UK General Election opinion polling problems

The opinion polls at the last general election were famously wrong. After they all consistently predicted a neck & neck result, the Conservatives won about 100 seats more than Labour, and easily won an overall majority. There are a series of different industry enquiries into the reasons behind this failure, one of which reported yesterday.

"Polling expert" Professor John Curtice, in a classic example of stating the bleeding obvious, has come up with the startling conclusion that not enough Conservative voters were included in the samples:

Quote:

Prof Curtice, who wrote the report for research agency NatCen, suggested polling difficulties arose "primarily because [pollsters] interviewed too many Labour supporters and not enough Conservatives".

Quote:

The failure of opinion pollsters to predict the outcome of May's general election may have been because Conservative voters are harder to track down, a report suggests. The disparity between forecasts and the eventual Conservative majority has been blamed on "shy Tory" voters or a late swing to David Cameron's party. But polling expert John Curtice highlighted sampling "deficiencies". His report said "more time and effort" was needed to find Conservative voters.

His report actually isn't as facile as the above suggests, and cites the much respected British Social Attitudes Survey which, although not in the game of political polling, put the Tories 6 points ahead of Labour, much the same as the actual winning margin.

The key finding is that the errors weren't due to "shy Tory" voters or a late swing to David Cameron's party.

BBC

via International Skeptics Forum http://ift.tt/1KeQpdC