A bit of a "breaking" story; but in short;
from my tweet thread
You know what you can't find easily in most BC polls? how people voted in 2013. Why? Cause they know these numbers are fishy.
If they showed you, they'd say 40% voted NDP in 2013 (they did) but only 34% voted Liberal. (it was 44%)
What this tells me is a very very simple fact; a LOT of people who voted Liberal last time are not answering polls
This was the key stat that told me the early Ontario and Quebec polls were wrong; but not the later (and correct) ones
And this is the key stat that showed me after the fact that both Alberta and BC were wrong in 2012/2013
It's been purposely hidden it seems because the BC pollsters likely know their numbers are way way off. Libs have ~10 additional pts
From a chat room
I just did some math
that shows the BC polls are ******* ********
this happened in 2013. and in alberta in 2012 with pc. I noticed it after the fact
it also happened in early polls with the liberals in ontario and quebec
it also happened in early polls with the liberals in ontario and quebec
Q: what are the amounts for other parties?
well as shown above, a poll that says the topline is 34L-40N-14G-10C should actually read 45L-41N-8G-6C
same with 36L-40N-14G-8C should read 45L-41N-8G-5C
Q: why?
WHY they don't answer? no ******* clue. but at least I've found a way to determine WHO is not answering.
TLDR: BC Polls are just as far off as they were last time, if not more so.
More to come in future posts.
Very good work, have you been able to correlate response rates to things like party/leader approval ratings?
ReplyDeleteNope. No clue; in general my focus has been on how to take a current poll and using numbers within it; tell that it is wrong in advance of the election.
Delete