h1

Contrasting LAB-CON figures from Mori and YouGov

February 24th, 2006

big ben collage.jpg

    But good news for whoever wins the Lib Dem contest

Two new polls this morning paint a very different picture of how the Tories and Labour are doing but both show a Lib Dem recovery.

  • Mori in the Sun has with changes on the pollster’s last survey a month ago CON 35 (-5): LAB 38 (nc): LD 20 (+3).
  • YouGov in the Telegraph has with changes on the pollster’s last survey 12 days ago CON 38 (+1): LAB 36 (-3): LD 18 (+3).
  • So with only five day left before their leadership contest ends the Lib Dems will be pleased to see that all the pollsters are now showing significant recoveries compared with what things were like in the immediate aftermath of the Charles Kennedy, Mark Oaten and Simon Hughes revelations.

    The YouGov changes are broadly in line with what ICM recorded in the Guardian on Tuesday – the Tories staying relatively constant but Labour slipping back as the Lib Dems recover.

    Mori continues its reputation as the most variable of the pollsters. Its last four polls have shown: November LAB+10, December LAB -9: January 17th LAB level: January January 23rd LAB -2.

    Pollster variations since November 2005
    Mori LAB lead/deficit in a range from +10 to -9
    YouGov LAB lead/deficit in a range +2 to -2
    Populus LAB lead/deficit in a range +8 to -1
    ICM LAB lead/deficit in a range +5 to -4

    The main reason for the differences is down to the methodologies the pollsters use:

  • YouGov tends to be the most stable because it carries out its surveys online amongst its polling panel and weights its samples in line with recorded party allegiances. YouGov, except in its pre-General Election polls – does take into account the issue of whether people are certain to vote. Lasy May it did and this reduced the Labour margin.
  • Populus and ICM seek to ensure they have representative samples by weighting in accordance with how people said they voted last May. their figures take into account answers to the “how likely is it that you will vote?” question?”
  • Mori do not make recall or allegiance calculations. The headline figures are based on those saying they are certain to vote.
  • Mike Smithson






    Comments are closed.