Have the “funny surveys” been over-stating Labour?

The table has been adapted from Wikipedia and shows the incredible variation in the polling surveys of the regional or list vote ahead of next Thursday’s crucial election for the Scottish Parliament.

On the one hand there have been the surveys by those pollsters which are listed as members of the British Polling Council which are required as part of their membership to follow strict transparency rules. Within two working days of a poll being published the BPC code requires basic detail to be made public so it is possible to analyse how the headline figures were arrived at.

From the above table YouGov, ICM, Populus, and TNS System Three are BPC members and have been carrying out political opinion polling for some time. Just compare their findings for Labour with those from the non-listed firms – MRUK and Scottish Opinion.

During April the BPC members have found Labour shares within a solid 25-27% band. Three of the surveys from the non-listed firms have found shares of 38%, 37%, and 34%. These numbers stretch credulity when you consider that in the Regional list in 2003 Labour chalked up 29.3% of the vote.

A lot has happened over the past four years to Labour and the idea that the party might be securing more than a quarter more votes this time is a total and utter nonsense.

A couple of weeks ago I proposed a wager of £1,000 here to MRUK and the Herald that the predicted Labour shares would be proved wrong. Surprise surprise this has not been taken up.

Once this election is over I plan to press the Press Complaints Commission to clampdown on papers describing as opinion polls surveys where the BPC transparency rules do not operate.

Mike Smithson