stewartb
The ease of asking questions of these AI tools is ‘interesting’!
My question: ‘How much difference does it make to the results of political opinion polling in Scotland if the polling company does or does not apply a 2014 weighting to its survey findings?’
ChatGPT said:
‘Applying or not applying a 2014 independence referendum weighting to political opinion polls in Scotland can significantly affect the reported results, especially when measuring support for Scottish independence.’
Here’s a breakdown of what this weighting does and why it matWhat is the 2014 weighting?
This weighting adjusts the results of a poll to reflect how respondents say they voted in the 2014 independence referendum (Yes vs. No). Since the actual result of that vote is known (55% No, 45% Yes), pollsters use it as a benchmark to correct for sampling biasWhy use 2014 weighting?
Mitigating recall or enthusiasm bias: People who support independence may be more motivated to respond to political polls. Over time, some may also misremember or reinterpret how they voted.
Correcting sample bias: Without weighting, your sample might have too many pro-independence or pro-union respondents relative to the actual population.
Consistency over time: It allows pollsters to track long-term changes more reliably.Impact on polling results
- Support for independence (Yes vs. No): Polls that do not apply 2014 weighting often show higher support for Yes.
- Applying 2014 weighting usually reduces the Yes percentage by a few points—sometimes by 2–4 percentage points or more.
- Voting intention (e.g., Holyrood elections) – The effect is less dramatic but still present, especially for pro-independence parties (e.g. SNP, Greens). Weighting can slightly reduce support for those parties and increase support for pro-union parties like Labour or the Conservatives.
Example: Suppose a raw poll (unweighted by 2014 vote) shows:
52% Yes
48% No
After applying 2014 weighting, the same sample might show:
49% Yes
51% No
The specific shift depends on the sample and how skewed it is in terms of 2014 recall.
Why is this controversial?
Some argue that recall of 2014 vote may be inaccurate or that people’s identity has changed over the last decade, so weighting by that vote might distort current opinion. Others say it’s still a useful corrective tool for ensuring polls don’t unintentionally overstate a side.
Summary
Polling Approach Effect on Independence Support
With 2014 weighting Slightly lower Yes %
Without 2014 weighting Slightly higher Yes %
It’s a methodological choice that can shift the headline figures, especially on constitutional questions. When comparing polls, it’s important to check whether 2014 weighting was used.
‘Would you like a comparison of specific polling firms and whether they use this weighting? ‘
Talking-up Scotland 'Albatraz' 
Would you like a comparison of specific polling firms and whether they use this weighting?
Yes please Stewart.
LikeLiked by 5 people
‘Yes please Stewart.’
Not to be unhelpful, BUT it’s so easy to do this – and indeed get an ‘answer’ to any other question – for yourself. And for free.
Go here to the Chat GPT/Open AI search facility and try it out: https://chatgpt.com
One possible wording of a suitable question to ask is: ‘Which polling companies use a 2014 weighting when reporting results of their political opinion surveys conducted in Scotland and which do not? ‘
But I encourage you: to construct your own question/s on this and other topics; to try out different online free AI search tools and compare and contrast what you get back; challenge responses by asking ‘tighter’ follow up questions, including asking ‘why’ – and have endless free fun!
And then ponder what to do with – and how much to rely on – what you get back. But do have fun!
LikeLiked by 1 person
The demographic has changed over the 10 + years since the referendum. With an average of 50,000* deaths per year 500,000 older voters have left the building and been replaced by as many younger voters. Older voters were typically NO voters. The younger generations are typically YES voters. I think weighting by 2014 recall is indefensible.
*I can’t recall where the figure for annual deaths comes from.
LikeLiked by 2 people
Aye mind you fact is, the richest live longest and they most likely are more inclined to be unionists at least the ones around us are, hinging on and they still believe the Britnat propagnda via TV and daily rags. Sigh.
Let’s hope younger voters do turn out in huge numbers next May!
LikeLiked by 1 person
so, what you’re saying is, we should hold back major improvements to SNHS until after independence?
only joking – sort of.
John Lawson
LikeLike
weighting by 2014 recall is indefensible. – I agree, particularly if the weighting is not made clear IN LARGE PRINT.
John Lawson
LikeLiked by 1 person
Here’s a post I made to Richard Murphy’s blog late in March 2025 on the impact of weighting in polls concerning Scotland’s independence:
I’d like to draw attention to the widespread weighting distortions in polling results relating to Scottish Independence and to Holyrood matters. This surfaced in the pro-Independence blogsite Talking-up Scotland yesterday – Latest Holyrood poll adjusted to minimise support for pro-independence parties – Talking-up Scotland – which pointed out that most of the English-based polling companies are still using weightings relating to the 2014 Independence Referendum and yet, in the interim, c500k older voters (who in the main were No voters) have dropped out of the electorate and been replaced by a similar number of young voters the majority of whom are Yes voters.
A Survation poll between 6th & 13th March asked for voting intentions if an independence referendum were to be held. Once Don’t Knows are removed the Weighted scores were 46.1% for Yes and 53.9% for No, but the Unweighted scores were 60.8% for Yes and 39.2% for No. That’s a massive Yes majority of 21.6% A similar poll by Norstat on 11th February produced a similar large Yes majority of 14% (57% Yes and 43% No).
We all get it that polls aren’t precise indicators but it’s clear that the use of inappropriate and outdated data for weighting skews the outcome significantly. In addition LINO’s adoption of Tory economic policies and demonstrable ineptitude in office has changed political thinking in Scotland. It’s also clear that support for Reform is significantly lower here than in England – memories of Farage taking refuge in an Edinburgh pub (where I’m sure he wasn’t welcomed) from a jeering mob and having to be rescued by police – are still prevalent here. Given all that it’s likely that support for independence is running above 10% higher than staying in the UK.
Apologies for saying it again, but it’s hard to imagine us making a bigger mess of running our own affairs and Robin McAlpine’s 2018 book ‘How to Start a new Country’ lays out what needs to be done. We’ve got a government which puts Scotland’s people first, but is hamstrung economically by Westminster. What are we waiting for?
Ken Mathieson
LikeLiked by 2 people
Pollsters have been censored and fined many times. Gerrymandering. Breaking data protection Laws. Using Twitter and Facebook. To illegally buy people details and send illegal propaganda. Cambridge Analytica. Fined but no one went to jail.
LikeLike
Annual death figures can be looked up. Support for Independence is increasing. Migrants want to rejoin the EU. Scotland did not vote for Brexit.
LikeLike