Friday, April 22, 2016

The vast majority of voters don't vote in the primaries and we don't know what they are thinking.

The following is from the latest column by Timothy Egan.

Almost two-thirds of voters — Democratic and Republican majorities — agreed with the statement that “The old way of doing things no longer works and we need radical change,” when asked in a recent Quinnipiac University poll. This is not a frustrated fringe.

I like Egan and enjoy reading his pieces.   But he is making an error here.  And I believe it is an error that is propagated over and over again, election after election, though it matters more so here because of the sort of inference that Egan makes. The error is based on WYSIATI (what you see is all there is) a cognitive bias we as humans are inclined to make, as discussed in his book Thinking Fast and Slow.

In this case the issue is whether people who respond to being polled, as a group, are similar to those who would not respond even if they were polled.  WYSIATI then encourages us to look at poll results and make inferences about the entire population, including those who would never respond to a poll.  This is okay when the two populations are essentially the same in their preferences.  It produces a biased conclusion, however, when the two populations are quite different.

Under 35% of the electorate in New York participated in the recent primary.  This is especially noteworthy because it was reported that participation was high.  (The measure is relative to participation rates in previous primaries.)   Presidential elections have been running at somewhere between 50% and 60% in voter participation.   Are those who don't vote in the primaries but do in the general election different from those who do?  One might guess that the former group has many more independents.  Does that matter?

Another issue is how preferences are elicited in a survey.  What are the alternative possible responses?  Do those alternatives allow us to discern voter preference well?   Here let's observe that what Egan presented is a conjunction of two distinct thoughts: (1) the old way of doing things is not working, and (2) we need radical change to fix the problem.  Were people asked about the possible alternative conjunction where (2) is replaced by: (2') I don't know how to fix the problem and I'm frightened that others will try things in an attempt to fix the problem but actually make things worse.
Again I'm guessing here without data, but my supposition is that the voters, even those who wouldn't participate in the poll, would agree with (1) in great numbers.  As to how they'd split between (2) and (2') if offered those alternatives, let me just say here that it is this question where I suspect being an independent matters, a lot.  My supposition is that among the independents there are a lot who'd agree with (2').  But I'm quite willing to admit that is just a guess.

Let me conclude with a brief consideration of the upwards of 40% of the electorate that won't even participate in the election this November.  Do their preferences matter and, if so, how should they be accounted for?  It is necessary to include this group to make the title of my post accurate, if for no other reason. 

Elections in our country have turned into very nasty affairs of smear and disinformation.  I watched this panel on the Charlie Rose show discussing the election.  One of the panelists was Ed Rollins, a Republican consultant, and he predicted that the coming campaign is likely to be much worse on this score than anything we've seen previously.  Nobody else on then panel challenged Rollins on this point.  Such a dirty campaign is a turnoff to many.  Yet it seems to be the old way of doing things, now on steroids.  And it seems to be inevitable.  Even the race for the nomination within the Democratic party, which started out in a fairly collegial manner - Americans are sick of hearing about Hillary Clinton's damn emails - has gotten much nastier as of late.  What economists would call a revealed preference argument as applied to the campaign itself, suggests there might be some creativity in how one candidate can be nasty to opponents, but on using the campaign to actually educate the public there is no change at all.

The non-participants aren't seeing anything that would change their minds on that score.  And, frankly, neither are the rest of us. 

No comments: