h1

The Indy’s CR poll methodology under “constant review”

November 29th, 2006

communicate - the indy1.jpg

    Encouraging response to PBC’s observations

Following yesterday’s article and thread on the methodology being used for the new month Communicate Research poll in the Independent I have had an encouraging response from both the pollster and the newspaper.

I wrote to both making the same point: “What convinces me about the need for past vote weighting is the sheer consistency of the figures that both Populus and ICM get when they ask how people voted last time. I’ve been monitoring this since the General Election and although there might be the odd spike in almost every three month period on a rolling average since May 2005 the proportion from both pollsters recalling that they voted Labour has not strayed outside the 44-45% range. Given that 36.2% actually voted for the party on May 5th 2005 there would seem to be a strong case for some sort of remedial measure.”

This is a point I have been making irrespective of the poll shares that the surveys produce. The same observation appeared last month when there was a Tory lead as this month with Labour ahead.

This is the response I have got from the paper’s political editor, Andrew Grice. “We will take the point you make into account. The Independent worked out and agreed the methodology jointly with Communicate Research but we have an open mind about it and keep it under constant review.”

    This is clearly very welcome and should be applauded. We should also be delighted that the paper has now decided to commission a poll every month adding to the overall amount of polling information that will be available.

Although I have been critical of CR about the lack of past vote weighting the pollster is using some innovative measures and it will be fascinating to watch how these pans out. The firm also gets it detailed poll data out on its website very quickly which is very helpful.

Mike Smithson






Comments are closed.