As some of you know, I’m writing a book about ecology, aimed at researchers and grad students. It’s about how to harness diversity of ideas, goals, and research approaches in ecology. Once again, I’m asking you to share your favorite examples, so as to overcome my own limitations as an example compiler. Today, I’m looking for examples of selection of questions, methods, research approaches, and results in ecology.
What do I mean by “selection”?
One way to harness diversity is to pick and choose the best option from among a diverse range of options, and discard the others. This is how evolution by natural selection works: the fittest individuals, and their phenotypes, increase in frequency at the expense of the less-fit individuals and their phenotypes.
Scientific research often proceeds in a manner analogous to evolution by natural selection. Someone invents a new research approach, or a new method, or etc., and shows (or at least claims) that it’s superior to others, at least in some respects. The new research approach or method gets taken up by other scientists, at the expense of other approaches or methods regarded as inferior. Model selection is perhaps the most familiar example: you choose the best model of the data, according to some appropriate criterion, from among a set of alternative models. Or think of doing strong inference among several alternative scientific hypotheses in order to identify which one is correct.
You can also think about “selection” in scientific research in a broader and looser sense. For instance, maybe a new research approach or method gets taken up, without ever being compared to competing approaches or methods. Perhaps because the new approach or method performs some novel task, that couldn’t previously be done at all. That’s still “selection” in the broad sense of “evaluating whether or not a given option is fit for purpose, using it if it is, and not using it if it’s not.”
“Selection” in a broad sense also crops up in the context of presenting scientific results to others. For instance: which results do you select for highlighting in your title and abstract, and which ones do you bury in online appendices? These aren’t merely idiosyncratic personal choices; there can be patterns in selective presentation of results at the level of entire scientific subfields. For instance, Fahrig (2017) showed that habitat fragmentation studies systematically under-emphasize positive effects of fragmentation on biodiversity in their titles and abstracts.
You can even think about selection at the level of what sort of research the entire field of ecology chooses to pursue, or not pursue. For instance, Anderson et al. (2019) used text mining to show that, since the late 1980s, ecology as a whole has selected for research on applied topics related to global change, invasive species, conservation, ecosystem services, etc. At the expense of research on older fundamental topics like competition, predation, and life history theory. Or think of frequent criticisms of ecology, and other scientific fields, for selecting for novel studies over replications of previous studies (e.g., Fraser et al. 2020).
So, lots of potential examples of “selection”! Not all of which will necessarily be successful. Selection in science often is challenging, and doesn’t always work out well. Ecologists sometimes don’t even try to do selection when they arguably should. Or when they do attempt selection, sometimes they select for the wrong “traits”–selecting for things that arguably shouldn’t be selected for. For instance, ecologists rarely even attempt to select among alternative hypotheses (Betini et al. 2017). Instead, they just focus narrowly on whether the evidence is consistent with their own preferred hypothesis, while ignoring alternatives with which the evidence also is consistent (Teixeira et al. 2020). Which leads to field-wide biases if everybody prefers the same hypothesis. As another example of unsuccessful selection, newly proposed methods and research approaches in ecology sometimes get widely taken up, only for them to turn out not to work as advertised (e.g., inferring species interactions like competition from species co-occurrences doesn’t work; Barner et al. 2018). Presumably, those methods and research approaches get “selected” initially because it seems like they should work, or because they have other virtues (e.g., ease of use).
So, in the comments, please share your favorite examples of “selection” in ecological research. Successful or not. Could be a single paper, could be a whole bunch of papers from a whole bunch of different people. Could be your own work, or the work of others. Ideally, it’ll be an example that has some “hook” that will make it fun/interesting/provocative/surprising to others. But if in doubt, just share the example and let me decide if it’ll work for the book. I can’t promise I’ll use all your examples, but I’m looking forward to reading them and learning from them all the same. The more examples, the merrier! Commenters on the previous post in this little series came through with some great examples I’m planning on using. So I’m really looking forward to the comments on this one!