Data & Civil Rights: Employment Primer

Abstract

Employees and prospective employees produce more data than ever - in the workplace, on social media, and beyond. Employers and the third party companies that assist them increasingly apply analytical tools to these various data streams to measure factors that influence employee performance, attrition rates, and workplace profitability. While some of the data - such as past performance - are unquestionably relevant to such analysis, other data that produces strong correlations to performance are more surprising. For instance, Evolv, a recruiting software company, analyzed 3 million data points about 30,000 hourly employees and identified that those who installed newer browsers, like Chrome or Firefox, onto their computers stay at their jobs 15% longer than those who use default browsers that come pre-installed on their computers, like Safari for Macs. Job candidates may rightly worry that they will be excluded from or included in job opportunities based on data that seem arbitrary and are outside their field of vision. For example, a job candidate’s resume could be excluded from a talent pool because of her online browsing habits, but she is unlikely to find that out directly. The complexity of hiring algorithms which fold all kinds of data into scoring systems make it difficult to detect and therefore challenge hiring decisions, even when outputs appear to disadvantage particular groups within a protected class. When hiring algorithms weigh many factors to reach an unexplained decision, job applicants and outside observers are unable to detect and challenge factors that may have a disparate impact on protected groups.