REDOING THE MATH

[Maureen Johnson]

One of our first posts, Do the Math by Scott Westerfeld, was very popular. In fact, it was so popular that people started e-mailing it around . . . and it turned McCain voters into Obama voters!

Well, yesterday, a blogger tried to say that this was all wrong.

You can read the entire exchange here. But below is Scott's full reply, in which he makes his case point by point.

I don't know the man who posted this--I'm sure he is very nice. I'm not posting this because of anything he's done. I'm posting this to make a point about something that's been going on in this election all the time for the last few weeks.

You have to have noticed by now that one party seems to have a bit of a disregard for facts, realities, even science. From the denials about global warming (remember Sarah Palin not wanting to get into a discussion about the causes of that? And Joe Biden saying that if we didn't accept the reason, there was no way we could fix it?) . . . to the outright lies and fear-mongering.

But there is truth out there. There are realities that we have to be willing to fight for, and we have to be ready to defend our positions.

This is Scott defending his position.

A Ridiculously Detailed Response

Hey, thanks for taking the trouble to look at and analyze my post. Obviously, I was simplifying some of things (it's a site for YA readers) but I think my points stand. In fact, looking at the data with your observations in mind, it's even more convincing.

First: Yes, I did mistype 4.326 into 4.236, but my one-year-shifted numbers (45% D advantage) are still right, and make more sense anyway. (It's hard to blame a president for an economic year in which he took office on Jan 20.)

And good catch on the 1994 election results, but we still wind up with a D-control advantage of 3.55 to 3.12. (I figured congressional control would be a wash in any case, for reasons stated below.)

You make a big point of including 1946-47 data, using the post-war crash to drag down all D numbers. Newsflash: where you start does change the data. But that big revelation doesn't pay off in the end. Here's why:

I went through my spreadsheet, deleting years from the top. In other words, I checked all 54 data sets from 1948-2007 to 2000-2007. And the D's still win them all.

The closest results are 1979-2007, with D growth at 3.28 and R growth at 2.74, a mere 20% better for the Dems. Or you could do 1983-2007, the ONLY starting year where the R's crawl over 3%, at 3.03! But D's are back to 3.72 by then, so it's still a 22% win for the D's. And R's fall swiftly back to 2.65 by 1985.

So here's the bottom line: No matter what year you begin from 1946 to 2000, the D's win.

It's one thing to wave your hands and say, "It all depends on what data you select." But when you check 54 data sets and the D's win them all, the hand-waving looks a bit more frantic.

So your last resort is to claim that party control is unconnected to economic cycles. A bit forced, when all your data sets give one answer to varying degrees, but possible. But it seems sad only to remember that first day in stat class, when the amusing correlations are on the blackboard, and then forget all the *other* days in stats class, when you looked for answers to the questions the stats raised.

Like asking, "Why DO the Republicans suck so hard?" Rather than, "Oh, there's SOME way to explain the last 60 years, like, um, tech bubble!"

But even statistical nihilism is useful to me. Most of the R's I know say, "The economy always grows better under R presidents. It's a fact!" What I was mostly doing was giving ammunition to kids whose parents and friends spout similar shibboleths.

In any case, it's always nice to know that my side has the statistical evidence, while the other side has strategies for blowing smoke at evidence.

As for your sports examples, those are ludicrous not because *all* statistical correlations are lame, but because sports teams don't make economic policy. The president *does* make economic policy, so those data *are* worth looking at.

Of course, you do become pro-stats in your congressional analysis. But the words "congress controls the purse strings" are fifth-grade Civics-class hooey. In the modern era, presidents have dominated the spending process. Reagan passed his tax cuts and budget increases with near-unanimous R support, peeling a few House D's away to get a majority (see Bowevils). Clinton famously triangulated, working with centrist R's and D's and leaving out the extremes. (After the 1995-6 shutdown, Gingrich was vaporware.) And the Bush II era, with R congressional control and truly staggering deficits, removes any residual ambiguity. Republicans borrow more.

I also note that you avoided my last point: R control of both the presidency and one or more houses of congress for any six-year stretch leads to an economic meltdown. It's happened every time they've been given that much power in the last century. Maybe you could find some lemming migration stat to correlate those crashes to, but lemmings don't deregulate markets. Republicans do, and proudly so. And the effects can't be hand-waved away.

(By the way, I checked and I still get ".212" on D borrowing. Perhaps you're still adding in 1946-47 without saying so? In any case, see if you can find any starting year after which the D presidents actually borrow more than the R's. My guess is, thanks to Bush II, you won't be able to.)

[Ellen Hopkins]

Woot! Audrey... you rock...

[Audrey]

do NOT argue with the scottmaster....

jk jk that's actually a really interesting article. I'll have you know that I was watching over my dad's shoulder as he filled out his ballot last night, making sure he voted for Obama and against Propositions 4 and 8.

[Zara]

Very nice.

[Tamora Pierce]

Ummm, lemmings don't do the migration stuff. It's a myth.

For the rest? You make me proud to have taken statistics, even though I flunked the second section of the course twice. (Dyscalculia, unheard of in 1974-76.)