re: #275 sagehen

But MCAT and LSAT scores aren’t everything, just like SAT scores aren’t. Above a certain threshhold, they’re all more than smart enough to learn the material, to correctly diagnose and know the right treatment, with the manual dexterity to do the work and the personality to be able to make patients feel comfortable.

Every highly selective college or grad school has 2 to 3 times as many “ideal” applicants as there’s slots for; at some point, the admissions office also has to consider that patients and clients would like a range of options when seeking professional services, that maybe they’re more likely to speak freely to someone from a similar background.

It’s ridiculous to be offended that people ranked 98.76 got affirmative actioned over people ranked 98.83, when any applicant ranked over 96 is superb.

Here we go again with the dramatic, and wrong, number games. Last time around, it was a claim that there must be something wrong with a test that 90% of whites, and 10% of blacks, pass. But where is such a test to be found? If one did pop up, it would be quite an anomaly, and quite rightly, people would be looking for the error in construction or scoring.

This time, it’s two claims. First, that the differences we’re talking about are a matter of 98.76 percentile vs. 98.83 percentile, where in fact they’re larger than that, and second, that above the 96th percentile, everybody’s pretty much equivalent.

The second claim flies in the face of the fact that many a doctor gets a diagnosis wrong that a better doctor gets right. Doctors are not demigods, and the bottom ten percent are not really much use to their patients. The bottom half, as a whole, are good at their jobs and deserving of society’s respect, but when faced with a toughie, they’re more likely to miss the call than those up in the top 10%.

It actually matters who you get. It matters who gets in.

As to the first claim, the test score gap tends to run at about 1 standard deviation. As the proportion of admittees to medical school does not perfectly track proportions in the overall population, the test score gap among the population in question, those admitted to medical school, is certainly less. For the sake of argument, let’s split the difference and say it’s 1/2 of a standard deviation. That would mean that the counterpart to the guy at the 98.83rd percentile stood not at the 98.76th percentile, but at the 93rd percentile. Bright. Quite bright. But—-not likely to find his/her way into the top 90% of doctors, at least, not the top 90% as far as perceptiveness in diagnosis, say, is concerned. And, on the far side of your own 96th percentile suggested cutoff.