in Proceedings of the Conference on Empirical Methods in Natural Language Processing. 2011, Association for Computational Linguistics: Edinburgh, United Kingdom. p. 1301–1309.
• 75. Badjatiya, P., M. Gupta, and V. Varma, Stereotypical Bias Removal for Hate Speech Detection Task using Knowledge-based Generalizations, in The World Wide Web Conference. 2019, Association for Computing Machinery: San Francisco, CA, USA. p. 49–59.
• 79. Basta, C., M. Costa-jussa, and N. Casas. Evaluating the Underlying Gender Bias in Contextualized Word Embeddings. in First Workshop on Gender Bias in Natural Language Processing. 2019. Florence, Italy: Association for Computational Linguistics.
• 80. Hitti Y., et al. Proposed Taxonomy for Gender Bias in Text; A Filtering Methodology for the Gender Generalization Subtype. 2019.
• 81. May, C., et al. On Measuring Social Biases in Sentence Encoders. in Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Languages Technologies. 2019. Minneapolis, Minnesota: Association for Computational Linguistics.
• 82. Zhou, P., et al. Examining Gender Bias in Languages with Grammatical Gender. in 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019. Hong Kong, China: Association for Computational Linguistics.
• 83. Zhao, J., et al. Learning Gender-Neutral Word Embeddings. in Conference on Empirical Methods in Natural Language Processing. 2018. Brussels, Belgium: ssociation for Computational Linguistics.
• 84. Bordia, S. and S. Bowman. Identifying and Reducing Gender Bias in Word-Level Language Models. in NAACL HLT 2019–2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies—Proceedings of the Student Research Workshop. 2019. Association for Computational Linguistics (ACL).
• 85. Leavy, S. Gender Bias in Artificial Intelligence: The Need for Diversity and Gender Theory in Machine Learning. in 2018 IEEE/ACM 1st International Workshop on Gender Equality in Software Engineering (GE). 2018.
• 88. Kim, Y., Convolutional Neural Networks for Sentence Classification. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, 2014.
• 95. Basile V., et al. SemEval-2019 Task 5: Multilingual Detection of Hate Speech Against Immigrants and Women in Twitter. in SemEval@NAACL-HLT. 2019.

Suggestions or feedback?

MIT News | Massachusetts Institute of Technology

• Machine learning
• Sustainability
• Black holes
• Classes and programs

Departments

• Aeronautics and Astronautics
• Brain and Cognitive Sciences
• Architecture
• Political Science
• Mechanical Engineering

Centers, Labs, & Programs

• Abdul Latif Jameel Poverty Action Lab (J-PAL)
• Picower Institute for Learning and Memory
• Lincoln Laboratory
• School of Architecture + Planning
• School of Engineering
• School of Humanities, Arts, and Social Sciences
• Sloan School of Management
• School of Science
• MIT Schwarzman College of Computing

Study finds gender and skin-type bias in commercial artificial-intelligence systems

Press contact :, media download.

*Terms of Use:

Images for download on the MIT News office website are made available to non-commercial entities, press and the general public under a Creative Commons Attribution Non-Commercial No Derivatives license . You may not alter the images provided, other than to crop them to size. A credit line must be used when reproducing images; if one is not provided below, credit the images to "MIT."

Previous image Next image

Three commercially released facial-analysis programs from major technology companies demonstrate both skin-type and gender biases, according to a new paper researchers from MIT and Stanford University will present later this month at the Conference on Fairness, Accountability, and Transparency.

In the researchers’ experiments, the three programs’ error rates in determining the gender of light-skinned men were never worse than 0.8 percent. For darker-skinned women, however, the error rates ballooned — to more than 20 percent in one case and more than 34 percent in the other two.

The findings raise questions about how today’s neural networks, which learn to perform computational tasks by looking for patterns in huge data sets, are trained and evaluated. For instance, according to the paper, researchers at a major U.S. technology company claimed an accuracy rate of more than 97 percent for a face-recognition system they’d designed. But the data set used to assess its performance was more than 77 percent male and more than 83 percent white.

“What’s really important here is the method and how that method applies to other applications,” says Joy Buolamwini, a researcher in the MIT Media Lab’s Civic Media group and first author on the new paper. “The same data-centric techniques that can be used to try to determine somebody’s gender are also used to identify a person when you’re looking for a criminal suspect or to unlock your phone. And it’s not just about computer vision. I’m really hopeful that this will spur more work into looking at [other] disparities.”

Buolamwini is joined on the paper by Timnit Gebru, who was a graduate student at Stanford when the work was done and is now a postdoc at Microsoft Research.

Chance discoveries

The three programs that Buolamwini and Gebru investigated were general-purpose facial-analysis systems, which could be used to match faces in different photos as well as to assess characteristics such as gender, age, and mood. All three systems treated gender classification as a binary decision — male or female — which made their performance on that task particularly easy to assess statistically. But the same types of bias probably afflict the programs’ performance on other tasks, too.

Indeed, it was the chance discovery of apparent bias in face-tracking by one of the programs that prompted Buolamwini’s investigation in the first place.

Several years ago, as a graduate student at the Media Lab, Buolamwini was working on a system she called Upbeat Walls, an interactive, multimedia art installation that allowed users to control colorful patterns projected on a reflective surface by moving their heads. To track the user’s movements, the system used a commercial facial-analysis program.

The team that Buolamwini assembled to work on the project was ethnically diverse, but the researchers found that, when it came time to present the device in public, they had to rely on one of the lighter-skinned team members to demonstrate it. The system just didn’t seem to work reliably with darker-skinned users.

Curious, Buolamwini, who is black, began submitting photos of herself to commercial facial-recognition programs. In several cases, the programs failed to recognize the photos as featuring a human face at all. When they did, they consistently misclassified Buolamwini’s gender.

Quantitative standards

To begin investigating the programs’ biases systematically, Buolamwini first assembled a set of images in which women and people with dark skin are much better-represented than they are in the data sets typically used to evaluate face-analysis systems. The final set contained more than 1,200 images.

Next, she worked with a dermatologic surgeon to code the images according to the Fitzpatrick scale of skin tones, a six-point scale, from light to dark, originally developed by dermatologists as a means of assessing risk of sunburn.

Then she applied three commercial facial-analysis systems from major technology companies to her newly constructed data set. Across all three, the error rates for gender classification were consistently higher for females than they were for males, and for darker-skinned subjects than for lighter-skinned subjects.

For darker-skinned women — those assigned scores of IV, V, or VI on the Fitzpatrick scale — the error rates were 20.8 percent, 34.5 percent, and 34.7. But with two of the systems, the error rates for the darkest-skinned women in the data set — those assigned a score of VI — were worse still: 46.5 percent and 46.8 percent. Essentially, for those women, the system might as well have been guessing gender at random.

“To fail on one in three, in a commercial system, on something that’s been reduced to a binary classification task, you have to ask, would that have been permitted if those failure rates were in a different subgroup?” Buolamwini says. “The other big lesson ... is that our benchmarks, the standards by which we measure success, themselves can give us a false sense of progress.”

“This is an area where the data sets have a large influence on what happens to the model,” says Ruchir Puri, chief architect of IBM’s Watson artificial-intelligence system. “We have a new model now that we brought out that is much more balanced in terms of accuracy across the benchmark that Joy was looking at. It has a half a million images with balanced types, and we have a different underlying neural network that is much more robust.”

“It takes time for us to do these things,” he adds. “We’ve been working on this roughly eight to nine months. The model isn’t specifically a response to her paper, but we took it upon ourselves to address the questions she had raised directly, including her benchmark. She was bringing up some very important points, and we should look at how our new work stands up to them.”

Share this news article on:

Press mentions.

Mashable reporter Adele Walton spotlights Joy Buolamwini PhD '22 and her work in uncovering racial bias in digital technology. “Buolamwini created what she called the Aspire Mirror, which used face-tracking software to register the movements of the user and overlay them onto an aspirational figure,” explains Walton. “When she realised the facial recognition wouldn’t detect her until she was holding a white mask over her face, she was confronted face on with what she termed the ‘coded gaze.’ She soon founded the Algorithmic Justice League, which exists to prevent AI harms and increase accountability.”

Los Angeles Times

Los Angeles Times reporter Brian Merchant spotlights Joy Buolamwini PhD '22 and her new book, “Unmasking AI: My Mission to Protect What is Human in a World of Machines.” “Buolamwini’s book recounts her journey to become one of the nation’s preeminent scholars and critics of artificial intelligence — she recently advised President Biden before the release of his executive order on AI — and offers readers a compelling, digestible guide to some of the most pressing issues in the field,” writes Merchant.

The Boston Globe

Joy Buolamwini PhD '22 speaks with Brian Bergstein of The Boston Globe’s “ Say More” podcast about her academic and professional career studying bias in AI. “As I learned more and also became familiar with the negative impacts of things like facial recognition technologies, it wasn’t just the call to say let’s make systems more accurate but a call to say let’s reexamine the ways in which we create AI in the first place and let’s reexamine our measures of progress because so far they have been misleading,” says Buolamwini

Joy Buolamwini PhD '22 writes for The Boston Globe about her experience uncovering bias in artificial intelligence through her academic and professional career. “I critique AI from a place of having been enamored with its promise, as an engineer more eager to work with machines than with people at times, as an aspiring academic turned into an accidental advocate, and also as an artist awakened to the power of the personal when addressing the seemingly technical,” writes Buolamwini. “The option to say no, the option to halt a project, the option to admit to the creation of dangerous and harmful though well-intentioned tools must always be on the table.”

Diverse Issues in Higher Education

Joy Buolamwini PhD ’22 has been named one of Diverse: Issues in Higher Education’s Top Women for 2023 for her work in developing “more equitable and accountable technology.” Buolamwini “uncovered racial and gender bias in AI services from high profile companies such as Microsoft, IBM and Amazon. Now a sought-after international speaker, Buolamwini continues to advocate for algorithmic justice,” writes Diverse: Issues in Higher Education .

New York Times

Graduate student Joy Buolamwini joins Kara Swisher on The New York Times' “Sway” podcast to discuss her crusade against bias in facial recognition technologies. “If you have a face, you have a place in this conversation,” says Buolamwini.

NBC News reporters Lindsay Hoffman and Caroline Kim spotlight graduate student Joy Buolamwini’s work uncovering racial and gender bias in AI systems in a piece highlighting women who are “shattering ceilings, making groundbreaking discoveries, and spreading public awareness during the global pandemic.” Hoffman and Kim note that Buolamwini’s research "helped persuade these companies to put a hold on facial recognition technology until federal regulations were passed.”

Fast Company

Fast Company  reporter Amy Farley spotlights graduate student Joy Buolamwini and her work battling bias in artificial intelligence systems, noting that “when it comes to AI injustices, her voice resonates.” Buolamwini emphasizes that “we have a voice and a choice in the kind of future we have.”

Quartz reporter Nicolas Rivero notes that IBM’s decision to end its facial recognition program was inspired by “one influential piece of research: the Gender Shades project, from MIT Media Lab’s Joy Buolamwini and Microsoft Research’s Timnit Gebru.” Buolamwini and Gebru found that “commercial facial recognition software was significantly less accurate for darker-skinned women than for lighter-skinned men. “

Verge reporter Nick Statt notes that, “Much of the foundational work showing the flaws of modern facial recognition tech with regard to racial bias is thanks to Joy Buolamwini, a researcher at the MIT Media Lab, and Timnit Gebru, a member at Microsoft Research.”

Financial Times

Graduate student Joy Buolamwini has been named to the Financial Times ’ list of change-makers, which highlights “30 of the planet’s most exciting young people.” Financial Times reporter India Ross notes that Buolamwini, “identified gender and racial biases in artificial intelligence, and her efforts have prompted technology companies such as IBM to upgrade their software accordingly.”

Graduate student Joy Buolamwini writes for TIME about the need to tackle gender and racial bias in AI systems. “By working to reduce the exclusion overhead and enabling marginalized communities to engage in the development and governance of AI, we can work toward creating systems that embrace full spectrum inclusion,” writes Buolamwini.

Fortune- CNN

Fortune reporters Aaron Pressman and Adam Lashinsky highlight graduate student Joy Buolamwini’s work aimed at eliminating bias in AI and machine learning systems. Pressman and Lashinsky note that Buolamwini believes that “who codes matters,” as more diverse teams of programmers could help prevent algorithmic bias. 

In a recent blog post, Microsoft’s president and chief legal officer, Brad Smith, references research by MIT graduate student Joy Buolamwini while calling for government to regulate the use of facial recognition software. Buolamwini’s work “showed error rates of as much as 35% for systems classifying darker skinned women,” reports Dina Bass for Bloomberg.

Katharine Schwab of Co.Design highlights graduate student Joy Buolamwini and Visiting Scholar J. Nathan Matias as “design heroes” for their commitment to keep technology fair. Schwab writes that Buolamwini has forced companies “to develop better, more equitable technology” while Matias helped “reduced the prevalence of fake news.”

In an article for The New York Times , graduate student Joy Buolamwini writes about how AI systems can often reinforce existing racial biases and exclusions. Buolamwini writes that, “Everyday people should support lawmakers, activists and public-interest technologists in demanding transparency, equity and accountability in the use of artificial intelligence that governs our lives.”

A recent study from Media Lab graduate student Joy Buolamwini addresses errors in facial recognition software that create concern for civil liberties. “If programmers are training artificial intelligence on a set of images primarily made up of white male faces, their systems will reflect that bias,” writes Cristina Quinn for WGBH.

Boston Magazine

Spencer Buell of Boston Magazine speaks with graduate student Joy Buolamwini, whose research shows that many AI programs are unable to recognize non-white faces. “‘We have blind faith in these systems,’ she says. ‘We risk perpetuating inequality in the guise of machine neutrality if we’re not paying attention.’”

Dave Gershgorn writes for Quartz, highlighting congress’ concerns around the dangers of inaccurate facial recognition programs. He cites Joy Buolamwini’s Media Lab research on facial recognition, which he says “maintains that facial recognition is still significantly worse for people of color.”

The Economist

An article in The Economist states that new research by MIT grad student Joy Buolamwini supports the suspicion that facial recognition software is better at processing white faces than those of other people. The bias probably arises “from the sets of data the firms concerned used to train their software,” the article suggests.

Marketplace

Molly Wood at Marketplace speaks with Media Lab graduate student Joy Buolamwini about the findings of her recent research, which examined widespread bias in AI-supported facial recognition programs. “At the end of the day, data reflects our history, and our history has been very biased to date,” Buolamwini said.

Recent research from graduate student Joy Buolamwini shows that facial recognition programs, which are increasingly being used by law enforcement, are failing to identify non-white faces. “When these systems can’t recognize darker faces with as much accuracy as lighter faces, there’s a higher likelihood that innocent people will be targeted by law enforcement,” writes Katharine Schwab for Co. Design . 

New Scientist

Graduate student Joy Buolamwini tested three different face-recognition systems and found that the accuracy is best when the subject is a lighter skinned man, reports Timothy Revell for New Scientist . With facial recognition software being used by police to identify suspects, “this means inaccuracies could have consequences, such as systematically ingraining biases in police stop and searches,” writes Revell.

A study co-authored by MIT graduate student Joy Buolamwini finds that facial-recognition software is less accurate when identifying darker skin tones, especially those of women, writes Josh Horwitz of Quartz . According to the study, these errors could cause AI services to “treat individuals differently based on factors such as skin color or gender,” explains Horwitz.

Writing for Gizmodo , Sidney Fussell explains that a new Media Lab study finds facial-recognition software is most accurate when identifying men with lighter skin and least accurate for women with darker skin. The software analyzed by graduate student Joy Buolamwini “misidentified the gender of dark-skinned females 35 percent of the time,” explains Fussell.

The New York Times

Steve Lohr writes for the New York Times about graduate student Joy Buolamwini’s findings on the biases of artificial intelligence in facial recognition. “You can’t have ethical A.I. that’s not inclusive,” Buolamwini said. “And whoever is creating the technology is setting the standards.”

Previous item Next item

Related Links

• Paper: " Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification."
• Gender Shades project website
• Joy Buolamwini
• Center for Civic Media
• School of Architecture and Planning

Related Topics

• Artificial intelligence
• Computer science and technology
• Diversity and inclusion
• Technology and society

Related Articles

Cornel West advocates the “examined life” on campus

Joy Buolamwini wins national contest for her work fighting bias in machine learning

Institute launches the MIT Intelligence Quest

Designing virtual identities for empowerment and social change

President Obama discusses artificial intelligence with Media Lab Director Joi Ito

More mit news.

Engineering proteins to treat cancer

Read full story →

MIT team wins grand prize at NASA’s First Nations Launch High-Power Rocket Competition

Nurturing success

For developing designers, there’s magic in 2.737 (Mechatronics)

Study: Transparency is often lacking in datasets used to train large language models

How MIT’s online resources provide a “highly motivating, even transformative experience”

• More news on MIT News homepage →

Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA, USA

• Map (opens in new window)
• Events (opens in new window)
• People (opens in new window)
• Careers (opens in new window)
• Accessibility
• Social Media Hub
• MIT on Facebook
• MIT on YouTube
• MIT on Instagram

• My Shodhganga
• Receive email updates
• Edit Profile

Shodhganga : a reservoir of Indian theses @ INFLIBNET

• Shodhganga@INFLIBNET
• Hemchand Yadav Vishwavidyalaya
• Department of Education

Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

Gender Discrimination Essay

500 words gender discrimination essay.

Gender discrimination is when there are unfair rights between male and female. It differs because of their gender roles which ultimately leads to unequal treatment in life. Gender discrimination has been around for many centuries. However, as we are evolving, it is time to do away with such notions of gender roles. Thus, gender discrimination essay will take us through it in detail.

Causes of Gender Discrimination

There are many causes of gender discrimination. The first one has to be illiteracy . When people do not educate themselves, they continue to live in the old times. Thus, they follow the old-age sexist traditions and norms.

Education can bring about a change in this mindset because educated people will less likely partake in gender discrimination. Further, poverty is also another reason which is interlinked in a way.

It is the root cause in many places because the economic dependence remains on the male counterparts mostly. Thus, women suffer a lot from it because of the same reason. They never get out of this and stay financially dependent on men.

Furthermore, the patriarchal setup in our society plays a big role. In this setup, the male dominates almost every aspect of life. Thus, they consider themselves to be superior to others.

This way, a lot of violence and injustice is meted out against females. Thus, when there is a gender considering themselves to be superior, it becomes difficult for everyone to avail equal opportunities.

Impact of Gender Discrimination

Gender discrimination has a deep impact on society as a whole. It does not just impact a specific section of the society but every part of it. First of all, it impacts children as they fall prey to gender stereotypes from a young age.

Further, it impacts young people because it impacts their behaviour, study choices, ambitions, attitudes and more. Thus, many girls do not participate in many sports and women experience physical violence more than men.

Next up, we have gender discrimination affecting adults because there is a gender pay gap between the working class. Men earn more for doing the same work as women. In addition, older women have more risk of becoming homeless than men.

It also impacts the aboriginal women because they have it a lot worse. It is more likely to happen that they can die from family violence, 11 times more than men. Even for men, it is not beneficial as it sets difficult standards for men to follow.

It draws a line on men being emotional. Thus, they can never showcase their emotions truly without being judged. Similarly, men do not parental leave in many places. Ultimately, all this results in more suicide in men. Thus, it impacts everyone.

Get the huge list of more than 500 Essay Topics and Ideas

Conclusion of Gender Discrimination Essay

Gender discrimination must be checked at every stage so that no person should be denied a chance to learn and grow. Thus, everyone, no matter male or female, must get a start in life in terms of educations and other opportunities. We must come together as a society to do this.

FAQ on Gender Discrimination Essay

Question 1: Who is affected by gender inequality?

Answer 1: Gender inequality affects everyone, which includes men as well. Stereotypes about how women and men, girls and boys should be, start from their childhood and follow us to adulthood. Thus, it does not affect just one but all.

Question 2: Give an example of gender discrimination.

Answer 2: There are many examples of gender discrimination. For instance, restriction on clothing. If a man wears shorts, no one will bat an eye. However, if a woman wears shorts, she will be seen in a bad light and be called names. Similar is the case for housework.

Customize your course in 30 seconds

Which class are you in.

• Travelling Essay
• Picnic Essay
• Our Country Essay
• My Parents Essay
• Essay on Favourite Personality
• Essay on Memorable Day of My Life
• Essay on Knowledge is Power
• Essay on Gurpurab
• Essay on My Favourite Season
• Essay on Types of Sports

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Download the App