Thursday, February 23, 2017

Beware of Pollsters Forecasting (Wrong) Election Results

Pollsters have fallen on hard times.  They were wrong on Brexit (Britain’s withdrawal from the European Union).  They were wrong on Trump.  They are likely to be wrong on Marine Le Pen, who is campaigning for president of France.  Why have the pollsters been, and are likely to continue to be, wrong?

Polls are taken for two purposes.  One is to sample public opinion on candidates, elected officials, policy choices, etc.  The second is to make news, to use the results of polls to emphasize a political trend or point of view for or against a specific politician or policy.

Years ago, most polls consisted of the first kind, to know what the public really thought.  Today, most are of the second kind, to try to sway public opinion in favor of one candidate or set of issues over others.

But it’s not the case that poll results are wrong half the time in one direction (say, liberal) and half in the opposite direction (conservative).  Rather, they are increasing wrong in favor of the liberal-centrist established order.

There is in the United States and other Western democracies a left-leaning (with a bit of centrism mixed in) “Academic, Media, Political Industrial Complex,” APMIC for short.  This interwoven political system is guilty of lying to the electorate, as President Trump would say, BIG TIME.  The voter’s trust in politicians sits at, or near, an all-time low, mirroring that of the main street media and leftist professors holding court in colleges and universities.

The pollsters, with few exceptions, are an integral part of APMIC.  The political establishment is its clients.  It’s hard, for some impossible, to acknowledge that one is part of a dishonest (lying) enterprise.

Pollsters have gotten recent elections wrong because voters have taken to lying to them, either face-to-face, over the phone, or online.  The pollsters don’t know how to compensate for lying.  It’s not part of standard polling methodology.

If pollsters were honest and admitted that they have no scientific basis to correct their measurements for lying, they would be hard pressed to earn a living.  Better to blame a wrong forecast on a “shock” they couldn’t measure or anticipate, and which their colleagues also got wrong.

With few exceptions, pollsters want their results to support APMIC.  As members of APMIC, they are contemptuous of nationalist movements, denigrating them and their leaders by calling them populist.  A survey of the pollsters themselves would likely show that they accept an outcome in which Marine Le Pen will come in first in the initial round of the April French presidential election.  However, they will forecast that she will lose by a large margin in the second round to the more mainstream second place winner in a runoff election two weeks later in May.  Your friendly proprietor thinks that pollsters will be wrong on Le Pen, much as they were on Brexit and Trump.

Stay tuned.

No comments :