Alberta's Chaotic Middle

On Polls And Projections

There were four public polls released on Monday, three of which featured an NDP lead in Calgary (two small, one 10%). There was an internal NDP poll released showing a double digit NDP lead in Calgary. And we still have David Coletto’s Abacus poll from this weekend that points to an easy NDP win.

But Janet Brown had a poll released she conducted from May 1st to May 11th that has the UCP up 12 in Calgary, so this election has been tossed into chaos. Or, more accurately, it’s thrown me into despair, because of course the best pollster in Alberta shows a 12 point UCP lead in Calgary the day after I say Notley’s favoured to win now. Of course she did.

And now we get our favourite tradition, What To Do When The Best Pollster Does Outlier Things? And unfortunately, I don’t have a great answer.

…

So, we could just toss the Janet Brown poll and have a solid justification for it - the majority of the poll was done before Smith’s Hitler comments came out. The video came out the night of the 7th and got media coverage on the 8th, so of the 11 days of polling, 4 were post tape. The problem with this is that I don’t want Danielle Smith to be leading, and as much as I like to think I’d act the same way if this were inverted, I’m not sure I would be - and I know the people in my mentions dismissing it also desperately hope it isn’t true.

There’s six data points for this kind of situation that immediately spring to mind. There’s Ann Selzer, the matron saint of polling in Iowa, for whom Janet Brown is the best comp. Four of her polls fit the template - her 2008 Iowa caucus poll where she had Obama winning when he was expected to come third, there’s her 2014 Senate poll where she correctly predicted the GOP would win by close to 10, there’s Selzer’s 2020 Pres poll which I famously said was wrong and then was spot on, and then her second to last 2022 poll that said Chuck Grassley was in a 3 point race. Expand the horizons, and you have Leger’s final Quebec-only 2021 poll showing a Bloc surge that many others didn’t have (which ended up correct) and Janet Brown’s own 2019 poll, which had the race at a 20 point lead when everyone else said it would be UCP by 10.

What should we do with this poll? Wait is the actual answer - I don’t really consider this reporting, but I do happen to have it on good authority the CBC will be commissioning another Janet Brown poll in the final week of the campaign – but that’s no fun so let’s work through this campaign as if we had no polling. And if we did, nothing would point to Danielle Smith being the favourite in this campaign.

As a frequent critic of Rachel Notley and her campaigning style, her campaign so far has been immaculate – from the pledge to not raise income taxes to the pledge to cut small business taxes and the focus on reducing power bills, she is approaching a fairly left wing agenda of outcomes – government intervention and support for small businesses over multi-nationals – in a fairly centrist tone. Pitching the effort to reduce power prices as almost a tax cut, Notley is showing she knows the audience she needs to win over.

Take the Brown poll and the way the NDP reacted to it – immediately giving Don Braid an internal of their own showing a big NDP lead – and compare it to the way the UCP/the broader Albertan right handled David Coletto’s poll, and you see a campaign that’s confident on the left and one that’s scared on the right. I know David Parker – head of Take Back Alberta and confidant of Smith – isn’t technically an agent of the UCP, but he would know if the UCP polls showed good results and he wouldn’t have freaked out on Coletto if he thought everything was fine.

Does any of this mean the NDP will win? No, and if you read my column from Sunday morning you’ll notice that it wasn’t nearly as bashful or as arrogant as past me would have been, because this is a really close run thing and the UCP could easily win. After all of today’s polls the NDP still lead 44-43 in seats, because between the raft of polls and the fact that Brown’s poll is arguably out of date it’s hard to argue that we should bend over backwards for Brown’s poll, but it’s not a solid situation when the NDP win the tipping point seat by 0.056%. If you want to call it a tossup be my guest, but I don’t do that cause I’m not a coward.

Are the NDP still favoured? I think so, because outside of this Janet Brown poll everything points to the NDP being favoured. For the first time in a week the UCP has gotten good news and that matters a lot to the tone and tenor of conversations we have about this race. But what doesn’t change is that if this weren’t Alberta, and the pollster wasn’t Janet Brown, we’d be laughing that poll and the idea that data from pre-scandal was meaningful out of the room.

This election is incredibly close and anyone who tries to tell you it’s been won or lost is lying to you. I said I think Notley’s favoured and after the suite of polls today I think she still is, but the idea that Notley has won it for sure is for the birds. Notley’s in a slightly more favourable view than Smith, but it’s nothing that can’t reverse itself in 14 days, and if Brown is even close to right, it might not even take any reversing.

Would it shock me if tomorrow I have the NDP down? No, God no. Janet Brown is better at her job than 99% of people are at their own, and betting against her is a task undertaken at one’s own peril. But for now, I think Rachel Notley is still ever so slightly favoured to win in 2 weeks.