So you think the Abu Ghraib prison torture scandal has broken the back of President Bush's popularity? Well, I did too. But then I did a reality check.
Last February I ran an experiment with the first 37 months of Bush's approval ratings. I found that 90 percent of the variations in the month-to-month change in approval of Bush could be traced to just five months. These were the two months following 9/11, the two months of the Iraq war and the month that Saddam Hussein was captured. Those months were big positives for Bush. But they were his only positives. Apart from them, Bush's approval showed a remarkably stable declining trend, which averaged 1.6 percentage points every month.
Tick, tock -- no matter what Bush said or did, Americans seemed to come to their senses about him at a steady rate. Except, of course, in the presence of a galvanizing foreign event or crisis.
A very big question hung over that finding. As of January, there had never been an average approval rating for Bush below 50 percent. Did there exist a hardcore group or "red zone" of Bush supporters who would stick with him through thick and thin? Would those who actually voted for him in 2000 hang tough? If so, his approval ratings might hit a floor, say, around 48 percent, and hold from there onward -- until the election or the next galvanizing event.
Now we have four more months of data. They show an average decline, across nine polls, of 3.7 points in February, 1.8 points in March, 0.58 points in April and 2.4 points (with eight polls so far) in May. The average monthly decline is 2.1 percentage points. This occurred over the period when the Democrats selected John Kerry, and when the Bush campaign spent some $70 million on advertising to build up Bush and knock down Kerry. The perceptible effects: a fluctuation at best.
The four-month decline is a bit higher on average than Bush's long-run downward trend. But it is not much higher. It is not enough higher to show that anything exceptional has happened. In particular, February's decline is not significantly greater than normal. And May's decline is within the normal range of 0.6 percent, give or take, around the standard minus 1.6 percent -- the 95 percent confidence interval. This suggests that Abu Ghraib has not had any special effect on public opinion. Not yet, anyway.
The figure below shows the evidence to date. Note the vertical line drawn to mark the time of the first experiment. Does the period afterward look different? A little more noisy, a little more variation between polls than is usual, perhaps. But the striking thing is how little the pattern has changed. If Bush's approval has a hardwood floor, we didn't hit it at 48, 47 or 46 percent. Things could still change. They could change if Bush fires Secretary of Defense Donald Rumsfeld -- or even Vice President Cheney -- and remakes his government over the summer. There is talk of a possible al-Qaida attack on the Olympics in Athens. There is talk of terror at the Democratic and Republican National Conventions. And, of course, there is continuing chatter about the capture of Osama bin Laden. Whether any of these will happen, and how they will play, no one knows.
But the numbers so far suggest that the underlying political reality may be rooted in something else: that the public isn't paying that much attention, at least to the daily news -- or to the advertising.
Instead, Americans could be coming to a deeper judgment on Bush -- perhaps about his competence, or trustworthiness, or character. And we could be coming to that judgment as a whole people. It could be that we are not irrevocably divided down the middle between blues and reds. Maybe some of us just take a bit longer than others to think things through.
The statistics could mean something else. We'll keep looking as time passes. But if they mean what they appear to mean, and if the patterns hold up, it's not good news for Bush.