This is what happens when you poll primarily through Twitter Ads.
April 21st - April 27th, n=398 (MoE: +/- 5%)
Re-weighted for gender and geographic region, but still completely trash. Funny enough though, the Atlantic sample seems normal and would be a viable result tomorrow.
Maybe if you added the results from this page's poll to mine it'd even out to something a bit more accurate
We were talking about how this sort of reflects on how people who are stuck in echo chambers can slowly fall down the path of election denialism. If all you see is positive conservative content (on Twitter) or positive Liberal content (Reddit), then when the results don't match what you've immersed yourself in then those types of conspiracy theories can rear their ugly heads.
I'll give a national prediction later tonight, but rest assured this twitter poll does not weigh into my prediction