Four days before Monday’s election, which saw a Progressive Conservative majority government elected in Alberta, the CBC News election team held its projection meeting.
We do this just before every election. About 20 people, including journalists, analysts and political experts, gather to discuss every single riding, choose the party we think is most likely to finish first, and assign a potential margin of victory.
The purpose of the meeting is to set a baseline for our computer system. That helps our decision desk editors analyze the vote count, make projections about who will be elected in each riding and make an accurate call as to who will form the next government. It also helps shape our election rehearsal strategy to plan coverage based on the most likely outcome.
At the meeting, we talked about riding history, the effect of redistribution, the insight of senior campaign reporters and what we’d heard from party insiders.
In short, the Wildrose Party thought they had it won. PC contacts told us on background, however, to be cautious because something was happening.
But even the Tories were not confidently predicting a big majority win.
We talked – a lot – about the polls.
By the end of the meeting, I was convinced that this was not going to be a slam-dunk Wildrose win. In fact, our CBC projection meeting ended up deciding on a PC minority government as the most likely outcome. That was four days before the Monday vote.
'Hail Mary' outcome
In more than a decade of doing elections at CBC, I know that no one has ever won the office pool using the results of the projection meeting. It often favours the incumbent party and is usually cautious when it comes to suggesting a change in government.
Most of us thought a Wildrose Party win was still the most likely outcome in a race we now saw as too close to call.
At this meeting, pollster Bruce Cameron from Return On Insight, who was working as our on-air analyst on the election night broadcast, told us the Wildrose party was losing momentum. He told us the race was tightening and any outcome was possible. Some of us, including me, questioned the possibility of "any outcome," but Cameron was certain he saw a wave back to the PCs.
Cameron offered that the PCs could win a majority, but it would be like the team that throws the "Hail Mary" pass on the final play to score a touchdown and win the game.
So, based on that meeting, we rehearsed every outcome, including a Sunday afternoon rehearsal of a PC majority government with 47 seats. Some of those who watched that rehearsal suggested we had wasted everyone’s time practising that result.
As it turns out, we were way too cautious once again. On Monday night, the PCs won a 61-seat majority government.
Cameron got to use his Hail Mary pass analogy on the air.
Polls are a snapshot in time. There is no doubt that the Wildrose party was headed for an easy, dynasty-ending, majority government victory one week before voting day.
Polls up to the final week showed them with a solid 42 per cent or more in the popular vote. That’s a majority.
Then, in the final week, the polls showed the race tightening. Still, a win seemed likely – but a minority was now in the mix.
It's true that no poll showed a big PC majority, but there was a clear trend in the final week indicating Wildrose support was slipping.
Polls don't predict results. The best they can do is suggest what the result would be if the election were held at that moment.
The old cliché that the only poll that counts is on election day was never more true than on Monday, April 23.
A lot changed on the single day.
The 'retriever' campaign
Campaigns matter. Strategy and tactics in war rooms and on the buses win or lose elections.
It would be an understatement to say the PC team got off to a bad start in this campaign. Despite the fact that they controlled the timing and had all the advantages of the governing party, they stumbled out of the gate. The "do nothing" committee controversy was, perhaps, the most damaging of several stumbles.
A PC strategist put it to us this way: if a campaign can be compared to a game of fetch, the PC team had just taken the stick and thrown it away. Now, they had to retrieve it.
To their credit, they did.
While the candidates fought hard in the ridings, the war room team led by people such as Stephen Carter and Susan Elliot attacked Wildrose on television with ads questioning Danielle Smith's leadership and her party's values and choices. They took advantage of controversial statements by Wildrose candidates to sow doubt in voters' minds.
By the final week, the dog had retrieved the stick and the only question was whether he could race home with the prize before time ran out.
'Flora syndrome' strikes again
In 1976, MP Flora MacDonald was a candidate for the leadership of the federal Progressive Conservative party.
More than 300 delegates pledged to support her on the first ballot. Pollsters and pundits ranked her as a top contender.
On the first ballot, she got just 214 votes. She dropped out after the second ballot.
People who said they would vote for MacDonald picked someone else when they got to the ballot box.
This happened Monday night in Alberta. People told pollsters they were voting Wildrose. People donated money to the Wildrose campaign. People put lawn signs in their front yards.
But when they got to the voting booth, they changed their minds.
About one in five poll respondents said they were undecided.
Usually, the undecided vote splits along the lines of the decided vote.
There is no way to measure this, but I think it’s clear that the majority of the undecided voters picked the PC party.
The questions the PC campaign raised about the Wildrose candidates and policies helped those voters make that decision.
It appears that people who might have voted Liberal or New Democrat decided to vote PC for the first time in their lives in order to stop the Wildrose party.
The Liberal vote dropped about 11 per cent. It's a fair guess that it all went to the PCs, especially in Calgary. The Liberals did well to hang on to five seats.
Voters always right
The answer to the "what happened?" question is not that the pollsters and the media got it wrong.
Many voters changed their minds at the last minute.
People who had not made a choice picked PC candidates on election day and many people made the toughest choice of all, to abandon their preferred choice and vote strategically to stop a Wildrose majority win that appeared all but certain.
If there was a mistake, it was to forget that the voters make the final decision.