The Alibi — Summary

Paul Meehl published the finding in 1954 and kept seeing patients for another forty years. Compare a clinician’s judgment against a simple statistical rule, on the same information, predicting the same outcome, and the rule wins or ties almost every time. He was a practicing psychotherapist who went looking for evidence that his own judgment was worth something, found the opposite, published it, and went back to work on Monday.

By 1996 the count was 136 studies, the rule equal or better in roughly nineteen out of twenty, and Dawes showed that a linear model with randomly assigned weights also beats the clinician. One finding is rarely quoted. Told what the rule predicts and permitted to adjust, clinicians adjust, and the adjusted prediction is worse than the rule alone. Machine proposes and human disposes was tested early and did not hold. What the human contributes there is not correction. It is variance.

Almost nothing changed, and what the record holds is the sequence of reasons. The criteria were inadequate, answered over a decade across many domains. The settings were artificial, and field studies followed with clinicians given everything including the interview, and more information sometimes made the prediction worse. Then the broken leg: a table cannot know the man broke his leg this morning. Correct in principle, and answered in practice, because a clinician allowed an exception finds that most cases feel exceptional. Then these patients are different from the ones in the studies. A fifth arrived late: the rule is opaque, or encodes the biases of its data. That one has teeth, and it is the first objection a clinician could raise without any evidence about clinicians, which is useful in an argument you are losing.

An objection that is real names the evidence that would satisfy it. An alibi produces a new objection when the evidence arrives. Three things still run the other way, and an account that omits them is contempt with citations: the metric can be wrong in a way the professional tracks and the study does not, the broken leg is unanswered rather than answered on average, and institutions asked professionals to surrender the judgment and keep the liability, which a reasonable person would refuse.

The income theory is not sufficient, because most of these people were not going to lose their jobs. A clinician’s self-concept is not that they hold a job. It is that they can see something, and the threat is that the thing they believe makes them who they are turns out to be a slight negative coefficient. A person defending an income can be bought out. A person defending a self cannot, which explains what the income theory does not: resistance was strongest among the best practitioners. Yagn’s framing, from the anthropology rather than the psychology, is that a profession is a claim about what a certain kind of person can perceive. Take the tasks and the guild adapts. Take the monopoly on seeing and there is no guild left.

I wonder whether any discipline has ever identified its own alibi from the inside, or whether the recognition only arrives from people who were never in the room.

The test applies here. This corpus claims capability commoditizes and deployment does not, because deployment is tacit local knowledge held by particular people in particular places, which has the shape of the clinician’s objection. It has been amended under challenge more than once, and it is held by two authors with identities built on it and one instrument with reasons of its own to say humans remain necessary. The strongest counter-case was put inside this corpus by one of its own authors and has not been answered. The claim does specify evidence that would move it: a pattern library transferring across firms in one industry, priced as a product rather than staffed as a service. Nobody has named the condition that would count as failure.

Meehl saw patients into his eighties. The closest he came was a remark that a psychologist could hold two things at once: that the actuarial method should decide, and that a person in a room with another person is doing something the decision does not describe. That may be an evasion, and it may be the only honest position available to somebody who proved the thing and still had to work.