Att OIs mätningar börjar närma sig varandra, och i detta fall ev uppjustera SD, har antagligen lite med herding att göra. ex:
As I mentioned before, an 800-person poll has a standard error of 3.5 percentage points because of sampling error alone. A related calculation is the average error introduced by sampling. As a rule of thumb, the average error is equal to about 80 percent of the standard error — or in this case, about 2.8 percentage points.
This is the theoretical limit on how accurate polls can be. Even if pollsters knew, for instance, that David Perdue would win by exactly 7.9 percentage points in Georgia (as he did), they’d still miss this result by 2.8 points on average when collecting 800-person samples.
In fact, however, the new polls deviated from the polling average by less than 2 points by the end of the campaign. How did that happen? To be clear, I’m not accusing any pollsters of faking results. But some of them were probably “putting their thumbs on the scale,” manipulating assumptions in their polls such that they more closely matched the consensus.11
In some cases, the pollsters’ intentions may have been earnest enough. Perhaps they ran a poll in Iowa and it came back Ernst +7. That can’t be right, they’d say to themselves. No one else has the race like that. So they’d dig into their crosstabs and find something “wrong.” Ahh — that’s the problem, not enough responses from Ames and Iowa City.12 Let’s apply some geographic weights. That comes out to … Ernst +3? We can live with that.
Hela artikeln:
https://fivethirtyeight.com/features/heres-proof-some-pollsters-are-putting-a-thumb-on-the-scale/