I've been thinking a lot lately about what's wrong with science in the 21st century. Part of the problem is sloppy thinking that becomes apparent when you realize how many widely believed models are inconsistent with what we know about biology. I assume that similar problems occur in other disciplines.
One wonders if the proliferation of papers with huge numbers of authors is part of the problem. Maybe this fad of "multidisciplinary" science is part of the problem and not part of the solution? Is it possible to be an expert in two or more different disciplines?
I've seen plenty of example of biochemists and molecular biologists who publish papers about evolution without knowing much about evolution. Is this an isolated example?
Speaking of "big science," I was reminded of a paper published by Bruce Alberts back in 1985 in Cell. The title was "Limits to growth: In biology, small science is good science" (Alberts 1985). These days, many people grow up believing that bigger is better. Giant factories that produce Wonder Bread have replaced thousands of corner bakeries, driven by the increased efficiency of scale. There is an unfortunate tendency to extend this view to the biological research community, and I have on occasion heard a major symposium speaker introduced in glowing terms as the coauthor of more than fifty papers per year. While I can admire the energy and management skills required to maintain a very large laboratory, the best biology is rarely done in this way. With a few notable exceptions, the biochemists and molecular biologists I most respect run relatively small laboratories and publish when they have something important to report. As I shall argue here, doing good science is very different from producing bread, and there are compelling reasons why large laboratories are in general less efficient and less interesting than smaller ones. To reflect this fact, I believe that changes in funding patterns and expectations would be useful in the biological sciences.Some "big science" is good. The sequencing of the human genome, and other genomes, for example, was a big science project that benefited the entire biological community. But I'm not sure that significant advances in our understanding of how life works come from big labs. Does anyone have examples? What are the most significant conceptual advances to come out of big labs?
Alberts, B.M. (1985) Limits to growth: In biology, small science is good science. Cell 41:337-338. [PubMed] [doi:10.1016/S0092-8674(85)80001-5]
10 comments :
What I would like to see is a graph of scientific productivity (number of papers weighted by impact factor?) per lab versus grant funding per lab. My impression is that labs with more money produce more papers, but not necessarily in proportion to their funding. The more students and postdocs in a lab, this less time each gets with the professor. So we might get more research per dollar by spreading the wealth more.
I'm less interested in "productivity" per se, than with Alberts described as "good science." A big lab might very well turn out more papers per capita or per dollar.
But my experience (limited) with large labs is that they highly discourage open debate and skepticism. When every member is on the same funding source, there is a strong incentive to ignore concerns that might cause the research project to be re-evaluated. Maybe a few interesting scientific exchanges happen within the group, and are never published. Many more probably never happen, even though they might have done had the same people been funded separately.
I can honestly say that I did not like what I encountered in the large lab setting, although I cannot say that my experiences were not simply the result of one anomalous lab culture. (Anomalous lab culture? Ewwwww...)
In my experience, when people say "big science" what they really mean is "big bio-medical engineering". Take the Human Genome Project for example: that is engineering rather than science. It didn't require 500 people thinking up the best way to do it; it required 500 pairs of hands to do something that was actually pretty trivial to accomplish.
But my experience (limited) with large labs is that they highly discourage open debate and skepticism.
That's my impression too
The problem is that there are things that must be done and can not be done in a small lab setting. Genomic papers do not have so many authors just for the sake of filling the whole page with the authors list and their affiliations.
And we should keep the "big biology vs small biology" and "big labs vs small labs" discussions separate, because with the advancement of technology more and more small labs will be doing "big biology"
a graph of scientific productivity (number of papers weighted by impact factor?) per lab versus grant funding per lab
Such a graph will no doubt show higher output from big labs an average. Money makes money, it's simple as that. David Baltimore has an h-index of 156 (meaning he has authored 156 papers that have been cited at least 156 times; in his case, that's over the past ~ 24 years with 514 papers in all). That kind of thing is absolutely impossible without an army of people working for you and a lot of money to pay to these people and their expensive kits to beat out a competition. (For comparison's sake, Larry Moran has an h-index of 16 over 21 years; both are ISI data).
Which does not necessarily says much about quality of science, BTW. A big name has infinitely higher chance of getting through Nature editors to have a paper reviewed. And a Nature paper, other things being equal, is cited a lot more than one in Biochemistry.
This week in evolution wrote, “What I would like to see is a graph of scientific productivity (number of papers weighted by impact factor?)”
I would drop the impact factor thing as a measure of the quality of the paper. There are many reasons why citations may not correlate to quality. My PhD research made me familiar with a paper that was very highly cited, but all citations were negative. The paper was eventually retracted.
Thank-you for posting this topic, Larry. This is a very important question for the future of science. We seem to be investing a high proportion of scientific resources into a smaller elite group of labs. I am positive this is not the best way to spend the money.
And for the record, I am post-doc'ing in a medium-large lab, which is currently quite well funded. I came from a PhD in a lab with only 1 or 2 PhD graduate students, and poor funding. I'm not convinced that the new lab is "worth" 1000+ times what the old lab was worth.
Personally, I find multidisciplinary co-operation to be refreshing, as it forces you to consider your own views from a different perspective.
For example, since an understanding of evolutionary change ultimately involves understanding its genetic and mechanistic underpinnings, expertise of biochemists and molecular biologists is indispensable.
I don't believe multidisciplinary science leads to sloppy thinking, rather, it exposes it.
I've been thinking a lot lately about what's wrong with science in the 21st century.
If I ever again see a lab head ("PI") spending a regular iota of his/her time doing experiments with his/her own hands in the lab, I will know things are back on the right track.
Post a Comment