Clinical and Statistical Prediction, a controversial book by clinical psychologist Paul Meehl, was released in 1954. Meehl argues that statistical approaches for diagnosing patients and selecting therapies surpass human physicians in this book. Meehl went on to win awards for his work, but his theories had almost little practical influence. While later study backs up Meehl’s findings, he received severe opposition from his peers. Others just refused to accept that a statistical approach could do better than them.
This is perhaps unsurprising. Clinicians may be offended by the implication that what they previously thought of as an art form of diagnosis or therapy may be reduced to a set of rules. They may be afraid of being fired, or they may believe their knowledge is being called into doubt. All of these factors would have been expected to pale in contrast to the concern for the patients’ well-being, but given what we know about human psychology, it’s not surprising that they didn’t.
More intriguingly, recent research suggests that non-experts have a predisposition to distrust algorithms in situations when their professional credentials aren’t on the line. Why do ordinary people prefer their own conclusions to those of a more advanced algorithm?
Continue reading on the next page
Sharing is caring!