The Teflon Mayor? Reading Rob Ford's Poll Numbers
Torontoist has been acquired by Daily Hive Toronto - Your City. Now. Click here to learn more.




The Teflon Mayor? Reading Rob Ford’s Poll Numbers

A closer look at Rob Ford's seemingly scandal-proof approval ratings.

Rob Ford at the second of two council meetings during which he was stripped of most powers  Photo by Greg Stacey from the Torontoist Flickr Pool

Rob Ford at the second of two council meetings during which he was stripped of most powers. Photo by Greg Stacey from the Torontoist Flickr Pool.

A now-familiar sequence of events: a polling company releases the latest results about Rob Ford’s approval, quickly followed by people across the political spectrum typing furiously on Facebook and Twitter, usually something along the lines of “OMG, who the hell are these ppl who still support the crack mayor?” or “Why aren’t more ppl supporting our mayor, he’s the best ever.” Which, in turn, is quickly followed by eager dismissals—polls are useless, forget about those numbers—until we do it all again two weeks later.

Despite this cycle, there are legitimate questions that emerge from all these surveys. With the usual caveat that they are an imperfect tool for understanding the electorate, we crunched some numbers from Toronto mayoral polls over the past two years to try to answer those questions and set the stage for the 2014 municipal campaign.

Rob Ford’s Curious Approval Rating

Much is made about Rob Ford’s approval rating, and how it just doesn’t seem to move. After Bill Blair announced that the police were in possession of the crack video—the one whose existence Ford had denied for six months—the mayor’s approval went up five points. This type of movement gives the impression that Ford is somehow a Teflon politician, and baffles both keen and casual City Hall observers.

In a recent poll, Forum Research refers to Ford’s approval as “resilient.” A more accurate word might be sticky: no matter what he does, he’s unable to move the needle up or down beyond a narrow range of support. Here’s a look at the mayor’s approval ratings from 35 Forum polls over the past two years; we’ve noted a few polls in particular, which were conducted immediately after some key events:

ford approval trend line

There are widespread questions about Forum’s methodology; in their Toronto polling they tend to produce higher results for Ford than other firms, but they also provide a consistent question and the greatest number of data points, so we can extrapolate some trends from their results.

The first thing to note is that Ford’s approval over the past two years has been generally stable. His support averages 43 per cent and ranges from 37 to 49 per cent; more than two thirds of the results fall within 41-45 per cent approval. What’s key is that there is no clear upward or downward trajectory: Ford varies within that range, but he’s not trending up or down overall. Torontonians have made up their minds about the city’s most polarizing politician, and they’re probably not going to change with much new information.

The problem for Ford is that one politician’s “resilient” poll numbers are another politician’s rut. Canadian mayors tend to have much higher approval ratings than their provincial or federal peers, and low- to mid-40s range is not good at all. Results from across the country give us a rough idea of the kind of approval ratings we should expect from an incumbent mayor, which is at least north of 50, and probably somewhere in the mid-50s to mid-60s range.

As for the local context, here’s how other potential candidates compare to Ford:

candidate approval nov 20

In this context Ford’s numbers don’t look too hot: he trails Olivia Chow by 17 points and John Tory by 10. It’s not as bad as you’d expect given everything that’s happened, but still not good. Worse still for Ford is a November 18 Ipsos poll that had Ford’s approval at 40 per cent. In that survey, voters approved of council as a whole at 57 per cent, and of their local councillor in particular at 74 per cent—again, it’s in light of the comparison that we get a sense of how Ford is really faring.

The Matchup Game

While the mayor’s approval numbers can tell us a bit, there are better ways to gauge a candidate’s chance of success in an election. As Ipsos pollster John Wright recently cautioned, this metric tends to indicate the upper band of a candidate’s support; actual voting numbers are almost always lower, and often substantially so.

There’s every reason to think this will hold true for Ford. Despite near constant campaigning over the past year and a half, the mayor struggles to win when people are surveyed about plausible mayoral matchups. Over the past two years, Forum has conducted 39 polls asking about Ford vs. Chow, with and without variations of other candidates included. Ford has led Chow a grand total of two times, both by one per cent; he tied her in a third. In the unlikely event Chow does not run, Tory has led Ford in 11 of 16 polls that include him but do not include Chow. In polls where Chow and Tory are not included but Ford and Karen Stintz (Ward 16, Eglinton-Lawrence) are, Stintz wins 10 of 12 times. Over the past two years of polling, one thing has been made clear: there are very few scenarios in which Rob Ford beats a serious candidate.

A big part of this problem is captured in a December 13 Ipsos poll question. Asked whether voters “would consider voting for Mayor Rob Ford in the next municipal election,” 61 per cent said no, and 39 per cent said yes. This might seem like another example of Ford’s famously enduring support—39 per cent is a good starting point for a campaign—but that’s looking at the glass half full. The question does not ask “will you,” but “would you consider.” In other words, only 39 per cent can see some scenario where they would vote for Rob Ford in 2014, while 61 per cent cannot. That 61 per cent is particularly firm: 48 per cent of the total number surveyed didn’t just disagree but strongly disagreed when asked whether they’d consider Ford. These numbers are absolutely awful for an incumbent.

In order to win in 2014, Ford will need to win over some people who right now say they would not consider voting for him, or win an extremely high percentage of those who say they would consider it. This appears unlikely: that latter group will also be able to look to candidates like Tory, Stintz, and David Soknacki who offer similar policies but none of the drama. Even if Ford did win every single one of these voters, it still might not be enough: no Toronto mayor has won with less than 40 per cent of the vote in recent history.

The mayor’s numbers are sticky—he has a high floor of support, and he’s not dropping below that baseline—but they are sticky at the upper end of the range, too. It will be very hard for him to break through that ceiling.

The Name Recognition Problem

At this point, with election day nearly 10 months away, polls aren’t treating all candidates equally. There’s a significant advantage that comes with high name recognition, as prospective voters are likely to prefer candidates they know—and that means candidates that already have a high profile will fare better in polls than those who don’t. Over time, the name recognition problem starts to correct itself as lesser-known candidates have a chance to introduce themselves to voters. Right now, support for candidates like former Scarborough councillor David Soknacki (who isn’t well known) tends to be understated, while world-famous mayor Rob Ford enjoys near universal name recognition.

There is a method for adjusting the polling results we have, to compensate for this. (The technical details: Bayesian math allows us to create a snapshot of what per cent of voters who have heard of each candidate so far make them their first choice; this gives a sense of how well other people might respond to each candidate in the future. What’s represented in the charts below is the percentage of people who have heard of the individual candidate who would vote for him or her. If all candidates were equally known, this would add up to 100 per cent. However, since name recognition is unequal, and that’s what we’re correcting for, it makes the totals in the below charts are greater than 100.) On November 14 Ipsos released a poll in which they asked people who they would vote for if the election were held tomorrow, and also measured how familiar people were with each candidate. If we run the numbers, we wind up with this:

recognition adjusted 5 candidate

There aren’t any fundamental shifts here—Chow still leads and Tory is still a strong second—but we see the race is more competitive, and split into two tiers. Rob Ford is in the second tier: because he is already close to 100 per cent name recognition, his adjusted support level doesn’t reflect any big changes. By contrast Soknacki, the least known candidate, experiences the biggest change in support.

The John Tory Question

John Tory, who is hemming and hawing in that special John Tory way, is currently the biggest unknown in this race: his presence or absence has the largest impact on results, to the extent that we can predict them now. If he were to decide to sit things out, here are the adjusted results for the other main contenders:

recognition adjusted 4 candidate

Soknacki once again show the biggest jump—one that vaults him into the top tier of candidates. Of the people who are already familiar with him, he represents the second choice for a lot of John Tory voters: enough that 44 per cent of people who have heard of him would rank him first in a race against Chow, Stintz, and Ford.

Side by side, here are the name-recognition-adjusted levels of support:

recognition adjusted chart

If Soknacki is able to raise his own profile over time, he stands to gain tremendously. This pattern has been true of recent successful mayoral candidates: think David Miller in 2003, Rob Ford and Calgary’s Naheed Nenshi in 2010, or Edmonton’s Don Iveson in 2013. Each of these candidates had limited name recognition early on, but very good poll results among those who knew them. Soknacki’s problem is not that he isn’t a competent or experienced candidate, or that people don’t like him. His problem is that most people don’t know him in the first place. That’s something that can be fixed with a smart campaign. On the other hand, when polling results show an incumbent with near universal name recognition polling in the 20s and 30s, that is very difficult to fix.