Oh great, here comes another polling conspiracy theory, regarding the latest Kantar poll. Long time followers will be able to guess how this ends, but here's the detail. 1/8
reddit.com/r/ukpolitics/c…
reddit.com/r/ukpolitics/c…
Someone on Reddit decided to divide the turnout-plus-standard weighted by the unweighted counts to imply "turnout". In fairness, the OP did say "unless I've massively misread the tables", but of course the conspiracy theorists (the usual suspects) aren't being so cautious. 2/8 
First of all, as several people on the thread pointed out, the maths is wrong. If you want to back out turnouts, you need to divide the compound weights by the nationally representative weights. Otherwise it's distorted by whatever falls out of the raw sample. 3/8
Secondly, one estimate of 2017 turnout is given, as if to be taken as gospel. The problem is that poll estimates of turnout are almost always too high. Conventional poll samples tend to be too politically engaged, and people tend to remember voting when they didn't. 4/8
Both of these biases have a disproportionate effect on lower turnout demographics, which in age terms means young people. The vote-validated (ie checked against the electoral register/"roll") random probability BES – the gold standard – estimated 18-24 turnout as 43% in 2017. 5/8
Kantar's model – which was in fact one of *most* accurate in 2017 (see below) – aims to adjust for these biases while taking into account how likely voters say they are to vote. (Keep in mind too that subsamples, which is what these are, have relatively big margins of error). 6/8
Also, because most polls sample all adults (18+) including those not eligible to vote, we need to divide these turnouts by the estimated eligibility rate, which is of course lower in the (younger) age groups where the non-naturalised, non-Irish/Commonwealth immigrants are. 7/8
So, if we do this exercise properly, we find that, yes, implied turnouts among young people are below what they were last time. But they are lower across ALL age groups besides over 65s (as often happens several weeks out). So there really ISN'T anything dodgy going on here. 8/8
