I remember some solid refutation on why the USC/LAT poll was biased since it was a tracking poll with a biased sample and gave a couple of voters on niche demog

I remember some solid refutation on why the USC/LAT poll was biased since it was a tracking poll with a biased sample and gave a couple of voters on niche demographics too much weight, etc etc. Turns out they were pretty much right which makes us question all the social science accepted as valid.