Apr 16, 2018

The Undecideds: Could they make the difference in the Ontario election?

Authored by:

David Coletto

David Coletto

Too often we (pollsters and election watchers) ignore survey respondents who tell us they are undecided when we ask how they will vote. Our intense focus on the horse race clouds our judgement and we forget that often anywhere between one in five and one in four eligible voters in our surveys tell us they are uncertain who they will vote for.

Many undecided respondents are not likely to vote. But sometimes, they decide elections. An analysis I did following the 2014 Ontario election suggests they were critical in lifting the Wynne Liberals to their majority in the final days of the campaign.

So what are the undecideds thinking as we head into the 2018 Ontario election?

Here’s what our survey of almost 4,200 eligible voters in Ontario told us:

One in five Ontarians are undecided at this point. That’s as many who say they would vote Liberal or NDP if an election was held today.


Undecideds look similar to the rest of the electorate except that they are much more likely to be female (63% are women), slightly more likely to be middle-aged (aged 30 to 59) and are somewhat more likely to have obtained lower levels of education (high school or less).  Beyond gender (which is important), undecideds share many of the same demographic characteristics and live in the same parts of the province as those who have decided who they will vote for.


There’s no clear issue set they are more likely to care about. When we ask them to rank their top three issues, the issue set is almost exact to respondents who have decided who they will vote for. Health care, housing affordability, and jobs are the top three, but large numbers care about taxes, electricity prices, provincial finances, and honesty in government.


When we asked how they voted in the last provincial election, 34% report not voting, 26% can’t remember, and 20% said they voted Liberal. Nine percent voted PC and NDP respectively.


Like those who have decided, the clear majority of undecideds think it’s time for a change in the province. More than half think it’s definitely time for a change and another 27% think change would be good but is not important. Perhaps most worrying for the Liberals is the fact that 79% of undecideds who voted Liberal in 2014 think it’s time for a change.


This intense desire for change makes sense given that only 9% of undecideds approve of the job performance of the provincial government. Six in ten disapprove.


And it’s not just the government’s performance they rate negatively. 58% of undecideds have a negative impression of Kathleen Wynne compared to 33% who feel the same way about Doug Ford and only 9% who feel the same way about Andrea Horwath.


These results suggest that for the most part, the undecideds are open for persuasion. But as with other data we have shared thus far, the Liberals and Kathleen Wynne will have a much harder time convincing them to support the Liberals this time.

The real opportunity among undecideds rests with the NDP. Most undecideds have no impression of Andrea Horwath, most want change, and most are women. That’s an attractive combination for the NDP looking to grow support.

In a future post, I’ll nerd out a bit and share some experimental questions we asked that attempt to measure the intensity of support someone might have for a party which I think will add further insight the potential dynamics of the coming campaign.

But for now, let's not forget about those undecideds. They could decide the outcome of the 2018 Ontario election.

About the survey

The survey was conducted online with 4,177 Ontarians aged 18 and over, from March 29 to April 8, 2018. A random sample of 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 +/- 1.55%, 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: