PA: Muhlenberg/MC: Sestak 40/Toomey 48 (10/24-10/27) DO NOT ENTER UNTIL 27 DROPS

[The Vorlon]

Hmmmm, there seems to have been an avalanche of 'interesting' polls in the past day or so.

Everybody and their dog will be releasing a poll in the next ten days.

Colleges, Universities, Newspapers, Zoltan The Magnificent, Pete's All Beef Double Cheeseburger poll, Zogby, The Psychic Hotline....

The list of crap will get very thick and deep....

[The Vorlon]

Hmmmm, there seems to have been an avalanche of 'interesting' polls in the past day or so.

Everybody and there dog will be releasing a poll in the next ten days.

Colleges, Universities, Newspapers, Zoltan The Magnificent, Pete's All Beef Double Cheeseburger poll, Zogby, The Psychic Hotline....

The list of crap will get very thick and deep....

But are Zogby's polls more or less reliable than those by Zoltan The Magnificent or the Psychic Hotline?

Pete's all beef Double Cheeseburger poll actually has a pretty decent record.....

[The Vorlon]

Rather a lot of attention being paid to a University poll with a daily sample of about 100 people.
Tthey have a decent record, but still, a sample of 400 is pretty noisy - not quite investment grade information here.

[The Vorlon]

Sestak is still pretty competetive with Independents: his share actually increased from 35 to 44% in this releases, with Toomey fell from 55 to 48%.

Given that only 7% self identify as "independent" in the way Muhlenburg is doing this, in a sample size of a bit over 400 people, 7% translates into 30 or so interviews.

Drawing any conclusions from a subsample of 30 people is tenuous to say the very least.

Not sure I want to go into the cross tabs of this poll for any nuggets of Gold.

[The Vorlon]

Okay, maybe I won't take them.

If Sestak is still behind with that sample... well... you know... it doesn't look too good.

It's a reason why I prefer "Party id reweighting" and so Rasmussen. The move showed by Muhlenberg doesn't exist inr reality. It's just due to the party id sample.

10/28: R+5
10/31: D +4

Wow - this poll bounces around so much you'd almost think it was some tiny sample poll run by a bunch of University students or something...

Oh wait... that's exactly what it is....

But seriously...

In 2008 actual turnout in Pennsylvania was about 44/37 to the Dem side in terms of party ID. (ie Dems +7)

This poll has it Dems +4, so in terms of turnout, things have declined by only 3% for the Dems versus 2008.

Does this seem reasonable to you?