Calling it a "difficult decision," Frank Graves, the president of Ekos Research Associates, said he and his media clients, the Toronto Star and Montreal's La Presse, agreed to do further polling yesterday to increase the sample size to 1,200 respondents. He confirmed the weekend findings -- from a sample of 500 calls -- indicated Stephen Harper and the Conservatives were on their way to forming a majority government similar to the ones enjoyed by Brian Mulroney in 1984 and Jean Chretien in 1993.
"In a normal setting we would have released it, but in the context of a 500-case poll with the results we had in last night, we didn't think it would be responsible on the day of the debate," Mr. Graves said.
By waiting an extra day and polling an additional 700 respondents, the margin of error falls to 2.9 percentage points, 19 times out of 20, from plus or minus 4.5 percentage points, 19 times out of 20.
The newspapers did not release the results from the smaller sample yesterday because they planned to publish the results from the larger sample today.
They weren't afraid to publish polls last month that had the Liberals 10 or 12 points up, with banner front-page headlines. And they didn't sample more voters to make sure, either.
Makes you wonder what else the papers think is too "irresponsible" to publish.
If the Tories do win a majority in two weeks time, perhaps the press can lean on Elections Canada not to release the results until they've had another two weeks to sample those who didn't vote, just to be sure.
Source: Ottawa Citizen