But will we learn to love the internet pollster?
Peter Kellner’s sometimes controversial internet polling organisation, YouGov, which was floated on the financial markets in May, has launched a series of measures designed, it appears, to get people to think of it more favourably. These include:-
Because the firm can run polls cheaper and faster than conventional interview-based surveys it has never been popular with the rest of the polling industry who see it as eating into their business. Competitors are quick to point out that because YouGov political surveys are restricted to selections from the 40,000 – 50,000 members of its “polling club” then those that are surveyed are self-selecting and not fully representative.
A sign of the antagonism is that the MORI web-site does not include YouGov surveys at all in its otherwise excellent compilation of what it describes as “all companies’ polls” – no doubt because of its view of the internet pollster. With UK PollingReport covering everything the usefulness of the MORI table has started to decline.
We had a bit of fun on the site last March with an article on how you could “improve your YouGov ratings” by telling the pollster that you read the Sun and never watched Newsnight. This was because they usually have a shortage of Sun readers in their samples so that the views of those who are registered as reader’s of the paper are given a much bigger rating than those who read the Guardian.
As a punter I’ve made big money betting against the polls but I am less inclined to risk cash in defiance YouGov compared with the other firms. The self-selecting nature of those surveyed is a minor problem compared with the challenges facing the phone pollsters. Using randomised unsolicited phone calls a pollster needs to make five or six calls for every successful interview and for some reason those they get through to are much more likely to be Labour supporters than the population as a whole.
Like the other pollsters YouGov when tested against real results is a Labour over-stater, but not by very much. Its final three General Election surveys gave Labour leads of 3, 4, and 5 points against the 3 points that actually happened. ICM, by comparison, had 7, 8 and 6 point margins all of which had an impact on the Labour vote share spread-betting markets providing profit opportunities.
Until now Anthony Wells has operated UK Polling Report and his associated blog as a hobby. His “day job” had been with the Conservative Party working in the Leader’s office. It is a tribute to him that he kept his site very objective – let’s hope he is able to do the same within the YouGov stable. Knowing Anthony I am sure that he will.
Tory leadership polling news – another boost for Clark A Populus survey of Tory party chairman in the Times has them supporting Ken Clarke over David Davis by a ratio of 2-1.