EnricoPallazzo said:
The problem is those polls as far as I know do not capture the hidden conservative votes. This is why he won in 2016 and this is why we had brexit. We also might have other impacts such as poor people not going to vote despite being captured in the polls. Considering this and what we saw in 2016, a 52x48 result to me seems more like a 50x50. Also lets not forget what happened in Michigan in 2016. Definitely Biden advantage now is higher than what Hilarry had but I dont think it is so much higher. |
The final polling average had Clinton at 48.5% of the popular vote, and Trump at 44.9. The actual results were 48.5 to 46.4 . So they were off by 1.5% which is within the margin of error.
A key difference this time around is that Biden is at 51.4% of the popular vote with Trump at 42.1% (fivethirtyeight average). If you're claiming that it's closer to 50-50, you're actually proposing that the polls are WAY less accurate than they were in 2016. They would have to be undercounting conservatives by 7.1% (more than 4x the amount they undercounted them in 2016). And, they'd have to be overcounting Biden voters by 1.4% which is possible, but they were bang on in 2016 with that. Overally, we're talking about polls being off on total by 8.6%. So, the polling data would have to be getting much worse, whereas intuition would tell you they'd have learned something from 2016. Even if they're undercounting Trump supporters, if 51.4% is accurate for Biden, a Trump victory is very unlikely.
Also, Fivethirtyeight's popular vote on that page is not strictly based on polling data but it also incorporates trends. The current polling data has Biden at 49.6 and Trump at 43.7. So, their model is already predicting that the race will be closer than the current polling indicates.
Of course, this is with the caveat that things can change between now and November and blah blah blah.