This mentee took on an especially tough project... testing the impact of a policy measure. In most cases these students do the sort of thing students do for their papers in intro statistics or econometrics classes: they look at the relationship between some broad social or economic variables and pontificate on what that means for society, policy, etc. etc. This mentee had a very specific program that she wanted to get a treatment effect for. It was great to have her interested in that, but as anyone who does this sort of thing knows - it's a tough slog. Her results (both in the summer and in the expanded analysis she finished this spring) were abysmally insignificant. It's a fact of life. It happens. You have to stay agnostic when that happens - often you have to stay agnostic on a conclusion you really would have liked to have found. She was disappointed at first, but ended up taking it in stride after about the tenth time I insisted to her that failing to reject the null is an important scientific finding.
I recently stumbled across this discussion by Neil deGrasse Tyson about the measure of a scientist - you really tell a scientist by the way they react to what a more naive observer would characterize as "failure" (but really isn't).
Just like in everyday life, we learn the most in science when we learn the least. I have grown to love non-significant results after 5 years of them--it even shocks me now when I find what I hypothesized!
ReplyDelete