TRC

Precious few of us are completely unbiased. Many of us would like to believe that we are unbiased when it comes to questions of race, gender, appearance etc, and would even say so when asked in surveys. But evidence indicates that at the aggregate level bias is real and prevalent. For instance, research by Marianne Bertrand and Sendhil Mullainathan has shown very clear bias in the selection of candidates for job interviews. They ran an experiment where they sent in resumes to real jobs where nothing was different except for how African-American some of the applicants' names sounded. In spite of equal qualifications, white names received 50% more interview calls. Sadly this was true even for federal jobs where "Equal Opportunity" is explicitly advertised. So, if bias exists but people won't say they are biased, how does one go about measuring it? Cleverly, of course. Here we will discuss two interesting indirect ways of measuring bias: Implicit Association Test (IAT) and Conjoint Analysis (CA).

Comment (0) Read More...