Tuesday, October 18, 2016

USA: how the polls could be wrong

There are three "ways" the US election could play out from here. We've seen that Trump support is relatively solid despite recent events, but that this level of support is not great enough to win.

As such I want to examine 3 scenarios that are all plausible and possible.

1 - The polls are right

This is the simplest one. Polls are correct, and the states that people expect will be won by one or the other candidate, will, in fact, be won by that candidate. Not much more to say beyond that.

2 - The polls are over-estimating Clinton

This presumes that a lot of people like Trump, but given all the negative coverage, are unwilling to admit it either to others or themselves. If this is correct, Trump could still win the Presidency. The explanation is simply that there are enough "old white male racists" who vote to push him over the top.

3 - The polls are over-estimating Trump

The basis for this conclusion is that a lot of Republicans think Trump is insane, but like above, don't want to admit to others or themselves that they are willing to vote for Clinton to stop him. If this is correct, the reason would simply be that people are "willing to vote for a crook" to stop Trump.

The fact that we have two candidates that are so controversial is why I'm unwilling to call a winner with any certainty. I'm thinking the polls likely are wrong, however, if both 2 and 3 are happening at the same time, the end result would likely be similar to what we are expecting. One way to see if we are seeing event #1 or event #2+3 is what happens in the "solid" states. States like Alabama, California, Oklahoma, and New York are all expected to be won by a lead of between 20 and 30 points. If these states are only won by 10 points, or even less, and that is not biased to one side (IE both 'losing' parties do better, not just one of them) then we can say that we are actually looking at the polls being wrong twice.

Either way, this chaotic election won't be decided until election day, and, as always, there's a chance that even then, we won't get our final answer.

No comments:

Post a Comment