Another way of looking at the ICM poll

Another way of looking at the ICM poll

The statistics they never tell you
As we predicted yesterday’s ICM poll in the Guardian pushed up Labour prices and depressed Tory ones on the spread, seat and overall winner markets.

  • The latest IG spreads are:- LAB 342-350 (+2): CON 198-206 (-2): LD 69-73 (+1).
  • SportingIndex is LAB 342-350: CON 205-213: LD 70-74 – the discrepancy based on the fact that they are using the old larget House of Commons – not the new reduced sized one.
  • The best boomaker price on Labour winning most seats has tightened further from 1/6 to 1/7
  • The Labour seats market now has the 360+ band as 5/4 favourite
  • But would the markets have moved if the Guardian had plastered all over their front page that the equivalent of nearly a quarter of those surveyed who said they had voted Labour last time were not going to do so again?

    The party’s ICM poll rating of 38% assumes that just one in ten of 2001 voters would not be going with them again. Yet the data from the poll seems to show otherwise.

  • The Labour vote in the November ICM survey was just 77% of those who said they had voted for the party in 2001
  • The Tory vote in the survey was 101% of those who said that had voted for the party last time.
  • The Lib Dem share, was 117% of those who said they voted for the party in 2001.
  • Opinion polls are not usually reported like this but you cannot easily dismiss the fact that such a large proportion of those saying they were Labour last time have apparently deserted the party.

    A lot of people would argue about the way we”ve worked out these number. Many of those surveyed say they’ve voted when they haven’t and Lib Dem supporters have a strong tendency to “forget” when polled where they put their cross. But the difference between those who told ICM they had voted Labour last time and those saying they will do so in the General Election is so substantial that it says something about they way polling surveys are processed.

    The ICM past voting weighting methodology has recently been criticised by Populus who say that if applied to their latest poll it would have increased the Labour margin by 4%. We think that the Labour position is weaker than ICM suggests.

    BETTING STRATEGY Given that in the main the polls that come out at the turn of each month, Populus and YouGov, tend to show lower Labour totals then punters wanting to bet on the party would be advised to wait for these when the prices might have eased. Similarly those wanting to bet against Labour or for the Tories should do so at this time in the month though we see no urgent reason to do so now.

    The division between the pollsters will reflect itself in the markets and you should time your bets accordingly.

    Comments are closed.