Almost all published public opinion polls have shown
significant improvement for Republican nominee Donald
Trump in the past ten days or so. Does this mean that
these polls are now accurate?
Not at all. I have argued that most, if not all, public polls
are flawed this cycle not only because of technical
deficiencies and some pollster pollster bias, but also because
there is a “mutiny of the masses” this particular year which
is aimed at, among other targets, polling. I think there is quite
a bit of evidence that during the primary/caucus campaign
many voters who favored Bernie Sanders and Donald Trump
decided that they were not going to be successfully polled.
With a potentially significant number of disaffected voters
on both the left and the right, there is some probability that
most public polls are still not giving us a true picture of voter
sentiment.
This could change. Most polls become much more accurate
just before election day (when there are much fewer
“undecided” voters --- and when pollsters take extra steps
not to be embarrassed a few days later when the votes are
counted).
But the basic problem remains. If some pollsters get it right
this cycle while many others get it wrong, we will want to
know what it was they did to reach the mutineers and measure
what they were going to do.
-----------------------------------------------------------------------------------
Copyright (c) 2016 by Barry Casselman. All rights reserved.
No comments:
Post a Comment