Tuesday, March 1, 2016

j reacts to the common polling error that correlation implies causality [over race]

please remember, as these results come in and the media tries to create a race war in the democratic base, that the following truth is true: obama was and remains the more conservative of the two candidates from 2008, and the results in many places were a reflection of this fact - at least as much as his skin colour was.

bernie's a liberal. he's not a socialist. but, he's at least a liberal. and, in my lifetime, i've never seen a liberal take a real shot at the presidency.

so, it's a different dynamic than we've seen since...what? dukakis? see, the earliest i can really remember in any seriousness is bill clinton's second term...

so, you've got a legit liberal and a centrist. last time, it was a clear conservative and a centrist. and, so you're seeing the spectrum shift to the right, as you'd expect - if you were analyzing results based on meaningful things, rather than convenient categories.

i mean, it's easy to ask people what colour they are or what their gender is, and then chart a graph. you may even pull out a correlation. it's a lot harder to establish any meaningful causality.

if you gathered data on shoe size or hair length or eye colour, you may very well see correlations develop, too. what does that mean?

don't get me wrong: i'm not exactly faulting anybody. people ask these questions. that's the data that exists. if exit polls asked questions about favourite soda or preferred ice cream flavour, you'd have that data instead, and you'd measure charts and try to figure things out.

based on heavy support from sprite drinkers...

and, i'm not denying that there's some value in the data as proxies, either. black southern democrats tend to be conservative, for example. so, the data bunches the way i'm saying, by measuring it the way you're claiming is important, right?

it's just that there's an actual election about ideas going on right now. it's a little unusual. it'll probably go back to normal next cycle, too, even.

but, for now, it would behoove everybody to focus a little more on ideology and a little less on identity.

the results, when they come in, will uphold this. i'll help you out a little by explaining them tomorrow or the next day.