Friday, January 11, 2013

On My Problem with Nate Silver

When you look at this graph, what percentages do you think is being represented?





For the first, it's kind of hard to eyeball.  I'd probably guess about 1%.  But what about the second one?  If you said 50%, then you're probably an average American, and your selection makes perfect sense.  But you're also wrong.  Though my issue is more with the graph than with you.  That graph represents 28%, but appears to represent nearly half.  (For the record, that first one represents .12%.)

At first glance, Nate Silver seems like the kind of guy with whom I would get along quite well.  With a shared penchant for poker and for proving people wrong, and the understanding and admiration for probability and statistics that comes with those things, he actually seems like he could be one of my major nerd heroes.

And because of THAT graph, I find myself exceedingly annoyed with him.  That graph was pulled from Chapter 1 of Nate Silver's The Signal and the Noise, a book about how bad or poorly-read data can lead to bad predictions.  And in this midst of all of this warning about how when you only choose to see the data the way you want to see it, your predictions will end up being not only wrong but potentially really harmful (this section in particular was on the housing bubble), he chose to use a graphic that was willfully manipulative and presented in a way which fools the average thinker into being a lot more concerned than they need to be.  Don't get me wrong, the fact that the number of securities which failed was 500 times the predicted number is a pretty damning statistic.  But don't compound that with intentionally misleading representation.  You're giving credence to the idea that there are "lies, damned lies, and statistics" in a book that is supposed to be defining the traps that our predictions fall into.  One such trap is EXACTLY what you did just there.  And it's unfair and manipulative.

I have since come to enjoy quite a bit of the book, which I am nearly through, but I can't help but remain hung-up on that one glaring point (and copious typos - did you even HAVE a copy-editor?) and getting really disgruntled about how knowingly misleading it is.

Irritably yours,
Rachel Leigh

No comments:

Post a Comment

Please let me know what you think!