What is the Diversity-Validity Dilemma in Selection?

Due to the diversity-validity dilemma, employers may have to sacrifice the validity of their assessments to maintain and increase the diversity of their workforce. This is particularly important for compliance with equal opportunity legislation, which prohibits discrimination in employment decisions based on protected characteristics such as race and ethnicity.
Employers must also ensure that the selection procedures they are using have acceptable levels of validity; under the Equal Employment Opportunity Commission’s Uniform Guidelines on Employee Selection Procedures, employers in the US are required to provide evidence of the validity of an assessment in the event that it is found to result in an adverse impact in order to justify its continued use.

Key takeaways:  

• The diversity-validity dilemma describes the trade-off between validity and adverse impact
• The most predictive measures of performance can be among the procedures resulting in the greatest adverse impact, so one outcome may have to be sacrificed for the other
• Ways to overcome the diversity-validity dilemma include combining predictors to assess a range of KSAOs, adjusting predictor weightings to reduce the influence of measures associated with adverse impact, and using a pareto-optimal approach to determine the best weightings
• Bias mitigation approaches from computer science can also help to mitigate bias associated with automated assessments while retaining model performance, reducing the trade-off necessary between validity and adverse impact
• An interdisciplinary approach that combines insights from psychology and computer science can be particularly useful for addressing the diversity-validity dilemma for automated tools