Nik Nanos is The Globe and Mail's pollster and executive chairman of Nanos Research
Empirically, it can be sheer folly to predict the outcome of party leadership races. Why? Because they are one part tribal politics, one part political ambition, and one part bare-knuckle organizational fight. They are politically and emotionally charged crucibles that often yield surprise outcomes. Throw in a preferential ballot and you get a data curveball that is very hard to directionally predict. Those reading data entrails with predictions should do so with extreme caution. This is especially true for the current federal Conservative Party race.
The only empirical and publicly "known" piece of data in the race is that Maxime Bernier has raised the most money, meaning he has wherewithal to deploy, to identify, to persuade and to deliver more votes than the other contenders. Having the most money means he should, but does not necessarily have, the best chance to win.
The other "known" is that there are now almost 260,000 members in the Conservative Party of Canada. However, who is ahead, in second or third should be considered complete conjecture. Individual campaigns know how many memberships they have sold but are in a data black hole when it comes to guessing the true composition of the race, because only the central party could know how many memberships have been submitted by each campaign.
Read more: Who's running for the Conservative leadership?
Campbell Clark: Conservative leadership race divides along new poles
Why doubt the claims of front-runner or challenger status by any candidate?
First, only a survey of actual Conservative Party members on the list currently could be reliable. Proxies such as donor lists or old membership lists are poor in their coverage of the target population to be surveyed and fundamentally flawed because they do not survey the true target population: Conservative Party members today.
Second, preferential ballots can have wickedly wild outcomes. You start with the first choice, and as candidates drop off, second choices kick in. In a field of 13 candidates, the computation of the cascading choices as candidates drop out can be very fluid. It is not always the candidate who tops the first choice that always ends up winning; they do win, but not always.
Leadership conventions where candidates drop out, forcing party members to turn to their second choices, can be very unpredictable. In 1996, Dalton McGuinty started off as fourth on the first ballot to be leader of the Liberal Party of Ontario and he won on the fifth ballot. In 2006, Stéphane Dion trailed both Michael Ignatieff and Bob Rae on the first ballot and won on the fourth ballot. Party members can be firm on their first choice but mushy on their second until they actually cast a ballot.
Third, who knows how many signed-up members will actually vote. Voter turnout and the ability of individual campaigns to deliver signed-up members will make or break the leadership aspirations of contenders. It will also be interesting to see how many of the 260,000 odd Conservative members decide to vote. The voter turnout will be a true indicator of the level of party commitment. A high party-member turnout will suggest an engaged, reinvigorated party ready to take on the Trudeau Liberals. A low party-member turnout will suggest party members are not enthusiastic about any of the choices. In that sense, the vote will not only select a leader but will be a vote of confidence on the party itself.
Nanos has done research about what Canadians think about the leadership race. That gives us an understanding of who might have traction among Canadians but does not inform us as to who might win the leadership.
The best advice? Without independent and reliable research of party members and factoring the fluidity of second-ballot preferences, forget about trying to predict who will win the Conservative leadership race.