Message to all the people who will say "it's tightening!": These differences are not statistically significant.

https://twitter.com/MorningConsult/status/1299735650783178752

PS- that's without using a design effect, which would add to the error. ALL pollsters should be reporting MOE with design effect to account for weighting.

The correct interpretation here is that there is no substantial movement.

cc @NateSilver538 @Nate_Cohn @gelliottmorris

I really wish so-called data people would check numbers before tweeting out a narrative of a bump that is NOT STATISTICALLY SIGNIFICANT to a gazillion followers.

Also I would note that opt-in panels like Morning Consult, particularly those using marketplace sample, tend to have substantial design effects, so in reality this is likely nowhere near significant.

Irresponsible all around.