May 07, 2018

New data: As first debate looms, Ford and PCs dip as NDP rises.

Authored by:

David Coletto

David Coletto

Ihor Korbabicz

Ihor Korbabicz

Campaigns matter. We hear it all the time. And it seems more and more they do.

Partisan attachments among voters are weaker. The news cycle is shorter. Politics is more leadership and personality driven and as a result, voters are more volatile than they used to be. All this makes what happens during a campaign matter more than in the past.

That’s why the research team at Abacus Data, in partnership with our friends at Summa Strategies and Spark*Advocacy decided to approach our polling in this election differently. Yes, we will ask people how they plan to vote. But more important, we wanted to understand campaign dynamics better than pollsters usually do. We want to go beyond the horse race.

So we did something that’s rarely done in Canadian public polling, we designed and executed a panel study.

So we did something that’s rarely done in Canadian public polling, we designed and executed a panel study.

In our first survey, we interviewed 4,685 Ontarians eligible to vote. All the content we have released on the ONPulse website over the past month comes from that survey conducted from late March to early April.

Last week, we invited as many of those respondents as possible to come and answers more questions about the election. In total 1,275 completed our second survey. The benefit of a panel study is we can track changes in attitudes and intended behaviours at the individual respondent level. We know how some felt about Kathleen Wynne four weeks ago and can see how impressions of her changed with the exact same respondents over time. Our goal is to track the views of a sample of respondents throughout the campaign.

It’s the way we approach our work with clients and part of Abacus Data’s record of innovation in our industry.

This way, we will be able to track the evolution of Ontarians’ minds in a way not typically done in publicly released election polling. It’s the way we approach our work with clients and part of Abacus Data’s record of innovation in our industry.

To ensure that our results are representative and are capturing broader shifts in the electorate, we also interviewed a new sample of 500 Ontarians eligible to vote. Both samples, the panel and the new sample, are weighted independently so each matches the Ontario population. Results are comparable across the two samples and I’m confident they provide us credible estimates of what the public is feeling and thinking about this election.

Wave 1 of our study was conducted from March 29 to April 8, 2018 with a random sample of 4,177 panelists.

Wave 2 was conducted from April 30 to May 6, 2018. 1,275 respondents from Wave 1 were recontacted and completed the survey. From May 4 to 6, an additional 500 new respondents completed the survey to create a total sample of 1,775 respondents.

Now, to the results!

Interest in the election is up slightly from the first wave. 67% of Ontarians are following the election very closely or somewhat closely, 4 points higher than in early April.

Wave2 1

The NDP’s pool of accessible voters has grown while the PC pool has shrunk. The Ontario NDP now has the largest pool of accessible voters (62%), growing 4 points since our first wave. The PC pool is down 5 to 54% while the Liberal pool marginally increased by 2 points to 46%.

Wave2 2

The PCs still lead on vote intention, but the lead is down to 6 as NDP support increases. When asked how they would vote, among decided voters, the PCs would get 35% (down 5) with the Liberals and NDP tied for second at 29%. The Greens get 5%. The NDP vote share is up 5 points while the PC vote is down 5. See below for comparisons across sample types (panel vs. new sample)

Wave2 3

Much of the NDP gains have come from previously undecided voters. One of the benefits of a panel study is we can see how preferences and intentions change over time among the same respondents. Among those who completed both waves, we find that the vote has been fairly stable from one wave to the next. Each main party has seen some minor attrition to the other parties. More notable that about 8 to 9% of previous party supporters now say they are undecided while a significant portion of previously undecided voters now say they are have a choice.

The chart below reports current and previous vote intention of the panel study respondents. For example, the top left cell (25.1%) can be interpreted as 25.1% of the sample said they would vote PC in Wave 1 and 2. The cell below that (1.0%) can be interpreted as 1.0% of the sample previous said they would vote PC but now say they will vote Liberal.

The bottom row (Total Attribution) represents the movement in vote intention from the first wave to the second. The PCs and Liberals had about equal numbers of people shift intentions (4.6% to 5.0%) while the most movement occurred among those who previously said they were undecided (8.8%).

The NDP gained vote share largely because it held more of its previous support while attracting more new supporters to its ranks (6.8% gain – 3.4% attrition)

Wave2 7

Regionally, the Liberals lead in Toronto and Eastern Ontario, the Tories have a sizable lead in the GTA while the NDP is ahead in southwestern Ontario.

Wave2 8

Demographically, the Tory lead province wide is built almost entirely on its lead among older men.  The PCs have a 14-point lead among men 45 and older while they are tied or trail slightly among younger men and all women.

  • Among those aged 18 to 29, the Tories are in third behind the Liberals and NDP who are tied. Discounting gender differences, the Tories lead among every other age group.
Wave2 10
Wave2 9

Comparing current vote intention to vote in the 2014 provincial election, we find that the Tories and NDP are doing the best retaining their previous support while the Liberals are only holding two thirds of their past supporters. Among those who said they did not vote in 2014, the parties are essentially tied.

Wave2 11

Horwath is reaping most of the benefits of Ford suppression/fatigue, but Wynne has also benefited. Ford’s base hasn’t shrunk, but the Liberal emphasis on the Tory leader, and some of Ford’s doing (flip flopping on the Greenbelt), appears to have increased his negative rating. Horwath has been the greatest beneficiary as her favourability rating bumped up 10 points, and even Wynne has seen a 5-point bump in her positive assessment and a softening of her negative numbers. 

Wave2 13

Finally, and perhaps most important, we find no shift in the desire for change among the electorate. Today, 80% want change, 59% intensely so and this explains why support for the Liberals and Kathleen Wynne has moved little despite a drop in PC vote share. Unless this desire change dissipates or becomes less intense, it will be difficult for the Liberals to gain any traction. Efforts to discredit and demonize Doug Ford appear to have so far helped the NDP and have done little to improve Liberal support.

Wave2 12


We will have more analysis of Wave 2 of our survey over the coming days including a look at issues, perceptions of the leaders, and what issues are grabbing and holding people’s attention during the campaign.

But as the leaders are set to square off in their first debate, a several things appear clear:

  1. The desire for change remains high and intense. 80% of Ontarians still would prefer a change in government.
  2. The change vote could be volatile. PC support is down largely because undecided voters are moved substantially to the NDP in the past month. While the PCs have lost some support over this period, most of the NDP gains have come from previously undecided voters moving to the NDP.
  3. As Mr. Ford’s negatives rise (more to come on this later this week), the Liberals are not likely to benefit. The NDP continues to be in the best position to take advantage of the desire for change and growing concerns about Mr. Ford.
  4. The PCs still hold the advantage. Their support is higher among more reliable (ie. older) voters and they continue to hold a large lead in the GTA.
  5. But this campaign is far from over and we’ll be here to capture the dynamics as they emerge.

If there’s a large audience for tonight’s debate on CityNews, the stakes could be high.

Ford needs to demonstrate he’s up to job of being premier and not offensive to the 10% of the electorate that is persuadable and needed for a PC majority.

Kathleen Wynne needs to find a way to get people to see her in a different light. She has to reconnect her motivation and drive to be a public servant with her vision for the province.

Andrea Horwath, more than anyone, must find a way to connect with the large number of Ontarians open to voting NDP. Right now, they are inclined to like her but those impressions are built on flimsy foundations.

Watch this space throughout the week for more on what Ontarians are paying attention to and how party tactics and message frames have been affecting the race. 


For those interested, there was no statistically significant difference in vote intention among the panel respondents and new respondents. 

Wave 2 4


The survey was conducted online with 1,755 Ontarians aged 18 and over, from April 30 to May 6, 2018. The sample is composed of two parts. 1,255 respondents who took part in our first survey of the election were completed the second wave. An additional random sample of 500 panelists were invited to complete the survey from a set of partner panels based on the Lucid exchange platform. These partners are typically double opt-in survey panels, blended to manage out potential skews in the data from a single source.

The Marketing Research and Intelligence Association policy limits statements about margins of sampling error for most online surveys.   The margin of error for a comparable probability-based random sample of the same size is +/- 2.4%, 19 times out of 20. 

The data were weighted according to census data to ensure that the sample matched Ontario’s population according to age, gender, educational attainment, and region. Totals may not add up to 100 due to rounding

Access more, more often

Subscribe to follow driving factors, receive customized insights and research findings that help provide important context along the 28 day journey to Queen's Park.

Never miss an update.

Join the ONPulse mailing list for updates as they occur — you can unsubscribe any time: