By JESSICA BRUNO, KRISTEN SHANE
May 9, 2011
Election pollsters and observers say pollsters generally did well to provide an accurate picture of election day results—even with the late-stage NDP surge—but if they were off, their results often showed lower Conservative support than what occurred.
“The pollsters did a pretty good job despite all the novel changes going on. Compared to final results, there was a slight under-estimate of the Conservative popular vote in many cases and a slight overestimate of the Liberal popular vote,” Carleton University Professor Scott Bennett said in an email response to questions from The Hill Times last week. Mr. Bennett specializes in survey research and public opinion.
Angus Reid Public Opinion, which worked with the Toronto Star and La Presse; Nanos Research, linked to CTV and the Globe and Mail; and Ipsos Reid, which partnered with Postmedia News and Global National, all came within a few points of the final popular vote tally in their final snapshots of voter intention taken during the campaign’s last week. Nanos, Ipsos and Abacus Data Inc. all indicated results reflecting the actual popular vote, within their margins of error.