Friday, 31 October 2008



http://politicalbetting.com/index.php/archives/2008/10/30/and-what-about-the-weekly-bbc-polling-scandal/
h1

And what about the weekly BBC polling “scandal”?

October 30th, 2008


    When are they going to clamp-down on pro-Labour samples?

One most Fridays when parliament is sitting BBC2’s “Daily Politics” programme announces the results of its latest ComRes poll on matters of current concern. This generally gets picked up by other parts of the media who treat it, quite naturally, like any other ComRes political poll. It’s not and the BBC should say so.

For the Daily Politics polls have one fundamental difference compared with the standard ComRes voting intention surveys - there’s no effort to ensure a politically balanced sample which almost inevitably means that its skewed towards Labour supporters.

    In every single ComRes voting intention poll since they introduced past vote weighting the influence of Labour supporters has had to be scaled back when working out the final figures - yet the BBC allows its political surveys on matters a significant current political interest to be carried out without this adjustment.

It’s been known since the early 90s that a disproportionate number of Labour supporters respond to unsolicited randomised polling phone calls. Who knows the reason? Ipsos-MORI think it’s because they get too many public sector workers answer the phone. Whatever - those involved in polling recognise the issue and do something about it.

In fact for the first time ever we will be going into a general election when every single pollster that carries out voting intention polling will be taking measures of one sort or another to avoid political sample bias.

This should be applied to the Daily Politics surveys even though BBC rules prevent voting intention questions. If that’s not possible then Andrew Neil needs to explain that no effort has been made to make such an adjustment and their poll cannot be compared with a standard ComRes poll.

And other parts of the media need to treat these surveys with the utmost caution. Without this they are misleading their viewers, readers and the public at large.