What to Make of Opinion Polls and By-Elections
As regular readers will know I follow and report on opinion polls. I have done so for the last 30 years and have developed a keen sense of which polls can be trusted and which have less credibility. At present the pollsters which have the most credibility, based on a reputation earned in the run up to recent elections, are ICM and YouGov. They should therefore be taken more seriously than some other pollsters.
However, even the best pollster can get things wrong and public opinion can shift significantly between a poll being conducted and actual votes being cast. Polls are a snapshot of current opinion not a predictor. Some people therefore argue that votes cast in real elections are the only true indicator of public opinion. That of course is a statement of the obvious as far as it goes but there are few real elections between General Elections and even those we do have, such as local elections, by-elections and Euro elections, are often affected by unusual factors. Euro elections have little impact on the individual elector and they are conducted under the party list system of PR, they also suffer from a pretty low level of turnout, together making them a very unreliable indicator of likely General Election voting.
Parliamentary by-elections are often given a great deal of scrutiny by the media and commentators. However, the best that can be said for them is that they are reflections of local opinion in the affected seat, levened with a hefty dose of protest voting, and are no more than a very rough guide to the wider national mood. Governments (of all parties) in mid-term tend to lose seats to the principal opposition party and opposition seats are normally held with increased majorities.
Local Council by-elections (which occur every week up and down the land) can be a reflector of national mood as well but they are usually very heavily affected by local factors. An individual local council by-election is of little significance in determining the national mood but every now an again one occurs which stands out from the crowd. One such occurred this week in Merseyside when Conservative Ian Lewis gained a seat from Labour with a swing of over 15% on a turnout of over 30%. The gain in a ward which voted Labour in 2010, 2011 and 2012 was remarkable, even more so in the north of England and in Merseyside.
Whilst the Conservative gain in Merseyside was noteworthy it should not be taken as an indicator of national mood. For that we would need a series of by-elections up and down the country in authorities of all political persuasion and in wards of all types. Fortunately such a set of statistics exist. I am grateful to the reader who has sent me a link to an article on PoliticalBetting.com where the votes cast in all 198 local Council by-elections held in 2012 have been added up and the vote shares calculated. These 300,000 real votes show the Conservatives leading Labour by 34% to 29%, with the Lib Dems on 19% and UKIP way behind on just 6%. This suggests that the Lib Dems are far better at holding onto their support in real elections than in the hypothetical situation of an opinion poll interview. If this is the case in the next General Election then it is very bad news for the Labour Party, as it suggests that the 10% transfer to them from the Lib Dems may revert when people come to decide who they want to represent them in their local Parliamentary seat. It is also a far more realistic indication of likely UKIP support at the next election when many seats will probably not even have a UKIP candidate.