h1

Have Iain Dale’s voters been a bit premature?

December 29th, 2009

Iain Dale’s end of year elections:-
Pollster of the Year
1. YouGov
2. Angus Reid
3. ICM

How could they support an untested firm?

I’m sure that my friends at PB’s pollster, Angus Reid, won’t be too upset to learn that I did not vote for them as “pollster of the year” in Iain Dale’s annual elections. The Vancouver-based firm only started publishing UK voting intention surveys in October and clearly have yet to go through the process of having their findings tested against real results in an actual election here.

The only test that we’ve had during 2009 was the Euro elections on June 4th when both YouGov and Populus came pretty close in the battle for top pollster. My vote went to one of these two.

Although AR has a big international reputation it is new to the UK and its methodology challenges some British polling orthodoxies. A major area is the way it operates it past vote weighting formula. ICM, Comres and Populus don’t weight to the actual result of the last election but a notional one designed to deal with the problem of “false recall”.

For some reason more people say they remember voting Labour than actually did so and the pollsters weight to a notional 2005 outcome that usually has Labour about six points ahead against the actual margin of 2.9%.

Not so Angus Reid. It weights to the actual 2005 result – and this is one of the main reasons why it usually reports smaller shares for Brown’s party than the other firms.

Are they right? Who knows at this stage – but this is how the AR approach did at the last Canadian elections just over a year ago.

The UK of course, is different, but I was hellish impressed by this table when the arrangement between PB and Angus Reid was being set up.

Mike Smithson