Thursday, March 27, 2014

On Foxes, Hedgehogs, and Toads

Much ado now about the arrival of Nate Silver's new media venture--FiveThirtyEight. The reviews are in. They're not so good. The best take-down is by Noah Smith. In turn, Silver has "lashed out" at his critics, especially Paul Krugman. Here's a blow-by-blow.

The heart of most people's criticism, with which I fully concur, is that the website and articles therein have some serious flaws.

First, as Stefan Fatsis said on Slate's Sports Podcast "Hang Up and Listen" this week, the headlines basically tell the whole story. So why read further? Titles like "Residents of Struggling Cities Opt to Skip Town" and "GOP Field Hasn't Been This Split in 40 Years" doesn't make me really want to spend my scarce attention time on reading the whole damn article. Now I can switch to more interesting pieces in Salon or Slate that have catchier headlines, because I know they're about to tell me something I didn't know or expect. This is really a pitch or style issue.

Worse is that the pieces or arguments themselves aren't that interesting. Most of the pieces in the politics section at least [don't know as much about the other sections] say nothing new or counterintuitive. Even his much ballyhooed election predictions, for which he gained his legion of fans and celebrity, aren't very original. The Republicans have a 60% of taking the US Senate? Anyone keeping up with the political headlines and rough understanding of who's up for reelection this year in the Senate in an off year would know this (which is mainly his audience anyway).

Then there's the issue with the quality of the writers themselves. For example, many have piled on Roger Pielke. It appears that Silver is suffering from the same growing pains in the hiring department as Ezra Klein did when he hired a "contrarian" homophobe.

Plus, the website is just a fancy blog (i.e., not much better than this one). Drop-down menus, ugly, uninteresting fonts, and a layout that looks like it's from the early 00's. Really, with the likely millions of dollars from ESPN and ABC (read: Disney) backing him up, I would expect something more eye-catching that pulls me into the website and keeps me there. He and his financial backers are after clicks and page views. Myself? I just check the headlines once a day and leave.

As Smith points out, there's not a lot of analysis going on, at least in any sophisticated way, with the possible exception of the NCAA Final Four postings [Great timing BTW: No better time to bring in stats than when people are obsessed with the odds of winning their office pool NCAA basketball bracket contests; leverage the pseudo-science of "bracketology"]. And then just presenting data is sufficient journalistic insight.

So Silver is learning, I hope, that he's not doing a good job as an editor, manager, or lead web designer or at least providing a vision for his blog. But is there something more wrong here than just his management style?

Unfortunately, I think it's his complete misunderstanding of the whole Fox and Hedgehog metaphor. He introduced it to us in his greeting to readers. I think it's Silver's belief (and maybe his employees' too) that "data" always speak for themselves. Most people have leveled the legitimate criticism that Silver is not really even doing data analysis, which is what he promised when he laid down the gauntlet on "narrative" journalism.

Yet, it's more than that. Even if he and his writers started actually doing sophisticated analytical work [in a space that's crowded with some excellent blogs], it seems that he's ignoring some important elements of what it really means to be a "fox."[For a quick but awesome summary of the "fox vs. hedgehog allegory/metaphor, go here].

Briefly, a fox knows much about a lot of things; a hedgehog knows a lot about one great thing.

Revised: the fox is skillful at many tasks, but the hedgehog is a genius in one thing.

Extended: The hedgehog tries to fit all data, including contradictory information, into her unified vision of the world such that that view remains unscathed and pure. A fox, in contrast, will learn new things from a variety of fields or streams of knowledge and then update her beliefs.

The great advantage of the fox was that she is openminded and sees how her vision can constantly be updated by new information or alternative explanations by visiting other patches of the field. A hedgehog, in contrast, will just keep her nose to the ground. If conflicting information falls into her hole, it will either be discarded, ignored, or re-shaped to fit his original vision.

But, to be a good fox you need at least some working knowledge of the topic! To be a good fox you need at least a little bit of theory/hedgehog inside of you! Even if you're a fox, constantly searching for new information located in the heavens, you need to know how to interpret evidence, why the information is relevant, and how it fits or not with existing, empirically valid and rigorous explanations. A fox has a good understanding of what the puzzle or problem is, what explanations have been offered in the past, and what the data look like. So when new information or insights come down from the sky, the fox can update her thinking. There needs to be some working knowledge, at least on the surface, of the puzzle you want to explain.

As a fellow U of C grad (we actually graduated the same year, but did not know each other), I hoped Silver would have absorbed this lesson. But Silver and his employees seems to think presenting data on its own is enough to make some insight. "See the regression line I drew. That proves it!" What it represents and why I should care is not really discussed; nor the contribution being made.

To some extent, this worked when he was doing the presidential forecasting. It appeared like he was just sharing numbers. But even there he was astute and in-tune enough to political science scholarship to read other people's models when constructing his own. So, while he was often just "pointing to the data," and that's what it often looked like, he had a ton of models going on in the background. And those models come from theory; simple explanations about how things work--namely, presidential elections.

What it comes down to is that Silver and his writers need to take their paws and dig down a little bit. They need to anchor themselves in the ground, before they start roaming around for new information. You don't need to become a hedgehog to do that. You don't need a certificate, degree or some other qualification to take on a topic. But when you're pulling down data. You need to tell the layperson why it's interesting given what we supposedly already know about it.

My biggest concern, however, is that he's taking all the criticism personally. Silver posted a stupid little piece about how Krugman's views of fivethirtyeight changed when Silver set up his own shop elsewhere. For some it appears that Silver is holding a grudge against anyone at his former employer, the NY Times. He also has a special dislike for those notorious pundits, like Peggy Noonan and George Will, who spit out just blather from their honorary privileged seats on the Sunday Morning talk shows. I understand and share his frustration with "pundits" and the pundit-ocracy. These people aren't hedgehogs. They're toads. Just burping useless hot air and comments full of warts.

Silver should realize that the criticism that's been leveled from most is from people who admired his work in the past and see a very bright future ahead for him. We want him to succeed! As long as it doesn't take things too personally and is openminded, like a fox should be, he'll make it a great site and contribute to the growing wave of "data-driven journalism."

No comments:

Post a Comment

There was an error in this gadget