Wednesday, April 28, 2021

Ashcroft poll shows statistical tie on independence

Thanks to Marcia for pointing me in the direction of the new Ashcroft poll, which reinforces the impression from the recent Survation poll that the traditional dip in Yes support caused by election campaigns (ie. because the SNP usually campaign on bread and butter issues rather than making the case for independence) is very slight on this occasion, and that public opinion is still essentially 50/50.

Should Scotland be an independent country? (Lord Ashcroft poll)

Yes 49%
No 51%

I've had a very quick skim through the report on the poll, and there doesn't appear to be conventional Holyrood voting intention numbers.  However, there are questions that ask respondents for their percentage likelihood of voting for each party, and 4% of people say they are at least 70% likely to vote Alba on the list.

Patrick Harvie's name is misspelt as 'Harvey', which makes me wonder how well-grounded in Scottish politics Lord Ashcroft (or whoever wrote the report) really is.

Meanwhile, Newsnet Scotland has published a very silly article about the new Scot Goes Pop / Panelbase poll.  I'm not sure whether it's worth the bother of rebutting it, given that most people are intelligent enough to spot that Newsnet under its new ownership has become a propaganda site and that any claims of fact it makes have to be treated with enormous suspicion.  But just in case anyone mistakes the article for a credible piece of polling analysis, I'll just explain that the point it's overlooking is that the seats projection from any poll has to take into account constituency results as well as list results.  That's why Alba are doing better in the seats projection than in the previous Panelbase poll, in spite of the fact that the list vote percentages have barely changed.  The SNP are projected to have fewer constituency seats now, which means unionist parties don't have to be compensated with quite as many list seats.  That leaves more list seats available for smaller parties, and Alba benefit.

That's the point I was making the other week - Alba are running a self-sacrificing campaign by urging an SNP vote in the constituencies.  That will maximise pro-indy representation but it may cost Alba themselves the odd seat here and there.


17 comments:

  1. Thanks to Marcia for the heads up! Yes, I've just read the short Newsnet piece. Needless to say I guessed it was nonsense. I have just this moment finished a BMG poll on the Scottish election and independence, so hopefully we will have more data soon. :)

    ReplyDelete
  2. As ever, James , your analysis is spot on.

    ReplyDelete
  3. Imagine for a moment a world in which the media was truly diverse and representative. I wonder how many people would support independence. 65 percent? 75 percent?

    ReplyDelete
  4. Aye, there are some VI numbers, as per my (yet to pass filter) post on the previous thread. They're in the tables and talked about in the Ashcroft report.

    While the method of assessing VI is a little different to other polls, it seems a valid enough approach, with parties all treated equally, as evidenced by the numbers looking identical to other polls and making the sum total of zero difference to my PoP averages.

    ReplyDelete
    Replies
    1. Whether they 'look' identical or not is irrelevant to whether they're genuinely voting intention numbers or not. You excluded a poll earlier in the campaign on much more dubious grounds.

      Delete
    2. I initially excluded that poll because it used a very different prompt for one single party, making it questionable to compare shares for it with the other parties within the same poll. That was a valid statistical concern; apples vs pears. Turns out the impact wasn't much if any, which we found out soon enough.

      This poll uses the same approach for all parties it seems, so shares should be comparable. All the pollsters use different methods; MORI and Comres take quite different approaches to Yougov etc in terms of assessing how people will vote and how likely they are to do so. There isn't a 'correct' method which everyone must follow; but a range of in-house methods.

      Ashcroft's in-house method is very similar to most polls in weighting, just different on LTV and approach to VI. If it turned out he predicted the result most closely, then maybe his method is the correct one!

      I'm happy to include in my PoP for that reason, particularly as the results don't appear anomalous. Others can exclude it as they see fit for their own reasoning. Makes zero difference as noted.

      Delete
  5. Such a shame about Newsnet. When I was in the process of discarding my “Union” allegiance in 2013/14, really enjoyed Derek Bateman’s podcast. Miss hearing his contribution to Scottish politics now.

    ReplyDelete
  6. I'm 60/40 in favour of an Alba vote, I'd say.

    My one worry is that the UK media only understand SNP majority (quite deliberately and makes no sense). Not pro-indy party majority.

    If snp fail to get a majority that's what will be spun.

    Parliament should be stronger than party but they are deliberately dense these folk.

    ReplyDelete
  7. Newsnet article is shitty just for shiftiness' sake. Very odd to even make the effort.

    ReplyDelete
  8. God, spellchecker! I used the word "shittiness'"

    ReplyDelete
  9. Just visited Newsnet Scotland - what a load of moronic crap it's become

    ReplyDelete
  10. I gave upon Newsnet ages ago. It has become another smear Salmond rag.

    ReplyDelete
  11. Yet another poll indicating the woefulness of Nicola Sturgeon's failure. With circumstances as they have been over the past six years Yes should have been polling at least ten points higher by now. That is the cost of Sturgeon's catastrophic approach to the constitutional issue.

    ReplyDelete
  12. Brexit: The great big British [sic English] wrecking ball.

    https://www.bbc.co.uk/news/uk-northern-ireland-56910045

    Arlene Foster announces resignation as DUP leader and NI first minister

    ReplyDelete
  13. These numbers are exactly what I would expect. Am I correct that by dint of a sample size twice as large as normal, Lord Ashcroft’s polls can be given a commensurate respect? My prognosis, hotly disputed by James, is that all the ‘OOHS’ and ‘AHHS’ of the campaign for independence count for little compared to the demographic drift. Crudely put, most of the old people who die each year are unionists, and most of the 16-year old’s who join the electoral role are not. My calculations are that this is ‘worth’ .5% per year. Four and a half years after we were at 45%, 49% is pretty close to the mark. Technically, we do not need to convert anyone if we just wait. I think there is little to no chance that life in the UK will get better anytime soon. There is no Unionist option for young people that looks attractive. Now this is not to debase the role of a vigorous campaign, but it does look like we have only really added .5% to the yes-count other than at the registry office.

    ReplyDelete
    Replies
    1. Agreed, it literally is a matter of time.

      Delete
    2. The damage the Yoons will do in the meantime though isn't worth thinking about. There needs to be a full on Indy Campaign and the sooner the better. Every single day that passes that we are ruled by the Brit Tories is one more day of damaged lives.

      Delete