Bad polls

I love polls, I really do. I think they often turn up fascinating results, and a dream job of mine would be to conduct polls all day about all sorts of things. This might be why I hate it when polls are incorrectly used and bandied about. Over the weekend the Washington Post did a poll asking whether people approved or disapproved of Obama's handling of health care. Since his approval rating in this regard was has gone from 57 to 53 to 49% approval from April to June to July, this finding was treated as evidence that, well, Obama and health care reform were in trouble, that the country, the median voter, the people who will boot the Democrats out of Congress in 2010 were turning against the Administration.

But that's not what the poll said. The poll said that 49% of people approved of the way Obama was handling health care (with 44% disapproval). It did not say WHY. Paul Krugman, for example, probably disapproves of the way Obama has been dealing with health care. Does that mean he's going to vote Republican in the midterms? I don't think so. Indeed, the same poll asked who you thought was doing a better job on health care, Obama v. GOP Congressmen -- Obama won hands down, 54 to 34. Funny I didn't see that number reported anywhere.