Polls apart
One day it’s a PC majority, the next it’s the Liberals riding to
victory. Ontario election 2014 is doing little to boost our faith in
public opinion surveys. The polls may be all over the place, but it’s
not too early to spot trends. Here’s where the pollsters stood as of
press time Wednesday afternoon.
Key metric Upswing in likely voters who say they’ll cast their ballot for the Liberals; Tory support dipping to 30 per cent after sitting at 35 per cent in an Abacus poll only days earlier.
Methodology gap Abacus’s was an online survey, which tend to skew right politically, but each of its last three polls has shown a trend: an upward trajectory for the Grits.
Key metric Three-quarters of voters polled say they want change.
Methodology gap Small sample size (868 voters); margin of error was 4 per cent.
Key metric Significant sample size of 1,215 respondents.
Methodology gap The results of the Interactive Voice Response poll are from a random sample of 927 people and were heavily weighted. Most of the other polls are of likely voters. And the 17 per cent NDP support must be a blip, right?
Key metric Zero margin of error.
Methodology gap The Oracle survey was of likely voters and conducted over four days with live interviewers over the phone, usually considered the best method.
Key metric Dip for Libs, which polled at 41 per cent in a much earlier Forum survey.
Methodology gap Forum’s polls are predicting a Liberal minority when vote concentration is factored into the equation.
Abacus Data
Predicting Liberal majority (maybe).Key metric Upswing in likely voters who say they’ll cast their ballot for the Liberals; Tory support dipping to 30 per cent after sitting at 35 per cent in an Abacus poll only days earlier.
Methodology gap Abacus’s was an online survey, which tend to skew right politically, but each of its last three polls has shown a trend: an upward trajectory for the Grits.
Ipsos-Reid
Predicting PC majority with 41 per cent of the vote and Libs at 29.Key metric Three-quarters of voters polled say they want change.
Methodology gap Small sample size (868 voters); margin of error was 4 per cent.
EKOS Research
Predicting Liberal majority with 39 per cent of the vote.Key metric Significant sample size of 1,215 respondents.
Methodology gap The results of the Interactive Voice Response poll are from a random sample of 927 people and were heavily weighted. Most of the other polls are of likely voters. And the 17 per cent NDP support must be a blip, right?
Oracle Poll Research
Predicting PC minority with 36 per cent of the vote.Key metric Zero margin of error.
Methodology gap The Oracle survey was of likely voters and conducted over four days with live interviewers over the phone, usually considered the best method.
Forum Research
Predicting A toss-up, with PCs and Libs each at 36 per cent.Key metric Dip for Libs, which polled at 41 per cent in a much earlier Forum survey.
Methodology gap Forum’s polls are predicting a Liberal minority when vote concentration is factored into the equation.
• NOW | June 5-12, 2014 | VOL 33 NO 40
RECENT STORIES BY ENZO DIMATTEO
Can Wynne convince voters that government is still a force for good?
Kathleen Wynne or lose?
Rob Ford: from distraction to abstraction
The hex on Hudak
No comments:
Post a Comment