Posts

Showing posts with the label bias

What I Learned from a Year Spent Studying How to Get Policymakers to Use Evidence

Image
Source: http://cognitive-edge.com/blog/on-evidence/ The past year I was a senior research analyst at Northwestern University's Global Poverty Research Lab on a study of evidence-based policy. Specifically, our goal was to work on a question often on researchers' minds: how can I get my ideas acted upon? To do this, I dug through a number of bodies of evidence on how science influences policy. One area I looked at is what is called "implementation science" in medicine, which looks at how to get doctors, nurses, and hospital administrators to adopt evidence-based practice. Another was a series of papers by social scientist Carol Weiss and her students on how policymakers in government agencies claim to use evidence. There is also a small literature on how to implement evidence-based policy in public schools, and a little work on policymaker numeracy. I've included a bibliography below that should be helpful for anyone interested in this topic. Most of my yea

Data and Racism in Machine Learning?

Image
We often hear stories these days about racism in machine learning algorithms. The subtlety in these stories is often missing. I've been reading about this recently and found this quote very telling: A wave of scholarship, triggered by the ProPublica report , illuminated the statistical challenge at the heart of the argument: Given that the underlying “base rate” of rearrest is higher for blacks than for whites, it is mathematically inevitable that the burden of false positives will fall more heavily on black defendants than on white ones. In other words, given that more black defendants than white defendants actually do have a high risk of reoffending, a “high risk” label that is correct 70% of the time for both white and black defendants will still mis-label more black than white defendants as high risk. A study titled “Inherent Tradeoffs in the Fair Determination of Risk Scores” proved mathematically that when rearrest rates are not equal between races, a well-calibrated tool