Of Polling, Pollsters, and Perplexed Political Pundits

Another election; another “shockwave” and another round of mea culpas. The May 14 election in British Columbia caused consternation among the punditry and, coming just a year after a similar surprise in the Alberta election results, led to another round of calls for pollsters to explain themselves and to account for their (alleged) incompetence. Not just pollsters, but political scientists and pundits too: all these talking heads who predicted an NDP government led by Adrian Dix should be ashamed of themselves, right?

I am a political scientist – and a political junkie. I confess that I never went on television to speak about the election in my neighbouring province. Moreover, a good friend of mine is a pollster and I would regularly ask him how things are going in BC – he never personally conducted any polls there, but he is a guy I turn to for these things. But Twitter and Television were calling for “our” heads last night.

To all, I say this: I am taking a stand, and the pollsters are not to blame, for their job is not to elect the politicians. Instead, why not congratulate the thousands of BC Liberal Party volunteers who worked on getting out the vote (GOTV)? Why not acknowledge that Christy Clark travelled more and impressed the voters (and editorial boards) more than Dix did? And why not acknowledge that somewhere around four or five percent of the voters simple decided, in the last days, that economic stability was more important than “throwing the bums out”, the baby and the bathwater as well?

I say again: the pollsters were not the ones electing the candidates. Yet, pundits in the media profession seem to interpret polls that way. Polls are a snapshot of citizens’ attitudes and mood at a given time; and not all citizens either: online polls don’t sample older folks as well as they should, given their voting behaviour; land line phone surveys don’t capture all the under-35s who don’t and never have had a land line at home; and we all hate robo-surveys and we might not be giving the right information when the robots call.

Pundits and political scientists are not entirely to blame, either. Media need to fill 24 hours of news chatter each and every day, and they ask these folks to comment on the horse race – and many if not most political scientists have never been to the racetrack, so to speak. I don’t know many university professors who have run for public office, who have worked the phone banks, or who have door-knocked or “main-streeted” with politicians (although Yours Truly and a few others such as my colleague Dr. Ted Morton actually do). They usually don’t know much about political operations and war rooms on a personal level. They read the polls; infer what they can, and try to satisfy the Media Beast’s hunger for on-air punditry.

I will be a minority holding the opinion I do about polling, pollsters and the political pundits: most journalists and citizens will demand answers, better methodologies and mea culpas. These analyses won’t hurt, and improving polling models and methodologies will contribute to better social science. But Economists get GDP predictions wrong all the time; meteorologists get tomorrow’s weather wrong as often as not; sports experts predict the wrong winner of various NHL playoff series on a regular basis; and what to say about today’s Horoscope – should I buy that lotto ticket or not? So why pick solely on the pollsters? The other night, the voters elected the politicians, and the voters got it right.