Speaker 4
This is the bb s. This podcast is supported by advertising outside the u. K. Bbs sounds music radio pod casts.
Speaker 1
Welcome to think with pinker. I am professor stephen pinker, and i hope to get you thinking about thinking, and perhaps to help you think better. To day. Thinking is on everyone's mind. We live in an era of breath taking scientific discovery, and also an era of fake news, quack cures, conspiracy theories and post truth rhetoric. How can we make sense of making sense and its opposite? Today's episode is, don't expect a zebra. After the advice given to medical students, when you hear hoofbeats behind you, don't expect to see a zebra. In other words, don't be too quick to diagnose an exotic disease just because it fits the patient's symptoms. Start with the basse rates, like the fact that horses are common where you work and zebra's rare. We'll be thinking about how to adjust your beliefs in light of the relevant evidence. It's called basian reasoning, after a famous little formula set down theeighteenth century by the reverend thomas bays. Let's treat our degree of credence in a hypothesis, such as that a patient has a particular disease, as a probability a number between zero and one. A rational believer should condition her credence by the strength of the evidence or data, such as a patient's symptoms or test results. Reverend bays tells us that the probability of a hypothesis, given the data, should be estimated by combining three numbers. Start with the priors. Your credence in the hypothesis, before you've even looked at the data, such as the patient's symptoms. For a disease, the prior might be the base rate of the disease in the population. That's where the horses and zebras come in. All things being equal, it's likelier that someone has indigestion than eosin phylic interopathy. Now multiply that prior by the likelihood if the hypothesis were true, what are the chances you'd be seeing the data you're seeing? It's not unreasonable to think that a patient has rocky mountain spotted fever. If he's just returned from the rocky mountains and has spots and fever finally, divide by the commonness of the data. How probable it is across the board among all people, healthy and sick. If a symptom of aa disease is common, like fatigue, that big, fat denominator will shrink your credence that the patient has the disease. If the symptom is rare, like blue urine, that small denominator should increase your credence. In other words, bays rule just says that you should believe a hypothesis to the degree that it was creditable to start with, that it's consistent with the evidence at hand, and that the evidence is unco across the board. Neglecting base rates, guessing zebras rather than horses isn't just an occupational hazard of medical residence and their hypochondriac patients. It's a widespread human affliction, namely, reasoning from stereotypes, or, as cognitive scientists call it, the representativeness huristic. Consider amelia, a college student described by her friends as impractical and sensitive. She speaks french and italian fluently and wrote a sonnet for her boy friend as a birthday present. What do you think is her major? Psychology or art history? Well, it seems like a no brainer, art history, of course. But wait isn't it a tency bit relevant that there are a hundred and 50 psychology majors for every art history major? Basrule says she's likelier to be a psyche major. Sonnets be damned. If this lopsidedness never entered your mind, you are a victim of base rate neglect. Joining me to day to discuss zebras and horses in evaluating ideas. Ar tolithia williams, professor of mathematics at harvey mud college in clermont, california, and the author of power in numbers, the rebel women of mathematics. And sidartha mukerje, professor of medicine at columbia university and the author of the polet prize winning the emperor of all maladies, a biography of cancer. Welcome to lythia and sidartha.
Speaker 3
Thanks you for having us, professor
Speaker 1
mukje, sid have you ever been involved in a case where a doctor mistook a horse for a zebra? I had a patient who came in with fatigue,
Speaker 3
weakness and weight loss, elderly man in his sixties. And of course, me being an ocologist, my first thought was, you know, what kind of cancer is this? Why is this man losing weight and having fatigue without any other problems. I asked him many questions about his exposures, his history, et cetera, et cetera. And finally we thought a little bit about the possibility that he had v and i said to him, well, do you use drugs? Have you ad unprotected sex? Have you had any exposures that would be irrelevant? He said, no, no, no, no, no. So i said, ok, well, let's go through the first series of tests, you know, the typical stuff, cancer, arimmune diseases, et cetera, et cetera. I was leaving the hospital that evening, an i saw the same guy talking to some one that i knew was a drug dealer.