Journal Courier | Commentary: Polling is the Magic 8 Ball of news


The last 12 months have been a catastrophe for pollsters, and I couldn’t be happier.

A little over a year ago, millions of people in the United Kingdom went to bed assured that the Brexit vote would fail and Britain’s connection to the European Union would be secured. The polls were fantastically wrong and the forces behind the Remain campaign have been in shock ever since.

Last September, CNN, Fox News, the New York Times, The Washington Post and others cited polls indicating that Donald Trump’s campaign was doomed and that the Hillary Clinton juggernaut was unstoppable. We now know that the polls were wrong and when the totals were summed up after Nov. 8, the Electoral College totals weren’t even close.

In four special congressional elections, some news organizations were predicting close votes or losses for the GOP. On what did they stake their claims? Polling data. They were wrong, again.

In April, the ruling Tories in Britain called a snap election based on polls that showed them way ahead of their Labour rivals. The polls were wrong again. The Tories lost their majority in Parliament and had to settle for a coalition government at the very moment when the government was about to enter into negotiations with the European Union.

A few weeks after Trump’s inauguration, a reporter for The Washington Post announced on NBC’s “Meet the Press” that Trump’s favorability numbers were simply terrible. On what was he basing these numbers? Polls.

How many more times will these supposed experts get it wrong before readers and listeners start treating their prophesies like horoscope predictions instead of news?

The only thing that makes polls credible are their predictive value and the last 12 months have demonstrated that the predictive value of political polling has sunk to the level of Ouija boards and tarot cards. Simply put, polls are not news, they’re speculations propped up by statistics and shouldn’t be presented on news programs and in newspapers as if they were matters of fact.

Unlike weather forecasts, which take into account measurable factors like temperature, pressure and humidity among others, political polling deals with that most maddening of complex creatures, human beings. Weather fronts are not known to be consciously deceitful, but people sometimes are. Trickery and deceit can go into the questions asked as well as in the answers given. And then, of course, the numbers can be fudged.

When evaluating political polling results we should continually remind ourselves of the principle behind the Harvard Law of Biology, which states: Under the most rigorously controlled conditions of pressure, temperature, volume, humidity and other variables, the organism will do as it pleases.

If the numbers of those who refuse to participate in polling goes up, the accuracy of results about complicated attitudes of the electorate may be compromised. As the number of telephone landlines goes down relative to the population, the potential sample of respondents may make polling via landline nothing more than a fallacious hasty generalization.

I’m a critic of using polling as news partly because the results do not give good reasons for selecting this issue or candidate over another, it’s a sort of institutionalized example of the ad populum fallacy.

Polling only tells us which way the herd is going at a given moment — and now it seems they can’t even get that right.

Jacksonville resident Jay Jamison writes each Friday for this page.

Source