Gender Violence and Work: Reckoning with the Boundaries of Sex Discrimination Law GENDER VIOLENCE AND WORK: RECKONING WITH THE BOUNDARIES OF SEX DISCRIMINATION LAW JULIE GOLDSCHEID* It is uncontroverted ... that Hossack was terminated because management feared her husband's threats and that he might very well cause workplace disruption in the future ... [consequently] no reasonable jury could find that the defendant terminated [her] employment because she is a woman. 1 Workplace inequality based on sex, as well as discrimination based on other protected characteristics, persist notwithstanding several decades of antidiscrimination laws.2 An extensive body of scholarship examines this * Julie Goldscheid, Associate Professor, CUNY School of Law. Thanks to Dean Michelle Anderson and the CUNY Law School Faculty Development program for supporting this project and thanks to participants in the CUNY faculty workshop, in the Second Annual Colloquium on Current Scholarship in Labor and Employment Law, at the 2008 Law & Society Conference, and at the New York City-area LGBT faculty workshop, at which previous drafts were presented. Thanks as well to Rachel Arnow-Richman, Sally Goldfarb, Nancy Levitt, Rick Rossein, Robin Runge, Catherine Smith, and Deb Widiss for their helpful comments. Thanks to Morgan Cashwell, Seeta Persaud, Daniela Arregui, and Nate Treadwell for their able research assistance. 1 Hossack v. Floor Covering Assocs., 492 F.3d 853, 863 (7th Cir. 2007), reh 'g & reh 'g en banc denied, No. 04-3990, 2007 U.S. App. LEXIS 23794 (7th Cir. Oct. 2, 2007). 2 According to the most recent census data, women working full time year-round earn approximately seventy-seven cents for every dollar earned by men. U.S. CENSUS BUREAU, HISTORICAL INCOME TABLES-PEOPLE, TABLE P-40: WOMEN'S EARNINGS AS A PERCENTAGE OF MEN'S EARNINGS BY RACE AND HISPANIC ORIGIN: 1960 TO 2007, available at http://www.census.gov/hhes/www/income/histinc/p40.html. Minority women fare significantly worse, with African American women earning just fifty-two cents for every dollar earned by a white man, while a Hispanic woman earned only forty-three cents on the dollar. Compare U.S. CENSUS BUREAU, CURRENT POPULATION SURVEY, 2006 ANNUAL SOCIAL AND ECONOMIC SUPPLEMENT, TABLE PINC-05: WORK EXPERIENCE IN 2005-PEOPLE 15 YEARS OLD AND OVER BY TOTAL MONEY EARNINGS IN 2005, AGE, RACE, HISPANIC ORIGIN, AND SEX, FEMALE 15 YEARS AND OVER BLACK A.O.I.C. (2006), available at http://pubdb3.census.gov/macro/032006/perinc/new05_ 113.htm (reporting mean income for black females as $26,916), and U.S. CENSUS BUREAU, CURRENT POPULATION SURVEY, 2006 ANNUAL SOCIAL AND ECONOMIC SUPPLEMENT, TABLE P[NC-05: WORK EXPERIENCE IN 2005-PEOPLE 15 YEARS OLD AND OVER BY TOTAL MONEY EARNINGS IN 2005, AGE, RACE, HISPANIC ORIGIN, AND SEX, FEMALE 15 YEARS AND OVER HISPANIC (OF ANY RACE) (2006) 62 Columbia Journal of Gender and Law [Vol. 18:1 and related forms of persistent inequality and identifies a number of approaches, such as reforms that would target cognitive bias, entrenched structures of decision-making, and unconscious discrimination, to redress what has been termed "second generation" employment discrimination.3 This Article takes that project in an unexplored direction. The treatment of domestic and sexual violence 4 survivors at work often reflects a subtle5 form of sex discrimination that antidiscrimination law currently fails to adequately reach.6 This subtle bias inevitably informs and distorts workplace decisions involving domestic and sexual violence victims. 7 This Article proposes a "second generation" discrimination solution that requires employers to engage in an interactive process with survivors of gender- based violence before taking adverse actions against them. This would bring the issue to the surface, reduce stigma, and ensure that any adverse employment action is based on legitimate reasons, as opposed to subtle biases. By identifying and accounting for this nuanced manifestation of sex discrimination, antidiscrimination law can more fully advance its goal of eliminating sex-based inequality at work.8 available at http://pubdb3.census.gov/macro/032006/perinc/new05 1 7.htm (reporting mean income for Hispanic women as $22,550), with U.S. CENSUS BUREAU, CURRENT POPULATION SURVEY, 2006 ANNUAL SOCIAL AND ECONOMIC SUPPLEMENT, TABLE PINC-05: WORK EXPERIENCE IN 2005-PEOPLE 15 YEARS OLD AND OVER BY TOTAL MONEY EARNINGS IN 2005, AGE, RACE, HISPANIC ORIGIN, AND SEX, MALE 15 YEARS AND OVER WHITE ALONE, NOT HISPANIC (2006), available at http://pubdb3.census.gov/macro/032006/perinc/ new05_058.htm (reporting mean income for white, non-Hispanic men as $52,147). 3 Susan Sturm, Second Generation Employment Discrimination: A Structural Approach, 101 COLUM. L. REV. 458, 460 (2001). See infra Parts I.A. and III.A. for further discussion of this scholarship. 4 This Article references domestic and sexual violence alternately as gender violence, gender-based violence, and, in the most descriptive though somewhat cumbersome form, domestic and sexual violence. 5 This Article uses the term "subtle bias" generally to reference expressions of gender bias that are not captured under current interpretations of antidiscrimination law. 6 For a discussion of covered and excluded cases, see Part II, infra. 7 For a discussion of the ways this subtle bias informs workplace decisions, see Part lI.B., infra. 8 By focusing on the ways antidiscrimination law responds to the problem, this Article's proposal would directly affect those who work in the formal labor market, in jobs governed by Title VII, and, potentially, by state analogs that follow Title VIl's approach to statutory interpretation. Therefore, this Article does not address the important question of how the law could better address the impact of domestic and sexual violence on the work lives of informal labor market employees and those who work for small employers outside 2008] Gender Violence and Work 63 Feminist advocacy increasingly emphasizes the economic dimensions of gender violence as a complement to the traditional focus on criminal justice, family law, and social service reforms. 9 United States v. Morrison, one of the Supreme Court decisions that most explicitly analyzed domestic and sexual violence, hinged on the proper Congressional response to the economic effects of gender-based violence.10 Nevertheless, the legal impact of domestic and sexual violence on women's"1 market-based employment has received limited scholarly attention.' 2 In the last decade, policymakers, advocates, and human resources professionals have begun to recognize the impact of domestic and sexual violence on women's employment.' 3 A growing literature documents the ways abusers use the workplace to perpetuate abuse, the costs of abuse to employers, and the detrimental effects of abuse on women's employment. 14 For example, abusers may stalk their partners at work, call incessantly to distract them from their jobs, or use coercive measures to keep them from the reach of antidiscrimination law. See generally ANNETTE BERNHARDT ET AL., THE GLOVES-OFF ECONOMY: WORKPLACE STANDARDS AT THE BOTTOM OF AMERICA'S LABOR MARKET (2008); Shirley Lung, Developing a Course on the Rights of Low-Wage Workers, 54 J. LEGAL EDUC. 380 (2004). 9 See infra notes 23-26 and accompanying text. 10 United States v. Morrison, 529 U.S. 598 (2000). 1 Although not all domestic and sexual violence is committed against women, this Article primarily uses female references as a generalization because the vast majority of domestic and sexual violence victims are women. See, e.g., CALLIE MARIE RENNISON, U.S. DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS, INTIMATE PARTNER VIOLENCE, 1993-2001, at 1 (2003), available at http://www.ojp.usdoj.govibjs/pub/pdf/ipv0l.pdf (concluding that eighty-five percent of all victimizations by intimate partners in 2001 were against women); BUREAU OF JUSTICE STATISTICS, U.S. DEPT OF JUSTICE, CRIMINAL VICTIMIZATION 1N THE UNITED STATES, 2006 STATISTICAL TABLES 15 tbl.2 (2008), available at http://www.ojp.usdoj.gov/bjs/pub/pdf/cvus060l.pdf [hereinafter CRIMINAL VICTIMIZATION] (reporting that eighty-eight percent of all rapes and sexual assaults in 2006 were committed against women). However, the proposal laid out in this Article would apply to at least a subset of the men who are victims of domestic or sexual violence as well. See infra Part II.A. 12 See infra notes 97-103. 13 See infra Part I.B. 14 id. Columbia Journal of Gender and Law succeeding at work.15 Employers have begun to develop policies and programs to assist victims and to promote workplace safety.16 However, even as awareness of the impact of domestic and sexual violence in the workplace grows, and as appropriate responses expand, women continue to be terminated, denied positions, and subjected to other adverse job actions because of their experiences with domestic or sexual violence. 7 Given the deep connections between gender bias and domestic and sexual abuse, in at least some of these cases subtle bias may skew both an employee's perception of available choices and an employer's response if she self-identifies as experiencing abuse. As a practical matter, the growing experience of survivors1 8 and their employers demonstrates that both parties' interests are best served when employees feel safe enough to disclose their experiences of abuse and work with their employer to determine how they can most safely and productively navigate both the abuse and their employment. 19 Yet current legal frameworks neither require nor encourage such an approach. To some extent, antidiscrimination law has long recognized that sexual violence at work is a form of sex discrimination. It has been more than twenty years since the Supreme Court ruled that sexual assault at work creates a hostile environment that violates Title VII. 20 But domestic and sexual violence affect women's employment in a range of ways, which include, but are not limited to, sexual assault at work.21 Inevitably, the ingrained, gender-based biases that inform both the circumstances in which abuse is committed and the response victims receive from officials, including law enforcement, health care, and employers, will inform at least '5 See infra Part I.B. 1. 16 See infra Part I.B.3. 'v See infra Part I.B.I & II.B. 18 This Article uses the terms "survivor" and "victim" interchangeably, reflecting both the common perception of those who experience domestic and sexual violence as "victims" and the importance of recognizing that those who have weathered the abuse should accurately be described as "survivors." "9 See infra Part I.B.3. 20 Meritor Sav. Bank v. Vinson, 477 U.S. 57 (1986) (holding that repeated sexual assaults by supervisor constitute impermissible sex discrimination). See also infra notes 160-164. 21 See infra Part 1.B.l. [Vol. 18:1 Gender Violence and Work some decisions about survivors' employment status.22 This Article proposes a more robust interpretation of antidiscrimination law's sex bias prohibitions that would require engagement and negotiation between employer and employee about workplace issues that result from abuse. Requiring antidiscrimination law to more fully account for the nuanced expressions of sex discrimination is important for a number of reasons. As an initial matter, it allows antidiscrimination law to more meaningfully fulfill its goal of eliminating discrimination in the workplace. This approach also incentivizes workplace policies that encourage interactive problem-solving. Businesses and human resources professionals increasingly recognize that such policies promote safe and productive workplaces. Moreover, it holds transformative potential by exposing the gender bias that continues to inform domestic and sexual violence. Part I of this Article provides an overview of the literature on unconscious bias and how it can be applied to the circumstances that survivors of gender violence encounter when they try to maintain their jobs in the face of abuse. It summarizes that body of literature and links it to the extensive scholarship framing domestic and sexual violence as a form of sex discrimination. It then grounds the problem of gender violence and work in studies and statistics detailing the ways gender violence impacts the workplace. It places the discussion of antidiscrimination law in context by surveying the range of applicable laws and recent legislative initiatives as well as the emerging scholarly commentary. It describes the workplace policies and practices that best address the issue. This discussion highlights the importance of workplace policies that encourage victims to disclose their experiences of abuse, but do not require them to do so. That disclosure enables employees and employers to engage in good faith dialogue about how best to implement safety measures that can both reduce the risk of violence in the workplace and promote victims' employment security. Part II discusses the application of antidiscrimination law. It analyzes the common situations in which domestic and sexual violence result in adverse actions against victims. It categorizes cases that are covered, versus those that are excluded, under current antidiscrimination paradigms. These cases demonstrate the ways that subtle gender bias negatively impacts women's employment in cases involving abuse-related workplace issues. Part III proposes a framework under which antidiscrimination law can better account for subtle forms of sex discrimination associated with domestic and sexual violence. Antidiscrimination law should recognize 22 See infra Part I.B. 2008] Columbia Journal of Gender and Law unexplained adverse actions taken against victims of domestic and sexual violence as sex discrimination if those actions are taken without a good faith negotiation between the employer and employee. Assuming an employee facing employment issues has voluntarily disclosed her experience with abuse, her employer would engage in an interactive process about employment issues and whether a modest accommodation would facilitate her ongoing safe and productive employment. By requiring a conversation with the victim, this approach would interrupt the operation of subtle bias. Absent a required interaction, employers often react reflexively and overreact to survivors' employment-related issues. Although a similar result could be accomplished through statutory reform that, for example, explicitly prohibited adverse actions against survivors absent reasonable accommodation, the proposed framework advances both the instrumental goal of better ensuring women's equality, and the normative goal of interpreting antidiscrimination law in a way that responds to the nuanced realities of discrimination. I. UNCONSCIOUS BIAS, GENDER VIOLENCE, AND WORK Domestic and sexual violence advocates, academics, and policymakers have in recent years begun to highlight the economic impact of domestic and sexual abuse. This shift has been sparked by a growing recognition of the limitations of the criminal justice system23 and also by mounting evidence that a woman's economic independence is one of the best predictors of whether she will seek and maintain safety in the aftermath of abuse.24 Since work is central to economic security, and given the 23 See, e.g., Naomi Cahn, Policing Women: Moral Arguments and the Dilemmas of Criminalization, 49 DEPAUL L. REV. 817 (2000); Donna Coker, Crime Control and Feminist Law Reform in Domestic Violence Law: A Critical Review, 4 BUFF. CRIM. L. REV. 801, 840-41, 858-60 (2001); Jenny Rivera, Domestic Violence Against Latinas by Latino Males: An Analysis of Race, National Origin, and Gender Differentials, 14 B.C. THIRD WORLD L.J. 231, 243-51 (1994); Nina W. Tarr, Employment and Economic Security for Victims of Domestic Abuse, 16 S. CAL. REV. L. & SOC. JUST. 371 (2007); Deborah M. Weissman, The Personal is Political-and Economic: Rethinking Domestic Violence, 2007 B.Y.U. L. Rev. 387, 398-403. 24 See, e.g., HUGH WATERS ET AL., WORLD HEALTH ORGANIZATION, THE ECONOMIC DIMENSIONS OF INTERPERSONAL VIOLENCE (2004), available at http://whqlibdoc.who.int/publications/2004/9241591609.pdf, CTR. FOR DISEASE CONTROL & PREVENTION, NAT'L CTR. FOR INJURY PREVENTION & CONTROL, COSTS OF INTIMATE PARTNER VIOLENCE AGAINST WOMEN IN THE UNITED STATES (2003), available at http://www.cdc.gov/ncipc/pub-res/ipv cost/IPVBook-Final-Feb18.pdf [hereinafter CDC STUDY]; DONNA GRECO & SARAH DAWGERT, PA. COAL. AGAINST RAPE, POVERTY AND [Vol. 18:1 2008] Gender Violence and Work 67 prevalence of gender violence--committed predominantly by men against women 25-the impact of domestic and sexual violence on women's work is a key issue for both safety and equality. 26 Moreover, survivors of gender violence inevitably will be subjected to the outdated, gendered stereotypes that surround abuse, both inside and outside the workplace.2 7 Data indicating that survivors of domestic and sexual violence often lose their jobs or suffer other employment-related losses as a result of the abuse highlight the importance of further inquiry into the ways that workplace responses to survivors may perpetuate women's inequality. At the same time, recent employment law scholarship recognizes and critiques the limits of antidiscrimination law's traditional disparate treatment, disparate impact, and facial discrimination theories for ridding the workplace of impermissible bias.28 Instead of deliberate discrimination, cognitive bias, structures of decision-making, and patterns of interactions increasingly are recognized as driving and perpetuating inequality at work.29 Numerous scholars draw on social psychology studies as a basis for arguments that antidiscrimination law's traditional focus on intentional discrimination is inadequate and instead should take into account the role of implicit bias.3° Indeed, Christine Jolls and Cass Sunstein have characterized the legal literature on implicit bias as "enormous.' SEXUAL VIOLENCE: BUILDING ADVOCACY AND INTERVENTION RESPONSES (2007), available at http://www.pcar.org/resources/poverty.pdf. 25 See CRIMINAL VICTIMIZATION, supra note 11, at 15 tbl.2. 26 See Emily F. Rothman et al., How Employment Helps Female Victims of Intimate Partner Violence: A Qualitative Study, 12 J. OCCUPATIONAL HEALTH PSYCHOL. 136, 138-42 (2007) (concluding based on empirical study that work helps victims by improving finances, promoting physical safety, increasing self esteem, improving social connectedness, providing mental respite, and providing motivation or a "purpose in life"). See also, e.g., infra Parts l.B.1 & I.B.3. 27 See infra Part II.B. 1. 28 Susan Sturm has termed these challenges "second generation" manifestations of workplace bias. Sturm, supra note 3, at 460. 29 Id. 30 See, e.g., Tristin K. Green, Discrimination in Workplace Dynamics: Toward a Structural Account of Disparate Treatment Theory, 38 HARV. C.R.-C.L. L. REV. 91, 95-99 (2003); Anthony G. Greenwald & Linda H. Krieger, Implicit Bias: Scientific Foundations, 94 CAL. L. REV. 945 (2006); Melissa Hart, Subjective Decision-Making and Unconscious Discrimination, 56 ALA. L. REV. 741, 745-50 (2005); Linda H. Krieger & Susan T. Fiske, Behavioral Realism in Employment Discrimination Law: Implicit Bias and Disparate Treatment, 94 CAL. L. REV. 997 (2006); Linda H. Krieger, The Content of Our Categories: A Columbia Journal of Gender and Law This Part first demonstrates the importance of antidiscrimination law to domestic and sexual violence survivors. It shows how emerging understandings of implicit bias open the door to re-examining the subtle operation of impermissible biases, including gender bias. It then grounds the argument for analyzing the reactions that survivors of gender violence face at work through the lens of antidiscrimination law by reviewing the global literature recognizing gender bias as a form of sex discrimination. This Part next describes the problems that gender violence victims face at work more fully, by presenting social science and statistical data, legal and scholarly responses, and the growing body of human resources materials recommending best practices in response to the problem. A. Surfacing Subtle Discrimination: Why Anti discrimination Law Matters Recent scholarship describing the role of unconscious or implicit bias adds a new lens through which to reconsider the role and application of antidiscrimination law. At the risk of oversimplification, a central premise of the implicit bias literature is that stereotypes and biases are natural cognitive responses employed by everyone, not just "prejudiced" people, to categorize information in a complex world.3 2 The research demonstrates that these stereotypes and biases affect inter-group judgment and decision- making, and generally operate unintentionally. 33 Stereotypes subconsciously predispose the stereotype holder to perceive, characterize, 34 and behave toward a stereotyped target in stereotype-consistent ways. Legal rules can counter these unconscious biases by, for example, prohibiting implicit biases from affecting workplace decisions. 35 Research Cognitive Bias Approach to Discrimination and Equal Employment Opportunity, 47 STAN. L. REV. 1161 (1995); Charles R. Lawrence 11, The Id, the Ego, and Equal Protection: Reckoning with Unconscious Racism, 39 STAN. L. REV. 317 (1987); Ann C. McGinley, i Viva La Evoluci6n!: Recognizing Unconscious Motive in Title VII, 9 CORNELL J. L. & PUB. POL'Y 415, 421-26 (2000); David B. Oppenheimer, Negligent Discrimination, 141 U. PA. L. REV. 899, 901 (1993). 31 Christine Jolls & Cass R. Sunstein, The Law ofImplicit Bias, 94 CAL. L. REV. 969, 975 n.31 (2006). 32 Krieger, supra note 30, at 1187-88. 33 Id. 34 Krieger & Fiske, supra note 30, at 1032-33. 35 Jolls & Sunstein, supra note 31, at 977-88. Jolls and Sunstein group their proposals into three categories: those in which law or policy insulates outcomes from the [Vol. 18:1 Gender Violence and Work indicating that contact with people who are the subject of stereotypes helps to reduce bias offers support for initiatives that increase population diversity in workplaces, educational institutions, and other organizations. 36 Similarly, research indicating that recurrent positive imagery can reduce implicit bias suggests the use of workplace training that includes positive imagery of historically stereotyped groups. 37 Along similar lines, parties in workplace discrimination litigation should be permitted to argue that implicit stereotypes produced purportedly discriminatory employment decisions. 38 Other approaches focus on structural forms of bias and demonstrate that patterns of interaction lead to the exclusion of non-dominant groups.3 9 Under that view, complex organizational cultures perpetuate exclusion; accordingly, discrimination can be fully discerned only when examined in relation to broader patterns of conduct and access. 40 The absence of systematic institutional reflection about these insidious patterns contributes to their cumulative effect. 41 Therefore, some scholars advocate a problem- solving approach that relies on institutional actors to systemically reflect on the patterns of exclusion and to create institution-specific responses.4 2 Viewed broadly through the lens of this body of scholarship, the workplace impact of domestic and sexual violence can be seen as a form of "second generation discrimination" that reflects and perpetuates women's inequality. Domestic and sexual violence may appear to be "gender neutral," in that these acts may be committed by and against both women and men. In reality, though, they are inextricably connected to gender discrimination in a general, rather than an individual sense, by virtue of their disproportionate impact on women as victims, the surrounding social and historical context, and the responses victims often receive from law effects of implicit bias; those in which legal or policy dictates directly reduce implicit bias; and those termed "indirect debiasing mechanisms" in which law encourages or enables regulated actors to take steps that reduce implicit bias. Id. at 989. 36 Id. at 981. 31 Id. at 983-84. 38 Krieger & Fiske, supra note 30, at 1061. 39 Sturm, supra note 3, at 468-69. 40 Id. at 471. 41 Id. 42 Id. Sturm goes on to offer examples of the "dynamic regulatory regime" she proposes. Id. at 489-520. 2008] Columbia Journal of Gender and Law enforcement, health care personnel, and other institutional actors.43 To better address the full range of cases in which women's work is affected by bias, antidiscrimination law should require a more searching analysis of whether domestic and sexual violence victims' employment is adversely affected because of gender. Another reason to closely investigate and analyze the ways workplace treatment of domestic and sexual violence victims constitutes sex discrimination lies in the value of publicly and formally framing domestic and sexual violence in terms of its connection with sex discrimination.44 In the past decade, employer awareness of the workplace impacts of domestic and sexual violence has expanded, workplace policies have been developed, and public policy increasingly addresses the problem. These are critical developments. However, much of this reform is cast in terms of cost reduction and safety.45 Those concerns are critically important, both as ends in themselves and for their instrumental value in producing critical programmatic and policy developments. But to the extent that workplace interventions are framed solely, or even primarily, in terms of workplace safety and productivity, and not in terms of sex discrimination, something is lost. Though it may be politically unpopular to talk in terms of discrimination, given backlash against existing protections, employer fatigue about antidiscrimination laws, and fears of unchecked liability, something is lost if the problem, and the accompanying legal responses, are framed solely in the gender-neutral terms of services and safety.46 These concerns are of course appropriate and important, but not to 43 See infra notes 47-55 and accompanying text. 44 For further discussion on how framing domestic and sexual violence as a problem of equality can help shift public perception of the problem see, e.g., ELIZABETH M. SCHNEIDER, BATTERED WOMEN AND FEMINIST LAWMAKING 27-28, 34-49 (2000), Sally F. Goldfarb, Applying the Discrimination Model to Violence Against Women: Some Reflections on Theory and Practice, 1 J. GENDER, SOC. POL'Y & L. 251, 254-59 (2003); G. Kristian Miccio, A House Divided: Mandatory Arrest, Domestic Violence, and the Conservatization of the Battered Women's Movement, 42 Hous. L. REV. 237 (2005). See also, e.g., Julie Goldscheid, Elusive Equality in Domestic and Sexual Violence Law Reform, 34 FLA. ST. U. L. REV. 731, 763-67, 776-77 (2007). 45 See infra Part 1.B.3. 46 For discussion of the mainstreaming of the anti-domestic violence movement and its attendant limitations, see, for example, LISA A. GOODMAN & DEBORAH EPSTEIN, LISTENING TO BATTERED WOMEN 36-47 (2008). Recent initiatives reassert the importance of addressing the social context of abuse. See, e.g., NATIONAL DOMESTIC VIOLENCE HOTLINE, DECADE FOR CHANGE SUMMIT REPORT 10, 29-31 (2007), available at http://www.ndvh.org/decadeforchange/Decade%20for/ 20Change%20SUMMIT%2Report. [Vol. 18:1 2008] Gender Violence and Work the exclusion of dialogue about the ways domestic and sexual violence is rooted in gender-based inequality. Authorities ranging from the United Nations to the United States Congress recognize that domestic and sexual violence are inextricably linked to formal and informal manifestations of sex discrimination.47 International legal instruments recognize domestic and sexual violence as forms of sex discrimination.48 Domestic violence continues to be committed overwhelmingly by men against women, in the United States49 and elsewhere around the world. 50 The nature of the violence itself also is deeply gendered. It is committed in the context of cultural norms sanctioning or ignoring male violence against women, and in a legal tradition that historically countenanced such abuse. 5' The circumstances pdf [hereinafter DECADE FOR CHANGE] (recognizing the importance of social context in addressing domestic violence). 47 See, e.g., Declaration on the Elimination of Violence Against Women, G.A. Res. 48/104, U.N. Doc. A/RES/48/104 (Dec. 20, 1993) (stating that violence against women is an "obstacle to the achievement of equality" and that it "constitutes a violation of the rights and fundamental freedoms of women," and recognizing that "violence against women is a manifestation of historically unequal power relations between men and women, which have led to domination over and discrimination against women by men."); Global AIDS & Tuberculosis Relief Act of 2000, Pub. L. No. 106-264, § 122, 114 Stat. 748, codified as 22 U.S.C. § 6822 (2009) (grant program implicitly recognizing rape and sexual assault as a form of gender-based violence); see also, e.g., International Violence Against Women Act of 2007, S. 2279, 110th Cong. § 4(1) (2007) (defining gender-based violence as including, inter alia, domestic and sexual violence). 48 See, e.g., Convention on the Elimination of All Forms of Discrimination Against Women, Dec. 18, 1979, 1249 U.N.T.S. 13 (entered into force Sept. 3, 1981); Committee on the Elimination of Discrimination Against Women, 11 th Sess., General Recommendation No. 19 (1992), available at http://www.un.org/womenwatch/daw/cedaw/recommendations/ recomm.htm. 49 See RENNISON, supra note 11, at 1. 5 0 See, e.g., WORLD HEALTH ORGANIZATION, MULTI-COUNTRY STUDY ON WOMEN'S HEALTH AND DOMESTIC VIOLENCE AGAINST WOMEN (2005), available at http://www.who.int/gender/violence/whomulticountrystudy/en/index.html (comparing data from fifteen cites in ten countries). 51 For a discussion of the gendered context of domestic and sexual violence, see, e.g., SCHNEIDER, supra note 44, at 13-23; SUSAN SCHECHTER, WOMEN AND MALE VIOLENCE: THE VISIONS AND STRUGGLES OF THE BAT-FERED WOMEN'S MOVEMENT (1982); Sally F. Goldfarb, Violence Against Women and the Persistence of Privacy, 61 OHIO ST. L.J. 1 (2000); Catharine A. MacKinnon, Reflections on Sex Equality Under Law, 100 YALE L.J. 1281 (1991); and see also, e.g., Julie Goldscheid, Domestic and Sexual Violence as Sex Discrimination: Comparing American and International Approaches, 28 T. JEFFERSON L. REV. 355, 358-62 (2006) (collecting sources). 72 Columbia Journal of Gender and Law [Vol. 18:1 associated with particular acts of abuse often, though not always, reveal perpetrators' conscious or unconscious biases.52 The attitudes of law enforcement officers, health care providers, and employers also reflect gender-based biases that continue to animate treatment of survivors.13 Authorities including the Centers for Disease Control recognize that the culture of masculinity and enduring biased beliefs, attitudes, and perceptions of women fuel a culture in which domestic and sexual violence, committed primarily against women, persist.54 Employers' responses to domestic and sexual violence victims inevitably will be informed, and, in at least some cases, skewed, by the gender-based biases that continue to surround domestic and sexual violence. 55 These attitudes may be explicit in some cases, but in others may underlie what may appear to be gender-neutral actions. Discerning actionable bias is complex, but it lies at the heart of antidiscrimination laws' mandate. 52 See, e.g., Goldscheid, supra note 44, at 773 n.196 (citing cases); see also Walter S. DeKeseredy & Molly Dragiewicz, Understanding the Complexities of Feminist Perspectives on Woman Abuse, 13 VIOLENCE AGAINST WOMEN 874, 875, 877-78 (2007) (summarizing research and refuting arguments that domestic violence is not gendered). 53 See, e.g., Nat'l Judicial Educ. Project, Gender Bias Task Forces: Summary and Recommendations, http://www.nowldef.org/html/njep/findings.shtml (last visited Apr. 8, 2009) (summarizing reports of thirty-one state and federal gender-bias task forces, including findings about gender bias in cases involving violence against women); see also, e.g., Goldscheid, supra note 44, at 773 n.197 (citing examples from law enforcement); Judicial Inquiry & Rev. Comm'n of Va. v. Shull, 651 S.E.2d 648, 659 (Va. 2007) (removing judge from bench following judge's admission that he "probably" advised a female litigant, who alleged that her boyfriend had inflicted bruises on her, that "if you married this guy, it would remove an impediment [regarding custody of your children]"; and that he had required a woman to pull down her pants in court so that he could see her wound before he agreed to extend a protective order against her estranged husband). 54 See, e.g., Ctr. for Disease Control and Prevention, Domestic Violence Prevention Enhancement and Leadership Through Alliances (DELTA), http://www.cdc.gov/ncipc/DELTA/default.htm (last visited Apr. 8, 2009) (describing prevention-oriented program involving ten demonstration projects addressing, inter alia, the societal factors underlying domestic and sexual abuse). For a discussion of one program that treats domestic violence as a form of sexism, see Cindy Rodriguez, Class Teaches Respect for Women to Batterers (WNYC radio broadcast Aug. 14, 2008), available at http://www.wnyc.org/news/articles/105891. 15 See infra Part II.B. 1. 2008] Gender Violence and Work 73 B. Grounding the Problem of Gendered Violence and Work To more fully describe the connections between gender bias, gender violence, and work, this Section reviews social science and statistical data, legal and scholarly responses, and the growing human body of resource materials recommending best practices in response to the problem. 1. Backgro und Data A growing literature documents the complex and often hidden ways that domestic and sexual violence affect the workplace.56 Statistics consistently show that one in four women will be assaulted by an intimate partner in her lifetime, while one in six women will experience a completed or attempted rape.57 Given the high rate of women's workplace participation, the odds are high that at any time a sizable percentage of a given employer's workforce will be subjected to domestic or sexual violence. This impacts the workplace in a variety of ways. Abusers may commit acts of domestic or sexual violence at the workplace, sometimes with tragic results.58 Workers may be sexually assaulted at work by supervisors, coworkers, or clients. Workplace violence is increasingly 56 See, e.g., CDC STUDY, supra note 24, at 30, 42; CHRISTINE LINDQUIST ET. AL., CTR. FOR DISEASE CONTROL & PREVENTION, INVENTORY OF WORKPLACE INTERVENTIONS DESIGNED TO PREVENT INTIMATE PARTNER VIOLENCE 1-1 (2006) [hereinafter CDC INVENTORY]; Press Release, Bureau of Labor Statistics, U.S. Dep't of Labor, Survey of Workplace Violence Prevention 2005 (Oct. 27, 2006) 2 tbl.l, available at http://www.bls.gov/iif/oshwc/osch0033.pdf [hereinafter Survey of Workplace Violence Prevention] (showing percent of establishments reporting an incident of domestic violence in the previous year); DECADE FOR CHANGE, supra note 46 (recognizing impact of domestic violence on employment and recommending set of responses). For a collection of studies, see, e.g., Deborah A. Widiss, Domestic Violence and the Workplace: The Explosion of State Legislation and the Need for a Comprehensive Strategy, 35 FLA. ST. U. L. REV. 669, 675-80 (2008). 57 PATRICIA TJADEN & NANCY THOENNES, U.S. DEP'T OF JUSTICE, FULL REPORT OF THE PREVALENCE, INCIDENCE, AND CONSEQUENCES OF VIOLENCE AGAINST WOMEN 13, 26 (2000), available at http://www.ncjrs.gov/pdffilesl/nij/183781.pdf. 58 See, e.g., Robin H. Thompson, Domestic Violence and its Effects on the Workplace, in THE IMPACT OF DOMESTIC VIOLENCE ON YOUR LEGAL PRACTICE 364 (Am. Bar Ass'n Comm'n on Domestic Violence ed., 2d ed. 2004) [hereinafter IMPACT ON YOUR LEGAL PRACTICE]; Stacey P. Dougan & Kimberly K. Wells, Domestic Violence: Workplace Policies and Management Strategies, 7 A.B.A. COMM'N ON DOMESTIC VIOLENCE E-NEWSLETTER I (2007). See also infra note 85 (sources citing cases). 74 Columbia Journal of Gender and Law [Vol. 18:1 recognized as a drain on productivity and morale; 59 domestic and sexual violence figures prominently as part of that equation. For example, homicide is the leading cause of death for female workers, 60 and sixteen percent of those homicides reportedly were the result of domestic violence. 61 Nearly a quarter of businesses surveyed by the United States Department of Labor in 2000 reported having experienced an incident of workplace violence stemming from domestic violence over the previous year. 62 Domestic and sexual violence that occur outside of work also affect the workplace. The central issues in domestic violence are power, dominance, and control, rather than violence per se.63 A batterer's need to control may cause him to challenge steps his target makes towards financial independence.64 Accordingly, he may engage in a range of behaviors that will undermine her ability to succeed in work, school, or other training programs.65 Researchers increasingly are identifying economic abuse as a distinct, but pervasive, form of abuse.66 59 See infra notes 77-82 and accompanying text; see also Nicole B. Porter, Victimizing the Abused? Is Termination the Solution When Domestic Violence Comes to Work?, 12 MICH. J. GENDER & L. 275, 276-77 (2006) (referencing studies). 60 NAT'L INST. FOR OCCUPATIONAL SAFETY & HEALTH, HOMICIDE IN THE WORKPLACE (1996), available at http://www.cdc.gov/niosh/violhomi.htmi (analyzing data from 1980-1992). 61 Bureau of Labor Statistics, U.S. Dep't of Labor, Women Experience Fewer Job- related Injuries and Deaths than Men, ISSUES IN LABOR STATISTICS (July 1998), available at http://www.bls.gov/opub/ils/pdf/opbils23.pdf. 62 Survey of Workplace Violence Prevention, supra note 56, at 2 tbl. 1. 63 See, e.g., EVAN STARK, COERCIVE CONTROL: HOW MEN ENTRAP WOMEN IN PERSONAL LIFE (2007); Martha R. Mahoney, Legal Images of Battered Women: Redefining the Issue of Separation, 90 MICH. L. REV. 1, 53-60 (1991); Tarr, supra note 23, at 385 (citing sources). 64 Tarr, supra note 23, at 386. Some programs have begun to address the ways abusers use economic control to prevent victims from attaining independence by offering financial and economic empowerment information to victims. See, e.g., Economics Against Abuse Program, http://www.econempowerment.org (last visited Apr. 8, 2009). 65 Tarr, supra note 23, at 376-77 (listing tactics such as: harassing her in the workplace in person, on the phone, and by email; causing her to be late; undermining her transportation; interfering with her child care arrangements); see also Maria A. Calaf, Breaking the Cycle: Title VII, Domestic Violence, and Workplace Discrimination, 21 LAW & INEQ. 167, 170-71 (2003) (same); accord GEN. ACCOUNTING OFFICE, DOMESTIC VIOLENCE: PREVALENCE AND IMPLICATIONS FOR EMPLOYMENT AMONG WELFARE RECIPIENTS (1998), 2008] Gender Violence and Work 75 Recent studies offer a snapshot of the ways batterers interfere with their partners' work. One revealed that abusers' "pre-work" tactics, such as physically restraining a partner from going to work, beating her severely enough that she could not or did not want to go to work, keeping her from sleeping, making a car unavailable, or cutting up work clothes, prevented fifty-six percent of women in that study from going to work.67 Abusers also interfered at work by showing up at the workplace, making harassing phone calls to victims and their supervisors, and stalking from places outside, but nearby, the workplace.68 In one example, a batterer kept his target on the phone for hours.69 These findings confirm previous studies finding that between thirty-five and fifty-six percent of employed battered women were harassed at work by their batterers, and that fifty-five to eighty-five percent missed work because of the abuse.7° Similarly, the Society for Human Resource Management reported that eleven percent of employees reported facing violence from a girlfriend or boyfriend at work, while ten percent 71reported violence from a spouse and seven percent from a former spouse. Abusers' tactics jeopardize women's jobs in both direct and indirect ways. One recent study found that over ninety percent of the employed battered women surveyed had either resigned or been terminated as a result available at http://www.gao.gov/archive/1999/he99012.pdf [hereinafter GAO STUDY] (describing abusers' tactics and citing studies); see also, infra notes 67-71. 66 Adrienne E. Adams et al., Development of the Scale of Economic Abuse, 14 VIOLENCE AGAINST WOMEN 563 (2008) (reviewing literature and explaining development of instrument to measure economic abuse). 67 Jennifer E. Swanberg & T.K. Logan, Domestic Violence and Employment: A Qualitative Study, 10 J. OCCUPATIONAL HEALTH PSYCHOL. 3, 6-8 (2005). 68 Id. at 6-8. 69 Id. at 10. 70 GAO STUDY, supra note 65, at 7-9, 18-19 (summarizing studies). For other studies similarly recounting abusive tactics that interfere with women's work, see Angela M. Moe & Myrtle P. Bell, Abject Economics: The Effects of Battering and Violence on Women's Work and Employability, 10 VIOLENCE AGAINST WOMEN 29 (2004) (detailing tactics and demonstrating that battering impacts women's work and employability across various employment levels and backgrounds). See also Carol Reeves & Ann M. O'Leary-Kelly, The Effects and Costs of Intimate Partner Violence for Work Organizations, 22 J. INTERPERSONAL VIOLENCE 327 (2007) (reporting on a workplace-based survey that revealed that victims of intimate partner violence were more likely than non-victims to be absent, and that current victims report higher levels of distraction at work than non-victims). 71 SOC'Y FOR HUMAN RES. MGMT., WORKPLACE VIOLENCE SURVEY 5 (2004). Columbia Journal of Gender and Law of the abuse in the previous two years.72 Virtually all of those who resigned did so without giving adequate notice, for reasons including their shame about their appearance after being physically abused, fears for their children's safety if they had to leave them with the batterer, or embarrassment about their batterer's harassing behavior on the job.73 Over forty percent of women in that study were terminated from at least one job in the previous two years, either due to abuse-related poor attendance or poor performance, or their abuser's harassment at work.74 This is consistent with previous studies finding, for example, that twenty-four to fifty-two percent of battered women surveyed had lost their jobs as a result of the abuse.75 It is not surprising that the Centers for Disease Control recently estimated that among the costs of domestic violence were $5.8 billion in productivity losses and health care costs. 76 Even if a batterer does not harass her at work, a woman navigating an abusive relationship may need time off for court dates or medical appointments. She may need time to find new housing, childcare, or other services if she is relocating in an effort to separate from the abuser. This may jeopardize her status at work, because the time off may cause performance problems or may force her to exceed sick day or leave allotments. Her need for time off will be particularly problematic if she does not feel safe enough to disclose the reason for her absence. Ultimately, like an abuser's harassment at work, lateness and absence associated with abuse also may cause her to lose her job. Other studies focus on the impact of sexual assault. The Center for Disease Control found that over twenty percent of women who were raped and over seventeen percent of those physically assaulted by an intimate partner lost time from paid work, totaling over 561,000 lost days of work per year. 77 Victims who were raped or physically assaulted lost an average 72 Swanberg & Logan, supra note 67, at 9. " Id. at 10. 74 Id. 75 GAO STUDY, supra note 65, at 7-9, 18-19 (summarizing studies). 76 See CDC STUDY, supra note 24; see also BUREAU OF NAT'L AFFAIRS, SPECIAL REPORT No. 32, VIOLENCE AND STRESS: THE WORK-FAMILY CONNECTION 2 (1990) (estimating that domestic violence costs employers between three and five billion dollars annually). 77 CDC STUDY, supra note 24, at 19, 42 tbl.14. [Vol. 18:1 2008] Gender Violence and Work of seven to eight days of paid work as a result of the crime. 78 The overall value of rape victims' lost productivity resulting from lost days of work has been estimated to be over $580 million each year.79 Few studies have analyzed the specific impact of sexual assault on victims' continued employment. However, the one published study to address the issue found that one half of sexual assault victims studied reported losing their job in the aftermath of the assault.80 This is consistent with an informal study of sexual assault victims who were working at the time of the incident, which also revealed that over half of them either lost income, had to take time off without pay, or had to quit as a result.81 Many employers have begun to recognize that domestic and sexual violence negatively affect their "bottom line" and expose them to potential liability. 82 A number have accordingly adopted policies and developed programs to assist domestic and sexual violence victims.8 3 Notably, the human resources literature predominantly addresses the impact of domestic violence, and not sexual assault. 84 In part, this may be explained by the 78 id. '9 Id. at 40 tbl. 12. 80 S. REP. No. 103-138, at 54 (1993) (citing Elizabeth Ellis et al., An Assessment of the Long Term Reaction to Rape, 90 J. ABNORMAL PSYCHOL. 263, 264 (1981)); see also Karen S. Calhoun et. al, Social Adjustment in Victims of Sexual Assault, 49 J. CONSULTING & CLINICAL PSYCHOL. 705, 710 (1981) (analyzing impact of sexual assault on victims' social adjustment and finding that work was the area of functioning affected for the longest period of time following the assault, as compared with social and leisure activities, and relations with marital, parental, and family units, or extended family). 81 Safe Horizon, Rape and Sexual Assault: Effects of Incident on Employment and School (2003) (survey results on file with author). 82 See, e.g., SOC'Y FOR HUMAN RES. MGMT., supra note 71, at 6. 83 Perhaps the most visible and organized employer response to the problem to date is the Corporate Alliance to End Partner Violence, http://www.caepv.org (last visited Apr. 8, 2009). 84 This disparity is troublesome. Of course, the distinction between domestic and sexual violence is, to some extent, misleading, given that most sexual assaults are committed by acquaintances. CRIMINAL VICTIMIZATION, supra note 11, at 56 tbl.43 (approximately sixty-eight percent of rapes and sexual assaults are committed by acquaintances, relatives, and intimates). However, not all sexual assaults fit into the rubric of domestic violence, and victims may have different needs and experiences. An analysis of why domestic violence seems to have garnered more traction than sexual assault in public discussions no doubt would shed light on the enduring nature of the gender biases associated with both forms of violence. That analysis, however, is beyond the scope of this Article. 78 Columbia Journal of Gender and Law [Vol. 18:1 publicity surrounding cases in which abusers commit acts of violence at work, prompting the phrase that domestic violence "comes to work., 8 ' The impact of sexual assault on the workplace has commanded less public attention, although it is no less important. Legislative and public policy initiatives increasingly address both domestic and sexual violence. 86 Although some of the workplace issues surrounding domestic violence and sexual assault differ, this Section addresses them together where appropriate. This is based on the premises that domestic and sexual violence are committed on a continuum of violence that is committed primarily by men against women, that both are rooted in gender-based bias, and that the responses of officials, including employers, to both reflect related gender- based stereotypes. 2. Law Reform and Legal Scholarship The complex dynamics produced when abuse impacts the workplace implicate tort law, workers' compensation, workplace safety regulations, leave provisions, unemployment insurance, and wrongful discharge and antidiscrimination laws.87 Courts grapple with the 85 For a discussion of those cases, see John E. Matejkovic, Which Suit Would You Like? The Employer's Dilemma in Dealing with Domestic Violence, 33 CAP. U.L. REV. 309 (2004); Stephanie L. Perin, Employers May Have to Pay When Domestic Violence Goes to Work, 18 REV. LITIG. 365 (1999); see also, e.g., Tarr, supra note 23, at 377 (citing State v. Byars, 823 So.2d 740, 741 (Fla. 2002), in which a defendant violated a domestic violence injunction prohibiting him from entering the shop where his wife worked, and murdered her); Christi Lowe, Shootings Shine Spotlight on Domestic Violence, WRAL, Apr. 2, 2009, http://www.wral.com/news/local/story/4876987/ (reporting investigation into whether domestic violence was cause of shootings at North Carolina nursing home, given reports by gunman's estranged wife and her mother that the couple had a history of domestic violence). 86 See, e.g., Violence Against Women and Department of Justice Reauthorization Act of 2005, Pub. L. 109-162, 119 Stat. 2960 (2006). 87 For articles surveying the applicability of existing laws, see Calaf, supra note 65 (focusing on Title VII remedies); Andrea Giampetro-Meyer et al., Raped at Work. Just Another Slip, Twist, and Fall Case?, 11 UCLA WOMEN'S L.J. 67 (2000); Matejkovic, supra note 85; Sandra S. Park, Working Towards Freedom from Abuse: Recognizing a "Public Policy" Exception to Employment-At- Will for Domestic Violence Victims, 59 N.Y.U. ANN. SURV. AM. L. 121 (2003) (addressing wrongful discharge claims); Perin, supra note 85; Porter, supra note 59; Jill C. Robertson, Addressing Domestic Violence in the Workplace: An Employer's Responsibility, 16 LAW & INEQ. 633 (1998); Robin R. Runge, Employment Rights of Sexual Assault Victims, 40 CLEARINGHOUSE REv. 299 (2006); Tarr, supra note 23, at 391; Wendy R. Weiser & Deborah A. Widiss, Employment Protection for Domestic Violence Victims, 38 CLEARINGHOUSE REV. 3 (2004); Widiss, supra note 56; see also, e.g., Ellen M. Bublick, Tort Suits Filed by Rape and Sexual Assault Victims in Civil Courts: Lessons for 2008] Gender Violence and Work 79 applicability of these laws to cases in which domestic and sexual violence victims, for example, allege that their employer bears liability for damages when negligent security practices permit an abuser to commit an act of violence at work;88 seek leave from work to cope with the abuse; 89 file for unemployment benefits if they have been terminated as a result; 90 or assert that their termination violated public policy. 91 Recently, a number of jurisdictions have considered, and in some cases, enacted, abuse-specific legislation that provide specific protections for victims. 92 For example, some new laws explicitly mandate provision of leave time to address the abuse,93 payment of unemployment insurance benefits to survivors who leave their jobs or are terminated as a result of abuse,94 and require development of workplace policies addressing the problem. 95 Other provisions prohibit employment and housing discrimination against victims. 96 Courts, Classrooms and Constituencies, 59 SMU L. REV. 55 (2006) (discussing employer tort liability for rape and sexual assault). 88 These cases tend to bring the most media attention to the issue. For a fuller discussion of the tort liability that may arise when an abuser commits an act of violence at work, see generally, Matejkovic, supra note 85; Perin, supra note 85. 89 See Tarr, supra note 23, at 413; Weiser & Widiss, supra note 87, at 7-10; Widiss, supra note 56, at 695. 90 Rebecca Smith, Richard W. McHugh & Robin R. Runge, Unemployment Insurance and Domestic Violence: Learning from our Experiences, 1 SEATTLE J. FOR Soc. JUST. 503 (2002). 91 See Park, supra note 87; Widiss, supra note 56, at 713. 92 See Widiss, supra note 56. For fact sheets listing new and proposed laws, see Legal Momentum, Employment and Housing Rights for Victims of Domestic Violence: State Law Guides, http://www.legalmomentum.org/site/PageServer?pagename=erhdv_16 (follow "For Survivors: State Law Guides" hyperlink) (last visited Apr. 8, 2009). 93 See Legal Momentum, Time Off from Work for Victims of Domestic or Sexual Violence (Aug. 2004), http://www.legalmomentum.org/site/DocServer/leave.pdf. 94 See Legal Momentum, Unemployment Insurance Benefits (Feb. 2008), http://www.legalmomentum.org/site/DocServer/UItfeb_08.pdf. 95 See Legal Momentum, Domestic and Sexual Violence Workplace Policies (July 2007), http://www.legalmomentum.org/site/DocServer/policies.pdf. 96 See Legal Momentum, Employment Rights for Victims of Domestic and Sexual Violence (May 2008), http:l/www.legalmomentum.org/site/DocServer/Employment Rights. May.08.pdfdocID=2721; Legal Momentum, Housing Discrimination Against Victims of Columbia Journal of Gender and Law Of the small but developing body of legal scholarship addressing the appropriate legal response, a number of articles focus on employers' potential tort liability when an abuser commits an act of violence at the workplace. 97 Others analyze the applicability of antidiscrimination laws and reach varying conclusions about the current availability of remedies under Title VII's traditional theories of recovery. For example, Tarr concludes that federal law offers few, if any, remedies for domestic violence.98 She argues instead that legal responses that strike a balance between protectionism and autonomy will best serve survivors' needs. 99 Calaf argues that the disparate impact theory offers the most useful approach for employees terminated as a result of abuse."'0 Porter focuses on the dilemmas employers face when balancing the workplace safety risks domestic violence can pose. 0 1 After identifying the weaknesses in traditional disparate treatment and disparate impact theories, she draws from the Americans with Disabilities Act and argues that a victim of abuse permissibly could be terminated if her abuser creates a sufficiently serious workplace safety risk.10 2 Widiss describes the possibilities and limitations of current disparate treatment and impact theories in her analysis of the emerging legislative responses.l13 This Article argues that these scholars too quickly dismiss the applicability of existing antidiscrimination law. Survivors necessarily will invoke a range of traditional and recent legislative and common law remedies to redress the harms resulting from the impact of gender violence Domestic and Sexual Violence (Aug. 2008), http://www.legalmomentum.org/site/ DocServer/Housing_02 27 07.pdf'?doclD=1041. 97 See, e.g., Matejkovic, supra note 85; Perin, supra note 85. 98 Tarr, supra note 23, at 391-95. 9' Id at 426-27. 100 Calaf, supra note 65, at 186-91. See infra notes 155-57 and accompanying text. 101 Porter, supra note 59, at 320-25. 102 Id. at 327-29. Porter would support a survivor's termination if the abuser poses a "direct threat." Id. at 328. According to Porter, the "direct threat" analysis requires that four factors be met: 1) the duration of the risk; 2) the likelihood that the potential harm will occur; (3) the nature and severity of the potential harm; and 4) the imminence of the potential harm. Id. The question of the appropriate balance between employee job security and employer discretion to terminate an abuse victim is addressed in Part III, infra. 103 Widiss, supra note 56, at 684-85. [Vol. 18:1 Gender Violence and Work on their work lives. Antidiscrimination law is a critical piece of this remedial mix both because it complements other available remedies and, as a normative matter, because it is charged with reaching the full range of discriminatory practices that perpetuate impermissible inequalities at work. 3. Workplace Responses Over the last decade or so, employers have come to recognize the widespread impact of domestic and sexual violence on the workplace. 0 4 The impact is both direct, as when the victim is harassed, threatened, or attacked at work, and indirect, as through absenteeism and decreased productivity of victims and perpetrators. 10 5 A growing consensus of management and human resources experts urge employers to take a proactive approach to the issue and to develop comprehensive policies, with the goal of reducing the risk of workplace violence and retaining valued employees. 1 06 Model policies have a dual focus on safety and productivity. Experts encourage employers to provide information about domestic and 104 See, e.g., Corporate Alliance to End Partner Violence, CEO & Employee Survey 2007: Corporate Leaders and America's Workforce on Domestic Violence, http://www.caepv.org/about/program detail.php?reflD=34 (last visited Apr. 8, 2009) (a significant majority of corporate executives and employees recognize the harmful and extensive impact of domestic violence on the workplace, while recognizing many employers' reticence to take action); Kelley Holland, Strife at Home Affects the Office, Too, N.Y. TIMES, Oct. 28, 2007, at C17. 105 CDC I'VENTORY, supra note 56; accord Soc'Y FOR HUMAN RES. MGMT., supra note 71, at 5-6, 9-15; Lois G. REcKTT & LAURA A. FORTMAN, MAINE DEP'T OF LABOR, IMPACT OF DOMESTIC OFFENDERS ON OCCUPATIONAL SAFETY & HEALTH: A PILOT STUDY (2004), available at http://www.maine.gov/labor/labor-stats/publications/dvreports/domestic offendersreport.pdf. '06 See, e.g., CDC INVENTORY, supra note 56, at 1-5; Dougan & Wells, supra note 58, at 2; Stacey P. Dougan, Employers May Face Liability When Domestic Violence Comes to Work, EMP. BENEFIT PLAN REv., Feb. 2003, at 3, available at http://www2.gtlaw.com/pub/articles/2003/dougans03a.asp; Jane A. Randel & Kimberly K. Wells, Corporate Approaches to Reducing Intimate Partner Violence Through Workplace Initiatives, 3 CLINICAL OCCUPATIONAL & ENVTL. MED. 821, 824 (2003); Jennifer E. Swanberg et al., Intimate Partner Violence, Employment, and the Workplace: Consequences and Future Directions, 4 TRAUMA, VIOLENCE & ABUSE 21 (2005). In addition, see generally Corporate Alliance to End Partner Violence, http://www.caepv.org (last visited Apr. 8, 2009); Family Violence Prevention Fund, Strategic Employer Responses to Domestic Violence, http://www.endabuse.org/workplace (last visited Apr. 8, 2009); Legal Momentum, Domestic and Sexual Violence Workplace Policies, http://www.legalmomentum.org/ site/DocServer/recommendedpolicyprovisions.pdf?doclD=521 (last visited Apr. 8, 2009). 2008] Columbia Journal of Gender and Law sexual violence resources in the community and to offer support to employees who are victims, while maintaining policies that clearly prohibit use of employer resources to commit acts of violence.10 7 This type of comprehensive approach can reduce an employer's potential liability by reducing the risk of violence occurring at work.108 Best practices for larger employers include establishing a "multidisciplinary response team" with representatives including senior management, human resources, health/medical, legal, security, internal communications and media, community, employee assistance programs, and unions. 10 9 Smaller employers can develop analogous groups designated to respond to incidents or threats of violence when they occur. 1 ° Experts emphasize that employee benefits such as leave policies, flexible hours, and working shift assignments should be administered flexibly to enable survivors to tend to abuse-related legal, medical, and social service needs."' They recommend training and awareness initiatives that alert employees about the nature of the problem and the availability of resources for those affected by it. "12 An essential part of this approach is employers' role in encouraging victims to seek help." 3 While employers cannot and should not take on the role of counselor, they can make a critical difference for employees coping with abuse by encouraging them to develop an individualized workplace safety plan." 4 For domestic violence victims, safety planning is key to seeking and maintaining safety. 1 5 In the workplace, this means evaluating 107 See, e.g., Dougan & Wells, supra note 58, at 3-4. 108 See, e.g., Dougan, supra note 106, at 3; CDC INVENTORY, supra note 56, at 1-7. 109 See, e.g., Dougan & Wells, supra note 58, at 2-3. 110 See, e.g., Family Violence Prevention Fund, Solutions for Small and Large Employers: Safety and Security Concerns, http://endabuse.forumone.com/workplace/ display.php?DoclD=33008 (last visited Apr. 8, 2009). 111 See, e.g., Randel & Wells, supra note 106, at 826. 112 See, e.g., id. at 827-28. 113 See, e.g., id. at 828; Robertson, supra note 87, at 654. 114 Randel & Wells, supra note 106, at 828; Dougan & Wells, supra note 58, at 3. 115 See, e.g., Donna Mathews & Deborah M. Goelman, Safety Planning, in IMPACT ON YOUR LEGAL PRACTICE, supra note 58, at 40; A.B.A. Comm'n on Domestic Violence, Domestic Violence Safety Plan: Safety Tips for You and Your Family, http://www.abanet.org/tips/publicservice/DVENG.pdf (last visited Apr. 8, 2009) (including tips for safety planning at work). [Vol. 18:1 Gender Violence and Work what steps may be needed to enhance a survivor's safety at work. A victim is in the best position to assess when she is at risk and to identify the strategies that will increase her safety and decrease the possibility of abuse. 1 16 Safety planning entails conducting a risk assessment for each step of a survivor's daily routines. With respect to the workplace, this can include considerations such as: the need to change regular travel habits; keeping emergency numbers and a copy of orders of protection, where applicable, 117 on file at work; giving a copy of a picture of the abuser to security and others at work; and evaluating whether other modifications, such as changing work locations or phone numbers, would increase an employee's safety. 118 For example, an employee who is continually harassed by an abuser's phone calls can have her phone number changed. Someone who needs time off from work to meet with her lawyer or a domestic violence or sexual assault counselor may be able to keep her job if she is afforded the time off. She may reduce the risk that an abuser will commit an act of violence at work if she is able to register an order of protection with security.1 19 However, all of these approaches are premised on the employee's disclosure of the abuse to someone at work, presumably a person in a position of some authority. 20 The decision to disclose any personal problem to someone at work is difficult; the decision to disclose one's experience with domestic or sexual violence is all the more complicated by fear of the stigma that continues to be associated with the experience of abuse.' 2' At least one recent study indicates that telling someone at work increases the likelihood that a victim will be able to retain her job in the face of abuse. 122 116 Sally F. Goldfarb, Reconceiving Protection Orders for Domestic Violence: Can Law Help End the Abuse Without Ending the Relationship?, 29 CARDOZO L. REV. 1487, 1541-42 (2008). 117 Id. at 1542. 118 See Dougan & Wells, supra note 58, at 6. 119 Only if she discloses the abuse can she work with other appropriate staff to develop a workplace response strategy. See Dougan & Wells, supra note 58, at 2-3. 120 Id; see also Randel & Wells, supra note 106, at 828. 121 The role of silence and disclosure is discussed more fully in Part II.B.2., infra. 122 Jennifer Swanberg, Caroline Macke & T.K. Logan, Working Women Making It Work: Intimate Partner Violence, Employment, and Workplace Support, 22 J. INTERPERSONAL VIOLENCE 292, 305 (2007). 2008] Columbia Journal of Gender and Law The receipt of workplace support, such as schedule flexibility, phone call screening, and employer assistance with a work-related safety plan, also has been shown to increase women's chances of keeping their jobs in the face of abuse.1 23 Nevertheless, current legal frameworks fail to incentivize disclosure and offer no assurance that a woman will be immune from retaliation if she does disclose her experience with abuse. II. GENDER VIOLENCE AND WORK: ANTIDISCRIMINATION LAWS' REACH AND LIMITS As one commentator recently noted, common sense would dictate that when an employee is terminated because she is a victim of domestic violence, that termination would be seen as impermissibly discriminatory.124 In some cases, discrimination against a domestic or sexual violence victim can form the basis for a relatively straightforward claim of sex discrimination. 12 However, antidiscrimination laws operate to preclude claims by domestic and sexual violence victims whose experiences do not fit prescribed circumstances. Notwithstanding the documented ways in which domestic and sexual violence survivors lose their jobs in the aftermath of abuse,' 26 notably few lawsuits have asserted employment discrimination claims on their behalf. This may be explained by a combination of factors, such as victims' reluctance to self-identify; their lack of understanding that their employment rights may have been violated; and lawyers' failure to recognize that a domestic violence victim may have suffered employment discrimination based on her sex. Her counsel may correctly determine that antidiscrimination law as it is presently applied would not cover her claim. The failure explicitly to recognize the insidious role sex discrimination plays in these cases is particularly troubling given: 1) the persistence of sex- based differentials in pay and position notwithstanding antidiscrimination laws; 2) the invisibility of the connections between domestic and sexual violence and sex discrimination; 3) domestic and sexual violence victims' continuing reluctance to disclose that they have suffered abuse; and 4) the mounting documentation that the predominantly female victims of domestic and sexual violence lose their jobs, at least in part, due to the abuse. 123 Id. at 304, 306. 124 See Porter, supra note 59, at 292. 125 See infra Part II.A. 126 See supra Part I.B. 1. [Vol. 1 8:1 Gender Violence and Work These considerations raise the question of whether antidiscrimination law accurately captures the discriminatory workplace harms of domestic and sexual violence. The following discussion describes the categories of cases that fit comfortably in current antidiscrimination models, as well as those that do not. It then argues that the cases that fall outside traditional frameworks often reflect the operation of subtle biases and invisible stereotypes, which antidiscrimination law can and should address. A. Covered Cases It will be no surprise that several categories of cases in which domestic or sexual violence victims are subjected to adverse employment actions may be actionable under federal or state antidiscrimination laws. This section will review how disparate treatment, disparate impact, and sexual harassment theories currently apply to those circumstances. 1. Similarly Situated Employees The hallmark of disparate treatment claims is proof that an employee suffered an adverse job action because of her membership in a protected class. 127 Without statutory employment protection for domestic or sexual violence victims, 128 a domestic or sexual violence victim's most likely claim (absent evidence of discrimination based on race, religion, or another protected category) would be that by virtue of suffering an adverse action due to her status as a domestic or sexual violence victim, she suffered discrimination based on sex. In some cases, a domestic violence victim may be able to establish that she suffered disparate treatment through comparator evidence showing that a female victim of domestic violence was treated differently than a similarly situated male. This will most frequently arise in cases in which both the perpetrator and victim work for the same employer, where the 127 See, e.g., Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e (2009) (prohibiting unlawful employment practices "because of" an individual's race, color, religion, sex, or national origin). For a discussion of the frameworks used to establish disparate treatment claims, see, e.g., LEx LARSON, EMPLOYMENT DISCRIMINATION § 8 (2008). 128 See supra notes 93-97 and accompanying text (discussing reform at the state level). 2008] Columbia Journal of Gender and Law employer knows of the abuse and nevertheless penalizes the victim (typically the woman) but not the perpetrator.1 29 For example, in Rohde v. K.O. Steel Castings, Inc.,' 30 Rohde, a secretary who had worked for K.O. Castings for six years, was involved with Arnulfo Lopez, a cleaning room foreman at the same company. Company officials were aware of their relationship. 13' At some point, Lopez became abusive towards Rohde. Rohde told the company's personnel director of two incidents in which Lopez assaulted her, and the director allowed her to take time off (counting against her vacation time).1 32 However, when the company president and vice president learned of Rohde's absence, she was terminated. 33 In upholding her disparate treatment claim, the Fifth Circuit Court of Appeals concluded: Where two employees were engaged in an altercation and the aggressor went unpunished while the victim, a member of a minority protected by the Act, bore the full brunt of retribution, it is clear to us that Congress intended a cause of action to lie in the absence of a sufficient explanation of nondiscriminatory reasons for the disparity. We are unable to discern such an explanation, as was the district court, and its judgment must be affirmed. 134 Other plaintiffs have presented similar facts, but their cases were resolved before a final ruling on their disparate treatment claims. 35 129 As Porter recognizes, disparate treatment theory might also apply when a battered woman seeks a benefit given to the men in the workplace. Porter, supra note 59, at 293. However, I have yet to encounter a case involving these facts. 130 649 F.2d 317 (5th Cir. 1981). Others have noted the applicability of this case. See, e.g., Porter, supra note 59, at 293; Tarr, supra note 23, at 393; Widiss, supra note 56, at 684 n.42. 131 Rohde, 649 F.2d at 319. 132 Id. 133 Id. 134 Id. at 323. 135 See, e.g., Valdez v. Truss Components, No. CV98-1310-RE, 1999 U.S. Dist. LEXIS 22957 (D. Or. Aug. 15, 1999) (rejecting employer argument that plaintiff was an independent contractor not covered by Title VII but rejecting wrongful discharge claim); see also cases discussed in Weiser & Widiss, supra note 87, at 5-6. Other cases brought under Title VII have resulted in jury awards but no written decisions. See, e.g., Complaint, Steele v. Snowline Mfg., Inc., No. 06CV0555-MA (Deschutes, Or., Dec. 13, 2007) (on file with author) (alleging, inter alia, gender discrimination when a female employee was fired after [Vol. 18:1 2008] Gender Violence and Work 87 Notwithstanding the evolution of disparate treatment doctrine since the 1981 Rohde decision, 136 its central holding-that adverse action against the victim of aggression who is a member of a protected class constitutes impermissible discrimination-stands as good law. 137 2. Recognized Sex-Based Stereotypes Disparate treatment claims also may be based on evidence that adverse actions were rooted in outdated sex-based stereotypes.138 For example, an employer may assert that a victim of domestic or sexual violence "deserved it," that she is unsuited for employment because she "allowed" herself to be abused or that, if she wasn't strong enough to handle her home life, she would not be strong enough to perform a particular job satisfactorily. 139 Alternatively, an employer might deny a job filing a protective order to keep her abusive coworker from coming to the workplace), reviewed in 10 A.B.A. COMM'N ON DOMESTIC VIOLENCE E-NEWSLETTER (2008), available at http://www.abanet.org/domviol/enewsletter/vol I 0/cases.html (describing jury award). 136 For discussion of the evolution of disparate treatment doctrine, see Martin J. Katz, Unifying Disparate Treatment (Really), 59 HASTINGS L.J. 643 (2008). 137 The Rohde case was not analyzed under the burden-shifting framework for proving a Title VII violation. Under that framework, an employee first would have to establish a prima facie case of disparate treatment by showing that: 1) she is a member of a protected class (women); 2) that she was qualified for her position; 3) that she suffered an adverse employment action; and 4) that the circumstances give rise to an inference of discrimination. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 (1972). The burden of production then shifts to the employer to prove a legitimate nondiscriminatory reason for the adverse action. The employee then has the burden of proving that the employer's stated reason was a pretext for impermissible discrimination. Texas Dep't of Cmty. Aff. v. Burdine, 450 U.S. 248 (1981). If her employer had been able to establish a legitimate nondiscriminatory reason for her termination, such as performance problems, Rohde still might have established that this reason was a pretext for discrimination. Terminating a female victim of abuse while not disciplining a male perpetrator may evince discrimination, particularly if her performance issues were no worse than those of her abusive coworker. Alternatively, such an argument might be analyzed under a mixed motive theory. 138 See Price Waterhouse v. Hopkins, 490 U.S. 228, 251 (1989) (plurality opinion). 139 See, e.g., Eileen Kwesiga et al., Exploring the Literature on Relationships Between Gender Roles, Intimate Partner Violence, Occupational Status, and Organizational Benefits, 22 J. INTERPERSONAL VIOLENCE 312, 317-21 (2007) (linking gendered stigma associated with domestic violence with women's decisions whether to access benefits at work); Swanberg & Logan, supra note 67, at 11-12 (finding that women did not disclose the abuse to someone at work due to fear of job loss, shame about their situation, and a sense that they "should" be able to handle the situation independently). Columbia Journal of Gender and Law or promotion to a sexual assault victim based, for example, on stereotypes that she "asked for it" and consequently could not be counted on to act appropriately in a particular position. In cases involving evidence of those or other sex-based stereotypes, plaintiffs should prevail under established doctrine.1 40 This is not unlike the ways that adverse employment actions taken against caregivers may constitute disparate treatment in violation of Title VII. 141 Nevertheless, current interpretations may miss the subtle operation of bias, thus omitting cases that nevertheless reflect gendered stereotypes. 42 3. Identifiable Policies Against Domestic or Sexual Violence Victims Domestic or sexual violence victims also may assert that an adverse employment action taken against them because of their experience with abuse violates antidiscrimination prohibitions because the actions reflect a policy that, while facially gender-neutral, has a disparate impact based on sex. 143 If an employee can establish such a policy, she need not prove discriminatory intent or motive. 44 Certainly, it is not difficult to establish that domestic and sexual violence have a disparate impact based on the victim's sex. 145 Consequently, if an employer maintained a policy authorizing adverse actions against domestic or sexual violence victims, whether framed as a domestic or sexual violence policy or as an application of a workplace violence policy, it might impermissibly discriminate based on sex. 140 See Price Waterhouse, 490 U.S. at 258. 141 See, e.g., JOAN WILLIAMS, UNBENDING GENDER 101-10 (2000); EEOC, Enforcement Guidance: Unlawful Disparate Treatment of Workers with Caregiving Responsibilities (May 23, 2007), http://www.eeoc.gov/policy/docs/caregiving.html; Eyal Press, Family-Leave Values, N.Y. TIMES MAG., July 29, 2007, at 37 (reviewing developing recognition that discrimination against caregivers violates antidiscrimination laws). 142 See infra Parts II.B. 1-3. 143 Griggs v. Duke Power Co., 401 U.S. 424, 432 (1971). Generally, to establish a Title VII disparate impact claim, a plaintiff would have to prove: 1) a particular employment practice; 2) that causes a disparate impact; 3) based on a protected characteristic. See Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-2(k)(l)(A)(i) (2009). An employer then can rebut the employee's prima facie case by establishing that the policy nevertheless was "job related for the position in question and consistent with business necessity." Id. '44 Griggs, 401 U.S. at 432. 145 See supra note 11 (citing statistics). [Vol. 18:1 Gender Violence and Work Analogous arguments have proved successful against landlords when domestic violence victims lose housing based on policies disadvantaging domestic violence victims. 146 Similar arguments should apply to the workplace in cases in which an employment policy with punitive effects toward domestic violence victims can be shown. 147 However, the realities of employment, as opposed to housing, may explain why fewer such cases have been brought involving work. Notwithstanding the widespread and severe impact of domestic and sexual violence in the workplace, it will be more commonplace and more visible in the home. Landlords and public housing authorities had developed explicit policies, nominally aimed at violence prevention and maintaining safe residences, which require victims of abuse to be evicted.148 In many cases, those formal policies have now been struck down, based on determinations that they discriminate based on sex.149 Despite the technical availability of disparate impact claims in workplace cases, several factors render such claims more difficult to sustain. First, employers are unlikely to maintain explicit policies mandating termination (the employment equivalent of eviction) or other adverse action against domestic violence victims. This may stem in part from employers' limited recognition of the scope of the impact of domestic 146 See, e.g., Bouley v. Young-Sabourin, 394 F. Supp. 2d 675 (D. Vt. 2005) (upholding sex discrimination claim under Fair Housing Act when landlord sought to evict domestic violence victim less than 72 hours after her husband assaulted her); Alvera v. Creekside Village Apts., HUD ALJ No. 10-99-0538-8 (U.S. Dep't of Hous. & Urban Dev., Portland, Or. Apr. 13, 2001) (policy discriminating against domestic violence victims violates prohibitions on sex discrimination); see also, e.g., Emily J. Martin, Fair Housing for Battered Women: Preventing Homelessness Through Civil Rights Laws, 27 CORNERSTONE 11 (2006), available at http://www.aclu.org/pdfs/fairhousingforbatteredwomen072806.pdf (discussing theories and citing cases); Weiser & Widiss, supra note 87, at 6-7 (same); 42 U.S.C. §1437f(c)(9) (2009) (providing, inter alia, that applicants for Section 8 housing vouchers cannot be denied housing because of their experiences with domestic violence, and that an abuser's actual or threatened violence cannot be the basis for adverse decisions regarding the victim's housing); U.S. DEP'T OF Hous. & URBAN DEv., PUBLIC HOUSING OCCUPANCY GUIDEBOOK 216-19 (2003) [hereinafter HUD] (recommending preferences for domestic violence victims and discussing best practices for maintaining victims' housing status); see generally Lenora M. Lapidus, Doubly Victimized: Housing Discrimination Against Victims of Domestic Violence, 1I AM. U.J. GENDER, SOC. POL'Y & L. 377 (2003). 147 See Calaf, supra note 65, at 186-91; Matejkovic, supra note 85, at 336; Porter, supra note 59, at 294-95; Tarr, supra note 23, at 393-94; Weiser & Widiss, supra note 87, at 6-7. 148 See, e.g., Martin, supra note 146, at 6; HUD, supra note 146. 149 Martin, supra note 146, at 6. 2008] Columbia Journal of Gender and Law and sexual violence on the workplace. It may also reflect an employer's correct and laudable recognition that termination is not always necessary or proper. Either way, in the absence of a stated policy to terminate or otherwise disadvantage domestic or sexual violence victims, a plaintiff would have to establish that an adverse action taken against her nevertheless reflected an actionable employment practice.1 50 Although at least one court has recognized that a single decision by an employer qualifies as an actionable employment practice, others indicate that an employment decision that can be shown to have affected only one employee falls outside the reach of a disparate impact claim. 5 ' A plaintiff likely would have to establish that an employer would also treat another domestic violence victim adversely. It may be difficult for a victim to obtain that type of information, particularly given victims' reluctance to self-identify and disclose their status to others in the workplace,1 52 and an employer's easy and plausible defense that it was an employee's unique circumstances, rather than a policy against domestic or sexual violence victims, that underlay its decision. Moreover, a plaintiff may have difficulty establishing the requisite statistical impact. Disparate impact theory arose, and is most often used, in the context of exclusionary hiring practices.' 53 Accordingly, courts generally look to a comparison of the percentage of employees in the protected class in the workplace under the challenged policy with the percentage of qualified individuals in the relevant population pool. 15 4 With an employment (as opposed to hiring) practice, the comparison would work somewhat differently. Here, a court might compare the percentage of IS0 Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-2(k)(l)(A)(i) (2009). A plaintiff might also sustain a claim if she could establish that a (gender neutral) workplace violence policy was applied in a way that disadvantaged domestic or sexual violence victims. I have not identified any such case to date. 151 Compare, e.g., Council 31, AFSCME v. Ward, 978 F.2d 373, 378 (7th Cir. 1992) (upholding disparate impact claim based on single employer action), with Bramble v. Am. Postal Workers Union, 135 F.3d 21, 26 (1 st Cir. 1998) (rejecting disparate impact claim based on evidence that pay decision to date had only affected one employee). 152 See Swanberg & Logan, supra note 67, at 11-12. 153 Calaf, supra note 65, at 187; see also, Michael Selmi, Was the Disparate Impact Theory a Mistake?, 53 UCLA L. REv. 701, 707, 733-53 (arguing that disparate impact theory has not been successful outside of cases challenging employment tests). ' See, e.g., LEX LARSON, EMPLOYMENT DISCRIMINATION § 22.02 (2007). However, not all courts have required comparator evidence. See infra note 155. [Vol. 18:1 Gender Violence and Work employees in a protected group who would be affected by the policy with the percentage of employees in the general workplace. Calaf is correct that generally, this would lead to a conclusion that a policy adversely affecting domestic or sexual violence victims would disproportionately affect women given that women are disproportionately victims of domestic and sexual violence. 155 However, as Porter and Tarr note, at least some courts have required evidence that the policy disproportionately affected members of a protected class in the plaintiffs workplace; in other words, proof of general statistical disparity would not suffice. 156 Nevertheless, not all courts have required proof that a neutral policy disproportionately impacted the protected group in a particular workplace. 157 Although I do not agree with Porter that the question of the appropriate comparison data is settled and that a disparate impact theory categorically is unavailable absent a particularized showing in a given workplace, she is correct that commentators such as Calaf have glossed over this distinction recognized by some courts and accordingly may have overstated the availability of the theory. 158 4. Sexual Harassment The Supreme Court has explicitly recognized that sexual harassment and sexual violence constitute forms of sex discrimination under Title VII. Sexual harassment is actionable under Title VII and analogous state antidiscrimination laws either when the employer conditions employment on sexual favors or where the abuse creates a 155 Calaf, supra note 65 at 187. 156 Porter, supra note 59, at 294-95 (citing, inter alia, analogous cases); Tarr, supra note 23, at 393-94; accord, Ruth Colker, Anti-Subordination Above All: Sex, Race and Equal Protection, 61 N.Y.U. L. REV. 1003, 1034 (1986) (criticizing disparate impact theory for challenging only policies within a particular employer's workplace over a limited time period). 157 Some courts have refused to require a workplace-specific showing. See, e.g., Bradley v. Pizzaco, 939 F.2d 610, 612-13 (8th Cir. 1991) (upholding disparate impact claim based on no-beard policy, relying on studies and expert testimony that such a policy discriminates against black males, who disproportionately suffer from a skin disorder brought on by shaving); accord Abraham v. Graphic Arts Int'l Union, 660 F.2d 811, 819 (D.C. Cir. 1981) (upholding disparate impact claim concerning pregnancy-related leave without analysis of statistical impact in plaintiff's workplace). 158 The difficulties in establishing disparate impact liability in this seemingly appropriate context raise questions whether the doctrine needs revisiting overall. However, such questions are beyond this Article's scope. 20081 Columbia Journal of Gender and Law hostile environment. 159 The landmark case, Meritor Savings Bank v. Vinson, was a case of repeated sexual assault. 160 The Supreme Court concluded without dissent that "[w]ithout question, when a supervisor sexually harasses a subordinate because of the subordinate's sex, that supervisor "discriminate[s] on the basis of sex." 161 The Court easily accepted that forcing employees to "run a gauntlet of sexual abuse" as a condition of employment constituted sex discrimination that violated Title VII. 162 Although commentators have rightly criticized the Court's easy equation of sexual harassment with sex discrimination for inadequately theorizing the ways in which sexual violence constitutes sex discrimination, 163 the Court's recognition of the connection paved the way for lower courts to treat at least some workplace manifestations of sexual violence as sex discrimination. Accordingly, rape or sexual assault in the workplace may violate antidiscrimination laws.' 64 The employer may be liable if the abuse is committed by a supervisor or other agent,165 by a coworker, or by a non- employee such as a customer, provided that the harassment involved the workplace and that the employer knew or should have known of the abuse and failed to take prompt and appropriate remedial action.166 Because 159 See Faraghar v. City of Boca Raton, 524 U.S. 775 (1998); Burlington Indus. v. Ellerth, 524 U.S. 742, 751-66 (1998) (articulating standards for employer liability in sexual harassment claims). 16' 477 U.S. 57, 60 (1986). 161 Id. at 64. 162 Id. 163 See, e.g., Kathryn Abrams, The New Jurisprudence of Sexual Harassment, 83 CORNELL L. REV. 1169, 1173-88 (1998); Katherine M. Franke, What's Wrong with Sexual Harassment, 49 STAN. L. REV. 691, 729-62 (1997); Vicki Schultz, Reconceptualizing Sexual Harassment, 107 YALE L.J. 1683, 1692-1713 (1998). 164 See, e.g., Meritor, 477 U.S. at 57, 66 (supervisor sexually assaulted employee). Sexual assault is also actionable in schools. See Doe v. Sch. Admin. Dist. No. 19, 66 F. Supp. 2d 57, 62 (D. Me. 1999) (sexual assault of student by high school teacher). 165 See, e.g., Smith v. Sheahan, 189 F.3d 529, 533-34 (7th Cir. 1999) (single incident of sexual assault would be enough to establish claim); Brock v. United States, 64 F.3d 1421, 1423 (9th Cir. 1995) (stating that "every rape committed in the employment setting is also discrimination based on the employee's sex"); Jones v. United States Gypsum, No. C99-3047-MWB, 2000 WL 196616 (N.D. Iowa Jan 21, 2000) (upholding sexual harassment claim based on assault in genital area, and citing cases). 166 See, e.g., Little v. Windermere Relocation, Inc., 301 F.3d 958, 964, 968 (9th Cir. 2002) (stating that an employer may be liable for harassment by failing to act after one [Vol. 18:1 Gender Violence and Work sexual harassment laws apply regardless of the personal relationship between the harasser and victim, employers may be liable for sexual harassment when anyone abuses an intimate partner at work, provided the abuse otherwise meets the requirements for a sexual harassment claim.' 67 For example, in Excel Corp. v. Bosley, the Eighth Circuit upheld the sexual harassment claim of Kristine Bosley, whose harassment by her ex-husband and coworker Rock Johnson created a hostile environment at the meat-packing plant at which they both worked. 68 Johnson repeatedly harassed Bosley by calling her names such as "bitch," "slut" or "whore."' 69 He threatened to kill one of Bosley's friends and her future husband, and violated work rules by throwing meat and animal organs at her.170 Bosley repeatedly reported the harassment to management, whose initial response of its employees reported being raped by a client); Hall v. Gus Constr. Co., 842 F.2d 1010, 1012, 1014-16 (8th Cir. 1988) (stating that sexual harassment laws applied where female employees were subjected to unwanted touching and offensive comments by coworkers); Menchaca v. Rose Records, Inc., No. 94-C-1376, 1995 WL 151847, at *2 (N.D. Ill. Apr. 3, 1995) (stating that an employer is not shielded from liability where the harasser is a customer if he "knew or should have known of the harassment"); Otis v. Wyse, No. 93-2349-KHV, 1994 WL 566943, at *6 (D. Kan. Aug. 24, 1994) (stating that an employer may be responsible for alleged harassment by an independent contractor); Hernandez v. Miranda- Velez, No. 92-2701, 1994 WL 394855, at *6 (D.P.R. July 20, 1994) (acknowledging potential employer liability for sexual harassment by one of its clients); Powell v. Las Vegas Hilton Corp., 841 F. Supp. 1024, 1028 (D. Nev. 1992) (stating that a hotel employer could be held liable for the sexual harassment of employees by customer); see also 29 C.F.R. § 1604.1 1(d)-(e) (2006) (EEOC guidelines confirming employers' liability for sexual harassment by coworkers and customers). But see, e.g., Lapka v. Chertoff, 517 F.3d 974 (7th Cir. 2008) (rejecting sexual harassment claim based on sexual assault and subsequent workplace visits by co-worker when employer initiated investigation). 167 For example, the First Circuit recently recognized that an employee's former partner harassed her "based on her sex," even though it rejected her claim after concluding that the employers' response was prompt and appropriate. Forrest v. Brinker Int'l Payroll Co., No. 07-1714, 2007 U.S. App. LEXIS 29300 (1st Cir. Dec. 19, 2007). The court reasoned that "[w]hether a harasser picks his or her targets because of a prior intimate relationship, desire for a future intimate relationship, or any other factor,... should not be the focus of the Title VII analysis." Id. at * 10. It correctly concluded that the focus instead should be on whether gender bias can be inferred from conduct. Id. at *10. See also, e.g., Fuller v. City of Oakland, 47 F.3d 1522, 1525-27, 1529 (9th Cir. 1995) (holding city liable for failing to take steps to stop a police officer from harassing another officer after she ended their relationship). 168 165 F.3d 635 (8th Cir. 1999). 169 Id. at 638. 170 Id. 2008] 94 Columbia Journal of Gender and Law [Vol. 18:1 was to tell both Bosley and Johnson to "keep their disputes at home.' 7 ' Bosley took personal leave due to the stress, but when she returned to work, Johnson continued to harass her. 172 Bosley asked for temporary relief from her workstation to get away from Johnson's harassment; Bosley's supervisor twice refused her requests despite knowledge of Johnson's harassment. 173 Bosley became frustrated, pushed Johnson in the chest, and told him to get out of the area and go back to his assigned post.174 Bosley's supervisor placed her on "indefinite suspension" for violating a work rule prohibiting physical contact between employees. 175 As Bosley was leaving the work area, she saw Johnson in another room. 176 She pushed past a supervisor to enter the room in which she saw Johnson.177 Her employer terminated Bosley, but did not sanction Johnson for his role in any of those events. 1 78 The court upheld Bosley's sexual harassment claim. 179 It found sufficient evidence to support the jury's conclusion that Bosley had been subjected to a hostile work environment; Excel did not appeal that finding. 80 The court concluded that the termination process was tainted by sex discrimination based on management's refusal to act to stop the harassment, even though Bosley had complained repeatedly. 18' In so doing, 171 Id. 172 id. 173 Id 174 Id. 175 id. 176 Id. 177 Id. 178 Id. 179 Bosley apparently did not appeal the court's rejection of her disparate treatment claim, and the Eighth Circuit Court of Appeals does not detail the trial court's reasoning. Excel Corp. v. Bosley, 165 F.3d 635, 638 (8th Cir. 1999). 180 Id. at 639. 181 Id 2008] Gender Violence and Work 95 the court found that this case fit readily into the familiar framework for analyzing sexual harassment claims. 182 B. Excluded Cases Notwithstanding the apparent applicability of antidiscrimination law to some cases involving domestic or sexual violence victims, many cases fall outside of antidiscrimination laws' comfortable reach. The dynamics of abuse, infused with the legacy of historic gender biases, operate to render the role of abuse invisible, creating a dynamic that perpetuates sex-based inequality. 1. Invisible Stereotypes Despite decades of reform, domestic and sexual violence victims continue to be subjected to historic, sex-based stereotypes. These stereotypes are well described in the literature, as are their links to the formal and informal legal and cultural norms sanctioning male violence against women.1 83 My purpose here is to highlight common ways in which 182 The concurring opinion of Judge Loken, which expresses concern about "marital spat[s]" that "spill... over into the workplace," is discussed infra notes 222-24 and accompanying text. 183 For discussions of common stereotypes associated with domestic violence, see SCHNEIDER, supra note 44, at 62, 75, 80 (discussing stereotypes of helplessness, passivity, and purity); Lenore Walker, How Battering Happens and How to Stop It, in BATTERED WOMEN 59 (Donna M. Moore ed. 1979) (describing the response to battering as "learned helplessness" that renders domestic violence victims unable to leave); Naomi Cahn & Joan Meier, Domestic Violence and Feminist Jurisprudence: Towards a New Agenda, 4 B.U. PUB. INT. L.J. 339, 353-55 (1995) (describing pedagogical approaches to dispel the common stereotypes of domestic violence victims as weak, passive, and pathological for "staying" with abuser); Mahoney, supra note 63, at 24-26 (describing stereotypes). For discussions of the somewhat different stereotypes associated with sexual assault, see Michelle J. Anderson, From Chastity Requirement to Sexuality License: Sexual Consent and a New Rape Shield Law, 70 GEO. WASH. L. REV. 51, 54, 64-69 (2002); Kimberl Crenshaw, Mapping the Margins: Intersectionality, Identity Politics, and Violence Against Women of Color, 43 STAN. L. REV. 1241, 1266-67 (1991) (describing stereotype of "innocent" victims of rape as opposed to women who "got what they were asking for"). See generally PEGGY R. SANDAY, A WOMAN SCORNED: ACQUAINTANCE RAPE ON TRIAL (1996). For discussions of the ways these stereotypes are racially skewed, see Sharon A. Allard, Rethinking Battered Woman Syndrome: A Black Feminist Perspective, I UCLA WOMEN'S L.J. 191 (1991) (critiquing the "battered woman syndrome" based on Lenore Walker's "learned helplessness" as consistent with stereotypes of white, not African-American, women); Linda L. Ammons, Mules, Madonnas, Babies, Bathwater, Racial Imagery and Stereotypes: The African-American Woman and the Battered Woman Syndrome, 1995 WIs. L. REV. 1003; Crenshaw, supra, at 1266-68 (describing racial and gendered stereotypes that inform popular conceptions of Columbia Journal of Gender and Law those stereotypes invisibly impact survivors' work experiences, rather than to describe the stereotypes in detail. The stereotypes ascribed to domestic and sexual violence victims are somewhat in tension, with domestic violence victims often cast as weak and passive, and rape victims generally seen as deviant, sexual, and provocative. Nevertheless, there are commonalities. For example, stereotypes operate to lay blame for the abuse on both groups of victims (to the extent they are distinct groups). They cause victims to experience shame and stigma that leave them reluctant to disclose the abuse to family, friends, law enforcement, and other community members, including employers. Both the stereotypes themselves, and the fear of being the target of stereotypes, produce silence and shame that operate to disadvantage women at work. In focusing on the subtle ways in which sex-based biases and stereotypes impact survivors' employment, this Article does not contend that actionable stereotypes operate in every case involving domestic or sexual violence or that every case in which a victim-blaming stereotype operates should give rise to an actionable Title VII violation. Instead, it describes a structural dynamic that illustrates the subtle operation of bias and that often skews domestic and sexual violence survivors' treatment at work. a. Victim-blaming Although the specific expression of the stereotype may vary, domestic violence victims often are blamed for the abuse they experience. 184 Employers may inadvertently base employment decisions on a judgment that a woman poses a risk to the workplace due to her partner's violence, as though she were accountable for or could control his abuse. A recent decision by the Seventh Circuit illustrates how the stereotype operates. In Hossack v. Floor Covering Associates, the court upheld the termination of an exemplary female employee because of her husband's threats and abuse after he learned of her affair with a coworker.18 5 The court rejected her argument that she was similarly situated rape); Zanita E. Fenton, Domestic Violence in Black and White: Racialized Gender Stereotypes in Gender Violence, 8 COLUM. J. GENDER & L. 1 (1998); Paula R. Gilbert, Discourses of Female Violence and Societal Gender Stereotypes, 8 VIOLENCE AGAINST WOMEN 1271 (2002) (discussing gender stereotypes and violent women). 184 Researchers have posited that stigma surrounding abuse negatively impacts women's willingness to access benefits and policies, even when they are available. See Kwesiga et. al., supra note 139. 185 492 F.3d 853, 856 (7th Cir. 2007). [Vol. 18:1 Gender Violence and Work to the (male) employee with whom she was involved for several reasons. First, management feared that "her husband might very well cause workplace disruption."' 86 In addition, the employee with whom she had the affair was "the top salesman and producer" at the store in which they worked. 187 After noting that the employer did not discipline either employee for having an affair, the court concluded that she would be similarly situated "only to other employees who threatened to cause workplace disruption."'188 It therefore rejected her efforts to establish a prima facie case of sex discrimination under the McDonnell Douglas framework.' 89 The court went on to evaluate whether she could present direct or circumstantial evidence of bias. After the trial court found in Hossack's favor and awarded her $250,000, the Seventh Circuit inquired whether she had produced "sufficient evidence for a rational jury to conclude that she was discriminated against and discharged because she is a woman."1 90 The court reversed the jury verdict, finding that Hossack did not produce sufficient evidence of intentional discrimination. Even though the company did not dispute that Hossack's work record was exemplary, and that she was terminated because management feared her husband's violence, the court upheld her termination because the male employee with whom she had the affair was the store's top salesman, and thus, was "more important to the organization."' 91 This case reflects several dimensions of bias. The fact that the male employee was retained due to his superior revenue-producing position may reflect historic, gender-based pay differentials. It essentially solidifies the statistical probability that men work in higher earning positions.,9 2 It perpetuates men's job security at women's expense, even when the woman 186 Id. at 861. 187 id. 188 Id. 189 Id. at 860. The court analyzed her claim as a mixed motive claim; however, it evaluated whether she had established a prima facie case under McDonnell Douglas in order to determine whether she could establish a sex discrimination claim based on the "indirect method" of proof established by McDonnell Douglas. Id. (citing McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-04 (1973)). 190 Id. at 862. 191 Id. at 863. 192 See supra note 2 (citing statistics). 2008] Columbia Journal of Gender and Law was a top performer in her, albeit lower revenue-generating position. But even more troubling for the purposes of this Article is the fact that the court allowed the employee to be terminated based on her husband's threatening abusive behavior. The court thus equated Hossack with her husband, and held her directly accountable for her husband's threats. The court's treatment of Hossack's husband's actions as her own harkens back to coverture, when a husband and wife were treated as one legal person. 193 Moreover, the court blames the victim, jeopardizing the innocent party's employment status for conduct over which she has no control.' 94 The implications of this decision are limited in several respects. First, the opinion does not reveal whether Hossack's husband's abusive behavior extended beyond the threats he made to her coworker. In addition, the employer here, and other employers in similar circumstances may have had legitimate concerns about workplace safety that may, in some cases, warrant altering the conditions of a victim's employment. However, the reflexive termination of the female employee, without consideration of less onerous alternatives that would address any safety concerns without jeopardizing her employment, is both unnecessary and problematic. By failing to acknowledge that she should not be held accountable for her husband's threats, never mind that she may be the victim of his threats as well, the court perpetuates a common sex-based stereotype and contributes to a woman's economic disadvantage. Other cases similarly illustrate employers' reflexive overreaction to an employee's abusive partner at the expense of the female employee's job. Philloria Green's employer terminated her-allegedly due to workplace safety concerns-after she reported that she was raped and beaten by her estranged husband. 195 The employer's concern may have been understandable, given its obligation to take reasonable steps to safeguard 193 See, e.g., 1 WILLIAM BLACKSTONE, COMMENTARIES *430-32 (setting forth system of coverture); SCHNEIDER, supra note 44, at 13-14; Reva B. Siegel, "The Rule of Love ": Wife Beating as Prerogative, 105 YALE L.J. 2117, 2122 (1996). 194 See e.g., Cass R. Sunstein, Civil Rights Legislation in the 1990s: Three Civil Rights Fallacies, 79 CAL. L. REv. 751, 759-60 (1991) (discussing victim-blaming). See generally WILLIAM RYAN, BLAMING THE VICTIM (1976). Although Ryan's critique that the oppressed are blamed for conditions of which they are in fact victims originated in the context of race and poverty, it applies as well to popular attitudes towards subordination of women, including abuse. See, e.g., Mahoney, supra note 63, at 26-27 (describing the victim- blaming mentality applied to battered women); Miccio, supra note 44, at 255 (describing how women have been blamed for their own victimization). 195 Green v. Bryant, 887 F. Supp. 798, 800 n.2 (E.D. Pa. 1995). Ms. Green did not argue that the termination violated antidiscrimination laws. [Vol. 18:1 Gender Violence and Work the workplace from known threats.196 However, nothing in the decision indicates that her former partner threatened the workplace, or that her removal would remove the safety threat. This is another case in which the reflexive termination of the (female) victim of abuse has the effect of punishing the victim without ascertaining whether it in fact would render the workplace any safer. Other examples illustrate that employers' actions may be based on well-meaning efforts to promote workplace safety but may nevertheless excessively discipline the (usually female) victim. For example, a childcare worker alleged that she was fired after telling management that she had an order of protection against her ex-husband. 97 The woman allegedly told security that her ex-husband had called her at work despite a restraining order prohibiting him from doing so. She alleged she was terminated because management was "unsure the campus was safe."' 198 Although an employer should take steps to enhance the safety of the workplace, removing the worker who is the survivor of violence as an initial matter is not the best approach. 199 Removing the worker will not necessarily remove the risk of violence, since her abusive partner could still show up at the workplace and commit an act of violence regardless of whether she still worked there. Accordingly, the response may not be an appropriately calibrated remedy to the perceived problem. The overuse of discipline against domestic or sexual violence victims should raise a red flag about whether the action was motivated, at least in part, by sex-based bias. A similar threat of violence by a stranger to the workplace likely would not lead an employer to terminate the employee who received the threat. 200 Even in a hypothetical case in which an employee knew the person issuing the threat (but was not involved in an intimate abusive relationship with him or her), one would expect the employer to inquire or otherwise engage with the employee about the best 196 Green, 887 F. Supp. at 800. For a description of employers' obligations under tort law, see supra note 88 and accompanying text. 197 Family Violence Prevention Fund, supra note 110 (citing Day care worker sues over firing, CHI. SuN-TIMES, Sept. 3, 2004, at 26). 198 Id. 199 Id. 200 As the Family Violence Prevention Fund materials point out, if a customer obsessed with a sales clerk in a retail business repeatedly threatens the sales clerk, terminating the clerk would not be the appropriate approach to addressing the risk. Family Violence Prevention Fund, supra note 1 10. 2008] Columbia Journal of Gender and Law way to reduce the risk of violence at work. Responses far less intrusive than termination, such as changing an employee's telephone number, or assigning someone to screen calls, may be more closely tailored to addressing the problem. Termination of a domestic violence victim under these circumstances conjures inaccurate sex-based stereotypes that a woman is able to control her abuser's violence, views that may be traced to legal approval of male violence against their partners. 20 1 Alternatively, reflexive termination of the female employee, as a way of "eliminating" the problem, evokes the historic attitude that domestic violence is a private matter that should not be addressed in the public sphere of employment.20 2 b. Misconceptions In other cases, employers' misunderstandings about the nature of abuse disadvantage women's employment. For example, a variation of the victim-blaming stereotype manifests in assumptions that if the victim leaves an abusive relationship the problem will be solved.20 3 Consequently, a survivor may fear that she will be treated differently at work if coworkers find out. Part of this fear may be a concern that she will be blamed for staying in the relationship.20 4 In addition, she may be held accountable for the violence if she does not leave. An employer may make an employee's continued employment contingent on her leaving her abuser. This reflects an inaccurate understanding of the nature of abuse, since separation often increases, rather than decreases, the risk of further abuse. 20 5 Nor is this misconception benign, as it blames the victim and may put her at risk of further violence. 201 Current views of domestic violence victims as responsible for or able to control the abuse may be rooted in historic depictions of domestic violence as caused by women victims. See GOODMAN & EPSTEIN, supra note 46, at 52 (describing psychological theories explaining battering, for example, by women's provocation of the violence). 202 See, e.g., id. at 29-30; SCHNEIDER, supra note 44, at 13-20. 203 This assumption has been criticized as negatively affecting the treatment of domestic violence victims in a wide range of contexts. See Goldfarb, supra note 116; see also, e.g., SCHNEIDER, supra note 44, at 77-79. 204 Robin R. Runge, Employer Liability and Domestic Violence Victim Advocacy, in IMPACT ON YOUR LEGAL PRACTICE, supra note 58, at 376, 377. 205 See, e.g., GOODMAN & EPSTEIN, supra note 46, at 76, 98; Mahoney, supra note 63, at 64-65; Margo Wilson & Martin Daly, Spousal Homicide Risk and Estrangement, 8 VIOLENCE AND VICTIMS 3 (1993) (correlating elevated risk of homicide with separation); Goldfarb, supra note 116, at 1520. 100 [Vol. 18:1 Gender Violence and Work In another variation of the ways subtle biases about abuse impact women's employment, courts may misinterpret a victim's actions and ignore the realities of abuse. For example, in RAP, Inc. v. District of Columbia Commission on Human Rights, Seneta Rose, who worked at a private nonprofit organization with her husband, Greg, was terminated after she and her husband had a heated argument at work.2 °6 Their dispute escalated to the point where Mr. Rose shoved Ms. Rose and knocked her to the ground, upon which Ms. Rose drew a knife from her purse and swung it at her husband.20 7 After investigating the dispute, RAP, Inc., their employer, discharged Ms. Rose and reprimanded Mr. Rose.20 8 It justified its disparate treatment by the fact that she had used a dangerous weapon, whereas he had not.2°9 The District of Columbia Commission on Human Rights agreed with Ms. Rose that the disparate treatment was discriminatory, since the two employees were similarly situated.210 However, the Court of Appeals reversed its findings, and concluded that there was insufficient evidence of discrimination. 211 The appellate court rejected the Commission's conclusion that Ms. Rose used the knife in self defense, and upheld the disparate treatment of these two employees on the basis that Ms. Rose used a knife, 212and Mr. Rose had not. The court ignored the fact that Mr. Rose apparently instigated the violence. More importantly, it failed to credit the substantial data describing the dynamics of abuse that may lead a woman to use a weapon in self defense.21 3 A fuller understanding of the nature of abuse would recognize that an employer should at least inquire whether she used the weapon in self defense, particularly when the facts indicated that it was her husband who was the aggressor. 206 485 A.2d 173, 175 (D.C. Ct. App. 1984). 207 Id. 208 Id. at 176. 209 id. 210 Id. at 177. 211 Id. at 179-80. 212 Id. at 179. 213 See, e.g., Leigh Goodmark, When is a Battered Woman Not a Battered Woman? When She Fights Back, 20 YALE J.L. & FEMINISM 75, 92 (2008). 2008] 102 Columbia Journal of Gender and Law [Vol. 18:1 2. The Bias in Silence Historically, domestic and sexual violence have been shrouded in silence and shame. Victims fear self-identifying because they do not want to be associated with the many labels ascribed to those who are victims of violence and abuse. 214 This self-censure can be traced to the historic relegation of domestic and sexual violence to the private sphere.215 As a result, victims often are reluctant to call law enforcement or to follow up with prosecutors in cases in which law enforcement has been called.216 At work, fear of the stigma associated with abuse often leads women to remain silent about abuse.217 The historic shrouding of domestic violence in silence may contribute to employers' impulse to rid the workplace of the problem, rather than to engage with the victim, as she or he might in other, non- gendered, circumstances. 214 See, e.g., SCHNEIDER, supra note 44, at 61 (discussing negative connotation associated with battering). 215 See generally, e.g., Elizabeth M. Schneider, The Violence of Privacy, 23 CONN. L. REV. 973 (1991); MARTHA A. FINEMAN & ROANNE MYKITIUK, THE PUBLIC NATURE OF PRIVATE VIOLENCE: THE DISCOVERY OF DOMESTIC ABUSE (1994) (collecting excerpted articles). 216 CALLIE MARIE RENNISON, U.S. DEPT. OF JUSTICE, BUREAU OF JUSTICE STATISTICS, SELECTED FINDINGS, RAPE AND SEXUAL ASSAULT: REPORTING TO POLICE AND MEDICAL ATTENTION, 1992-2000, at 2 tbl.3 (2002), available at http://www.ojp.usdoj.gov/ bjs/pub/pdf/rsarp00.pdf (noting that between 1992 and 2000, sixty-four percent of rapes, sixty-six percent of attempted rapes, and seventy-four percent of sexual assaults were not reported to police); LAWRENCE A. GREENFELD ET AL., BUREAU OF JUSTICE STATISTICS, U.S. DEPT. OF JUSTICE, VIOLENCE BY INTIMATES (1998), available at http://www.ojp.usdoj.gov/ bjs/pub/pdf/vi.pdf (stating that between 1992 and 1996, 47.8% of victims of intimate partner violence did not report the incident to police). Of course, the decision to involve law enforcement is complex and often informed by law enforcement's historic biases based on race, nationality, and immigration status, as well as gender bias. See, e.g., Ammons, supra note 183, at 1023; Crenshaw, supra note 183, at 1257; Beth E. Richie, A Black Feminist Reflection on the Antiviolence Movement, 25 SIGNS 1133, 1136-37 (2000), available at http://www.jstor.org/stable/3175500; Holly Maguigan, Wading into Professor Schneider's "Murky Middle Ground" Between Acceptance and Rejection of Criminal Justice Responses to Domestic Violence, 11 AM. U. J. GENDER SOC. POL'Y & L. 427, 431-32 (2003); Jenny Rivera, Intimate Partner Violence Strategies: Models for Community Participation, 50 ME. L. REV. 283, 294-305 (1998). 217 See Swanberg et al., supra note 106, at 12, 21 (explaining that victims remain silent at work in part because of stigma associated with abuse, and arguing that employers should take steps to destigmatize abuse and encourage victims to disclose). Gender Violence and Work This self-censorship can be problematic. 218 Employers are increasingly becoming aware of the detrimental impact of abuse on the workplace and are taking steps to address the problem. 219 However, a survivor will only access programs and services if she is assured that it is safe to disclose the abuse. Absent affirmative assurances from her employer that she will not be penalized if she discloses, a victim is likely to stay silent. If she does, she may suffer unexplained performance problems, absences, or other workplace infractions that may subject her to disciplinary action or termination. So by maintaining the status quo and not adopting a domestic and sexual violence policy, an employer inadvertently creates an environment in which the effects of abuse remain unchecked. 3. Victim Status Cases in which an adverse action was based solely on an employee's status as a domestic or sexual violence survivor present more difficulty than those involving a similarly situated employee not in the protected class, or those involving obviously outdated stereotypes. Absent those types of evidence, a case may either not be commenced or may fail. On the one hand, this is not incorrect, since domestic and sexual violence are not per se, sex-based characteristics: both men and women can be victims or perpetrators. Consequently, courts could determine that an employee was terminated for the gender-neutral reason that she was a victim. Analogous arguments have doomed many, though not all, claims by domestic violence victims claiming that law enforcement policies treating domestic violence claims less seriously than other analogous claims discriminate on the basis of sex.22 ° 218 By exposing the link between victims' silence at work and sex-based biases I do not mean to suggest that gender bias drives all cases in which a victim chooses not to disclose her circumstances at work. Certainly, non-gendered reasons, such as a preference not to reveal personal circumstances, may define an individual's choice. However, the connection helps explain the structural dynamic in which, absent an explicit policy, victims are less likely to disclose and therefore more likely to be disadvantaged at work as a result of the abuse. 219 See supra Part I.C. 220 See, e.g., Ricketts v. City of Columbia, 36 F.3d 775, 781-82 (8th Cir. 1994) (finding a policy of fewer arrests in domestic abuse cases than nondomestic cases could be rationally explained by "inherent differences" between those cases and rejecting circumstantial evidence of officers' statements about not arresting batterers as "unreliable hearsay" and insufficient to establish discriminatory animus toward women); Eagleston v. Guido, 41 F.3d 865, 878 (2d Cir. 1994) (rejecting claim due to lack of evidence beyond statistical evidence of low arrest rates in domestic violence cases); McKee v. City of 2008] Columbia Journal of Gender and Law Although not framed as a question of whether "victim status" constitutes a form of protected status, several courts have raised a related question about whether abusive conduct stemming from a "personal relationship gone sour" falls within Title VII's sexual harassment prohibitions. Ironically, according to some judges, the fact that the violence and abuse was committed within an intimate relationship would render it less likely that a subsequent employment action would be found to be discriminatory than if the abuse had been committed between non- intimates. For example, in the Excel Corp. v. Bosley decision discussed previously,22 2 the Eighth Circuit upheld a sexual harassment claim based on an employer's failure to remedy ongoing harassment by an employee's former husband.223 In his concurring opinion, Judge Loken raised concerns, however, about whether a "marital spat" that spills over to the workplace is a proper basis for a sexual harassment claim. Quoting an unpublished Sixth Circuit decision, he stated that "personal animosity is not the equivalent of sexual harassment and is not proscribed by law. 224 Yet he recognized that the same "misconduct" between unrelated coworkers would "quite properly" be the basis for an inference of discrimination.2 25 Under that view, the fact that some form of abusive conduct occurred within the context of an intimate relationship, albeit one that ended, reduces rather than suggests the operation of sex-based bias. Rockwall, 877 F.2d 409, 416 (5th Cir. 1989) (finding insufficient evidence of a police department's policy of treating domestic violence victims differently than other similarly situated victims). But see, e.g., Navarro v. Block, 72 F.3d 712, 715-17 (9th Cir. 1995) (upholding an equal protection claim that domestic violence victims were treated less seriously than similar non-domestic violence victims under rational basis review); Thurman v. City of Torrington, 595 F. Supp. 1521, 1529 (D. Conn. 1981) (rejecting a motion to dismiss an equal protection claim challenging police policy of differential treatment of domestic violence victims). 221 See, e.g., Galloway v. Gen. Motors Serv. Parts Operations, 78 F.3d 1164, 1167 (7th Cir. 1996) (holding that repeated remarks in which former partner called employee "sick bitch" and subjected her to other verbal abuse and obscene gestures were not sex- or gender- related); Lowry v. Powerscreen USB, Inc., No. 4:98CV00377(MLM), 1999 WL 1127409 (E.D. Mo. July 15, 1999) (rejecting sexual harassment claim based on former partner and co- worker's abuse, concluding that harassment was based on prior relationship and not on gender). 222 See supra notes 168-82 and accompanying text. 223 Excel Corp. v. Bosley, 165 F.3d 635, 639 (8th Cir. 1999). 224 Bosley, 165 F.3d at 641 (Loken, J., concurring) (quoting Rothenbusch v. Ford Motor Co., No. 93-3945, 1995 WL 431012 (6th Cir. July 20, 1995)). 225 Bosley, 165 F.3d at 641 (Loken, J., concurring). [Vol. 18:1 Gender Violence and Work 4. Legitimate Employer Concerns An employer may cite legitimate concerns that arguably justify adverse actions against domestic or sexual violence victims. 226 It may state that it terminated an employee because of the risk posed by her abusive partner's threats.227 Or an employer may claim that an adverse action was 22justified due to an abuse victim's work performance issues.228 These are legitimate concerns that employers reasonably should, and in fact must, take into account. However, since these workplace performance and safety concerns may be generated by the sex-related dynamics of abuse, a better understanding of these problems can help antidiscrimination law separate cases that reflect impermissible bias from those that do not. a. Performance Issues It is not surprising that domestic or sexual violence victims may experience job performance issues. Studies show that abusers use the workplace to perpetuate abuse. 229 For example, one common form of harassment is for an abuser to make incessant daily phone calls to his partner's desk, keeping her on the phone for hours.23 ° It is not hard to imagine the difficulty of performing a job in the face of such distraction. If an employee chooses not to disclose the reasons for her performance issues, the ramifications of abuse will appear to be her performance problem, rather than an employment-related manifestation of abuse. To account for the role of abuse and maintain productive employees, employers should affirmatively adopt policies and conduct education programs so that employees who are victims of abuse feel assured that they can self-identify and take appropriate steps to reduce the impact of the abuse on the workplace without fear of retaliation.23 1 Regardless of whether an employer maintains such a policy, once an employer learns that a performance 226 These arguments could be raised either in service of proof of a legitimate business reason defense to a disparate treatment claim, or in support of a business necessity defense to a disparate impact claim. 227 See, e.g., Green v. Bryant, 887 F. Supp. 798 (E.D. Pa. 1995). 228 See, e.g., Rohde v. K.O. Steel Castings, Inc., 649 F.2d 317, 322 (5th Cir. 1981). 229 See supra Part I.B. 1. 230 See supra notes 67-69. 23 1 For a discussion of these policies, see supra Part I.B.3. 20081 Columbia Journal of Gender and Law problem may be the result of domestic or sexual violence, the employee should be given an opportunity to develop a safety plan before being subjected to adverse action. To do otherwise blames the victim and compounds her abuse. 232 The dynamics of abuse operate to produce performance problems for victims in additional ways. In some cases, an employee's frustration with her abuser's workplace harassment, coupled with an employer's failed responses, may lead the victim to take some action that violates a workplace conduct rule and subjects her to discipline or termination. For example, in the Excel Corporation v. Bosley case discussed above, an employee who was repeatedly harassed by her coworker ex-husband was terminated for "having struck a supervisor" when she pushed past a supervisor to enter the room where her ex-husband was located.233 Although she may have violated a workplace rule, she did so in the context of a longstanding pattern of abuse, her supervisor's refusal to accommodate her requests for relief from the harassment, and a meeting in which she was subjected to discipline for pushing her ex-husband in an attempt to get him out of her area and back to his appointed workplace. 34 Similarly, in Lowry v. Powerscreen USB, Inc., an employee was terminated for falsifying a "return to work" slip after she took a job leave to address the medical ramifications of her abuse.235 Although the petitioner did not contest the fact that she violated a workplace rule, the court failed to recognize the difficulty of complying with workplace rules while simultaneously negotiating abuse. b. Workplace Safety Employers also may justify adverse actions against domestic or sexual violence victims based on concerns about workplace safety and avoiding workplace disruption. For example, the employer cited safety concerns as the reason for Philloria Green's termination from her job in a doctor's office after she informed her employer that her estranged husband raped and severely beat her with a pipe at gunpoint.236 The court upheld her termination, which was based solely on her employer's apparent fears of the 232 For a fuller discussion of this dynamic, see supra Part I.B. 1. ... 165 F.3d 635, 638 (8th Cir. 1999). 234 Id. 235 No. 4:98CV00377(MLM), 1999 WL 1127409 (E.D. Mo. July 15, 1999). 236 Green v. Bryant, 887 F. Supp. 798, 798-800 (E.D. Pa. 1995). 106 [Vol. 18:1 Gender Violence and Work potential physical or emotional danger to other employees or patients if plaintiff's estranged husband came to the workplace and engaged in further violence.237 Similarly, in Hossack v. Floor Covering Associates, the Seventh Circuit upheld the termination of an employee based on her employer's concerns that her husband's threats would disrupt the workplace. 238 An employer's concern is understandable given its obligation to take reasonable steps to safeguard the workplace from known threats.239 However, the reflexive termination of the victim of abuse without inquiring whether removing her in fact would increase workplace safety is not the appropriate response. Abuse in this case distorts the analysis, and increases the likelihood that adverse actions will be taken against the victim, regardless of whether termination is warranted under the circumstances. Workplace safety concerns should not be underestimated or summarily dismissed.240 Nevertheless, a stated concern for safety may also constitute a pretext for discrimination. Ingrained attitudes and lack of understanding about the nature of abuse and the utility of safety plans, as well as legitimate fears about safety, may lead employers to reflexively penalize the victim without inquiring whether the adverse action is required, or whether some other, less extreme intervention could both protect workplace safety and the employee's job status. A stated concern for safety should not justify terminating an employee absent further inquiry and negotiation about whether termination in fact is the appropriate remedy. IlI. PROPOSAL: ACCOUNTING FOR THE REALITIES OF DISCRIMINATION The previous Parts demonstrate how employment decisions disadvantage victims of domestic and sexual violence without adequately accounting for the role of abuse and its connections with sex discrimination. These cases raise the question of how connections between domestic and sexual violence and sex discrimination should be taken into account in cases in which the victim suffers some adverse job action but traditional 237 Id. at 800-01. 238 492 F.3d 853, 863 (7th Cir. 2007). 239 See supra note 87 (discussing employers' obligations under tort law). 240 Id. Experts recommend workplace policies that encourage employee disclosure and safety planning to reduce the risk that an abusive partner will commit an act of violence at work. See supra Part I.B.3. 2008] Columbia Journal of Gender and Law indicia of discrimination, beyond the victim's abuse experience, are not present. All too often, the victim, most often a woman, is reflexively disciplined or terminated with no analysis of the connection between the abuse and prohibited discrimination. This renders invisible both the relationship between domestic and sexual violence and sex discrimination and the ways gender bias informs and distorts employers' reactions to victims. 24' On the other hand, an absolute equation of domestic or sexual violence with sex discrimination would be inappropriate. In addition to concerns that may be raised about whether an adverse action taken against a domestic or sexual violence victim necessarily reflects discrimination "because of sex," a flat equation would not adequately account for the cases in which domestic or sexual violence survivors nevertheless were justifiably disciplined or terminated. This Part offers a proposal for how antidiscrimination law can better account for the subtle bias associated with domestic and sexual violence while recognizing employers' legitimate interests in employee performance and workplace safety. It does this by making the implicit, hidden dynamics explicit, thus creating a check that impermissible bias is not at play. Consequently, the proposal requires employers and employees to discuss the actual, rather than presumed, role of abuse. This would both counter the operation of subtle stereotypes and ensure that adverse actions would only be taken when justified by the facts. Accordingly, unexplained adverse actions taken against survivors of domestic or sexual violence would raise an inference of discrimination that could be rebutted by proof that the employer engaged in a good faith negotiation with the employee and nevertheless determined that the adverse action was required. This Part will discuss each aspect of this proposal. A. Adverse Actions and Inferences of Discrimination The close connection between domestic and sexual violence and sex discrimination raises the difficult question of whether the fact that an employee was abused, without more, should lead to finding adverse treatment impermissible because of her sex. If the charge of antidiscrimination law is to eliminate discrimination "because of" protected characteristics such as sex, and if domestic and sexual violence victims are subjected to abuse because of their sex, and face responses at work that are, at least in the aggregate, overly determined by gender, the answer to this question must be "yes" in at least some subset of these cases. On the other 241 For further discussion of these dynamics, see supra Part I.B.1. [Vol. 18:1 2008] Gender Violence and Work 109 hand, not every adverse action taken against a domestic or sexual violence victim is taken "because of sex"; legitimate concerns about performance and safety may require disciplinary action in some cases. Evolving understandings about the operation of implicit bias suggest that pervasive, albeit unconscious, stereotypes about victims may permeate decisions regarding their employment status. The cases described above reveal at least two types of problems with employer responses to employees who are coping with abuse. In one set of cases, biases impact employment decisions through nuanced and unrecognized stereotypes that are not captured under current antidiscrimination frameworks.242 In the other, biases and stigma operate to silence victims and keep them from disclosing their circumstances and negotiating for modest workplace modifications that could facilitate job retention.243 As described in Part I.A, recent scholarship on "second generation" discrimination and "implicit bias" endorses using the law to more effectively address structural discrimination and subtle bias. For example, Linda Krieger and Susan Fiske propose prohibitions on implicitly biased behavior as one way to broaden the reach of antidiscrimination law. 2" Christine Jolls and Cass Sunstein argue that legal rules can address implicit bias either by insulating employment decisions from the impact of bias, or by directly targeting implicit bias.245 Susan Sturm proposes problem-solving solutions that are directed at the structural biases that maintain discriminatory practices.246 These approaches need not be mutually exclusive.247 These approaches suggest that antidiscrimination law can do a better job of making visible and policing the implicit sex bias that impacts domestic and sexual violence victims' employment.248 Statutory reform explicitly prohibiting discrimination against victims could accomplish this 242 See supra Part II.B. 1. 243 See supra Part II.B.2. 244 Krieger & Fiske, supra note 30, at 1060-61. 245 Jolls & Sunstein, supra note 31, at 977-87 (discussing "direct debiasing"). 246 See Sturm, supra note 3, at 468-71. 247 See Jolls & Sunstein, supra note 31, at 985 (discussing how affirmative action plans can both insulate decisions from the impact of bias, and can directly reduce bias). 248 See supra Part 11.13 for a discussion of workplace biases. Columbia Journal of Gender and Law directly.249 But employers' treatment of domestic and sexual violence victims implicates sex discrimination laws as well. As argued in Part I.A, there is a value in making explicit the links between the adverse action and sex discrimination. 250 As the cases and previous discussion demonstrate, the circumstances in which domestic and sexual violence are committed, the reactions of victims, and the responses victims receive from institutional officials, including employers, may be inextricably infused with sex-based bias. Just as with other allegations of bias, it may be difficult to assess whether a particular employment decision was consciously motivated by bias. Yet the charge of antidiscrimination law is to make those distinctions. In other contexts, antidiscrimination law overcomes the challenge by assigning liability where there is some evidence of bias (for example, either an unexplained differential treatment of similarly situated employees, circumstantial evidence of bias, or a facially neutral policy with a disparate impact on a particular group), and the elimination of valid business justifications. 25 1 Accordingly, disparate treatment claims fail if the employer establishes a legitimate business reason and disparate impact claims fail if the employer can prove business necessity. 252 In the context of cases involving gender violence, antidiscrimination law can directly target implicit bias by recognizing the ways sex-based biases infuse employer decisions concerning domestic and sexual violence. It can do this by recognizing a cause of action where there is an unexplained adverse action taken against a domestic or sexual violence victim because of her status as a victim. This would have the effect of requiring employers to engage with the employee and to consider the actual, rather than presumed, role of the abuse. A second part of this 249 See supra Part I.B.2 for a discussion of gaps found in existing statutes. 250 See supra text accompanying notes 32-43. 251 See, e.g., Furnco Constr. Corp. v. Waters, 438 U.S. 567, 580 (1978) (explaining that the McDonnell Douglas prima facie case raises an inference of discriminatory motives because the acts, if unexplained, are likely based on impermissible factors). 252 Sexual harassment claims do not fall easily into either category. Instead, the Supreme Court's treatment of sexual harassment claims more closely resembles the high explanatory burden courts impose on employers upon a showing of facial discrimination. Presumably, sexual harassment in the workplace will never have a legitimate business justification, thus vicarious liability is warranted where a supervisor takes an adverse action against a target of harassment, and imposing liability where there is no tangible employment action, but where the employer fails to satisfy the affirmative defense. See Burlington Indus. v. Ellerth, 524 U.S. 742, 764-65 (1998); see also Faraghar v. City of Boca Raton, 524 U.S. 775, 807-08 (1998). 110 [Vol. 18:1 2008] Gender Violence and Work 111 proposal would deem it to be a form of discrimination if an employer refused to make modest accommodations to a victim in order to facilitate her safe and productive employment in the face of the abuse. Employers that take an adverse action after ascertaining that no such modification was possible, or that a performance or safety issue persisted notwithstanding the modification, would not be subject to liability. This modification would constitute an additional way of establishing sex discrimination in disparate treatment cases involving adverse actions against victims of domestic or sexual violence.2 1 3 Courts would recognize an unexplained adverse action against a victim of domestic or sexual violence as raising an inference of discrimination that the employer then could rebut through proof that it engaged in an interactive process with the employee and nevertheless could not resolve the employment-related concem. Employers can draw from established and workable definitions used in other contexts to determine what circumstances, and which employees, would be covered. Standard definitions of domestic and sexual violence include physical and sexual violence, as well as threats of violence and associated emotional abuse. 4 Claims by male victims, both 253 Although there is variation in the terms courts and commentators use, disparate treatment can be established through the McDonnell Douglas burden shifting framework, mixed-motive frameworks, or through "direct" evidence of discrimination. See, e.g., LARSON, supra note 127, at § 8.07. This proposal raises the question whether the modified framework proposed in this Article should apply in cases other than those involving gender violence. Although a full analysis of that question is beyond the scope of this Article, the approach may be useful in other circumstances in which the conduct at issue is highly correlated with gender or another form of bias, such as cases involving dress codes and grooming standards, or adverse actions based on an employee's family or caregiver responsibilities. 254 For example, the Centers for Disease Control and Prevention recently issued guidelines that define intimate partner violence to include physical and sexual violence, threats of physical and sexual violence, and psychological and emotional abuse that occur in the context of prior physical or sexual violence. LINDA E. SALTZMAN, ET AL., CENTERS FOR DISEASE CONTROL AND PREVENTION, INTIMATE PARTNER VIOLENCE SURVEILLANCE: UNIFORM DEFINITIONS AND RECOMMENDED DATA ELEMENTS 11-13 (2002), available at http://www.cdc.gov/ncipc/pub-res/ipvsurveillance/Intimate%20Partner%20Violence.pdf The Violence Against Woman Reauthorization Act of 2005 similarly defines "domestic violence" to include felony or misdemeanor crimes of violence committed by a "current or former spouse of the victim, by a person with whom the victim shares a child in common, by a person who is cohabitating with or has cohabitated with the victim as a spouse, by a person similarly situated to a spouse of the victim under the domestic or family violence laws of the jurisdiction receiving grant monies, or by any other person against an adult or youth victim who is protected from that person's acts under the domestic or family violence laws of the jurisdiction." 109 Pub. L. 162, 119 Stat. 2965, codified at 42 U.S.C. § 13925 (2009). New Columbia Journal of Gender and Law heterosexual and gay, would be subject to the same analysis, as stereotypes about abuse apply to them as well.255 However, employers should not require a criminal indictment, prosecution, or conviction, recognizing that these crimes continue to be underreported and under-prosecuted. 5 Critics may raise concerns that this approach is overly broad, since not every act of domestic or sexual violence can be shown to be informed by gender bias. That critique misses the point. This proposal does not require courts to investigate whether there is sufficient evidence of gender bias in a particular act of domestic or sexual violence to warrant a finding of sex discrimination. 257 It focuses instead on institutional, i.e., employer, responses to circumstances in which employees are grappling with domestic and sexual violence. It rests on the premise that domestic and sexual violence is sufficiently over-determined by gender, both in the commission of the act or acts and in institutional responses, to warrant an inference that adverse actions may be grounded in bias.2 58 That inference invites inquiry and examination, rather than reflexive actions that may reflect and perpetuate bias. A showing that the adverse action was justified would negate the inference. By incorporating a threshold requirement that an employee in good faith asserts that she has suffered acts that constitute domestic or sexual violence, this proposal ensures that cases of de minimus "personal" disputes between intimate partners would not alone bring an employee within the reach of antidiscrimination laws. 59 York State's Model Domestic Violence and the Workplace Policy similarly defines domestic violence as "a pattern of coercive tactics, which can include physical, psychological, sexual, economic and emotional abuse, perpetrated by one person against an adult intimate partner, with the goal of establishing and maintaining power and control over the victim." State of N.Y. Model Domestic Violence and the Workplace Policy, http://www.opdv.state. ny.us/workplace/statepolicy.html (last visited Apr. 8, 2009). 255 Male victims may be even more vulnerable to biased stereotypes since their experience defies the traditional role of male as aggressor. See infra notes 260-262 and accompanying text. 256 See supra notes 213-215 and accompanying text. 257 In this way, this proposal is different from schemes, such as that incorporated in the federal Violence Against Women Act's civil rights remedy declared unconstitutional in United States v. Morrison, 529 U.S. 598 (2000), which predicated civil rights liability on a finding of gender motivation in acts such as domestic or sexual violence. 258 See supra notes 44-54 and accompanying text. 259 Accordingly, this proposal addresses the concerns of those such as Judge Loken in the Bosley case discussed supra, at notes 223-24 and accompanying text. [Vol. 18:l Gender Violence and Work Nor would this proposal render the employer responsible for employees' personal choices or relationships. As with other areas in which employees' personal circumstances affect the workplace, the employer would be responsible only for ensuring that it does not discriminate, either intentionally or through the operation of subtle biases, against the employees. This approach would apply to adverse employment actions taken against male as well as female victims of domestic and sexual violence.260 Gender bias skews traditional views about male as well as female victims of domestic violence, particularly in cases of gay male victims.261 Since male victims often defy the common conception of men as aggressors and women as victims, male victims may be subjected to biased responses by employers. Consequently, cases in which male victims claim adverse treatment as a result of their abuse also may be unduly skewed by gender bias.262 As a result, these cases should be subject to the same analysis, to ensure that the decision is grounded in legitimate concerns, rather than in impermissible bias. Under this framework, employers would be liable only for unexplained adverse actions taken against domestic or sexual violence victims because of their status as victims. Accordingly, it would encourage employers to engage in an interactive process with an employee about the effect of abuse on her employment. Through this process, which is described more fully in the next subpart, the employer and employee could ascertain whether any modest workplace modifications could address legitimate workplace concerns such as performance or safety. In that way, the law woulO mirror and incentivize recommended practices, which Allegations based on "personal animosity" or of a "relationship gone sour," absent other indications of abuse, would not bring an employee into the ambit of this proposal. 260 As discussed above, women constitute the vast majority of victims of domestic and sexual violence. See supra note 11 and accompanying text. 261 See, e.g., NATIONAL COALITION OF ANTI-VIOLENCE PROGRAMS, LESBIAN, GAY, BISEXUAL AND TRANSGENDER DOMESTIC VIOLENCE IN THE UNITED STATES IN 2006 6-7, 12-13 (2007), available at http://www.avp.org/publications/reports/documents/2007NCAVP DVREPORT.pdf (reporting data about domestic violence in same sex relationships and discussing the role of gender, as well as other forms of bias, in the treatment of lesbian, gay, bisexual, and transgender victims). 262 See, e.g., Schwenk v. Hartford, 204 F.3d 1187, 1202-03 (9th Cir. 2000) (recognizing that prison rape of male-to-female transsexual was motivated by gender, and reviewing literature detailing gendered stereotypes associated with prison rapes of male inmates). 2008] Columbia Journal of Gender and Law promote safety in the workplace and maintain valued employees' economic and physical security. B. Scope of the Interactive Process This proposal posits that treating domestic or sexual violence as a neutral circumstance unconnected to sex discrimination effectively colludes with and perpetuates gender bias. Requiring an employer to engage with an employee about how best to address the abuse would help ensure that a legitimate reason rather than subtle or implicit bias underlies an adverse 263employment action. This type of interactive negotiation, in combination with a requirement that employers implement modest workplace modifications to facilitate employees' safe and productive employment, is consistent with the model practices advocated by human resource and management professionals. 264 The approach draws from the interactive process and reasonable accommodations required under the Americans with Disabilities Act (ADA),26 5 as well as the related process associated with Title VII's prohibition of religious discrimination.2 66 Accordingly, when an employee discloses that she has experienced abuse, the employer would be required to engage with her in a specific and explicit discussion of the workplace responses that could assist her in 263 This proposal is similar to reforms put forth to address persistent discrimination against caregivers. See Rachel Arnow-Richman, Public Law and Private Process: Toward an Incentivized Organizational Justice Model of Equal Employment Quclity for Caregivers, 2007 UTAH L. REV. 25 (advocating requirement that employers engage h, interactive process with employees as an "organizational justice" approach to advancing workplace equality for caregivers). 264 See supra Part I.B.3 (discussing model policies); see also Sturm, supra note 3, at 559 (noting cases in which courts look to "best practices" to inform legal norms in particular cases). 261 See 42 U.S.C. § 12112(b)(5)(A) (2009) (defining "discrimination" to include failure to make reasonable accommodation to the known physical or mental limitations of an otherwise qualified individual with a disability, unless the accommodation would impose an undue hardship); 29 C.F.R. § 1630.2(o)(3) (defining "reasonable accommodation" and suggesting that in order to determine the appropriate accommodation, the covered entity may have to initiate an "informal, interactive process" with the qualified individual). 166 See 42 U.S.C. § 2000e(j) (defining "religion" to include all aspects of religious observance and practice, as well as belief, unless an employer demonstrates that it is unable to reasonably accommodate an employee's religious observance or practice without undue hardship). [Vol. 18:1 2008] Gender Violence and Work negotiating both the abuse and her job.267 Thus, this approach would remove the reflexive judgments and misconceptions from employment decisions concerning abuse, and instead would ground those decisions in a realistic and practical balancing of employee and employer needs and concerns. This approach is consistent with feminist theory, in its requirement of interaction and dialogue.268 It is also in keeping with the dynamic approach advocated by Sturm, in which "regulatory" participants within the workplace engage with the specific dynamics at play in each case to solve workplace problems.269 It exemplifies an intervention that would directly prohibit implicitly biased behavior, as advocated by Jolls and Sunstein.27 ° The proposed interactive process and the possibility of modest modification would help defeat the operation of bias in at least two ways. First, it would help to ensure that the employer's actions were based on legitimate business concerns, rather than on stereotypes about abuse by requiring employers to make their rationale for adverse actions against 267 See 29 C.F.R. § 1630.2(o)(3); 29 C.F.R. § 1630.9, at 414 (Appendix: Interpretive Guidance). The EEOC Reasonable Accommodation Guidance recommends the following, flexible, four-step process: (1) Analyze the particular job involved and determine its purpose and essential functions; (2) Consult with the individual with a disability to ascertain the precise job-related limitations imposed by the individual's disability and how those limitations could be overcome with a reasonable accommodation; (3) In consultation with the individual to be accommodated, identify potential accommodations and assess the effectiveness each would have in enabling the individual to perform the essential functions of the position; and (4) Consider the preference of the individual to be accommodated and select and implement the accommodation that is most appropriate for both the employee and the employer. Id. The scope of required modifications is discussed under the approach proposed infra at Part lll.C. 268 See Carlos A. Ball, Looking for Theory in All the Right Places: Feminist and Communitarian Elements of Disability Discrimination Law, 66 OHIO ST. L.J. 105, 147-49 (2005). 269 Sturm, supra note 3, at 462-63. 270 See Jolls & Sunstein, supra note 31, at 986. Columbia Journal of Gender and Law domestic or sexual violence victims explicit. 271 It would help counter the operation of bias in a second way as well. As described in Part II, supra, part of the legacy of gender bias and abuse is its silencing effect. This silence is often counterproductive, because it prevents an employee from enlisting the assistance of her employer, which can be critical.272 By requiring a legitimate justification and modest accommodations for abuse- related employment decisions, antidiscrimination law would signal to employees that they would not be retaliated against merely for disclosing the abuse. This interpretation of the sex discrimination prohibition in the case of domestic and sexual violence would counter the silence historically associated with abuse and would encourage the self-disclosure that facilitates effective safety planning. Accordingly, as with the interactive process commended under the ADA, an employer would only be required to engage in an interactive process with an employee about her experience with abuse if the employee voluntarily disclosed the abuse and informed her employer of the need for an accommodation.273 This would ensure an employee's autonomy and privacy, and her ability to control her employer's intervention in her personal life.274 Of course, as Tarr argues, the question whether to disclose one's experience with abuse to an employer is not simple, and an employee's legal protection should not simply be cast in terms of whether she was "willing to leave her abuser or accept any help from her employer." 275 Nevertheless, an employee will be more likely to disclose in 271 See Sturm, supra note 3, at 496 (discussing Deloitte & Touche's strategy of making the performance review and promotion process visible and accountable). 272 Although disclosing abuse and enlisting an employer's assistance may be helpful, it is not appropriate in all instances. By endorsing policies that encourage disclosure, this proposal would in no way require employees to disclose their circumstances or to engage with their employer as a condition of continued employment or assistance. The decision of whether and with whom to share information about her abuse must at all times remain with the victim. See Tarr, supra note 23, at 409 (discussing importance of victim autonomy and agency in all policy approaches). 273 See 29 C.F.R. § 1630.2(o) (app.) (noting that an employer would not be expected to accommodate disabilities of which it is unaware); see also EEOC, Reasonable Accommodation and Undue Hardship Under the ADA, http://www.eeoc.gov/policy/docs/ accommodation (last visited Apr. 8, 2009) (discussing process by which employee must request accommodation, and noting that the employee need not mention the ADA or use the words "reasonable accommodation"). 274 See Tarr, supra note 23, at 398. 275 Id. at 373, 392. Tarr critiques Porter's framing of the problem as "pathologiz[ing]" the victim, assuming that separation is the only viable option, ignoring the 116 [Vol. 18:1 Gender Violence and Work an environment with a clearly articulated policy that explains the employer's recognition of the role of abuse, its support for victims, and that the employer will not retaliate in any way against an employee who discloses. This approach would not guarantee employment for all domestic and sexual violence victims. If an employer could establish that it engaged in a good faith negotiation with an employee and determined that no reasonable accommodation could address the concerns, any adverse action then would not be considered "because of sex." Thus, for example, if an employee's performance suffered and she claimed that the problems stemmed from her abusive partner's incessant phone calls, a workplace modification might entail changing her phone number so that he could not reach her.276 If her performance problems nevertheless persisted, and her employer established that she could not adequately or safely perform the job despite good faith negotiations and any reasonable attempt at workplace modifications, the employer would overcome the adverse inference and the claim would not succeed.277 The key to this modified approach is the requirement that the employer's response is explicit and considered. By requiring an employer to engage with an employee about how it could accommodate the abuse, antidiscrimination law would go a long way toward reducing the shame that still accompanies domestic and sexual violence while enhancing the possibility of employee retention and complexity of leaving an abuser and as giving the employer too much power over an employee's personal life and choices. Id. I agree that conditioning legal protection on a victim's disclosure to her employer, or on a blanket requirement that she take some specific step, such as obtaining an order of protection or leaving her abuser, dramatically and inaccurately oversimplifies the issue. Employers should encourage, not mandate, disclosure and can best do so through widely publicized and well-enforced policies assuring victims that they will not summarily be subjected to adverse actions without an assessment of whether modest accommodations were possible. Before an employer requires an employee to take a particular step the employer thinks would promote safety, the employer should engage with the employee about what step or steps she believes would best prevent violence at work, and best help her maintain required performance levels. 276 Examples of modifications that have been found to be effective are discussed in Part I.B.3, supra. 277 This requirement should be interpreted flexibly and reasonably. For example, in a case of domestic violence where the violence escalates over time, the employee and the employer would engage in a dialogue about how best to address the problem. The nature of the interactive process and the scope of modifications might be modified over time to account for the changing nature of the abuse. However, an employer could take into account previous accommodations when evaluating the appropriate response to a circumstance involving escalating abuse. 20081 Columbia Journal of Gender and Law workplace safety. In this sense, the approach is consistent with other process-oriented proposals to address discrimination.278 C. Scope of Required Accommodations This proposal advocates importing an ADA-like accommodation requirement into Title VII's sex discrimination prohibition to eliminate artificial barriers to women's equality created by domestic and sexual violence. 279 As with the disabled, or with those who require workplace accommodations of their religious beliefs, modest workplace modifications may be needed to ensure that victims of domestic and sexual violence can fully participate at work, notwithstanding their experience with abuse.2 s° The notion of importing an accommodation requirement to Title VII's sex discrimination prohibition is not novel. The accommodation could be required either through judicial interpretation or through statutory amendment. Although statutory amendment no doubt would make the obligation more explicit, judicial interpretation could reach the same result without the political difficulties of legislative reform.28' Nor is the notion of 278 See, e.g., Arnow-Richman, supra note 263. 279 For discussion of antidiscrimination law's accommodation requirements, see Christine Jolls, Antidiscrimination and Accommodation, 115 HARV. L. REV. 642 (2001); Peggy Smith, Accommodating Routine Parental Obligations in an Era of Work-Family Conflict: Lessons from Religious Accommodations, 2001 Wis. L. REV. 1443. Like childrearing accommodations, the accommodations suited to domestic and sexual violence victims may be more similar to those sought under religious accommodation case law than the structural and often physical modifications often sought under the ADA. See Smith, supra, at 1474-78 (discussing modifications such as lateral transfers, voluntary substitutes and swaps, and flexible scheduling). 280 The question of whether the accommodation model used in the religious discrimination or disability discrimination context should apply is difficult given the Supreme Court's cramped interpretations of both. See Joan Williams & Nancy Segal, Beyond the Maternal Wall: Relieffor Family Caregivers who are Discriminated Against on the Job, 26 HARV. WOMEN'S L.J. 77, 83-85 (2003) (reviewing literature and discussing problems with both statutory frameworks). 281 The history of Title VII's religious discrimination prohibition provides a useful example. The accommodation requirement originated in an EEOC guideline that interpreted Title VII's antidiscrimination prohibition to require employers to accommodate employees' religious needs. See Alan D. Schuchman, The Holy and the Handicapped: An Examination of the Different Applications of the Reasonable-Accommodation Clauses in Title VII and the ADA, 73 IND. L.J. 745, 748-51 (1998) (tracing the history of the religious discrimination accommodation requirement). The judicially imposed mandate was rejected due to concerns that requiring businesses to accommodate employees' religious preferences and practices would impose employee's religious beliefs on their employer, and could violate the 118 [Vol. 18:1 Gender Violence and Work a judicially created defense to liability novel; the Supreme Court has devised just that type of defense in other types of sex discrimination claims.282 In related contexts, commentators have proposed a general duty to reasonably accommodate caregiving needs, comparable to the accommodation requirement for disabilities and religious beliefs.283 The suggestion embodies a substantive equality interpretation of Title VII's antidiscrimination prohibition. It takes into account the realities of the workplace responses needed to place all workers on equal footing. It would normalize employees' requests for modest modifications to cope with abuse, and would incentivize employer policies granting accommodations. This proposal would not open the door to requiring an interactive process and accompanying accommodation in all Title VII cases. It is a limited proposal in that it would apply to circumstances so closely related to the experience of impermissible biases as to require a closer look before an adverse action is taken. The ADA provides the best guidance for the scope of the required modification, since it contemplates a multi-faceted, fact-intensive inquiry that balances financial cost, administrative burden, complexity of implementation, and other negative impacts on the employer's business.2 84 Of course, the ADA's accommodation requirement has been roundly criticized as, among other things, imposing uncertain and burdensome administrative and financial responsibilities on employers, 285 and as Establishment Clause. See, e.g., Dewey v. Reynolds Metals Co., 429 F.2d 324, 334-36 (6th Cir. 1970), aff'd 402 U.S. 689 (1971) (4-4 decision with Harlan, J., not participating). That concern about the appropriate limits of employer decisions about religion is not raised in the domestic violence context, and nothing in the decisions surrounding interpretation of Title VII's religious discrimination prohibition before Title VII's statutory amendment requiring accommodation would bar judicial requirement of an accommodation here. 282 For example, the Supreme Court has recognized an affirmative defense to sexual harassment claims. See Faraghar v. City of Boca Raton 524 U.S. 775, 804 (1998); Burlington Indus. v. Ellerth, 524 U.S. 742, 765 (1998). 283 See, e.g., Laura T. Kessler, The Attachment Gap: Employment Discrimination Law, Women's Cultural Caregiving and the Limits of Economic and Liberal Legal Theory, 34 U. MICH. J.L. REFORM 371, 457-59 (2001); Smith, supra note 278; cf Arnow-Richman, supra note 262, at 43-44 (critiquing proposals as creating significant uncertainty and administrative difficulties for employers, and as producing backlash reflecting judicial lack of commitment to the ADA's transformative mission). 284 Americans with Disabilities Act of 1990, 42 U.S.C. § 12112(b)(5)(A) (2009); 29 C.F.R. § 1630.2(p). 285 See, e.g., Samuel Issacharoff & Justin Nelson, Discrimination with a Difference: Can Employment Discrimination Law Accommodate the Americans with 2008] Columbia Journal of Gender and Law producing general backlash from the business community against 286antidiscrimination norms. As a substantive, as opposed to political, matter, these concerns should not prevent courts from requiring modest accommodations to advance antidiscrimination law's broad mandate to eliminate artificial barriers to equality. No doubt, the proposed interactive process and accommodations will impose costs on employers. However, those costs generally will be modest. Most employers have found that low- to no-cost modifications such as changing phone numbers, worksite locations, shift adjustments, or application of leave policies to domestic or sexual violence related appointments go a long way toward facilitating a victim's safe and productive continued employment.287 The difficulty lies in assessing the appropriateness of a particular accommodation, which would be evaluated based on the particular needs and resources of an employee and her employer. That assessment, including the assessment of when safety concerns outweigh an employee's interest in continued employment, no doubt may be difficult in any given case.288 But it is the type of assessment in which employers of all sizes routinely engage. As a matter of equality, an employer's response to a disclosure of abuse should be deliberate and rational, rather than reflexive and automatic, to guard against the potential to reinforce and reinscribe subtle gender-based bias. Disabilities Act?, 79 N.C. L. REv. 307, 347-55 (2001); Jerry L. Mashaw, Against First Principles, 31 SAN DIEGO L. REv. 211, 219 (2004). 286 See, e.g., Arnow-Richman, supra note 263, at 43-44. 287 See supra Part I.B.3. For example, common workplace adjustments, such as changing a phone number or worksite location, impose virtually no costs on an employer. Other adjustments, such as scheduling changes, may have some financial impact; their reasonableness would be evaluated on a case-by-case basis, depending on the size and type of employer, in the same way that accommodations are evaluated under Title VII's religious discrimination prohibition, and under the ADA. 288 In assessing the severity of safety threats stemming from gender violence, employers can rely on the same type of guideposts they employ in other cases involving risks of workplace safety. In this context, experts conclude that an explicit and comprehensive policy that takes seriously the victim's assessment of the severity of the risk is the approach best suited to prevent violent incidents from occurring at work and to reduce the risk of liability. See supra Part IIl.C. Ignoring the risk of violence, or automatically removing employees who are victims of abuse, is not rationally tailored to the risk. 120 [Vol. 18:1 Gender Violence and Work D. Why Statutory Interpretation and Not Legislative Reform? This proposal calls for a modified interpretation of antidiscrimination law's current concept of sex discrimination, not new statutory reform that would explicitly prohibit adverse actions against survivors of gender violence. Due to the uncertainty surrounding whether antidiscrimination laws will cover the circumstances faced by domestic and sexual violence victims, some jurisdictions have enacted abuse-specific laws to address the challenges victims face in the workplace. Some of these laws cover leave, unemployment, and other provisions that will assist domestic and sexual assault survivors. 89 Others specifically prohibit discrimination against domestic and sexual violence survivors.29 ° Strong arguments based on the traditional justifications for antidiscrimination protections support the inclusion of domestic and sexual violence in antidiscrimination provisions. For example, antidiscrimination protection has been justified based on factors such as a history of discrimination, economic disadvantage, and employer discrimination that causes or contributes to economic disadvantage. 29 1 The statistics described in Part I, detailing the biases against victims, the economic disadvantages domestic and sexual violence produces, and the nexus between treatment at work and economic disadvantage, support inclusion. On the other hand, a number of objections have been raised to these proposals. Some have argued that the coverage would be duplicative of the protection afforded under other antidiscrimination provisions, such as those prohibiting discrimination based on disability.2 92 It is true that some domestic and sexual violence victims may suffer physical or psychological disabilities as a result of abuse that would qualify under disability discrimination prohibitions.293 But not all will. Limiting coverage to those victims who experience a cognizable disability as a result fails to recognize the varied realities of abuse. 289 See supra notes 92-96 and accompanying text. 290 See supra note 96 and accompanying text. 291 MICHAEL J. ZIMMER, ET AL., CASES AND MATERIALS ON EMPLOYMENT DISCRIMINATION 32 (6th ed. 2003). 292 See Marta B. Varela, Protection of Domestic Violence Victims Under the New York City Human Rights Law's Provisions Prohibiting Discrimination on the Basis of Disability, 27 FoRDHAM URB. L.J. 1231 (2000). 293 See Porter, supra note 59, at 305-08. 2008] Columbia Journal of Gender and Law Others argue that antidiscrimination laws are not necessary because they inadequately safeguard workplace safety and because voluntary policies better allow employers to adopt flexible responses on a case-by- case basis. 9 While voluntary efforts should be applauded, history shows that not all employers will adopt recommended policies absent legal requirements. 95 With respect to safety concerns, antidiscrimination provisions should be interpreted to require an employer to respectfully acknowledge the impact of abuse on the battered employee, and to work with her to strategize how she can keep her job. This would not, however, result in blanket job protection for all employed victims of abuse, since employers could avoid liability if they terminated an abused employee after good faith efforts to secure her job and workplace safety proved unsuccessful. From a different perspective, some have raised concerns that domestic and sexual violence victims are unlike other groups covered by antidiscrimination laws. Widiss argues that "domestic violence victim status" is not a personal characteristic; it is a descriptive statement of a circumstance to which an individual has been subjected.296 However, as Widiss points out, immutability is not a requirement of protected class status.297 Indeed, as our understanding of the nature of impermissible bias evolves, it makes sense that the categories of protected groups also change to more effectively rout impermissible discrimination. Widiss also invokes the practical realities of the political opposition to expanding protected status classifications. 298 This is no small consideration when weighing strategic decisions about where to place scarce resources and lobbying efforts. However, this should not preclude legislative reform as a substantive or normative matter. 294 See, e.g., Sue K. Willman, Too Much, Too Long? Domestic Violence in the Workplace, Testimony Before the U.S. Senate Subcommittee on Employment and Workplace Safety 8-9, 12-13 (Apr. 17, 2007), http://help.senate.gov/Hearings/2007 04 17/ Willman.pdf. 295 For example, although some employers voluntarily adopted anti-sexual harassment policies, it was not until courts held that sexual harassment was a form of prohibited sex discrimination and, in particular, that policies could play a role in insulating employers from liability, that policies became widespread. See, e.g., Goldscheid, supra note 44 (tracing policy development); see also supra notes 82-83 and accompanying text. 296 Widiss, supra note 56, at 707-08. 297 id. 298 Id. at 708-09. [Vol. 18:1 Gender Violence and Work Legislative reform that explicitly prohibits discrimination on the basis of domestic or sexual violence victim status can help victims in a number of ways. Such laws will squarely define prohibited conduct. They may galvanize employer policies that assist, rather than penalize victims. They also can draw needed attention to the issue and can expand education and awareness campaigns. These are critical reforms, particularly if courts do not recognize the full scope of sex discrimination that domestic and sexual violence produces. However, amendments to antidiscrimination laws adding survivors of gender violence as an additional protected category could fuel arguments that the meaning of those categories has become hopelessly diluted and that the justification for additional new categories is unclear and amorphous. For those who favor coverage, adding a new protected category may also serve as a shortcut that allows employers, policymakers, and courts to sidestep grappling with the pernicious and persistent vestiges of sex discrimination as it manifests through domestic and sexual abuse and the treatment of survivors. From that perspective, statutory reforms should not substitute for a more robust and accurate interpretation of sex discrimination prohibitions. IV. CONCLUSION Workplace discrimination persists despite decades of antidiscrimination statutory protection. The workplace ramifications of domestic and sexual violence exemplify one way in which current antidiscrimination frameworks fail to address the realities of discrimination. Notwithstanding that domestic and sexual violence reflect historic sex discrimination, litigation brought by victims who have suffered adverse employment actions may not be successful under current legal approaches. This Article has demonstrated how subtle stereotypes and silence related to those stereotypes inform workplace decisions involving survivors of domestic and sexual violence. It proposes a framework in which the role of abuse and its connection with gender bias are made transparent, while employers retain the flexibility to terminate survivors when their safe and productive employment cannot reasonably be accommodated. By making explicit the ways in which domestic and sexual violence continue to be a site of sex discrimination, employment law can more effectively address impermissible discrimination, promote retention of valued employees, and help domestic and sexual violence survivors attain safety and economic security. 2008]