Two polls on the motion to censure Bush show very different results. Yesterday an ARG poll found Americans favor censure for the President by a 46-44% margin, i.e., a statistical tie. Today, Rasmussen found Americans oppose censure by 38-45%. Seems like a big swing for a single day, eh? Perhaps one of the polls is screwy? After all, ARG and Rasmussen use different methodologies.
There are number of plausible reasons for the apparently steep decline in support for censure. But one of the most obvious is that the polls used two different questions. Guess which question registered the higher number in favor of censure:
A. “Do you favor or oppose the United States Senate passing a resolution censuring President George W. Bush for authorizing wiretaps of Americans within the United States without obtaining court orders?”
B. “Senator Russ Feingold has introduced a measure to censure, or publicly reprimand, President Bush for authorizing the NSA wiretapping program. Should President Bush be censured for authorizing the NSA wiretapping program?”
If you guessed “A,” you’re right.
I don’t know a great deal about polling methodologies, but I did spend some time as a public-opinion research consultant (if I hadn’t decided to return to graduate school after my summer of focus groups, branding, and polling data I would probably be doing something in that field today). My hunch is that (A) better gauges public opinion on censure, since it provides crucial information about the grounds for censure. (B) is more obscure–“NSA wiretapping program”–and junks up the issue by placing an emphasis on the Senator championing the motion. But polls are snapshots. The more important lesson is that framing does matter in politics. It clearly makes a difference how these issues get presented to the public: public opinion isn’t entirely plastic, but it can be molded and shaped by relatively simple ways of presenting issues.