Injecting molecules from a sea slug that received tail shocks into one that didn’t made the recipient animal behave more cautiously.
The domination of model organisms and charismatic megafauna in the literature is a disservice to the life sciences.
May 30, 2017|
As researchers working to understand animal behavior, we have studied only a small subset of the more than 1.5 million described animal species. This is unavoidable, as there are many more species than scientists. But the animals we work with are our windows into nature, and it is increasingly clear that our field revolves around animal subjects that simply do not reflect the diversity of the natural world.
Studies in related fields such as ecology and conservation have repeatedly found that research effort is skewed towards warm-blooded vertebrates (birds and mammals) and against cold-blooded vertebrates and invertebrates in general. Our well-documented preferences for what we consider to be attractive or charismatic creatures may be limiting our contributions to a broad understanding of nature.
Does it matter that there are five times as many publications per year on primates as there are on beetles?
To test whether these skews exist in animal behavior research, we assigned taxonomic information to the 4,076 research articles published in the journal Animal Behaviour from 2000 to 2015 and compiled a data set that combined this with citation metrics for each article. By comparing this data set to actual species numbers, we were able to quantify the direction and magnitude of taxonomic skew in published papers.
Our findings can be summarized in two major points:
First: The warm-blooded vertebrate skew was intense. Almost 85 percent of described species are arthropods, but more than 70 percent of publications were on vertebrates. Birds and mammals alone accounted for well over 50 percent of publications, despite representing less than 2 percent of all animal species.
Second: In a world where citations are used to measure impact, publishing on understudied systems comes at a cost to the researcher. Publications on vertebrates received more citations on average than arthropod papers. They were also far more likely to be “blockbuster” publications with more than 100 citations.
How do we resolve this problem? First, we have to agree that there is one.There is no law that says we must invest as much effort studying dung beetles as we do studying chimpanzees. Some would argue that biologists’ work is intended to generate knowledge that can benefit humanity, and that it makes sense to focus on taxa that are best suited to addressing human-relevant questions. So does it matter that there are five times as many publications per year on primates (fewer than 500 species total), as there are on beetles (with a staggering 240,000 described species), as our study shows? We think it does.
For theorists, a narrow taxonomic focus does a disservice to all branches of the animal sciences whose goals are to understand the broad processes and patterns of the natural world. When we limit ourselves to a narrow subset of life, we generate narrow answers to broad questions. For pragmatists, a narrow focus makes it challenging to apply insights from animal behavior research to real-world problems. For example, insects are a critical part of most terrestrial food webs. If we study only a handful, how do we predict how changes in climate will affect their dispersal, or their foraging behavior, or the welfare of the vertebrates that depend on them as food? We are not saying that the intensity of study should match biodiversity. We are saying that it is critical to try to learn more about a broader range of organisms.
Truly resolving this issue will require an understanding of its causes. For example, might human preferences for charismatic species affect what we consider to be important or broadly relevant science? Previous work shows that papers on non-model organisms have more broadly framed introductions, suggesting that the bar for relevance is higher when the taxon is less appealing.
Our demonstration of uneven citation patterns is also consistent with this. We must ask some uncomfortable questions: Do papers that differ only in their study subjects get treated differently by reviewers or editors? Do grants? Do job applications? We can speculate, but these questions can, and should, be answered empirically. Luckily, we are members of a community that is equipped to do exactly that.
In the meantime, we suggest three ways to engage with this issue:
Be aware of your own potential bias when reviewing grants or papers: Ask yourself honestly whether your assessment of the quality and relevance of the work is a result of the quality of the science or of the taxon under study.
Be proactive when citing other publications: Publications on birds and mammals are more likely to cite within their taxon than are publications on non-model systems. In addition to reading broadly, make systematic efforts to consider the relevance of studies asking similar questions in other taxa.
Keep the conversation going: How big a problem is all of this? How can we determine the causes underlying these patterns? Should established researchers consider branching out taxonomically? What can my department do; what can my journal do?
We have strong evidence that the taxonomic research skew exists, and that it is severe. We have some hypotheses about causes. We have the data needed to answer some of the most pressing questions. Now, all we need is the will to explore the issue further.
Malcolm F. Rosenthal is a postdoc in Damian Elias’s lab at the University of California, Berkeley, and formerly a researcher in the lab of Maydianne C.B. Andrade, an evolutionary ecologist at the University of Toronto, Scarborough.
May 30, 2017
"Does it matter...? We think it does." Interesting perspective. But then, why should it be done this way? Rather than that other way? In the end, it's a matter of what one gets out of it, that is, what a funding taxpayer gets out of it. Seeing cool spiders on a National Geographic special is interesting and fun, but, sorry, that's as far as things go when it comes to paying for something for the vast majority of people. Please show us all the truly pertinent reasons why we should fund grants, and the words "Okay, fine." should be readily apparent. But, this article doesn't do that. So, what is viewed as bias from the writer will remain...even in his own scientific field. In the end, this sort of research will be funded solely by university endowment money, so those in the field should become fully prepared for that.
May 30, 2017
As mammals we tend to choose animals that look like us either as pets or scientific models. Since Claude Bernard in the 19th century, mammal models have been used to solve problems related to human physiology in a comparative way. Nevertheless, as a worm researcher I would like to call attention to the text below taken from a Johnathan Hodgkin paper published in Genome Biology in 2001 on the 20,000+ genes of Caenorhabditis elegans (Jonathan Hodgkin, Genome Biology 2(11):comment2008.1–2008.4, 2001): "... there is only one organism that we have any hope of understanding fully, in terms of biology. That organism is our own species. Only we are in a position to report on every disease or toxin we encounter. Similarly, only we can adequately monitor our own physiological and genomic responses to arctic blizzards, tropical heat, psychological stresses and social pleasures. And only for human beings will we ever have truly extensive data on genomic variability and molecular paleontology. The other tens of millions of species on this planet are likely to remain largely mysterious forever, specially as most of them may well go extinct during the present century, along with all their complex biotic interactions. The biological world will then become more comprehensible but incomparably poorer."
Maybe this is part of the explanation of the biases you find in your scientometric analysis: we are interested in ourselves (and anything that resemble us) more than any other living being on Earth.
May 30, 2017
By the reasoning of this piece, more than half of all biologists ought to be entomologists. That would be nice, but I doubt it will happen any time soon.
Perhaps a more (de)pressing issue is the observation that in 2016, the U.S. spent more ($76 billion) on R&D related to the military than it did on all other research combined ($68 billion) (1). Of the latter, a mere $6 billion went to the National Science Foundation, and of that, only $215 million went to Integrative Organismal Systems, which is the division that supports research on animal behavior of non-model organisms, as well as physiology, neurobiology and developmental biology (2). In sum, it seems that support for U. S. scientific research is driven almost entirely by human fears about death of one sort or another. For "basic" science - as they say, the knives are so sharp because the stakes are so small (and likely to get a lot smaller).
May 30, 2017
What an interesting paper and it brought back some fun memories from the past. Many years ago, probably around1960, when I was a graduate student at U. of Illinois, a fellow graduate student and I, after a bit of beer, addressed the same issue --that biology was being built on a few type organisms or parts of them. We decided to try to construct an evolutionary tree built on "standard organisms". I no longer recall all of the components but I know that after several days we had a chart done that we submitted to Science to use as a cover (to the dismay of our advisors).
At the time, more than 55 years ago, we had life beginning with the "primordial soup". For early life we had E. coli, of course and T2 phage. We pretty quickly moved up through Tetrahymena and Paramecium and went on to components like "squid giant axon",and for botanical things there was the onion root tip, and the garden pea, then a Planaria that was kind of split in the middle, then Drosophila--we argued about whether it would be bar eyed or bar. bar, infrabar, we kind of skipped over Ascaris and we moved on then to what was thought to be more important species and parts like, something about a frog with a leg clipped, the lab rat, rhesus monkey, ablated frog, a highly inbred mouse, beagles, etc., etc.
Science never accepted our submission for a cover chart but the exercise at the time was a blast and I encourage those of my colleagues with a spirit of fun to sit around in a group with some beverage and build their more modern tree of life.
May 31, 2017
I hadn't clearly thought along these lines, but the question might be related to some thoughts that have haunted me for several years by now.
In particular I am concerned with the question of brain size as related to mental power (intelligence if you like, or at least apparent intelligence).
I grew up repeatedly confronted with solemn assurances that brain size (or at the least, ratio of brain size to body size) was closely, and nearly universally, correlated with IQ. Well, there certainly is some correlation, but I am increasingly convinced that some anomalies in the trend are very significant.
For example, I have observed under uncontrolled conditions, some eerily intelligent behaviour among vertebrates including reptiles, that are supposed to be really dumb, and of course, birds achieve some startling intelligence with mere bird-brains.
However, consider some of the spiders, such as Portia, that can see a prey item in an inaccessible position, and work out a long and elaborate detour to find it directly. I have seen similar behaviour in chameleions.
Its brain is so tiny, that one can hardly imagine how it manages to coordinate its limb movements, let alone exhibit purposive behaviour.
In comparison, my own cranial bowl of porridge seems materially excessive for its performance.
Some cephalopod behaviour also is so impressive as to spoil one's appetite for sea food.
And other animals such as horses have shown behaviour patterns that no one who had read about "Clever Hans" would have credited it. And the nature of science is such that straws in the wind often get ignored. Not all material information comes in properly controlled packets.
In short, that strikes me as a field that would justify (and require!) a great deal of work to make much sense of the study of brains and of the basis for defining intelligence at all.