A survey run on Facebook the weekend before the Oct. 31 territorial election correctly predicted 16 of the 22 soon-to-be MLAs.
The survey, said its creator Robin Anawak, was a spur-of-the-moment thing. The budding pollster is a University of Ottawa law student originally from Rankin Inlet, now making Iqaluit his base. He is on leave from the GN’s Department of Justice.
He’s the first to say the survey is “unscientific”, without the necessary safeguards in place for credible results.
It lacked “the granularity of control over the survey that I would want next time,” he said.
And there will be a next time. The main goal was to see if others had his same interest in elections.
“I’m a bit of a political creature and I like reading polls, reading platforms and policies of political parties during elections across Canada. We don’t have anything like that in Nunavut. We don’t have any measures of popular support for candidates. We don’t have any mechanisms to gauge that kind of support,” said Anawak.
“But I find that those kinds of mechanisms not only give voters a better idea of the kind of support candidates have but it leads to dialogue and discussion about candidates and about issues.”
The fact that he received 133 responses in a quick weekend poll proved to him others shared his interest.
Anawak doesn’t think a respondent can submit the survey more than once and thereby skew the results, but he’s not sure.
“It’s my first time doing one of these surveys,” he said.
A quick trip to Survey Monkey, the site Anawak used, shows that the program’s default is to reject an IP address that shows up more than once.
“I don’t know who filled out the survey and in which riding they were,” he notes, adding the survey is anonymous. But he figures the 133 respondents were likely mostly Iqalungmiut because he posted it in the Iqaluit-based Nunavut General Elections 2017 Facebook group. Not all of its 1,452 members are located in Iqaluit, though.
Of the 22 ridings, Uqqummiut was the one the survey got the most wrong, predicting a Jerry Natanine landslide win over incumbent Pauloosie Keeyootak – 53.79 per cent of the vote for Natanine and 12.12 per cent for Keyootak.
“Pauloosie Keyootak seems to have a popularity that wasn’t represented in the survey. Probably because a large majority of people outside of that riding guessing that Jerry Natanine was going to win, but Pauloosie having a firmer local base who may not have participated in the survey,” said Anawak.
Uqqummiut is also a riding famously split into two communities, a circumstance which often sees split votes on the ground.
The poll did predict the George Hickes win over Jacopoosie Peter, with 77 per cent of the vote – to the percentage. It also predicted Elisapee Sheutiapik’s win over incumbent Paul Okalik, only off by two per cent of the vote.
In Cambridge Bay, the poll predicted a Jeannie Ehaloak win over Pam Gross, with a 14 per cent margin. In reality, the two candidates ran such a close race a judicial recount is in the works as per Nunavut legislation.
“The survey itself only took about 15 minutes to make. It wasn’t very comprehensive. I mean, it was just a list of candidates in different ridings,” said Anawak.
“But because of the interest I saw in it, next election I do want to work with others to develop a more comprehensive survey asking not only about popularity of candidates but about the popularity of different issues. What issues do people want to see discussed and addressed by our political leadership.”
He figures he’ll get started on a quality, scientific, credible survey a month or two before the next writ drops.
“The more people discussing these kinds of things, the better,” he said.