Well actually there is a second side to the data. Rasmussen Reports use a 'likely voter' model for their polls, rather than including all adults. This means two things. Firstly, in general, this results in a slight conservative bias. I read a university study once which concluded that Rasmussen typically has a ~3.9% Republican bias. Similarly in New Zealand; likely voter polls follow the same trend. Secondly, it means you need to stop saying that the majority dislike Obama. Given American voter turnout is dismal, nearly half of the country doesn't care or is indifferent.
Subtracting 3.9% from the result of the healthcare poll puts it at less than 50% favouring repeal (...and it was already within the margin of error, so you're premature to say most people favour repeal). Not using a likely voter model would most likely knock this down further.
Bookmarks