Her example was about prisons and who should get parole. So we can apply an algorithm and determine a potential recidivism score and make a decision about whether to grant parole
Henry,
Kristofer Fosmoe
11

Greatly appreciate your thoughtful comments!

Funny thing. Stephen Jay Gould used this very thing as an example of the kind of social Darwinist scientism (i.e. people misusing science to advance their social and political agendas) in his book The Mismeasure of Man. Your characterization of misusing SCIENCE!(tm) to shut up awkward and difficult to answer debates has a long history — in a sense, O’Neil has rediscovered a lot of old problems. This is something I deeply sympathize with.

There is an interesting argument that Gould makes, with regards much older use of “data,” how, from very early on, IQ tests became a tool of many political agendas, long before Murray and the Bell Curve. Gould goes on to claim that a great deal of good could be done by using IQ tests in context, but that was largely brushed aside in favor of a particular social-political agenda. Here, I use the word “claim” since this book led to a firestorm of controversy. Historians were very much sympathetic to Gould, but there were many others who were very much offended — they accused Gould of twisting intelligence tests and other applications of data to suit his social/political agenda. I’m a bit averse to say too much about it since I’m liable to step on to things that I don’t know very well. Until just now, I hadn’t thought about this, but there is a certain analogue to the critique by more data science oriented types towards O’Neil — that she is abusing and mischaracterizing data science techniques a bit. But it also remains true that the people she is critical of are at least as guilty of abusing and misusing data science techniques, and what’s more, like the eugenicists that Gould criticized, many of them are in position to do real damage through their abuse.

I suppose my objection to O’Neil and her arguments is that I think the best counter to the abuse of data and statistics is to better use data and statistics, with the recognition of their limits, whereas O’Neil engages in sleights of hands and moralizing to make her points more persuasive — which, in a sense, is exactly what Gould did before. It might be a bit too waffly an approach, though, if the goal is to change the way the world works, so to speak.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.