Populus shows big Tory bounce-back

Populus shows big Tory bounce-back

    Is a Tory BUY now good bet?

Today’s Populus Poll in the Times shows a big recovery for the Tories after a tricky autumn. The figures are:

LAB 34 (-1): CON 33 (+5): LD 22 (-3): UKIP 2 (-2)

This has been helped by the expcted decline in UKIP following the leadership rifts and Robert Kilroy-Silk’s decision to detach himself from his MEP colleagues in Brussels.

The poll itself and the big coverage being given to it by the Times almost certainly means that the current 200-208 spread betting price on the number of seats the Tories will win at the General Election will rise after being at its lowest levels since Michael Howard became leader.

The Labour rating is substantially lower than the 39% recorded 10 days ago by Commuicate Research, 37% by ICM and 36% by YouGov. Part of this might be driven by a change in Populus methodology rather than any change in the political climate. Today’s paper reports:-

After a review by The Times and Populus, a number of changes have been made to the way that the poll is undertaken, though they should make no more than one percentage point difference to the voting shares of each of the parties. First, the sample size on voting intentions has risen from 1,000 to 1,500. This should reduce the margin of error. Second, the formula for assessing likelihood to vote has been changed from just including 6 to 10 on a ten-point scale to weighting depending on the position on the scale. Third, the formula for adjusting for past votes recalled by interviewees has been modified to take account of more recent data. This is to ensure that the sample is politically as well as demographically representative.

These changes contrast sharply with the Independent’s new pollster, Communicate Research which ten days ago had LAB 39: CON 33: LD 17 . CR do not weight by recalled past vote. This is important because at the 2001 General Election Labour got 42% of the vote. So if you carry out an opinion poll and 58% of those surveyed told you, as in the last ICM poll, that they had voted Labour you would know that something was not quite right.

Of course the interviewees might have forgotten or not have been telling the truth but with such a disparity the sample is probably not representative. ICM and Populus make an adjustment to their raw data to take to take into account the recalled last vote answer. They do not weight for the full amount but do change their figures significantly. According to polling commentator Anthony Wells, CR’s approach leads to an of over-stating Labour and understating the Lib Dems and the Tories.

We have been a constant critic of the polls for the way they always understate the Tories and over-state Labour. These Populus changes are welcome and should lead to slightly smaller ratings for Labour.

    As a betting strategy BUY Tory or SELL Labour at the start of each month – after CR’s Indy poll which should have a higher Labour figure and before the Populus Poll comes out which is more likely to be lower

LINKS CHANGE We have now set up four new links to our coverage of the latest polls from ICM, CR, YouGov and Populus.


Comments are closed.