The poll carried out in March 2017, by the respected ORB International for the Daily Telegraph had 2 000 respondents of whom 175 were based in Scotland. Now, I’ve criticised the validity of polls before on the basis of typical sample sizes (1 000) but these are common and are defended vigorously by people like Professor Curtice as long as the sample construction is scientifically done to produce something representative of the population as a whole. With a UK electorate of around 46 000 000, you can see why I’m uneasy relying on 2 000. However, Prof Curtice assures us that with proper sampling you get pretty accurate results in most cases. He makes Brexit, Trump and the last Tory win exceptional cases. If 2 000 out of 46 million can be sampled to be scientifically valid then presumably a sample of 175 out of 3.9 million will have a reduced bit still similar level of validity. So, if the UK sample is useful then so must the Scottish one be? It was an online poll making it also more likely to be representative. Landline-based telephone polling tends to have a distorted sample with fewer young and poor respondents and face-to-face or even telephone questions tend to inhibit non-status quo or change responses. Both of the former groups are more likely to be pro-independence. See these results and conclusion from Orb:
Orb Daily Telegraph Poll March 2017
‘Q.1 Do you support or oppose Scotland becoming an independent country and leaving the UK?’
Respondents from Scotland region:
Unweighted – 185
Weighted – 175
Support – 58.7%
Oppose – 41.3%
Poll suggests most Scottish voters now back independence
I know, it’s only one wee poll but we take our pleasures where we can, given the inbuilt bias against us in many of the other polls and this was before the spectre of a Tory landslide in England, this June, had emerged.
Yes I agree with your conclusion. Opinion polling leaves me wondering how well the pollsters have adapted to the new age of mobile phones and the internet in all it’s guises plus the variations in voting age and the inclusion or exclusion of EU citizens.
Fingers crossed though for this poll.
2 000 is 0.0043% of 46 000 000 and 175 is 0.0045% of 3.900 000 so which is more credible? Is the choice of 1 to 2000 just a convention?