Do polls prove their point?

Reading a recent newspaper article about the pollsters congratulating themselves on their accuracy in the last provincial election was enough to make me ill. The more serious question on that provincial election was whether the polls followed the voters or did the voters follow the polls?

Opinion polls, focus groups and voter profiling used to be handy tools for campaign management but in the hands of the news media, they have been weaponized. Various media outlets seem to have their own tame pollsters to support their editorial stance. The Canadian Broadcasting Corporation (CBC) even has its own compiler who takes all the polls and comes up with an average. This brave soul also tries to forecast the number of seats in an imaginary legislature.

The problem with this is that 19 times out of 20, the pollsters manage to get certain things wrong. They say they are adjusting their algorithms to compensate for some of the errors. In that case, I suspect the errors must be growing faster than their corrections.

We have always known, for example, that the NDP vote will be exaggerated. Early in the campaign, it is the parked votes of people who are not sure of how they will ultimately vote. Later in the campaign, people just lie.

And they never have been able to determine who among our young people are likely to vote. The campaign manager that fails to develop a strong youth movement for the candidate is not very good at the job.

But who is likely to vote by election day is always the key question. And when you are dealing with 40 per cent or more non-voters, what can any opinion poll really tell you?

The guys who do surprise me are the interactive voice response (IVR) pollsters who use high volumes of calls to try to correct their built-in errors. Who do you think answers the hard-wired phone in a household replete with children and youths? If I was a four-year old picking up an IVR call, I would have just as much fun as an adult playing with the phone buttons.

Campaigns keep changing. You can never judge by what happened in the last election. As I always found as a campaign manager was that three or four afternoons with carefully selected poll sheets could give me the best idea of what was happening. It gave you direction.

-30-

Copyright 2018 © Peter Lowry

Complaints, comments, criticisms and compliments can be sent to  peter@lowry.me

Comments are closed.