I think Rasmussen distorts 538's

[Laura PourMeADrink]

Trump's approval #s. I have been watching these #s since day 1...no one except Rasmussen & affiliates and maybe throw in Harris, polls so high for trump. I know Silver adjusts for past accuracy, but what about frequency? Seems Rasmussen has a poll out constantly. Surely their consistent high approval affects Silver's average?

[Loki Liesmith]

1. They are weighted less into the average

[Laura PourMeADrink]

4. But isn't that due to o past inaccuracies?

[Loki Liesmith]

9. Also they don't adopt all Nate's best practices

They all figure into the pollster rating which helps determine the proper weighting.

[paleotn]

2. Always has.....

and I've always wondered why the keep including a group that obviously focuses on over 65's with land lines. The distortion can't be weighted out unless it's factored at 0.

[Laura PourMeADrink]

6. But aren't people with land lines over 65 hating on him too?

[Make7]

3. 538 ranks the pollsters they use. Rasmussen gets a C+.

They adjust the approval numbers based on the rankings. Rasmussen's reported 45% / 53% becomes 40% / 54% after adjustment.

538's overall calculated approval is 40.2% / 55.7% currently.

[Laura PourMeADrink]

5. Yes..but talking frequency of their polls. Even with dings

It's consistently adding high approval #s

[grantcart]

7. No because the weight also declines rapidly by date after a few days they

Have a lot less impact.

The real value of any of these pollsters whether PPP or Rasmussen is not the number but the trend


In other words if both Rasmussen and PPP (and the others) all show Trump consistently trending down over a 10 week period, say 7 % , then they are confirming each other even if one is at 40 and the other is 44.

If Rasmussen shows a decline over time using the same turnout projections then his results are as helpful as the others, in some ways more.

[BGBD]

8. They don't just weigh..

they also adjust, so a Ras poll at 48 goes into the average at like 41.

[Laura PourMeADrink]

10. Right but nevertheless frequent Rasmussen poll #s

Go into the avg. Even if they come in 41.

[BGBD]

11. They do

but even if you took them out, the change to the average would be minor.