|
‘BRACE FOR IMPACT’ — POLITICO contributor PHILIPPE J. FOURNIER predicts Monday night could be a long one.
He’ll publish his final 338Canada projection Sunday. Before then, he has insights to share with Playbook readers: The likely scenario: DANIELLE SMITH and the United Conservative Party should win a short majority government. But it’s not a shoo-in: The NDP has held its own, and appears to have significantly grown its support in Battlefield Calgary while holding its Edmonton fortress, according to every pollster in the field. The 338Canada model counts 14 toss-up districts. To win, the NDP has to run the table. It can be done, but every domino has to fall just right. Trends and outliers: Analysis benefits from hindsight. Case in point: When Abacus Data released a poll over Mother’s Day weekend showing a 10-point province-wide lead for the NDP, readers wondered if momentum had shifted in favor of RACHEL NOTLEY’s team after months of deadlock. The 10-point NDP lead appears to have been an outlier — the infamous 20th poll out of 20. Outliers are statistically inevitable. Such is the nature of statistics and the art/science of aiming at a constantly moving target. I’d worry more about polling firms that hide outliers than those that lay them out for the world and then go back to the field to correct course. In a political ecosystem in which the main parties are perhaps 2-4 points apart, the odd outlier showing a clear lead for either party is bound to happen. Looking for signs: Early turnout at advance polls is above the 2019 high mark — not surprising because it’s a close race. While high overall turnout often signals a desire for change, it can also represent people who just want to avoid long lines on voting day. Quebec broke its advance poll turnout records, for example, but overall turnout was 66 percent — historically low for Quebec standards, but close to the previous general elections. Elephant in the room: During the 2019 Alberta election, none of the campaign’s final polls showed the UCP as high as it ended on election night, just below 55 percent of the popular vote. A rule of thumb in polling is that a party is never as strong as its very best polls and never as low as its worst ones. In 2019, the UCP beat all the polls. Past errors are not necessarily precursors of future ones. Since the spring of 2019, we have had two federal elections, along with elections in populous provinces such as Quebec, Ontario and British Columbia, where Canadian pollsters have fared, on average, more than well. But a pollster recently told me, “Alberta is such a bitch to poll,” citing rapidly changing demographics and several hard-to-reach, but critically important, voters in the province’s vast rural areas. We’ll find out soon if pollsters have tamed the beast. This just in: Mainstreet Research President and CEO Quito Maggi surprised more than a few readers Tuesday when he predicted an NDP victory. The update from Mainstreet rolling polls gave the UCP a province-wide lead over its NDP rival (49 to 45 percent). Maggi cited his firm’s Calgary numbers and riding polls that heavily leaned NDP. He predicts Notley’s party will make significant gains in outer Edmonton ridings, and that the margins in Calgary should be enough to push it past the 44-seat mark. No matter how close — a 44-43 final tally is not out of the question — the post-election dynamics could out-drama what we’ve witnessed so far. What comes next: If it’s a razor-thin majority, individual MLAs would suddenly yield far more power over their party, forcing the winning leader to try to please multiple factions. It would arguably be far harder for Smith to whip a 44- or 45-seat caucus to accept some of her more controversial policies. In such a scenario, Smith would hold the premiership by a thread and prominent MLAs would start polishing their credentials for a leadership race sooner rather than later. | 
