December 10, 2013
(Larry Downing/Reuters)
(Larry Downing/Reuters)

How’s Barack Obama doing? “Back To Pre-Rollout Levels,” says Talking Points Memo, based on a new Pew poll. Pew reports a 45 percent approval rating; not great, but not bad, and up from last month. But just a few hours ago, Obama was in terrible shape; a new Quinnipiac survey had him at 38 percent, down a point since November (to TPM’s credit, they included that Q number in their story, albeit far down).  But that’ s not all! Today’s Gallup number had Obama at 40 percent, while Rasmussen (for whatever it’s worth) actually has Obama spiking up to 47 percent — almost healthy. And there was also a McClatchy-Marist poll released yesterday showed Obama at 43 percent.

Which only serves as a reminder: if you really want to know what’s going on, use polling aggregators. HuffPollster, for example, has Obama at 41.5 percent approval. That’s down just under two full percentage points since October 1; on the other hand, change the settings to their most sensitive level and it estimates that Obama is a bit lower, at 41.2 percent, and has lost almost three percentage points since October 1.

The point is that there’s really no excuse, years after Nate Silver became famous, for ignoring the plain truth that single polls can be extremely misleading. We know better, and we have better resources; we should use them.

All that said, it’s pretty clear that Obama’s year-long slide has continued, and may have accelerated recently. If it has accelerated, it’s not clear (despite all the certainty that you’ll read by pundits) exactly why. It could be the Obamacare rollout — perhaps the effects of the rollout itself, but more likely the effects of three weeks or so of a press frenzy about the rollout. On the other hand, it’s entirely possible that Obama’s approval was hit more by the government shutdown. That could be more or less a direct effect, with people blaming House Republicans more but unhappy with how everyone acted. It could also be indirect; Gallup’s economic confidence index plunged around the shutdown, and has only recovered about half of its losses; it’s certainly possible that most or all of Obama’s falling approval ratings since September could be entirely caused by economic pessimism.

Which brings up a second point: not only should the press be more careful about attributing too much confidence to any one poll, and should avoid constructing narratives around single polling sequences, but the press should also be far more modest about assigning causes to fluctuations.

Again: we know this stuff. Far too many are still getting it wrong.