Er, I don’t think averaging five results–from a …

Comment posted on Russell Bruce: What do the Guardian monthly polls tell us? by Stephen Mackenzie

Er, I don’t think averaging five results–from a small sample with significant uncertainty in them–will produce a more precise set of numbers. Just the opposite in fact.
Also, what kind of average? Yes, there is more than one.

Stephen Mackenzie also commented

  • Thanks Russell. I’m still a little concerned about the methodology, but I should make clear I don’t disagree radically with your actual conclusions.

Recent comments by Stephen Mackenzie

powered by SEO Super Comments

Share and Enjoy:
  • Digg
  • del.icio.us
  • Facebook
  • Google Bookmarks
  • email
  • LinkedIn
  • Technorati
  • TwitThis
  • Ma.gnolia
  • NewsVine
  • StumbleUpon
  • SphereIt
  • Reddit
  • Slashdot
  • Print

5 Responses to Er, I don’t think averaging five results–from a …

  1. I think we have to insert a chilling insight into the power of the media over the SNP crusade. The SNP is no longer fighting the Labour Party (or the defunct LibDems or the paralysed Tories). They are all politically inert in Scotland and increasingly reliant on destructive copy against the SNP in the papers and on TV.
    Cast your eyes back to the Scottish Parliament Election in 2011.
    It may be almost impossible to believe this now but every newspaper in Scotland except the Mirror group titles (Record, Sunday Mail) supported the SNP in that election to a greater or lesser extent. They were joined by a number of high profile, well known unionists like Sir David Murray. This suport suddenly started about a month before the election and was, I’m sure, orchestrated.
    The calculation was that a new Westminster Tory/Lib Dem coalition struggling in Government couln’t afford a Labour win in the Scottish election. A narrow SNP win was the lesser of two evils.
    They succeeded too well however and gave the SNP an overall majority (they probably believed the distorted polls they kept commissioning to start with). The unintended consequence of MSM support for the SNP was an unprecedented SNP victory – and an indication of where we would be with an honest press.
    By contrast the recent six week concentrated campaign of demonisation of Alex Salmond (which is being extended to Nicola Sturgeon) probably cost us about 10 points in actual ratings and 50 to 80 councillors.
    The good news is that AS dealt with this attack sensibly by basically ignoring it and remaining friendly, progressive and positive. It is a tactic they can probably only use effectively very rarely.
    It has diminishing returns and if pushed too hard has the opposite effect to its intention. This happened at the last Westminster election at which the appalling ad hominen attacks on Gordon Brown began to backfire and almost saved Labour in the final count and provided a hung result.
    This suits the powers that be and who own the media. They don’t really care which right wing grouping – Tory or Labour – rules UK as long as they are in control of it – which is the present situation.
    We live in a very imperfect democracy.

    Like or Dislike: Thumb up 0 Thumb down 0

  2. Er, I don’t think averaging five results–from a small sample with significant uncertainty in them–will produce a more precise set of numbers. Just the opposite in fact.
    Also, what kind of average? Yes, there is more than one.

    Like or Dislike: Thumb up 0 Thumb down 0

    • Stephen
      You are quite right about the dangers of reading too much into small samples which is why this analysis was carried out by aggregating five months of Scottish responses and averaging the results for each party. The correlation with the actual local election results in Scotland this month does support the methodology employed here.
      ICM have a fair track record and the weighting they use recognises the distinctive nature of politics in Scotland compared with England and other parts of the UK.
      YouGov poll weightings in contrast are distinctly suspect in comparison as they consistently predict Conservative support in Scotland at around 20% and it is a very long time since the Tories achieved that in Scotland.
      A poll of polls as a broader indicator of trends is a long established method of ironing out individual sampling deficiencies.
      The article only commented on the Scottish samples. A interesting feature of the latest May poll is that despite the recent few bad weeks for the Tories, the all England breakdown puts The Tories on 40% and Labour on 42%. Within the margin of error and suggests the Labour Party are not making any real headway

      Like or Dislike: Thumb up 0 Thumb down 0

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>


All the latest comments (including yours) straight to your mailbox, everyday! Click here to subscribe.