By Lisa Harpe and David Cohen
The concept of adverse impact under Title VII of the Civil Rights Act (Title VII) is widely misunderstood. Adverse impact occurs when a seemingly neutral employment practice disproportionately (i.e. statistically) screens out members of a protected class. To be clear, under Title VII, adverse impact is legal if it can be shown to be job-related and consistent with a business necessity. In other words, statistical differences in selection rates are perfectly legal if the selection procedure is “merit-based” and can be justified.
As the regulatory landscape continues to evolve under the Trump administration’s emphasis on illegal diversity, equity, and inclusion (DEI) practices, one interesting issue to monitor is the application of disparate impact theory. This issue is particularly relevant given the theme of “merit-based” decision-making within the new regulatory landscape, which DCI recently blogged about.
The most important context when considering disparate impact theory in the changing regulatory landscape is Project 2025, an influential federal policy agenda produced by the Heritage Foundation[1], a policy think tank in Washington D.C. While Project 2025 does not propose the outright repeal of Title VII of the Civil Rights Act of 1964[2]—which prohibits employment discrimination based on race, religion, national origin, and color—it does call for amending Title VII. Specifically, it advocates for eliminating the collection of demographic data and the elimination of the disparate impact theory of employment discrimination. These changes would significantly reduce the enforcement strength of Title VII.
The disparate impact theory of discrimination was codified by the Supreme Court’s decision in Duke v. Griggs (1971)[3]. In that case, Duke Power Company imposed a high school diploma requirement for employees seeking to transfer to higher-paying departments. Those without a diploma could substitute a passing score on an intelligence test. These requirements disproportionately affected Black employees. The legal issue was not that the procedure negatively impacted Black employees; rather, it was that White employees without a high school diploma were able to perform the job adequately. Duke Power failed to demonstrate the necessity of either the diploma or the intelligence test score to perform the job (i.e., they were not job-related or merit-based).
Many well-intentioned employers implement selection procedures in hiring, promotion, and other employment decisions without evaluating their job-relatedness. When these procedures disproportionately affect members of a protected class, they have an adverse impact, supporting the first phase of a disparate impact claim of discrimination[4]. However, statistical disparities alone do not constitute discrimination. In such cases, employers must demonstrate the job-relatedness of the selection procedures.
We cannot stress enough that statistical disparities do not denote discrimination. In our experience, there is confusion on this matter. There are various reasons why we may expect disparities against demographic groups in a particular context, which may include Whites and/or males. We may expect different demographic distributions on particular characteristics based on well-established social scientific research literature, the method of measuring adverse impact, and/or the size of the pool being analyzed. For example, social science research literature suggests that men typically score higher on physical ability tests, while women typically score higher on measures of the personality characteristic Agreeableness. These differences may produce an adverse impact in decisions, yet particular use cases of these measures may be job-related, valuable to an employer, and defensible. That is, a job-related (i.e., merit-based) selection procedure that results in an adverse impact on a protected class is legal. Conversely, a non-job-related (non-merit-based) procedure with an adverse impact is illegal. In other words, what causes the disparity matters in disparate impact theory, and if the cause is job-related (i.e., merit based) that disparity is likely justified.
In fact, the Supreme Court ruled in 2009 in favor of a class of White firefighters when the New Haven Fire Department discontinued the use of a job-related promotion test due to its obvious adverse impact on Hispanic and Black test takers[5]. The Court found that by eliminating the test based solely on race-related outcome differences, the Fire Department violated Title VII. That is to say, by giving more weight to the adverse impact of the selection procedure over the job-relatedness, the Fire Department had discriminated against the White promotion candidates.
Misunderstood Relationship Between Disparate Impact and DEI
The concern expressed in Project 2025 regarding disparate impact as a Diversity, Equity, and Inclusion (DEI) tool to increase the representation of females and minorities seems to be misplaced. Disparate impact theory protects any demographic group subjected to a non-merit-based selection procedure that disproportionately screens out that group because Title VII protects everyone.
Consider the following scenarios: An employer prefers to hire women over men and implements an assessment that measures Agreeableness because women tend to score higher on measures of Agreeableness. While Agreeableness may be a positive attribute, it may or may not be relevant to any specific job. By imposing such a requirement, an employer could limit male representation in the workforce. In a different scenario, an employer seeking higher representation of Hispanic employees might implement a Spanish-speaking requirement. In client-facing roles in a community with a significant Spanish-speaking population, Spanish speaking could be job-related. However, in non-client-facing roles, it may not be. Alternatively, an employer aiming to increase Hispanic or Black representation could limit the applicant pool to specific zip codes. Without the collection of race or gender data, without adverse impact analyses, and without the disparate impact theory of discrimination, a Male or White applicant – for example - may have little recourse to claim discrimination. Disparate impact theory allows for a legal evaluation in these examples, and again, the job-relatedness of the process is the key dimension.
Demographic Data is Critical for Protecting Everyone
An adverse impact analysis does not inherently favor any particular race or gender; the analysis is simply a tool for evaluating a selection procedure. The analysis does not care which racial or gender group is favored or disfavored. What is crucial is that employers review all adverse impact findings, regardless of the disadvantaged race or gender. This approach aligns with Title VII, which prohibits employment decisions based on any race or gender.
The perception that minorities and women receive more positive outcomes in employment decision making whether based on preferences or facially neutral selection procedures over non-minorities or men can be assessed, but only if race and gender data are available and adverse impact analyses, consistent with the disparate impact theory of discrimination, are conducted. Without the underlying data, employers lose a critical piece of analytics to help determine whether decision making was merit-based. Given the many nuances related to legal theories of discrimination and analytics, employers should involve legal counsel when deciding how to proceed.
For additional insight on adverse impact and related matters, keep an eye on this space. The DCI Blog will continue to address industry best practices and offer resources as needed.
[1] Mandate for Leadership: The Conservative Promise (2023)
[2] Civil Rights Act of 1964 § 7, 42 U.S.C. § 2000e et seq (1964)
[3] Griggs v. Duke Power Co., 401 U.S. 424 (1971)
[4] Disparate impact is a theory of discrimination. Adverse impact is a negative outcome such as a statistical disparity in selection rates that may support a Disparate Impact claim.
[5] Ricci v. DeStefano, 557 U.S. 557 (2009)
