Lessons for why the polls seem to be so wrong lately?
. . . But changes in communications are threatening the viability of public election polling in many developed countries where the landline phone was once a reliable medium for representative surveys
Like their counterparts in the United States, Britons are abandoning their landlines. Telephone pollsters have shifted, contacting an increasing percentage of their respondents via mobile phone, but perhaps not enough, according to many pollsters. In the United States, polling mobile phones is far more expensive than relying on landlines, because survey researchers are required by law to enter cellphone numbers manually when dialing. . . .
Alberto Nardelli, data editor at The Guardian newspaper, noted last month that phone polls tended to favor the Tories, while Labour performed better in Internet polls. That phenomenon existed as late as a week ago, according to the New Statesman’s “May 2015” website.
“I would be concerned about what looks like systematic errors in the British polling — under-representation of Conservative voters and overestimation of support for the Labour Party,” said University of Michigan professor Michael Traugott. “I haven’t seen any polls that suggested that the Scottish National Party was going to do extremely well in Scotland at the expense of Labour primarily. So, systematic errors like that, patterned errors like that, suggest methodological problems.”
(Similarly, in Israel, some observers believe Web polling was biased against Netanyahu’s Likud Party and swayed the overall balance of pre-election forecasts.) . . .
Telephone polls are increasingly unreliable and too expensive, while Internet polling isn’t yet able to replace truly random surveys. . . .