Wanderings – What are political polls really worth?

Opinion polls are fascinating things to read about. Take a random group of people, ask them some questions, then maths people do some work and extrapolate this tiny sample of people to project a result on a much larger audience. It’s voodoo math for politicos and marketing people.

We’ve seen a lot of polls in the past 12 months leading up to the recent federal election call. Before Justin Trudeau announced his resignation, it looked like the guy who rescued the federal Liberals from third-party status to government was going to sink his party as low as the Ontario Liberals did in 2018. Forget a small caucus room or office space, the party was heading for a future of pushing two tables together at the Tim Hortons to hold their meetings. If anyone has ever told you that quitting does no one any good, they don’t follow Canadian politics.

Conservative leader Pierre Poilievre and his party led the polls by over 20 points. The party was on auto-pilot for a super-majority of Brian Mulroney proportions. In two months however, that has all changed – sort of.

Donald Trump returned to the White House and started a trade war against the U.S.’s continental best friend – us. People realized the outgoing Trudeau could stand up for himself, and Canadians – despite the opinions of some. And Canadians realized that we should have some national pride and not just buy our cheap knock-off stuff at American-owned stores, instead of supporting Canadian businesses where possible.

The polls have reversed, and the leaderless Liberals ascended – mostly at the expense of the NDP and Bloc Québécois – and that cut the polls to a veritable tie with the Conservatives.

Once Mark Carney won the Liberal leadership, pollsters reported that the Liberals were not just tied with Conservatives, but led the polls – red team members rejoice. In some cases, the red and blue teams are within the plus/minus error margin of the poll. But what does it all really mean?

It means that the small sample of people from which polling companies draw say that there is change afoot. Of those small groups surveyed, they are all changing their minds, or they are messing with the pollsters.

When the polls first started reversing in favour of the Liberals, most polling firms did not reflect this. There was just one. As the weeks went on, more firms’ results showed change was on the rise. But it’s still a small group. So how much faith should we put in what polls say? Little to none, except when they are right.

In grade school, students are told to come up with a survey about something: who likes Strawberry Jam and who likes Grape Jelly on their peanut butter and jam sandwiches. The students ask each other, take the information, and put the data in a graph or a chart. Fifteen students in Ms. Perkins’ Grade 3-4 class liked Grape Jelly, 12 students said Strawberry, and one kid said he preferred Nutella. From that survey, the students can say a majority of kids like grape; Or that Strawberry is on the decline; Or that one person thinks outside the box, questioning the existence of peanut butter in the equation. On a different day, those same people could answer in a different way.

Take that analogy, and reflect on politics again. Swap jams for politicians, and it’s the same inaccurate mess. A poll is a snapshot in time the second the question is answered. Those opinions change at the drop of a hat. A person who just got their tax bill in the mail may not feel favourable towards a governing party; just like a person may not feel happy about a politician after hearing them speak, or seeing a commercial. The person polled may be fighting to get their kids and their dog in from the car and is annoyed being asked questions by some random caller. The only poll that is 100 per cent accurate is the one on voting day.

Should you take stock in polls? No. Nor should you take stock in these recent polls. Poilievre’s support has not dropped much since Trudeau announced he was leaving. That means his base is solid. Support returned to the Liberals at the expense of the NDP and others, meaning some Liberal supporters were parking their vote elsewhere. That shows that the Liberal base is coalescing around an “anyone but…” campaign.

Fairweather support is a fickle thing. If those voters don’t think the bandwagon they are jumping on is actually going to win, many will not turn out to vote. Polls or no polls, converting votes and higher turnout on the ground is the key to winning any campaign.

George Gallup, considered one of the fathers of public opinion polling said that “Polling isn’t about predicting the future; it’s about understanding the present.” Polls showing the Liberals and Mark Carney leading should be taken with a grain of salt. In a 37-day election campaign, that’s a lot of days to string together for a win.