New Report: How HackerRank Prioritizes Assessment Fairness & Sensitivity
At HackerRank, we continuously take equality and fairness in our assessments very seriously. We believe removing bias from the hiring process is key to matching developers with the right companies.
That’s why we’ve launched a report with more information on test fairness and sensitivity.
The goal of fairness in testing should be ensuring that test properties are as barrier-free as possible and fair for all test takers. To this end, we are committed to helping hiring managers and recruiting teams attract and select a diverse and talented workforce.
To ensure that any unintended bias is identified and removed from our library of test items, we regularly conduct a careful Fairness and Sensitivity Review of our assessments by experts including senior item writers/designers, content experts, psychometricians, and Industrial/Organizational psychologists.
Types of bias our experts seek to identify in the review include:
- Offensive Items
- Racial or Ethnic Bias
- Gender Bias
- Socioeconomic Bias
- Stereotypical Bias
Highlights from the report include:
- Why building unbiased assessments can be challenging for test developers.
- How HackerRank uses the Adverse Impact framework to assess test fairness and equality.
- Results of several studies conducted to demonstrate the fairness of HackerRank assessments.
As a polished, dynamic, and highly skilled industrial/organizational psychologist with many years of experience in all aspects of the assessment industry, Fred Rafilson has authored and published over 30 employment exams that assess cognitive abilities, skills, personality traits, motivation and attitude, etc. He has developed and implemented assessment processes for hundreds of companies/organizations and federal, state, and county/municipal agencies. He also conducts scientific research and program evaluations and provides presentations and training to executive-level private and public-sector groups. In addition, he serves as an expert for employment litigation cases, often working closely with agencies under DOJ consent decrees.