The ‘Other’ thing about polls

Some opinion polling earlier this week showed that Labour’s lead over the Conservatives had dropped to 1 or 2%. This generated excitement among those who like the narrative that recent indications of economic improvement in the UK should be followed by an upturn in Tory support. Over-excitement in the case of the normally level-headed and fastidious Iain Martin:

At the time I was working on the previous post about the consequences of the main parties having the same share of the vote, so some statistics to be kept in mind were readily available and are described below.

In the UK we have the two main parties who have had vote shares in general elections since 1945 in the 30 to 50% range, and the Liberal Democrats (Liberals up to 1987) who have been somewhere between 2.5% and 25%. In the 2005 election UKIP, a party founded in 1993, had 2.2% of the votes and 3.1% in 2010. However, in the opinion polls for the last three years, UKIP’s support has been similar to that of the Lib Dems, as the chart below shows (from the School of Politics and International Relations at the University of Nottingham blog, Ballots and Bullets):

It could be argued that we now have two main parties and two secondary parties – but then what about the rest? The next chart shows the size of the “Other” vote in elections from 1945, defining “Other” as being not Conservative, Labour or Lib Dem, and not UKIP after 2005:

In no election since 1945 has “Other” been more than 10%. The Ballots and Bullets chart above implies that on a smoothed basis “Other” is 11.5% currently. So what about individual recent polls? Some from UK Polling Report for the last couple of weeks are shown in the table below, concentrating on the polling organisations which generated the ‘excitement’ referred to at the start of this post.

Those two polls are highlighted and, interestingly, show anomalous levels of support for “Other” which is usually 7 to 9% - the level in the last four general elections. The 20 January YouGov/Sun poll showing an 8% Labour lead also looks like an oddity.

Some advice for the commentariat, don’t get too excited about a poll if “Other” is less than 7% or more than 9%.


UPDATE 31 JANUARY 

Two more YouGov/Sun polls have appeared. The second shows a 10 point Labour lead, something which Iain Martin seems to have missed.

Scepticism is in order for both by my criterion above, certainly for the second one.