FAQs

Practical Concerns

• Other than complying with specific regulatory requirements, what do I need to consider?
• What types of issues arise during the preparation of an affirmative action plan?
• How concerned should I be if my analysis reveals statistical disparities?

Other than complying with specific regulatory requirements, what do I need to consider?

OFCCP regulations explain that affirmative action plans are more than a paperwork exercise. The plan must evaluate workforce statistics and must address underutilization through action-oriented programs describing specific practical steps. This process should be undertaken with considerable care. If an affirmative action plan states unequivocally that certain problems exist, the employer may find it difficult at a later date to dispute these written statements or to offer qualifications or explanations that were not spelled out in the document. Similarly, if an affirmative action plan commits to specific remedial steps, the employer will not easily be able to justify a failure to carry them out, even if it turns out that these commitments were unrealistic or incompatible with day-to-day management practices.

What types of issues arise during the preparation of an affirmative action plan?

Many difficult issues may arise, particularly in connection with the statistical analysis. For example, in gathering and assembling the underlying data, a contractor must resolve such questions as the exact stage at which an individual becomes an applicant and the way in which job titles should be arranged into groups for purposes of statistical review. In preparing the utilization analysis, a contractor also must determine census categories to use as a basis for comparison, must identify feeder groups for promotions to higher job groups, and must strike the correct balance between external hires and internal promotions as a source for comparative data. All of these questions require judgment calls for which neither the regulations nor any pre-determined formula can provide a universal answer.

How concerned should I be if my analysis reveals statistical disparities?

It is quite normal for a statistical analysis to reveal one or more significant disparities. If this occurs in your plan, you should not leap to any strong conclusions. Instead, keep two facts in mind. First, a significant disparity is often defined as a difference exceeding two standard deviations, indicating about a 95% probability that something other than random chance explains the outcome. Even if you chose employees by the most color-blind and gender-blind means imaginable flipping a coin, for example you would still expect a significant disparity about 1 in 20 times. Second, no analysis ever can achieve perfect accuracy. There may be no census data category, for example, that matches the exact mix of skills you require in a position. In such situations, a statistical disparity may reflect problems with the comparative data rather than actual discrimination. This does not mean that you should disregard statistical disparities by failing to address them in your plan. It does mean that you should avoid overreacting by treating availability percentages as rigid quotas or by altering your hiring or promotion standards. An affirmative action plan is a management tool designed to identify areas where the employer can strive to improve. It should not be transformed into an unachievable quest for statistical perfection.

Return to top