Affirmative Action

Affirmative action has been one of the most controversial public policies of the past 40 years. A conceptual definition of affirmative action is any measure, beyond a simple termination of discriminatory practice, adopted to correct for past or present discrimination or to prevent discrimination from recurring in the future. In practice, organizational affirmative action programs (AAPs) can and do encompass a multitude of actions. These actions are shaped by federal, state, and local laws and regulations. Although some educational institutions apply affirmative action to student admissions and many countries have corresponding laws and regulations, this entry is limited to workplace affirmative action in the United States.

Legal Issues in Affirmative Action

Affirmative action law in the United States is jointly determined by the Constitution, legislative acts, executive orders, and court decisions. It is complex, incomplete, and open to revision. The Office of Federal Contract Compliance Programs is responsible for developing and enforcing most AAPs, although the Equal Opportunity Employment Commission (EEOC) enforces AAPs in the federal sector.

A distinction exists between so-called set-aside AAPs and organization-specific AAPs. Set-aside AAPs exist when a pubic organization (e.g., a municipality or federal agency) is required to set a goal for directing a certain percentage of its budget to qualified firms—typically those owned by members of an underrepresented group.

In contrast, organization-specific AAPs are created for one of three reasons. First, some organizations are required by a court order or an EEOC consent decree to establish an AAP to compensate for illegal discrimination. These AAPs are relatively rare. Second, many organizations establish AAPs to satisfy regulatory requirements. Specifically, the Rehabilitation Act of 1973 and the Vietnam Era Veterans’ Readjustment Assistance Act of 1974 require certain federal contractors to take affirmative action to employ individuals with disabilities and certain veterans, respectively. Most important, Executive Order 11246, signed by President Lyndon Johnson in 1965 and subsequently amended, requires federal contractors to take affirmative action to eliminate discrimination on the basis of race, color, religion, sex, or national origin. Along the same lines, state and local laws and regulations may require organizations to take affirmative action to improve the employment opportunities of various groups. Third, some organizations establish AAPs on a fully voluntary basis.

Precisely which organizations are required to establish AAPs and which actions are required, permitted, or forbidden varies with the legal basis for the AAP. Furthermore, actions of state and federal governments are limited by the U.S. Constitution, whereas actions of firms in the private sector are constrained by state and federal legislation (e.g., the Civil Rights Acts of 1964 and 1991). The following brief and incomplete description focuses on affirmative action as required by Executive Order 11246, because that is the primary source of AAPs in the United States and is the basis of much of the controversy.

Organizations with annual federal contracts of at least $10,000 are required to take affirmative action to eliminate discrimination on the basis of race, color, religion, sex, or national origin. They must establish nondiscrimination policies and communicate those policies to employees and applicants. Organizations with at least 50 employees and contracts above $50,000 are further required to perform and report the results of utilization analyses in which they compare the gender and racial distributions of their workforce to the relevant labor markets. The relevant labor market for any position includes only those individuals who are qualified for that position and who reside in the recruitment area. If the utilization analysis reveals that all groups are reasonably represented, no further actions are required. If the utilization analysis reveals that any group defined by gender, race, or ethnicity is underrepresented, the firm must establish flexible goals to eliminate the underutilization and must make a good faith effort (i.e., take affirmative actions) to meet those goals. Utilization analyses are not required for other protected dimensions (i.e., disability, veteran status, religion), so it is impossible to determine whether underrepresentation exists along these dimensions.

An important question is which actions are permitted when underutilization is revealed. Federal regulations strongly emphasize nonpreferential actions such as the elimination of barriers and the use of targeted recruitment or training. Because these approaches may fail to eliminate substantial underrepresentation, some organizations may want to take stronger actions. In so doing, the firm must not violate the constraints established by the Constitution and antidiscrimination law. It is clearly illegal to use quotas or to give preferences to unqualified members of the underrepresented group (e.g., through race norming of the selection test). Furthermore, Supreme Court decisions have determined that any AAP that gives a positive weight to racial minority status is subject to “strict scrutiny.” Such an AAP must be remedial, narrowly tailored, and temporary; must not trammel the rights of others; and must further a compelling governmental interest. Note that the final requirement can be satisfied only within the public sector. Although it has been suggested that private-sector organizations might use the economic value of diversity to justify positive weighting of racial minority status, it is not clear that such an argument would be approved by the Supreme Court. Although positive weighting of gender requires only intermediate scrutiny rather than strict scrutiny, it would still be a risky approach.

Empirical Research of Affirmative Action

As mentioned previously, affirmative action is a controversial public policy. The debate regarding whether it should be eliminated, perpetuated, or expanded is complex. For example, philosophical arguments have been offered regarding the appropriateness of using race-conscious approaches to attain a race-blind society. These arguments tend to focus on race-based affirmative action and occasionally gender-based plans; they rarely mention policies that target veterans or individuals with disabilities. These debates also focus on preferential forms of affirmative action rather than the more common, and legal, nonpreferential forms. Empirical research, in contrast, has focused on the consequences of affirmative action and on predictors of attitudes toward affirmative action.

Consequences of Affirmative Action

For Organizations

A logical analysis shows that affirmative action could either help or hurt organizational performance, depending on details of the AAP and on the procedures used by the organization in the absence of affirmative action. Positive effects should occur if the AAP increases the organization’s access to the labor market (e.g., through intensive recruitment) or decreases discrimination against women or racial or ethnic minorities. Negative effects should occur if the organization uses a preferential AAP that supplants a nondiscriminatory procedure. In addition, organizations must bear the costs of administering AAPs. Consistent with the logical uncertainty, empirical research has failed to demonstrate any substantial effect of affirmative action on organizational performance.

For Target Groups

A different line of research has assessed the economic impact of affirmative action on the targeted groups. Affirmative action appears to have improved employment outcomes of targeted groups, but the effects have varied in complex ways depending on factors such as the targeted group, geographic region, time period, and type of position. For example, affirmative action had a substantial positive impact on African Americans in the South between 1965 and 1975, presumably because that time and place offered a substantial opportunity for improvement. On the other hand, affirmative action had virtually little or no effect during the 1980s, perhaps because the Reagan administration decreased support for the regulatory agencies and substantially revised those agencies’ policies and procedures. Little or no research exists on the effects of affirmative action on employment of individuals with disabilities or veterans. Because organizations are not required to report employment statistics of these groups, such effects would be difficult to document.

There is also some evidence that affirmative action may lead to stigmatization of individuals who belong to the AAP target group. The logic is consistent with the discounting principle of attribution theory. When targeted individuals are selected in the context of an AAP, others are uncertain about whether their selection was because of their performance or the AAP. In the absence of affirmative action, this uncertainty disappears and the individual is assumed competent. Research reveals such stigmatization when observers believe or are told that the AAP involves preferences. It can be eliminated or greatly reduced by providing compelling evidence that the AAP is nonpreferential or that the selected individual is fully qualified or has performed well.

A related stream of research deals with self-stigmatization by target group members. According to the logic outlined above, members of AAP target groups may doubt their own competence and consequently lose confidence and interest in the job. Although this effect has been observed, almost all supportive evidence has come from laboratory research in which White female college students are explicitly told that they have been preferentially selected on the basis of their gender. There is little evidence for this effect among racial or ethnic minorities, and the effect is absent or much smaller when participants are given clear evidence of their competence or are told their selection was based in part on merit.

For White Males

A final question concerns the impact of affirmative action on White males. Although there are many reports of backlash—opposition by White males based in part on the belief that they have been hurt by affirmative action—there is surprisingly little research on this question. Logically, the effect should be negative if affirmative action reverses traditional biases that favor White males or if preferential forms of affirmative action replace nondiscriminatory procedures. The limited research that exists reveals such a negative effect. Of course, this assumes a “fixed pie” situation; if implementation of an AAP enhances organizational performance because of the increased diversity, that increased performance may help all organization members.


Perhaps the largest body of empirical research on affirmative action has dealt with public attitudes toward the policy. This work has assessed the effects of structural predictors, perceiver variables, and psychological mediators of the effects.

The structural predictor that has received the most attention is AAP strength—the weight given by the AAP to demographic characteristics. Specifically, the public strongly supports AAPs that require only the elimination of discrimination. Support decreases somewhat for AAPs that are designed to enhance target group opportunities—for example, by requiring targeted recruitment. There is a further drop in support if the AAP requires selection of underrepresented group members when their qualifications are equivalent to those of other applicants. Note that such an AAP would rarely if ever pass legal muster. Finally, there is strong opposition to AAPs that require preferential selection of underrepresented group members even when their qualifications are inferior to those of other applicants. Although such an AAP would be illegal, many scholars who study attitudes toward affirmative action attitudes have described it in those terms, and many people believe preferences are common. Indeed, although most research on AAP strength has involved manipulation of the AAP, research on public beliefs reveals corresponding effects, so that individuals who believe affirmative action merely requires the elimination of discrimination have more positive attitudes than those who believe it involves preferences.

The only other structural predictor that has received enough attention to merit conclusions is the identity of the target group. It appears that attitudes, at least of White respondents, are more negative when the AAP is described as targeting African Americans or minorities than when it is said to target women or individuals with disabilities.

The two perceiver variables that have received the most attention are respondent race and gender. In general, African Americans report the strongest support for affirmative action and Whites the strongest opposition, with Hispanics and Asians reporting intermediate levels of support. However, this race effect is moderated by AAP strength, increasing in size as the AAP gives greater weight to demographic status. The effect of gender on attitudes is much smaller, but in general, women report more support than do men.

Attitudes are also associated with several opinion variables. Most significantly, opposition increases with the respondent’s racial prejudice and sexism. In addition, those who subscribe to a conservative political ideology or who identify with the Republican Party report greater opposition than do those who are politically liberal or who identify with the Democratic Party. Opposition also increases with the level of the respondent’s social dominance orientation—an individual difference variable that represents a general opposition to equality and support for group-based dominance. Finally, support for affirmative action is associated with the belief that the target group experiences discrimination and thus that affirmative action is needed.

Research on psychological mediators finds that support for affirmative action is positively associated with anticipated positive effects of the AAP on the respondent’s personal self-interest and on the respondent’s demographic group. But the strongest association of all is with perceived fairness of the AAP—people support AAPs they consider fair and oppose those they consider unfair. As this would suggest, providing a justification increases support for affirmative action, but only if the justification refers to the value of diversity or the need to make up for past discrimination; simply citing underrepresentation typically decreases support instead of increasing it.


  1. Crosby, F. J. (2004). Affirmative action is dead; long live affirmative action. New Haven, CT: Yale University Press.
  2. Crosby, F. J., & VanDeVeer, C. (2000). Sex, race, and merit: Debating affirmative action in education and employment. Ann Arbor: The University of Michigan Press.
  3. Doverspike, D., Taylor, M. A., & Arthur, W., Jr. (2000). Affirmative action: A psychological perspective. Huntington, NY: Nova Science.
  4. Edley, C., Jr. (1996). Not all Black and White: Affirmative action, race, and American values. New York: Hill & Wang.
  5. Gutman, A. (2000). EEO law and personnel practices (2nd ed.). Thousand Oaks, CA: Sage.
  6. Holzer, H. J., & Neumark, D. (2000). Assessing affirmative action. Journal of Economic Literature, 38, 483-568.
  7. Kravitz, D. A., Harrison, D. A., Turner, M. E., Levine, E. L., Chaves, W., Brannick, M. T., et al. (1997). Affirmative action: A review of psychological and behavioral research. Bowling Green, OH: Society for Industrial and Organizational Psychology.
  8. Leiter, S., & Leiter, W. M. (2002). Affirmative action in antidiscrimination law and policy: An overview and synthesis. Albany: State University of New York Press.
  9. Rubio, P. F. (2001). A history of affirmative action, 1619-2000. Jackson: University Press of Mississippi.
  10. Spann, G. A. (2000). The law of affirmative action: Twenty-five years of Supreme Court decisions on race and remedies. New York: New York University Press.

See also: