i'm not saying that the online "polling" is wrong, what i'm pointing out is that if you're not sampling then you're not polling. it could turn out that one of the online "focus groups" (and that is what they are doing - not polling) is the most accurate in the pile, but it would just be by sheer luck.
so, you can tell me i'm not doing any math, but i keep telling you that your data is useless - and the theory says that it has no predictive value, whatsoever.
there was one other actual, real poll released by mainstreet today and while the headlines say one thing, the data says another:
that's about consistent with what i pulled out of the leger poll regarding the ndp, although it has the liberals coming in a lot lower. the only way to make sense of that is to conclude that the liberals are deeply restless on their right flank, right now. the "another party" line is mostly the bc conservatives - but the undecided is going to be either the furthest right or the furthest left. the swing vote in the centre in canada is very, very small - a truth that holds at every level, in every region.
according to the media, the bc conservatives don't exist. but, they're a major factor in the liberals trying to hold their base, which would be considered conservatives almost anywhere else, in canada.
if they're not afraid of horgan (because he's not very left-wing), or think that wilkinson is basically the same as the rest (and he is.) then it opens up huge levels of uncertainty as to how that coalition holds itself together. that is more likely to determine the outcome than anything else - including restless environmentalists trying to make a balanced choice.
my position from the start is that we have almost no data to base any meaningful projections on. the "online polling" is not math, and not useful in making predictions. we had one poll, and i pulled what i could from it.
this poll backs up the basic analysis, but opens up some greater levels of uncertainty and some wider error bars.
i wish i had more data, but there isn't much, and you shouldn't be surprised if the outcome is a surprise.