Quote (KrWWW @ Jun 18 2016 06:21pm)
my link shows that Ontario is semi conservative province, but you completely ignore it
Were you meaning to quote me? Considering I'd actually agree that Ontario is semi-conservative, I'm not sure lol .
Quote (dxlightning @ Jun 18 2016 06:24pm)
Again, that's the industry assumption, but I have reservations against using it for politics - the industry of stats tends to use sample size as a be all end all, the same sample size equation applies to every industry. I don't buy it.
Most industries are 5-10 years behind on stats when it comes to sampling.
I think the perfectly random sample is the end all be all, only problem is actually achieving that. Yes the equation does change in most industries, but an underlying principle remains. I'm usually skeptical of quantitative evidence, as you said it's often taken to be a tool that's viewed as far more perfect than it actually is, but the (theoretical) math behind sample sizes and the normal curve has me convinced, mostly, that sufficient sample sizes properly surveyed are generally sound.
I do agree that polling is deficient in politics. Polling has never interested me that much mostly because I find it to be an inaccurate tool of making specific analytical judgements. It's OK for making generalizations and big picture observations, but it's not good at telling us how people actually think.