(12-08-2020 02:43 PM)Redwingtom Wrote:  It's only an anomaly if you assume that the voters who voted in both 2016 and 2020 all voted the exact same way...and that's ridiculous and you know that.

Not really

Quote:I know the polls aren't fully to be relied on at all, but they showed that trump lost voters from 2016. Not to mention the higher turnout can most rightly be attributed to those who wanted trump out of office. And that's further illustrated in the fact the the GOP did well down ballot.

Unless you look at the large number of 'top of ballot only' votes... which again were a statistical anomaly.

Quote:Your problem is the same one team trump has had in court when they bring in experts. They testify to something that looks interesting or odd, but they fail to show that it proved anything.

Except that I've admitted that the statistical analysis alone doesn't prove anything.... meanwhile you've made wild ass assertions with no PROOF at all of them, and when challenged, your PREMISE is the PROOF. I.e... You believe Trump lost votes because the polls tell you that... but you ignore that the polls are based on a small number of actual responses and then numerous assumptions about voter behavior... AND you ignore that after being so wrong that they missed their prediction on the last election... so they adjusted their biases... and then STILL missed the outcome quite badly. By that I mean they had Biden +12 and he won by +3 in some states. Had they kept the last elections models, they would have had Biden +20, and he STILL would have only won by +3.

Trump once again outperformed the individual state polls by a WIDE margin, so why are you so comfortable using that demonstrably flawed polling data to imply that he 'lost' votes?

2020 Trump gets 5.9mm votes. Biden gets 5.3mm votes. 2016, Trump gets 4.7mm votes and Hillary gets 3.9mm. Interestingly, Trump got 52% of the vote both times. Biden outperformed Hillary by almost 3.5%. Meanwhile we look at the polls, who on average (according to 538) had Trump winning 48.6 to 47.4. They got Biden's percentage almost right... and missed Trumps by the same 3.5 points... so even with the upward adjustments from 2016, they under-estimated Trump.

and YOUR conclusion is that Trump did worse, yet somehow got exactly the same percentage as last time... because the polls tell you that??

Statistical analysis as I said IMPLIES things, but PROVES absolutely nothing... but without an allegation, it is almost impossible to do the sort of investigation that COULD prove something. That's a big part of what statistics is used for. Computers do it constantly... and they would balk at this.