"You don't like me! You really don't like me!"--UPDATED

August 10, 2005


Well, they never call me--Too much is made of polling, or at least of individual polls. There is the common tendency to reject the ones that come to conclusions you don’t like or agree with (”the only poll that counts is the one on election day”) and embrace the ones that show you are definitely among the intelligent elite. And then there is the disbelief that you can make 1,500 telephone calls and tell what 300 million people are thinking. [Your humble servant was actually trained to do exactly that at a program at Northwestern University back in the last Nixon year]. Whether you believe individual polls or not, you sometimes run into a meta-analysis of polls--a study of lots of polls asking similar questions--and you come up with some interesting and believable stuff. Mark Blumenthal and his excellent blog, Mystery Pollster, has done just that. His essay was triggered by an op-ed piece in the New York Times by the Pew’s Andrew Kohut and Peter Hoey, that compares public opinion and major events during the first half-year of the second term for Presidents Nixon, Reagan Clinton and Bush. As Blumenthal points out, the graphic that goes with it is amazing. Essentially, it shows that President Bush’s approval rating is the lowest of any president in a second term with the exception of Richard Nixon who was being swallowed by Watergate. It’s down to 44 percent in the Pew poll. If, however, you go and combine all the respected national polls and plot them on one chart, they all--with one exception--show essentially a similar result:

Bush’s approval spiked naturally at 9/11 and quickly began to slide. That’s expected; no one would think any president could keep that high an approval as we moved away from the attack. It took another uptick with the capture of Saddam Hussein. It rose 5-10 points around the election. But the trend throughout Bush’s presidency has been steady and consistent in all but one of the polls--downward. [The exception is the Rasmussen poll, which uses machine voices for interviews in a tracking poll, and neither Blumenthal nor anyone else can explain why it is an outrider, besides the fact is uses robot interviewers. Blumenthal guesses it is because Rasmussen weights his polls by party which doesn’t catch changes in public’s party affiliation. More people are now associating themselves with Democrats then they did several years ago.)

One possible explanation for the downward slide, suggested by Slate’s blogger, Mickey Kaus, is that the polls are simply reflecting the natural approval rating Bush would have if 9/11 never happened. It’s not the war or the economy. It’s Bush. Think, Kaus says, of a baseball hitter who has a hot streak and then regresses back to his natural batting average. Bush has. What he sees it what he gets.

Works for me.

UPDATE: On 8/14, the Rasumussen robots fell into line, making it unanimous.