why retractions aren't really a bad thing

Don't worry about retractions or bad papers being discovered by scientists trying to build on and replicate each other's work. That's exactly how science advances.

Carl Zimmer has a feature on the recent rise of paper retractions in the NYT, citing it as a symptom of what he calls a dysfunctional climate in the scientific world. Much of the dysfunction in question is made up of valid concerns raised many times before, concerns such as the use of grad students as cheap labor instead of a grad student's lab work translating into a scientific career, the dearth of tenure track jobs, and the reduction of the tenure process to meaningless numbers, all of which fuels a cloud of doubt and uncertainty about an alarming number of STEM careers.

By now, pretty much every pundit realizes the need for more STEM majors and even the reliable citadel of STEM-bashing humanities scholars, the Chronicle of Higher Ed, has began to reverse course and promote scientific and engineering fields. But once you have a budding scientist entered in a graduate program, what are this scientists' prospects? That's the dysfunction here, not the retractions. In fact, the retractions are actually a good thing because scientific papers aren't supposed to be set in stone and the scientific process not only allows, but demands papers to be retracted if they're wrong or fraudulent.

Fact is that there are profoundly influential papers which never make it past the preprint stage and there are a number of splashy papers in the highest of the high profile journals not up to snuff, like the GFAJ paper. This is why tenure committees' current laser-like focus on the homogeneity of research and impact factors don't just miss the forest for the trees, but go far beyond it.

Instead of measuring how ideas are received and who's working on extending them, they drone on how many other papers cite a scientist's work even though citations are not the same as additional study and experimentation. If we were to focus on what research really matters and why, and weren't afraid to pull papers when they're found to be wrong or shoddily researched, we'd have a much saner scientific climate in the first place. Rather than look at the spikes in retractions as a symptom of a disease in science, we should be looking at them as the immune response of the methodology and promote it while relieving the actual symptoms through new policies, better tenure processes, smarter funding, and an approach to scientific papers as constant works in progress, not monolithic declarations of results. We don't need ten thousand new papers to be written each year. We just need a few really novel and good ones…