but, the methodology is sound, and the results are consistent - if a little pro-republican. investor's business daily? but, that's useful variance if it's mild and the methodology checks out. it does.
so, we have four polling firms now that are worth aggregating.
the polling story for the week is that stein has seen a measurable bump. let's see if it holds into next week.
--
actually, i missed the mclatchy poll, which is also good. this is an outlier right now - clinton+14? she's not up much. it's more that trump is barely pushing 30.
they claim it's a straight sample. that's interesting to me, because i'd have expected the opposite phenomenon - that weighting would be helping clinton. weighting is another valid debate. i don't think that we can get away with not weighting at all, which is to say the data there is probably a little too raw, but i'd think that the general trend right now is to fuck with the data too much. so, again: that's probably wrong, but it's useful. and, as mentioned, it's surprising. what it suggests is that the other polling firms may be exaggerating trump's lead with white people.
so, we've got five polls, now. that's good enough, i think.