Jonathan Chait had a nice item up yesterday about Karl Rove's strategy for getting people to hate Barack Obama's jobs plan by cleverly reframing the popular components of it: for example, hiring teachers and cops becomes rewarding union cronies, or building roads (popular) becomes stimulus (unpopular).
But what Chait (and perhaps Rove) don't really consider is whether all of this is doing anyone any good. And it sure seems to me that the answer is: no. It's hard to get good, consistent polling results, but Gallup had the jobs bill at 45/32 in mid-September when it was new (and presumably before the GOP counterspin had really started), compared with 30/22 in a WSJ/NBC poll a month later. Similar ratio, with the undecideds perhaps an artifact of question wording. As far as the effect of all this on Barack Obama, he's been in the low 40s this week per Gallup, which is just about where he was before his speech to Congress.
And of course part of the reason that this kind of clever wordplay doesn't work is because both sides are playing it. Obama isn't rescuing the states, and certainly not bailing anyone out; he's hiring cops and teachers. "Stimulus" has disappeared from the White House vocabulary -- easier to avoid a demonized word than to explain to everyone that the 2009 stimulus did what it was supposed to do.
The other reason it doesn't do all that much is that the audience is small; partisans are prepared to support opinion leaders on their own side, for the most part, even if the spin they're given is implausible, and true independents are a small group that mostly isn't paying attention. Of course, there's more to it than that (since otherwise every issue would wind up with the same polling split reflecting nothing but the underlying partisan breakdown), but the more issues get engaged, the more likely they'll end up close to it.
Now, you certainly can produce great polling results on any issue by manipulating a question to get the most favorable result. But it's not at all clear that underlying opinions are changing at all. When it comes to elections, it's a good enough result if you can reframe things to get that surface shift in voting. But for public opinion between elections? It's just real easy to overplay the importance of this kind of shift.