IN a recent article, Professor Bartlett argues that modifying legal tools in order to reduce implicit race and gender bias is a worthy goal, but one that is almost certainly unattainable.1 The modern workplace, in her view, is populated mostly by individuals who are well intentioned and committed to nondiscrimination. Legal "coercion" threatens their autonomy and sense of personal efficacy and, as a result, generates a backlash that is more likely to increase workplace bias than to reduce it. Instead of strengthening antidiscrimination laws, the most effective way to create workplace fairness is to restructure work in a way that facilitates our tolerant instincts. Doing so will reduce actions that are discriminatory and unlawful. At the same time, it will make us feel better about ourselves.2
I. IMPLICIT BIAS AND THE "COGNITIVE TURN"
I agree with Professor Bartlett that strengthening laws in order to reduce "hidden" bias is a bad idea. But to me (and no doubt to other organizational sociologists), the flaw in the legislative approach is that legal scholars, litigators, human resource professionals, and diversity consultants have become so enamored with the notion of ubiquitous unconscious, implicit, or hidden bias that they are quick to attribute systemic workplace racial and gender inequality to what is going on in people's heads. Instead, it is vital to consider what is built into organizational structures, processes, and routines.
The "cognitive turn" in workplace bias discourse is reflected in the business press, in professional human resources discourse, in legal scholarship, and in the legal strategies embraced in employment discrimination litigation. For example, "The War Over Unconscious Bias," a 2007 Fortune article about the Dukes et al. v. Wal-Mart gender class action litigation,3 proclaimed that "the problem isn't [Wal-Mart's] policies, it's their managers' unwitting preferences."4 A year earlier, Business Week framed the issue similarly in an article titled "White Men Can't Help It."5
The Society for Human Resource Management's website contains numerous links to articles about the dangers of unconscious bias and prescriptions for how to identify and eliminate it.6 In legal circles, the National Employment Lawyers Association has sponsored seminars for plaintiffs' attorneys to educate them on how expert testimony on the science of implicit bias can be used in litigation, and their management-side opponents offer seminars on how to counter it.7 Legal battles over the application of implicit bias testimony in litigation draw upon writings by legal scholars on the topic, providing fodder for further scholarly analysis in leading law reviews.8
In short, business and legal professionals today usually attempt to understand workplace discrimination in terms of processes and mechanisms that are cognitive and psychological, rather than organizational and institutional.9 Thus it is not surprising that proposals for legal reforms to combat "invisible, deep, and pervasive" biases10 focus on changing what happens in peoples' heads rather than identifying and remedying the subtle and hidden features of organizational policies, practices, and structures that create and sustain unlawful racial and gender inequality at work. The irony revealed by Professor Bartlett's cogent analysis is that those who view discrimination in contemporary workplaces as grounded in psychological processes of implicit bias offer remedies that are based on a shallow reading of the relevant psychological science. By ignoring motivational issues, the solutions they propose are as likely to intensify cognitive bias as to counteract it.
II. MOTIVATING NONDISCRIMINATION
Professor Bartlett's solution for minimizing implicit bias in organizational settings assumes that organizations can create strong norms of nondiscrimination and that their employees can and will internalize those norms. The organizations' employees (including those responsible for hiring, job assignment, pay, training, and promotion decisions; as well as those who control access to and distribution of valuable resources) will act in a nondiscriminatory manner not because of organizational policy directives and oversight, but because they are motivated to "do the right thing." Because they are internally rather than externally motivated, their actions will reinforce their personal commitments to nondiscrimination norms. Furthermore, by modeling moral behavior for others, they will re-educate colleagues who come to the organization with less enlightened views about race, gender, and nondiscrimination.11
Professor Bartlett's proposals would intervene into processes that operate at the organizational, work group, and individual levels; and their effectiveness assumes and requires strong linkages within and across levels of social organization. Implicit bias will remain unchecked if the organization only nominally embraces nondiscrimination goals, if those goals are decoupled from organizational norms, or if significant numbers of employees either fail to internalize those norms or engage in counter-normative behavior. Each of these linkages is potentially problematic because both individuals and organizations are heterogeneous in their responses to the kinds of interventions Professor Bartlett advocates. At the individual level, if discrimination is to be reduced by changing "the preferences and intentions that lead to it,"12 we need to know whose intentions are benign and whose are not, so that we can reinforce the former and change the latter. That is a difficult challenge for two reasons. First, those with prejudicial attitudes and intentions are often unwilling to reveal them.13 Second, as Professor Bartlett points out, those who are prejudiced do not respond to feedback, guidance, cues, and oversight regarding nondiscrimination in the same way as those who have internalized motivations to avoid prejudice.14
At the organizational level, organizations differ in the strength of their cultures and the degree to which shared and internalized norms are designed to and actually do guide the behavior of employees. A large body of social science theory and research (building on ideas from transaction costs economics) analyzes the circumstances under which markets, rules, and shared values are efficient for coordinating activities in pursuit of an organization's goals.15 For many organizations, it is neither economical nor feasible to coordinate activities around a strong uniform culture with widely and deeply shared norms.16
Just as structural reforms can be "co-opted by risk-management principles, amounting to 'symbolic responses' without advancing the goals of equality and inclusion,"17 so too can reforms that are based on transforming organizational cultures and norms. Indeed, one argument is that reforms using organizational norms to shape employees' attitudes and commitments are by their nature symbolic, and as a result they are even more vulnerable to being co-opted. Organizations today readily embrace and invoke the rhetoric of diversity and inclusion, and some do so with little impact on human resources practice.18 Moreover, Professor Bartlett's cautions about the dangers of backlash associated with legal mandates and top-down compliance structures could be read as a prescription to avoid the kind of monitoring and oversight required to diagnose and remedy a breakdown in the very multi-leveled mechanisms upon which her proposals depend.
Although Professor Bartlett offers her remedies as an alternative to "coercive" mandates to comply with civil rights laws, her program for reducing workplace bias is in a way much more radical, with intervention extending considerably beyond transforming workplace cultures and norms. Instead of modifying employers' human resources practices, her proposal for changing the psychological motivations and cognitive associations that contribute to workplace racial and gender discrimination is to profoundly reorganize the structure of organizations, work, and jobs. Professor Bartlett cites research suggesting that restructuring companies by flattening authority hierarchies, breaking down narrowly defined job categories, implementing team-based work and job rotation, and promoting collaborative workplace culture will encourage internal motivation to purse shared goals (including diversity goals) and reduce the impact of stereotyping.19
However, research on "high performance work systems" and "high-commitment human resource management" shows that the costs of moving to these kinds of organizational structures can be substantial, while the benefits to the organization can vary considerably depending on the company's technology, market, and industrial environment.20 For example, while such restructuring may indeed have positive effects on diversity, it is probably futile to preach about the benefits of collaborative, high performance work systems to a labor-intensive company that competes in its marketplace primarily on cost. Finally, even when companies do embrace collaborative and team-based work structures, the consequences for employees extend beyond diversity and are not always positive. High performance structures sometimes lead to an intensification of work and increased job stress,21 and they are often embraced as part of a union-avoidance strategy designed to undermine employees' ability to pursue collective goals.22
III. THE CASE FOR ACCOUNTABILITY AND ENFORCEMENT
Professor Bartlett and I are largely in agreement regarding the necessity of organizational accountability for nondiscrimination and enforcement of equal employment opportunity laws and regulations, although we differ on the extent to which the focus should be on organizational versus psychological processes. She writes that "clear, enforceable, and enforced standards prohibiting workplace discrimination are necessary to reducing it,"23 so long as rules are perceived as fair and do not threaten decision-makers' sense of autonomy, competence, and relatedness. She advocates accountability systems that motivate compliance with nondiscrimination goals in a "positive, constructive way,"24 and her analysis of why such an approach is far superior to heavy-handed compliance efforts is compelling.25 However, the line between oversight approaches that are appropriately positive and those that are "unnecessarily coercive" and threatening is rather vague. Is any approach to oversight of human resources decision-making that is perceived by a manager, supervisor, or employee to be undermining his or her autonomy, competence, and relatedness by definition coercive and counterproductive? That definition ignores the fact that systemic discrimination sometimes is sustained by processes of "social closure" through which high status employees consciously or unconsciously isolate or exclude outsiders, monopolize access to the most desirable jobs via closed social networks, and develop trust and a sense of mutual obligation ("relatedness") based on social similarity.26
It also should be understood that whether accountability and oversight for nondiscrimination is framed as coercive, unfair, and threatening, or as positive and integral to deeply held organizational values, is to some extent socially constructed and shaped by top management. When the catalyst for addressing pervasive "hidden" discrimination is a high profile lawsuit, for example, it is common for top management to respond defensively and to frame plaintiffs' allegations as illegitimate and externally motivated.27 This poses a significant challenge for effecting positive change through Title VII, for all the reasons described by Professor Bartlett. However, at the same time, it is unclear how one can incentivize internally motivated change in organizations that lack the kind of nondiscriminatory leadership that Professor Bartlett identifies as critical. I agree with Professor Bartlett that stronger rules, especially those that impose liability for discrimination for which no causal link has been established (or for discrimination that has not occurred), are not the answer.28 But sometimes external pressures for establishing effective accountability are the only alternative, and the law is a tool for implementing such pressure. Under some circumstances it may indeed be possible to deter discrimination while improving attitudes about race and gender, and when that is possible the proposals set forth by Professor Bartlett are very much worth pursuing. But for the reasons outlined above, I believe it is risky to invest in nondiscrimination efforts that require internally motivated well-intentioned acts in order to be effective.29
My view of accountability and enforcement is consistent with the organizational sociology research cited by Professor Bartlett on effective nondiscrimination interventions. As she notes, the research of Alexandra Kalev, Frank Dobbin, and Erin Kelly shows that diversity training and evaluations (practices designed to alter the attitudes and behavior of individuals by providing education and feedback) are generally ineffective and, in some circumstances, counterproductive.30 Professor Bartlett also relies on this research (and the work of Susan Sturm) in support of the assertion that
responsibility for diversity should be spread across the institution rather than focused in a single individual or administrative office; that in-house experts have advantages over outside consultants in building institutional diversity goals; that top management should be both diverse and committed to diversity; and that use of positive program incentives is more productive than legal compliance mechanisms, such as EEOC charges and lawsuits.31
I read the same research as saying a bit more and a bit less than this. For example, Kalev et al. conclude that "[s]tructures that embed accountability, authority, and expertise (affirmative action plans, diversity committees and taskforces, diversity managers and departments) are the most effective means of increasing the proportions of white women, black women, and black men in private sector management" and that the presence of these structures also makes diversity training and evaluations more effective.32 They also note that some effective accountability structures, like affirmative action plans and diversity staff, centralize authority and accountability for diversity, while others, like task forces and committees, locate it more broadly across the organization.33 The important point is that both approaches embed responsibility into the structure of the organization; they do not simply articulate an organization-wide norm in an effort to shape individual attitudes and behavior, and they do not simply create decentralized accountability and feedback at the individual levels (given diversity evaluations proved to be ineffective).34 Kalev et al. also found that Title VII lawsuits and affirmative action compliance reviews lead to increases in managerial diversity, "especially in periods and judicial circuits wherein civil rights enforcement was strong."35 Finally, they found that one sector where diversity training programs are effective is among government contractors, presumably because those organizations almost always have accountability structures that assign responsibility to a specific office or person.36 I interpret this as strong support for the importance of meaningful accountability and oversight built into the structure of the organization, and for the effectiveness, at least in some circumstances, of legal compliance mechanisms.
In sum, it is indeed important to "accentuate the positive" in designing, articulating, and implementing diversity and non-discrimination interventions, and Professor Bartlett's splendid analysis of individuals' motivations to act fairly deepens our understanding of why that is the case. She also shows us that in certain kinds of organizational environments it may be possible for organizations to pursue nondiscrimination goals in a way that both builds upon and enhances members' motivations to act fairly. But it is also important to recognize and understand that workplace bias (even "hidden bias" that operates outside the conscious awareness of the individuals who make decisions about hiring, pay, promotion, training, and access to career-enhancing resources) operates at the level of organizations and institutions, not just inside people's heads. Failure to attend to the multi-level nature of workplace bias only serves to perpetuate the status quo.
Preferred citation: William T. Bielby, Accentuate the Positive: Are Good Intentions an Effective Way to Minimize Systemic Workplace Bias?, 95 Va. L. Rev. In Brief 117 (2010), http://www.virginialawreview.org/inbrief/2010/02/28/bielby.pdf.
[Back to this issue]