Operating in an Empirical Vacuum: The Ellerth and Faragher Affirmative Defense OPERATING IN AN EMPIRICAL VACUUM: THE ELLERTH AND FARAGHER AFFIRMATIVE DEFENSE ANNE LA WTON* "Oh dear, what nonsense I'm talking. , There is a classic management article by Professor Steven Kerr entitled On the Folly of Rewarding A, While Hoping for B. 2 Its thesis is simple: reward systems often are structured so that they discourage desired behavior and encourage unwanted behavior. The federal courts' decisions on the affirmative defense in cases of sexual harassment by supervisors illustrate Kerr's thesis. In Burlington Industries v. Ellerth,3 and Faragher v. City of Boca Raton,4 the Supreme Court created a two-pronged affirmative defense to employer liability in certain cases of workplace harassment by supervisors. The Court did so in order to effectuate what it identified as Title VII's "primary objective"-the prevention of illegal workplace discrimination, including sexual harassment.5 The affirmative defense requires an employer * Associate Professor, Roger Williams University School of Law. A.B., M.B.A., J.D., University of Michigan. This Article is dedicated to Manuel Carrstell Little. I want to thank Michael Yelnosky, a colleague at Roger Williams, for helping me to work through the ideas in this Article; Kiran Advani, Nicholas Kapsa, and Liannie Parahoo for their diligent research assistance; and the library staff at RWU's law school, in particular Emilie Benoit, for their determined efforts to locate non-legal resource material for me. 1 Lewis Carroll, Alice's Adventures in Wonderland 10 (1865). As the findings in this Article suggest, Alice is in good company. 2 18 Acad. Mgmt. J. 769 (1975). In 1995, the Academy of Management Executive published an update of Kerr's article as an "Academy Classic." See Steven Kerr, An Academy Classic: On the Folly of Rewarding A, While Hoping for B, 9 Acad. Mgmt. Exec. 7 (1995) [hereinafter An Academy Classic . ' 524 U.S. 742 (1998). 4 524 U.S. 775 (1998). 5 Farajher, 524 U.S. at 806 (quoting Albermarle Paper Co. v. Moody, 422 U.S. 405, 417 (1975)). It is questionable whether deterring discriminatory conduct, such as sexual harassment, is Title VII's primary goal. Professor Grossman argues that prior to Ellerth and Faraher, deterrence and compensation were equally important goals of Title VII. See Joanna L. Grossman, The First Bite is Free: Employer Liability for Sexual Harassment, 61 Columbia Journal of Gender and Law to prove both that (1) it exercised reasonable care to prevent and correct workplace harassment ("prong one"), and (2) the victim employee unreasonably failed to take advantage of the preventive and corrective mechanisms established by the employer ("prong two"). 6 If an employer can satisfy the affirmative defense, it can limit or altogether avoid liability for the harassing conduct of its supervisors. By offering employers the possibility of limited or no liability for workplace harassment by supervisory personnel, the Court hoped to encourage employers to take steps to prevent and correct harassing behavior, for example, by adopting effective grievance procedures. 8 Thus, in Ellerth and Faragher, the Court created a reward system which ties an employer's liability to its efforts to reduce the incidence of workplace harassment. The problem is that the lower federal courts have interpreted the elements of the affirmative defense so as to reward employers for engaging in behaviors that have little effect on the incidence of workplace harassment. The courts reward employers for developing and distributing nicely worded harassment policies and procedures and, in some cases, providing sexual harassment training to their employees. The empirical literature does not support the federal courts' assumption that paper policies and procedures, even when coupled with training, deter sexual harassment in the workplace. Rather than providing employers with incentives to address the predictors of workplace harassment, such as the organization's culture and the job gender context, 9 the courts reward employers for "file cabinet compliance." The fault lies not only with the lower federal courts' implementation of the affirmative defense, but also with the Supreme Court's articulation of the defense itself. The success of the affirmative defense as a means to increase deterrence of workplace harassment rests in part on behavior that occurs infrequently-formal reporting of sexual harassment. By hinging liability on a response to harassment that is uncommon, especially in cases involving supervisors, the Court created a legal rule that from its inception was unlikely to promote the stated goal of prevention. In addition, the limited guidance offered by the Court on what U. Pitt. L. Rev. 671,720-21 (2000). Professor Grossman contends that the Supreme Court in Ellerth and Faraeher "elevated deterrence to the primary goal and left compensation by the wayside." Id. at 721. 6 Ellerth, 524 U.S. at 765; Farahe, 524 U.S. at 807. 7 Professor Grossman contends that while the Supreme Court intended for the affirmative defense to "sometimes affect damages and sometimes affect liability," the lower federal courts have interpreted the defense so "that it always operates to eliminate liability." See Grossman, supra note 5, at 676-77 (footnote omitted). 8 Ellerth, 524 U.S. at 764. 9 These terms are defined and discussed infra in Parts III.A.2.b (Organizational Culture) and III.A.2.c (Job Gender Context). [Vol. 13:2 Operating in an Empirical Vacuum constitutes prevention, coupled with the Court's emphasis in Ellerth and Faragher on policies and procedures, paved the way for the lower federal courts to interpret the affirmative defense in ways that further undermine, rather than facilitate, the goal of deterring sexual harassment in the workplace. The end result is a system in which an employer is able to limit its liability for sexual harassment by doing little more than creating and distributing an anti-harassment policy and grievance procedure that allows employees to bypass a harassing supervisor. There is little incentive for an employer to keep records of harassment complaints, to implement post- complaint follow-up procedures, to periodically assess and revise the firm's anti-harassment policies and procedures, or to evaluate supervisory personnel on their compliance with and implementation of the employer's policies and procedures. If the employer can escape liability for workplace harassment by doing less rather than more, why should it expend the time and energy in developing evaluative mechanisms that actually may expose it to greater liability? This Article, through an analysis of the federal court decisions issued since Ellerth and Faragher, explores the incentives that the federal courts are creating for employers to prevent and correct workplace harassment. In Part I, I begin with a brief overview of three Supreme Court decisions on employer liability for supervisory harassment-Meritor Savings Bank v. Vinson,' ° Ellerth, and Faragher. I conclude that the affirmative defense rests on two flawed assumptions, which made it unlikely that the defense would serve the Court's intended goal of deterring workplace harassment. Part I explains how I arrived at the body of lower federal court cases on which the Article is based. In Part ILI.A, I examine the lower federal courts' analysis of employers' prevention efforts and conclude that the courts are creating incentives for employers to engage in conduct that has no empirically documented impact on the prevention of sexual harassment in the workplace. In addition, the courts' focus on paper policies and procedures impermissibly shifts the burden of proof on prevention from the employer to the employee, contrary to the Supreme Court's decisions in Ellerth and Faragher. Part II.B of the Article critiques the federal courts' analysis of the reasonableness of employee decisions not to report sexual harassment using their employers' internal grievance machinery. The courts "punish" harassment victims by relieving employers of liability without evaluating the reasonableness of victim responses, either in the context of the empirical literature or the facts of the individual case. Moreover, by doing so, the courts once again improperly shift the burden of proof on reasonableness from the employer to the employee. In Part IV of the Article, I conclude with some recommendations for change, keeping in mind that the liability paradigm itself is fundamentally flawed. 10477 U.S. 57 (1986). 2004] 200 Columbia Journal of Gender and Law [Vol. 13:2 I. THE LIABILITY TRILOGY In Ellerth and Faragher, the Supreme Court created an affirmative defense to employer liability for certain cases of supervisory sexual harassment." The Court concluded that an affirmative defense would further Title VIl's primary goal of prevention. 2 By allowing employers to limit their liability through invocation of the affirmative defense, the Court hoped to encourage employees to report harassment before it became severe or pervasive, thereby "serv[ing] Title VIl's deterrent purpose."' 13 In order to understand why the affirmative defense exists, however, it is necessary to go back to the Supreme Court's earlier decision in Meritor. It is in Meritor that the Court lays the groundwork on which the affirmative defense rests.' 4 A. Meritor Meritor is an interesting case. By recognizing that a claim based on a hostile work environment is actionable under Title VII, the Court expanded the concept of sexual harassment beyond the more limited category of quid pro quo harassment,' 5 which occurs much less frequently " 524 U.S. at 765; 524 U.S. at 807. While this Article was in the final editing process, the Supreme Court handed down its decision in Pennsylvania State Police v. Suders, 124 S.Ct. 2342 (2004). In Suders, the Court resolved a circuit split "on the question whether a constructive discharge brought about by supervisor harassment ranks as a tangible employment action and therefore precludes assertion of the affirmative defense articulated in Ellerth and Faragher." Id. at 2350 (citations omitted). (For a brief discussion of what constitutes a tangible employment action and how proof of a tangible employment action affects employer liability, see infra notes 47-50 and accompanying text.) In Suders, the Supreme Court held that "when a supervisor's official act precipitates the constructive discharge," the employer may not invoke the affirmative defense because the constructive discharge amounts to a tangible employment action. Id. at 2351. But, the Court concluded that not all cases of constructive discharge involve tangible employment actions. See id at 2355. For example, a supervisor's harassment may result in a constructive discharge even without a precipitating "employer-sanctioned adverse action," such as "a humiliating demotion." Id. at 2347; see id. at 2355. Therefore, "when an official act does not underlie the constructive discharge," the employer may assert the affirmative defense because the plaintiff has not shown the existence of a tangible employment action. Id at 2355. 12 See Faraher, 524 U.S. at 806. 3 Ellerth, 524 U.S. at 764. 14 See Faraeher, 524 U.S. at 792 (explaining that "Meritor's statement of the law is the foundation on which we build today"). 15 Quid pro quo harassment involves an "employer demand[ing] sexual favors from an employee in return for a job benefit." Ellerth, 524 U.S. at 752. Operating in an Empirical Vacuum in the workplace than hostile environment harassment. 6 At the same time, however, the Court rejected a rule of automatic liability for hostile environment cases involving supervisors and, while not a model of clarity, the opinion suggested that liability hinged on notice to the employer of the harassing conduct, a requirement at odds not only with the EEOC Guidelines at the time,17 but also with the way in which most victims respond to sexual harassment. 8 Meritor involved a claim by Mechelle Vinson against her employer Meritor Savings Bank ("Bank") based on what Vinson claimed was constant sexual harassment by Sidney Taylor, a Bank vice president and Vinson's supervisor. 19 After an eleven-day bench trial, the district court concluded that Vinson was neither "the victim of sexual harassment [nor] of sexual discrimination." 20 Though unnecessary to its decision, the district court also found that the Bank could not be held liable for any harassment because it was without notice since no employee, including Vinson, had ever filed a sexual harassment complaint against Taylor.2' The Court of Appeals for the D.C. Circuit reversed and remanded the case, concluding that the district court had failed to consider whether Vinson had made out a claim for sexual harassment based on a hostile work environment.22 In addition, the Court of Appeals held that an employer's 16 See Anne Lawton, The Emperor's New Clothes: How the Academy Deals with Sexual Harassment, 11 Yale J.L. & Feminism 75, 100-01 (1999) (discussing emergence of hostile work environment cause of action and noting that "far more women experience hostile environment harassment in the form of sexist remarks and behaviors, than quid pro quo harassment"). 17 The Guidelines differentiated between the liability standards applicable to harassment by supervisors as opposed to co-workers, applying strict liability to the former and a rule of actual or constructive notice to the latter. See 29 C.F.R. § 1602.1 1(c)-(d) (1985). What is interesting is that the Court in Meritor chose to disregard the EEOC Guidelines on the issue of employer liability, but to adopt the position of the Guidelines on whether a hostile work environment constituted an actionable claim for sexual harassment pursuant to Title VII. See Meritor, 477 U.S. at 65, 71. In National R.R. Passenger Corp. v. Morgan, 536 U.S. 101 (2002), a recent decision on the continuing violations doctrine, the Court explained, in a footnote, that the EEOC's interpretive guidelines do not "warrant deference under Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 827, 81 LEd.2d 694, 104 S.Ct. 2778 (1984)." Instead, the EEOC's interpretive guidelines are "'entitled to respect' under [the Court's] decision in Skidmore v. Swift & Co., 323 U.S. 134, 140, 89 LEd. 124, 65 S.Ct. 161 (1944), but only to the extent that those interpretations have the 'power to persuade.' Christensen v. Harris County, 529 U.S. 576, 587, 146 L.Ed. 2d 621, 120 S.Ct. 1655 (2000)." Id. at 110 n.6. 18 See infra notes 73-80 and accompanying text. 19 477 U.S. at 60. 20 Id. at 61. 21 1d. at 62. 22 Id. at 62-63. 2004] Columbia Journal of Gender and Law notice of a supervisor's harassment is irrelevant to liability 23; employers were strictly liable for such harassment.24 The Supreme Court rejected both the district court's and the Court of Appeals's liability formulations.2 ' First, the Court concluded that by defining the term "employer" to include "agent" in Title VII, Congress "evince[d] an intent to place some limits on the acts of employees for which employers under Title VII [could be] held responsible. 26 The Court explained that common law agency principles, while not "transferable in all their particulars to Title VII, ' 27 should guide courts when evaluating claims of employer liability for sexual harassment by supervisors. With little explanation other than a general citation to §§ 219-237 of the Restatement (Second) of Agency, which govern a principal's liability for its agents' torts against third parties, the Court held that the Court of Appeals erred in setting forth a rule of automatic liability for supervisory sexual harassment.28 Second, the Court also rejected the Bank's contention (and the district court's conclusion) that the Bank escaped liability because Vinson had failed to use its anti-discrimination policy and procedure.29 The Supreme Court explained that the Bank's policy and procedure were deficient in two respects: (1) the policy did not specifically address sexual harassment, and (2) the grievance procedure required Vinson to first complain to Taylor, her alleged harasser. 30 Had the Bank's procedure been "better calculated to encourage victims of harassment to come forward," the Court opined that the Bank would have had a "substantially stronger" argument for avoiding liability. 3' In the wake of Meritor, the lower federal courts applied different standards of liability to cases of sexual harassment involving supervisors, depending on whether the plaintiff asserted a claim based on quid pro quo or hostile work environment harassment.32 In the case of quid pro quo harassment, "the employer was subject to vicarious liability. 33 In cases of 23 Id. at 63. 24 See id. at 70. 25 See id. at 72-73. 21 Id. at 72. 27 id. 21 See id. 29 See id. at 62, 72. 30 Id. at 72-73. "' Id. at 73. 32 See Burlington Indus. v. Ellerth, 524 U.S. 742, 752-53 (1998). 33 Id. at 753. 202 [Vol. 13:2 Operating in an Empirical Vacuum hostile work environment, most courts applied an "actual-or-constructive knowledge standard," regardless of whether the harasser was a supervisor or a co-worker. 34 Moreover, anti-harassment policies and procedures became more important in the liability calculus. Employers that established policies and procedures meeting Meritor's minimums-a policy specifically addressing sexual harassment and a procedure allowing employees to bypass harassing supervisors-frequently escaped liability after Meritor.35 B. Ellerth and Faragher On June 26, 1998, the Supreme Court handed down its decisions in Ellerth and Faragher. Both decisions set forth the same affirmative defense to employer liability, but each case focused on a somewhat different aspect of the liability equation. In Ellerth, the Court held that the labels quid pro quo and hostile work environment do not determine employer liability, while in Faragher, the Court set forth its rationale for adopting the affirmative defense. 1. Ellerth: Labels Do Not Control Liability Ellerth involved an appeal of a federal district court's order granting summary judgment to Burlington Industries ("Burlington") on Kimberly Ellerth's claim of sexual harassment by Ted Slowik, a mid-level Burlington manager.36 Ellerth contended that Slowik had harassed her and threatened "to deny her tangible job benefits," although such threats never materialized.37 Even though it found that "there was a quid pro quo 'component' to the hostile environment," the district court analyzed Ellerth's claim using a negligence standard. 38 The court held that while Slowik's conduct was sufficiently severe and pervasive to create a hostile work environment, Burlington was not liable because it had neither actual nor constructive knowledge of the harassment. 39 During her fourteen-month 34 Michael J. Phillips, Employer Sexual Harassment Liability Under Agency Principles: A Second Look at Meritor Savings Bank, FSB v. Vinson, 44 Vand. L. Rev. 1229, 1238 (1991). 35 See Ann Juliano & Stewart J. Schwab, The Sweep of Sexual Harassment Cases, 86 Cornell L. Rev. 548, 591 (2001) (finding that in cases decided after Meritor but before Ellerth and Faraghe, "when an employer ha[d] a program that allow[ed] victims to report sexual harassment, plaintiffs [were] successful barely one-third of the time, far less than in other cases"). 36524 U.S. at 746. 37 Id. at 748. 381 d. at 749. '9 See id. 2004] Columbia Journal of Gender and Law tenure at Burlington, Ellerth had never lodged a complaint of harassment against Slowik, although Burlington had an internal grievance procedure.4n Sitting en banc, the Seventh Circuit reversed.41 The court agreed that vicarious liability, not negligence, governed Ellerth's claim.42 Nine of the eleven judges also agreed that Ellerth's claim of unfulfilled threats fell into the category of quid pro quo harassment.43 The judges otherwise failed to reach consensus on the standard for employer liability. 44 The Supreme Court affirmed the Court of Appeals's decision reversing the district court's order granting summary judgment to Burlington.45 The Court held that vicarious employer liability no longer turned on identification of the harassment as quid pro quo.46 Instead, liability now turns on the existence of a tangible employment action,47 which the Court defined as "a significant change in employment status, such as hiring, firing, failing to promote, reassignment with significantly different responsibilities, or a decision causing a significant change in benefits. 48 If a supervisor's harassing conduct culminates in a tangible employment action, such as termination, the employer is strictly liable.49 Otherwise, the employer may assert an affirmative defense to liability or damages. 50 In order to prevail on the affirmative defense, the employer must prove by a preponderance of the evidence both prongs of the defense. Prong one requires the employer to demonstrate that it "exercised reasonable care to prevent and promptly correct sexually harassing behavior," while prong two requires proof that 4 id. 41 Id. 42 id. 43 See id. at 750-51. 4 See id. 451 d. at 766. 46 Id. at 754. Ellerth had "focused all her attention in the Court of Appeals on proving [that] her claim fell within [the quid pro quo] category." Id. at 765. Therefore, the Court remanded the case to give Ellerth "an adequate opportunity to prove she ha[d] a claim for which Burlington [was] liable." Id. at 766. 47 See id. at 765. Because Ellerth's claim involved only unfulfilled threats by Slowik, she could not show a tangible employment action under the new liability paradigm. See id. at 753-54. In its recent decision in Pennsylvania State Police v. Suders, 124 S.Ct. 2342 (2004), the Supreme Court held that sexual harassment by a supervisor which results in constructive discharge may in certain circumstances amount to a tangible employment action. See supra note 11. 48 Ellerth, 524 U.S. at 761. 49 See id. at 765. 50 Id. [Vol. 13:2 Operating in an Empirical Vacuum the employee "unreasonably failed to take advantage of any preventive or corrective opportunities provided by the employer., 51 2. Faragher: The Rationale for the Affirmative Defense Faragher involved a claim by Beth Ann Faragher against the City of Boca Raton ("City") for sexual harassment by two of her supervisors, Bill Terry and David Silverman, while she worked summers as a lifeguard for the City's Marine Safety Section of the Parks and Recreation Department.52 Faragher never complained to higher management about Terry's or Silverman's conduct.5 3 She and other female lifeguards did speak with Robert Gordon, another of Faragher's immediate supervisors. Gordon, however, never reported these conversations to any City official.54 The City learned of Terry's and Silverman's harassment through a complaint filed by a former lifeguard, and after conducting an investigation, sanctioned both Terry and Silverman. 55 Following a bench trial, the district court found the City liable for Terry's and Silverman's harassment.56 A panel of the Court of Appeals for the Eleventh Circuit reversed, concluding that the City could not be held liable for the harassment.57 Sitting en banc, the Eleventh Circuit affirmed the panel's decision.58 The Supreme Court reversed.59 The Court held that an employer may be liable for sexual harassment by a supervisor, even though such conduct falls outside the scope of employment, if the employee was aided in accomplishing the harassment by virtue of the agency relationship.60 Faragher had argued that harassing supervisors misuse the power conferred upon them by the agency relationship to force subordinate employees to endure harassing behavior.6' The Court agreed, acknowledging that victims of harassment may deal with harassment by supervisors differently than co-worker harassment because 51 id. 52 Faragher v. City of Boca Raton, 524 U.S. 775, 780 (1998). 531d. at 782. 541d. at 782-83. 51 See id. at 783. 56 1d. 17 Id. at 783-84. 58 Id. at 784. 59 ld. at 786. 60 See id. at 799-803. 61 Id. at 801. 20041 Columbia Journal of Gender and Law of "the risks of blowing the whistle on a superior."62 The Court expressed concern, however, that if aided by the agency relationship meant "the unspoken suggestion of retaliation by misuse of supervisory authority, [then] the risk of automatic liability [was] high. 63 In order to square the aided by the agency relationship foundation for employer liability with Meritor's holding that employers are not automatically liable for sexual harassment by supervisors, the Court created an affirmative defense to liability when such harassment does not culminate in a tangible employment action. 64 In applying the affirmative defense to the facts in Faragher, the Court concluded that a remand to allow the City the opportunity to present evidence on the affirmative defense was unnecessary because the City could not satisfy the first prong of the defense.65 First, the City "had entirely failed to disseminate its policy against sexual harassment among the beach employees and . ..made no attempt to keep track of the conduct of supervisors like Terry and Silverman. 66 Second, the employer's policy did not provide employees with a means to bypass a harassing supervisor in 67order to lodge a complaint of sexual harassment, a procedural deficiency first noted by the Supreme Court twelve years earlier in its decision in 68Meritor. The Court concluded that these shortcomings meant that the City could not, as a matter of law, "be found to have exercised reasonable care to prevent the supervisors' harassing conduct," and, thus, the affirmative defense was foreclosed to it.69 C. Flawed Assumptions: The Supreme Court's Adoption of the Affirmative Defense In Faragher, the Supreme Court stated that Title VII's "primary objective" was prevention of workplace discrimination. 70 The affirmative defense was the Court's attempt to create incentives for both employers and 62 Id. at 803. 63 Id. at 804. 64 See id. at 804, 808. 651 Id. at 808-09. 66Id. at 808. 67 id. 68See Meritor Say. Bank v. Vinson, 477 U.S. 57, 73 (1986) (stating that it was "not altogether surprising that [Vinson] failed to invoke the procedure and report her grievance" because the "bank's grievance procedure apparently required an employee to complain first to her supervisor," who was Vinson's alleged harasser). 69 Fa ,her 524 U.S. at 808. 70 d. at 806 (quoting Albermarle Paper Co. v. Moody, 422 U.S. 405,417 (1975)). 206 [Vol. 13:2 20041 Operating in an Empirical Vacuum 207 employees to effectuate that primary statutory goal. By limiting liability for employers based, in part, on their implementation of effective anti- harassment procedures, the Court sought to provide incentives for employers to develop proven and effective sexual harassment policies and procedures. The Court concluded that if effective procedure would "encourage employees to report harassing conduct before it bec[ame] severe or pervasive, it would also serve Title VII's deterrent purpose.' The Court's reliance on the affirmative defense as the means to effectuate Title VII's primary goal of prevention, however, rests on two flawed assumptions. First, there is no empirical evidence as to what constitutes an "effective" policy or grievance procedure. The EEOC has published guidelines, as have various authors (see e.g., Gutek, 1997), regarding sexual harassment policies and procedures. For example, policies should include a strong statement that sexual harassment will not be tolerated, explain sanctions, and should be clearly and regularly communicated via multiple methods. Procedures should include informal and formal mechanisms, be designed to encourage victims to come forward, and should not require complaining to an offending supervisor. They should ensure confidentiality as much as possible and include protection against retaliation. Although these guidelines exist, there are very little data regarding the effects of these various suggestions. For example, I was unable to locate any studies that assessed the effects of variations in policies and procedures on the incidence of sexual harassment. 72 So, how did the Supreme Court expect the lower federal courts to evaluate the effectiveness of an employer's policy and procedure when there is no empirical evidence as to which provisions, for example, promises of confidentiality, are necessary for a policy and procedure to be effective? Did the Court expect employers to generate studies of their own internal policies and procedures? That would make sense; if an employer expects to relieve itself of liability based on its internal grievance mechanisms, then the employer should have some obligation to show that 71 Burlington Indus. v. Ellerth, 524 U.S. 742, 764 (1998). 72 Patricia A. Frazier, Overview of Sexual Harassment from the Behavioral Science Perspective, 1998 A.B.A. Center for Continuing Legal Education, National Institute on Sexual Harassment B-10 (1998); see also Deborah Erdos Knapp et al., Determinants of Target Responses to Sexual Harassment: A Conceptual Framework, 22 Acad. Mgmt. Rev. 687, 701 (1997) (stating that while the sexual harassment literature was "replete with recommendations for developing [sexual harassment] policies in organizations ...the influence that different aspects of policies and procedures may have on [sexual harassment] reporting behavior has not been explored extensively"). 208 Columbia Journal of Gender and Law [Vol. 13:2 those procedures actually work. The Court in Ellerth and Faragher, however, did not address that question, no doubt because it never acknowledged that we simply do not know what constitutes an effective anti-harassment policy and grievance procedure. Second, in order for the employer's grievance mechanism to work, the employee must alert the employer to the harassment, normally by filing a complaint. The research over the past twenty years, however, has consistently shown that "filing a formal complaint or reporting the harassment to an authority appears to be a very uncommon occurrence. 73 In its three large-scale studies of sexual harassment in the federal workplace, the Merit Systems Protection Board ("MSPB") found that only between two percent and six percent of male and female victims of harassment took some kind of formal action, such as requesting an investigation or filing a grievance or complaint, in response to the harassment.74 Schneider, Swan, and Fitzgerald found in their study of sexual harassment at a large, private-sector company and a large Midwestern university that only 13.3 percent of the women in the private- sector sample and six percent of the women in the university sample filed a formal complaint in response to the "worst" harassment experience that they had encountered at work.75 In other studies published between 1982 and 1994, researchers found reporting rates ranging from a low of five percent to a high of eighteen percent.76 When the harasser is a supervisor, 73 Bonnie S. Dansky & Dean G. Kilpatrick, The Effects of Sexual Harassment, in Sexual Harassment: Theory, Research, and Treatment 152, 158 (William O'Donohue ed., 1997); see Mindy E. Bergman et al., The (Un)reasonableness of Reporting: Antecedents and ConseNuences of Reporting Sexual Harassment, 87 J. Applied Psychol. 230 (2002) (citations omitted) (noting that "[m]ost harassment targets do not report their experiences ... despite the supposed benefits of doing so"); Frazier, supra note 72, at B-15 (stating that "the one consistent finding [of the studies of responses to sexual harassment] is that very few women make a formal complaint or report") (emphasis omitted); Lawton, supra note 16, at 86-87 (summarizing studies showing low incidence of formal reporting of sexual harassment in the workplace). 74 See MSPB, Sexual Harassment in the Federal Workplace: Is it a Problem? 71 (1981) (finding that two percent of male and three percent of female victims filed formal complaints); MSPB, Sexual Harassment in the Federal Government: An Update 27 (1988) (reporting that five percent of victims took formal action); MSPB, Sexual Harassment in the Federal Workplace: Trends, Progress, Continuing Challenges 33, 34 tbl.10 (1995) [hereinafter 1995 MSPB] (finding that six percent of victims took formal action). 75 Kimberly T. Schneider et al., Job-Related and Psychological Effects of Sexual Harassment in the Workplace: Empirical Evidence from Two Organizations, 82 J. Applied Psychol. 401, 406-08 & tbl.2 (1997). 76 See Dansky & Kilpatrick, supra note 73, at 158 (summarizing reporting rates of various studies). Operating in an Empirical Vacuum the victim is even less likely to formally report the harassment.77 The Supreme Court acknowledged as much in Faragher: The agency relationship affords contact with an employee subjected to a supervisor's sexual harassment, and the victim may well be reluctant to accept the risks of blowing the whistle on a superior. When a person with supervisory authority discriminates in the terms and conditions of subordinates' employment, his actions necessarily draw upon his superior position over the people who report to him, or those under them, whereas an employee generally cannot check a supervisor's abusive conduct the same way that she might deal with abuse from a co-worker. When a fellow employee harasses, the victim can walk away or tell the offender where to go, but it may be difficult to offer such responses to a supervisor, whose "power to supervise - [which may be] to hire and fire, and to set work schedules and pay rates - does not disappear ...when he chooses to harass through insults and offensive gestures rather than directly with threats of firing or promises of promotion. 78 Yet if formal reporting is an uncommon occurrence, in particular when the harasser is the victim's supervisor, then how is deterrence increased by adopting a rule whose success depends, in large part, on formal reporting? Thus, the empirical evidence on reporting raises an interesting question: if the vast majority of harassment victims do not report harassment, then the reasonable response is not to report harassment.79 Under what circumstances, then, could an employer ever satisfy prong two of the affirmative defense? If the Supreme Court did not intend the lower federal courts to measure the reasonableness of victims' responses against the empirical data, at least some of which the Court was aware of,80 then 77 See Knapp et al., supra note 72, at 694 (citations omitted) (stating that "[tihe research indicates that, with few exceptions ..., a target of [sexual harassment] is less likely to report the behavior if the initiator is her/his superior"); James E. Gruber & Michael D. Smith, Women's Responses to Sexual Harassment: A Multivariate Analysis, 17 Basic & Applied Soc. Psychol. 543 (1995) (percentages omitted) (finding that "[w]omen were less apt to respond directly to a supervisor [ ] than a coworker [ ] or client or customer [ ] [and] were also more likely to have quit a job when the harasser was a supervisor [ as opposed to either a coworker [ ] or client/customer[ ]"). 78 Faragher v. City of Boca Raton, 524 U.S. 775, 803 (1998) (quoting Susan Estrich, Sex at Work, 43 Stan. L. Rev. 813, 854 (1991)). 79 See Frazier, supra note 72, at B-16, B-21 (arguing that "since reporting is very uncommon, not reporting does not seem unreasonable," and concluding, on the basis of the empirical evidence on reporting, that "it would be difficult to argue that a victim acted unreasonably by not reporting"). so In their amicus brief to the Court in Faragher, the National Women's Law Center, Equal Rights Advocates, and Women's Legal Defense Fund cited to several studies, including the MSPB's 1995 report, which showed the low rate of victim reporting of sexual harassment, and argued that the Court should craft liability standards that recognize the 2004] 209 Columbia Journal of Gender and Law what standard of behavior were the lower federal courts to apply in evaluating the plaintiff s duty to avoid harm? In an earlier article, I questioned whether the Court's decisions in Ellerth and Faragher would give employers incentives to do more than merely promulgate anti-harassment policies and grievance procedures: The problem is that the Court in Ellerth and Faragher left important interpretive questions to the lower federal courts to resolve, on a case-by-case basis. Thus, even though employers have the burden of proving that an employee unreasonably failed to use a proven, effective procedure that is not unduly risky or expensive, lower federal courts may interpret this language in such a fashion as to effect little change in current employment practices. While Ellerth and Faragher raise the issue of effective procedures, they fail to enunciate workable rules for employers in constructing effective procedures. As a result, only time will tell, as the lower courts elaborate on the Court's opinions in Ellerth and Faragher, whether these decisions effectively maintain the status quo or effect a major change in employers' incentives for eliminating sexual harassment in the workplace.8 My concern was well-founded. An examination of 200 federal court cases addressing the new Ellerth and Faragher affirmative defense reveals that many federal courts have interpreted the Court's decisions in Ellerth and Faragher to require little more than what the Court in Meritor commanded: promulgate an anti-harassment policy that specifically addresses sexual harassment and a grievance procedure that allows an employee to bypass a harassing supervisor. As a practical matter, the Court's decisions in Ellerth and Faragher did little to change employer incentives to reduce the incidence of sexual harassment by supervisors in the workplace. II. THE LOWER FEDERAL COURT CASES In order to evaluate how the lower federal courts have interpreted the two prongs of the affirmative defense, I Shepardized the Supreme Court's decisions in Ellerth and Faragher on LEXIS, using the Custom function to divide the cases by circuits. Within each circuit, I narrowed the body of cases by conducting a Focus search using the phrase "affirmative defense." The district court results are limited to a thirty-month period, from January 1, 2001 through June 30, 2003. There is no time limitation on the Court of Appeals cases; the results include cases from June 26, 1998, reticence of victims to report harassment, especially when the harasser is their supervisor. Brief of Amici Curiae National Women's Law Center et a., Faracher v. City of Boca Raton, 524 U.S. 775 (1998) (No. 97-282). 81 Lawton, supra note 16, at 110. 210 [Vol. 13:2 2004] Operating in an Empirical Vacuum 211 the date of decision for both Ellerth and Faragher, through June 30, 2003, a five-year period. This search produced a total of 452 cases.8 2 I then eliminated the following types of cases: (1) hostile work environment claims based on anything other than sex, e.g., harassment based on race, national origin, or disability; 83 (2) other claims, including those for sex discrimination, not premised on sexual harassment; and (3) co- worker harassment cases.84 This process of elimination produced a total of 272 cases involving claims of sexual harassment by supervisors: 126 Court of Appeals and 146 district court decisions. 85 The courts in 52 of these 272 cases decided plaintiffs' claims of sexual harassment without addressing the new liability framework enunciated by the Supreme Court in Ellerth and Faragher. In 24 of the 52 cases, the court's decision turns on the plaintiff's ability to make out a prima facie case of sexual harassment, e.g., severe or pervasive conduct. 86 In 28 cases, the court does not discuss the affirmative defense, for example, because the employer did not raise or brief the issue.87 In the remaining 220 cases (out of 272), the court bases its decision on the Supreme Court's framework in Ellerth and Faragher. I divided these 220 decisions into two categories: (1) the 20 cases in which the court's decision turns on the existence of a tangible employment action,88 and (2) 82 See infra tbl.I, col.A, Total. 83 Most federal courts analyze employer liability for harassment based on race, national origin, religion, age, or disability the same way that they analyze sexual harassment cases. See, e.g., Abramson v. William Patterson Coll. of N.J., 260 F.3d 265, 276 n.5 (3d Cir. 2001); Jackson v. Ouanex, 191 F.3d 647, 658 (6th Cir. 1999). I deleted other harassment cases from the body of cases because the empirical literature on which this Article is based only discusses sexual harassment. 84 See infra tbl.I, cols.B-D. The vast majority of federal courts apply the affirmative defense only to claims of supervisory harassment. The Western District of Kentucky, however, has applied the affirmative defense in a co-worker harassment case. See Good v. MMR Group, Inc., No. 3:00-CV-182-H, 2001 U.S. Dist. LEXIS 20036, at *12 (W.D. Ky. Dec. 4, 2001) (applying affirmative defense to a claim of co-worker harassment because "allowing an affirmative defense in cases of supervisor harassment, but not coworker harassment, would produce the anomalous result of making liability 'stricter' in coworker cases than supervisor cases"). I included Good in the group of cases involving harassment by supervisors because it sheds light on how the court interprets the affirmative defense. 85 See infra tbl.I, col.E, Total. I included remand decisions in this total because some remand decisions contain meaningful discussions of the employer's burden under prongs one and two of the affirmative defense. See, e.g., Lissau v. S. Food Serv., 159 F.3d 177 (4th Cir. 1998) discussed infra Part III.A.3. 86 See infra tbl.II, col.B, Total. 87 See infra tbl.II, col.C, Total. 88 See infra tbl.II, col.D, Total; supra notes 47-51 and accompanying text for a discussion of the role a tangible employment action plays in the Supreme Court's new Columbia Journal of Gender and Law the remaining 200 cases in which the court's decision rests on the employer's ability to satisfy the two prongs of the affirmative defense.89 While some of the twenty tangible employment action cases appear in the footnotes of this Article, for example, in support of assertions about the impact of organizational culture and job gender context on the likelihood of sexual harassment, I base my findings principally on those 200 cases in which the courts' decisions rest on the elements of the affirmative defense. One caveat is necessary before moving on. Cases obtained from LEXIS, or any other electronic retrieval source, represent only a fraction of the total number of cases involving claims of sexual harassment by supervisors. Victims of harassment who do not file with the EEOC, who do not file suit in federal court, or who settle their cases are not included in the results obtained from any LEXIS search. The research is nonetheless valuable. Employers and their counsel rely on the published cases to discover what behaviors are necessary in order to satisfy the employers' obligations to prevent and correct workplace harassment. Because attorneys, employers, and academics all labor under the same constraint, i.e., the inability to access and analyze all relevant sexual harassment claims, the cases that are published provide insight into the kinds of conduct that the federal courts reward. It is precisely these incentives that this Article explores. III. FLAWED INCENTIVES: INTERPRETING THE AFFIRMATIVE DEFENSE From the start, it was questionable whether the affirmative defense was the best way to achieve what the Court labeled as Title Vl's primary goal: prevention of workplace harassment. The success of the defense depended on the lower federal courts' ability to determine the effectiveness of employers' policies and procedures-a task made more difficult by the absence of empirical evidence on what makes a policy and procedure effective. It also required employees to formally report harassing conduct, which the research shows is an uncommon response to sexual harassment. While the affirmative defense rests on flawed assumptions, the lower federal courts could have interpreted the employer's obligations and burden of proof under the two prongs of the defense so as to give employers incentives to reduce the incidence of workplace harassment. Instead, the lower federal courts have interpreted the employer's obligations and burden of proof so that employers have little incentive to do anything besides promulgate policies and procedures that look good on paper. liability paradigm under Ellerth and Faragher. See also supra note 11 on the Supreme Court's recent decision in Pennsylvania State Police v. Suders, 124 S.Ct. 2342 (2004). 89 See infra tbl.II, col.E, Total. 212 [Vol. 13:2 2004] Operating in an Empirical Vacuum 213 The courts concentrate on "simple, quantifiable standards against which to measure" the employer's efforts to fulfill its obligation to deter sexual harassment in the workplace. 9° By doing so, they objectify the inquiry required under both prongs of the affirmative defense, contrary to the express language of the affirmative defense, which focuses on the reasonableness of both the employer's and the employee's conduct. The frequency in some circuits, in particular the Second and the Fourth, with which employers prevail on motions for summary judgment or judgment as a matter of law on the basis of the affirmative defense attests to the objectification of the reasonableness inquiries in prongs one and two of the affirmative defense. In fact, it is not uncommon for an employer to concede or a court to assume that the workplace conduct of which the plaintiff complains was severe or pervasive. Employers and courts do so in order to dispose of the plaintiff's case on the basis of the affirmative defense on a motion for summary judgment or judgment as a matter of law. 91 90 See An Academy Classic, supra note 2, at 12. Professor Kerr attributes the prevalence of "fouled-up reward systems," in part, to the appeal of objective evaluation criteria. 91 See, e.g., Wallace v. San Joaquin County, No. 02-15278, 2003 U.S. App. LEXIS 3540 (9th Cir. Feb. 13, 2003) (holding, in summary opinion containing no discussion of whether supervisor's conduct was severe or pervasive, that employer was entitled to summary judgment based on the affirmative defense); Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1292 n.1 (lth Cir. 2000) ("assum[ing] arguendo that [the supervisor's] conduct constituted sexual harassment" and affirming trial court order granting summary judgment to employer on the basis of the affirmative defense), cert. denied, 531 U.S. 926 (2000); Leugers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *7 (6th Cir. Feb. 3, 2000) (noting that the "parties agreed for purposes of summary judgment that [the supervisor's] actions [had] created a hostile working environment"); Wright v. Anixter. Inc., No. 98-17164, 1999 U.S. App. LEXIS 19962, at *2 (9th Cir. Aug. 18, 1999) (assuming the existence of a hostile work environment and resting its decision, as had the district court, on the employer's satisfaction of the affirmative defense); McPherson v. City of Waukegan, No. 01 C 9264, 2003 U.S. Dist. LEXIS 9098, at *14 (N.D. Ill. May 29, 2003) (stating that employer did not challenge the existence of a hostile work environment, instead arguing it was entitled to summary judgment based on the affirmative defense); Sutton v. Zemex Corp., 261 F. Supp. 2d 392, 400-01 (W.D.N.C. 2003) (stating that "even if Plaintiff could assert a timely claim of hostile work environment sexual harassment" based on her supervisor's conduct, "the undisputed evidence establish[ed] that Defendant may not be held liable for that conduct under Ellerth and Faraghe'); Stone- Graves v. Coop. Elevator Co., No. 01-CV-74280-DT, 2003 U.S. Dist. LEXIS 5834, at *14 (E.D. Mich. Mar. 12, 2003) (assuming that plaintiff could satisfy the elements of her prima facie case and concluding that defendant was entitled to summary judgment based on the affirmative defense); Arnold v. Yale New Haven Hosp., 213 F. Supp. 2d 142, 143 n.5 (D. Conn. 2002) (stating that employer moved for summary judgment on the basis of affirmative defense and granting employer's motion on that ground); Roelen v. Akron Beacon Journal, 199 F. Supp. 2d 685, 695 n.5 (N.D. Ohio 2002) (stating that because plaintiff had not used the employer's sexual harassment policy, the court did not have to address whether the harasser's conduct was severe or pervasive, but noting that it likely was "not sufficiently severe or pervasive" to create a hostile work environment); Taylor v. Nickels & Dimes, Inc., No. 3:00-CV-1461-L, 2002 U.S. Dist. LEXIS 14549 (N.D. Tex. Aug. 7, 2002) (The 214 Columbia Journal of Gender and Law [Vol. 13:2 This is odd. If anything, it should be more difficult for the employer to prevail on such motions on the basis of the affirmative defense than on the ground that the conduct was not severe or pervasive (or some other element of the plaintiff s prima facie case) because the employer bears the burden of proof on the former, 92 not the latter.93 The fact that some courts jump right to the question of liability demonstrates that it is relatively easy for the employer to win on the affirmative defense on summary judgment or judgment as a matter of law in some circuits. The lower federal courts are accomplishing the objectification of the affirmative defense in several ways. First, federal courts in virtually every circuit interpret the employer's obligation to prevent workplace harassment under prong one of the affirmative defense as satisfied by proof Northern District of Texas authored two separate Taylor opinions, Taylor I and Taylor II, granting the employer's motions for summary judgment on plaintiffs Erika Square's and Kimberly Taylor's hostile work environment claims, respectively) (Taylor 1-containing no discussion of whether the objectionable conduct was severe or pervasive and granting employer's motion for summary judgment on the basis of the affirmative defense), aff'd, No. 02-10950, 2003 U.S. App. LEXIS 13103 (5th Cir. June 3, 2003); Walton v. Johnson & Johnson Servs.. Inc., 203 F. Supp. 2d 1312, 1315 (M.D. Fla. 2002) (containing no discussion of whether the conduct was severe or pervasive, the employer having conceded that plaintiffs supervisor had "committed 'unwelcome sexual advances,"' and granting employer's motion for summary judgment on the basis of the affirmative defense), affd, 347 F.3d 1272 (11 th Cir. 2003); Duviella v. Counseling Serv., No. 00-CV-2424, 2001 U.S. Dist. LEXIS 22538, at *32 (E.D.N.Y. Nov. 20, 2001) (granting employer's motion for summary judgment, stating that while "defendants do not technically concede that Duviella has established a hostile work environment, they focus their entire argument on whether liability can be imputed to the employer"), aff'd, No. 02-7019, 2002 U.S. App. LEXIS 24050 (2d Cir. Nov. 21, 2002); cf Daniels v. Home Depot, Inc., No. 01-466, 2002 U.S. Dist. LEXIS 11990, at *18-19 (E.D. La. June 26, 2002) (granting employer's motion for summary judgment in a race harassment case, noting that while the employer had claimed that the behavior was severe or pervasive, the court would "pretermit that argument and assume without deciding that plaintiff could prove such an environment" because the employer was entitled to summary judgment on the basis of the affirmative defense). 92 See Burlington Indus. v. Ellerth, 524 U.S. 742, 765 (1998) (stating that employers "may raise an affirmative defense to liability or damages, subject to proof by a preponderance of the evidence"). 93 Because the employer bears the burden of proof at trial on the affirmative defense, it should rarely prevail on a motion for summary judgment based on the affirmative defense. See 11 James Win. Moore et al., Moore's Federal Practice 156.31[2] (3d ed. 2003) (footnote omitted) (explaining that it is easier for a party defending a claim to succeed on a motion for summary judgment because it "need only show that the claimant's case falls short on a single element of the showing required to sustain the cause of action," but when the party moving for summary judgment bears the burden of persuasion, it not only must "make a prima facie showing of an absence of genuine factual disputes, [it] must also demonstrate that the facts set forth by the movant are so firmly established that no reasonable trial fact finder could reject them and would be required to grant judgment as a matter of law at trial"); see also Jack H. Friedenthal et al., Civil Procedure 9.3, at 463 (3d ed. 1999) (footnote omitted) (stating that "courts should be cautious in granting summary judgment for a party with the burden of proof"). Operating in an Empirical Vacuum of a sexual harassment policy and grievance procedure. Many then assume, based on little more than an examination of the policy and procedure on paper, that the policy and procedure are effective and reasonable. The employer, therefore, can satisfy the prevention element of prong one with "objective" evidence-a paper policy and procedure-that is easy to produce. Second, some federal courts allow the employer to define what constitutes notice from the employee of harassing behavior. An employer then can restrict the avenues and methods available for filing complaints as a way to limit liability. Without notice of the harassment, courts hold that an employer's duty to respond under prong one is not triggered. Therefore, a reporting case is transformed into a non-reporting or delayed-reporting case, thereby making it easier for employers to satisfy the correction element of prong one. Moreover, because prong one's correction element is tied to prong two's reporting element, allowing an employer to restrict when it obtains notice of harassing behavior also makes it easier for employers to satisfy prong two of the affirmative defense. Third, many federal courts assume that if an employer has an anti- harassment policy and procedure, that policy and procedure are reasonable. Therefore, a plaintiff's delay in reporting or failure to report harassment is deemed per se unreasonable. Reducing the inquiry to whether the plaintiff reported at all, or if she did report, whether she did so immediately after the first incident of objectionable conduct, as opposed to a month, two months, three months, or even six months later,94 eliminates the need to evaluate the reasonableness of the plaintiffs actions in the context of the specific case facts and the empirical evidence on reporting. This, in turn, transforms a fact-intensive inquiry into one susceptible of easily provided, objective proof. By transforming the issue of reasonableness in both prongs of the affirmative defense from fact-intensive to "objective," factually insensitive inquiries, the lower federal courts have created a "fouled-up reward system[]" for employers.95 Employers have no incentive to focus on the actual predictors of workplace harassment, e.g., organizational culture and job gender context.96 Instead, formulating and disseminating policies and procedures normally suffices to satisfy the employer's prevention obligation. Because the employer's obligation to correct harassment generally is not triggered until it obtains notice, an employer also has an incentive to narrow the avenues available for employees to file complaints. The courts do not require employers to track what happens to harassment victims, to survey employees to determine the effectiveness of internal 94 See infra notes 251-256 and accompanying text. 95 See An Academy Classic, supra note 2, at 12. 96 See infra Parts III.A.2.b, 1II.A.2.c. 2004] 215 216 Columbia Journal of Gender and Law [Vol. 13:2 grievance procedures, to evaluate supervisors on their reporting of harassment and compliance with the employer's anti-harassment policies, or to engage in any meaningful evaluation of training programs offered. Thus, employers have no reason to assume any of these responsibilities. In fact, the courts have made it so easy for employers to satisfy the two prongs of the affirmative defense that they effectively have shifted the burden of proof on the reasonableness inquiries of both prongs of the defense from the employer to the employee. By doing so, the courts give employers incentives to engage almost exclusively in liability-avoiding strategies, including using the affirmative defense primarily as a means to prevail on motions for summary judgment, rather than as a mechanism for reducing the incidence of workplace harassment, as intended by the Supreme Court in Ellerth and Faragher.97 A. Prong One: An Ounce of Prevention Isn't Worth a Pound of Cure The lower federal courts have consistently held that an employer satisfies its obligation to prevent harassment by doing little more than creating and disseminating (or posting) an anti-harassment policy with a grievance procedure.98 This phenomenon of equating prevention with paper policies and procedures is occurring in virtually every circuit. Even courts, such as those in the Second Circuit, that acknowledge that "the existence of an anti-harassment policy with complaint procedure" is "not necessarily dispositive" 99 on the first prong of the affirmative defense, nonetheless proceed to analyze the employer's obligation to prevent harassment by looking no further than the employer's paper policies and procedures. 0 97 The interpretation given to the elements of the affirmative defense by the lower federal courts also tends to blur the line between the standard of liability for harassment by co-workers as opposed to supervisors. See, e.g., Wilson v. Tulsa Junior Coll., 164 F.3d 534, 541 n.4 (10th Cir. 1998) (explaining that while "many of the same facts will be relevant to both negligence and vicarious liability claims, in asserting that the employer was negligent, the plaintiff bears the burden of establishing that the employer's conduct was unreasonable; while in a misuse of authority claim, the employer bears the burden of establishing as an affirmative defense that it exercised reasonable care to prevent harassment"); Bartniak v. Cushman & Wakefield, Inc., 223 F. Supp. 2d 524, 529 (S.D.N.Y. 2002) (explaining that "the major difference" in employer liability standards between a co-worker and supervisory harassment case is the burden of proof). The lower federal courts apply a negligence standard to co-worker harassment cases. See Faragher v. City of Boca Raton, 524 U.S. 775, 799 (1998) (citing cases). 98 See infra notes 101-105, and notes 151-156 and accompanying text. 99 Caridad v. Metro-N. Commuter R.R., 191 F.3d 283, 295 (2d Cir. 1999), cert. denied, Metro N. Commuter R.R. v. Norris, 529 U.S. 1107 (2000). 100See, e.g., Leopold v. Baccarat, Inc., 239 F.3d 243, 245 (2d Cir. 2001) (affirming district court order granting employer's motion for summary judgment, stating that "the existence of an anti-harassment policy with complaint procedures" was "not necessarily dispositive" on prong one, but in the next sentence "conclud[ing] that [the employer] had in 2004] Operating in an Empirical Vacuum 217 Some courts explicitly hold that creating and distributing an anti- harassment policy and procedure fulfill prong one's element of prevention.101 Most federal courts, however, do not clearly delineate place an anti-harassment policy and an accompanying complaint procedure sufficient to satisfy the first prong of the affirmative defense"); Caridad, 191 F.3d at 295 (affirming district court order granting summary judgment to employer, noting that the existence of an anti-harassment policy is "not necessarily dispositive" on prong one of the affirmative defense, yet concluding that employer satisfied prong one of the defense by having "an anti- harassment policy with a procedure for filing complaints" and "endeavor[ing] to investigate and remedy problems reported by its employees"). See also Hairston-Lash v. R.J.E. Telecom, Inc., 161 F. Supp. 2d 390, 394 (E.D. Pa. 2001) (granting employer's motion for summary judgment, stating that having an "extensive antiharassment policy is evidence" of the employer's preventative efforts, but then engaging in no further analysis of the employer's preventative efforts and concluding that because the plaintiff had not invoked her employer's complaint procedure, the employer had established the affirmative defense). 101 See, e.g., Kohler v. Inter-tel Techs, 244 F.3d 1167, 1180 (9th Cir. 2001) (affirming trial court order granting summary judgment to employer, finding that affirmative defense would apply to claims under California employment discrimination law and that the employer's "policy and its efforts to disseminate the policy to its employees establish that [the employer] exercised reasonable care to prevent sexual harassment in the workplace"); Matvia v. Bald Head Island Mgmt., Inc., 259 F.3d 261, 268 (4th Cir. 2001) (affirming district court order granting summary judgment to employer and finding that "[in the face of a policy that clearly defines sexual harassment and to whom harassment should be reported," plaintiff could not demonstrate that her employer "lacked an effective preventative program"); Barrett v. Applied Radiant Energy Corp., 240 F.3d 262, 266 (4th Cir. 2001) (stating that there was "little difficulty [ ] in concluding that [the employer] took reasonable steps to prevent sexual harassment" because the employer distributed an anti-harassment policy that allowed employees to bypass their supervisors in order to file a complaint and many of the managers with whom a complaint could be filed worked in the same building with and, thus, were accessible to the plaintiff); Lissau v. S. Food Serv., 159 F.3d 177, 182 (4th Cir. 1998) (explaining, in dicta, that upon remand the district court might be able to grant summary judgment to the employer on plaintiff s supervisory harassment claim, noting that "evidence that [the employer] had disseminated an effective anti-harassment policy provides compelling proof of its efforts to prevent workplace harassment"); Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1297-99 (1lth Cir. 2000) (affirming district court's order granting summary judgment to employer on plaintiff employees' sexual harassment claim and finding that employer had exercised reasonable care to prevent harassment because it had created and distributed a policy and complaint procedures that were not "inherently defect[ive]"), cert. denied, 531 U.S. 926 (2000); Montero v. AGCO Corp., 192 F.3d 856, 862 (9th Cir. 1999) (emphasis in original) (affirming trial court order granting employer motion for summary judgment and concluding that employer's "policy and its efforts to ensure that all employees were aware of the policy establishes that [the employer] exercised reasonable care to prevent sexual harassment in its workplace); McCombs v. Chrysler Corp., No. IP 99-0697-C-T/K, 2003 U.S. Dist. LEXIS 6448, at *19-20 (S.D. Ind. Mar. 4, 2003) (granting employer's motion for summary judgment on plaintiffs claim of a hostile work environment, explaining that inclusion of employer's anti-harassment policy in the collective bargaining agreement satisfied employer's obligation to exercise reasonable care to prevent harassment); Smith v. Am. Int'l Group, Inc., No. 01 Civ. 1652, 2002 U.S. Dist. LEXIS 7472, at *20-21 (S.D.N.Y. Apr. 26, 2002) (recommending that employer's motion for summary judgment be granted and holding that "by establishing an effective sexual harassment policy and providing [plaintiff] with a copy, [the employer] unquestionably took reasonable steps to prevent improper conduct"), adopted by No. 01 Civ. 1652, 2002 U.S. Dist. LEXIS 8954 (S.D.N.Y. May 21, 2002); cf. Jones v. Rent-A-Center, 218 Columbia Journal of Gender and Law [Vol. 13:2 between their analysis of the correction and prevention elements of prong one. Instead, they combine into one discussion the employer's prevention and correction efforts, concluding that the employer satisfies prong one of the affirmative defense by distributing an anti-harassment policy and responding promptly once a complaint is filed. 102 Because the employer's 240 F. Supp. 2d 1167, 1180 (D. Kan. 2002) (denying employer's motion for summary judgment on the basis of the affirmative defense because there were questions as to the employer's efforts to correct the harassment, but concluding that employer had "exercised reasonable care to prevent sexual harassment by maintaining a complaint procedure, an Open Door Policy, posting the policy at the workplace, and distributing the policy to its employees"); Morton v. Steven Ford-Mercury, 162 F. Supp. 2d 1228, 1245 (D. Kan. 2001) (denying employer's motion for summary judgment on the basis of the affirmative defense because there were questions about employer's ability to satisfy prong two, but concluding that the employer's "efforts to establish a company wide policy and to ensure that employees were aware of the policy constitutes an exercise of reasonable care in the prevention of sexual harassment"). 102 See, e.g., Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *4-5 (4th Cir. Nov. 13, 2001) (affirming district court order granting summary judgment to employer, rejecting plaintiffs argument that her employer had not satisfied prong one because it had not shown that her harasser had received sexual harassment training by concluding that "'distribution of an anti-harassment policy provides compelling proof that the company exercised reasonable care in preventing and promptly correcting sexual harassment"' (quoting Barrett v. Applied Radiant Energy Corp., 240 F.3d 262, 266 (4th Cir. 2001)); Gawley v. Ind. Univ., 276 F.3d 301, 311-12 (7th Cir. 2001) (affirming district court order granting summary judgment on plaintiffs sexual harassment claim, concluding that plaintiff had "no evidence that the [employer] failed to exercise reasonable care in preventing and correcting the harassing behavior" because the employer "had a system in place for employees to report sexual harassment" and, once reported, the employer "took action and the harassment stopped"); Hill v. Am. Gen. Fin., Inc., 218 F.3d 639, 643-44 (7th Cir. 2000) (affirming district court order granting summary judgment to employer on plaintiff's claim of sex and race harassment, noting that while the employer's policies left "room for improvement," the employer "took immediate corrective action" once it learned of the harassment and had posted "in a 'public access type place"' its anti-harassment policies), en banc reh'g denied, No. 99-2682, 2000 U.S. App. LEXIS 18367 (7th Cir. July 27, 2000); Leugers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *8 (6th Cir. Feb. 3, 2000) (affirming district court order granting summary judgment to employer on plaintiffs claim of supervisory harassment, concluding that employer had satisfied prong one of the affirmative defense by "distribut[ing] an anti-harassment policy with a complaint procedure ... that included ways to bypass an immediate supervisor in making a complaint" and by taking prompt corrective action once it learned of the harassment); Casiano v. AT&T Corp., 213 F.3d 278, 286-87 (5th Cir. 2000) (affirming district court order granting summary judgment to employer on plaintiff s sexual harassment claim, concluding that it would waste judicial resources to remand the case for trial given that AT&T could satisfy the affirmative defense and explaining that AT&T had several publications articulating its anti-harassment policy, plaintiff was aware of those policies, and AT&T acted promptly once plaintiff complained of harassment); McPherson v. City of Waukegan, No. 01 C 9264, 2003 U.S. Dist. LEXIS 9098, at *19-20 (N.D. I11. May 30, 2003) (granting employer's motion for summary judgment, noting that employer exercised reasonable care to prevent and correct harassment "through its policy that prohibited sexual harassment"); Taylor v. Nickels & Dimes, Inc., No. 3:00-CV-1461-L, 2002 U.S. Dist. LEXIS 14549, at *16-17 (N.D. Tex. Aug. 7, 2002) (Talor 1l-granting employer's motion for summary judgment and concluding that the employer had satisfied the first prong of the 2004] Operating in an Empirical Vacuum 219 affirmative defense because it had posted its anti-harassment policy in the workplace and plaintiff was aware of the policy and complaint procedures), aff'd, No. 02-10950, 2003 U.S. App. LEXIS 13103 (5th Cir. June 3, 2003); Olsen v. H.E.B. Pantry Foods, 196 F. Supp. 2d 436, 440 (E.D. Tex. 2002) (granting employer's motion for summary judgment on sexual harassment claim, concluding that the employer had met the "first prong of the affirmative defense by having a sexual harassment policy and taking immediate remedial action as soon as [the plaintiff] complained" of the harassing behavior); Pickett v. Colonel of Spearfish, 209 F. Supp. 2d 999, 1007 (D.S.D. 2001) (granting employer's motion for summary judgment on plaintiffs sexual harassment claim and noting, in dicta, that employer could prevail on the affirmative defense, in part, because the employer "had in place policies and procedures for handling sexual harassment [and] [t]hese policies and procedures were put into effective use through the Hothem complaint, the immediate investigation, and the prompt firing of Jones some three months after the alleged rape of Pickett"); Clardy v. Silverleaf Resorts, Inc., No. 3:99-CV-2893-P, 2001 U.S. Dist. LEXIS 16481, at *28-30 (N.D. Tex. Oct. 10, 2001) (granting employer's motion for summary judgment on plaintiff's supervisory hostile work environment claim, explaining that the employer had exercised reasonable care to prevent and correct harassment and noting that employer had provided employees with its handbook containing both a sexual harassment policy and grievance procedure, had posted copies of "sexual harassment policy statements on the bulletin boards," and had taken immediate action once it learned of plaintiff's harassment complaint); Green v. Wills Group, Inc., 161 F. Supp. 2d 618, 625 (D. Md. 2001) (granting defendant employer's motion for summary judgment on sexual harassment claim, stating that "distribution of an anti-harassment policy provides 'compelling proof' that the defendant exercised reasonable care in preventing and promptly correcting sexual harassment unless plaintiff shows that the employer adopted or administered the policy in bad faith or that the policy was otherwise defective or dysfunctional"); O'Dell v. Trans World Entm't Corp., 153 F. Supp. 2d 378, 388-90 (S.D.N.Y. 2001) (granting employer's motion for summary judgment on plaintiff's hostile work environment claim, concluding that employer had satisfied prong one of the affirmative defense by offering proof of its anti-harassment policy, which plaintiff could rebut with evidence that the policy was ineffective), afid, 2002 U.S. App. LEXIS 14446 (2d Cir. July 16, 2002); Slay v. Glickman, 137 F. Supp. 2d 743, 752 (S.D. Miss. 2001) (granting employer's motion for summary judgment on plaintiff's hostile work environment claim, concluding there was no genuine issue of material fact on prong one of the affirmative defense because the employer had a sexual harassment policy and investigated plaintiff's complaint); see also Lewis v. Forest Pharm., Inc., 217 F. Supp. 2d 638, 649 n.5 (D. Md. 2002) (noting, in dicta, that had plaintiff not suffered a tangible employment action, the employer would have prevailed on the affirmative defense, and concluding that the employer satisfied prong one of the defense because its anti-harassment policy was "designed to deter sexual harassment," even though no evidence of how the policy's design did, in fact, deter harassment was provided); Duviella v. Counseling Serv., No. 00-CV-2424, 2001 U.S. Dist. LEXIS 22538, at *36 (E.D.N.Y. Nov. 20, 2001) (stating that "many district courts have held that the first prong of the affirmative defense has been satisfied when evidence exists that the employer had an anti-harassment policy in place that was readily available to the complainant"), afTd, No. 02-7019, 2002 U.S. App. LEXIS 24050 (2d Cir. Nov. 21, 2002); cf Marrero v. Goya, 304 F.3d 7, 21-22 (1st Cir. 2002) (affirming trial court order denying employer's motion for judgment as a matter of law, explaining that the employer had not shown that the facts at trial compelled a decision on the affirmative defense because the employer had not shown "the existence of an antiharassment policy with a known complaint procedure"); Johnson v. West, 218 F.3d 725, 731-32 (7th Cir. 2000) (reversing trial court decision in favor of employer on plaintiffs sexual harassment claim because trial court's analysis was incomplete on prong two, but concluding that there was "nothing reversible" in trial court's conclusion that employer had satisfied prong one of the affirmative defense by having an "established harassment policy" and "respond[ing] adequately" when plaintiff reported the harassment). 220 Columbia Journal of Gender and Law [Vol. 13:2 response to a report of harassment clearly falls into the category of correction, the employer's anti-harassment policy must fulfill the duty of prevention. Cases in which the plaintiff fails to invoke her employer's internal grievance procedure make clear the courts' assumption that having an anti- harassment policy satisfies the employer's prevention obligation.10 3 In these cases, the employer's obligation to correct the harassment is not at issue because the plaintiff did not report the harassment to her employer,1°4 or did so through means not sanctioned by the employer's policy. 0 5 Thus, the only 103 See, e.g., Leopold, 239 F.3d at 245 (affirming district court order granting summary judgment to employer on plaintiffs hostile work environment claim in case in which plaintiff failed to invoke employer's complaint procedure, noting that the employer "had in place an anti-harassment policy and an accompanying complaint procedure sufficient to satisfy the first element of the affirmative defense"); Riffle v. Sports Authority, No. 99- 2192, 2000 U.S. App. LEXIS 5236, at *3 (4th Cir. Mar. 28, 2000) (stating, in summary decision affirming trial court order granting summary judgment to employer in a case in which plaintiff apparently did not report harassment to her employer, that the employer "by virtue of its sexual harassment policy, established the first element" of the affirmative defense); Ritchie v. Stamler Corp., No. 98-5750, 2000 U.S. App. LEXIS 568, at *6-8 (6th Cir. Jan. 12, 2000) (affirming district court order granting summary judgment to employer on hostile work environment claim in case in which victim of harassment did not report harassment according to terms of employer's policy and concluding that employer had fulfilled prong one of the affirmative defense by "publish[ing] and promulgat[ing] an antiharassment policy with a procedure for reporting sexual harassment complaints," which was not ineffective simply because it required employees to report harassment in writing only to the firm's president); Good v. MMR Group, Inc., No. 3:00CV-182-H, 2001 U.S. Dist. LEXIS 20036, at *13-14 (W.D. Ky. Dec. 4, 2001) (granting employer's motion for summary judgment in co-worker harassment case in which court applied affirmative defense, finding that employer had satisfied prong one of the defense because plaintiff was aware of the employer's policy and procedure and had not reported the harassment to the person specified in the employer's policy, even though a foreman witnessed one incident of harassment and plaintiff's husband informed a company superintendent of the harassment); Ellenboeen v. Proiection Video Servs., No. 99 Civ. 11046, 2001 U.S. Dist. LEXIS 8852, at *37-38 (S.D.N.Y. June 29, 2001) (granting employer's motion for summary judgment on hostile work environment claim, concluding that employer fulfilled prong one of the affirmative defense by publishing in its Personnel Policies and Procedure Guide an anti- harassment policy and having in place a complaint procedure); cf Woodcock v. Montefiore Medical Center, No. 98-CV-4420, 2002 U.S. Dist. LEXIS 2965, at *22-23 (E.D.N.Y. Jan. 28, 2002) (granting employer's motion for summary judgment on plaintiff s claim, pursuant to § 1981, of a racially hostile work environment, noting that "[tlhe existence of Montefiore's anti-harassment policy is sufficient in this case to satisfy the first prong of the affirmative defense to employer liability"); Anderson v. Deluxe Homes of PA, Inc., 131 F. Supp. 2d 637, 649-51 (M.D. Pa. 2001) (denying employer's motion for summary judgment on plaintiffs supervisory harassment claim, in a case in which plaintiff failed to report harassing behavior, based on prong two of the affirmative defense, but finding that the employer had "satisfied its burden on the first element of the affirmative defense, as it indisputably prove[d] the existence of a sexual harassment policy with a complaint procedure"). 104 See, e.g., Leopold, 239 F.3d at 245; Riffle, 2000 U.S. App. LEXIS 5236, at *3. 'o5 See infra Part HI.B.2 for cases. Operating in an Empirical Vacuum discussion is of prevention, and the analysis is limited to an examination of what the employer's policy and procedure looks like on paper. There are several problems with equating an employer's creation and distribution of an anti-harassment policy with prevention of harassment. First, such an equation is arguably inconsistent with the language of Ellerth and Faragher. Second, by focusing on the employer's paper policy and procedures, the courts provide employers with little incentive to reduce and eliminate those conditions in the workplace that predict the incidence of sexual harassment. Finally, the framework devised by the courts for analyzing prevention under prong one of the affirmative defense impermissibly shifts the burden of proof from the employer to the employee on the question of whether the employer exercised reasonable care to prevent workplace harassment. 1. Ellerth and Farageher: Are a Policy and Procedure Sufficient? One explanation for why so many federal courts equate having an anti-harassment policy and complaint procedure with prevention of harassment lies with the Supreme Court's decisions in Ellerth and Faragher. The Court in Ellerth and Faragher stressed the importance of "antiharassment policies and effective grievance procedures," noting that if "employer liability [were] to depend in part on an employer's effort to create such procedures, it would effect Congress' intention to promote conciliation rather than litigation in the Title VII context."'' 0 6 In addition, in reaching its conclusion in Faragher that the City could not, as a matter of law, "be found to have exercised reasonable care to prevent the supervisors' harassing conduct," the Court relied almost exclusively on deficiencies related to the employer's anti-harassment policy.107 The Court noted that the City "had entirely failed to disseminate its policy against sexual harassment among the beach employees."' 0 8 The Court also pointed out that the employer's policy did not provide employees with a means to bypass a harassing supervisor in order to lodge a complaint of sexual harassment.'°9 Given the limited guidance provided by the Court on what constitutes prevention," 0 and the fact that the discussion of prevention in Faragher largely centered on the employer's anti-harassment policy, it is not entirely surprising that the lower federal courts have equated an anti- 106 Burlington Indus. v. Ellerth, 524 U.S. 742, 764 (1998). 107 Faragher v. City of Boca Raton, 524 U.S. 775, 808 (1998). 108 d. 9MId. See supra notes 67-68 and accompanying text. 11 See Lawton, supra note 16, at 108-10 (explaining the various ways in which the Supreme Court left unanswered important questions about the affirmative defense). 2004] Columbia Journal of Gender and Law harassment policy with the element of prevention in prong one of the affirmative defense. Nonetheless, the Court in Ellerth and Faragher never held that the mere existence and distribution of an anti-harassment policy sufficed to satisfy the employer's obligation to prevent workplace harassment. The Court spoke in terms of "effective" and "proven" employer mechanisms."' The words "proven" and "effective," by definition, require some examination of effects." Therefore, even assuming the Court intended that an employer could satisfy its prevention obligation by promulgating an "effective" and "proven" policy and procedure," 3 it is unlikely that the Court intended the lower federal courts to judge effectiveness based solely on the language of the employer's policy and grievance procedure. Moreover, the Court in Faragher did not hold that if an employer distributes an anti-harassment policy with a complaint procedure allowing the employee to bypass a harassing supervisor, then the employer has met its prevention obligations under prong one of the affirmative defense. Simply because the Court in Faragher enumerated two deficiencies in the employer's prevention efforts does not mean that the Court intended that an employer whose prevention efforts do not suffer from those same deficiencies necessarily satisfies the prevention element of prong one. While there is no empirical evidence on which provisions make a policy effective," 4 let us assume that an effective policy contains six salient provisions, including providing multiple avenues of complaint, and protection against retaliation; if any one provision is missing, the policy is considered ineffective. If the Court identified only one missing provision, e.g., providing multiple avenues of complaint, and concluded that the policy was ineffective, it would not necessarily follow that another policy having that provision but missing one or more of the other five, e.g., protection against retaliation, is effective. In other words, the lower federal courts have assumed that in Faragher the Supreme Court listed those provisions that are "' Ellerth, 524 U.S. at 764; Faragher, 524 U.S. at 806. 112 The problem, of course, is that there is no empirical data on what makes a policy effective. See supra note 72 and accompanying text. 113 The Supreme Court's opinion in Pennsylvania State Police v. Suders, 124 S.Ct. 2342 (2004), suggests that an employer is not obligated to do more than create an accessible and effective anti-harassment policy and grievance procedure in order to satisfy its obligation to prevent workplace harassment by supervisors. See id. at 2347 (stating that for those constructive discharge cases in which the employer may assert the affirmative defense, prong one is satisfied by proof that the employer "had installed a readily accessible and effective policy for reporting and resolving complaints of sexual harassment"). 14 See supra note 72 and accompanying text, and infra note 162 and accompanying text. 222 [Vol. 13:2 Operating in an Empirical Vacuum sufficient for an effective policy, even though the Court never expressly so held." 5 Admittedly, the Court's decisions in Ellerth and Faragher are not models of clarity on the parameters of the employer's obligation to prevent workplace harassment. There is language in Faragher, for example, that a court intent on narrowing the employer's responsibilities under the affirmative defense can use in order to conclude that an employer satisfies prong one by promulgating an anti-harassment policy and grievance procedure.1 6 Limiting the employer's obligations under prong one to the promulgation of a policy and procedure, however, allows the employer to ignore those factors in the workplace that the empirical literature shows increase the likelihood of harassment occurring in the workplace. 2. The Predictors of Sexual Harassment and Employer Incentives There are several predictors of sexual harassment in the workplace. They include the following: (1) an individual's propensity to harass; (2) the organization's culture; and (3) the job gender context in the workplace.1 7 a. Individual Propensity to Harass John Pryor has developed a scale entitled "Likelihood to Sexually Harass ("LSH")," which measures "the likelihood of [ ] performing acts of quid pro quo sexual harassment."'" 8 Pryor and his colleagues have found that men who score high on the LSH scale "tend to identify with 115 See Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1299 (1 1th Cir. 2000) (holding that the employer's grievance procedure met "the minimum requirements for the Faher defense because the procedures did not require that the employee complain to the offending supervisor or through the supervisor's chain of command and the procedures provided multiple avenues of lodging a complaint to assessable [sic], designated representatives"), cert. denied, 531 U.S. 926 (2000). 116 See Faragher, 524 U.S. at 806-07 (stating that if an employee unreasonably fails to use her employer's "proven, effective mechanism for reporting and resolving complaints of sexual harassment ... she should not recover damages that could have been avoided" if she had used the employer's procedure). 117 Cf Matthew S. Hesson-Mclnnis & Louise F. Fitzgerald, Sexual Harassment: A Preliminary Test of an Integrative Model, 27 J. Applied Soc. Psychol. 877, 897 (1997) (concluding that victim vulnerability, job gender context, and organizational tolerance for harassment predicted the incidence of harassment within a workplace). 118 John B. Pryor & Nora J. Whalen, A Typology of Sexual Harassment: Characteristics of Harassers and the Social Circumstances under which Sexual Harassment Occurs, in Sexual Harassment, supra note 73, at 129, 131. The LSH scale "has consistently demonstrated high reliability." Id. at 132. 2004] 223 224 Columbia Journal of Gender and Law [Vol. 13:2 hypermasculine stereotypes" and "to describe their own sexual behavior as motivated by a desire for dominance." 19 The fact that certain men have a propensity to harass suggests that there is little that employers can do in advance to deter harassment. Short of administering the LSH scale to prospective employees, how is an employer able to know who will or will not engage in workplace harassment? Pryor and his colleagues, however, have also found that high-LSH men take their cues from the workplace environment. 20 When a man who scores high on the LSH scale is exposed to a harassing role model, such as a manager, he is more likely to engage in sexual harassment himself.' 2 ' When "local norms condone professional behavior," however, men who score high on the LSH scale tend not to engage in harassing behavior. 22 Pryor and his colleagues hypothesize that the role model: influence[s] the perception of both descriptive and injunctive norms - everybody's doing it and it's okay to do it. So, it cannot be assumed that sexual exploitation is solely the result of deviant personality tendencies. Men who want to sexually exploit do so when the circumstances or local social norms permit such behavior. 123 Therefore, the organization's culture, not simply an individual's proclivity to harass, predicts the likelihood of harassment occurring. The same individual may feel "free" to harass in one organization, but feel constrained from doing so in another because management sends a clear message that sexual harassment in the workplace is not tolerated. 119 Id. at 132. See also John B. Pryor & Lynnette M. Stoller, Sexual Cognition Processes in Men High in the Likelihood to Sexually Harass, 20 Personality & Soc. Psychol. Bull. 163, 167 (1994) (concluding that "sexuality and dominance are closely related concepts in the minds of high-LSH men"). 120 See John B. Pryor et al., A Social Psychological Analysis of Sexual Harassment: The Person/Situation Interaction, 42 J. Vocational Behav. 68 (1993). 121 See id. at 76-77 (describing the function that the graduate student played in their experiment as "analogous to the function of managers" and concluding that exposure to a harassing manager increases the chances that a man who scores high on the LSH scale will also harass); see also Pryor & Whalen, supra note 118, at 133. 122 Pryor et al., supra note 120, at 77; see Pryor & Whalen, supra note 118, at 133. 123 Pryor & Whalen, supra note 118, at 133-34. 2004] Operating in an Empirical Vacuum 225 b. Organizational Culture The empirical research shows that the more tolerant an organization is of harassment, the more likely it is to occur.124 Researchers measure organizational tolerance based on employees' perceptions of organizational conduct, including (1) the organization's refusal to take complaints seriously, (2) the risk to the victim of reporting, and (3) the absence of meaningful sanctions. 125 Some might critique data based on employee perception, arguing that it is unreliable because it is "subjective" and may reflect the views of 124 See Louise F. Fitzgerald et al., Sexual Harassment in the Armed Forces: A Test of an Integrated Model, 11 Mil. Psychol. 329, 340 (1999) (citation omitted) (concluding that harassment is more common in the military, as in the civilian organizations studied, when "personnel perceive that such behavior is tolerated"); Louise F. Fitzgerald et al., Antecedents and Consequences of Sexual Harassment in Organizations: A Test of an Integrated Model, 82 J. Applied Psychol. 578,579-80,586 (1997) [hereinafter Antecedents] (study of more than 350 female employees at a "large, regulated West Coast utility," which found that "women employees who believe their organization is tolerant of sexual harassment ... experience considerably higher levels of harassment"); Hesson-McInnis & Fitzgerald, supra note 117, at 897 (stating that "[w]e are not the first researchers to demonstrate a relationship between organizational tolerance for sexually harassing behavior and the prevalence of such behavior"); Greetje Timmerman & Cristien Bajema, The Impact of Organizational Culture on Perceptions and Experiences of Sexual Harassment, 57 J. Vocational Behav. 188, 201 (2000) (finding that "unwanted sexual conduct is less a problem in organizational cultures that are perceived by employees as creating equal opportunities for women and men"); cf. Mia L. Cahill, The Social Construction of Sexual Harassment Law: The Role of the National, Organizational and Individual Context 44-45, 86 (2001) (finding, in study of five firms-one Austrian domestic company, two U.S. domestic firms, and two U.S. multinationals operating in Austria-that "the more likely it is that an individual believes something is not allowed under an organization's rules, the more likely they also believe it is illegal under national law, and the more likely they will support punitive organizational action"). 125 See Antecedents, supra note 124, at 580 (describing methodology to include employees' completion of the Organizational Tolerance for Sexual Harassment Inventory ("OTSHI"), which includes three subscales, "Risk of Reporting, Likelihood of Being Taken Seriously, and Probability of Sanctions," and "asks participants to report their perceptions of the likelihood of organizational reactions when superiors and coworkers engage in various forms of harassment"); Theresa M. Glomb et al., Ambient Sexual Harassment: An Integrated Model of Antecedents and Consequences, 71 Organizational Beh. & Hum. Decision Processes 309, 315-16 (1997) (explaining that participants in study completed OTSHI and describing the nature of the OTSHI measurement instrument); see also Hesson-McInnis & Fitzgerald, supra note 117, at 882 tbl.I (examining the following factors in evaluating organizational tolerance: "the presence or absence of 5 organizational remedies; victim's [sic] ratings of the effectiveness of these remedies; victim's [sic] subjective ratings of organization's willingness to investigate and punish harassers; and perceptions of the effectiveness of these actions"). Columbia Journal of Gender and Law those overly sensitive to the issue of sexual harassment-what Fitzgerald and her colleagues label the "whiner hypothesis.' 26 This research, however, is important for two reasons. First, if employees perceive their organization to be tolerant of sexual harassment because harassers are not punished or complainants suffer retaliation, then victims are less likely to come forward and use their employer's internal grievance procedure. Second, Fitzgerald and her colleagues, as well as Glomb and her colleagues, have cast doubt on the validity of the so-called "whiner hypothesis." In their study of female employees at a large West Coast utility, Fitzgerald and her colleagues found that "female employees whose coworkers believe that their organization is tolerant of sexual harassment - are significantly more likely to be sexually harassed."1 27 Glomb and her colleagues, in a study of 300 employees at a large utility company and a food processing firm, found that "employees' perceptions that the organization is tolerant of sexual harassment are positively related to their experiences of sexual harassment as well as to the ambient level of sexual harassment in their work group."'128 The researchers defined "ambient" sexual harassment as "the frequency of sexually harassing behaviors perpetrated against others in a woman's work group. 129 In other words, both studies examined not only individual women's experiences of sexual harassment, but also co-workers' perceptions of the organization's tolerance for harassing behavior. The researchers found that women reported experiencing more harassment when their co-workers perceived the organization as more tolerant of harassing behavior. 30 Thus, the research suggests that employee perceptions are an accurate barometer of organizational tolerance for harassing conduct. c. Job Gender Context Job gender context is comprised of two elements: (1) the male to female ratio in the workgroup, and (2) the gendered nature of the work performed. 13' Gendered jobs are those strongly identified with one sex, for example, police work. 126 See Antecedents, supra note 124, at 586-87; see also Glomb et al., supra note 125, at 322-23. 127 Antecedents, supra note 124, at 587. 128 Glomb et al., supra note 125, at 321. 1'29 Id. at 314. 130 See Antecedents, supra note 124, at 586-87; Glomb et al., supra note 125, at 322-23; see also Frazier, supra note 72, at B-8 to B-9 (describing studies). 131 See Antecedents, supra note 124, at 579; Hesson-Mclnnis & Fitzgerald, supra note 117, at 882 tbl.1; see also James E. Gruber, The Impact of Male Work Environments [Vol. 13:2 2004] Operating in an Empirical Vacuum 227 Sufficient evidence exists to suggest the gender composition of an individual's work group will affect the frequency with which [sexual harassment] occurs. Individuals in gender-imbalanced work environments (particularly women in male-dominated work environments) are more likely to experience unwanted sexual attention at work. . . . Alternatively, sexually integrated workplaces exhibit lower levels of [sexual harassment] than workplaces dominated by either gender. 32 Fitzgerald and her colleagues found that "a male dominated workplace [is] significantly implicated in high levels of sexual harassment."' 3 3 They concluded that "it is not skewed gender ratio per se, but rather the presence of large numbers of male workers, combined with traditionally male- oriented tasks" that increases the risk of workplace harassment. 34 Gruber found that women are most likely to be the targets of harassment "when their visibility in the work environment centers on gender-based status differences, that is, they have a nontraditional job in a male-dominated work environment ... or their traditionally female job is located in a work environment that involves a high degree of contact with men.' Therefore, the research indicates that women holding jobs in fields dominated by men, construction or police work, for example, are more likely to encounter harassing behavior because they work in male-defined jobs in a male- dominated work environment. 36 and Organizational Policies on Women's Experiences of Sexual Harassment, 12 Gender & Soc'y 301, 309-10 (1998) (evaluating impact of occupational sex composition, contact with men in the workplace, and gender predominance on the incidence of workplace harassment). 132 Knapp et al., supra note 72, at 706-07 (internal citations omitted); see Lawton, supra note 16, at 122-23 (summarizing studies). 133 Antecedents, supra note 124, at 586; see Hesson-Mclnnis & Fitzgerald, supra note 117, at 889-90. 134 Antecedents, supra note 124, at 586. 135 Gruber, supra note 131, at 311 (percentages omitted). 136 See, e.g., Caridad v. Metro-N. Commuter R.R., 191 F.3d 283, 290 (2d Cir. 1999) (only woman in thirteen-person crew working the shift as an electrician), cert. denied, Metro N. Commuter R.R. v. Norris, 529 U.S. 1107 (2000); Mikels v. City of Durham, 183 F.3d 323, 326 (4th Cir. 1999) (one of five female officers on police squad); Hurley v. Atlantic City Police Dep't, 174 F.3d 95, 103 (3d Cir. 1999) (explaining that plaintiff was both the first woman to graduate from the Atlantic City Police Academy and the first woman to achieve the rank of sergeant at the Atlantic City Police Department), cert. denied, 528 U.S. 1074 (2000); Richmond-Hopes v. City of Cleveland, No. 97-3595, 1998 U.S. App. LEXIS 29572, at *2 (6th Cir. Nov. 16, 1998) (noting that plaintiff, who was an electric meter service installer for Cleveland Public Power, was "only one of two women working in non- secretarial positions in the department"); Brissette v. Franklin County Sheriff's Office, 235 F. Supp. 2d 63, 68-69 (D. Mass. 2003) (describing long history of sex discrimination at sheriffs office, noting that at the time of the lawsuit no women occupied "rank in line" positions at the sheriff's office); Good v. MMR Group, Inc., No. 3:00CV-182-H, 2001 U.S. Dist. LEXIS 20036, at *2-3 (W.D. Ky. Dec. 4, 2001) (sole woman in the work crew at Columbia Journal of Gender and Law d. What the Research Tells Us About Structuring Employer Incentives What the research shows is that an organization's culture and the job gender context predict the likelihood of sexually harassing behavior in the workplace. While some men have a greater propensity to harass than others, they do so only when local norms, as communicated by workplace personnel such as supervisors, condone harassing behavior. The question then is: how are sexual harassment policies and procedures related to the predictors of harassment? In other words, do policies and grievance mechanisms operate to deter harassing behavior in the workplace? In his study of the workplace experiences of almost 2,000 Canadian women, based on telephone interviews conducted in 1992, Gruber found that anti-harassment policies and procedures do have an impact on men's behavior in the workplace. 137 But it is not the mere promulgation of an anti- harassment policy and grievance procedure that deters harassment. 138 Gruber based his findings on women's awareness of, as opposed to company documents detailing, the employer's policies and procedures.' 39 Therefore, he hypothesized that when an organization makes a "concerted" and "highly visible effort" to address sexual harassment, women will be more aware of the employer's policies and procedures."40 For Gruber, women's awareness of their employers' policies and procedures served as a proxy for the organization's commitment to reducing workplace harassment.1 41 Current theory and research suggests that employee perceptions of organizational tolerance of sexual harassment (Hulin, Fitzgerald, and Drasgow 1996), their beliefs about leaders' stances on the problem (Pryor, Giedd, and Williams 1995), or their concerns about procedural justice (Rudman, Borgida, and employer's construction site); Truax v. City of Portsmouth, No. 00-63-B, 2001 U.S. Dist. LEXIS 9190, at *3 (D.N.H. June 18, 2001) (first two female police officers in the department). 137 See Gruber, supra note 131, at 316. 138 Research in a related field-the impact of formal hiring and promotion rules on gender discrimination-indicates that "the mere existence of a specialized personnel function and formalized hiring and promotion procedures does not ensure that those procedures will be followed. These procedures may constrain decision makers only when the personnel function is powerful enough to reward those who comply and sanction those who do not." Elissa L. Perry et al., Explaining Gender-Based Selection Decisions: A Synthesis of Contextual and Cognitive Approaches, 19 Acad. Mgmt. Rev. 786,805 (1994). 139 Gruber, supra note 131, at 316. 14o Id. 141 See id. 228 [Vol. 13:2 Operating in an Empirical Vacuum Robertson 1995) rather than the mere objective existence of formal rules and regulations strongly and routinely influence their attitudes and behaviors on matters of sex discrimination and harassment. 142 Williams, Fitzgerald, and Drasgow also found that it is not the mere promulgation of policies and procedures, but rather the organization's efforts to implement anti-harassment strategies that affected the incidence of sexual harassment in the workplace. 143 The researchers based their findings on the results of the 1995-Form B-Gender Issues survey completed by more than 28,000 American armed services personnel. 144 They divided organizational practices into three groups: (1) implementation, (2) education, and (3) resources.1 45 Implementation practices included prevention, which the researchers defined to include "promoting professional work environments, integrating women into jobs where they are traditionally underrepresented, examining the effectiveness of current sexual harassment programs, and evaluating supervisory work performance in part by enforcement of the sexual harassment program." 14 6 The researchers defined education to encompass training, and resources to include advice and support services related to sexual harassment.1 4 7 Williams and her colleagues found that the three groups of organizational practices-implementation, education, and resources- "differentially affected the incidence of harassment in the military, with implementation practices having the greatest effect and resources and training having the least."' 48 In fact, of the three types of organizational practices identified, "only the Implementation Practices scale significantly contributed to the prediction of the incidence of harassment for both sexes. . . [so that] [a]s the rate of reported organizational implementation increased, the frequency of sexual harassment decreased in a linear fashion.' ' 149 The researchers also concluded that training had "little effect on ... the frequency of harassment."' 50 142 id. 143 See Jill Hunter Williams et al, The Effects of Organizational Practices on Sexual Harassment and Individual Outcomes in the Military, 11 Military Psychol. 303, 322 (1999). 144 Id. at 308-09. 141 Id. at 306, tbl.1. '46 Id. (citations omitted). 147 See id. 141 Id. at 315. 149Id. at 322. 150 Id. at 325. 2004] 229 230 Columbia Journal of Gender and Law [Vol. 13:2 Notwithstanding the empirical evidence, the lower federal courts routinely look no further than the employer's policy and procedure on paper when evaluating the employer's preventative efforts,15 ' often concluding, with no evidentiary support other than the words of the policy or the description of the grievance procedure, that the policy and procedure are "reasonable,"'' 52 "effective,' ' 153 or "designed to deter harassment."'' 54 Allowing victims to bypass a harassing supervisor, assurances of confidentiality, prohibitions against retaliation, and descriptions of potential 151 See supra notes 101-105 and accompanying text for cases. 152 See Zelava v. E. & W. Hotel Corp., No. 99-16179, 2001 U.S. App. LEXIS 3492, at *4 (9th Cir. Mar. 5, 2001) (affirming district court order granting employer's motion for summary judgment and concluding in one sentence that the employer had satisfied the prevention element of prong one by "posting a reasonable anti-harassment policy in Spanish and in English in multiple locations accessible to housekeeping staff"); Scrivner v. Socorro Indep. Sch. Dist., 169 F.3d 969, 971 (5th Cir. 1999) (affirming trial court order granting summary judgment to employer and concluding that the employer's policy and response to the harassment was "both reasonable and vigorous," without offering any evidence, not even the language of the employer's policy or procedure, in support of this conclusion). 153 See, e.g., Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *5 (4th Cir. Nov. 13, 2001) (affirming trial court order granting summary judgment to employer, "agree[ing] with the district court that [the employer's] policy was proven and effective because the policy provided... employees with many avenues through which to report harassment and [the employer] promptly investigated the complaints against [the harasser]"); Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1298-99 (lth Cir. 2000) (affirming trial court order granting employer's motion for summary judgment, explaining that the Supreme Court had "provide[d] some guidance regarding the minimum requirement for an anti-harassment policy's complaint procedures to be considered effective" and the employer had met that minimum because its grievance procedure afforded employees multiple avenues for filing a sexual harassment complaint), cert. denied, 531 U.S. 926 (2000); Smith v. Am. Int'l Group, Inc., No. 01 Civ 1652, 2002 U.S. Dist. LEXIS 7472, at *20-21 (S.D.N.Y. Apr. 26, 2002) (recommending grant of defendant's motion for summary judgment and concluding that the employer had "establish[ed] an effective sexual harassment policy" because it described prohibited conduct, created two avenues for filing complaints, and ensured confidentiality and protection against retaliation), adopted by No. 01 Civ. 1652, 2002 U.S. Dist. LEXIS 8954 (S.D.N.Y. May 21, 2002); cf Citroner v. Progressive Cas. Ins. Co., 208 F. Supp. 2d 328, 336-37, 341 (E.D.N.Y. 2002) (granting employer's motion for summary judgment on plaintiffs claim of racial harassment and concluding, after reviewing the text of employer's anti-harassment policy, that employer's "Code of Conduct and Open Door Policy . .. provided an effective mechanism for employees to complain about harassment, confidentially if necessary"). 154 See Brown v. Perry, 184 F.3d 388, 396 (4th Cir. 1999) (affirming district court order granting summary judgment to employer on plaintiff's sexual harassment claim, concluding that employer had "an anti-harassment policy (including a complaint procedure) designed to deter harassment," but offering no evidence of how the policy's design actually deterred harassment); Lewis v. Forest Pharm., Inc., 217 F. Supp. 2d 638, 649 n.5 (D. Md. 2002) (noting, in dicta, that had plaintiff not suffered a tangible employment action, the employer would have prevailed on the affirmative defense, and concluding that the employer satisfied prong one of the defense because its anti-harassment policy was "designed to deter sexual harassment," even though no evidence of how the policy's design did, in fact, deter harassment was provided). 20041 Operating in an Empirical Vacuum sanctions are frequently cited as evidence that the employer's policy is reasonable, viable, or effective. 155 Yet there is no empirical evidence to show that the mere inclusion of such provisions in a policy automatically decreases the incidence of sexual harassment on the job. 156 For example, promises of confidentiality may increase reporting, but what effect does confidentiality have on the incidence, as opposed to the reporting, of workplace harassment? We simply do not know. How, then, can a federal court determine that a policy which promises confidentiality ensures that the employer is doing its part to deter workplace harassment? The story is similar with training. Providing some sort of sexual harassment training, while not required, is viewed favorably by the courts as part of an employer's prevention efforts and helps an employer prevail on a motion for summary judgment or judgment as a matter of law. 157 Those 155 The Second Circuit, however, has held that "guarantee[s] [of] confidentiality and non-retaliation" are not "mandatory" features of an anti-harassment policy. See Leopold v. Baccarat, Inc., 239 F.3d 243, 245 (2d Cir. 2001). 156 See supra note 72 and accompanying text. 157 See, e.g., Idusuvi v. Tenn. Dep't of Children's Servs., No. 00-6324, 2002 U.S. App. LEXIS 2463, at *13 (6th Cir. Feb. 11, 2002) (affirming trial court order granting judgment as a matter of law to defendant employer, concluding that employer had a "zero- tolerance" policy on sexual harassment, which was passed out to employees by e-mail and posted at job sites, and plaintiff attended a two-hour training session on sexual harassment); Watkins v. Prof'1 Sec. Bureau, Ltd., No. 98-2555, 1999 U.S. App. LEXIS 29841, at *7, 14 (4th Cir. Nov. 15, 1999) (affirming trial court order granting judgment as a matter of law to defendant employer, noting that the employer had a "comprehensive policy against sexual harassment," which included providing employees with "training about the policy as part of the orientation process"), cert. denied, 529 U.S. 1108 (2000); Shaw v. AutoZone, Inc., 180 F.3d 806, 809, 811-12 (7th Cir. 1999) (affirming trial court order granting summary judgment to employer and concluding that employer exercised reasonable care to prevent harassment because it had disseminated a detailed policy with multiple avenues to file a complaint and the harasser had "attended approximately twenty management meetings which included training or conversations on [the employer's] sexual harassment policies and procedures and the handling of sexual harassment claims"), cert. denied, 528 U.S. 1076 (2000); Samedi v. Miami-Dade County, 206 F. Supp. 2d 1213, 1220 (S.D. Fla 2002) (granting employer's second motion for summary judgment, noting that the employer's policy was "reasonable on its face" and that the employer provided sexual harassment training to supervisors, had a separate office for investigating sexual harassment complaints, and posted its policy and grievance procedure on "a public bulletin board"); Walton v. Johnson & Johnson Servs., Inc., 203 F. Supp. 2d 1312, 1322 (M.D. Fla. 2002) (granting employer's motion for summary judgment, explaining that the employer had exercised reasonable care to prevent harassment because it had a sexual harassment policy, notified employees of the policy through an annual letter sent to employees, and provided sexual harassment training to supervisors, even though plaintiff claimed that "her supervisor [had] not attend[ed] training until after the alleged sexual harassment"), aff'd, 347 F.3d 1272 (11 th Cir. 2003); Roelen v. Akron Beacon Journal, 199 F. Supp. 2d 685, 693-94 (N.D. Ohio 2002) (granting employer motion for summary judgment, noting that the employer newspaper had policies and procedures for dealing with harassment and had conducted a one-time mandatory training session on sexual harassment conducted by the state Civil Rights Commission, which plaintiff had attended); Church v. Maryland, 180 F. Supp. 2d 708, 737 (D. Md. 2002) (granting employer's motion for summary judgment, concluding that having a Columbia Journal of Gender and Law [Vol. 13:2 courts that mention training in their opinions, however, never require the employer to demonstrate that its training program actually works to deter harassment, or even that the employer has tried to measure the effectiveness of its own training efforts. The MSPB identified a similar lack of evaluation of training programs by federal agencies offering such training.1 58 When we asked Federal agencies whether they had conducted any evaluations of the content, effectiveness, or quality of the sexual harassment training they offer employees, the responses were similar from agency to agency. Most of the evaluations consist of participant critiques completed at the end of the training. Course evaluations such as these are very common, but are of limited use. As noted in the Board's recent report on training in the Federal Government, "[W]hile this sort of evaluation probably says something about how participants felt immediately after having received the training, it most likely says very little about what the participants took back to the job." None of the agencies described evaluations that examined whether the training had any effect on particular problems that had been identified in their organizations. Without more meaningful evaluations, agencies won't really know how much difference training makes, and what content changes might further curtail sexual harassment.159 policy and "adequate complaint procedure," determined based on the language of the policy and procedure, coupled with "training on the policy and complaint procedure for all incoming officers at the correctional training academy, and periodic training for supervisors (denoted as lieutenants and above) on the sexual harassment policy and procedures by the Department's EEO office demonstrate that the State exercised reasonable care to prevent sexual harassment"), afftd, No. 02-1179, 2002 U.S. App. LEXIS 25909 (4th Cir. Dec. 17, 2002); Carl v. Parmely, 188 F. Supp. 2d 991, 1005 (S.D. Ill. 2001) (granting employer's motion for summary judgment, concluding that there was "no evidence that the [employer] failed to exercise reasonable care to prevent sexually harassing behavior" because it had a sexual harassment policy, which it had posted, and within two weeks of the first complaint of harassment met with employees to remind them of the policy and a month later held a training session on the policy); Pfau v. Reed, No. A-95-CA-135 AA (W.D. Tex. June 16, 2000) (unpublished opinion on file with author) (granting employer's motion for summary judgment and concluding that employer satisfied prong one of the affirmative defense by having "an accessible grievance procedure," "a specific policy against sexual harassment and ... sexual harassment training [for] its employees"); cf Stricker v. Cessford Constr. Co., 179 F. Supp. 2d 987, 1008-10 (N.D. Iowa 2001) (denying both plaintiffs and defendant's motions for summary judgment on the affirmative defense, noting that defendant had raised a genuine issue of material fact as to whether it exercised reasonable care to prevent harassment based on its annual training sessions on sexual harassment conducted by "'outside' specialists," even though the employer's "written policies [were] woefully inadequate"). '5 8 See 1995 MSPB, supra note 74, at 42-43. 159 Id. at 43-44 (footnote omitted). 2004] Operating in an Empirical Vacuum 233 It appears that employers and courts alike simply assume that training works. Yet there is scant empirical data on the consequences of training' 60 and no research on the longitudinal effects of training on workplace behavior.' 6 For example, if a male employee views a video on sexual harassment during his job orientation, does that exposure have any long- term effects on his behavior one month, six months, or even a year later? The courts cannot answer that question because there is no data on it. Two things are clear. First, the empirical evidence does not, at this time, support the assumption by federal court judges that distributing a sexual harassment policy and grievance procedure, even when accompanied by training, deters sexual harassment in the workplace. In general, the current set of policies, procedures, and training used in organizations has not solved the problem of sexual harassment. Nor is there any evidence that they have or will lead most targets of harassment to come forward or that they are greatly inhibiting would-be sexual harassers. While they may be having a positive impact and are probably leading more victims to come forward, such effects have not been documented so far.1 62 Second, the courts are asking employers the wrong questions. Why are employers not required to show evidence of their efforts to evaluate and assess the impact of their organization's culture on the incidence of workplace harassment? If role models, such as supervisors, set the tone in the workplace, 163 as Pryor's research suggests, then should the employer not 160 Elizabeth O'Hare Grundmann et al., The Prevention of Sexual Harassment, in Sexual Harassment, supra note 73, at 175, 182 (noting that "various institutions and individuals offer diverse prevention programs but seem much less concerned with systematically evaluating these programs"); Knapp, supra note 72, at 716 ("very little research has been reported that systematically evaluates the effectiveness of the myriad [sexual harassment]-related training programs developed"); Robert S. Moyer & Anjan Nath, Some Effects of Brief Training Interventions on Perceptions of Sexual Harassment, 28 J. Applied Soc. Psychol. 333, 334 (stating that "the unpleasant empirical truth is that almost nothing is known about the effects of sexual harassment education and training programs"). 161 See Grundmann et al., supra note 160, at 182 (stating that "despite an extensive literature search," the authors "could find no published outcome study on the effectiveness of a sexual harassment prevention program"); Pryor & Whalen, supra note 118, at 145-46 (explaining that the existing studies of training test only "knowledge, opinions, and perceptions of workshop participants" and that to their knowledge, there were no studies of "how educational efforts have affected the incidents of sexual harassment in an organization"). 162 Barbara A. Gutek, Sexual Harassment Policy Initiatives, in Sexual Harassment, supra note 73, at 185, 196. 163 In some cases, it is clear that management has set a tone in which harassing behavior is tolerated. See Marquis v. Tecumseh Prods. Co., 206 F.R.D. 132, 174-79 (E.D. Mich. 2002) (describing pervasive harassment by supervisors and co-workers, including allegations against human resources director Ike Growe, who was responsible for 234 Columbia Journal of Gender and Law [Vol. 13:2 evaluate its supervisors, in part, on their efforts to enforce and comply with the employer's anti-harassment policy?' 64 Why are employers with highly sex-segregated workplaces not required to demonstrate their efforts to decrease sex segregation within jobs? Noting the connection between high levels of harassment and low numbers of women in the workplace, Hesson-Mclnnis and Fitzgerald suggest "that strong affirmative action programs and moving (large numbers of) women into jobs traditionally held by men may prove a reasonable organizational strategy for reducing harassment.' ' 65 If harassment is more likely to occur in "male bastions," then should employers in those fields not be held to a higher standard of prevention? [I]t is not unforeseeable that in some work environments certain employees will engage in sexual harassment, and it is even more predictable in a male bastion such as a police department. Indeed, the need for a vigilant and responsive approach to sexual harassment ought to be more obvious to managers in such investigating some of the plaintiffs' complaints of harassment); Gawley v. Ind. Univ., 276 F.3d 301, 306 (7th Cir. 2001) (noting that university police department "had a history of sexual harassment of its female employees dating back to the early 1980s" and that "[o]ther female employees [besides the plaintiff had] related a number of incidents over the years involving [plaintiff's harasser] and other male employees"); Katt v. City of New York, 151 F. Supp. 2d 313, 320-21 (S.D.N.Y. 2001) (describing work atmosphere as "rowdy," explaining that plaintiff experienced a "lot of sexual innuendos, sexual comments, questions, [and] intrusive questions" about her personal life and sexual habits, and noting that the "chief perpetrator" of the harassing behavior was Lieutenant DiPalma, who was "generally in charge of all employees at the Seventh Precinct during [plaintiff] Katt's shift"), affd, Krohn v. N.Y. City Police Dep't, Nos. 01-7827, 01-7875, 01-9023, 2003 U.S. App. LEXIS 6409 (2d Cir. Apr. 2, 2003); Dinkins v. Charoen Pokphand USA, Inc., 133 F. Supp. 2d 1254, 1258-60, 1269 (M.D. Ala. 2001) (describing workplace as "a seething cauldron of bigotry, misogyny and transferred self-hate," in which harassment was pervasive, and male supervisors and co-workers alike engaged in the conduct). 164 See Seldomridge v. Uni-Marts, Inc., No. 99-496, 2001 U.S. Dist. LEXIS 9491, at *9 (D. Del. July 10, 2001) (stating that prior to exposing himself at work to plaintiff, supervisor had received an evaluation "plac[ing] him in the top category of excellent," but noting that the evaluation did not have any "evaluation criteria based on anti-discrimination or harassment prevention"). 165 Hesson-Mclnnis & Fitzgerald, supra note 117, at 898; cf Audrey J. Murrell et al., Sexual Harassment and Gender Discrimination: A Longitudinal Study of Women Managers, 51 J. Soc. Issues 139, 146 (1995) (citation omitted) (concluding in study of more than 250 female MBAs surveyed over seven-year time period that "people do not routinely see the connection between gender-related discrimination in the workplace and sexual harassment"); Barbara A. Gutek & Mary P. Koss, Changed Women and Changed Organizations: Consequences of and Coping with Sexual Harassment, 42 J. Vocational Behav. 28, 32 (1993) (citing two studies for proposition that "[s]exual harassment and sex discrimination appear to go together"). Operating in an Empirical Vacuum environments than it might be to managers of more refined or gender-integrated workforces. 166 Finally, if employers depend on their anti-harassment policies and procedures to deter workplace harassment, should they not be required to evaluate those policies and procedures? Why do federal courts not require an employer to ask its own employees about their perceptions of the employer's procedures? If employees believe the employer does not punish harassers and penalizes victims who come forward with complaints of sexual harassment, why should an employer be allowed to prevail on the affirmative defense for simply drafting a policy and procedure that workers are reluctant to use? In conclusion, equating prevention with paper compliance provides employers with few incentives to address the predictors of workplace harassment. In addition, by equating prevention with the dissemination of a policy and procedure, the federal courts lower the bar for employers and impermissibly shift the burden of proof from the employer to the employee on the reasonableness of the employer's efforts to combat harassment in the workplace. 3. Shifting the Burden of Proof on Prong One In both Ellerth and Faragher, the Supreme Court clearly stated that the employer has the burden to prove, by a preponderance of the evidence, both prongs of the affirmative defense. 167 Yet, the vast majority of federal courts have shifted the burden of proof on prevention to the plaintiff.168 The Fourth Circuit has done so explicitly. The process of rewriting Ellerth and Faragher in the Fourth Circuit began with Lissau v. Southern Food Service. 169 In Lissau, the Fourth Circuit 166 Mancuso v. City of Atlantic City, 193 F. Supp. 2d 789, 802 (D.N.J. 2002) (quoting Hurley v. Atlantic City Police Dep't, 933 F. Supp. 396, 422 (D.N.J. 1996)); cf Katt, 151 F. Supp. 2d at 325-27 (discussing expert testimony offered at trial to explain plaintiff's failure to use employer's grievance procedure and noting that expert had testified that "the paramilitary structure of police departments and the general stress of police work create a culture in which police officers so value loyalty and mutual support that complaints or truthful testimony about misconduct by officers is strongly disfavored and met with ostracism or other retaliation"). 167 Burlington Indus. v. Ellerth, 524 U.S. 742, 765 (1998); Faragher v. City of Boca Raton, 524 U.S. 775, 807 (1998). 168 See infra note 181 and accompanying text. 169 159 F.3d 177 (1998). Judge Michael, who concurred in the judgment in Lissau, dissented from the majority's discussion of employer liability in Part III of the opinion. See id. at 183 (Michael, J., concurring in part and dissenting in part). He argued that it was "premature" of the majority "to suggest ways in which summary judgment might once again be granted to Southern, the employer." Id. He also believed that the majority was "ill-advised to encourage the district court to consider, on summary judgment, whether [the employer] 2004] 236 Columbia Journal of Gender and Law [Vol. 13:2 reversed the district court's order granting summary judgment to the employer on plaintiffs claim of supervisory harassment and remanded the case to the district court for reconsideration based on the Supreme Court's decisions in Ellerth and Faragher.170 In its remand opinion, the Court of Appeals cited to the Supreme Court's decisions in Ellerth, Faragher, and Meritor, stating that evidence by an employer that it "had disseminated an effective anti-harassment policy provides compelling proof of its efforts to prevent workplace harassment." '1 7' This "compelling proof' language, which does not appear in Ellerth, Faragher, or Meritor,172 created a presumption that an employer satisfies prong one of the affirmative defense by having an "effective anti- harassment policy." This presumption might not work to the detriment of Title VII plaintiffs if the Fourth Circuit required employers to provide evidence that their anti-harassment policies are effective. In practice, however, courts in the Fourth Circuit rarely require the employer to offer any proof, beyond the words of its policy, to demonstrate that the policy has any deterrent effect on workplace harassment. 73 has established the new Faragher-Ellerth affirmative defense . . . because the defense is especially fact intensive." Id. at 183-84. ' 7 °1d. at 183. 171 Id. at 182 (emphasis added). 172 The Supreme Court in Ellerth, Faracher, and Meritor never said that an employer's anti-harassment policy provided compelling proof of the employer's efforts to prevent harassment. In fact, the word "compelling" is not used anywhere in the Court's decisions in Ellerth, Faragher, or Meritor. But, in its recent decision in Pennsylvania State Police v. Suders, 124 S.Ct. 2342 (2004), the Court does suggest that an employer only needs to provide an accessible and effective anti-harassment policy in order to satisfy its prevention obligation. Compare id. at 2347 (stating that for those constructive discharge cases in which the employer may assert the affirmative defense that prong one is satisfied by proof that the employer "had installed a readily accessible and effective policy for reporting and resolving complaints of sexual harassment") with Ellerth, 524 U.S. at 765 (stating that while a sexual harassment policy and complaint procedure were "not necessary in every instance as a matter of law, the need for a stated policy suitable to the employment circumstances may appropriately be addressed in any case when litigating the first element of the defense"). 173 See, e.g., Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *4 (4th Cir. Nov. 13, 2001) (affirming district court order granting employer's motion for summary judgment on the basis of the affirmative defense, explaining that the employer's distribution of its anti-harassment policy was compelling proof that the employer had satisfied prong one of the affirmative defense); Matvia v. Bald Head Island Mgmt., Inc., 259 F.3d 261, 268 (4th Cir. 2001) (affirming district court order entering summary judgment for employer on plaintiffs sexual harassment claim and rejecting plaintiffs argument that employer's anti-harassment policy did not effectively prevent harassment, noting that dissemination of a policy was compelling proof of reasonable care and that plaintiff failed to show how the policy, as written, was "ambiguous or difficult to follow ... [or] could have been made any clearer"); Barrett v. Applied Radiant Energy Corp., 240 F.3d 262, 266 (4th Cir. 2001) (affirming district court decision reversing a jury verdict for plaintiff and granting defendant employer's motion for judgment as a matter of law, rejecting as "incompatible with the law of [the] circuit" plaintiff's argument that the employer could not show that it Operating in an Empirical Vacuum In Brown v. Perry, the Fourth Circuit used the presumption created in Lissau to then explicitly shift the burden of proof on prevention from the defendant employer to the plaintiff employee. 74 In Brown, the Fourth Circuit affirmed a trial court decision granting the employer's motion for summary judgment." 5 In its discussion of the prevention element of prong one of the affirmative defense, the court noted that the employer had promulgated "an anti-harassment policy (including a complaint procedure) designed to deter harassment."'' 76 It concluded, citing to both Faragher and Lissau, that when: there is no evidence that an employer adopted or administered an anti-harassment policy in bad faith or that the policy was otherwise defective or dysfunctional, the existence of such a policy militates strongly in favor of a conclusion that the employer "exercised reasonable care to prevent" and promptly correct sexual harassment. 77 This language from Brown not only does not appear in Faragher, it runs counter to the Supreme Court's allocation of the burden of proof to the employer. When would an employer ever introduce evidence that its policy was adopted in bad faith, or was defective or dysfunctional? This is evidence that the plaintiff would provide, but the plaintiff does not bear the burden of proof; that burden is the defendant's. Of course, if the defendant employer comes forward with evidence that its policy and complaint procedures are effective and work in practice, then it makes sense to offer the plaintiff the opportunity to offer evidence in rebuttal. The court in Brown, however, concluded that the employer's policy was "designed to deter harassment" without ever delineating the evidence on which that conclusion was based. The court did not even specify which aspects of the policy and complaint procedure were designed to deter harassment. Given the fact that social science researchers do not know which specific provisions in sexual harassment policies and procedures actually deter harassment, 178 how is it that the Fourth Circuit knows when an anti-harassment policy is designed to do so? By combining the holdings in Lissau and Brown, courts in the Fourth Circuit now routinely shift the burden of proof on prong one from exercised reasonable care to prevent and correct harassment because the employer did nothing more than distribute its anti-harassment policy). 174 184 F.3d 388 (1998). 175 id. 1 76 Id. at 396. 177 Id. 178 See supra note 72 and accompanying text. 2004] 237 Columbia Journal of Gender and Law [Vol. 13:2 the defendant employer to the plaintiff employee. 179 It would be a mistake, however, to assume that only the Fourth Circuit is engaged in shifting the burden of proof on prong one from employers to employees. The vast majority of federal courts are doing so.' 179 See, e.g., Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *4-6 (4th Cir. Nov. 13, 2001) (affirming district court order granting summary judgment for employer, rejecting plaintiffs claim that distribution of an anti-harassment policy did not suffice to satisfy employer's obligation under prong one and finding no evidence that the policy was adopted or administered in bad faith, or was defective or dysfunctional); Matvia v. Bald Head Island Mgmt., Inc., 259 F.3d 261, 268 (4th Cir. 2001) (affirming trial court's order granting employer's motion for summary judgment, citing to both Lissau and Brown and concluding that plaintiff had failed to rebut the employer's proof, based solely on the employer's policy, of an "effective preventative program"); Barrett v. Applied Radiant Energy Cor., 240 F.3d 262, 266 (4th Cir. 2001) (affirming trial court order granting judgment as a matter of law to employer, citing to both Lissau and Brown and holding that the plaintiff had failed to provide evidence that the employer's policy was adopted in bad faith, and deciding, based on a review of the language of the policy, that there was no evidence that the policy was defective or dysfunctional); Murray v. City of Winston- Salem, 203 F. Supp. 2d 493, 500-01 (M.D.N.C. 2002) (granting employer's motion for summary judgment, concluding that even if plaintiff could establish the existence of a hostile work environment, the employer still prevailed on summary judgment on the basis of the affirmative defense, in part, because the employer had "disseminated an effective anti- harassment policy," and the plaintiff had "presented no evidence that [her employer] adopted or administered its sexual harassment policy in bad faith"); Green v. Wills Group, Inc., 161 F. Supp. 2d 618, 625 (D. Md. 2001) (granting employer's motion for summary judgment, stating that Fourth Circuit "refinements" on the Supreme Court's decisions in Ellerth and Faragher include recognition that an employer's "distribution of an anti-harassment policy provides 'compelling proof that the defendant exercised reasonable care in preventing and promptly correcting sexual harassment unless the plaintiff shows that the employer adopted or administered the policy in bad faith or that the policy was otherwise defective or dysfunctional"). 180 Some federal courts simply make it clearer that it is the plaintiff employee, not the defendant employer, who bears the burden of proof on prong one. See, e.g., Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1299 (11 th Cir. 2000) (affirming trial court order granting employer motion for summary judgment and citing to the Fourth Circuit's decision in Brown v. Perry for the proposition that because there was no defect in the employer's complaint procedures and no evidence that its policy was administered in bad faith, the employer had exercised reasonable care to prevent harassment), cert. denied, 531 U.S. 926 (2000); Breeding v. Cendant Cor., No. 01 Civ. 11563, 2003 U.S. Dist. LEXIS 6558, at *18- 19 (S.D.N.Y. Apr. 17, 2003) (granting employer's motion for summary judgment and stating that once employer presented evidence of a policy that guaranteed confidentiality, "promised careful investigation, followed by appropriate remedial action," and prohibited retaliation, the employer had shown "reasonable care as a matter of law," which the employee could then rebut "by presenting evidence that the policy is ineffective, that she had no notice of the policy, or that defendant failed to follow its own procedures"); Arnold v. Yale New Haven Hosp., 213 F. Supp. 2d 142, 148 (D. Conn. 2002) (granting motion of employer for summary judgment, and while stating that having a sexual harassment policy and grievance procedure is "not necessarily dispositive" on prong one of the affirmative defense, nonetheless making the policy dispositive by explaining that the "employee may then rebut the employer's proof by showing that the sexual harassment policy is not effective"); O'Dell v. Trans World Entm't Corp., 153 F. Supp. 2d 378, 388 (S.D.N.Y. 2001) (granting employer's motion for summary judgment, noting that the employer had a sexual harassment policy (which the 2004] Operating in an Empirical Vacuum 239 In fact, in seventy-three percent of the cases in which the court discusses the employer's prevention obligation under prong one, the court shifts the burden of proof from the defendant employer to the plaintiff employee. 181 In most of these cases, the courts are not explicitly shifting the burden of proof, as in the Fourth Circuit. Instead, they are doing so by equating an employer's prevention obligation with the dissemination of an anti-harassment policy and a grievance procedure affording multiple avenues of complaint. This equation of prevention with paper policies and procedures is problematic. If an employer does not have to show that its policy and procedure works in practice, then, as a practical matter, the plaintiff must demonstrate that the policy is ineffective in order for the employer to lose on the prevention element of prong one of the affirmative defense. An examination of the cases shows that this is exactly what is happening. The courts' discussion of prevention reveals that the employer is not offering evidence that its policy and procedure are effective in practice and work to deter harassment. Instead, the plaintiff shoulders the burden of offering evidence that the employer has somehow failed in its obligation to prevent harassment. 182 If the plaintiff offers no evidence that the policy is defective court appeared to assume was effective based on the language of the policy) and explaining that plaintiff could "rebut the employer's proof by showing that the sexual harassment policy is not effective, which can be demonstrated, for example, by evidence that the employer did not disseminate its policy to its employees"), afftd, No. 01-7807, 2002 U.S. App. LEXIS 14446 (2d Cir. July 16, 2002); Pickett v. Colonel of Spearfish, 209 F. Supp. 2d 999, 1007 (D.S.D. 2001) (granting employer's motion for summary judgment, noting that employer had "in place policies and procedures for handling sexual harassment" and that the plaintiff "failed to produce conclusive evidence that the policies and procedures were not available to her or that they were somehow ineffective"); Good v. MMR Group, Inc., No. 3:OOCV-182- H, 2001 U.S. Dist. LEXIS 20036, at *14 (W.D. Ky. Dec. 4, 2001) (granting summary judgment motion by employer, explaining that the employer had a "clearly explained policy" and complaint procedures, as well as an EEO officer to receive complaints, and plaintiff had offered "no evidence to suggest that [her employer] would be unresponsive[,I that the policy was an insincere one, or that the procedures were unlikely to be effective"). 181 This figure is based on 132 of the total 200 cases of supervisory sexual harassment. See supra Part II. I arrived at the pool of 132 cases by eliminating cases in which the court either (1) did not discuss the employer's prevention obligation, instead basing its decision on the employer's correction obligation or prong two of the affirmative defense, or (2) discussed the affirmative defense in such a summary fashion that the court's views on prevention were impossible to discern. In 97 of these 132 cases (73 percent), the court either expressly shifted the burden of proof on prevention to the employee, as in the Fourth Circuit, or effectively did so by requiring nothing more than file cabinet compliance by the employer to satisfy prong one's prevention obligation. 182 See, e.g., Ritchie v. Stamler Corp., No. 98-5750, 2000 U.S. App. LEXIS 568, at *7-8 (6th Cir. Jan. 12, 2000) (affirming district court order granting employer's motion for summary judgment, rejecting plaintiffs argument that defendant's policy was ineffective because it required employees to report harassment in writing only to the firm president, and noting that there was no evidence that the president would have been unavailable had the victim tried to report the harassment to him); Hill v. Am. Gen. Fin., Inc., 218 F.3d 639, 643- 44 (7th Cir. 2000) (affirming district court order granting employer motion for summary Columbia Journal of Gender and Law [Vol. 13:2 or ineffective, then the court concludes that the employer has satisfied either its prevention obligation or prong one of the affirmative defense. 83 Not one judgment, noting that employer had "a number of policies in place at the time," which got "the job done" even though they left "room for improvement" and rejecting plaintiff's claim that she had not received employer's policy, stating that the policies "were kept in a set of notebooks in a 'public access type place' where the employees could look at them"); McCombs v. Chrysler Corp., No. IP 99-0697-C-T/K, 2003 U.S. Dist. LEXIS 6448, at *19-20 (S.D. Ind. Mar. 4, 2003) (granting employer's motion for summary judgment on plaintiff s claim of a hostile work environment, rejecting plaintiff's claim that the employer had not distributed the sexual harassment policy to its employees and concluding that plaintiff had "insufficient evidence to raise a triable issue as to whether the Defendant exercised reasonable care to prevent harassment"); Duviella v. Counseling Serv., No. 00-CV-2424, 2001 U.S. Dist. LEXIS 22538, at *36-42 (E.D.N.Y. Nov. 20, 2001) (granting employer's motion for summary judgment, noting that employer had a sexual harassment policy and grievance procedure, and rejecting plaintiffs arguments (1) that the employer had not taken preventative measures once it learned of the harassment and (2) that the employer's policy violated the law because it required a written complaint before the employer would commence an investigation), aff'd, No. 02-7019, 2002 U.S. App. LEXIS 24050 (2d Cir. Nov. 21, 2002); Walton v. Johnson & Johnson Servs., Inc., 203 F. Supp. 2d 1312, 1322 (M.D. Fla. 2002) (granting employer's motion for summary judgment, explaining that the employer had exercised reasonable care to prevent harassment because it had a sexual harassment policy, notified employees of the policy through an annual letter sent to employees, and provided sexual harassment training to supervisors, and rejecting plaintiffs claims that she had not received the letter and that "her supervisor [had] not attend[ed] training until after the alleged sexual harassment"), aft'd, 347 F.3d 1272 (11 th Cir. 2003); Clardy v. Silverleaf Resorts, Inc., No. 3:99-CV-2893-P, 2001 U.S. Dist. LEXIS 16481, at *28-31 (N.D. Tex. Oct. 10, 2001) (granting employer's motion for summary judgment, finding that employer had posted copies of its sexual harassment policy and had a complaint procedure for reporting harassment, and rejecting, as irrelevant to plaintiffs claim of harassment, affidavits from former co-workers who also alleged sexual harassment by employer's supervisors); cf Hardy v. Univ. of Ill. at Chi., 328 F.3d 361, 365 (7th Cir. 2003) (reversing, on the basis of prong two of the affirmative defense, district court order granting summary judgment to employer, but rejecting plaintiffs argument that employer's policy was not "reasonable enough to prevent harassment because the reporting procedures were merely 'suggestions' and the University did not adequately respond to her complaints," because the University had a "formal sexual harassment policy" allowing an employee to use the method of reporting with which she was comfortable, the harassment stopped once the plaintiff reported the harassment, and there was "no indication that the University's policy was ineffective or merely a disguised run-around"). 183 See, e.g., Zelaya v. E. & W. Hotel Corp., No. 99-16179, 2001 U.S. App. LEXIS 3492, at *4 (9th Cir. Mar. 5, 2001) (affirming trial court order granting summary judgment to employer and holding that employer had exercised reasonable care to prevent harassment because it had "post[ed] a reasonable anti-harassment policy in Spanish and in English in multiple locations accessible to the housekeeping staff"); Casiano v. AT&T Corp., 213 F.3d 278, 286-87 (5th Cir. 2000) (affirming trial court order granting summary judgment to employer, pointing to employer's policies (of which plaintiff was aware) forbidding harassment and encouraging employees to inform supervisors or its EEO officers of any workplace harassment); Leuiers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *8 (6th Cir. Feb. 3, 2000) (affirming trial court order granting summary judgment to employer on claim of supervisory harassment, concluding that employer had met prong one of the affirmative defense by passing out a sexual harassment policy with complaint procedure and disciplining harasser after completing its investigation); Wright v. Anixter, Inc., No. 98-17164, 1999 U.S. App. LEXIS 19962, at *4-5 (9th Cir. Aug. Operating in an Empirical Vacuum of the cases examined for this Article mentions that an employer has offered evidence that it has conducted surveys of its own employees to determine how the sexual harassment policy works in practice, or evaluated whether its policy has any deterrent effect on the incidence of harassment in its own workplace. As a practical matter, shifting the burden of proof to plaintiffs is inefficient. The defendant employer controls the process and the information generated by the process. Information about how the policy and procedure work in practice, and how employees fare when they invoke the internal procedures is readily accessible to the defendant employer, not the plaintiff employee. Thus, it makes no sense to require the plaintiff, who has neither control over the process nor ready access to the information generated thereby, to prove that the employer's policy and procedure are ineffective, defective, or dysfunctional. In addition, while employers may maintain comprehensive records of sexual harassment complaints, there is no legal requirement to do so.' 84 Therefore, placing this burden on a plaintiff when her employer may not even keep relevant records is simply unfair. How can a plaintiff show that the employer's internal grievance procedure is defective or ineffective if the employer does not keep records about (1) the average length of an investigation, (2) the disposition of prior sexual harassment complaints, (3) the sanctions imposed for various offenses, and (4) the current status of men who have violated the firm's sexual harassment policy and women who have reported harassment? Even more important, how can an employer 18, 1999) (affirming district court order granting employer's motion for summary judgment and holding that employer had satisfied its obligation to prevent harassment by having a sexual harassment policy, posted in the kitchen of the office where plaintiff worked, which contained contact names to report harassment, and that plaintiff had not argued that the policy or procedure was "inadequate or ineffective"); Ogden v. Keystone Residence, 226 F. Supp. 2d 588, 601-02 (M.D. Pa. 2002) (granting employer motion for summary judgment, holding that as a matter of law, the employer had satisfied the affirmative defense because it had a sexual harassment policy through which employees could report harassment to a "Human Resources Specialist," plaintiff had signed a form stating that she knew of the employer's procedures, and yet she did not report the harassment to the Human Resources Specialist); Olsen v. H.E.B. Pantry Foods, 196 F. Supp. 2d 436, 440 (E.D. Tex. 2002) (granting summary judgment for employer on sexual harassment claim, concluding that employer had met prong one of the affirmative defense because it had a sexual harassment policy and its human resources department acted promptly, terminating the harasser, once plaintiff reported the harassment); Slay v. Glickman, 137 F. Supp. 2d 743, 752 (S.D. Miss. 2001) (granting employer's motion for summary judgment, concluding that because the employer had a sexual harassment policy of which the plaintiff was aware and promptly investigated plaintiffs allegations of harassment, plaintiff had "failed to establish a genuine issue of material fact with regard to whether the Defendant exercised reasonable care to prevent and correct promptly sexual harassment in the workplace"); Hairston-Lash v. R.J.E. Telecom, Inc., 161 F. Supp. 2d 390, 394 (E.D. Pa. 2001) (granting summary judgment in favor of employer and stating that plaintiff did not contest that she had received employer's "extensive policies and procedures on handling sexual harassment"). 18429 C.F.R. § 1602.12 (2002). 2004] Columbia Journal of Gender and Law evaluate the effectiveness of its own policies and procedures if it does not maintain comprehensive complaint records? Finally, employers currently have little incentive to maintain such records. Many federal courts allow them to satisfy their obligation to prevent harassment with little more than a policy and complaint procedure. In fact, maintaining detailed evaluative records leaves the employer open to discovery requests that may produce information damaging to the employer's attempt to invoke the affirmative defense.'85 B. Prong Two: The Platonic Ideal of Victim Reasonableness In Ellerth and Faragher, the Supreme Court held that the defendant employer also must prove by a preponderance of the evidence that the plaintiff employee "unreasonably failed to take advantage of any preventive or corrective opportunities provided by the employer or to avoid harm otherwise."' 86 Typically, once an employee reports harassing conduct, the employer's obligation to correct the harassing behavior is triggered. Therefore, the employer's duty under prong one to correct harassment and the employee's duty under prong two to report harassment are interrelated. Many federal courts assume that if the employer's policy and procedure are reasonable and effective, then the plaintiff s failure to use that policy and procedure must be unreasonable as a matter of law. Therefore, when the plaintiff does not report the harassment to her employer, the employer often prevails on its motion for summary judgment or judgment as a matter of law by offering evidence sufficient to satisfy prong one of the affirmative defense: typically, evidence that it disseminated an anti- harassment policy and complaint procedure.' 87 Even when a plaintiff does report harassment, however, many federal courts interpret prong two of the affirmative defense in ways that make it easier for an employer to prove that the plaintiff employee unreasonably failed to avail herself of her employer's internal reporting 185 See, e.g., Spina v. Our Lady of Mercy Med. Ctr., 97 CIV 4661, 2001 U.S. Dist. LEXIS 7338, at *3, 8-9 (S.D.N.Y. June 7, 2001) (denying plaintiff's motion to compel discovery of harassment records related to another employee who had been discharged by employer and concluding that allowing discovery "could have a chilling effect on an employer's readiness and willingness to take action in harassment or discrimination cases, for fear of any documentation being used against them"); cf. McGrath v. Nassau County Health Care Corp., 204 F.R.D. 240, 246 (E.D.N.Y. 2001) (overruling employer's objection to magistrate judge's order directing employer to produce reports and notes by Amy Ventry, employer's outside counsel who had conducted an internal investigation of female employees' sexual harassment claims, concluding that the employer, by invoking the affirmative defense, had "waived the work product privilege under the facts of this case"). 186 Burlington Indus. v. Ellerth, 524 U.S. 742, 765 (1998); Faragher v. City of Boca Raton, 524 U.S. 775, 807 (1998). 187 See generally supra Part III.A. 242 [Vol. 13:2 Operating in an Empirical Vacuum mechanisms. In fact, in sixty-seven percent' 88 of the cases addressing prong two, the courts improperly interpret the employer's obligation under prong two. First, a handful of courts re-write the legal test enunciated in Ellerth and Faragher so that an employer may prevail on the affirmative defense by demonstrating nothing more than a prompt and appropriate response to a complaint of harassment. Second, some courts defer to the employer's policy to determine when the employer obtains notice of harassing conduct. Finally, other courts engage in fact-finding on motions for summary judgment and judgment as a matter of law, concluding that any "delay" in reporting equals an unreasonable failure on the part of the plaintiff to use her employer's reporting machinery. There are several problems with the federal courts' interpretation of prong two of the affirmative defense. First, allowing employers to prevail on the affirmative defense by demonstrating a prompt response to a complaint of harassment effectively eliminates prong two of the affirmative defense. Second, allowing employers to define how and when they obtain notice of workplace harassment encourages the creation of a private system of rules over which there is no effective oversight, and provides incentives to employers to narrow the avenues available for employees to file complaints. Third, the empirical research shows that while formal reporting of sexual harassment is an uncommon occurrence, retaliation in response to such complaints is not. This raises an important question: why is not reporting harassment considered unreasonable? Finally, by requiring almost nothing more from employers on prong two than what the employer offers to satisfy the very low burden set by the courts for prong one, the courts, once again, effectively shift the burden of proof on prong two from the defendant employer to the plaintiff employee. 188 This figure is based on 111 of the total 200 cases of "supervisory" sexual harassment. See supra Part II. I arrived at the pool of 111 cases by eliminating cases in which (1) the plaintiff made no report of harassment, (2) the court based its decision on prong one of the affirmative defense, or (3) the court's opinion provided little dependable guidance as to its views on the correct interpretation of prong two. In 74 of these 111 cases (67 percent), the court interpreted prong two so as to make it easier for an employer to prove that an employee had unreasonably failed to avail herself of the employer's corrective or preventive opportunities. The pool of seventy-four cases includes a handful of decisions in which the plaintiff prevails, for example on a motion for summary judgment, but the court's opinion indicates that had plaintiff not followed her employer's policy the court may have reached a different result on prong two. See, e.g., Haugerud v. Amery Sch. Dist., 259 F.3d 678, 699-700 (7th Cir. 2001) (reversing trial court order granting summary judgment to defendant on plaintiff's hostile work environment claim, explaining that failure of an employee to use her employer's grievance procedure would indicate that she had failed to use reasonable care in avoiding harm, but that because "plaintiffs decision to file an external complaint rather than filing a complaint with the Board was in accordance with the Board's own policy, the Board can not now allege that plaintiff did not fulfill her obligation of reasonable care"). 2432004] Columbia Journal of Gender and Law 1. Re-writing the Affirmative Defense In Ellerth and Faragher, the Court made clear that an employer must prove both that it exercised reasonable care to prevent and correct harassing behavior by supervisors and that the plaintiff employee unreasonably failed to use the employer's reporting machinery. 8 9 In fact, in his dissent in Ellerth, Justice Thomas expressed concern that even if the employer exercised reasonable care to prevent and correct harassment, the employer would still be liable (unable to satisfy the affirmative defense) if the plaintiff "fulfilled her duty of reasonable care to avoid harm."' 90 In other words, an employer cannot escape liability simply by showing that it has satisfied prong one of the affirmative defense. A small group of courts in the Second, Fourth, and Fifth Circuits, however, has re-written the Supreme Court's decisions in Ellerth and Faragher to allow an employer to prevail on the affirmative defense so long as it can demonstrate a prompt response to the employee's complaint of sexual harassment. For example, in both Van Alstyne v. Ackerley Group, Inc.,'91 and Yerry v. Pizza Hut,' 2 the Second Circuit and the district court for the Northern District of New York re-wrote the Ellerth and Faragher affirmative defense to eliminate its second prong. Both courts concluded that an employer may escape liability for harassment by a supervisor by showing that (1) it exercised reasonable care to prevent and correct harassment and either (2)(a) the employee failed in her duty of reasonable care or (2)(b) the employer promptly corrected the harassing behavior once the employee complained. 9 3 The problem with this formulation is that prong (2)(b) simply restates the employer's correction obligation under prong one. In other words, the employer satisfies the affirmative defense by offering proof only as to prong one. Similarly, in Indest v. Freeman Decorating, Inc.,' 94 Judge Jones of the Fifth Circuit affirmed the district court's order granting judgment as a matter of law to the employer, holding that the employer's "prompt remedial response relieve[d] it of Title VII vicarious liability."' 95 Judge Jones concluded that "[i]mposing vicarious liability on an employer for a supervisor's 'hostile environment' actions despite its swift and appropriate 8 9 See Ellerth, 524 U.S. at 765; Faragher, 524 U.S. at 807. 190 524 U.S. at 773 (Thomas, J., dissenting). '91No. 00-9340, 2001 U.S. App. LEXIS 10588 (2d Cir. May 21, 2001). 192 186 F. Supp. 2d 178 (N.D.N.Y. 2002). 193 Van Alstyne, 2001 U.S. App. LEXIS at *10; Yerry, 186 F. Supp. 2d at 185 (citing Van Alstyne). 194 164 F.3d 258 (5th Cir. 1999). 195 Id. at 267. [Vol. 13:2 Operating in an Empirical Vacuum remedial response to the victim's complaint would ... undermine not only Meritor but Title VII's deterrent policy.' 96 Judge Jones's decision in Indest is not precedent in the Fifth Circuit because Judges Wiener and Furgeson concurred only in the judgment.1 97 Indest, however, has had an impact outside the Fifth Circuit. In Watkins v. Professional Security Bureau, Ltd.,' 98 the Fourth Circuit cited favorably to Indest in dicta. The court concluded that even if the plaintiff had taken advantage of her employer's reporting procedures, it still would have affirmed the trial court's order granting the employer judgment as a matter of law, or in the alternative a new trial, on plaintiffs hostile environment claim.199 The Fourth Circuit found it inconceivable that an employer could be held liable for harassment even when it responded promptly to a complaint of harassment simply because it could not "satisfy the literal terms of the second element of the affirmative defense." 2°° The court concluded that such a result would undermine the purpose of the affirmative defense's limitation on liability-the promotion of conciliation 201 Although only a handful of courts interpret the affirmative defense so as to eliminate prong two, the decisions are troubling for two reasons. First, they reveal a bias in favor of employers. Courts like Watkins are less concerned with the plaintiffs ability to secure a remedy under Title VII than with what they consider a "fair" result for the employer. Second, eliminating prong two of the affirmative defense effectively deprives a plaintiff of her federal remedy for harassment, even when she acts promptly to report the harassing conduct. Suppose an employee were assaulted at work and promptly reported the assault. The Van Alstyne, Indest, and Watkins courts would relieve the employer of all liability for the assault so long as the employer responded promptly and appropriately to the plaintiffs complaint. Such an interpretation of Title VII effectively robs a plaintiff of her Title VII remedy, notwithstanding her exercise of reasonable care. 202 '961d. at 266. 197 See Indest v. Freeman Decorating, Inc., 168 F.3d 795, 796 n.12 (5th Cir. 1999) (Wiener, J., specially concurring). '9' No. 98-2555, 1999 U.S. App. LEXIS 29841 (4th Cir. Nov. 15, 1999), cert. denied, 529 U.S. 1108 (2000). '99 See Watkins, 1999 U.S. App. LEXIS, at *21 n.16. 200id. 201 See id. 202 See supra note 7. 2004] 246 Columbia Journal of Gender and Law [Vol. 13:2 2. Allowing Employers to Define When They Obtain Notice Some courts allow the employer to define how and when it obtains notice of harassment from an employee. 20 3 If an employee reports harassing behavior to a manager not designated in her employer's policy, some courts deem the plaintiff's notice to be ineffective. Because the employer is not on notice, it is not obligated to respond to the plaintiff's complaint. In addition, the plaintiffs failure to follow the dictates of her employer's grievance procedure means that she either has not reported the harassment or impermissibly delayed in doing so, allowing the employer to argue that the plaintiff unreasonably failed to invoke its internal procedures, as required 204by prong two of the affirmative defense. 203 See, e.g., Ritchie v. Stamler Corp., No. 98-5750, 2000 U.S. App. LEXIS 568, at *7-8 (6th Cir. Jan. 12, 2000) (rejecting plaintiffs argument that employer's policy was ineffective because it required employees to complain of harassment in writing to the president of the company); Coates v. Sundor Brands, Inc., 164 F.3d 1361, 1364 (1 1th Cir. 1999) (noting that the question of whether the employer had adequate notice of the plaintiff s harassment was "made easy" because the employer, through its policy, had "answered the question of when it would be deemed to have notice of the harassment"); Roelen v. Akron Beacon Journal, 199 F. Supp. 2d 685, 694 (N.D. Ohio 2002) (stating that plaintiff "never filed a complaint under Defendant Akron Beacon Journal's sexual harassment policy" and her complaint to Home Delivery Manager James DeLuca did not indicate that her complaint stemmed from "sex discrimination, as opposed to general bad temperament"); Duviella v. Counseling Serv., No. 00-CV-2424, 2001 U.S. Dist. LEXIS 22538, at *40-42 (E.D.N.Y. Nov. 20, 2001) (rejecting plaintiffs argument that a "material question of fact exist[ed] as to the lawfulness of [employer's] complaint procedures," which required "a written and signed statement before [the employer]" could investigate, stating that the policy allowed victims to discuss with anyone in management their concerns about harassment and noting that the court was "not aware of any cases which have found a complaint procedure that requires a written statement before the commencement of an investigation to be an ineffective procedure"), aff'd, No. 02-7019, 2002 U.S. App. LEXIS 24050 (2d Cir. Nov. 21, 2002); Madrid v. Amazing Pictures, No. 99-1565, 2001 U.S. Dist. LEXIS 23942, at *29 n.5 (D. Minn. July 23, 2001) (stating that while assistant manager, to whom plaintiff had complained about harassing conduct, "could have responded more promptly and aggressively, the fundamental fact remains that she was not the person designated by the company to receive sexual harassment complaints"). 204 See, e.g., Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *3-5 (4th Cir. Nov. 13, 2001) (holding that plaintiff unreasonably failed to avail herself of her employer's procedures because Morgan, a Chief Probation Officer, to whom plaintiff had complained about the harassment, had no corrective authority over the harasser and "was not one of the avenues through which the policy provided Dowdy should report the harassment"); O2den v. Keystone Residence, 226 F. Supp. 2d 588, 601-02 (M.D. Pa. 2002) (holding that because the employer had an anti-harassment policy allowing employees to complain to a Human Resources Specialist, plaintiffs failure to do so, even though she may have informed her immediate supervisors of the harassment, "allow[ed] [the employer] to prevail on the affirmative defense"); Good v. MMR Group, Inc., No. 3:OOCV-182-H, 2001 U.S. Dist. LEXIS 20036, at *4-5, 14-15 (W.D. Ky. Dec. 4, 2001) (applying affirmative defense to co-worker harassment case and concluding that even though a foreman saw plaintiff being harassed and plaintiff's husband reported the harassment to a company superintendent plaintiff had unreasonably failed to use her employer's grievance procedure because she provided no reason why she did not report the harassment to her employer's Operating in an Empirical Vacuum By deferring to employers' private definitions of notice for purposes of the affirmative defense, these courts permit a private entity, the employer, to unilaterally define what constitutes notice in the context of Title VII sexual harassment law. Employees, who are required to invoke their employer's internal grievance procedure, are given no voice. Moreover, by deferring to employers' policies on notice, the courts abdicate their responsibility to evaluate the reasonableness of employers' preventive and corrective mechanisms. In effect, "federal sexual harassment law... incorporate[s] organizational ideology into state law," 205 with no examination of whether the goals of private employers are consistent with those of Title VII law. Courts that allow employers to define notice fail to recognize that employers are not neutral participants in the process; they are potential defendants.206 Thus, their primary goal is to avoid liability, not to ensure Title Vil's "primary objective" of preventing workplace harassment. The end result is a scheme in which outcomes turn not on federal courts' efforts, at least within a circuit, to render consistent decisions that effectuate Title VII's goals, but rather on the balance struck by each individual employer between the competing goals of increasing reporting and limiting liability. In addition, allowing employers to define what constitutes notice creates improper incentives. The Eleventh Circuit's decision in Madrav v. Publix Supermarkets, Inc., 207 illustrates the problem. Madray involved a sexual harassment suit filed by Connie Lynn Madray and Melody Holden against their employer, Publix Supermarkets, based on the conduct of their supervisor, Publix store manager Ralph Selph. °8 Selph's conduct included "suggestive and provocative comments," "designated compliance official"); see also Mukaida v. State, 159 F. Supp. 2d 1211, 1232-33 (D. Hi. 2001) (concluding that if alleged harasser was plaintiffs supervisor, employer nonetheless was entitled to the affirmative defense because plaintiff's complaint to her immediate supervisor, Michael Hamnett, was not pursuant to her employer's policy); cf Frederick v. Sprint/United Mgmt. Co., 246 F.3d 1305, 1316 (1lth Cir. 2001) (concluding that there were genuine issues of material fact about plaintiffs reasonable care to avoid harassment; if the employer's 1994 policy were in effect, which allowed complaints of harassment to be made to any manager, then plaintiff's complaints sufficed to show reasonable care, but if the 1990 policy were in effect, then plaintiff had failed to complain in accordance with her employer's policy). 205 Cahill, supra note 124, at 87. 206 Cf Richard Edwards, Rights at Work: Employment Relations in the Post-Union Era 216-17 (1993) (explaining that internal complaint and grievance procedures implemented by employers with non-union workforces "justifiably bring upon themselves the suspicion of conflict of interest, inasmuch as company officials are likely to think beyond the actual merits of the grievance immediately before them and worry about such matters as the besmirching of the firm's reputation, showing support for the supervisory staff, and the implications of a decision for other employees"). 207 208 F.3d 1290 (2000). 20 1 See id. at 1292-93. 2004] 247 Columbia Journal of Gender and Law as well as "groping, full-body hugs, rubbing his body against [plaintiffs'] in such a way that his genitals made contact with their bod[ies], kissing, blowing in their ear[s], wetting his finger in his mouth and sticking it in their ears, and rubbing their shoulders, backs, hips, and legs." 2 9 The district court granted the employer's motion for summary judgment, concluding that the employer was entitled to judgment based on the affirmative defense.210 Madray and Holden appealed, contending that Publix could not satisfy the elements of the affirmative defense. 21' The Eleventh Circuit disagreed and affirmed the trial court's order granting the employer's motion for summary judgment. 212 The court first determined that Publix had exercised reasonable care to prevent harassment by virtue of its sexual harassment policy. 213 Publix's policy provided that "[i]n order for the Company to deal with [sexual harassment], we must report such offensive conduct or situations to the Store Manager, District Manager, or Divisional Personnel Managers."'1 4 The court also found that Publix had acted promptly to correct the harassment because District Manager Richard Rhodes "was responsive to [plaintiffs'] complaints" 215 once they filed a formal report with him. Plaintiffs claimed, however, that Rhodes's response to their formal complaint was not prompt corrective action because Publix had notice of the harassment through plaintiffs' earlier complaints to mid-level managers about Selph's harassing conduct. 1 6 About six months before Holden complained to Rhodes, she spoke with Second Assistant Manager Darlene Clark, telling Clark 'that it made me sick for [Selph] to hug me and touch me and kiss me.' 217 Approximately one to two months later, Holden told Second Assistant Manager Gary Priest, while dining out with Priest and several other store employees, that Selph had "'grabbed me and ducked me over and kissed me on the neck.' 218 Several weeks before Holden complained to Rhodes, the 209 Id. at 1293 n.2. The court explained that it was not giving a "detailed account" of the harassment because the issue of whether the conduct was severe or pervasive was not before it. Id. 210 Id. at 1293 n.1, 1295-96. 211 Id. at 1297. 212 Id. at 1293. 213 Id. at 1298-99. 214 Id. at 1294 (citation and footnote omitted) (emphasis in original). 215 Id. 216 Id. at 1297. 2117Id. at 1293 (citation omitted). 218 Id (citation omitted). [Vol. 13:2 Operating in an Empirical Vacuum Bakery Manager David Neff saw Selph engage in inappropriate conduct toward Holden.21 9 Neff told Selph that his behavior was harassment and informed Holden that as a manager he had to report harassment, so if the behavior did not stop, Holden should let Neff know. 220 Holden thanked Neff for his help and told him that she and Madray would arrange to speak with Rhodes. 22' Finally, several days before Holden reported Selph's conduct to Rhodes, Priest saw Selph act inappropriately toward both Holden and Madray.222 Upon witnessing the incident with Madray, Priest told Holden that they needed to act; Holden asked Priest to contact Rhodes, who then met with the plaintiffs and began an investigation. 223 The Eleventh Circuit rejected plaintiffs' argument that Publix was on notice of the harassment based on the plaintiffs' complaints to Clark and Priest, and Priest's and Neff's witnessing Selph's harassing conduct. 224 The court explained that Publix had disseminated a sexual harassment policy and procedure that allowed employees to bypass a harassing supervisor by providing multiple avenues of complaint, including a toll-free number for employees to register complaints.225 Thus, the court concluded that Publix's policy satisfied "the minimum requirements for the Faragher affirmative defense. 226 This was a critical step in the court's analysis. Because the policy was reasonable, employees had to follow it. "[O]nce an employer has promulgated an effective anti-harassment policy and disseminated that policy and associated procedures to its employees, then 'it is incumbent upon the employees to utilize the procedural mechanisms established by the company specifically to address problems and grievances. 227 Because Madray and Holden had not complained to a Store Manager, District Manager, or Divisional Personnel Manager, as required by Publix's policy, their complaints to Clark and Priest, and Priest's and Neff's witnessing of Selph's harassment, did not put Publix on notice.228 As a result, Publix had 219 id. 220 id. 22 1 id. 222 ld. at 1293-94. 223 Id. at 1294. 224 Id. at 1300. 225 Id. at 1298. 226 Id. at 1299. 227 Id. at 1300 (quoting Farley v. Am. Cast Iron Pipe, 115 F.3d 1548, 1554 (11th Cir. 1997)). 228 id. 2492004] Columbia Journal of Gender and Law no duty to correct Selph's behavior before the plaintiffs filed a formal complaint with District Manager Rhodes.229 The court also held that "the plaintiffs unreasonably delayed utilizing the complaint procedures established by Publix's sexual harassment policy. '230 It rejected plaintiffs' argument that they had reported the harassment in a timely manner by using Publix's Open Door Policy, which "encouraged employees to talk to a manager about any 'problems or misunderstandings.' '2 3 ' The court found that plaintiffs did not "reasonably utilize the complaint procedures provided by Publix's Open Door Policy" because plaintiffs made their complaints "during informal, often social, circumstances" and "did not fully inform the mid-level managers of the extent of the harassment Selph directed towards them., 232 The decision in Madray illustrates the problem with allowing employers to define notice. Why did neither Clark nor Priest pass along Holden's comments about Selph to Rhodes? The Eleventh Circuit simply re-wrote Holden's deposition testimony in addressing that issue. It held that Holden only told Clark that Selph's behavior made her sick.233 But Holden alleged that she had told Clark "'it made me sick for [Selph] to hug me and touch me and kiss me.' 234 The court's conclusion that Holden had not "disclosed the extent or precise nature of Selph's behavior" is simply 235disingenuous. Moreover, even had Holden simply said that Selph's behavior made her sick, why was Clark under no obligation to ask why Selph's conduct made Holden sick? "[Tihere is no legal mandate that an employee use the term 'sexual harassment' in order to inform his or her employer about a harasser in the workplace." 236 Holden's comment to Clark certainly suggests that Selph may have been engaging in harassing conduct in the workplace. If Holden's statement does not suffice to put a manager on notice of potential sexual harassment, then what would do So?237 229 See id. 230 Id. at 1302. 231 Id. at 1294; id. at 1301-1302. 232 Id. at 1302. 233 Id. at 1300. 234 1d. at 1293 (citation omitted). 235 Id. at 1300. 236 Gentry v. Exp. Packaging Co., 238 F.3d 842, 849 (7th Cir. 2001). 237 In Coates v. Sundor Brands, Inc., 164 F.3d 1361 (11th Cir. 1999), Vickie Coates, the plaintiff, showed a note that Ernie Long, the alleged harasser, had written to her to Lloyd McLean, the manager of the plant at which Coates and Long worked. Id. at 1365. The note read as follows: "From the Desk of Ernie Long, Hey Sweetheart $100 for 45 minutes of hugging and kissing or $100 for stop loving Vickie guarantee." Id. The Eleventh Circuit held that Coates's interaction with McLean "did not put McLean on notice that she was being sexually harassed by Long." Id. The court explained that McLean was getting 250 [Vol. 13:2 2004] Operating in an Empirical Vacuum 251 The court also concluded that Priest "could not have been expected to take action" based on what Holden had told him because she failed to tell Priest that the "incident was part of an ongoing pattern of sexually harassing behavior or that she wanted Priest to take action to stop Selph's inappropriate behavior. 238 Once again, why was Priest under no obligation to explore further with Holden what had been happening with Selph? Holden's description of Selph's behavior was not ambiguous; she told Priest that Selph had "'grabbed [her] and ducked [her] over and kissed [her] on the neck.' ''239 She also told Priest that she did not know what to do about 240Selph's conduct. Why does the fact that her complaint occurred during a social setting relieve Priest of any obligation to make further inquiries or to pass along her complaint to a person with authority to investigate? Nothing in Publix's Open Door Policy limited an employee to disclosing problems 241to managers only on the work site. Why must an employee victimized by harassment set forth all the particulars of the harassment and explicitly ask the manager to take the complaint forward before the employer is required to take any action?242 The Eleventh Circuit's decision in Madray allows an employer to sit passively by until an employee files a formal complaint using the employer's designated complaint channels. It does not encourage employers to require supervisory personnel to report harassment that they witness or that is reported to them.243 Even if a victim of harassment complains to her ready to leave on a business trip to Japan and that Coates gave McLean no indication that the note was part of a pattern of harassing behavior by Long. See id. 238 Madrav, 208 F.3d at 1300-1301. 239 id. at 1293 (citation omitted). 240 id. 241 See id. at 1294-95. 242 What is interesting about the court's conclusions is that plaintiffs claimed that Rhodes was upset when he heard about their complaints because .' he said that the managers knew better and should have let him know what was going on."' Id. at 1294 (citation omitted). 243 See, e.g., Watkins v. Prof 1 Sec. Bureau, Ltd., No. 98-2555, 1999 U.S. App. LEXIS 29841, at *15-16 & n.10 (4th Cir. Nov. 15, 1999) (concluding that even if plaintiff's testimony that she had reported the harassment to Dowling, a site supervisor, in July were true, it was "irrelevant to the question of whether [the employer] acted reasonably to prevent and promptly correct" harassment because Dowling, who did not pass on the information until August, could neither investigate harassment nor discipline harassers), cert. denied, 529 U.S. 1108 (2000); Good v. MMR Group, Inc., No. 3:OOCV-182-H, 2001 U.S. Dist. LEXIS 20036, at *4-5, 14-15 (W.D. Ky. Dec. 4, 2001) (applying affirmative defense to co-worker harassment case and concluding that even though a foreman saw plaintiff being harassed and plaintiff's husband reported the harassment to a company superintendent, the employer did not have notice of the harassment until a journeyman electrician employed at the plant informed a member of the company's personnel department who, in turn, transferred him to the company's EEO officer). 252 Columbia Journal of Gender and Law [Vol. 13:2 immediate supervisor about harassing behavior, if that supervisor is not a designated reporting person, then neither the supervisor nor his employer has any further obligation to the harassment victim.244 Managers not in the official chain of reporting are not required to follow up with even the simplest of questions when an employee complains about harassment to them. Instead, the burden rests solely with the employee to clearly set forth the entire pattern of workplace harassment and request the manager to take further action. Simply because the supervisor to whom harassment is initially reported does not have the authority to investigate or discipline245 does not explain why that supervisor cannot report the harassment to a member of management who then can take action.24 By permitting employers to restrict the methods by which victims can report harassment, courts make reporting more difficult. 247 Making reporting more difficult, in turn, ensures that fewer employees will do so. Formal reporting is already a rarely used means for dealing with sexual 2 44 See, e.g., Ogden v. Keystone Residence, 226 F. Supp. 2d 588, 601-02 (M.D. Pa. 2002) (holding that because the employer had an anti-harassment policy allowing employees to complain to a Human Resources Specialist, plaintiffs failure to do so, even thoughi she had informed her immediate supervisors of the harassment, "allow[ed] [the employer] to prevail on the affirmative defense"); Mukaida v. State, 159 F. Supp. 2d 1211, 1232-33 (D. Hi. 2001) (concluding that even if plaintiffs harasser were a supervisor, plaintiff had not reported the harassment until July 1, 1998, despite her allegation that she had informed her direct supervisor, Michael Hamnett, about the harassing conduct in December of 1996, because Hamnett was not designated under the employer's policy as a person to whom harassment could be reported). 245 Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *3-5 (4th Cir. Nov. 13, 2001) (rejecting plaintiffs contention that her employer's policy was ineffective because she had reported the harassment to Morgan, a Chief Probation Officer, who had not reported it to the employer, because Morgan had no corrective authority over the harasser and "was not one of the avenues through which the policy provided Dowdy should report the harassment"); Watkins, 1999 U.S. App. LEXIS 29841, at *15-16 & n.10 (4th Cir. Nov. 15, 1999) (concluding that even if plaintiffs testimony that she had reported the harassment to Dowling, a site supervisor, in July were true, it was "irrelevant to the question of whether [the employer) acted reasonably to prevent and promptly correct" harassment because Dowling, who did not pass on the information until August, could neither investigate harassment nor discipline harassers). 246 Cf Wilson v. Tulsa Junior Coll., 164 F.3d 534, 541-42 (10th Cir. 1998) (stating, in a case involving a claim of employer liability premised on negligence, that a "procedure that does not require a supervisor who has knowledge of an incident of sexual harassment to report that information to those who are in a position to take action falls short of that which might absolve an employer of liability" (quoting Varner v. Nat'l Super Mkt., 94 F.3d 1209, 1214 (8th Cir. 1996)). 247 See, e.g., Ritchie v. Stander Corp., No. 98-5750, 2000 U.S. App. LEXIS 568, at *7 (6th Cir. Jan. 12, 2000) (capitalization in original) (finding acceptable employer's policy requiring "employees to bring grievances regarding sexual harassment 'to the attention of the President IN WRITING"'). 2004] Operating in an Empirical Vacuum 253 harassment.248 If the purpose of requiring internal reporting is to "encourage employees to report harassing conduct before it becomes severe or pervasive," 249 then it makes sense to read notice more expansively, obligating supervisors to pass along reports of harassment to the appropriate personnel. If the employer can avoid liability by narrowly defining when it obtains notice, then an employee can repeatedly complain, albeit to the "wrong" people, and obtain no relief. The objectionable conduct continues, thereby frustrating Ellerth's and Faragher's stated goal of deterrence. 3. Why is Not Reporting Considered "Unreasonable"? The empirical evidence over the past twenty years has consistently shown that few victims of workplace harassment report the harassment or 250file a formal complaint. Yet, some federal courts, on motions for summary judgment, conclude that a plaintiff's "delay" in reporting harassment is unreasonable as a matter of law. Courts have held that plaintiffs who have waited six weeks,251 two months,252 three months,253 248 See supra notes 73-80 and accompanying text. 249 Burlington Indus. v. Ellerth, 524 U.S. 742, 764 (1998). 250 See supra notes 73-80 and accompanying text. 251 See Clardy v. Silverleaf Resorts, Inc., No. 3:99-CV-2893-P, 2001 U.S. Dist. LEXIS 16481, at *3-6, 28-29 (N.D. Tex. Oct. 10, 2001) (granting employer's motion for summary judgment and holding that plaintiffs "failure to indicate to Silverleaf management in a timely fashion the problems that she was experiencing with her supervisor constitute unreasonable failure to take advantage of any preventive or corrective opportunities provided by her employer," even though plaintiff reported her supervisor's conduct in mid-January of 1999, no more than six weeks after starting work under his supervision after Thanksgiving of 1998). 252 See, e.g., McCombs v. Chrysler Corp., No. IP 99-0697-C-T/K, 2003 U.S. Dist. LEXIS 6448, at *23-34 (S.D. Ind. Mar. 4, 2003) (granting employer's motion for summary judgment and concluding that employer satisfied prong two of the affirmative defense because the plaintiff "waited at least two months before reporting" the harassing conduct to her employer); Olsen v. H.E.B. Pantry Foods, 196 F. Supp. 2d 436, 440 (E.D. Tex. 2002) (granting summary judgment to employer and concluding that employer met prong two of the affirmative defense because plaintiff "failed to take advantage of HEB's sexual harassment policy and complaint procedures by waiting two months after the initial comments were made.., to complain"). 253 See, e.g., Taylor II, 2002 U.S. Dist. LEXIS 14560, at *18 (N.D. Tex. Aug. 7, 2002) (granting employer's motion for summary judgment and finding that employer satisfied prong two of the affirmative defense because plaintiff "offer[ed] no evidence to explain why she waited approximately three months before reporting the harassment"), aft'd, No. 02-10950, 2003 U.S. App. LEXIS 13103 (5th Cir. June 3, 2003); Walton v. Johnson & Johnson Servs., Inc., 203 F. Supp. 2d 1312, 1324-25 (M.D. Fla. 2002) (granting employer's motion for summary judgment, finding that "[t]he three month delay by Plaintiff in reporting the first alleged act of harassment (and thus allowing it to continue) initially weighs against her" on prong two of the affirmative defense and determining that plaintiffs fear of retaliation did not excuse the delay), affd, 347 F.3d 1272 (11th Cir. 2003); Green v. Wills 254 Columbia Journal of Gender and Law [Vol. 13:2 five months, 54 six months,255 or seven months25 6 from the first incident of harassment have unreasonably failed to avail themselves of their employer's grievance mechanism. If the typical victim of harassment does Group, Inc., 161 F. Supp. 2d 618, 621, 626 (D. Md. 2001) (granting employer's motion for summary judgment and concluding that "[p]laintiff did not pursue the available avenues of corrective action in a timely manner," even though plaintiff reported the harassment in August of 1998 and it had begun in May of 1998, only three months before); cf Phillips v. Taco Bell Corp., 83 F. Supp. 2d 1029, 1034 n.3 (E.D. Mo. 2000) (finding, after bench trial, that employer had satisfied affirmative defense because, in part, plaintiff had waited three months and seven days from the first incident of harassment to report it, and explaining in a footnote that such a "delay" "may not alone prove that the plaintiff employee unreasonably failed to take advantage" of the employer's grievance machinery, but that it did so in this case); but see Hardy v. Univ. of Ill., 328 F.3d 361, 365 (7th Cir. 2003) (reversing district court order granting summary judgment to employer on the affirmative defense, concluding that on summary judgment the court could not "say as a matter of law that it was unreasonable for Hardy to wait until May 3, 2000, to report Green's behavior" and even though she complained "almost three months after Green's first alleged act of harassment, it was only eight days after his last"). 254 See Taylor I, 2002 U.S. Dist. LEXIS 14549, at *16 (N.D. Tex. Aug. 7, 2002) (granting employer motion for summary judgment, explaining that an employer cannot remedy harassment until the victim makes a complaint and concluding that the plaintiff waited a little over five months to report the harassment, did not report the threats that she received even though the investigation was ongoing, and resigned before the investigation was complete). 255 See, e.g., Reese v. Meritor Auto., Inc., No. 00-1604, 2001 U.S. App. LEXIS 3517, at *14 (4th Cir. Mar. 8, 2001) (affirming district court order granting summary judgment to employer and finding that employer satisfied prong two of the affirmative defense because plaintiff waited over six months from the time her supervisor's advances began to report his behavior); Leugers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *3-4, 8-9 (6th Cir. Feb. 3, 2000) (affirming district court order granting summary judgment to employer on plaintiff's supervisory sexual harassment claim and finding that plaintiffs delay in reporting for a period of six months from the beginning of the harassing conduct was unreasonable); Hill v. Am. Gen. Fin., Inc., 218 F.3d 639, 641, 644 (7th Cir. 2000) (affirming district court order granting employer's motion for summary judgment on plaintiff's sex and race harassment claims and concluding that employer satisfied prong two of the affirmative defense because it responded once plaintiff complained on April 14, 1995, approximately six months after the harassment began in the fall of 1994), reh'g en banc denied, No. 99-2682, 2000 U.S. App. LEXIS 18367 (7th Cir. July 27, 2000); Arnold v. Yale New Haven Hosp., 213 F. Supp. 2d 142, 143-44, 149 (D. Conn. 2002) (holding that plaintiff had unreasonably failed to avail himself of employer's procedures because employer claimed that plaintiff did not report harassment until July 8, 1998, when harassing conduct had begun some time in early 1998); Slay v. Glickman, 137 F. Supp. 2d 743, 752 (S.D. Miss. 2001) (granting employer's motion for summary judgment on plaintiff's claims of both quid pro quo and hostile work environment harassment, and concluding, on the latter theory, that plaintiffs delay of six months from the date on which her supervisor propositioned her for sex was unreasonable). 256 See Gawley v. Ind. Univ., 276 F.3d 301, 312 (7th Cir. 2001) (concluding that plaintiff had unreasonably failed to use her employer's complaint procedures because she waited about seven months before filing a formal complaint, during which time her harasser "made up to three inappropriate comments to her each day," and her requests to the harasser to stop were not working). Operating in an Empirical Vacuum not report harassment, then how is a court able to conclude, on a motion for summary judgment, that a plaintiffs failure to report harassment, or her delay in doing so, is unreasonable as a matter of law? It is difficult to reconcile the courts' conclusions that victims who report harassment to a manager, albeit one not designated in the employer's policy, 257 or who delay several weeks to several months from the first incident of harassing behavior to complain, have acted unreasonably as a matter of law, given that most victims never report harassment in the first instance. In cases of "delayed" reporting, the courts' rush to judgment is especially problematic. Typically, it is the accretion of events over time, not a single egregious incident, which makes a work environment hostile. Not every workplace slight is harassment and, as the courts often point out, Title VII is not a general civility code.25 9 Therefore, a reasonable jury certainly might conclude that a plaintiff acted reasonably by waiting several months from the first incident of objectionable conduct to report her supervisor's behavior. If the courts are not evaluating a plaintiffs response to harassment against the empirical evidence about victims' responses to harassment, then what are they using to determine that the plaintiffs conduct is unreasonable? The Fifth Circuit's decision in Wyatt v. Hunt Plywood Co., Inc.,26 illustrates the problem. In Wyatt, the plaintiff Alisha Wyatt complained about her supervisor John Thompson's harassment to Donald Gorum, who was Thompson's superior. 26' Not only were Gorum's "informal discussions with Thompson" ineffective in stopping Thompson's harassment, but Gorum 262himself also began to harass Wyatt. Approximately eight months after Thompson began his campaign of harassment and a little over four months after Gorum began harassing her, Wyatt filed a formal complaint with Buddy Rachal, Gorum's supervisor.263 She did so in response to a 257 See supra Part III.B.2. 258 See supra notes 73-80 and accompanying text. 259 See, e.g., Patterson v. County of Fairfax, No. 99-1738, 2000 U.S. App. LEXIS 11009, at *14-15 (4th Cir. May 18, 2000); Hocevar v. Purdue Frederick Co., 223 F.3d 721, 741 (8th Cir. 2000) (Gibson, J., concurring); Indest v. Freeman Decorating, Inc., 164 F.3d 258, 263 (5th Cir. 1999); Williams v. Gen. Motors Corp., 187 F.3d 553, 564 (6th Cir. 1999), reh'g en banc denied, No. 97-3351, 1999 U.S. App. LEXIS 26482 (6th Cir. Sept. 30, 1999); Mendoza v. Borden, Inc., 195 F.3d 1238, 1245 (1 1th Cir. 1999), cert. denied, 529 U.S. 1068 (2000). 260 297 F.3d 405 (5th Cir. 2002), reh'g en banc denied, 2002 U.S. App. LEXIS 18781 (5th Cir. Aug. 14, 2002), cert. denied, 537 U.S. 1188 (2003). 261 Id. at 407. 262 Id. 263 See id. at 407-08. 20041 255 Columbia Journal of Gender and Law particularly egregious incident in which Thompson pulled down her sweat pants while she was working, in full view of other employees. 26 The Fifth Circuit held that once it "became clear to Wyatt that Gorum was not only ineffective in dealing with Thompson's harassment, but that he (Gorum) himself was a sexual harasser, Wyatt's failure to report either Thompson's or Gorum's behavior to one or more among the other individuals listed in the sexual harassment policy was unreasonable. 265 Accordingly, the court affirmed the trial court's order granting summary judgment to the employer for the period of time after which Gorum began his harassment of Wyatt.266 The problem with the decision in Wyatt is that the Fifth Circuit, on a motion for summary judgment, determined that Alisha Wyatt's conduct was unreasonable without acknowledging the ways in which victims respond to sexual harassment or explaining the standard of reasonableness against which it measured Wyatt's conduct. The Fifth Circuit faulted Wyatt for failing to report up the chain of management when Gorum not only failed to stop Thompson's harassment, but began harassing her himself. Yet, the empirical research shows that victims do not report harassment for various reasons, including their belief that nothing will be done or that reporting will make the work environment unpleasant.267 On a motion for summary judgment, it is at least arguable that Wyatt believed, after having had no success with her prior complaints of harassment to Gorum, that nothing would be done if she filed yet another complaint of harassment with a different manager. Moreover, she may have feared making the situation worse, given Gorum's "unofficial and informal admonitions not to 'go over his head.' 268 As the harassment escalated, however, culminating in a publicly humiliating scene, Wyatt may have had enough and decided to report Thompson's and Gorum's conduct. The research shows that as the "frequency and severity" of harassment increases, "so does the likelihood that the target will report the behavior., 269 The court downplayed the fact that Wyatt waited only about four and a half months from the time that Gorum began harassing her to report both Thompson's and Gorum's conduct to upper management.27 0 264Id. at 407. '65Id. at 413. 266 id. 267 See 1995 MSPB, supra note 74, at 35 tbl. I1 (stating that twenty-nine percent of federal government employees subjected to unwanted sexual attention did not take formal action because they "[t]hought it would make [their] work situation unpleasant" and twenty percent did not do so because they "[d]id not think anything would be done"). 268 297 F.3d at 413. 269 Knapp, supra note 72, at 695 (citing studies). 270 297 F.3d at 407, 413. 256 [Vol. 13:2 2004] Operating in an Empirical Vacuum 257 Alisha Wyatt responded in a manner typical of many harassment victims. The empirical data shows that some victims do not report harassment because they believe that reporting will make the situation worse. For example, they fear that reporting may cause the harassment to escalate, may adversely affect their career, or otherwise may "invite" retaliation.27' Plaintiffs commonly cite fear of retaliation to explain their reluctance to report harassment,272 and sexual harassment claims are frequently coupled with allegations of retaliatory conduct.273 271 See 1995 MSPB, supra note 74, at 35 tbl. I1 (showing that twenty-nine percent of federal government employees did not report harassing conduct because they believed it "would make [their] work situation unpleasant" and seventeen percent believed that doing so "would adversely affect [their] career"); Frazier, supra note 72, at B-16 (stating that "[a]nother reason for not reporting is that it would make the situation worse in some way: it would make the work environment unpleasant, the situation might escalate, or the woman would be blamed or experience some kind of retaliation"); Lawton, supra note 16, at 126 (summarizing results of General Accounting Office survey of the military academies, which found that "more than one-third of all women who had experienced harassment at one of the three military academies did not report the incident for fear of reprisal"). Research outside the field of sexual harassment indicates that fear of retaliation is one reason why employees do not use their employers' internal grievance appeals systems. See David Lewin, Dispute Resolution in the Nonunion Firm: A Theoretical and Empirical Analysis, 31 J. Conflict Resol. 465,500 (1987). 272 See, e.g., Barrett v. Applied Radiant Energv Corp., 240 F.3d 262, 267 (4th Cir. 2001); Dowdy v. North Carolina No. 01-1706, 2001 U.S. App. LEXIS 24382, at *4-5 (4th Cir. Nov. 13, 2001); Cardenas v. Massey, 269 F.3d 251, 259 (3d Cir. 2001); Leopold v. Baccarat, Inc., 239 F.3d 243, 246 (2d Cir. 2001); Leugers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *4 (6th Cir. Feb. 3, 2000); Greenhorn v. Marriott Int'l, Inc., 258 F. Supp. 2d 1249, 1252 (D. Kan. 2003); Arnold v. Yale New Haven Hosp., 213 F. Supp. 2d 142, 150 (D. Conn. 2002); Bennett v. Progressive Corp., 225 F. Supp. 2d 190, 209 (N.D.N.Y. 2002); Walton v. Johnson & Johnson Servs., Inc., 203 F. Supp. 2d 1312, 1324 (M.D. Fla. 2002), affd, 347 F.3d 1272 (11th Cir. 2003); Anderson v. Deluxe Homes, 131 F. Supp. 2d 637, 650 (M.D. Pa. 2001); Carl v. Parmely, 188 F. Supp. 2d 991, 1006 (S.D. Ill. 2001); see also Kortan v. Cal. Youth Auth., 217 F.3d 1104, 1117 (9th Cir. 2000) (Fisher, J., dissenting in part) (noting that majority had failed "to address the 'chilling effect which [a supervisor's] retaliatory conduct might have on the remaining employees under his supervision"' (quoting Hashimoto v. Dalton 118 F.3d 671, 676 (9th Cir. 1997)) and that the California Youth Authority's employees would "likely be deterred from reporting sexual harassment out of fear of retaliation"); cf Murray v. Chi. Transit Auth., 252 F.3d 880, 889 (7th Cir. 2001) (rejecting plaintiff's "claims that she feared further harassment if she reported her actions"); Curry v. District of Columbia, 195 F.3d 654, 666 n.2 (D.C. Cir. 1999) (Wald, J., concurring) (noting in co-worker harassment case that plaintiff did not complain at first about the harassment because "she did not think it would be good for her career" based on what she had seen happen to other female police officers), cert. denied, 530 U.S. 1215 (2000). 273 See, e.g., Mast v. IMCO Recycling of Ohio, Inc., 2003 U.S. App. LEXIS 1940, at *2 (6th Cir. Feb. 3, 2003) (hostile work environment and retaliation claims under Ohio's employment discrimination statute, and several state law claims); Kohler v. Inter-tel Techs, 244 F.3d 1167 (9th Cir. 2001) (sexual harassment and retaliation claims brought pursuant to California Fair Employment and Housing Act); Gawley v. Ind. Univ., 276 F.3d 301 (7th Cir. 2001) (sexual harassment, retaliation, and Indiana tort law claim); Murray, 252 F.3d at 883 (sexual harassment, retaliation, and § 1983 claims); La Day v. Catalyst Tech., Inc., 302 F.3d 258 Columbia Journal of Gender and Law [Vol. 13:2 Employee fears of retaliation are not limited to sexual harassment claims, however. In his study of "nonunion appeal systems in three large U.S.-based companies over the four-year period from 1980 to 1983, ' '274 Lewin found that twenty-seven percent of management and administrative workers and twenty-one percent of non-management personnel identified "fear of reprisal" as the reason for not using their employer's grievance procedure.275 Three quarters of management and administrative employees and two-thirds of non-supervisory workers, however, believed that they had a claim that merited filing of an appeal.276 Moreover, employee fears of retaliation are not baseless. Lewin found that employees who used their employers' complaint procedures had "significantly lower promotion rates and performance ratings and significantly higher turnover rates" than employees who had not invoked their employers' grievance machinery. 277 Bergman and her colleagues evaluated data collected from almost 6,500 military personnel and found that "reporting [harassment] often triggers retaliation. 278 Stockdale concluded that the results of her evaluation of the 1988 MSPB data "appeared to defy conventional wisdom that assertive response to sexual harassment will improve (or not worsen) outcomes for targets... [and] lend support to the fear that targets using such tactics may experience retaliation." 279 Using the same MSPB data, Hesson-Mcnnis and Fitzgerald reached conclusions similar to those of Stockdale. Contrary to conventional wisdom, assertive and formal responses were actually associated with more negative outcomes of every sort. Women who reported harassment to their supervisors or 474, 476 (5th Cir. 2002) (claims of same sex harassment, retaliation, intentional infliction of emotional distress, and battery); Scrivner v. Socorro Indep. Sch. Dist., 169 F.3d 969, 970 (5th Cir. 1999) (sexual harassment and retaliation claims); Hocevar v. Purdue Frederick Co., 223 F.3d 721, 723 (8th Cir. 2000) (sexual harassment and retaliation claims); Greenhorn, 258 F. Supp. 2d at 1252 (sexual harassment, retaliation, and various state law claims); Taylor II, 2002 U.S. Dist. LEXIS 14560, at *2 (N.D. Tex. Aug. 7, 2002) (sexual harassment, retaliation, and various state law claims), aff'd, No. 02-10950, 2003 U.S. App. LEXIS 13103 (5th Cir. June 3, 2003); Walton, 203 F. Supp. 2d at 1314 (sexual harassment, retaliation, and state law claims of negligent retention and intentional infliction of emotional distress), affd, 347 F.3d 1272 (1 1th Cir. 2003). 274 Lewin, supra note 271, at 466. 2751 d. at 499-500. 276 Id. at 500. 277 See id. at 498; see also Lawton, supra note 16, at 126-27 (describing, in greater detail, results of Lewin's study). 278 Bergman et al., supra note 73, at 237. 279 Margaret S. Stockdale, The Direct and Moderating Influences of Sexual- Harassment Pervasiveness, Coping Strategies, and Gender on Work-Related Outcomes, 22 Psychol. Women Q. 521,532 (1998). 2004] Operating in an Empirical Vacuum 259 who filed complaints were more likely to quit, be fired, or be transferred; to need or utilize medical and psychological assistance; to feel worse about their jobs; and so forth.280 Yet a number of federal courts, particularly in the Second and Fourth Circuits, have held, in cases ranging from no report of harassment 281 to delayed reports 282 to reports to managers not designated in the employer's policy,283 that an employee's "generalized" or "speculative" fear of retaliation does not excuse her failure to report the harassment using her employer's internal grievance procedure. 284 Given the empirical research showing that few victims of sexual harassment file formal complaints and 280 Hesson-Mclanis & Fitzgerald, supra note 117, at 896. 281 See, e.g., Matvia v. Bald Head Island Mgmt., Inc., 259 F.3d 261, 269 (4th Cir. 2001) (affirming district court order granting employer's motion for summary judgment and concluding that plaintiffs fears of retaliation by co-workers did not excuse her failure to report her supervisor's harassment); Barrett v. Applied Radiant Enery Corp., 240 F.3d 262, 267 (4th Cir. 2001) (affirming trial court order granting judgment notwithstanding the verdict to defendant employer and rejecting, as speculative, plaintiffs argument that she feared retaliation); Leopold v. Baccarat, Inc., 239 F.3d 243, 246 (2d Cir. 2001) (affirming trial court order granting employer motion for summary judgment and finding that plaintiff's fear of being fired for speaking up was not credible and based only on her subjective beliefs); Breeding v. Cendant Corp., No. 01 Civ. 11563, 2003 U.S. Dist. LEXIS 6558, at *21 (S.D.N.Y. Apr. 17, 2003) (granting employer motion for summary judgment and rejecting as "conclusory" plaintiffs claims that she had not reported the harassment because she feared being fired or that doing so would limit her career opportunities). 282 See, e.g., Leugers v. Pinkerton Sec. & Investigative Servs., No. 98-3501, 2000 U.S. App. LEXIS 1831, at *9 (6th Cir. Feb. 3, 2000) (affirming district court order granting summary judgment to employer on hostile work environment claim, concluding that plaintiff delayed reporting only in part out of fear of reprisal); McCombs v. Chrysler Corp., No. IP 99-0697-C-T/K, 2003 U.S. Dist. LEXIS 6448, at *22-24 (S.D. Ind. Mar. 4, 2003) (granting employer's motion for summary judgment and finding that confidentiality and anti- retaliation provisions in employer's policy should have allayed plaintiff's fear of reporting); Arnold v. Yale New Haven Hosp., 213 F. Supp. 2d 142, 150 (D. Conn. 2002) (granting employer's motion for summary judgment and concluding that plaintiff had not offered sufficient evidence to show that his delay in reporting arose from a "credible fear that he would be subjected to an adverse employment action"); Walton v. Johnson & Johnson Servs. Inc., 203 F. Supp. 2d 1312, 1324 (M.D. Fla. 2002) (granting employer's motion for summary judgment and concluding that plaintiff had "failed to articulate a specific factual basis for being in fear of losing her job"), afJd, 347 F.3d 1272 (11 th Cir. 2003). 283 Dowdy v. North Carolina, No. 01-1706, 2001 U.S. App. LEXIS 24382, at *3-5 (4th Cir. Nov. 13, 2001) (affirming trial court order granting employer motion for summary judgment, explaining that plaintiffs generalized fear of retaliation did not excuse her failure to follow her employer's policy and report the harassment to the persons designated therein); Watkins v. Prof'1 Sec. Bureau, Ltd., No. 98-2555, 1999 U.S. App. LEXIS 29841, at *19-20 (4th Cir. 1999) (affirming trial court order granting judgment as a matter of law to defendant employer and concluding that plaintiff's "embarrassment and fear of reprisal" did not excuse her failure to report "promptly and fully" according to her employer's policy), cert. denied, 529 U.S. 1108 (2000). 284 Barrett, 240 F.3d at 267. Columbia Journal of Gender and Law that retaliation is a not uncommon response to formal reporting, the conclusion that a plaintiff acts unreasonably by not reporting harassment is certainly suspect. 4. Shifting the Burden of Proof on Prong Two What evidence must an employer produce in order to show compliance with prong two of the affirmative defense, other than demonstrating that it has exercised reasonable care to prevent and correct harassment? Suppose an employer not only creates and widely distributes its policy and procedure, e.g., on its web site, but also undertakes steps to prevent workplace harassment by integrating its workforce and creating a professional work environment in which sexual harassment is swiftly and severely punished. The employer conducts annual surveys of its employees in order to evaluate how its policy and procedure work in practice, maintains records and follows up with complainants to ensure no retaliation, and evaluates supervisors on commitment to a harassment-free workplace. What more, besides an employee's mere failure to use this employer's grievance procedure, would the employer have to show in order to satisfy prong two of the affirmative defense? In other words, did the Supreme Court intend that the employer's exercise of reasonable care in prong one also satisfies its burden of proof on prong two? 285 If that were the case, then why did the Court use the word "unreasonably" in describing the employee's conduct under prong two? Were the lower federal courts to assume that an employee acts unreasonably by not invoking a grievance procedure that the employer has shown is reasonable, proven, and effective? One answer may be that in cases in which the plaintiff employee completely fails to use the employer's reasonable, proven, and effective procedures, then the employer has no further burden on prong two than showing the plaintiffs decision not to complain internally. But in cases of "defayed" reporting, the employer bears the burden of demonstrating that the "delay" was unreasonable. Thus, the employee's decision to wait to file an internal complaint would require an examination of the circumstances surrounding her decision, such as the egregiousness and frequency of the conduct, coupled with any efforts she made to directly confront the harasser. The lower federal courts, however, are not requiring employers to demonstrate that their anti-harassment policies and procedures are reasonable, proven, and effective.28 6 Instead, file cabinet compliance is 285 See Faragher v. City of Boca Raton, 524 U.S. 775, 807-08 (1998) (stating that "while proof that an employee failed to fulfill the corresponding obligation of reasonable care to avoid harm is not limited to showing an unreasonable failure to use any complaint procedure provided by the employer, a demonstration of such failure will normally suffice to satisfy the employer's burden under the second element of the defense"). 286 See supra notes 100-105 and accompanying text. 260 [Vol. 13:2 2004] Operating in an Empirical Vacuum 261 commonplace. That means that in cases in which the plaintiff does not use the employer's procedures at all, the employer must offer evidence typically of nothing more than the following on prong two: (1) distribution of an anti- harassment policy and a grievance procedure allowing an employee to bypass a harassing supervisor, (2) provision of training, in some cases, and (3) the plaintiff's failure to complain.287 This shifts the burden of proof from the employer to the employee on prong two's requirement of reasonableness. Rather than requiring the employer to provide post- complaint data on complainants, for example, to determine whether retaliation occurred, the plaintiff employee must come forward with such 287 See, e.g., Matvia v. Bald Head Mgmt., Inc., 259 F.3d 261, 265, 268-70 (4th Cir. 2001) (affirming trial court order granting employer's motion for summary judgment, holding that employer was entitled to the affirmative defense because employer had disseminated an anti-harassment policy, which was "compelling proof' of its exercise of reasonable care, and plaintiff had not complained, the harassment coming to the attention of management from the alleged harasser); Kohler v. Inter-Tel Techs., 244 F.3d 1167, 1179-82 (9th Cir. 2001) (affirming trial court order granting summary judgment to employer, holding that the affirmative defense applies to California employment discrimination cases, and finding that employer satisfied the affirmative defense because it had disseminated its policy and procedure, conducted an "exemplary" investigation once the EEOC notified it of plaintiff's charges, and plaintiff had not complained to her employer); Leopold v. Baccarat, Inc., 239 F.3d 243, 245-46 (2d Cir. 2001) (affirming trial court order granting summary judgment to employer on affirmative defense and holding that employer had an anti- harassment policy and complaint procedure, even though neither guaranteed confidentiality or prohibited retaliation, and plaintiff had failed to use the employer's grievance machinery); Shaw v. AutoZone, Inc., 180 F.3d 806, 811-13 (7th Cir. 1999) (affirming trial court order granting summary judgment to employer, holding that employer had satisfied the affirmative defense because it had distributed an anti-harassment policy to its employees and provided training on its policy to managers, and plaintiff had failed to report the harassment, even though she was not isolated from upper management to whom complaints could be made), cert. denied, 528 U.S. 1076 (2000); Barrett, 240 F.3d at 266-68 (affirming trial court order granting employer's motion for judgment notwithstanding the verdict based on the employer's satisfaction of the affirmative defense, explaining that the employer's policy was "compelling proof' that it had met prong one, and concluding that plaintiff had offered no evidence that the policy was defective or dysfunctional, and had not complained to any of the firm's twelve managers, even though many worked in the same building); Breeding v. Cendant Corp., No. 01 Civ 11563, 2003 U.S. Dist. LEXIS 6558, at *18-19 (S.D.N.Y. Apr. 17, 2003) (granting employer's motion for summary judgment, holding that employer was entitled to the affirmative defense based on its sexual harassment policy and plaintiff's failure to invoke employer's grievance procedure); Duviella v. Counseling Serv., No. 00- CV-2424, 2001 U.S. Dist. LEXIS 22538, at *35-51 (E.D.N.Y. Nov. 20, 2001) (granting employer's motion for summary judgment, holding that it was entitled to the affirmative defense because (1) it had distributed an anti-harassment policy and procedure, which was not ineffective simply because it required a signed and written statement before an investigation would commence, (2) it had investigated plaintiffs complaint, made in her letter of resignation, and (3) plaintiff had not complained before resigning because her conversations with the site director did not indicate that she was complaining about sexual harassment), aff'd, No. 02-7019, 2002 U.S. App. LEXIS 24050 (2d Cir. Nov. 21, 2002). Columbia Journal of Gender and Law [Vol. 13:2 evidence, or otherwise demonstrate that her decision not to complain was reasonable.288 The story is no different in cases in which the plaintiff "delays" in reporting the harassment. Once the employer satisfies prong one by offering evidence that it has disseminated a policy and procedure and promptly corrected the complained-of behavior, some federal courts presume that any delay by the plaintiff in reporting is unreasonable.2 89 As a result, the party 288 See, e.g., Matvia, 259 F.3d at 265, 269-70 (rejecting plaintiff's argument that her failure to complain was reasonable because she needed time to collect evidence and she feared retaliation by co-workers); Barrett, 240 F.3d at 267-68 (rejecting plaintiffs contention that she did not report because she believed she would not be taken seriously and feared retaliation); Leopold, 239 F.3d at 245-46 (rejecting plaintiff's argument that she did not report out of fear of losing her job); Shaw, 180 F.3d at 811-13 (rejecting plaintiff's claim that her decision not to report was reasonable because she did not feel comfortable enough to discuss the harassment with someone from the company); Duviella, 2001 U.S. Dist. LEXIS 22538, at *49-50 (rejecting plaintiff's argument that she did not report because the harasser was "immune from charges of harassment" as the employer's executive director was the harasser's "protectorate"). 289 Hill v. Am. Gen. Fin., Inc., 218 F.3d 639, 643-44 (7th Cir. 2000) (affirming district court order granting summary judgment to employer on plaintiff's claim of sex and race harassment, noting that even though employer's policies left "room for improvement," the employer "took immediate corrective action" once it learned of the harassment, had posted "in a 'public access type place"' its anti-harassment policies, and while plaintiff's earlier anonymous letter complaining about the harasser's conduct indicated her apprehension about complaining, "apprehension [did] not eliminate the requirement that the employee report harassment"), reh'g en banc denied, No. 99-2682, 2000 U.S. App. LEXIS 18367 (7th Cir. July 27, 2000); Leugers v. Pinkerton Sec. & Investigative Servs., No. 98- 3501, 2000 U.S. App. LEXIS 1831, at *7-9 (6th Cir. Feb. 3, 2000) (affirming district court order granting employer's motion for summary judgment on hostile work environment claim, concluding that employer had distributed a sexual harassment policy with a grievance procedure allowing bypass of a harassing supervisor and had terminated the harasser once plaintiff complained, and faulting plaintiff for waiting six months to complain, even though plaintiff waited to report, in part, out of fear of retaliation); Casiano v. AT&T Cori., 213 F.3d 278, 286-87 (5th Cir. 2000) (affirming district court order granting summary judgment to employer on plaintiff's sexual harassment claim, concluding that AT&T could satisfy the affirmative defense because AT&T had several publications articulating its anti-harassment policy, plaintiff was aware of those policies, AT&T acted promptly once plaintiff complained of harassment, and plaintiff did not complain, despite being propositioned at least fifteen times, until "well after the harassment" had ended); Wright v. Anixter, Inc., No. 98-17164, 1999 U.S. App. LEXIS 19962, at *4-6 (9th Cir. Aug. 18, 1999) (affirming trial court order granting employer's motion for summary judgment on basis of affirmative defense, explaining that employer had a sexual harassment policy with multiple avenues to complain, which was posted, the employer responded promptly to plaintiff's complaint, but plaintiff "chose not to take advantage of the complaint procedure earlier" because "she [did] not like 'confrontation'); McPherson v. City of Waukegan, No. 01 C 9264, 2003 U.S. Dist. LEXIS 9098, at *19-23 (N.D. Ill. May 30, 2003) (granting employer's motion for summary judgment on basis of affirmative defense because employer had a sexual harassment policy and grievance procedure, promptly corrected the harassment, and plaintiff, who failed to complain about harasser's offensive comments for two years before it escalated to sexual assault, did not "dispute Waukegan's assertion that it [could] meet the second Ellerth element"); Walton v. Johnson & Johnson Servs., Inc., 203 F. Supp. 2d 1312, 1322-25 (M.D. Fla. 2002) (granting employer's motion for summary judgment, concluding that employer 2004] Operating in an Empirical Vacuum 263 with the burden of proof does not have to produce any evidence of the plaintiff's pre-complaint unreasonableness in order to prevail on a motion for summary judgment. Instead, the party without the burden of proof cannot get past summary judgment without evidence that she acted reasonably---evidence such as the company's inaction in the face of prior complaints of harassment, which may not even exist.2 ° As with prong one, shifting the burden of proving reasonableness to the plaintiff is inefficient. The employer controls the complaint procedure. If a plaintiff fears retaliation and therefore delays reporting, she must produce evidence that her fears of reprisal are justified.29' Her own allegations that she suffered retaliation for reporting do not suffice because post-complaint retaliation directed at her would not explain her pre- complaint fear of retaliation. Therefore, the plaintiff must produce testimony from other employees or records from her employer showing that retaliation results from reporting. The problem is that employers are not required to keep such records.292 As a result, it is not clear how the courts expect a plaintiff to substantiate her fear of reprisal. In addition, employers are not required to conduct any post- complaint follow-up with employees who use the employer's internal grievance procedure. Yet, measuring the effectiveness of a grievance had a policy and procedure, provided training, and promptly corrected the harassment by immediately suspending the harasser, and rejecting plaintiffs claim that she waited three months from the first incident of harassment to report because she was afraid she would not get a promotion), affid, 347 F.3d 1272 (1 1th Cir. 2003); Green v. Wills Group, Inc., 161 F. Supp. 2d 618, 625-26 (D. Md. 2001) (granting employer's motion for summary judgment, finding that employer had distributed an anti-harassment policy, and that "[p]laintiff [had] not pursue[d] the available avenues of corrective action in a timely manner and when she [had], prompt action resulted"); Slay v. Glickman, 137 F. Supp. 2d 743, 752 (S.D. Miss. 2001) (affirming trial court order granting summary judgment to employer on hostile work environment claim, concluding employer was entitled to the affirmative defense because it had a sexual harassment policy and had promptly responded once plaintiff complained, but that plaintiff waited six months from sole sexual proposition to report, even though she was a member of the Farmers Home Administration's EEO Committee); see also Frederick v. Sprint/United Mgmt. Co., 246 F.3d 1305, 1314 (11th Cir. 2001) (citing Madray v. Publix Supermarkets, Inc., 208 F.3d 1290, 1302 (11th Cir. 2000), cert. denied, 531 U.S. 926 (2000)) (stating that "[a]s to the second element of the defense, an employer's showing that the plaintiff-employee failed to follow its complaint procedures will often be sufficient [to] satisfy its burden"). 290 See Pennsylvania State Police v. Suders, 124 S.Ct. 2342, 2354 n.7 (2004) (stating that the employer bears the burden of proving whether its employee acted unreasonably in failing to use the employer's grievance machinery because "[tihe employer is in the best position to know what remedial procedures it offers to employees and how those procedures operate"); see also supra note 93 and accompanying text. 291 See supra notes 281-284 and accompanying text. 292 29 C.F.R. §1602.12 (2002). Columbia Journal of Gender and Law procedure without examining post-complaint consequences fails to capture the impact that retaliation can have on subsequent reporting behavior.293 Suppose Marta works for a large firm.294 Her employer distributes to each new employee a sexual harassment policy and accompanying grievance procedure, which informs employees that they may file complaints of harassment with any manager or a representative from human resources. The policy ensures confidentiality of complaints and prohibits retaliation. The firm also requires employees to attend a short twenty- minute video on sexual harassment when they are first hired. Marta files a formal complaint with the firm's human resources division, claiming that Andrew, her supervisor, has been sexually harassing her. While Andrew began harassing Marta ten months before she complained, she waited to report the harassment because, at first, the behavior was isolated or hard to pin down, e.g., an offensive comment or lascivious stare. As the behavior became more severe, e.g., improper touching, Marta hesitated to complain knowing that Andrew conducted her annual performance evaluations. Marta told her friends of Andrew's behavior, but did not immediately file a formal report with her employer. When the behavior does not stop, however, Marta relents and files a formal report with the human resources division. The firm quickly investigates, but because the conduct occurred when there were no witnesses around, the company is unable to substantiate Marta's claim. 295 Nonetheless, it cautions Andrew about improper workplace conduct, reminding him of the company's sexual harassment policy and its prohibition against retaliation. Six months later, Marta is due for her annual evaluation.296 Andrew completes her evaluation, rating her as "needs improvement" in three categories. Marta, however, has never received a rating of "needs improvement" before on an annual evaluation. She has. consistently earned ratings of "excellent," once or twice getting a "meets expectations" during 293 See Lewin, supra note 271, at 498 (concluding that "[t]he weakness of the due process perspective on workplace dispute resolution . . . is that it ignores postdispute resolution consequences or outcomes"). 294 The following example is loosely based on the facts of Kortan v. Cal. Youth Auth., 217 F.3d 1104 (9th Cir. 2000), in particular, the dissenting opinion of Judge Fisher. 295 In Kortan, the employer took three months to complete its investigation, and concluded that the harassment charges were "unsubstantiated." Kortan left the job about two and a half weeks later, was hospitalized, and was on a leave of absence when the Ninth Circuit heard her appeal of the district court's order granting summary judgment to the employer on her claims of sex discrimination, sexual harassment, and retaliation. See id. at 1108. 296 Kortan's evaluation took place while the investigation of her sexual harassment complaint was pending. See id. at 1107. [Vol. 13:2 Operating in an Empirical Vacuum her six-year tenure with the firm.2 97 In fact, two years before, she had received the employer's Outstanding Employee of the Year award. 298 Marta is ambitious and is concerned about the evaluation's impact on her chances for subsequent promotion. She complains about the evaluation, and another supervisor independently reviews Andrew's evaluation. The second supervisor decides to split the difference, increasing Marta's three scores for "needs improvement" to "meets expectations. 2 99 Marta contemplates filing suit against her employer for sexual harassment and retaliation. She faces two significant hurdles. The first is the fact that her employer has an anti-harassment policy and procedure that satisfies the minimum requirements for prevention imposed by most federal courts. Her employer also provides some training on sexual harassment. In addition, her employer responded quickly once she filed a complaint. This means that Marta now must come forward with evidence as to why she waited ten months to report Andrew's behavior. Her subjective fear of reprisal does not suffice. But her employer likely does not maintain records on what happens to sexual harassment complainants. As a result, she likely will lose on her hostile work environment claim, quite possibly at the summary judgment stage. The second hurdle confronting Marta relates to her retaliation claim. "Courts are split on whether unfairly evaluating performance alone can trigger retaliation. ' '3° If the unfair evaluation Marta received does not constitute an adverse employment action, she cannot make out a primafacie case of retaliation under Title VII. Even if the court considers the downgrade as a possible adverse employment action, it may conclude that the independent evaluation by another supervisor "fixed" the problem; after all, Marta was not fired, demoted, or denied a raise based on the 297 During her five-year tenure with the California Youth Authority, Kortan received only one average rating (an "M") and that occurred during her first year on the job in only one performance category. Id. at 1116 (Fisher, J., dissenting). 298 Kortan actually received "'Outstanding Employee of the Year' for 1993 (the evaluation period in question)..." Id. 299 In Kortan, Judge Fisher concluded that Schulman, the supervisor who conducted an independent review of Kortan's evaluation, "merely struck a compromise between the highest ("E") rating and [the harasser's] unwarranted "I" ratings - moving Kortan up one rating to "M" [an average rating]." See id. 300 Barbara Lindemann & Paul Grossman, Employment Discrimination Law 505 (3d ed. Supp. 2002) (footnote omitted) (citing cases); see Davis v. Town of Lake Park, 245 F.3d 1232, 1241 (1 1th Cir. 2001) (citing cases concluding that "criticisms of an employee's job performance-written or oral-that do not lead to tangible job consequences will rarely form a permissible predicate for a Title VII suit"); Anne Lawton, The Meritocracy Myth and the Illusion of Equal Employment Opportunity, 85 Minn. L. Rev. 587, 660 nn.328-32 (2000) (discussing Fifth Circuit's requirement that plaintiff show an ultimate employment decision in order to state an actionable claim for retaliation under Title VII). 2004] 265 Columbia Journal of Gender and Law evaluation. 30 ' The problem, however, is that the "revised evaluation largely relied upon [the] tainted evaluation and left [Marta] with the lowest performance rating of her career. ' 3 2 Either way, without evidence of an adverse employment action, Marta may be unable to make out a prima facie case of retaliation pursuant to Title VII. Marta decides not to file suit against her employer. She obtains other employment and quits. The message that is communicated to her co- workers is that reporting harassment is a risky endeavor. 30 3 The harassment may end, only to be replaced by career-damaging retaliation. Potential harassers will feel safe in continuing to harass because doing so is unlikely to result in formal reporting and employer sanctions. By failing to protect Marta from retaliation, the firm sends a message to employees about the organization's culture, a message that the affirmative defense, as currently interpreted by the federal courts, fails to capture. IV. CONCLUSION AND RECOMMENDATIONS As a vehicle for reducing the incidence of harassment by supervisors in the workplace, the affirmative defense is flawed. Its success rests, in part, on two assumptions-that victims report harassing behavior and that anti-harassment policies and procedures deter workplace harassment-neither of which is adequately supported at the current time by the empirical literature. Despite these shortcomings, it is still possible to interpret the affirmative defense so as to create a workable system of incentives for employers. In order to do so, however, the lower federal courts must move toward a more complex analysis of employer compliance, examining how the employer's anti-harassment policy and grievance procedure actually work in practice. [L]egal scrutiny might better focus on the organization's responsibility to create a climate in which victims have no reason to fear reporting mechanisms or their aftermath. That is, courts should require that the organization demonstrate a climate intolerant of sexual harassment through a documented history of taking complaints seriously, protecting complainants from retaliation, and holding perpetrators responsible for their actions. 30 4 Therefore, taking into account an employer's size and resources, the lower federal courts, in evaluating the two prongs of the affirmative defense, "See Kortan, 217 F.3d at 1112. 302 See id. at 11 14 (Fisher, J., dissenting). 303 See id. 3 04 Bergman et al., supra note 73, at 241. 266 [Vol. 13:2 Operating in an Empirical Vacuum should require employers to produce evidence of the following: (1) complaint records; (2) a system of post-complaint follow-up; (3) employees' evaluations of the employer's policy and procedure; and (4) a system for evaluating managers on their compliance with the firm's anti- harassment policy and procedure. In cases involving claims of harassment in a highly sex-segregated workforce, the courts should consider holding employers to a higher level of prevention. A. Complaint Records Without complaint records, a court cannot evaluate the effectiveness of an employer's procedure. For example, if an employer rarely sanctions harassers, even for egregious misconduct, employees may be reticent to invoke the employer's grievance machinery. If the court does not have access to records of prior complaints and how the employer disposed of those complaints, it cannot evaluate the reasonableness of an employee's decision not to use the employer's grievance procedure. Moreover, if the employer does not maintain such records, the employee cannot obtain them during discovery, making it difficult to rebut the employer's claim that the employee's failure to report or delay in reporting the harassment was unreasonable. Therefore, the courts should not allow an employer to prevail on the affirmative defense if the employer fails to maintain records of prior harassment complaints. These records should indicate the nature of the complaint, e.g., gender harassment versus improper touching, the length of the investigation, and the disposition of the complaint, e.g., whether sanctions were imposed and the nature of those sanctions. B. A System of Post-complaint Follow-up Courts currently impose no obligation on employers to follow up with victims; warning harassers that they are not to retaliate suffices to discharge any further obligation on the part of the employer. But it is not at all unlikely that a supervisor will try to retaliate against an employee who files a harassment complaint.305 Therefore, it makes sense to require the employer to follow up with the complaining employee periodically after the termination of the employer's grievance procedure. Doing so accomplishes two things. First, it puts the supervisor on notice that the employer is actively checking for retaliation and, therefore, might dissuade some supervisors from taking retaliatory action. Second, it places on the employer, not the employee, the burden of following up. The research shows that most victims of harassment do not file formal complaints. A system that affirmatively places on the employer the obligation to follow up 305 See supra notes 294-302 and accompanying text. 2004] 267 Columbia Journal of Gender and Law with complainants might make the employer's promise of protection from retaliation more credible. As part of the post-complaint system, employers also should maintain records of the current status of complainants, for example, whether they were fired, have quit, or have been promoted. Similar records should be kept on supervisors when the firm has investigated and verified a harassment complaint. Having this information is critical to an evaluation of the reasonableness of plaintiffs' reporting decisions. If reporting results in adverse consequences for victims, then employees will not use their employer's grievance machinery. Similarly, if harassers are rewarded, e.g., with promotions, instead of sanctioned for harassing behavior, then employees will view reporting as a futile endeavor. C. Employees' Evaluations of Employer Policies and Procedures It is surprising that the federal courts allow an employer to satisfy its prevention obligation under prong one of the affirmative defense without producing any evaluation of the employer's own policies and procedures. The policies and procedures are supposedly designed for use by employees. If employees do not know how to use the procedure, do not know to whom to complain, or consider reporting a risky proposition, then an employer should not be able to satisfy its obligation under prong one to prevent and correct harassment simply because it has a policy "on the books." D. A System of Management Evaluation If the employer does not evaluate and reward managers on their compliance with the firm's anti-harassment policies and procedures, then managers will not consider the policy or the reduction of harassment to be important firm goals. Suppose Supervisor A fails to pass along to appropriate personnel complaints of sexual harassment made to him about Supervisor B. 3°6 If Supervisor A receives a hefty pay raise notwithstanding this failure, the firm communicates to him and to other employees that supervisors are not accountable for compliance with the firm's anti- harassment policies and procedures. By doing so, the firm creates a culture in which reporting harassment is not a high priority, and employees quickly learn that management is not interested in knowing when harassment occurs. 306 See supra notes 207-242 and accompanying text for discussion of Madray v. Publix Supermarkets, Inc., 208 F.3d 1290 (11 th Cir. 2000). [Vol. 13:2 Operating in an Empirical Vacuum E. Higher Standard of Prevention for Sex-segregated Workplaces In hostile work environment cases involving highly sex-segregated workplaces, the courts should consider holding employers to a higher standard of prevention. For example, a court might want to examine an employer's affirmative action program or its efforts, if any, to reduce sex segregation within certain jobs. In such cases, evidence of a history of exclusion and sex discrimination are powerful indications that an environment ripe for sexual harassment exists within the organization. At a minimum, given the empirical research on the impact of male-dominated work environments, courts should be more vigilant in policing employers' anti-harassment efforts in hostile work environment cases involving "male bastions. 30 7 These suggestions do not correct the flaws inherent in the affirmative defense. But implementing them will accomplish at least three goals. First, as the Supreme Court intended, employers will bear the burden of proof on the affirmative defense. This, in turn, will increase plaintiffs' chances of surviving employer motions for summary judgment based on the affirmative defense. Second, requiring employers to provide complaint records, evidence of post-complaint follow-up, employee evaluations of procedures, and evaluations of managers on policy compliance ensures a more fact-intensive and nuanced approach to the reasonableness inquiries of prongs one and two of the affirmative defense. Finally, implementing these recommendations should alter employer incentives, rewarding conduct that bears a closer connection to the employer's efforts to eliminate workplace harassment. 307 See supra notes 165-166 and accompanying text, and supra Part III.A.2.c. 2004] 269 270 Columbia Journal of Gender and Law [Vol. 13:2 TABLE I (A) (B) (C) (D) (E) CIRCUIT TOTAL OTHER OTHER CW TOTAL CASES HARASSMENT CLAIMS HARASSMENT SUPERVISORY HARASSMENT CASES D.C. Circuit 2 0 0 1 1 D. Courts 3 1 0 0 2 D.C. Cir. Total 5 1 0 1 3 1st Circuit 5 0 0 1 4 D. Courts 16 3 2 2 9 1st Cir. Total 21 3 2 3 13 2d Circuit 15 1 1 0 13 D. Courts 51 16 6 8 21 2d Cir. Total 66 17 7 8 34 3d Circuit 7 3 0 1 3 D. Courts 26 6 2 1 17 3d Cir. Total 33 9 2 2 20 4th Circuit 26 2 1 1 22 D. Courts 11 3 1 0 7 4th Cir. Total 37 5 2 1 29 5th Circuit 19 4 3308 1 11 D. Courts 31 8 2 1 20 5th Cir. Total 50 12 5 2 31 6th Circuit 22 7 0 3 12 D. Courts 18 6 0 3 9309 6th Cir. Total 40 13 0 6 21 308 One of these cases involves claims of sexual harassment, but the court applied pre-Ellerth liability standards. 309 Includes Good v. MMR Group, Inc., No. 3:00-CV-182-H, 2001 U.S. Dist. LEXIS 20036, at *13 (W.D. Ky. Dec. 4, 2001), applying the affirmative defense in a co- worker sexual harassment case. 2004] Operating in an Empirical Vacuum (A) (B) (C) (D) (E) CIRCUIT TOTAL OTHER OTHER CW TOTAL CASES HARASSMENT CLAIMS HARASSMENT SUPERVISORY HARASSMENT CASES 7th Circuit 24 6 1 4 13 D. Courts 56 13 1 10 32 7th Cir. Total 80 19 2 14 45 8th Circuit 19 0 1 2 16 D. Courts 15 1 0 4 10 8th Cir. Total 34 1 1 6 26 9th Circuit 19 2 2 2 13 D. Courts 8 2 0 1 5 9th Cir. Total 27 4 2 3 18 10th Circuit 16 2 4 1 9 D. Courts 17 6 1 3 7 10th Cir. Total 33 8 5 4 16 1 1th Circuit 11 1 1 0 9 D. Courts 15 4 2 2 7 llth Cir. Total 26 5 3 2 16 Circuit Total 185 28 14 17 126 D. Court Total 267 69 17 35 146 TOTAL 452 97 31 52 272 Columbia Journal of Gender and Law [Vol. 13:2 TABLE II (A) (B) (C) (D) (E) CIRCUIT TOTAL PRIMA AD NOT TEA31 0 TOTAL SUPERVISORY FACIE RAISED AD HARASSMENT CASE OR CASES CASES DISCUSSED D.C. Circuit 1 0 0 0 1 D. Courts 2 1 1 0 0 D.C. Cir. Total 3 1 1 0 1 1st Circuit 4 0 0 0 4 D. Courts 9 0 1 1 7 1st Cir. Total 13 0 1 1 11 2d Circuit 13 2 3 1 7 D. Courts 21 1 4 0 16 2d Cir. Total 34 3 7 1 23 3d Circuit 3 0 0 2 1 D. Courts 17 1 1 0 15 3d Cir. Total 20 1 1 2 16 4th Circuit 22 0 0 0 22 D. Courts 7 0 1 0 6 4th Cir. Total 29 0 1 0 28 5th Circuit 11 1 1 2 7 D. Courts 20 3 3 1 13 5th Cir. Total 31 4 4 3 20 6th Circuit 12 1 1 0 10 D. Courts 9 1 0 0 8 6th Cir. Total 21 2 1 0 18 7th Circuit 13 0 2 1 10 D. Courts 32 5 2 3 22 7th Cir. Total 45 5 4 4 32 See supra note 11, and310 TEA is an acronym for a tangible employment action. notes 47-50 and accompanying text. 2004] Operating in an Empirical Vacuum (A) (B) (C) (D) (E) CIRCUIT TOTAL PRIMA AD NOT TEA TOTAL SUPERVISORY FAciE RAISED AD HARASSMENT CASE OR CASES CASES DISCUSSED 8th Circuit 16 2 2 1 11 D. Courts 10 0 0 3 7 8th Cir. Total 26 2 2 4 18 9th Circuit 13 2 1 0 10 D. Courts 5 1 0 0 4 9th Cir. Total 18 3 1 0 14 10th Circuit 9 1 1 1 6 D. Courts 7 1 1 1 4 10th Cir. Total 16 2 2 2 10 1 1th Circuit 9 1 2 2 4 D. Courts 7 0 1 1 5 11th Cir. Total 16 1 3 3 9 Circuit Total 126 10 13 10 93 D. Court Total 146 14 15 10 107 TOTAL 272 24 28 20 200