Wikipedia:Reference desk/Archives/Mathematics/2022 August 1

= August 1 =

Stats/Polling Methodology
Hi. At Wikipedia:Opinion Polling On Scottish Independence, we are debating the validity of a poll.

The pollsters appear to have created a 1029 person representative sample of Scotland, and then asked two randomly selected halves of the sample two differently worded questions concerning attitude to Scottish Independence. The Wikipedia page has tables listing results for both of these questions.

One option seem to be to include both of the halves of the polling in the relevant tables. That seems to imply that we are accepting that a randomly selected half of a representative sample is a representative sample. Our other option is probably to leave the poll(s) out entirely.

Do any maths/stats people have a view? If you like you could join in at the talk on that page, where the last section discusses the 'In these Islands' poll. RERTwiki (talk) 12:47, 1 August 2022 (UTC)


 * Courtesy link: Opinion polling on Scottish independence. --Lambiam 12:55, 1 August 2022 (UTC)


 * There are hundreds of polls listed in the article. Is this about the poll from 29–31 Mar 2022, labelled "YouGov/These Islands", listed in the section Polling on a second independence referendum? --Lambiam 13:09, 1 August 2022 (UTC)
 * Essentially, if you look at the pdf linked there you will see they also asked about independence preference. Those results are not in any of the tables pending some resolution on my question. RERTwiki (talk) 16:40, 2 August 2022 (UTC)
 * How these poll numbers are best presented in the article is not a maths question. A sample taken from a population is considered "representative" if the sampling procedure is such that the probability of a member of the population being included in the sample is uniform across the population: everyone has the same probability of being selected as anyone else. (In the practice of polling this is an ideal that is never fully attained.) If a representative sample is drawn from an existing representative sample, the new, smaller sample is also representative, since every member of the original population has the same probability of being included. --Lambiam 17:39, 2 August 2022 (UTC)
 * Thanks. I understand, but the argument was being made that the subsamples were not representative, which I think is a maths question. I will feed this back into our discussion, and it may well help. Thanks again, RERTwiki (talk) 09:14, 3 August 2022 (UTC)