A Manitoba election survey that originally showed Manitoba’s NDP and PC parties neck-and-neck with close to 30 per cent of decided voters’ support each was flawed due to overweighting of responses from Northern Manitoba, says the company that conducted the survey.
Converso Research, part of Converso Engagement Services, said Aug. 19 that they noticed a discrepancy between regional and provincial results after releasing the results of their survey Aug. 16.
“The discrepancy was caused by an overweighting of responses from northern Manitoba,” the company said in a statement Aug. 19. “We apologize for the error and recognize the importance of providing Manitobans with an unbiased, independent election survey. We feel confident that we have now fulfilled our intention.”
Converso released updated results of the survey, generated by a qualified third party, with their Aug. 19 statement. They showed 35 per cent support for the Progressive Conservative party, 21 per cent for the NDP, 12 per cent for the Liberals and eight per cent for the Green Party. Fifteen per cent of respondents were undecided, while five per cent didn’t want to say, two per cent said they weren’t going to vote and one per cent planned to vote for another party.
The interactive voice response (IVR) survey polled 1,127 Manitobans between July 8 and Aug. 7. A representative population sample of this size would be accurate to within three per cent, 19 times out of 20, and Converso said this survey, the company’s first public opinion research survey, was weighted in line with Statistics Canada’s 2016 census data for the populations sampled.
