Wednesday, April 30, 2014

conservative inference

I went to see a physical therapist this week after finally figuring out that I still have a torn tendon in my elbow after a rollerblading accident last September, and our conversation makes me think that I've become too conservative about statistical inference, based on my economics training, to converse effectively with other humans. He asked me all kinds of questions about what makes it hurt, and I couldn't really commit to any answers because it mostly gets sore after doing things, so I never know exactly what causes it, although there are some clear correlations, but we all know correlation doesn't imply causation. Then he asked me a bunch of questions about whether the cortisone shot, tennis elbow band, or avoiding backhanded maneuvers is helping, and I couldn't answer that either, because all three began at the same time after seeing the orthopedist, so who knows which is making it feel better? Finally he just said "Ok, but it did feel better after you got the shot, right?" "Yes" "Alright then."

I suppose I should have known better - doctors are regularly bombarded by people who have googled their symptoms and are convinced they have crazy diseases and convinced that ridiculous homeopathic cures have had certain effects. They're used to having to scale back what they hear, not the other way around. (This also makes it incredibly frustrating when I do know what's wrong and they don't believe me. Like when I had bacterial bronchitis and went in as soon as I had symptoms and they made me wait another 10 days for antibiotics since the viral variety is much more common, even though I obviously got it from Matt who was responding to antibiotics finally prescribed after two weeks of suffering...)

This also reminds me of the time I was pulled over for speeding in Texas and he claimed I didn't signal when changing lanes, and I said of course I did, so he asked me how sure are you that you did. I said about 99% because I didn't consciously remember pulling the lever at that exact moment (it's just muscle memory, after all) but I'm sure I don't fail to signal more than 1 in 100 lane changes (not signaling a huge pet peeve of mine). Whoops. Most people claim absolute certainty when they're nowhere near 100% confident, so claiming only 99% confidence was interpreted as obvious lying. And then again when someone came around a corner and hit me when I was pulling into the lane, and I told the insurance company I realistically thought the liability split was approximately 80/20... big mistake. You'd think your own insurance company would advocate for your side a little instead of immediately acquiescing as soon as you admit to anything.

From now on I'll consider the equilibrium.