New Populus, ICM and Panelbase surveys
The weekly Monday Populus poll is out, and has topline figures of CON 34%, LAB 38%, LDEM 12%, UKIP 8%. It was conducted over the weekend, and unlike the YouGov/Sunday Times poll obviously shows no impact on voting intention from the Syria defeat.
There is also a new ICM poll for the BBC. This doesn't have any voting intention figures (the BBC doesn't like commissioning them!), but has some more general questions on Syria. 71% think MPs were right to vote against military action (20% think they were wrong). Asked if they approve of how David Cameron has handled the situation in Syria 40% approve, 42% disapprove. The figures for Ed Miliband are 33% approve, 39% disapprove. Full tabs are here.
Finally today we have a Panelbase poll commissioned by the SNP that claims to show a Yes lead in the Scottish independence referendum, 44% to 43%. Now, I always advise extreme caution in dealing with polls commissioned by political parties - reputable pollsters should ensure the questions themselves are fair, but they don't decide what parties ask about, or how they spin the results.
On the Scottish referendum there are some big differences between the polling by Panelbase and the polling by other companies, but the trend data from each company is itself pretty steady. The regular polls from Ipsos MORI have bounced about a bit with NO leads between 20 and 28 points, but there is no obvious up or down trend. YouGov have only done a couple of polls, but have show consistent NO leads in the mid-twenties. There have only been two recent TNS BMRB polls (they are starting up a more regular series later this week) but they showed NO leads of 19 and 21 points. In contrast Panelbase has been tending to show leads of between 8 and 10 points. A much smaller lead, but again a very consistent one with no obviously trend towards yes or no - even by Panelbase's standards a Yes lead looks odd.
The results for the Panelbase/SNP survey are here. The referendum question itself is completely fair. However, it was not the first question asked. Typically polls always put voting intention questions (Parliamentary and referendum) questions right at the start in order to remove the risk that other questions in the survey could influence or skew voting intention results. Panelbase, for example, normally ask Holyrood VI, immediately followed by referendum VI. In the case of this poll the referendum voting intention question was asked at the end of the survey, after asking people if they thought Scotland could be "a successful, independent country" and whether people trusted the Scottish government or Westminster government more to take the right decisions for Scotland. Both questions had the potential to skew responses to the referendum question. We can't know for certain, but given the contrast with Panelbase's previous polls, my guess is that this is what has happened and Panelbase's next normal survey will be back to a NO lead.
