Conversation

Replying to

The same people who were polled the first time, should be polled the second time. It’s hard to draw a correlation if that is not the case.

Replying to

Opinion polls that measure political change across time very often don't ask the same people

Replying to

and

Except it’s much more likely that this is driven by the selection into your follower pool since these numbers diverge greatly from national random polls.

Replying to

and

Do they? I just looked at one and it was unclear

Replying to

and

nytimes.com/interactive/20

Replying to

and

Paywalled

Replying to

and

Or... Walled, it wants me to log in? Is it free?

Replying to

and

Oops. Didn’t realize NYT didn’t give free access.

Replying to

and

Two weeks isn't 3 years tho, do you have one that's longer?

Replying to

and

It appears to be monotonically increasing in nearly every demo, so you can just work backward by the column on the right.

Replying to

and

I mean that's a suggestion that might be right but you can't definitively conclude based on just assuming the pattern continues.

7:51 PM · Jan 19, 2021

Replying to

and

Lots of opinions change strongly in multiple dimensions over the course of 3 years, taking a 2 week sample during a particularly heated time is a terrible thing to extrapolate from

Show replies