Gallup's job approval numbers defy punditry that 'conventions don't matter'

President Obama's job approval, according to Gallup, from mid-August until the present. President Obama's lead over Mitt Romney has ebbed in both the Gallup daily tracking polls and the Rasmussen tracking polls over the past five days, as most of our readers almost certainly noticed.

Predictably, this has led to a flurry of conventional wisdom tweets and observations about how "the conventions don't matter."

The problem with that rapidly hardening CW is that it is almost certainly untrue.

Consider the graphic at the top of this article. The vertical line tagged with "Sep '12" marks an excellent dividing line for understanding the impact of the conventions. Sept. 1 lay in-between the GOP confab down in Tampa and the Democratic National Convention in Charlotte.

And in those two squiggly lines we can find three separate critical points of analysis that seem to speak to far more than ephemeral "bounce" in the numbers.

• Notice all the movement prior to Sept. 1. More appropriately, not the relative lack of consistent movement prior to Sept. 1. This is immensely telling. President Obama's approval numbers for the bulk of August were middling, at best, and pretty damned lousy, at worst. But notice that the lines immediately to the left of the August/September dividing line show no dramatic dip, and an immediate recovery. Translation: Any Republican attempts to define (or, more appropriately, ill-define) Barack Obama at the RNC failed miserably. His numbers not only held, but we see a clear movement towards approval parity immediately in the wake of Tampa. Did voters suddenly like Obama more, considering the alternative that had just been presented to them? It's hard to say, but what is easy to say is that three days of Chicken Little-esque rhetoric about how badly Obama had failed the nation did not translate to a change in voter perceptions about their president. Which would also explain why the RNC failed to produce a head-to-head bounce for Mitt Romney, to boot.
• But then, look what happens to the immediate right of that August/September dividing line. Obama's job approval ratings immediately lurch upward, first to parity (a place where they had not been in the previous month) and then into net positive territory. Within a few days of the close of the convention, Obama's job approval numbers were in a place that they had not been since Osama bin Laden was killed. Only a partisan cheerleader could dispute that the Democratic National Convention in Charlotte did a markedly better job of restoring positive sentiments about Obama than the RNC did at boosting Romney (or diminishing Obama).
• However, the most important statistic might well be the final one. President Obama has had net positive job approval for the last 12 days in a row. That has not happened since June of 2011, when the rally effect of the bin Laden death was still in full bloom. Indeed, then his job approval was net positive for over a month. Throughout this campaign cycle, however, it has not happened. And, given the fact that Gallup is a three-day tracking poll for job approval (as opposed to seven days for the head-to-head trial heats), it would be very difficult for critics to dismiss Gallup's numbers as a product of the DNC, given that it ended 10 days ago.

Taken as a whole, the picture painted by the tracking poll graphic at the top of this piece tell the tale of two very different conventions. No one has ever been elected president via decent September job approval numbers. But, some reelection bids have been thrown away in September. And the early signs are that the RNC failed to sink Obama, and the DNC may well have resurrected his image. The improvement may well be marginal, but when you are trying to lock down a narrow lead, a marginal improvement could be enough.