[oldalum]

 we know that the Trumpustani tend to answer the polls based on how it will be perceived, not necessarily as they really believe. We know this because Trump did much better that the polls predicted in the election. 

My take on the early polling postmortems is that this was not an important factor this time (not even clear it was important in 2016, maybe just a pop explanation); the bigger (and more troubling for the future of polling) problem is that more people, especially Trump supporters, did not answer the phone when pollsters call, and the adjustments made to the pollsters' models based on 2016 did not adequately anticipate or compensate for this. More definitive analysis should be forthcoming in the next few months. See., e.g.:

NY Times polling expert on what went wrong (again)

Also the Nov 10 NY Times "Daily" podcast with this same expert, which I found informative.

[oldalum]

 we know that the Trumpustani tend to answer the polls based on how it will be perceived, not necessarily as they really believe. We know this because Trump did much better that the polls predicted in the election. 

My take on the early polling postmortems is that this was not an important factor this time (not even clear it was important in 2016, maybe just a pop explanation); the bigger (and more troubling for the future of polling) problem is that more people, especially Trump supporters, did not answer the phone when pollsters call, and the adjustments made to the pollsters' models based on 2016 did not adequately anticipate or compensate for this. More definitive analysis should be forthcoming in the next few months. See., e.g.:

NY Times polling expert on what went wrong (again)

Also the Nov 10 NY Times "Daily" podcast with this same expert, which I found informative.