Frances Bula header image 2

NPA gives out internal poll numbers week before election: What do they mean?

November 11th, 2011 · 49 Comments

Poll stories are tricky, especially when those polls are coming from political parties rather than pollsters operating completely independently.

So what to think about the NPA poll splashed across the front page of the Sun today that says the NPA’s Suzanne Anton is closing the gap between her and Vision’s Gregor Robertson?

Here are my thoughts on what it means.

1. I have no doubt that the Occupy Vancouver situation is having an impact on voting choices for some city residents and that Suzanne Anton’s numbers have improved at various points since the camp started. There’s been heavy media coverage of the Occupy camp and a lot of negative public reaction to it. As well, the poll doesn’t seem that far off what I’ve been hearing all along about how many NPA council candidates might get elected. It says “on track for four,” — a little higher than the two or three I’ve been hearing, but not completely out of whack.

2. I have to wonder about the order of the poll’s questions.

Polling experts have always told me that a pollster should never ask any question before the one about voter intention, because it can have the effect of skewing the vote as you “prime” the pollee to think about a certain topic before answering. I understand the first question this poll asked was something like, “Do you think life in Vancouver is better or worse?” That’s a priming question that could have had an impact.

It also looks as though there was a second question before “who would you vote for,” with questions about whether the Vision administration has done a good or poor job and should be replaced. The first two questions in that set are the negative option (it should be replaced) and the second two are the positive. Again, seems like priming. CORRECTION: Jeff Lee posted the order of the poll questions after I wrote this. This second question was NOT asked before “who would you vote for.”

3. The polling that was released was done Nov. 3-5. Those were the three days with the most negative news about Occupy Vancouver. On Nov. 3, there was a non-fatal drug overdose. On Nov. 5, a young woman was found dead. That had to have affected people’s responses about OV and about the mayoral candidates. (I know from covering other campaigns that polling is often done on a daily basis so that parties can see the trends, not just a snapshot. If the NPA is doing rolling polls, it could have chosen the three worst days in order to highlight the most positive numbers for the party.)

4. Hamish Marshall, the pollster who did this, says the results are weighted to the last provincial election. I’d be curious to know what that means. BTW, Hamish used to be at Angus Reid and is a very experienced pollster, one whom I’ve relied on in the past.

5. The margin of error for the city sectors is very high, especially for the downtown region, where it’s 9 per cent. That’s an area with a lot of renters, a lot of disaffected Vision voters who may go for Neighbourhoods for a Sustainable Vancouver or stay home.

There are other things I could pick away at but I’ll leave it at that for now, except to say that I’ve done my own share of news stories about internal polls from parties that I’ve been criticized for. My hands are not clean on this issue and I’m not attacking my buddy, Jeff Lee, at the Sun. Just raising some thinking points.

In fact, I was trying to get this poll as well for weeks, but there was some kind of miscommunication between me and Norm Stowe at the NPA and Jeff got it instead of me. (Here it is, by the way, so you can do your own analysis.)

I did notice, though, that onetime Conservative strategist and commentator Norm Spector was critical of the Sun for running this internal poll all over its front page, a week before the election.

 

 

Categories: Uncategorized