The Saskatchewan Party membership will soon elect its new leader, and the province's next premier. But predicting which of the five-candidates aiming to replace outgoing Premier Brad Wall will come out on top on Jan. 27 isn't easy, as suggested by recent polls with differing findings.
Saskatchewan Party leadership candidates have been sharing various types of polls showing how they're doing in the race on social media. But the polls have all been conducted in different ways and suggest different outcomes.
Experts say those polls should be taken with a grain of salt.
"The mark of a good poll is how transparent it is," said Jim Farney, an associate professor in political science at the University of Regina.
'The mark of a good poll is how transparent it is.'- Political science professor Jim Farney
"There's really been no solid polling done for the leadership race that I could go back and say, 'Yes, that was well done. And we know who is in the lead as a result,'" said Farney.
Political analysts say a good poll will include basic information including how the respondents were contacted, how many respondents there were, when the survey happened, a margin of error and who, generally, was interviewed in the survey.
"There is no independent third-party poll that was not commissioned by one of the candidates," said David Valentin, a political analyst and former executive vice-president of Mainstreet Research.
"Because that hasn't been produced, essentially there's nothing to stop all of the campaigns from releasing their own poll numbers."
Some of the polls shared by candidates have been generated on social media or news websites.
"Anyone can vote in those surveys. They're just an online poll you click and then you're set. As a result, it's very hard to determine if those individuals are actually members of the Saskatchewan Party," said Valentin.
"We don't actually know or have a reasonable understanding that those people are actually eligible to vote in the leadership race."
Only leadership candidate Alanna Koch has released polling results that her campaign commissioned from a third-party polling firm, NRG Research Group, in October and December.
Valentin said that one advantage for a candidate commissioning a poll is that they have access to membership records to find eligible voters to respond to the survey.
In October, a campaign release about the poll indicated "Alanna is positioned in a three-way cluster at the top with Ken Cheveldayoff and Scott Moe within the statistical margin of error," with a large number of undecided voters.
The campaign said that survey sampled 500 Sask. Party members with demographic and geographic distribution across the membership. It said the results of the study have a margin of error of 4.38 per cent.
In December, her campaign said she had made significant gains since October and "there are only two candidates who are in contention."
"In October she didn't release the percentages, but she released a lot of useful information. In December, she didn't release the useful information, but she did release the percentages," said Valentin.
Valentin said for the poll to be transparent, both the results and useful information about how the poll was conducted should be released.
"We have to remember who commissioned the survey," said Valentin.
"There's a value to campaigns to spread that information to show how well they're doing, to try to help build a narrative that they have momentum, that they're a serious candidate. The only problem is you really can't rely on that information so much."
Farney said that for a race where some of the candidates may be close to one another in terms of policy, sharing polls may be a way to influence voters' perceptions.
"For one of them to be able to say, 'Hey, look. I'm ahead, I'm the winner' … that's a strategic gain for them, if there's less to differentiate you on the policy side."
Farney said that even though there has been little third-party polling in this election race and it's not clear who's in the lead, it's not all that bad.
"There's a growing literature that says what we call 'horse-race reporting' is actually harmful because it means you and I have a lot of conversations about who's winning rather than who stands for what policy, or what people's backgrounds are," said Farney.
"The substance of it gets obscured if we talk too much about polling."
CBC contacted each campaign to ask if they had commissioned any sort of polling. Koch's campaign said it had not done more polling since December. Scott Moe and Gord Wyant's campaigns did not respond in time for publication.
Ken Cheveldayoff's and Tina Beaudry-Mellor's campaigns said they had not commissioned any formal polls.
"Our polling has taken place on doorsteps and talking to grassroots members," said Cheveldayoff's campaign in an email.
"As many political campaigns have demonstrated, polling is not reliable. Connecting directly with voters is our preferred method of gauging voter sentiment," said Beaudry-Mellor's campaign in an email.