CBC's polls analyst Éric Grenier answers your questions about the polls

CBC's polls analyst Éric Grenier will take your questions on the election and the polls.

    Hello everyone, I'm Éric Grenier, the CBC's polls analyst.

    Thanks for joining me on your lunch.

    With the provincial election less than two months away, the CBC has launched an online tool — the Ontario Poll Tracker — to follow the trends and make estimates about the final outcome of the June 7th vote.

    If you have any questions about the Poll Tracker, the upcoming election or polls in general, ask away!

    You can check out the Poll Tracker (and bookmark it!) here:


    At first glance, more recent polls show greater variability, but confidence interval in your plots doesn't seem to change for the polls. It does change for the seats projection, but not the way I'd expect. Any comments on polls variability?

    Thanks everyone for your questions. We'll be getting to as many as we can over the next little bit!
    Hi DrV — for the averages, the confidence interval is constant despite the variability in the polls. It is based on the historical error polls have had at the end of an election. For the seat projection, though, it all depends on how many seats each party is capable of winning. So, if a party's support is concentrated in only a few regions of the province, an increase in their support might not put more seats in play. 
    For the variation in the polls themselves, that is a little more puzzling. We've seen a Forum poll showing the PCs under 40%, an EKOS poll putting the PCs in the mid-40s, and a Mainstreet poll putting them at 50%. Because of the different polling dates, they could technically all be accurate. I think things will clear up once we start seeing more polls, as well as when Ontarians perhaps start tuning into the campaign in bigger numbers.
    Hi Eric,How much credibility do you give the argument that the 'youth' vote is under-counted given an assumed reluctance to answer automated polls?Is there any recent evidence to suggest this may be the case?

    Sadly, we're running this live blog from Toronto while Eric is in Ottawa. Unclear if he has his famous calculator with him...

    Eric Grenier's Ancient Calculator (@CBCcalculator) | Twitter

    The latest Tweets from Eric Grenier's Ancient Calculator (@CBCcalculator). Serving @EricGrenierCBC for an indistinguishably long period of time. Taught your TI-83 everything I know. #StayAncient. Markham, Ontario
    Hi David — we did see something like that happening in the Calgary mayoral election in 2017. But I think it can vary from campaign to campaign. If youth turnout is low in the provincial election — which it often is — then under-counting the youth vote wouldn't have much impact on the outcome.
    Pollsters weight their samples, however, so even if it is harder to reach younger voters in polls, pollsters are still able to make sure they carry the appropriate weight in the sample. The question is whether the young people who do answer polls are different from the young people who go out and vote.
    Have you calculated the probability of a second-place NDP finish vs. second place Lib finish?

    Hi AnthonyCouto — I have! That is one of the things I'm going to be tracking during the campaign. Right now, I have the odds of the NDP finishing ahead of the Liberals in the seat count at about 52%. That's because, despite the wider edge the NDP has in the average seat projection, the Liberals have a significantly higher ceiling, while only a slightly lower floor.
    Hi Eric, thanks for taking questions. Is it possible that Liberal support might be underrepresented, and what does Kathleen Wynne need to do to turn things around?

    Our Queen's Park reporter Mike Crawley just walked past. His latest analysis of the Ontario race: 

    Kathleen Wynne's latest problem: 11 Liberal veterans not seeking re-election | CBC News

    CBCThe number of veteran Liberal candidates on Kathleen Wynne's team is shrinking and that will make her task of winning re-election even harder than it already is.
    Hey Chuck — I suppose that is possible, but I'm not sure why that would be the case. Incumbents do tend to out-perform their polling, but usually not enough to overcome the kind of 17-point deficit the Liberals currently have in the polls.
    How do they turn things around? Their best bet is to galvanize the anti-Ford vote behind the Liberals. We know Ford is a polarizing figure and that a lot of Ontarians have a very negative view of him, but the problem for Wynne is that she is even more unpopular. The risk for the Liberals is that if they make too good of a case that one party needs to be the anti-Ford option, that they might make the case for the New Democrats.
    Tom Mulcair often said after the 2015 federal campaign that one of the problems the NDP had is that he had argued too much for replacing Stephen Harper, rather than arguing for electing Tom Mulcair.
    Hi Braeden — it is difficult to make that kind of estimate based only on region-wide polling. So far, the Greens do not have the level of support needed in that part of the province to be in a position to win Guelph. But you can't discount the possibility of Green Party Leader Mike Schreiner out-performing expectations. It will be something to keep an eye on, however.
    So the NDP has a chance to overtake the Liberals in the June election, but can NDP Leader Andrea Horwath make her party the top anti-Ford vote? (Photo: John Rieti/CBC)
    Hey Kyle — we don't have it in the plans to show those riding-by-riding projections on the Ontario Poll Tracker. We have to figure out a way to communicate projections at that level with the right amount of uncertainty, so people can understand that our estimates have a big margin of error down at the riding level. Since people often use riding-level projections for strategic voting decisions, it is important to make sure we are doing it right and that we aren't giving people a false impression of certainty.
    But I'm always happy to "show my work" and answer questions about individual ridings on Twitter! You can follow me @EricGrenierCBC.
    Hi Eric, is this election really a story of what happens in the GTA-Hamilton-Niagara region?(as my reading of the political tea leaves says Eastern and Southwest Ontario will strongly go PCs, and Northern will be split between the parties.)Are you aware of any defining issues/reasons for the strong PC support in this region?

    Remember, you can check out our new Ontario Poll Tracker any time you want to get a sense of the polling picture. 

    CBC News Ontario Poll Tracker

    CBC NewsThe CBC News Poll Tracker is your guide to following the polls in the 2018 Ontario provincial election.
    Hi David — I think the GTA-Hamilton-Niagara region will be key, but I'd also point to Toronto itself. If the PCs do well in Toronto, they will almost certainly win a big majority. If the party fails to pierce the city (aside from, say, Etobicoke North, where Doug Ford will run) then a majority would be in doubt.
    But the PCs definitely need to win big in the GTA in order to secure that majority. I'm not sure if a specific issue resonates with voters there, though. This campaign is setting up to be a change election, and in change elections issues take the backseat to leadership and the question about whether people want to give Kathleen Wynne's Liberals another four years. If the election becomes about policy instead, the Liberals have a shot.
    Eric, the polls show Doug Ford is the clear frontrunner, though as you noted there is some fluctuation in those numbers. So, how strong is Ford's support?
    I'd say a good chunk of the PCs' support is pretty solid. The party has been polling well over 30% for a few years now. The dissatisfaction with Wynne's government is so high that the PCs have a pretty high floor. But the unpopularity of Doug Ford (about half of Ontarians disapprove of him) means the PCs could be vulnerable to shedding some of that support. But it would go to the Liberals, stay home, or switch over to Andrea Horwath's NDP?
    How do various players try to manipulate polls & how do you deal with such efforts?

    Hi LdZorTar — pollsters have mechanisms to protect against these kinds of tactics. If respondents answer a poll too quickly or if the answers are suspicious (i.e. always choosing the first option), they are filtered out of the final count. 
    Some people lie to pollsters, but they are very few and they would cancel each other out. Pollsters do rely on the good will of respondents, and the fact of the matter is that there is very little lying when it comes to polling. People like having their opinions known!
    Eric- Can you please explain how margin of error works in polls and how to compare them with 3+ parties? What are some tips on how to beat read them?

    Doug Ford leads in the polls, but as Grenier notes, that support may be for the party more than its leader. (Photo: John Rieti/CBC)
    Hi JG92 — simply put (and if you want a longer explanation: Margin of error - Wikipedia) the margin of error is the kind of error that is expected from a sample of X amount of people. Think of it this way: if you use a spoon to taste a pot of soup, you can be pretty certain that your spoonful will be representative of the entire pot. Maybe it might have a tad more paprika or not enough onions, but it will still be close enough that you can judge the seasoning of the soup. The margin of error is like that. If you sample the population correctly, you should get to within a few points of the actual result.
    How to read them in the context of three parties, though, is something else. Generally speaking, you double the margin of error when comparing the gap between two parties, because the margin of error applies to both. If a party is up five points when the margin of error is +/- three points, the lead is not statistically significant. Though, it is still more likely than not that the party ahead is actually ahead.
    For three parties, you have to take this into account. It is possible that the +/- only applies to two parties (i.e., the PCs are over-estimated by three points, all of it coming from the Liberals) but that is not necessarily the case.
    But don't get too hung-up on the statistical margin of error. There is also the practical margin of error — people changing their minds, turnout effects, etc. A lead is a lead, but unless it is significant you can't take it to the bank, particularly when you consider how a few points here and there can flip a lot of seats.
    Even with a depressed 2014 PC vote Sarnia-Lambton, and Chatham-Kent went to PC? What does would it take for NDP to win those Ridings?

    The NDP vote has been holding pretty steady in the southwest, while PC support has spiked. For the New Democrats to make some gains in the southwest, they will need to have their support increase and for the PCs to come down. Because as it stands, the NDP is at risk of losing some normally safe seats because the PC vote has increased so significantly. At these levels of support, the NDP needs to focus on holding its seats rather than thinking of new gains.
    Eric, thank you so much for taking the time to do this. We're going to wrap up, but before you get back to crunching the numbers I'd love to know: what are the 3 key storylines you'll be following with just two months to go until election day?
    Some breaking #ONpoli news happening now: 

    Ontario PC party removes Kitchener-Conestoga MPP Michael Harris from caucus | CBC News

    CBCThe Ontario Progressive Conservative Party says it has removed Kitchener-Conestoga MPP Michael Harris from caucus after a complaint from a former intern and the alleged discovery of text messages "of a sexual nature."
    Thanks John, the story lines revolve around the three leaders.
    Can Doug Ford avoid the same mishaps that have hobbled the PCs in the past three election campaigns? We know that Ford has some very engaged supporters, but he is also very polarizing and can get himself in trouble. Ford can ride a change wave to a big victory — as long as he doesn't give Ontarians too many reasons NOT to vote PC.
    For Kathleen Wynne, the question is whether she can corral the anti-Ford vote. She has a lot of baggage to carry in this campaign, not just five years of her own government but the fact that the Liberals have been in power since 2003. The odds are already stacked against re-election, nevermind how unpopular the Liberals are. Can she make the case that if you don't want a Premier Ford, then the Wynne Liberals are the only option?
    And finally, for Andrea Horwath. This is her third campaign as NDP leader and, if it doesn't go well, probably her last. She has the advantage of being the only leader of the three main parties that is generally well-liked. If people come to the conclusion that Ford is not an option and that another four years of the Liberals is also out of the question, can the NDP take advantage of that? Horwath's disapproval rating increased during the 2014 campaign. She can't have that happening again.
    Okay, that's it for today! Sorry we couldn't get to all of your questions. We'll have to do this again. Thanks to everyone who sent in their questions and to those who stuck with us during the live blog. It will be an interesting campaign!
    Don't forget to bookmark the Ontario Poll Tracker. We have an easy-to-remember URL: cbc.ca/polltracker

    CBC News Ontario Poll Tracker

    CBC NewsThe CBC News Poll Tracker is your guide to following the polls in the 2018 Ontario provincial election.
Powered by ScribbleLive Content Marketing Software Platform
302 Found


The document has moved here.