Wednesday Round Up
D minus 29 Tonight's polls: Angus Reid/Political Betting (6th Apr-7th Apr) CON 37%(-1), LAB 26%(-1), LDEM 22%(+2) YouGov/Sun (6th Apr-7th Apr) CON 37%(-3), LAB 32%(nc), LDEM 19%(+2) Populus/Times (6th Apr-7th Apr) CON 39%(-1), LAB 32%(+2), LDEM 21%(+1)
Three polls tonight. YouGov and Angus Reid polls can be compared with their polling from just before the campaign started, and both show a boost in Lib Dem support, though not really enough to distinguish a trend from random variation.
With YouGov's Tory score dropping below 40 again those pollsters who have reported in the last week or so are now reporting pretty similar figures for the Conservatives, all in the range of 37%-39%. There is rather more variation in the other parties - the more established pollsters (YouGov, ICM and Populus) have Labour between 32%-33%, Opinium and Angus Reid have them on 29% and 26%. For the Liberal Democrats Opinium tend to give them very low figures and are at 17%, the other companies vary between 19% and 22%.
Considering National Insurance has dominated the political debate over the past few days there is a surprising lack of polling on it. YouGov asked about it for the Sun last week and found the public evenly split 43% in favour of the Conservative policy, 43% against, but that was before the business leaders came out in support, and before a lot of debate on the topic. Populus may have asked about it, but their report only quotes the results amongst those people who wanted a change but not to the Tories - it is unclear whether the whole sample was asked.
Two other articles on polling today worth a read - Mark Pack here and John Curtice in the Indy here. the most important thing to take away from them (and something that really can't be mentioned often enough) is the volatily you should expect to see from normal sample error. Party support from a single pollster should randomly vary a couple of points in either direction from poll to poll (the lead will be even more volatile, since you've got random variation on two numbers). While I get sick of typing it, there is a reason why I end up stating in almost every post that the movement since the last poll is not in itself significant! It's the trends across many polls that count.
