Anatomy Of A Knee-Jerk Reaction

Our story begins with an innocent post on Sarah Palin. Everyone who was anyone was talking about Sarah Palin's sudden resignation that day, and little did we know that this one event would set off a chain reaction that would ricochet through the blogosphere for many, many minutes. For all we know the repercussions still echo through the series of tubes and their wooshing noises that we call The Internets.

At Angry Bear, cactus writes that Palin was wise to quit while she was ahead, and if only others had thought to do the same. Cactus mentioned Bush among other presidents, and somehow, somewhy, this dart pierced the armor over the heart of our heroine, Miss (soon to be Mrs.) Megan McArdle. She responds, in her own inimitable fashion:

Question Of The Day
What happens to the cottage industry among Democratic-leaning armchair
economists grinding out analyses proving that Democratic presidents are, like,
totally awesome for the economy? Presuming that we're stuck--as seem very
likely--in at least a couple of years of really grinding low-to-no growth, Obama
is going to destroy their figures. Are we in for a resurgence of belief in
exogenous growth factors?

Dishonest, vapid and spiteful in only two sentences. She's becoming quite succinct. Cactus, for the umpteenth time, says the facts speak for themselves. Matthew Yglesias seconds that post. McArdle responds to Yglesias and cactus by attempting to gum the latter's post to death. McArdle, from the comments:

All of these "studies" have the same problems: the number of data points is too
small, the start points are somewhat arbitrary, and the "D" label is very poorly
specified....Right, but he doesn't use appropriate tests. He uses tests designed
for independent distributions, and doesn't control for sample size.

She does not explain why these problems invalidate the data, or link to a post that does explain. That would be beside the point. The point is that you can never be sure and causation is not correlation and it's too difficult to predict and the plural of anecdote is not data. The point is obfuscation, and at that Megan McArdle is very successful indeed. The rocket scientists at Instapundit and The Corner will link and nod and McArdle will have earned her daily bread. It's a kind of rotten, day-old bread, but hey, it's a living.

UPDATE: Cactus responds to McArdle again, with better humor than I would have in the circumstances.