John DeCicco, back in the days of the famous Barbershop Poll. (Image: Mel Rothenburger file photo)

By Mel Rothenburger

ARMCHAIR MAYOR

ROTHENBURGER: Gut instinct might still be the best election poll of all

Oct 12, 2019 | 6:51 AM

WHILE THE REST OF US make election predictions based on the science of gut instinct, idle speculation and wishful thinking, there are actual experts at work using more sophisticated means.

They’re called pollsters. Over the past few elections they lost a lot of their credibility with some wildly inaccurate predictions. As the old landline faded from its place as the number one way we communicate, the pollsters fell behind.

I’m not sure they’ve gotten much better.

I’ll never forget the B.C. provincial election of 2013, when every pollster on the planet was predicting a relatively easy NDP win. Only John DeCicco got it right.

John, of course, used to run his “Barber’s Poll” at the Continental Barber Shop. People would come in and write their choice on a piece of paper and put it in his makeshift ballot box. The Barber’s Poll forecast a BC Liberal landslide. Everyone thought it was nuts.

1000 Ways to Heal

1000 Ways to Heal: Moons

Dec 10, 2024

Question of the Week

Should Mayor Reid Hamer-Jackson apologize or resign after it was found he breached a section of council's code of conduct for allowing false or misleading information to be communicated?

Dec 09, 2024

One Man's Opinion

COLLINS: Donald has it wrong - again

Dec 08, 2024

On May 14, 2013, the Liberals sailed to victory, increasing their majority, while the NDP dropped a couple of seats. Both Terry Lake and Todd Stone won better than 50 per cent of the popular vote, leaving their NDP opponents in the dust.

I tweeted at the time that the big polling outfits should have hired DeCicco.

Ever since, the pollsters have been struggling to get back in the game. They still seek representative samples, they still report results as plus or minus margins of error, and questions are still formulated the way they always were.

But they’ve broadened their sampling methods. They use things like opt-in on-line sampling, and have changed the way responses are weighted. It may not be any closer to the truth than armchair or coffee-shop predictions.

There was, for example, the embarrassment of Brexit in 2016. In 2015, the pollsters were close on the popular vote in our federal election but were way off on the seat count. Justin Trudeau was supposed to win a minority, not a clear majority.

So what about 2019? John DeCicco is retired, so in the absence of the Barber’s Poll I like to look at two aggregators, 338canada.com and CBC’s Poll Tracker. They compute averages from all the pollsters — Ekos, Angus Reid, Nanos, Leger, all of them.

The two sites, though they undoubtedly use different methods to get to their results, are very close in their daily snapshots. A couple of days ago their popular vote percentages for the Liberals and Conservatives were exactly the same — Liberals with 33.9 per cent and Conservatives with 33 per cent.

Those numbers have varied only slightly over the past few weeks, with occasional bumps whenever one leader or another has been hit with a controversy.

Friday, Poll Tracker showed the Cons down slightly to 32.7 per cent, but ahead of the Libs, who dropped to 32.2 per cent.

According to Poll Tracker, the reason is the uptick being enjoyed by the Bloc Quebecois, which is eroding Liberal support more than it’s eating into Conservative support. Poll tracker now says the Liberals and Conservatives have roughly the same chance of winning the most seats, though neither with a majority.

That’s a dramatic change in recent days from earlier polling that showed the Liberals with a comfortable edge in seats.

The 338Canada site, on the other hand, continues to show the Liberals winning more seats than the Conservatives. Only at one point since late August, on Sept. 24, did 338Canada’s aggregating show the parties in a dead heat for seats.

What’s interesting about Canada338 is that it breaks things down by province and even by riding. For example, as of Friday, it has the Liberals taking 13 seats in B.C., and the Conservatives 16 or 17, and the NDP eight or nine.

Kamloops-Thompson-Cariboo is labeled a “Likely CPC” win. Unfortunately, no sample sizes are given but, then, that prediction comes from collecting results from several pollsters.

Percentage-wise, 338Canada sees Cathy McLeod ahead with 37.7 per cent of the vote, followed by Terry Lake with 25.2 per cent, the NDP’s Cynthia Egli with 21.4 per cent (after the NDP vote dipped to 18 per cent Sept. 30, and down from 31 per cent in 2015), the Greens’ Iain Currie with 11.4 per cent (up from 3.5 per cent in 2015), and the PPC with 3.9 per cent.

The Communist Party and Animal Protection Party don’t show up.

I don’t believe it. I’m sure the Greens don’t, either, even though their numbers are up. They have high hopes for Currie, but they have to keep in mind that the maximum number of seats being forecast for the Greens across the entire country is three or four. Will Kamloops be one of them?

It’s the race between McLeod and Lake where the numbers don’t make sense. The 338Canada results show the Liberal share of the vote actually dropping from the 30 per cent it was in 2015, with the Greens increasing by as much as seven per cent. The Conservatives actually gain a couple of points over 2015.

This is where the science of gut instinct conflicts with the science of polling. I just don’t see Lake losing five points from what Steve Powrie got four years ago, or trailing McLeod by 9,000 votes.

One thing that looks more certain is a minority government of either Liberals or Conservatives, but we’ve been there before. Keep in mind that in 2015 Poll Tracker predicted the Liberals getting between 124 and 161 seats, the Conservatives between 100 and 139, and the NDP between 51 and 90.

The Liberals ended up with 184, the Cons 99 and the NDP 44.

This time, Poll Tracker is predicting between 92 and 203 for the Libs, the Cons between 94 and 184, the NDP between four and 50 and Bloc between 14 and 45. That’s a lot of wriggle room.and, of course, there’s still more than a week to go.

We’ll see on Oct. 21 if the pollsters are wishing John DeCicco would come out of retirement.

Mel Rothenburger is a former mayor of Kamloops and former newspaper editor. He publishes the ArmchairMayor.ca opinion website, writes five commentaries a week for CFJC and is a director on the Thompson-Nicola Regional District board. He can be reached at mrothenburger@armchairmayor.ca.

—

Editor’s Note: This opinion piece reflects the views of its author, and does not necessarily represent the views of CFJC Today or the Jim Pattison Broadcast Group.