I follow politics the way...

...most guys follow sports. (Regular readers of this blog will also be interested to know that the earth is round.) 

I remember watching some pre-game football show with my son once when he turned to me and said, "These guys sure like to talk about football!" I smiled and said, "Yeah, it's just like the political shows I watch. And, just as you couldn't make money betting on football based on the observations of these former coaches and players, so couldn't you predict the political future based on the musings of the chattering class." And it's true; remember how everyone thought the 2008 presidential race would be between Hillary Clinton and Rudy Giuliani?

That doesn't stop me, though, from watching every political show I can.

But I do listen with a critical ear. And some politicos are worth listening to more than others. Take Matthew Dowd, for instance, shown above with Donna Brazile and George Stephanopoulos. Although Dowd worked in the Bush White House (and had a famous falling out with the 43rd president), he impresses me as a particularly balanced analyst. And, even though he is often critical of President Obama (J'accuse!), I like his take on the current landscape.

Sometimes, however, Dowd's observations get me thinking. (Uh-oh!) For example, I've heard him mention this particular statistic more than once:

Since polling began, no president has won reelection with a Gallup approval rating under 47 percent going into Election Day, and no president with an approval rating over 50 percent has lost.

And I always think, Yeah, but how many elections are we talking about? Is it really a statistically significant sampling?

And then, this morning, I read on Nate Silver's blog in the New York Times (my emphasis):

These qualitative factors are important because a sample size of 16 elections since World War II is insufficient to provide for persuasive statistical evidence as to which economic indicators (from among dozens or hundreds of credible candidates) are really best for predicting election outcomes in the long-run.

And I think that's important to remember: we are talking about a very small number of elections.

Here's another tidbit I often hear:

Since the 1976 presidential election, every incumbent president, excluding Ronald Reagan, lost their bid for reelection when the unemployment rate exceeded seven percent.

So what? All that tells me is that Reagan was the exception in a statistically insignificant sampling. (Okay, that's the last time I'll use that phrase.)

But watch: If President Obama wins reelection this year (as I think he will), analysts will come out of the woodwork with breathless observations about how he's the first president to get reelected with such-and-such a statistic.

To which I say: Don't take any of this analysis too seriously until we've had about a thousand or so elections under our belts.