Polls are for dogs

I generally ignore polls.  Which is to say that I do not put any stock in the estimates contained in them or the "analysis" of pollsters and pundits of the estimates.  Polling has become way too predictable.

There are Conservative friendly polling companies that generally show the Conservative Party is in good shape.

There are Liberal friendly polling companies that generaly show the Liberal Party is in good shape.

Then there is Nik Nanos who I believe is not friendly to any party but who is still often guilty of reading way more into the estimates of his polls than what is there, just like all of the rest.

Overarching all of this is Mr. Grenier and his poll aggregator.  When he was independent I found his work interesting, although I still did not put much stock in it either, but when he joined the CBC I knew that going corporate would degrade the quality of his work and I was correct.  And it did not take long.  I stopped looking at his aggregator well before the 2015 election.

Polling used to be done to provide people with an idea of what people thought about things, whether is be politics or whether they prefer Coke or Pepsi.  The polling that is done for product market research is still useful but political polling ceased being about informing people about public opinion and became about trying to sway that public opinion a long time ago.

You see around 50 years ago some social science research was published that indicated that polls could swing a small percentage of voters to the party that the polls said was leading.  So naturally, it was only a matter of time before political parties and their allies in the media would attempt to use that to their advantage.

The problem is 50 years ago there was only one major polling firm.  That was Gallup.  If a Gallup poll stated that the Liberals of Pierre Trudeau was leading in an election then it was more than likely true.  Now of course there are about a dozen polling firms publishing polls, with some completely contradicting the others.  No longer is there one generally trusted polling company there are many polling companies only trusted by a segment of the population that would agree with their estimates.

In other words 50 years on the social science research indicating that polls could swing votes is no longer valid.  There are too many polling companies, releasing too many contradictory polls to swing any vote.  People are overloaded with polling information so when they process it they are going to process it according to their preconceived notions.  Conservatives will cheer polls indicating the Conservatives are doing well and stating that the polls that indicate the opposite as wrong.  Liberals will do the same for polls that indicate the Liberals are doing well.

Meanwhile, those with no preconceived ideas of who they support will process the polls by generally ignoring them.  It is hard to take polls seriously when companies that are polling one population, Canadians, come up with wildly different results.  These are the voters most coveted by political parties but they are also the voters less likely to take what the polls say about the popularity of the policial parties into account when making their decision at the ballot box.

So John Diefenbaker was right, although he was about 60 years ahead of his time.  Polls truly are just for dogs.