New Opinium and ComRes polls
ComRes's monthly online poll for the Indy on Sunday and Sunday Mirror has topline figures of CON 32%(+4), LAB 35%(-1), LDEM 9%(-1), UKIP 16%(-1). Changes are from their poll a month ago, conducted just after the Lib Dem conference. The 32% for the Conservatives is the highest ComRes have shown in their online polls since January, and the three point Labour lead the lowest since before the omnishambles budget in 2012 (in recent months ComRes's online polls have tended to show lower support for the Conservatives than their phone polls). Tabs are here.
In contrast the fortnightly Opinium poll for the Observer has no narrowing, with topline figures of CON 27%(-2), LAB 38%(+2), LDEM 9%(+2), UKIP 17%(+2), popping Labour back into a double-digit lead for the first time since July for Opinium.
There have been some apparently conflicting polls in recent weeks - some like MORI and ComRes showing things narrowing sharply, some like Opinium and Survation still showing double point leads, some like Populus showing all remaining steady. Remember that all polls have a margin of error, all are unavoidably subject to sample error, the ebb and flow of random chance. Don't focus too much on individual polls, look at the broad trend, which is a Labour lead in the mid-single digits. It's certainly down from the bigger Labour leads we were seeing this Spring, but more recent polls are not showing any strong trend either way.
ComRes's poll also had some quite fun split sample questions, seeing what effect mentioning the party leaders in questions and attributing policies had on answers. Party preference on economic issues showed no real difference between asking if people preferred Conservative or Labour or Cameron or Miliband (as YouGov and ComRes have found before, the more general finding was that Labour lead on "cost of living" type measures, the Conservatives on more general economic competence). They also asked about a couple of policies, attributing them to David Cameron and Ed Miliband in one half of the sample, presenting the bare policies to the other half. On the two Conservative policies it made no difference mentioning David Cameron's name, on the energy price question support was six percentage points lower when Ed Miliband was mentioned... interesting, but it's only one data point (and the policies were very high-profile ones that presumably lots of people in the control group knew were associated with Ed Miliband and David Cameron respectively).
