This is the latest in a series of monthly posts that report on the state of the parties as measured by opinion polls. By pooling together all the available polling evidence we can reduce the impact of the random variation each individual survey inevitably produces. Most of the short term advances and setbacks in party polling fortunes are nothing more than noise; the underlying trends – in which we are interested and best assess the parties’ standings – are relatively stable and little influenced by day-to-day events. If there can ever be a definitive assessment of the parties’ standings, this is it. Further details of the method we use to build our estimates of public opinion can be found here.
Based on polls taken before David Cameron’s Friday morning veto, our estimates suggest support for the main parties remains stable at the levels we have seen for much of the year: the Conservatives are up 0.3 points at 34.9%, Labour are down 0.5 points at 38.6 and the Lib Dems are up 0.7 points at 8.5%. As our chart shows, the central estimates of voting intention we develop by aggregating all the polling data together have been more or less flat since the early spring. Journalists keen for a story might despair, but none of the headline-grabbing events of a busy year in politics has yet led to a lasting change in British voter sentiment.
This picture of stability may seem at odds with some of the headlines in early December, when a poll by ICM suggested the Conservatives had returned to the lead. Such polls can be misleading when taken in isolation. When the two main parties are relatively close together in the polls, there will be the occasional poll which will show them trading places due simply to random variation, even if there is no underlying change in sentiment. Such polls tend to generate disproportionate attention, because journalists and the public at large focus heavily on the horse race, and get very excited when a new horse takes the lead.
The results from single polls can also be misleading because they do not take into account the systematic differences in how polling companies conduct their polls, which can lead to variation in the estimates of opinion they come up with. In the case of ICM, as Anthony Wells has pointed out in a recent blog post they reallocate those who say they don’t know how they would vote to the party they say they voted for in the previous general election. At present, that practice should boost the Liberal Democrats, whose support has fallen sharply since the general election, and reduce support for Labour, who have advanced since the 2010 poll. Our model confirms this is the case: we estimate that ICM polls show the Lib Dems 2.5% higher, and Labour 1% lower than the average (median) pollster. The net effect of this is that ICM polls tend to suggest a closer race than those conducted by other pollsters, and as a result they are particularly likely to show the leading two parties switching places: of the five polls to show a Conservative lead since the start of 2011, four have been by ICM.
It is impossible to know without a new election which pollsters’ practices actually produce the best estimate of party support, but there are reasons to think that ICM’s practice may be problematic in the current political environment. In previous election cycles, the Liberal Democrats have tended to lose some support outside of campaign periods, owing in particular to the difficulty third parties faces getting attention in a system dominated by two larger parties. The current situation is quite different: the Liberal Democrats are now part of the governing coalition, and the sharp fall in their vote share most likely reflects the loss of left-leaning voters who dislike their decision to ally with the Conservatives. Many of these voters have switched to Labour, while others may still be sitting on the fence.
It is quite plausible that this shift in voter sentiment, which reflects a dramatic and largely unanticipated shift in the political position of the Liberal Democrats, from centre-left opposition party to junior partner in a centre-right Coalition, will prove to be a lasting one. ICM’s assumption that the preferences of the undecided will revert to those held in the general election may therefore be leading them to underestimate the realignment of voter preferences. Time will tell, but the message for now is to treat any claims of a Conservative resurgence with caution, and wait for evidence from a wide range of polls.