[finnman69]

Great analysis and commentary.

FYI, I have read NRO commentary claiming sources are talking good internals, not bad.

[LS]

My point on "internals" is, if it's from polling, they wouldn't know if they are "good" or "bad." The numbers are simply against them. Again, the math is astounding: for a decent sample, you need 7,000 calls minimum in ONE NIGHT. I seriously doubt ANY polling firm is doing that for most House districts.

BTW, jmaroneps37, who is in the business, said that their research on Rob Simmons in CT showed that the larger the sample, the better the R did consistently. Most of these House samples are 500 peeps.