Stats vs. scouts, polls vs. pundits, and ecology vs. natural history

Attention conservation notice: In case you find this tl;dr, here’s the bottom line: I consider lots of different reasons for worrying about the status or role of natural history within ecology. I conclude that there’s little reason to worry.

Attention conservation notice #2: I just discovered a sort of Rorschach test that will let you figure out how you feel about this post without having to read it! Just read this tweet aloud. If you find yourself emphasizing the third word from the end of the quoted passage, you’ll love this post. If you find yourself emphasizing the last two words in the quoted passage, you’ll hate this post. 🙂

***************************

You’ve probably seen the movie Moneyball, or at least heard of it. Brad Pitt plays Billy Beane, the general manager who turns the underdog Oakland A’s baseball team into a winner by using newfangled statistical methods to make decisions. In many cases against the advice of his own scouts, experienced and traditional men who prefer to evaluate players and predict their future performance by carefully observing their play. If you’re a baseball fan, you may well have read the book on which the film was based. And if you’re a statistically-inclined baseball fan like me, you may also have read Bill James‘ old Baseball Abstracts, a pioneering effort to apply statistical approaches to explain and predict the performance of baseball players and teams.

If you follow US Presidential elections, you’re probably familiar with Nate Silver. He’s a former economist and baseball statistician who turned his attention to predicting election outcomes. His statistical models, based on polling data plus a bit of economic data, have predicted the outcome of the past two Presidential elections quite well, and much better than many professional political pundits with years of experience observing politics. His success has pundits afraid for their jobs. Silver went on to write a very good book on the general problem of making good predictions.

In baseball, the whole “stats vs. scouts” battle is now old news. Almost every major league team now employs several statisticians as well as numerous scouts (footnote 1), with everyone working closely together. For all the statistical information we now have about baseball, we don’t–and probably can’t–quantify everything relevant. So that, if you want to understand the past, and predict and control the future, you can often do better by combining information and ideas from statisticians and scouts. Further, the statisticians who analyze baseball are avid, knowledegable baseball fans pretty much without exception, and they all love just watching games (the oft-made joke that they’d prefer the games to be simulated on computers has no basis in reality). So with rare exceptions, it’s only people who don’t know what they’re talking about, and who have no real power or influence, who are still trying to fight the “stats vs. scouts” battle.

Similarly, when it comes to predicting election outcomes, “polls vs. pundits” is yesterday’s news. We have an increasingly good handle of the circumstances under which one can make very accurate, precise forecasts of election outcomes based just on polling data, and when it helps to also draw on other sources of information, including the qualitative impressions of experienced observers (e.g, Cook Political Report). It’s only those who don’t know what they’re talking about, or who have vested interests of various sorts, who deny this (fn 2).

But judging by “various random things I’ve seen on the internet in the last little while” (fn 3), ecology vs. natural history is still a live issue for numerous people (Lindenmayer & Likens, E. O. Wilson, Terry McGlynn, Jim Bouldin, Mark Bertness, Paul Dayton, Tom Webb, Alex Bond, others…). They express themselves differently and raise different issues (and indeed probably disagree on a lot). But it does seem like there’s a shared undercurrent of worry that ecologists as a group aren’t as grounded in natural history as they used to be, and that this has various negative consequences for both natural history and ecology. (At least, that’s my reading, sincere apologies if I’m putting words into anyone’s mouth here…)

Which puzzles me, because from my own perspective as an admittedly-crap natural historian, I see myself surrounded by colleagues who know lots of natural history! Indeed, I’ve had to work very hard to justify the sort of work I do, precisely because it’s not grounded in natural history. So have others whose work is only loosely anchored in natural history (e.g., R. A. Fisher, Simon Levin, Ben Bolker). As ant ecologist Mike Kaspari noted in a comment on an old post, and as statistical ecologist Ben Bolker wrote in a wonderful piece, different ecologists are drawn to ecology for very different and very personal reasons. I’m troubled that for some (far from all, but some), worries about the status of natural history-based work seem to go hand in hand with quite negative views about other sorts of work, particularly quantitative work like mathematical modeling and meta-analysis.

We’ve done several posts defending the value of various approaches to ecology, both singly and in combination (e.g., this and this). So rather than go over that ground again (which Robert MacArthur covered 50 years ago), I’m going to examine worries about the status of natural history itself, and about the status of ecology that’s strongly grounded in natural history. I’m trying to make a good faith effort to understand where the folks who hold those worries are coming from, while also subjecting their worries to a bit of a data-based reality check where that’s possible.

For purposes of this post, I’m going to define ecology operationally as “the sort of work that gets published in ecology journals”, and natural history as “close observation of specific types of organisms in their environments”. Obviously, one could quibble endlessly about definitions, but I think these are close enough for government work.

I suspect I’m not going to change many minds. But even if I don’t change many minds, I hope I’ll at least clarify why worries about the status of natural history mostly don’t resonate with me and folks like me. Maybe that will help the discussion move forward a bit, or else will identify issues on which no further progress is possible.

Below I’ve listed various concerns about the status of natural history and its relationship to ecology. I freely admit I have no idea how widely any of these concerns are felt. Probably some are felt by only a few people, while others are felt more widely. But the key thing is just that I’m focusing on concrete, specific concerns. Hopefully that cuts the issue down to size and allows some progress to be made. They’re not in any particular order–sorry about that. I tried to think of a sensible ordering, but failed.

• Natural history isn’t valued anymore. Ok, this is really too broad and vague a worry to be addressable (valued why? valued by whom? etc.) But for what it’s worth, if you search Google on “natural history” (without the quotes), you get about 872,000,000 hits. For comparison, if you do the same for “ecology” (again without the quotes), you get 28,800,000. Ok, this is kind of a silly measure of “value”–but I don’t think it’s totally silly. I mean, 872 million hits is a lot! 🙂 A very different and perhaps less silly measure of value is that the American Society of Naturalists has a major annual award that goes to an ecologist or evolutionary biologist whose work is firmly grounded in natural history. They have no equivalent award for, say, work grounded in mathematics.
• Ecology has grown at the expense of natural history. Again, rather too broad and vague to productively discuss (see below for discussion of narrower and more specific versions of this worry). But for what it’s worth, according to Google Ngrams, mentions of natural history (and botany) in books peaked in the first half of the 19th century and declined well before there were any mentions of ecology. Mentions of natural history are flat or even slightly increasing since 1900, during which time mentions of ecology took off. So at least by this one admittedly-limited measure, it’s simply not true that ecology has grown at the expense of natural history.
• People are publishing more ecology these days, instead of publishing natural history. The ecological literature, and citations of it, are not growing much faster than the natural history literature, at least not recently. Here’s a crude first pass. I used ISI’s Journal Citation Reports to look up data on number of journals, number of articles, and two measures of citation rate (median impact factor and aggregate impact factor) in each of several categories: ecology, plus four “natural historical” (well, organism-specific) categories: entomology, ornithology, mycology, and zoology. I picked those four just because they seemed like the most natural-historical categories ISI has; ISI doesn’t have a “natural history” category, or categories for other groups of organisms like mammalogy, herpetology, etc. I looked up data from the first and last years for which data are available: 2003 and 2012. Here’s what I found. In the ecology category, there were 30% more journals and 62% more articles indexed in 2012 than in 2003. The median impact factor of those journals increased by 58%, and the aggregate impact factor by 69%. The same numbers for the natural history categories: entomology: 36% more journals, 47% more articles, median and aggregate impact factor up by 39 and 48%, respectively. Mycology: 44% more journals, 64% more articles, median and aggregate impact factor up by 121 and 48%, respectively. Ornithology: 47% more journals, 52% more articles, median and aggregate impact factor up by 42 and 50%, respectively. Zoology: 36% more journals, 74% more articles, median and aggregate impact factor up by 35 and 17%, respectively. (note: I’m not differentiating here between growth in existing journals, founding of new journals, and ISI starting to index journals it previously didn’t index. But I don’t see why that should affect my point.)
• It’s hard to publish natural history these days. Thanks to the internet, it’s now easier and cheaper than it’s ever been to publish at least some types of natural history observations and make them widely available in usable form. Think of online databases like eBird, and the growing efforts of museums to digitize much of the information in their collections. And thanks to the rise of Plos One and other open-access journals that only judge papers based on technical soundness, it’s easier than it’s ever been to publish natural history papers that can be found by anyone who wants to search for them. And if the concern is with publishing natural history in more selective journals, American Naturalist, long one of the top journals in ecology and evolution, has a special section devoted to natural history papers. Similarly, Journal of Ecology, another leading ecology journal, devotes a special section to natural history articles on the plant species of the British Isles. And see also previous bullet.
• Ecology is now all about hypothesis testing, statistics, and mathematics, at the expense of natural history. It is true that, starting in the 1960s, ecological research began to put a greater emphasis on hypothesis development and testing, implying increasing emphasis on theoretical development, manipulative experiments, and statistical rigor. See data in Sam Scheiner’s recent piece, The Silwood Circle for a more narrative history, or this article by Jay Odenbaugh for a history focused more on the role of Robert MacArthur. But that trend is decades old now, plus it was one that absolutely needed to happen (fn 4). Ecology had to change in the way it did in order to become anything other than a semi-professional discipline of minor importance, especially compared to molecular biology. And it had to change in the way it did to respond to the increasing demands placed on it by the newly environmentally-conscious public. And don’t just take my word for it (or the word of The Silwood Circle, which tells this story well), see the Ecological Society of America’s own 1965 report on the future of ecology! (ht EEB and Flow). In other words, the choice (if that’s the right word) was not between “ecology remaining grounded in natural history” and “ecology placing increasing emphasis on hypothesis testing, statistics, and math”, it was between “ecology placing increasing emphasis on hypothesis testing, statistics, and math” and “ecology no longer existing as a professional activity.” (At least, that’s my reading of the history here; happy to be pointed to sources telling a different story) But here’s the thing: even though ecology is now more than just natural history, it’s hard to argue that it’s lost its roots and become divorced from natural history. Few of the papers published in ecology journals comprise nothing but natural history, but most certainly build on natural history. And just because the natural history knowledge on which most ecological studies are based isn’t spelled out in the resulting papers doesn’t mean it’s any less important to the conduct of those studies. I’d suggest that, for much of ecology, natural history is like the foundation or frame of a house: often out of sight, and valued for the purpose it serves rather than in its own right, but nevertheless essential to the house, which would otherwise fall down.
• The ecological literature is increasingly filled with mathematics and meta-analyses, which are crowding out papers in which people report field data they collected themselves. This is flat-out false. Here’s the data: most ecology papers, even in leading ecology journals, are about single species. The majority of ecology papers, even in leading ecology journals, are based primarily on observational data never before published (!), and the bulk of the remainder are based on experimental data never before published. Only small minorities of papers in ecology journals are based on mathematics, or on statistical analyses of previously-published data, and the frequency of such papers is unchanging or only slowly growing. All of this has been true since at least the 1980s. So here at least is one concern that can be totally dismissed by data.
• Ecologists these days care about generalities rather than the specifics that natural history is all about. Judging by a show of hands, the admittedly small (a few hundred people) and non-random sample of ecologists attending Tony Ives’ recent MacArthur Award lecture mostly don’t believe ecology is about the study of general laws. Which presumably implies that most ecologists think ecology is about specifics, such as the specifics of particular species in their particular environments. Anecdotal, I admit, but still. (see below for other versions of this worry)
• Ecologists don’t know natural history anymore. I’ve never met an ecologist who wasn’t a good natural historian of their own study system, and often of other systems as well! That even goes for microcosmologists–Peter Morin for instance is an amazing natural historian. Even many theoreticians are good natural historians (The Silwood Circle notes several, for instance). And given the sort of ecology I do, I think it’s pretty unlikely that the ecologists I happen to know know more natural history than the typical ecologist! 🙂 Plus, even if ecologists as a group now know more math, stats, programming, etc. than they used to, and so know less natural history if only because nobody can know an infinite amount of stuff, I still don’t see why that’s a bad trade on balance (see following bullet).
• Ecology students aren’t trained in natural history anymore. I’m sure undergraduate and graduate training in ecology now places more emphasis on statistics and mathematical theory than was the case decades ago. But surely that’s a good thing on balance! The wide availability of computers and software, and associated advances in computationally-intensive quantitative methods, has made it possible to use more and better quantitative methods than was the case previously. The training in every field of science changes as technology and techniques improve, and quite rightly so! Now, there’s always scope for discussion about precisely what things ecology students ought to be taught or trained in. And there are always hard choices and trade-offs to be made. There are many valuable things we could teach, and time spent teaching one thing is always time spent not teaching something else. But even if you think ecologists generally aren’t making the right trade-offs here, please do the world a favor and recognize that there are trade-offs! Rather than just talking as if changes over time in how ecology students are trained must be bad because students now spend less of their time learning natural history (as if students were sleeping or playing video games instead, rather than learning general concepts, statistics, theory, programming, etc.). Further, be aware that, anecdotally, lots of scientists, including ecologists, wish they’d learned more math, statistics, and programming as students, not less. So while it might be true that ecology curricula currently are suboptimal, that might be because they teach too little math, statistics, and programming! Personally I agree with what Jay Stachowicz said recently. I trust my colleagues to do their best at training the next generation of ecologists. And I trust the next generation to listen to the previous generation without taking it as gospel, and then go out and do the best ecology they can. And while I expect the next generation to make some mistakes (just as I would have), and to do some things differently than I would have, I also expect that, on balance, they’ll do things better than I would have.
• We aren’t able to address practical conservation issues due to a lack of ecologists with natural historical knowledge and field training. Yes, many practical conservation issues are challenging to address. Yes, among those challenges often is a lack of basic data on the species concerned (though I wouldn’t venture to guess how important “lack of basic data” is relative to other challenges. In Canada for instance, the biggest obstacle to conserving numerous species at risk is political). But where is the evidence that lack of data is due to lack of people who have the training to collect it, as opposed to lack of funding or whatever? I mean, there surely are far more ecologists in the world now than there were decades ago! Can anyone point to funding for research or salaries not being spent because no one with the appropriate expertise can be found to spend the money on? Further, given that there will always be situations when we’re short on data, I’m awfully glad we have general guidance to fall back on, and that general guidance often comes from theoreticians.
• There are no faculty jobs for natural historians anymore. It’s true that there aren’t many tenure-track jobs these days for, say, ornithologists or entomologists. But there aren’t many tenure-track jobs these days for population ecologists, or quantitative ecologists, or indeed any sort of ecologist, or any sort of non-ecologist. There aren’t many tenure-track jobs for anyone, relative to the demand for them, especially not at research universities (the supply-demand mismatch is less bad at teaching universities). You say that you, as a natural history-based, “muddy boots” field ecologist, are struggling to find a tenure track job at a research university? Well, so did I, and so did Brian, and neither one of us is a natural history-based, muddy boots field ecologist. (Indeed, one reason Brian struggled to get a job is because of the false impression that he couldn’t teach field courses!) And Meg was hired at Michigan in part because she is a muddy boots field ecologist. So if you find yourself often having to “pitch” yourself as something other than what you “really” are in order to compete for academic jobs, well, join the club! Everybody has to do that (I certainly did). Now, it’s true that you won’t get a job or tenure at a research university just by doing natural history. But how long has it been since you could? I doubt E. O. Wilson was tenured at Harvard back in the mid-1960s solely because of “pure” natural history work!
• You can’t get grants to collect basic natural history data on specific species or systems anymore. Yes, it’s hard to get a grant to collect basic natural history data on particular species or systems. But it’s hard to get a grant to do field experiments, or theoretical modeling, or meta-analyses, or anything else. Success rates at the NSF Division of Environmental Biology are less than 10%. Nobody is finding it easy to get funding to do any sort of research these days, no matter how much or little natural history is involved. Having said that, yes, it’s true that in order for you to get funding from a basic science agency like NSF, you are going to have to have hypotheses, aim for general applicability rather than narrow specificity, use statistical tools, etc. But I guess I’d say the same thing about that that I said about the content of the ecology literature above. I don’t see how it’s a bad thing that ecologists have added hypothesis development and testing, sophisticated statistics, mathematical modeling, etc. to their “toolbox” along with natural history knowledge, and that they now use that toolbox to do the sorts of science one can’t do solely by doing natural history (or field experiments, or modeling, or any one thing). (Aside: I’m curious what percentage of funded NSF grants in ecology propose to collect field data, of whatever sort. Judging by data on the content of ecology journals discussed above, I bet the percentage is really high.)
• You have to know a lot of natural history, and be out in the field collecting your own data, in order to be a good ecologist. Sorry, but this strikes me as an overly-restrictive definition. Not every ecologist is a great natural historian. Nor is every ecologist out in the field collecting their own data. But there’s nothing that every ecologist does! Surely everybody has their own strengths! And surely it’s better for ecology if ecologists, collectively, have more strengths rather than fewer. I mean, does anyone seriously think ecology would be better off if it somehow kept out those whose strengths lie in areas other than natural history (mathematics, programming, genetics…um, blogging…)? So that you couldn’t be an ecologist unless you were a good natural historian (or collected your own field data, or conducted your own experiments, or whatever)? Personally, I think we learn the most when we make effective use of all viable approaches (that’s true in science and elsewhere). And don’t just take my microcosmy word for it, or Brian’s statistically-sophisticated, meta-analytical word for it, or Robert MacArthur’s word for it–Meg’s a muddy boots field ecologist, and she agrees.
• You have to know at least some natural history in order to be a good ecologist. And the evidence that lots of ecologists (or even any!) fail to clear this very low bar is…? Jeez, even I know some natural history! 🙂 Further, even if, say, a few ecological theoreticians or statisticians really don’t know any natural history, is it really plausible that their contributions to ecology (which comprise a very small part of the field as a whole; see data above) are seriously damaging the field as a whole rather than improving it by bringing in ideas and approaches that ecology wouldn’t otherwise include?
• It can be helpful to know some natural history, or to have collected some of your own data at some point, in order to be a good ecologist. I agree 100%, but I’m not aware of anyone denying this modest and sensible claim. Also, it can be helpful to know some math and statistics to be a good ecologist. It can be helpful to know some programming. It can be helpful to conduct meta-analyses. It can be helpful to do laboratory microcosm experiments. Etc.
• You have to know a lot about the natural history of a system in order to conserve or manage it. That’s true. But while that’s necessary, it’s generally not sufficient. See discussion here and at the end of this post. Further, even if the world does need more applied ecology research, it’s not clear that the top priority is more natural history work. I mean, maybe that’s true–but it’s far from obvious. For instance, here’s Georgina Mace’s vision of what ecology needs to be in order to be of applied relevance today.
• Natural history is valuable for its own sake, not just as a means to the end of doing ecology. I’m happy to agree, and I’ve never met anyone who disagrees. But insofar as you think natural history is valuable for its own sake (or because of the pleasure it gives you, or the appreciation of nature it fosters, or for any other reason having nothing to do with its relevance to ecology), why do you care about whether ecologists know enough natural history, or value it sufficiently, or include it in their papers, or whatever? Art and literature are valuable for their own sake, but that’s no reason to think that ecologists should paint or write poetry. Plus, scientific research also is valuable for its own sake (and for other, instrumental reasons) and gives pleasure to its practitioners, even when it’s not based on natural history. For instance, see this passionate post from Hope Jahren. Sorry, maybe I’m being totally dense here (wouldn’t be the first time!). But I just can’t understand why the fact that natural history is valuable for its own sake (or gives pleasure to its practitioners, or whatever) has any implications for how ecologists should do ecology. It just seems like a total non sequitur to me. Indeed, insofar as natural history is valuable for its own sake, as an end in itself rather than as a means to the end of doing ecology, shouldn’t you worry that ecology will distort or degrade natural history by making use of it? Much as artists and writers worry that commercial pressures–which treat art and literature merely as a means to the end of making money–will distort and degrade art and literature?

In conclusion, I freely admit that I may well have missed, or mis-described, some issues. And maybe this whole post is just kind of misdirected (I really struggled with how, and even whether, to write it). Looking forward to your comments, as always.

fn 1: Sadly, my own Philadelphia Phillies are the last old-school, purely scout-reliant holdouts. Or at least they were until a few weeks ago. 😦

fn 2: Interestingly, “people who don’t know what they’re talking about, or who have vested interests of various sorts”, still includes lots of people with political power and influence, in contrast to what’s been the case for years in baseball.

fn 3: As usual, you get the background research you pay for on this blog. In seriousness, I actually have no idea whether concerns about the status of natural history, or field-based studies more generally, are at all widespread.

fn 4: Full disclosure: there’s a part of me–the snarky, nasty part–that just wants to respond to this particular worry by going “Come on, are we still having this discussion?!” But for better or worse we are still having it. And as someone who’s on record as saying that there are some discussions we need to keep having even if they never get resolved, I’m certainly in no position to tell anyone else “Just give it up already!” 🙂