Jan 25, 2008

Lies, damn lies, and...

I've been going through some statistics tonight, because I was gorging myself on Dr. Female A-M's developmental lectures, since they're like potato chips, thus putting me WAY too caught up in behavioral science at the expense of statistics.

Which is silly, really, because I think our stat prof is really good, goes through things at a slow pace (too slow for some, I know, but if there is one thing you do NOT want to race through with me, it's math), and overexplains things to a point where I feel pretty confident in the material even after only the lecture, but...

Everything was going perfectly well up to the "Power" lecture, again, not interesting, but relevant, and more importantly, things up to that lecture had been inherently logical. So maybe that line about "lies, damn lies, and statistics" is a little harsh, right?

Nope, it was just a matter of time. Suddenly, "So if you want this value and don't have enough participants, just move this one." "If your sample size isn't big enough, just increase your estimate of what you want the population differences to be." Or how willing you are to make a Type I error.

Hey! You can't do that!

What's sad is that despite a now-revealed tendency for statisticians to make stuff up, I think medical research/stat is still beating immunology and respiratory physiology for "things there would be a chance in hell I would do for a living". The wonders of a good professor.

Now, our stat prof has been really good about relating everything we're doing to published papers in the medical literature so we can see real-life application. Also, every time she says "But you don't have to calculate this number since in real life, you'd just plug it into a computer", I would start some wave action going on if I weren't alone in my apartment. However, all of that paled in favor of the class response, when for linear regression, her chosen example was "relationship of med school GPA to board scores". That shuffling you hear on Sonic Foundry is the sound of 100 students suddenly snapping awake.

1 comment:

enchantra108 said...

Goodness me but something is very wrong here. The statistics are clearly going to your head. Showing a preference for statistics. Ok, fine respiratory physiology is a fate worse than death (or broccoli,I might add); but still, it sounds like you're being taken over by the dark side. Have been seeing anything strange or unusual lately? Have the fluffy bunnies been telling you to do things for them?