Fact: PPP's sample size is LARGER than Rasmusen, Zogby and Suffolk COMBINED

[thesubstanceofdreams]

Of course they may still be wrong (and I believe they are) but I'm cautiously optimistic that Hillary will win by fewer than 5 points.

http://www.realclearpolitics.com/epolls/2008/president/pa/pennsylvania_democratic_primary-240.html

[DrDan]

1. but it has to be a representative sample - otherwise the results can certainly be flawed

[speedoo]

2. It's hard to argue with PPP's track record.

They have been very accurate in this election cycle. I feel confident the result will be within their MOE.

[NJmaverick]

3. I have my fingers crossed that this will put an end to the destruction

and we can unite and focus on defeating McCain.

[grantcart]

4. They seem to be very good at states with cross over like Wiscosin

where they predicted Obama winning first and later on were spot on.

[PoliticalAmazon]

5. Larger sample size doesn't mean more accurate...

...On a national poll, 1000 to 1400 respondents, IF the sample is constructed to actually reflect the opinions of the group they want to study, is sufficient for a valid study.

In a smaller sample like PA, 600 would be adequate.

More respondents doesn't mean it is more valid.

[Laura PourMeADrink]

6. PPP's median error rate, historically, is 7 points. That would mean

a range of O winning by 12 to Clinton winning by 4.

[RUMMYisFROSTED]

7. Quinnipiac 08/14 - 08/20 524 RV 42 12 Clinton +30.0

Wha' happened?!?

[george_maniakes]

8. what poll is that? for which state and when? n/t

[RUMMYisFROSTED]

9. Follow the link in the OP.