Rasmussen out with Ohio

[Whacker77]

Rasmussen released ne Ohio numbers.  He has Bush at 49 and Kerry at 47.  What happened here.  The numbers Pollwatch99 posted from the premium section didn't show this.  Does this a difference come from leaners?

[Beefalow and the Consumer]

Rasmussen released ne Ohio numbers.  He has Bush at 49 and Kerry at 47.  What happened here.  The numbers Pollwatch99 posted from the premium section didn't show this.  Does this a difference come from leaners?

With the Rasmusbots, it's really hard to say.

"Awwww, it's a friendly robot!

...this time."

[The Vorlon]

Rasmussen's Ohio, um, er, ah... "tracking poll" interviews a heart stopping 80 (that would be eighty, as in 4 score) or so likely voters per night, as part of a 7 day rolling sample.

And while I am not 100% sure, I think part of these interviews continue to be extracted from his main sample of his national poll.

A sample size of 80 has a MOE of +/- 11% per day, and because two samples are involved in the change each day (one rolls off, on rolls on) the average daily noise level on this poll is aboutr 1.6% per day over the 7 day roll.

The Ohio Microbot poll has Bush running in the +2 to +5 range, with Florida running maybe +3 to +4 as well.

Rasmussen says Bush is up a tad in both Florida and Ohio - beyond that you can't say too much.

80 interviews a day - there is a lot of noise

[Beefalow and the Consumer]

If he's using robots, can't he buy a few more dialers and phone lines and afford to make more than 80 calls per night per state?

What's the point of using bots if you can't use them to get a much bigger sample size than mere human interviewers can do?

[The Vorlon]

If he's using robots, can't he buy a few more dialers and phone lines and afford to make more than 80 calls per night per state?

What's the point of using bots if you can't use them to get a much bigger sample size than mere human interviewers can do?

Rasmussens state polls:

1) Tiny samples
2) Weighted all to hell
3) Extracted from a larger sample designed for a different purpose
4) Unproven track record
5) Unproven methodology

Who knows

[Whacker77]

Didn't know all that.  Thanks!

[lidaker]

Vorlon, I think Kerry's EV number on your map should be 216, not 226.

[The Vorlon]

Vorlon, I think Kerry's EV number on your map should be 216, not 226.

You are correct !

I had it correct, then somebody told me it was wrong, so I "fixed" it with the wrong numbers.

Bush is at 264 as well, not 254.

When I update my map of Saturday I will use a calculator and get the numbers (if not the states) correct

[Pollwatch99]

Vorlon, I think Kerry's EV number on your map should be 216, not 226.

You are correct !

I had it correct, then somebody told me it was wrong, so I "fixed" it with the wrong numbers.

Bush is at 264 as well, not 254.

When I update my map of Saturday I will use a calculator and get the numbers (if not the states) correct

I'm the guilty one.  Send you a note, and I typed the correction in wrong.   Modified it later but you may not have seen that

[ATFFL]

80 interviews a day - there is a lot of noise

80 per day does not lead to mere noise, it leads to a rock concert.

Rasmussen needs to rethink his state polling methods.