Ezra Klein writes:
A year ago, when pollsters asked [who people trust to deal with major problems], Obama won 50 to 35. That he’s now tied not just with Republicans but with congressional Republicans two months after they shut down the government shows how much damage the roll-out of the Affordable Care Act has done to perceptions of his administration’s competence.
Since the Post/ABC poll hadn’t asked the “who do you trust” question in a year, it’s very difficult to discern how much of the damage is because of the rollout fiasco. We have some clues, however, and they suggest that the answer is: not much. After all, that 50/35 poll last December also had Obama’s approval at 54 percent — 11 ticks higher than the current 43 percent. It’s likely that the “who do you trust” question doesn’t really get people to compare Obama with Republicans; instead, odds are that most people hear it as just another version of “do you like how Obama is doing?”
Nor did Obama’s numbers drop rapidly from the rollout. In the Post-ABC series, he was down to 47 percent approval in mid-September, before the exchanges opened (well, sort of) for business. Indeed, the December 2012 number looks in retrospect very much like a post-election honeymoon.
It’s certainly possible that the pollsters have found a real reaction to what happened (although it’s not clear whether that reaction would be to the very real rollout disaster or to the press frenzy that more or less overlapped with it), but it’s very difficult to know whether it’s anything more than a temporary blip.
This isn’t to say that Barack Obama is doing well in the polls; the long-term drop is real, and while it’s not enormous, it’s very different than the longer-lasting boosts provided to Bill Clinton and Ronald Reagan at the start of their second terms. Such drops can have real consequences, both for midterm elections and for a president getting his way on other things. But it’s really hard to know why that long-term drop has happened. (Regression to the mean? Continued economic weakness? Problems all year with health care?)
Either way, I’m for extreme caution both in interpreting the sub-questions in these polls, and also in drawing connections between changes in polling levels and any particular events.