In a recent episode of Analysis, David Edmonds digs into some of the conflicting evidence and critiques of implicit bias research. Does the Implicit Association Test stand up to rigorous statistical scrutiny? Do survey tools that measure implicit bias actually predict bigoted, prejudiced, or discriminatory behaviour in the real world? How explanatory is the data we use to describe implicit bias? Do workshops that teach people about implicit bias give participants effective tools to constructively change their unconscious thinking… or do they inadvertently foster an atmosphere of distrust between majority and minority groups?
This program provides a good reminder to maintain a healthy, indifferent skepticism about using indirect measures to infer or quantify the way that other human beings think about one another. At the risk of falling into an infinite rabbit hole of meta analysis, the key takeaway here is to remember that its quite possible to be bias about one’s understanding of bias.