Tuesday, May 13, 2014

Another one for the West Coast Stat Views lexicon: Jethro Models

[Part of an ongoing series]

The Jethro Model is a formal or informal model that leaves out a large number of necessary parts. The allusion was explained in a previous post.

From Anti-orthogonality at Freakonomics
In one of the many recurring gags on the Beverly Hillbillies, whenever Jethro finished fixing the old flatbed truck, Jed would notice a small pile of engine parts on the ground next to the truck and Jethro would nonchalantly explain that those were the parts that were left over. I always liked that gag and the part that really sold it was the fact that the character saw this as a natural part of auto repair: when you took an engine apart then reassembled it you would always have parts left over.

Sometimes I find myself having a Jed moment when I read certain pop econ pieces.

"What's that pile next to your argument?"

"Oh, that's just some non-linear relationships, interactions, data quality issues and metrics that won't reduce to a scalar. We always have a bunch of stuff like that left over when we put together an argument."
For a recent example, consider this quote from  George Mason University economist Robin Hanson (via Andrew Gelman):
If your main reason for talking is to socialize, you’ll want to talk about whatever everyone else is talking about. Like say the missing Malaysia Airlines plane. But if instead your purpose is to gain and spread useful insight, so that we can all understand more about things that matter, you’ll want to look for relatively neglected topics. . . .
Obviously, this is intended more as an observation than even an informal model, but we're still looking at a level of simplification that makes this rule pretty much meaningless; as soon as add any of the complexity of actual conversations, either with respect to why we converse or how we decide what to talk about, the whole argument just collapses. We converse for a long list of reasons. Sometimes we simply want company. Other times it's something more specific, to propagate our ideas, to amuse, to impress, to be liked, to establish individual and group identity, to get laid or, far more frequently, to convince ourselves that we could get laid if we wanted to. We could make similar list of reasons for picking conversational topics, but I think you get the point.

To reduce this down to social vs. informative motives and common vs. neglected topics, you either have to leave out important options or group together things so diverse as to make the definitions meaningless. What's more, by equating neglected topics with informative conversations, the model suggests some strange implications, such as that the person who just wants to be sociable will talk about racism and climate change, while the person who wants to be informative is more likely to discuss obscure distinctions between Phish bootlegs.

That's not to say that there's no extra value to bringing up neglected topics; it's just that Hanson's observation doesn't capture the fundamental relationships. I've been writing quite a bit recently on the importance of orthogonality and there's certainly a relationship between unique information and how much a topic has been discussed. Unfortunately there's also a great deal of collinearity. Lots of topics are relatively neglected because they don't contain that much interesting information.

To further complicate matters, under the right circumstances, you can gain considerable social cachet by knowing interesting facts about little known topics. The "interesting" part can be a bit of a hurdle, but I know  people who do it which puts yet another hole in the model. As do people who bring up obscure topics for the primarily social purpose of making themselves seem distinctive or erudite.

Another problem with Jethro models is the way that their oversimplified, overgeneralized approach can enable self-serving hero/villain narratives. Andrew Gelman made a related point about many popular economics books and articles -- "What strikes me about this discussion is the mix of descriptive and normative that seems so characteristic of pop-microeconomics." You don't have to look hard to see that mix here -- you can almost hear the inspirational music in the background while reading this "if instead your purpose is to gain and spread useful insight, so that we can all understand more about things that matter, you’ll want to look for relatively neglected topics."

It should be noted that Robin Hanson spends a great deal of time on out-of-the-mainstream ideas. Without putting too fine a point on it, when someone who "has elected to have his head cryonically preserved in the event of medical death" depicts in such glowing terms people who discuss neglected topics, I can't help but suspect bias.

And given Hanson's tendency to portray himself as being above this sort of thing...

No comments:

Post a Comment