Beginning of the End Games: Paths to Win in the NDP Leadership Race
[Welcome, National Newswatch readers!]
Just three NDP leadership candidates have a path to win, according to new poll results released by the campaign manager of one of the three, but naturally no-one else agrees with him on which three they are.
In an email sent to that campaign's inner circle Wednesday afternoon (see below), Dan Mackenzie outlined further details of the large-sample IVR poll commissioned on behalf of his candidate Paul Dewar, showing how the second-ballot preferences of other candidates' supporters broke down (opens PDF), and arguing that they show only Dewar, Nash or Mulcair as able to win.
The poll, as we reported earlier, is several weeks old now, and predated the withdrawal of Romeo Saganash, the Québec City debate, Dewar's Québec caucus endorsements, and the membership cutoff, but it is still the only evidence that has emerged to date as to the direction of other candidates' down-ballot support.
According to the N=6,373, Feb 8-9, 1.9 MoE IVR poll conducted by , the second-choices of first ballot supporters of all the then-contestants were as follows:
| First Choice |
Second Choice Candidate | Und | |||||||
|---|---|---|---|---|---|---|---|---|---|
| NA | NC | PD | TM | PN | RS | MS | BT | ||
| Wtd Total |
10.7% | 14.4% | 21.2% | 16.7% | 19.4% | 3.6% | 1.8% | 12.4% | |
| N Ashton | 14.9% | 25.0% | 19.7% | 18.0% | 4.4% | 0.9% | 4.8% | 12.3% | |
| N Cullen | 13.4% | 24.9% | 18.2% | 18.6% | 2.4% | – | 8.3% | 14.2% | |
| P Dewar | 9.0% | 12.5% | 22.0% | 27.7% | 2.9% | 1.1% | 8.8% | 16.0% | |
| T Mulcair | 7.8% | 15.3% | 21.1% | 26.0% | 4.2% | 0.5% | 16.4% | 8.7% | |
| P Nash | 11.2% | 11.1% | 25.5% | 19.4% | 3.4% | 1.4% | 13.4% | 14.5% | |
| R Sagan. | 16.5% | 10.3% | 14.4% | 23.7% | 15.5% | 1.0% | 1.0% | 17.5% | |
| M Singh | 21.1% | 15.8% | 15.8% | 10.5% | 10.5% | 10.5% | 2.6% | 13.2% | |
| B Topp | 7.8% | 5.0% | 20.8% | 23.0% | 30.4% | 2.2% | 0.9% | 9.9% | |
With ballots and voting instructions set to arrive in the mail of NDP members any day now, of course, all the campaigns are turning their minds to what they need to do to get over the finish line in first place.
But let's take a look at what it would take for each of the top five candidates to win, using those numbers as a starting point.
First of all, we need to keep in mind that while Romeo Saganash has suspended his leadership campaign (he continues to raise money for it, however), I understand that as he did not withdraw from the race before the deadline of January 24, his name will therefore remain on the printed version of the first ballot. How they handle the drop-offs after the first ballot I'm unsure, but let's work from the known at least, inasmuch as the bottom person and anyone having less than 1% of the vote must drop off, along with anyone else who voluntarily withdraws after that ballot.
Using the first and second choice data from the Feb 8-9 Dewar poll, as weighted using the February 2 membership provincial totals, here's a rough approximation of how things might unfold.
| Decided Vote% |
Leadership Contestants | Und. | |||||||
|---|---|---|---|---|---|---|---|---|---|
| TM | PN | PD | NC | BT | NA | MS | RS | ||
| 1st BALL | 25.5% | 16.8% | 15.1% | 12.8% | 12.7% | 9.5% | 4.1% | 3.6% | [31%] |
| fr Sagan. | 0.9% | 0.6% | 0.5% | 0.4% | – | 0.6% | 0.6% | ||
| fr Singh | 0.4% | 0.4% | 0.6% | 0.6% | 0.1% | 0.9% | 0.5% | ||
| 2nd BALL | 26.8% | 17.8% | 16.3% | 13.8% | 12.8% | 11.0% | 1.2% | ||
| fr Ashton | 2.2% | 2.0% | 2.7% | 1.6% | 0.5% | 1.3% | |||
| 3rd BALL | 28.9% | 19.8% | 19.0% | 15.5% | 13.4% | 2.5% | |||
| fr Topp | 3.1% | 4.1% | 2.8% | 0.7% | 1.3% | ||||
| 4th BALL | 32.0% | 23.8% | 21.8% | 16.1% | 3.8% | ||||
| fr Cullen | 2.9% | 3.0% | 4.0% | 2.3% | |||||
| 5th BALL | 35.0% | 26.8% | 25.8% | 6.1% | |||||
| fr Dewar | 5.7% | 7.1% | 4.1% | ||||||
| 6th BALL | 40.6% | 34.0% | 10.3% | ||||||
As you can see, any number of variables since the survey was conducted could change the very tight ranking between its 2nd-5th place finishers.
But according to these numbers, Dewar makes his first big move in this scenario as the leading second choice of Niki Ashton supporters. On the next round, however, he loses ground to Peggy Nash, as the result of her being the leading second choice of Brian Topp supporters. But then Dewar moves up to nearly match Nash after Nathan Cullen drops off, given that he is somewhat more favoured as the second choice of Cullen supporters.
[Click on image to open full-sized version]
Suppose, on the other hand, that Topp was placed ahead of Cullen. Now the second-choices of Cullen's supporters would kick in one ballot earlier, putting Dewar into second place ahead of Nash — at least according to these numbers.
On the other hand, if Dewar could push past Nash, he might be able to benefit from his above-average second choice support from Ashton, Cullen, and Nash herself, to offer a stronger challenge to Mulcair.
[Click on image to open full-sized version]
Finally, if Dewar were in fifth place (say with 12.5% on the first ballot, rather than 15.1%), his supporters' second choice votes would give slightly more to Peggy Nash than to Thomas Mulcair, but not enough to close the gap between the two front-runners in that case, and Mulcair goes on to win.
Careful readers will wonder why, in the last ballot for example, not all of Dewar's hypothetical 25.8% of the vote is distributed to the other two candidates. This is because those voters' stated second choices have already fallen off the ballot. In real life, their third or fourth choice would then be counted, as this system always counts the highest remaining preferential vote still on the voter's ballot,.
While these exact results would also be subject to a number of additional factors, such as:
- new membership sign-ups
- changes in support of the various candidates
- a shrinking undecided rate
- third and subsequent choices down-ballot
- early withdrawals by one candidate or another
- voting on Convention day itself
- etc., etc.
… they still give the reader an idea of the kinds of challenges involved in winning when needing to break out of a pack of four fairly equally weighted challengers to a front-runner. Subject to these considerations, I think we can say that the chance of any another candidate beating Thomas Mulcair at this stage will depend on their placing even slightly ahead of their competitors, and hoping the rest of the candidates drop off in the most opportune order.
For Peggy Nash, she is the leading second choice of Brian Topp's and Paul Dewar's supporters, but also Thomas Mulcair's in this poll. So long as Mulcair places ahead of her, she will never get to claim that tranche of support, and thus her overall second-choice tally is misleading. She could win if Mulcair were behind her, but has a much harder path to victory if he's ahead.
[Click on image to open full-sized version]
Brian Topp shows little second-choice support in this poll, while Nathan Cullen's first- and second-choice support in it is no doubt underestimated by now.
In any event, several other campaigns have been in the field since the membership cut-off, including a two-question IVR poll from the Nash camp a week and a half ago, an online poll that looked a lot like the Topp poll from last November, and another 5-question IVR poll the other day, asking for membership status, gender and the respondent's first, second and third choices. If any of them show the least break-out for any of the challengers, we can be certain we'll be seeing their results in some form or another in coming days.
——————————-
From: danxxxxxxxxx@xxxxxxxxxx.xxx
Date: Tue, 28 Feb 2012 14:40:46 -0500
Subject: Insider Update – Paul Dewar polling
To: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxDear Dewar Campaign Activist,
With just under 4 weeks until the leadership convention, I want to share some internal campaign information with you about Paul Dewar’s path to victory.
As you know, a few weeks ago we conducted a large IVR poll of NDP members to ask who they are supporting for NDP leader on the first and second ballot. We contacted 56,522 NDP member households and received 6,373 responses from every region of Canada. The poll had a margin of error of 1.19%, 19 times out of twenty. At the time we publicly released the raw and weighted (to reflect the membership figures) results of 1stand 2nd ballot support. Those results are at the bottom of this email.
As a key Paul Dewar supporter, I’d like to share some more key information from that poll showing that Paul is well-positioned to win. Attached here is a PDF that shows where the second choices of each candidate get distributed should they fall off the ballot. This is information we haven't yet shared publicly as it has a great impact on our campaign’s strategy. Please keep in mind that this data was collected before Romeo Saganash withdrew from the race, before leading members of Romeo's campaign came to Paul, and before Paul picked up some impressive Quebec MP endorsements.
What the numbers show is that three candidates are in a position to win: Thomas Mulcair, Peggy Nash and Paul Dewar. While the other candidates in this race have run strong campaigns, none of them have the combination of first and second ballot support needed to win.
While Mulcair is running first he is far from a first ballot victory. Our numbers show that as candidates lower in the polling drop off the ballot, and their second choices are distributed, the gap closes significantly. Based on the distribution of 2nd ballot choices, if the election were held today it would be a tight finish with either Paul, Peggy Nash, or Thomas Mulcair winning.
Here’s where you come in. With less than four weeks to go, our campaign is gaining momentum. We are in a strong position to win this race. But we need your help in this home stretch, to seal the deal.
Over the next week, when the ballots are mailed to members, it will be crucial for our campaign to identify as many voters supporting Paul on the 1st, 2nd and 3rd ballot.
We need to continue to contrast Paul’s experience, passion, popular appeal, and social democratic values with the other two leading candidates.
Now I know we ask a lot of you, but now is the time to re-double your efforts. In the phone banks, on the candidate’s tour, at events, by donating, tweeting, and through our canvassing – let’s give it one last push and win this.
Keep up the great work!
Dan Mackenzie
Campaign Manager
Paul Dewar CampaignRAW NUMBERS BEFORE WEIGHTING
Count %
Niki Ashton 466 7.3%
Nathan Cullen 470 7.4%
Paul Dewar 726 11.4%
Thomas Mulcair 1098 17.2%
Peggy Nash 807 12.7%
Romeo Saganash 171 2.7%
Martin Singh 154 2.4%
Brian Topp 508 8.0%
Undecided 1973 31.0%
Total 6373 100%
FIRST CHOICE (weighted to NDP membership numbers by province)
Thomas Mulcair 25.5%
Peggy Nash 16.8%
Paul Dewar 15.1%
Nathan Cullen 12.8%
Brian Topp 12.7%
Niki Ashton 9.5%
Martin Singh 4.1%
Romeo Saganash 3.6%
SECOND CHOICE (weighted to NDP membership numbers by province)
Paul Dewar 21.2%
Peggy Nash 19.4%
Thomas Mulcair 16.7%
Nathan Cullen 14.4%
Brian Topp 12.4%
Niki Ashton 10.7%
Romeo Saganash 3.6%
Martin Singh 1.8%
——————————-
For the latest on the NDP Leadership Race, don't forget to follow the half-hourly news updates, and social media tickers at the Pundits' Guide NDPLdr portal page: http://ndpldr.punditsguide.ca
I don’t want to start a fight, but this poll really isn’t worth your excellent analysis.
Then send me your campaign’s poll results, Kevin.
I must admit that it does suprise me that the only poll released with full details (sample size, scripts, raw and weighted data, dates of calls, margin of error, details on the firm) is the one other camps critisize.
My response is simply, if you don’t believe it (you should), release your own full polling information.
Alice, you are Canadian politics answer to Bill James! The list of methodological problems with this poll would keep a social science methods seminar busy for hours. In answer to your question, the reasons why a campaign releases a poll are different from a news organization. A campaign releases a poll when it is tactically advantageous to do so. On reading the convuluted logic of the Dewar campaign it seems obvious that they fear that their supporters are losing heart (who wouldn’t after Dewar’s Quebec debate debacle) and the poll is an attempt to reassure them that Dewar can somehow still eke out a win. As to the other campaigns, perhaps they think releasing a poll that favours their campaign, after NDP members have their ballots, makes more sense. I think Paul Dewar may turn out to be the Flora MacDonald of this leadership race.
Although I understand in lieu of worthwhile polls why pundits who love to play with numbers would analyze this one, I have to admit that after orchestrating a lot of IVRing, and corresponding phone follow ups, I laugh at the play pundits are affording this poll as something even remotely accurate. You don’t even know who picks up the phone (could be a child, visiting guest, neighbour, or even a clever dog or cat for all you know), and thus you don’t know if the survey respondent is the actual voter or not. If it had been conducted on just cell phones it might be a little more accurate but it wasn’t. In addition, what you are analyzing is a sample of people who voluntarily responded to the robo poll, which is not an accurate sample of voters since there are the many voters who hate robo-calling and would never respond to one. And there are so many ways to game the results, I mean you could simply give a heads up to some your supporters to respond and taint the results of even the already unusable voluntary robo poll respondent data set. These are just some of the many reasons why you normally never see that type of polling getting any media play in a general election, because it shouldn’t.
Peter, Jesse: If either Brian Topp’s or Nathan Cullen’s campaigns have better quality data to release, they both know where to send it: alice@punditsguide.ca.
Interesting work Alice. Hopefully you can do the same on future polls.
For what it is worth, Mulcair’s poll gives a completely different picture of second preferences with Dewar fourth instead of first: Nash 25.4, Mulcair 21, Topp 19.7, Dewar 13.7.
Its a useful thing to say Alice, but I dont think its likely anyone will take the bait.
Yes, and kind of obvious, campaign only release data if there is strategic advantage.
But there are umpteen reasons most campaigns do not go there, including not in response to the competition doing it and leaving you with the optics of looking weaker.
The methodology of this poll gives me reasonable confidence in the ‘big picture’ of first choices- Mulcair well ahead, then 4 more.
But I really cannot see much reliability in the down ballot numbers.
Ropshin, if Thomas Mulcair’s campaign is able to give me breakdowns of second choice by first choice, with decent enough sub-sample sizes to be valid, I would run the same analysis with that data. His poll had an N of 1,100 or so, which would make that harder. However, if there is any new data, please send it over.
If an IVR poll shows 31% undecided, this is certainly a low estimate. Likely the majority of undecided respondents hang up rather than keep answering “undecided” to the list of questions. A live poll that shows 30% undecided will also underestimate the undecideds, since some respondents will be nice to the caller — it’s the loving and optimistic thing to do. The real undecideds must have been close to 50%. Analyzing such polls is fascinating but, unfortunately, subject to the old GIGO rule — Garbage in, garbage out.