Cathy claims that Nate Silver confused cause and effect in his New York Times best seller "Signal and Noise". Cathy cites a few examples of this.
For instance, in Cathy's judgement, corruption caused bad models, not the other way around.
Reasonable people can disagree.
That isn't anything that anybody wants to hear though. In general, populist messages are positive.
Indeed, Nate is popularizing modelling.
Cathy doesn't like it:
"Silver is selling a story we all want to hear, and a story we all want to be true. Unfortunately for us and for the world, it’s not."
And, there's a call to action:
"It would be great if substantive data scientists had a way of getting together to defend the subject against sensationalist celebrity-fueled noise."
People want to hear what they want to hear.
It's akin to the person who sells beet juice across the street from somebody who sells Boylan's Root Beer.
Good on Cathy for voicing the concerns.
Is it really such a bad thing that Nate Silver is sort of like the Entertainment Tonight of bayesian modelling though?
***
I'm Christopher Berry
@cjpberry
2 comments:
Wow. It sounds like Cathy and I were reading the book at the same time and wound up with very different takeaways. She really honed in on his assessment of the financial meltdown and the details of his explanation. I don't have the modeling chops that she does by a long stretch, so, what I took away, was that Silver was actually making the point -- repeatedly and with numerous examples -- that predictive models are not "just" a matter of having a lot of data and a smart modeler. He actually highlighted multiple times how there is not a single, black-and-white approach that is applicable in all fields at all times. He made a convincing case that it's a messy world, that you need to recognize that no human being is 100% objective, and that prediction really is an "art and science" (and environmental luck) deal.
@TimWilson
I agree.
I don't know for sure if people knew what they were doing. Models help people figure that out. If they want.
I think it's a reinforcing variable.
Post a Comment