If I move to a more highly ranked institution will I do better science?
Or, if I move to a more lowly ranked institution will the quality of my work decline?
Some scientists are obsessed with "moving up", thinking that being at the "best" place is essential. They cannot fathom that one could do outstanding work at a mediocre institution.
However, consider the following. People at a high status university may get Nobel Prizes but that is not necessarily where they actually did the prize-winning work. Here are a few examples.
John Van Vleck: Wisconsin to Harvard
Joe Taylor: U. Mass to Princeton
Tony Leggett: Sussex to Urbana
William Lipscomb: Minnesota to Harvard
Can anyone think of other examples?
So can one actually measure how career moves affect the quality of science? One recent attempt is
Career on the Move: Geography, Stratification, and Scientific Impact
Pierre Deville, Dashun Wang, Roberta Sinatra, Chaoming Song, Vincent Blondel & Albert-László Barabási
The authors give an exhaustive analysis of the authors, affiliations, and citations of more than 400,000 papers from Physical Review journals, concluding
while going from elite to lower-rank institutions on average associates with modest decrease in scientific performance, transitioning into elite institutions does not result in subsequent performance gain.This made it into an article in the Economist magazine, entitled Why climb the greasy pole?
It is worth looking at the figure that this conclusion is based on, noting the size of the error bars.
I think for people that have a drive to climb up the greasy pole, the drive is not to be more productive, or do better science, but the drive is to be more famous.
ReplyDeleteA Harvard professor is more famous, just because he is at Harvard, than a professor in Sussex.
This is part of what's wrong in science now: while fame is useful to get better productivity (via more money, meaning more opportunities), a lot of people strive first of all for fame, and not first of all for good quality work.
And this is related to the publication pressure; writing 3 PRLs is good, but I like following each of those up with an "8 page PRB" where you can actually nail down all the details.
That does not happen often anymore as the drive is for the next (too short, hence less easily reproduced) PRL.