Why don't the networks pool their resources and have a 5000 sample poll?

[HeldsBelds]

Such a poll would have a margin of error of 1% or so.

They did it in the past for exil polling with VNS.

[LiberalFighter]

1. And what happened to VNS?

[A Simple Game]

2. Uhh, they got it right?

[LiberalFighter]

7. Their projections were messed up

[A Simple Game]

8. But the exit polls were right on.

In fact the exit polls were so accurate they discarded them. Seems they didn't support * winning.

[tritsofme]

10. It was disbanded

after 2000.

[Roland99]

3. They have to call about 10x as many people as the sample...

due to hangups, not home, etc.


Imagine calling 50,000 people?

[Mountainman]

4. On Countdown tonight Keith said that the pollsters can barely get 30%

of those they call to participate in a 3 day poll and less than 20% participation in a one day poll.

[Bill McBlueState]

5. lots of work, little payoff

Would you do five times the work to go from 3% to 1%? That's probably the issue.

[pabloseb]

6. Very simple

If the MoE is large, they can blame the MoE when reality doesn't fit their results. If MoE is just 1% and they're off by 3-5%, they have nowhere to blame but themselves.

[Technowitch]

9. Wouldn't work, really, unless they completely agreed on polling methods