Albert Einstein once expressed a crucial scientific truth: “No amount of experimentation can ever prove me right; a single experiment can prove me wrong.” This identifies the foundation of our faith in science: self-correction.
Scientific advances depend not just on one researcher’s breakthrough, but upon the work of others attempting to replicate the research and confirm the results. Confidence comes with replication and with the fact that failure to replicate results invalidates the research.
But there’s an erosion of credibility in scientific research. As The Economist recently pointed out, roughly half of biotechnology research can’t be replicated; the results of only six of 53 supposedly “landmark” cancer studies could be reproduced; some computer scientists believe that more than ¾ of the papers in that field are unverifiable.
Causes vary. Mishandled statistics compromise many studies because researchers don’t understand increasingly complex statistical techniques or lack skill deploying them. The huge growth in the number of researchers, the papers they produce, and the journals in which they publish, all make peer-review more superficial - people simply don’t have time for meticulous assessment of others’ work.
Perhaps the fundamental cause is the explosion in the number of scientists. In 1945, there were only a few hundred thousand research scientists; today over 6 million compete for jobs and grants. Advancement depends upon publishing new findings. Success requires discoveries - the more dramatic the better. But many of those supposed discoveries have flaws which remain undetected because the self-correction mechanism is broken.
This will get worse.
High school science enrollments boom, in Vermont and elsewhere, spurred by college demands that candidate transcripts emphasize so-called STEM courses (that’s Science, Technology, Engineering, and Math). Caught up in college admission competition, students increasingly migrate to so-called “right answer” departments, where grading seems clearer than in supposedly more subjective disciplines. In an increasingly competitive and over-crowded field, there’s less and less possibility for thoughtful self-correction.
Testifying in Congress last March, Bruce Alberts, editor of the distinguished journal Science , argued that young scientists must be schooled to be skeptical of their own results and those of others, and to focus on the quality, not the quantity, of their work.
Recently, Harvard art historian Jennifer Roberts made a strong case that only the humanities provide necessary training in “critical attention, patient investigation, and skepticism about immediate surface appearances” – in other words, the central skills of self-correction.
Roberts is on to something. Science and the humanities aren’t antithetical. To use a scientific term, they’re symbiotic.