About EKOS Politics

We launched this website in order to showcase our election research, and our suite of polling technologies including Probit and IVR. We will be updating this site frequently with new polls, analysis and insight into Canadian politics. EKOS's experience, knowledge and sophisticated research designs have contributed positively to many previous elections.

Other EKOS Products

In addition to current political analysis, EKOS also makes available to the public general research of interest, including research in evaluation, general public domain research, as well as a full history of EKOS press releases.

Media Inquires

For media inquires, please contact: Frank Graves President EKOS Research Associates t: 613.235-7215 [email protected]

2022 Ontario Election: A Retrospective

[Ottawa – June 6, 2022] Here at EKOS Research, we are proud to have successfully forecasted the outcome of the Ontario election correctly for the sixth consecutive time. We predicted that Doug Ford and the Progressive Conservatives would win a second majority based on our estimates of voting intention and seat projections. To their credit, they exceeded our estimates and won 83 seats.

Premier Ford has our deepest congratulations, and we wish him success as he leads Ontario over the next four years.

Our riding projections were reasonably accurate: we made 14 incorrect calls and had 89% accuracy. Historically, our riding projections have been just over 90% accurate. We are satisfied that we closely maintained that level of accuracy in what proved to be a trickier election to forecast.






Our prediction:






Final seat tally:






We knew early on that Doug Ford was cruising to a second win early on in the election and that he would win a majority within the final week of the campaign. That came as no surprise, as first term Premiers generally get re-elected in their first term.

What did come as a surprise was how it happened. The real story of the 2022 Ontario election is not that Doug Ford won a massive majority but rather that this election had the lowest voter turnout in the history of Ontario.

Low voter turnout, but good voter estimates:

While EKOS has an outstanding track record of producing accurate estimates of voting intention in Ontario (our 2018 estimates were almost perfect), we did underestimate the PC vote this time by 3.8% – which falls within the margin of error of that particular estimate (+/- 4.4%).

In 2018, however, 57% of eligible voters cast their ballots. This time it was 43%.

Anyone familiar with survey research will know the concept of coverage error. Coverage error is when one’s target population is different from the sampling frame. In this case, it would lead people who plan to vote in the election not to get surveyed. In the case of last week, what might have happened is that respondents who told us that they would vote ended up not going to the polls in large numbers. As most will know, sometimes people who have no intention of voting (and don’t participate in surveys) until the last minute will show up at the polls and vote. But that didn’t happen in this case.

We always ask if people intend to vote, and this question is particularly prone to social desirability bias. That is, people will often tell us that they are voting because they think that is the “right” thing to say and then don’t vote. We fully expect measurement error with this question. But we were pretty surprised to see it at this level with this level In fact, voter turnout was so low that it is remarkable that our voter estimates did not have greater differences between the actual voting estimates on election day. We know that our estimates can become askew when some demographic subgroups that usually do not vote (e.g., young people) go to the polls. If voter turnout were 80%, we would expect our estimates to be entirely off. However, that was not the case this time: the overall differences were rather small and concentrated on a single party.

What remains to be seen is whether this is a one-off or not. For the health of our democracy, we would strongly prefer to see voter turnout to be at least over 50%. And while we do not have mandatory voting in Canada, we would also prefer people to be able to give coherent accounts as to why they are voting for someone instead out of a weak sense of civic duty.

Looking back in retrospect, we found in our previous polling that neither Doug Ford, Andrea Horwath, nor Steven Del Duca had good favourability ratings. It is perfectly reasonable to believe that Ontarians decided not to go and select someone that they were not enthusiastic about supporting. Liberals weren’t keen on their leader, and neither were NDP supporters.

PC voters also expressed ambivalence towards Ford, but had somewhat less doubt working against him than Horwath and Del Duca.

In this one example that we hope never to see repeated, election polling can show some resistance to extreme low voter turnout. We hope that Ontario can do better next time and come out and vote in greater numbers. Maybe political parties can work harder and offer something that the voters might latch on to more. Or even better, the Ford government can introduce mandatory voting, so we survey researchers do not have to worry so much about coverage error next time.

Leave a Reply