On the night of Monday, May 2, Frank Graves, president of EKOS Research Associates, opened a bottle of Mer Soleil in his Ottawa home and settled in front of the television to watch the federal election results. He felt smug, very smug.
He had every reason to believe that he had surveyed as accurately throughout the 2011 campaign as he had in 2008, when he forecast that election's almost exact outcome.
He had been exceptionally careful with his methodology and analysis. His numbers had been stable for three days. To avoid being caught by last-minute voter shifts, he had surveyed as late as Sunday, later than in any previous campaign.
He pleasurably imagined incoming kudos from his clients. Instead …
http://www.theglobeandmail.com/news/politics/a-pollsters-painful-reckoning-how-could-i-have-screwed-up-so-badly/article2065573/http://www.ekospolitics.com/wp-content/uploads/accurate_polling_flawed_forecast.pdf