10/05/2009 08:33 am ET Updated Dec 06, 2017

Revisiting Obama's health care speech

Back on Sept. 9, I predicted that President Obama's speech to Congress on health care was "not likely to change much in terms of public opinion" based on previous political science research. A few days later, I noted weak and inconsistent evidence of an effect (a claim that was disputed by Nate Silver). University of Wisconsin political scientist Charles Franklin subsequently weighed in, finding that "Opposition [to health care reform] has grown but is now slowed to a near halt" while "[s]upport reversed its decline sometime in August and has begun an upturn" which was "probably driven by the speech."

So how do things look today? Here are estimated trend lines for Obama job approval and support for health care reform:



To maximize the likelihood of seeing an effect, I've restricted the date range to July 1-October 5 and used the most sensitive trend line estimator. Nonetheless, the effect of the speech on Obama's job approval is minimal -- the graph shows a small upward blip after the speech but the series quickly returned to its previous trajectory. There was a small bounce in support for health care reform after the speech, but part of the effect dissipated. Meanwhile, estimated opposition to reform, which dipped in the wake of the speech, quickly rebounded toward previous levels and is now greater than it was before the speech. When Charlie Rangel said before the speech that "this level of involvement from the president could well be a game-changer," I don't think these were the results he had in mind.

I'm emphasizing this point because there's a misperception among journalists that the president can easily move public opinion. As we've seen again and again over the years, it's simply not true, but the lack of followup by the press means that the lesson is never learned. (At most, a failure to move poll numbers is blamed on some specific aspect of president's message or strategy.) So we repeat the same cycle over and over again.

(Cross-posted to