[Personal Responsibility]

I still find it hard to understand how Rasmussen justifies using a D+5 model when his own party ID data says R+2 (actually, R+3 if you round up)

http://www.rasmussenreports.com/public_content/archive/mood_of_america_archive/partisan_trends/summary_of_party_affiliation

[MrChips]

Yes, very odd. Do we know for sure that he is using a D+5 sample?