1. A thousand people is enough if they are well chosen. A million people is not enough if their choice is not random. It's hard to know which is the case here

1. A thousand people is enough if they are well chosen. A million people is not enough if their choice is not random. It's hard to know which is the case here, but the general presumption should probably be that there is significant bias. They describe their methodology a bit here: https://uchicagopolitics.opalstacked.com/uploads/homepage/Disinfo-Poll-Methodology.pdf https://uchicagopolitics.opalstacked.com/uploads/homepage/Di... 2. The wording of that question seems to be "It may be necessary at some point soon for citizens to take up arms against the government." (https://uchicagopolitics.opalstacked.com/uploads/homepage/IOP-Poll-Topline.pdf https://uchicagopolitics.opalstacked.com/uploads/homepage/IO...) I'd call that softer, but not way softer. It would definitely alarm me if it was true. 3,4,5: Yes, it would be useful to know the answers here. They describe some of this in the Methodology page I linked in 1, but more detail would be helpful. Probably the best use of this poll result is to suggest ways for someone else to perform a better poll. But I did find it to be pleasantly well sourced for a newspaper with a poor reputation for accuracy.