Tuesday, October 30, 2018

Election predictions

A common theme here is that I do not trust the polls.  They have been increasingly inaccurate - ignoring the possibility that some of them are designed to mislead.  There seems to be an increasing resistance among the population to answering pollster calls - a New York Times poll had to call 38,000 people to get 700 who would reply.  Don Surber dissects this, and this is the key graf:
Gallup meanwhile has given up on election polling. The grand-daddy of polling admitted it cannot figure out how to get an accurate way to find out how people are voting.
Bottom line: the polls are not helpful forecasting this election.

It's made worse because the interesting race is for the House.  There are only 50 States to poll about Senate races, and there are only about a dozen that are close enough to bother.  In the House, there are maybe 50 districts that count.  Each of these districts contains 711,000 people.  To get a meaningful poll (i.e. one with a margin of error less than 5%), you might need to call 20,000 people.  The cost of polling the key House races will be perhaps ten times as expensive as polling the meaningful Senate races.

Since Senator are likely more powerful and important than Representatives, House polling is not going to happen as often, or be as thorough as Senate polling, because it's a lot tougher to justify the spend when compared to other options (say, Get Out The Vote efforts).

So the first takeaway is that nobody really knows what's going to happen in the House elections.  The polls are sporadic and of questionable accuracy, and so this is really in the realm of informed guesswork.  I have noticed that the closer we get to the election the less sure each side sounds - both the White House and the Mainstream Press/Democratic Party (but I repeat myself) have been making noises in the last couple days that seem intended to cool their supporters expectations.  Maybe the Democrats will take back control of the House, maybe not.  I don't think that anyone remotely knows.

The Senate has always looked like the Republicans would keep the Senate.  The closer we get to election, the more the Republicans look to pick up seats from vulnerable Democrats.  But again, the margin of error for some of the races is very thin, and the polls are of questionable value.  It seems certain that the GOP will pick up a couple, maybe 4 or 5.  It's unlikely they'll pick up the 9 they would need to theoretically avoid a filibuster.  I say "theoretically" because with the likes of the Susan Collins wing of the GOP there wouldn't be a solid 60 GOP votes for almost anything.

Increasing a GOP majority in the Senate will tend to weaken the Susan Collins types, though, which will help reduce gridlock there.  A thin GOP majority in the House will probably result in the rise of Susan Collins types there, and a Democratic majority will basically ensure gridlock for the next two years.

So it really comes down to the House which nobody can handicap with any sort of confidence.

No comments: