Do numbers matter?
OK, what numbers are we talking about? GG is considering academic numbers. What makes the best professor? If you read an obituary for a local academic, you likely will see the phrase “author of XXX papers and YY books…” stand in for something meaningful about their academic contribution.
As scientists we are always eager to have a quantitative measure to use. For many years, research productivity was measured in numbers of papers, with papers to Nature and Science being multiplied by some factor. But of course we all knew that a lot of these papers might be shingled, or might be tripe. You might adjust for being first author, or having a student as first author and so on. So along comes citation indices, and now the numbers that folks like to tout are citations–total citations, citations per paper, h indices, g indices and the like. But of course there is danger here too: papers can be widely cited for being wrong, or a paper just happily serves as a surrogate for a broader literature. And there are self-citation issues. Toss into the mix the variation in churn between different disciplines and citations can be a hard measure to use. Probably better than numbers of papers, but still imperfect.
Same thing goes on the teaching side. Is it better to teach lots more students in big classes? Financially for the university, yes. But does this provide the best learning experience? No. Similarly, does having a lot of grad students result in better experiences than only 1 or two? Maybe; it really depends on the situation.
So GG would like to argue that there is precisely one number than matters, and that number is zero.
Is it possible to have one scientific paper that justifies a professor’s research program? Sure. Take Einstein’s general relativity; that by itself so shook up physics that other contributions were icing on the cake. Or to look in geoscience, G.K.Gilbert’s work on Lake Bonneville was exceptional; it took years of work to get all that together, and the insights that emerged from that work were substantial [of course, Gilbert was not a professor, though he had similar burdens away from this research]. You can imagine a situation where there is one important paper that creates or redefines a field, but the h-index of that author will be stuck at 1.
Is it possible to only have one student and so make a larger impact? GG would say yes, that he has seen some professors who don’t have many students turn out some really good students that arguably outshine many out of some other programs. Not terribly likely but possible.
This is NOT to say that you cannot do really well by publishing a lot or having a lot of students. Certainly there are examples there too (Einstein, for instance, was no slouch in publishing papers). The point is that contributions can come in nearly any number, except one: zero.
Zero is really failing. No papers is no contribution for a research scientist. No citations over a sufficiently long period is saying that work had no real impact [this is a bit different in engineering, where a paper is more often used than built upon]. No students is nobody to continue to advance a field.
So GG advocates counting from zero to one and then looking to see just what else is going on. Yes, numbers can help to inform that judgement, but they cannot be the sole or even dominant criteria if the goal is more than cluttering up journals or enriching the university.