22-01-2024, 11:51 AM
(22-01-2024, 11:43 AM)WillPS Wrote: I like the idea but this might be tricky on election day itself. The reason the exit poll matters is that the sample is so much wider than any normal poll - and that's only possible because it's based on a single question. Is there sufficient resource inside polling in general to do both that and a normal 'smaller sample wider reaching' study?These days, with online panels it's possible to get huge numbers of people answering a poll within a matter of hours. Exit poll has to be done differently, because the methodology is dependent on using the same polling stations at each election, but there's no reason that you couldn't commission a separate on-the-day online poll from another pollster that is then reweighted to account for the projected shares for the exit poll (which they do produce, but don't publish).
It just seems a bit of an oversight for me, particularly considering the BBC used to do something similar and it is typical with exit polls in other countries to include extra info/questions. To be honest, it doesn't really even need to be a bespoke poll - there's so much polling data produced in that final week of the campaign, that it could easily be reanalysed for use on election night programmes.