Why do a Ph.D., given that the pay is low and academic jobs (and other jobs requiring a Ph.D.) are very hard to get? Well, maybe because it represents several years of secure employment (often with health benefits), doing something you enjoy doing, that has some non-zero chance of improving your future career prospects. Put that way, it’s easy to see why many people might consider a Ph.D. a better option than the alternatives. Nice piece at The Atlantic. It’s written about humanities Ph.D.’s, but applies to most Ph.D.’s, I think. Echoes my own thoughts.
Andrew Gelman’s latest comments on researcher degrees of freedom (i.e. data-dependent analytical decisions that compromise statistical validity) and the frequency of false discoveries in the scientific literature. I liked his comments at the end about how scientists are generally reasonable people and science as a whole is pretty sensible. So even if the typical scientific paper is much more likely to be wrong than is generally recognized, science as whole may still be, as Andrew says, “lurching toward the truth in some way”. Which I don’t think is an excuse to stop worrying about researcher degrees of freedom (like Andrew, I find myself worrying about them more and more). After all, we presumably want science to lurch towards the truth as quickly as possible, so it’s worth scrutinizing practices that may slow progress even if they don’t stop progress entirely.