stewartb:

I asked the question of Chat GPT (for free) as a place to start – other AI tools are available.

From Chat GPT – (see https://chatgpt.com )

My question: ‘When a polling company reports a survey of political opinion in Scotland weighted to 2014 what does it mean?‘

ChatGPT said in response: ‘When a polling company reports a survey of political opinion in Scotland “weighted to 2014,” it usually refers to weighting the data based on how people voted in the 2014 Scottish independence referendum.

I’n Detail: In Scotland, “2014” almost always refers to the independence referendum, held on 18 September 2014, when voters were asked: “Should Scotland be an independent country?” The result was 55% No and 45% Yes.

‘So what does “weighted to 2014” mean? It means the polling company adjusts its results so that the sample of respondents reflects the actual breakdown of Yes and No voters in the 2014 referendum. (my emphasis)

‘For example: If too many “Yes” voters respond to a survey compared to their actual proportion in the population, the polling company may down-weight those responses.

‘If too few “No” voters respond, their responses may be up-weighted to better reflect the 2014 outcome (55% No, 45% Yes).

‘Why is this done?

‘To reduce sampling bias: Some groups are more or less likely to respond to polls. Using past vote weighting helps correct that imbalance.

‘To ensure comparability: Weighting to a fixed, known result (like 2014) lets pollsters track shifts over time more accurately.

‘Potential issues:
‘Voter behavior changes over time: The population and political attitudes have shifted since 2014. Some people who voted No may now support independence, and vice versa.

‘New voters weren’t part of 2014: Young people who’ve become eligible since 2014 weren’t in the original referendum and may not fit the old weighting structure well.

‘Bottom Line:
‘When a poll in Scotland is “weighted to 2014,” it means the data is adjusted to match the Yes/No vote breakdown from the 2014 independence referendum—a common practice in Scottish political polling.’

When I asked the same question of Chat GPT but without any mention of Scotland, the response also included this:

‘Possible concerns:

‘Outdated weighting: If the weighting is still based on 2014 data in 2025, it might not accurately reflect the current political or demographic landscape. People’s voting habits and population structures change over time, so relying on old data could skew results.’