Sunday, August 7, 2016

The frustrating dearth of state-by-state polling

The theme of today's post was foreshadowed by the morning music.  If you head over to RealClearPolitics, you'll see one Florida poll with August data, and nothin' for Ohio with August data.  We are truly in the general election phase, and despite the fact that we don't actually have a national popular election, we have virtually nothing in the way of state-by-state polling yet.  How... irritating.

And yet, there's social science to discuss here.  Buzzword alert:  "endogeneity."  At this point, if my grad school advisor, Nelson Polsby, were around, he'd make me pay him a quarter for using it.  He hated the word, but it's kind of useful.  Basically, it refers to a situation in which two variables are affecting each other so that both are cause, and both are effect.

What makes Florida and Ohio special?  They are relatively large states with evenly divided populations, so smart campaigns will focus their attention here (OH) and there (FL) because that's where they get the most bang (electoral votes) for the buck, to the limited degree that the buck gets anything, which isn't much (I'll keep coming back to that, I'm sure).  But, if the campaigns also affect the votes in the states, then we have.... endogeneity!  The characteristics of the states draw the campaigns, and the campaigns affect the closeness of the states!

At this stage of the game, we don't know for certain where the campaigns will focus, so we don't know where to poll.  Now, we sort of  do.  Ohio, and Florida.  Beyond that?  There's a weird sort of dynamic process by which the campaigns feel out where to focus their attention.  A poll here, a focus group there, and a targeted ad campaign here, an independent expenditure campaign there...  Right now, we have some idea where the battleground states will be, but there is guesswork involved.  Except for Ohio and Florida.  So, you know, some polls here, please!

But to be blunt, polls cost money, and there is little payoff for the organizations to conduct extensive state-by-state polling at the moment.  Yes, junkies like me, and probably you, need our fix, but at this stage, until the precise battlegrounds begin to take shape, nobody has the incentive to do the kinds of polling that we will begin to see in September and October.  Until then, we're stuck with the national polls.

Those of us fixated on polls, though, can console ourselves with the fact that if one candidate has a lead as large as Clinton's is now, a popular vote-electoral vote split is nearly impossible.  Not completely impossible, but it would be really, really weird.  Then again...

For now, just understand that a popular vote-electoral vote split usually requires a closer election than the 6.9 point lead Clinton currently has at RCP.  Those of us craving state-by-state polling just need to understand that we are the ones with a problem.

I'm still annoyed.



Get it?  No?  Fuck.

No comments:

Post a Comment