(from Pollkatz via Washington Monthly)

What's striking to me, here, is that post-9/11, and after every big "positive" event thereafter, there's approximately the same trend: the downward slope is roughly reproduced after 9/11, the beginning of Gulf War II, the capture of Saddam Hussein, and the 2004 presidential election.

That weirds me out- as does the fact that the trend is approximately linear. What the heck? That's not natural in any way. What mechanism is responsible for a linear trend in public opinion? These things should be roughly exponential (like a lot of epidemiological phenomena)...

I won't complain about the results, though.