Poll results: here’s what our readers think about some of the most controversial ideas in ecology

Recently we polled readers on their views on some of the most controversial–or maybe just seemingly controversial!–ideas in ecology. Here are the results!

This is a long post, but you should stick with it. I found the results very interesting and I think you will too. Below the fold there’s an overview of the respondents, the results, and then some commentary.

Overview of the respondents

We got 251 responses (thanks everyone!), split about evenly between grad students (27%), postdocs (30%), and faculty (34%), along with a smaller number of other ecologists (10%). Obviously they’re not a random sample of ecologists, or even of readers of this blog, though they’re probably fairly representative of our regular readers. But they’re a big enough and diverse enough sample of ecologists to be worth talking about, I think.

I suspect the main way that they’re unrepresentative of ecologists more broadly is that they’re probably a bit more likely than randomly-chosen ecologists to agree with my views on topics I’ve blogged about. (Then again, some of the results suggest that’s not true…) They may also be a bit more likely to agree with one another than randomly-chosen ecologists. But I’m speculating on that.

The results!

Below are histograms of the distribution of opinion about each idea. Responses are on a 5-point scale from 1 (definitely false) to 5 (definitely true). Respondents were asked to skip any idea with which they were completely unfamiliar. Respondents also were asked to indicate their level of expertise about each idea, choosing from the options “I’m an expert”, “I know some”, “I know a bit (e.g., learned it in a class)”, and “I know nothing”. In the histograms, I omitted the opinions of anyone who indicated they knew nothing about the idea (those people weren’t supposed to provide an opinion, and when they did it was usually “3”). I color-coded the opinions of others by their level of expertise: light blue for experts, blue for people who know some, dark blue for people who know a little. Note that the y-axis scale varies among histograms, because the number of respondents varied among ideas.

As best I can tell, respondents honestly evaluated their own expertise. Nobody claimed expertise on most or all of the ideas (which would be implausible), a different set of people claimed expertise on each idea, and only a minority of respondents claimed to be experts on any given idea.

I’ve ordered the ideas roughly from most to least controversial. More controversial ideas are those with a mean opinion near 3, and with a high variance in opinion, especially a bimodal distribution.

So without further ado, the most controversial idea in ecology (well, at least among these respondents, from the choices I provided) is….

Species interactions typically are stronger and more specialized in the tropics!

In second place (or first place, if you prefer to quantify “controversy” via variance of opinion rather than bimodality of opinion): the idea that local biodiversity is declining in most or all localities:

Aside: I am proud that the two most controversial ideas in ecology are ideas we’ve blogged about. 🙂

Next, several ideas that are all quite controversial, but that you could rank in different orders depending on how exactly you quantified “controversial”:

Species’ poleward geographic range limits typically are set by abiotic factors, not species interactions:

(Aside: I did not expect ideas about range limits to be among the most controversial ideas on the list.)

The intermediate disturbance hypothesis (empirical version): species diversity typically peaks at intermediate disturbance frequency or intensity:

The intermediate disturbance hypothesis (one theoretical version): disturbances promote diversity by interrupting competitive exclusion and preventing the community from achieving equilibrium:

Biodiversity is among the most important determinants of ecosystem function:

Habitat fragmentation per se (as distinct from habitat loss) typically reduces biodiversity:

(Aside: that is a weird distribution of opinion–a very clear mode somewhere other than “3”, but also a high variance.)

Interspecific competition can be reliably inferred from information about functional trait similarity:

(Aside: that is another weird distribution of opinion.)

Species diversity typically peaks at intermediate productivity:

Next, in rough descending order, some ideas I’d call moderately controversial.

Species richness on continents is dominated by ecological limits, as opposed to evolutionary limits:

(Aside: apparently, the ASN debate on limits to continental-scale species richness was a tie in the minds of ecologists. 🙂 )

Higher biodiversity will buffer ecosystem function against climate change:

Predator functional responses typically are ratio-dependent:

(Aside: that last graph shows the biggest split between expert and non-expert opinion on the list. I’ll come back to that.)

Top-down effects typically are stronger and more important than bottom-up effects:

Species’ equatorward geographic range limits typically are set by species interactions, not abiotic factors:

The dilution effect: host biodiversity generally reduces disease prevalence or infection risk:

Ecological networks (e.g., food webs, plant-pollinator networks) have the structures they do because other structures would be unstable:

The enemy release hypothesis: invasive species often spread and become abundant relative to natives because enemies attack the invasives less:

And finally, in rough descending order, some ideas that are not particularly controversial:

Hubbell’s neutral model: community dynamics typically are dominated by drift, because all individuals of all species typically are demographically equivalent:

Interspecific competition and facilitation can be reliably inferred from co-occurrence matrices or correlation matrices of species’ abundances:

(Aside: that last one and the next one are the two for which I wonder the most if there’s more consensus among our readers than there is among ecologists as a whole. But that’s pure speculation on my part.)

Interspecific competition can be reliably inferred from information about phylogenetic relatedness:

(Aside: Is that last one a sign that the phylogenetic community ecology bandwagon–or at least, one big important piece of it–got stopped in its tracks by criticism?)

Population growth typically is density-dependent:

And finally, the least-controversial idea on the list: Earth is currently experiencing a 6th mass extinction:

Experts vs. others

There are some very interesting patterns in the views of experts. vs. non-experts. Below is a scatterplot of mean expert opinion on each idea vs. mean opinion of everyone (including experts). Each point is one idea, the line is the 1:1 line.

Expert views correlate quite well with everybody’s views; there’s no idea in this poll on which experts as a group have a massively different view than everyone else. And since experts are only a small minority of respondents for any idea, that positive correlation mostly isn’t an artifactual part-whole correlation. The positive correlation would still be about as strong if I was less lazy and plotted mean expert opinion against mean non-expert opinion.

But here’s the really interesting result: experts think ideas are more likely to be false (or less likely to be true) than do all respondents. That is, the points in the previous graph mostly fall below the 1:1 line. It’s generally not a huge effect, but it’s so consistent across so many different ideas that it’s clearly a real trend. Surprisingly, the trend holds for ideas that the crowd tends to believe, for ideas the crowd tends to disbelieve, and ideas for which the crowd is on the fence. So it’s not that experts are generally more cautious than non-experts–it’s not that mean expert opinion is always closer to 3 than mean crowd opinion. Nor is it that that experts are always more certain than non-experts–it’s not that the experts choose 1 when the non-experts choose 2 and the experts choose 5 when the non-experts choose 4. I did not expect this result, and I’m not sure what to say about it. Really looking forward to your thoughts on this result!

The next plot shows the variance in expert opinion about each idea vs. the variance in everyone’s opinion. Again, the line is the 1:1 line.

As you might expect, there’s a positive correlation, though it’s pretty noisy. Ideas that are controversial among all ecologists tend to be controversial among experts.

But here’s the interesting thing: expert opinion about any given idea almost invariably exhibits greater variance than does all opinion. I don’t think this is just a reflection of the smaller number of experts, at least not in all cases. If you estimate a population variance with a sample variance, your estimate is not biased upward by a small sample size. And in the most extreme cases, the high variance of expert opinion aligns with my own impressions of the literature, so I don’t think we’re just seeing sampling error here:

• Expert opinion about ratio-dependent functional responses in the poll is pretty bimodal, with the larger group of experts rubbishing the idea. That aligns with my subjective impression of the field, that expert opinion is split between a larger group of people who share Peter Abrams’ very negative opinion of ratio-dependent functional responses, and a smaller group of people who share Roger Arditi and Lev Ginzburg’s very positive opinion.
• Expert opinion about the ubiquity of local biodiversity declines is quite bimodal in the poll, with the larger group of expert respondents thinking that local biodiversity declines are not ubiquitous. The opinion of non-experts seems to skew a bit towards the opinion of the larger group of experts in this case. The poll results align with my impression that expert opinion on this issue really is divided into opposing camps. FWIW, in my view the experts arguing against ubiquitous local biodiversity declines (Mark Vellend, Maria Dornelas, our own Brian McGill, et al.) have much stronger evidence and arguments than the opposing experts (Brad Cardinale, Andrew Gonzalez, et al.).
• Expert opinion about the intermediate disturbance hypothesis also looks fairly bimodal, with one group of experts who think it’s definitely false and another group who think it might have some truth to it (hardly any expert respondents think it’s definitely true). I presume that’s a split between experts who think about the IDH more or less as I do, and those who think about it more or less as Doug Sheil and David Burslem do.
• Expert opinion about whether species interactions typically are stronger and more specialized in the tropics also is quite bimodal, although the two camps merely lean in opposing directions rather than being confident the idea is true or false. That’s the only idea on the list for which non-expert opinion also is bimodally distributed; no idea why.
• Expert opinion about the ubiquity of humped diversity-productivity relationships also is pretty bimodal. I’m not surprised, that’s a longstanding vociferous debate in the literature (see here and here).
• Our poll found a particularly high variance in expert opinion about whether Earth is in the midst of a 6th mass extinction. Not sure what to make of that. On the one hand, our poll respondents didn’t include many experts on that idea, so the high variance could well be a blip. On the other hand, the idea is at least somewhat controversial–prominent paleontologist and mass extinction expert Doug Erwin has publicly criticized the idea. I’d be very curious to know if the experts in our poll who disagreed with that idea are mostly paleontologists or people familiar with the paleo literature, and if the experts who agreed mostly work on extant species.
• There was one idea for which I expected a high variance of expert opinion, that didn’t show up in our poll: the dilution effect. I think most people familiar with that literature would say there are two opposing camps of experts on that issue, but you can’t see those camps in the poll data. But very few experts on the dilution effect completed the poll, so we didn’t have much power to detect bimodality of expert opinion on that idea.
• Hubbell’s neutral model is the only idea polled for which the variance in expert opinion is much lower than the variance in overall opinion. Expert opinion in our poll agrees quite strongly that community dynamics are not typically dominated by neutrality+drift. Which raises the question, why isn’t there an opposing camp of experts who think the world is dominated by neutrality+drift? Did the poll just miss them? Does Steve Hubbell not read this blog? 🙂 Or is there some more interesting reason? Speculation: perhaps no expert, not even Steve Hubbell or Graham Bell, ever really believed that the world is really neutral. Rather, maybe even neutral theory’s developers and staunchest proponents saw it not as true, but rather as a thought-provoking null model to consider, or as an approximation useful for certain limited purposes (e.g., explaining the shape of the species-abundance distribution). Maybe you only end up with polarized opposing camps of experts if you have polarized opposing camps from the get-go? That is, further research might sometimes fail to reduce polarization of expert opinion, but never creates it?

So one take-home message of this poll is that, if it seems to you like ecological opinion about topic X is divided into opposing camps, well, that’s probably just expert opinion. The bulk of ecological opinion on even the most controversial topics is hardly ever bimodally distributed. Presumably because non-experts who lack the time and/or inclination to do a deep dive into the literature on a controversial topic just withhold judgment or have vague mixed feelings. Which as an aside is perfectly rational of them. Nobody has time or a good reason to come to a deeply-informed opinion about every issue in ecology!

Commentary

• Most every idea on this list was at least slightly controversial, meaning that every choice from 1-5 was chosen by someone. Proof, if proof were needed, that disagreement is a normal part of science. So if it makes you anxious to know some ecologist out there somewhere disagrees with you on X, well, sorry but you’d better get used to it! Note that this does not mean that every idea in ecology is controversial. I chose most of these ideas because I thought they might be controversial. I didn’t include ideas that I’m sure are non-controversial, like “species richness typically increases with area sampled”.
• But just because disagreement is a normal part of science doesn’t necessarily mean it should be. After all, why should any idea be controversial, at least among experts? If the available evidence and arguments are incomplete or point in different directions, shouldn’t everyone who’s aware of all the evidence and arguments have chosen “3”? Indeed, there’s an old joke that the answer to every question in ecology is “it depends”. So how come there’s no question in this poll for which respondents mostly agree that “it depends” is the correct answer? And how come people who know more about the issue typically disagree more than people who know less about it? Maybe the disagreements revealed by this survey, are a symptom of the fact that we all have our own intellectual biases, and that those matter most when we’re thinking about the issues we care most about–i.e. the issues on which we’re experts. From one perspective that is a depressing thought; arguably scientists could and should be more objective than that. But on the other hand, one could argue that is science as a whole actually better off with some level of controversy, even controversy arising from different scientists having different intellectual biases, their own “pet” hypotheses, etc. Arguably, science as a whole doesn’t need individual scientists to be intellectually unbiased, it just needs effective procedures for selecting among the views of different opposing camps. Maybe the variation in opinion among scientists provides the raw material for a process analogous to natural selection to drive scientific progress.
• Trouble is, that selection process seems to be pretty slow and not proceed to “fixation” in many cases. Ecologists are reluctant to say that an idea is definitely true. I’m struck by the number of ideas for which the modal response was 4, but for which 5 got few or no responses. There was a similar reluctance to choose 1, but that reluctance was evident for fewer ideas. I think this illustrates Brian’s old complaint that ecologists lack a “problem solving mentality”, and my complaints about “zombie ideas”. Ecologists are reluctant to ever completely rule an idea in or out. In some ways, that shows admirable intellectual caution. But the downside is that ecological ideas tend to live on long past the point when you’d think the field would’ve settled the issue.
• Case in point: BEF. Shouldn’t ecologists collectively be kind of embarrassed that, after decades of BEF being perhaps the single most intensively-researched topic in the field, that there’s no consensus on an empirical question as basic as “biodiversity is among the most important determinants of ecosystem function”? Another case in point: the IDH. The IDH has been around for 40 years. It’s in all the textbooks and is the subject of uncountably many papers. Isn’t it embarrassing that simple descriptive statements about the IDH are still controversial? Or perhaps not. Perhaps the timescale on which we should expect to be able to definitively answer these sorts of questions is centuries, not decades? How quickly should ecology, or any field of science, be expected to progress? If science progresses by a process analogous to evolution by natural selection, how long should we expect it to take for an idea to “go to fixation”?
• Speaking of the IDH, I’ve been banging on against the IDH online for 7 years, and I also have a widely-read TREE paper that grew out of those blog posts. And yet a plurality of my own readers still mostly agree with the IDH! Assuming that the respondents weren’t just trolling me (and if you were, well played :-)), that illustrates that you shouldn’t overestimate the influence of this blog–or any one person or paper–on ecologists’ collective thinking. I know that some ecologists don’t like my “zombie ideas” rhetoric, which is fair enough. But if the reason you don’t like it is because you think this blog is massively influential, well, you’ll be glad to know this blog is not nearly that influential. I cannot appreciably shift ecologists’ collective thinking about an idea merely by calling that idea a zombie. Even though I wish I could! 🙂 (I’m kidding, I’m actually glad neither I nor anybody else has that much power.) More broadly, influential individual scientists have some power to generate discussion, but no individual has the power to dictate the field’s collective thinking. (Aside: I don’t know if many people actually worry about this blog being overly influential, so maybe this bullet is attacking a straw man.)
• My own views mostly agreed with the crowd’s modal view, or were only one step removed. But there were some exceptions, and I confess I find them some combination of surprising, frustrating, worrisome, and depressing (What can I say? I’m me.) The biggest exception was the least-controversial idea on the list, that the Earth is experiencing a 6th mass extinction. I’m not an expert on this one, but FWIW I’m pretty convinced by Doug Erwin’s arguments against that idea. I really worry that respondents are expressing a much stronger consensus in favor of a currently-ongoing 6th mass extinction than can be justified by the empirical evidence. And “1” (definitely false) is the correct view on that theoretical version of the IDH, but was chosen by only a minority of respondents (Chesson & Huntly 1997, Fox 2012, and see this blog post, this one, and the series that starts here). More broadly, the degree to which an idea is controversial doesn’t line up all that well with my own views. The consensus occasionally is wrong, some ideas that are quite controversial really should not be, and others that aren’t really should be! I admit that bugs me (I know, I know, I should remember my own advice–disagreement is a normal part of science…) Anyway, I may do a follow-up post where I reveal my votes and explain my reasons for them, with links to key citations.

In conclusion, really looking forward to your comments on this one. I am especially interested to hear from readers whose views are far from the consensus on any of these ideas. Genuinely curious why you think as you do, especially if your views place you in a minority.