Subtle biases with important consequences

Note from Jeremy: this is a guest post from Mark Vellend.

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

As a Ph.D. student (long ago) I was interested in studying ecological consequences of human land use that might last a really long time.  One of my Ph.D. advisors, Peter Marks, and colleagues had reconstructed the history of each bit of forest in Tompkins County, New York, providing a superb template for asking how biodiversity varies between forests never cleared for agriculture (primary forests) and those growing up on abandoned agricultural land (secondary forests) starting at different times.

From that point of departure, I then made a critically important decision.  I am a plant ecologist, but rather than survey the entire plant community in different forests, I decided to focus on “forest herbs”: herbaceous species whose persistence seems to depend on forest cover in the sense that their main and sometimes exclusive habitat is the forest understory. As a group, forest herbs tend to be very long-lived, most lack a persistent seed bank, and a great many lack mechanisms for seed dispersal beyond a few tens of meters.  (Many have seeds dispersed by ants.) As such, they are expected to be quickly eliminated when a forest is converted to agriculture, and very slow to return when the forest grows back.

Good decision, right? Well, it depends on the question.  If I wanted to ask, “are there some species whose (meta)populations require many decades to recover from historical disturbance?”, this was a wise choice. But if I wanted to ask if plant biodiversity in secondary forests is lower than in primary forests, my decision created an obvious bias. I deliberately chose the subset of the flora most likely to remain underrepresented in secondary forests for a long time. Indeed this is exactly what I found, as have many studies of a very similar nature in temperate forests.  In a review, we said: “For centuries after agriculture has ceased, plant communities on abandoned agricultural lands remain impoverished in herbaceous species characteristic of uncleared forests.” 

The problem is that the quote above is an answer to the first question (about a particular set of species), but it is routinely read as an answer to the second question (i.e., evidence of a negative effect of human actions on biodiversity). If we take a broader view of the landscape and how human activities have likely influenced plant biodiversity, the story is rather complex and nuanced. While the diversity of forest herbs might be lower in secondary forests, the diversity of all plants might be just as high or even higher in secondary forests. We have found that the composition of different secondary forest patches tends to be more similar than among different primary forests, suggestive of biotic homogenization caused by land use. However, if we take primary forests as representing what the landscape looked like prior to European influence, then beta diversity in the current landscape is almost certainly much greater than it was previously, given greatly increased environmental heterogeneity (= biotic differentiation). The cartoon version of these and other results (below) is modified from a recent paper in a philosophy journal that I wrote on this topic (same journal issue as Jeremy’s philosophy paper!).

Despite this highly context-dependent picture of how biodiversity has changed under human influence in a commonplace landscape like this one, our results for forest herbs appear consistent with the idea of a “biodiversity crisis” – a powerful and dominant narrative in conservation biology, a field with which many (most?) ecologists feel a strong alignment.  The existence of our science is often justified by how our research can help fight the crisis, and so we are quite adept at finding bad news stories.  My experience is that ecology undergraduates arrive at university already with a sense of certainty that any and all changes in nature caused by human actions are for the worse, and so essentially the conclusion sections of term papers and research projects can be written before any research is done: people are screwing up nature, probably more than we thought, and we’ve got to do something about it. In my view, reality is much less straightforward, and far more interesting.

The main point of telling this story is to argue that I think we have good reason to be concerned that as ecologists, we make decisions about what to study, how to study it, what to report, and how to interpret the findings, that seem likely to emphasize some conclusions more than others. As a collective, ecologists tend to share a set of values that might create an important source of bias in our science. If we each had individual biases in different directions – representative of the broader society – we might hope that they would cancel one another out and that a fairly accurate view of reality would emerge. I assume that most readers would agree that ecologists bring a set of values to the table that is not representative of the distribution of values found in the societies in which they live.  (As one example, just think about the distribution of votes among political parties wherever you live: ecologists vs. a random sample. Significant differences? For social scientists in the U.S. the difference is massive.)

The other point in telling this story is to draw attention to the article mentioned above, in a philosophy journal, that explores this topic in greater detail (not going to bother hiding it): “The behavioral economics of biodiversity conservation scientists”. I present additional examples where values entered into ecological science, and I draw on the literature from conservation and environmental science about common values shared by ecologists, from the social sciences on the relationship between values and science, and from psychology (the “behavioural economics” part) on some of the reasons we might have developed preferences for how nature ought to look (e.g., more like it did centuries ago than it does now). I hope that some of you find it thought-provoking and worthwhile and that you’ll let me know what you think. 

I think that as ecologists it would do us well to think more deeply about the values we bring to our science, how it might influence what we conclude about the world, and whether there are consequences for our credibility and our hopes to influence policy. The example I laid out here is, in some senses, subtle, with results represented accurately in publications, but with an overall message that emphasizes the “people are bad for biodiversity” message to a degree that seems misaligned with what has actually happened in the landscape. Just as a small selection coefficient and many generations in a large population can lead an allele to fixation, so might small biases in one direction, among many researchers, over many years lead us to a skewed view of the world. Is it happening? And if so, is there anything to be done?

29 thoughts on “Subtle biases with important consequences

  1. Nice post! I am, of course, on board with the idea that our values affect our science and am glad when more of us approach this from different angles (meta!), and I’m curious about whether fields such as science and technology studies (STS) and geography get at similar questions in probably-complementary ways. Excited to read the paper to explore those overlaps and divergences, thanks for the push to do so!

    • Thanks, Ambika! When one wades into literature far from their “home” discipline, I’m not always sure how to classify particular studies. As best I can tell, geography covers a truly massive range of things. I’ve read a decent number of things that would probably fall under STS, so I think that point of view is likely represented in the article. But of course others might have a different perspective (within disciplines too)…

  2. Mark, thanks for bringing this up with your personal anecdote. When I was a grad student in the 1980s, we were discouraged from doing work that might have an applied conservation component. The main internal bias we had to fight against was being in love with our own pet hypotheses. Although our advisors beat this into us, not everyone managed it. But, it was also fashionable to pick apart someone else’s pet hypothesis. So we did get the kind of competition of ideas that would presumably move us closer to the truth on a particular concept (departmental seminars more resembled gladiator matches than the hug fests they are today). But by the mid 90s, most ecologists embraced their motivation and moral obligation to understand and slow the biodiversity crisis. Our community pet hypothesis coalesced into “people do bad things to nature” – a hypothesis for which there was much obvious evidence, and relatively little dissent, at least within the academic community. I believe that we do have an obligation to understand the biodiversity crisis, but that this means doing good science, which of course means acknowledging and questioning our pet hypotheses. I think it would be a worthwhile endeavor to tabulate what those pet hypotheses are. Three obvious ones are that biodiversity generates ecosystem services, invaders impact native species, and climate change reduces biodiversity. In my study area, the main ones are that humans make infectious diseases worse through stressors like disturbance, biodiversity reduction, climate change, and pollution. All of these have substantial theoretical and empirical support, but challenging their generality ruffles feathers and has professional costs in terms of grants and prestige. A good first step is to have these conversations with our students and acknowledge that as people who generally care deeply about nature, it can be difficult to study it objectively. Your paper is a nice step in that direction.

    • Thanks, Kevin. Those are really interesting anecdotes as well. I think this sums it up nicely: “as people who generally care deeply about nature, it can be difficult to study it objectively”. You won’t be surprised to learn that I have these conversations with students (and others) quite frequently. I would hesitate to compile things under the heading of “pet hypothesis”, because it could be read as belittling, but I do think we’re primed and ready to “see” some kinds of evidence far more than others. “Climate change reduces biodiversity” is so interesting because in a great many places evidence seems to indicate that it’s the cold and short-growing seasons that keep biodiversity lower than it otherwise could/would be. In others at dry/hot limits the opposite is true, so overall there isn’t just one story there.

  3. Mark, I think you’re right that ecologists’ values skew in one direction. But, I’m not sure this is a big deal when it comes to science-based policy.

    As scientists, we have to get over the idea that our role is to provide objective and bias-free data to decision-makers, who then make logical science-based policy. Instead, we should see ourselves as partners at the negotiating table, alongside other stakeholders with their own values, biases and political priorities. Thus, the goal isn’t *science-based* policy, but rather *science-informed* policy. This perspective is consistent with post-normal science.

    In post-normal science, ‘objectivity’ is not what determines the usefulness of scientific contributions. Instead, ‘authenticity’ is what matters most because it means that scientists can be trusted to be negotiating in good faith. So, we shouldn’t try to be bias-free, but should rather aspire to be transparent with our biases. (Shameless plug: we have a paper on this topic: https://doi.org/10.1111/csp2.73).

    That said, you raise an interesting point about the role of value biases *within* ecology. Does the broader ecological community collectively misinterpret data because we all share the same values? I’m really don’t know, but I wonder if it is in the eye of the beholder? Since biodiversity is such a poorly defined concept, it is very difficult to conclude anything definitive about it. For example, if species richness increased across a land-use gradient, but evenness decreased, did biodiversity increase or decrease?

    Maybe a bigger problem than our subtle biases is that we are trying to answer intractable questions using tractable metrics?

    • Thanks, Falko. Good points. My concern here is largely with what happens in science itself, under the influence of our non-representative distribution of values. This might be related to political outcomes we desire, but how we approach the next step of informing policy is a different, if related, question. Indeed we are not bias-free machines providing information, and I totally agree that we should “aspire to be transparent with our biases”. But in order to do that we need to know and understand those biases, and my point (opinion) is that there’s a great deal of progress to be made in ecology on that front. We can’t be honest about biases when we are unaware of their existence. If biodiversity seems intractable, then just think about recent discussions of the Living Planet Index, which is based on pretty simple underlying data. Biodiversity is just one example, so I don’t think the values-bias problem is because things can be measured in multiple ways. In the example I presented, the issues arise even if one common metric was used always.

      • “But in order to do that we need to know and understand those biases, and my point (opinion) is that there’s a great deal of progress to be made in ecology on that front. We can’t be honest about biases when we are unaware of their existence. ”

        I tend to agree with you. I’ll even go out on a limb and hypothesise that researchers who define themselves as ‘conservation scientists’ are more likely to be aware of their own biases than researchers who define themselves as ‘ecologists’ (even though there is still a lot of overlap)

        Michael Soule’s landmark paper back in the 80s explicitly defined conservation biology as a “mission- or crisis oriented discipline” and since then the conservation literature has only grown closer to political, economic and social topics: https://doi.org/10.1111/cobi.13435 Self-identified ‘conservation scientists’ might feel more comfortable acknowledging the limits of scientific objectivity.

        I am super curious how people’s formal education might affect this? In my case, I came to ecology from a background in environmental management, so I was exposed to the messy world of value judgements early on. But I can imagine that ecologists who came via maths or physics (or even a theory-focused ecology department) might have different expectations of scientific objectivity.

    • Falko: I think those are really interesting questions about how our training will influence our perception of biases (in ourselves and others). I certainly don’t have answers. My own sense of bias across the collective of ecologists is no doubt itself biased by my own interactions! That said, I have interacted with probably thousands of people over the years (e.g., seminar audiences, chats at meetings) and I’m quite convinced there’s enough of it out there to matter, even if my quantitative guess of how much would be an overestimate. If you present a result that aligns with the “people are bad for biodiversity” narrative, you just get nods and little critical thinking about the underlying science. If you present a result that doesn’t align with that narrative (and point that out), all of a sudden people wake up and become quite adept at finding reasons you might be wrong. This applies to people in audiences ranging from first-year undergrads to senior scientists. To revisit my evolutionary analogy, one case of result X getting a free pass and result Y being scrutinized (selected against) is no big deal, but if it happens again and again and again, we will enter discussions inside and outside science with what we might call “consensus state of knowledge” that we don’t even realize is biased.

      • ” If you present a result that aligns with the “people are bad for biodiversity” narrative, you just get nods and little critical thinking about the underlying science. If you present a result that doesn’t align with that narrative (and point that out), all of a sudden people wake up and become quite adept at finding reasons you might be wrong. This applies to people in audiences ranging from first-year undergrads to senior scientists.”

        Yeah, that’s my (admittedly anecdotal) experience too. It’s not just an issue with “people are bad for biodiversity”, it’s a broader problem. Tell someone something that jives with the conventional wisdom, or their own personal experiences or intuitions, and they usually won’t think too critically about it.

      • Agreed, and in many ways that’s entirely natural and understandable: we have limited mental capacity and time, and so we spend more of it on situations that challenge us to see things differently. I can’t imagine any of us are immune to that. But I think it’s worth being aware of potential consequences. I imagine two buckets filling with water (evidence), and bucket A starts with more than bucket B. If we let new drops fall into A “easily”, but we scrutinize all drops headed for B and exclude any with trace impurities, then it won’t take long before A has *way* more water than B (far out of proportion to the rate of raw input).

      • “I can’t imagine any of us are immune to that.”

        Well, we’re probably doing it a bit right now. After all, when I read the post, you’re just telling me something I was already inclined to agree with.

  4. Nice post. I think ecology, conservation biology, and related fields have long grappled with this issue. This reminds me of an article in The Economist from over a decade ago, dealing with taxonomy and whether emerging information suggesting a rise in species numbers (not merely subspecies) was value-laden, or based on scientific truths. Were taxonomists and ecologists debasing their currency (the definition of what constitutes a distinct species) in the interest of conservation goals?
    Source: https://www.economist.com/leaders/2007/05/17/hail-linnaeus (paywall)
    One point is that to hear an ecologist make these arguments comes across somewhat differently compared to hearing them from a relatively conservative mainstream source like The Economist.

    Also interesting – one letter-writer, apparently a taxonomist, responded with a plea: taxonomists are scientists, not value-driven conservationists! May we all be so noble and pure.
    Source: https://www.economist.com/letters-to-the-editor-the-inbox/2007/05/28/hail-linnaeus-may-17th

    • Interesting story Dan, I wasn’t aware of that old Economist piece.

      I wonder, is there any way to quantify the extent to which taxonomists are “lumpers” vs. “splitters”, and whether the proportion of “splitters” has increased over time?

      • Thanks, Dan. That is an interesting example (even if I can’t access the original article). I hadn’t really thought of taxonomists in this context, although thinking of it now…I recall when teaching conservation biology years ago, I consulted the list of Canadian species at risk, and a surprising number were populations (e.g., of fish from one or a few lakes) deemed “distinct”. Memory admittedly hazy. When converted to numbers like the percent of species that are at risk (read as “threatened because of people”) it can make a non-trivial difference. Presumably those populations were always at high risk of extinction. Anyways, I’ve always thought of splitters as people who were just inherently atuned to and fascinated by small differences (not having any particular agenda).

  5. Mark, sorry I am late to the party, but thanks for the nice post and for taking on this topic. From my armchair here, it seems like the bias you are describing comes from a desire to do science that will motivate public action to save biodiversity: if I can demonstrate that people are harming nature, then I can help convince society to change. Not only does this kind of thinking lead to the bias you describe, but I doubt it would ever be effective for advocacy either! Does anyone really think that a little more evidence of human impact is going to sway public opinion? Lack of evidence of impacts is not the problem! But there are other motivations for tackling mission-oriented ecology that I don’t think would lead to bias, like improving management or optimizing investments of conservation dollars. The key difference is that these don’t have a goal, even if unconscious, of fueling an advocacy campaign.

    Separate topic: When results from the biodiversity ecosystem function experiments started rolling in, I was convinced that the results would not extend to more “realistic” settings, and that this line of research was just another example of ecologists trying to manufacture ammunition for advocacy. But comparisons of biodiversity effects to effects of other kinds of drivers and efforts to scale-up the results in space and time show that small-scale experiments actually underestimate biodiversity effects, and often quite dramatically. Have you been watching this line of research? Are you as surprised and impressed as I am? Or maybe you are reading it more closely than I am and finding reasons to still be skeptical?

    • “When results from the biodiversity ecosystem function experiments started rolling in, I was convinced that the results would not extend to more “realistic” settings, and that this line of research was just another example of ecologists trying to manufacture ammunition for advocacy. But comparisons of biodiversity effects to effects of other kinds of drivers and efforts to scale-up the results in space and time show that small-scale experiments actually underestimate biodiversity effects, and often quite dramatically. ”

      So Peter, how much younger than you and Mark and I do you think you have to be to not really be aware that BEF research was once a huge fight? I remember as a newly-minted postdoc, with the freedom to design my own research program, telling Shahid Naeem that there was no way I was going to work on BEF. It was too much of a polarized, combative topic for me to want to work on it (I later changed my mind).

    • “Not only does this kind of thinking lead to the bias you describe, but I doubt it would ever be effective for advocacy either! Does anyone really think that a little more evidence of human impact is going to sway public opinion?”\

      Peter Morin once told his lab “if you wanted to save the planet, you should’ve become a lawyer or gone into politics”. I actually have no idea how true that is. It’s hard to say just because most people’s individual impact on the planet’s future is going to be tiny, no matter what profession they go into. But I feel like he has a point, FWIW.

    • Peter – thanks for your great comments. You make several points. First, plenty of things get written in a way that suggests that yes, the authors believe that more evidence will help sway public opinion (to my eye anyway). To paraphrase: “look, things are bad, maybe worse than we thought, something needs to be done”.

      Second, in teaching I often make a distinction between two categories of link from ecological science to policy: (1) something you find *justifies* the need for nature protection to begin with (i.e., influencing what policy goals should be), and (2) society (or some agency or institution) already has goals and your science *guides* strategies towards achieving those goals. I think you’re arguing that 2 need not raise any problematic issues related to values, and I would agree.

      Finally, I haven’t read the papers you mention closely enough to comment (I would assume we’re talking about correlations and the possibility of unknown/unmeasured covariates?). Stepping back, the proposed links go like this: humans do stuff -> biodiversity declines -> ecosystem function declines -> ecosystem services are compromised -> people suffer. The nature of my critique has not been about BD -> EF (the studies you mention) but about how often biodiversity declines *at the relevant scales and in the relevant contexts* for the quantitative effects making it all the way through to be large (relative to other things) and negative. I don’t want to re-hash the details here, but the gist is described in this paper if you’re interested: http://mvellend.recherche.usherbrooke.ca/2017-03Vellend.pdf

  6. Pingback: Scepticism towards the Living Planet Index is unwarranted | The Solitary Ecologist

  7. Pingback: How has the frequency of “but see” citations changed over time in ecology and evolution? | Dynamic Ecology

  8. Hi Mark – Very thoughtful post! I can’t remember from your article (I read it in the ‘before times’!) what solutions do you propose? Pre-registration of methods/study goals seems like an obvious one – I think many ecologists face an uphill battle publishing papers that ‘go against the grain’. This is probably exacerbated if you are not a well-known and/or already respected name in the field. Is there a push for this? I could envision that many journals might being open to a handful or a section of pre-registered studies. It would also allow journals and editors to require or ‘pre-review’ some of the best practices that are being circulated (e.g., American Naturalist has a ‘Best practices’ checklist document that is excellent – but we can currently only apply it after the study has been done….).

    • Preregistration might help. But it’s not a cure-all (not that you or anyone else said it was!) I’m thinking of Lenore Fahrig’s work showing that habitat fragmentation researchers underreport positive effects of fragmentation on biodiversity in their paper abstracts. (Fahrig 2017: https://oxford.universitypressscholarship.com/view/10.1093/oso/9780198808978.001.0001/oso-9780198808978-chapter-5) How can you pre-register which results you’ll include in your abstract?

    • Risa – good question! It’s easier to point out problems (what I’ve done) than it is to come up with good solutions (what I haven’t done), but perhaps simply acknowledging that there is a problem is an important first step. I sense some resistance to just that first step, and without it moving further will be difficult. Pre-registration could help, but in a limited way I think. In my own examples I think it’s all very subtle, with the key bits coming at the stage of choosing the set of species that “count” and the habitats to look at. When I presented that work, I described the choice as a strength – choosing a “system” most likely to reveal if a certain phenomenon was important. Not sure how pre-registration would change anything. Pointing out the downside of such choices (not just benefits) might help, but (a) there’s little incentive to do so when you hope to get your paper accepted, and (b) it’s not clear it would make a ton of difference to how people “hear” the results. A friend sent me this jarring example of how, if you’re expecting to hear something, and you hear something close enough, you will hear precisely what you expected (worth clicking): https://twitter.com/SteveStuWill/status/1348049804832632832?s=20

      Anyway, I think your question is a great one we should all think about more…

  9. I very much agree that it’s important to examine one’s own biases, which so often go unacknowledged, and I thank Mark for delving into this important issue. In the interests of taking the conversation further, I’d like to push back on one assumption that seems central to Mark having reached the conclusions he’s reached: that increases in local species richness are “good for biodiversity”. I would be curious to know what the basis is for that*. As someone working in the conservation field, what I’m concerned about in relation to human impacts on landscapes are: (1) extinctions, and (2) loss of ecological functions, with extinctions being the more urgent priority because they are essentially irreversible, and a clear moral wrong. Local species richness is irrelevant to (1) and a very incomplete proxy for (2), so it is not at all clear to me why we should care about it.

    In the example of the woodland herbs, older Mark is right that younger Mark was wrong to extrapolate his result to biodiversity generally. However, younger Mark was right to see the result as showing a negative impact of habitat degradation. If those plants are at no global risk of extinction, then their replacement by other species may be no big deal. However, taken as a global pattern, the dominant change that we are seeing is the loss of more specialised species associated with relatively stable environments, such as lowland primary rainforest or ancient grasslands, and their replacement with a smaller global pool of species better adapted to disturbed landscapes. In that context, younger Mark’s finding is congruent with that pattern. Why is such a pattern “bad”? Because there are many more species in the world which are relatively specialised and have relatively small ranges than there are species which do well in human-dominated landscapes, and the overall result is that human transformations of the landscape are leading to an accelerated rate of global extinctions. I hope we can all agree with the normative statement that this is a bad thing.

    What does this mean for the discussion above? It is in fact quite sensible to take the approach that species not present in natural habitats in a region “count for less” in a conservation assessment (as done in the Barlow et al. paper Mark cites in the 2019 article), to the extent that those species have colonised from elsewhere, are widespread and adaptable, and are at low risk of extinction. Their expansion into a deforested landscape adds little or nothing in terms of global biodiversity, while the disappearance of a regionally endemic species is a big loss. It is of course of interest to study all the species, winners and losers, in the landscape, but it is perfectly rational to attach more importance to some changes than others – it is not just an example of irrational “the bad is stronger than the good” bias. What I’d ask Mark to do is to examine whether he is actually asking the same question as researchers who see increases in non-native species as a bad thing. I contend that he is not. For those researchers, we are interpreting our results through the lens of what it tells us about biodiversity change globally. An increase in invasive species in a new region adds essentially nothing to global biodiversity, while in some cases having serious negative impacts through extinctions and increased extinction risk of other species. From that perspective, it is not a matter of weighing up the positives and negatives. At global level, there is no positive. It seems to me that Mark is more interested in changes in gamma diversity at the level of a landscape, but let’s acknowledge that that is a different question, and it’s not even clear to me why we would want to ask it.

    There are three other important biases which have not, as far as I can see, been mentioned. One is that the species we, as ecologists, tend to work on and encounter most often, are not representative. Most species in any given ecosystem are rare and relatively specialised. But the species we tend to encounter are those which are common and have either been cultivated/domesticated, or have adapted to more anthropogenic landscapes. We even, often, remove rare species from ecological analyses, because sample sizes are too small, increasing this bias further. Then we assume that the patterns we find are general. Younger Mark’s understory herbs might have been more representative than older Mark is giving them credit for (although of course they should not be assumed to be representative).

    The second important bias is normative. We (humans generally) have an anthropocentric bias. That is, things are assumed to be good or bad only to the extent that they affect human interests. Mark’s point in the paper that human-modified landscapes are often better at supporting large numbers of people is, of course, true. From an anthropocentric perspective, it is largely a good thing that we have cleared forests for crops, drained wetlands for cities, and so on (even if we might now recognise that we need to do things differently to maintain stable conditions for continued human thriving on the planet). But from an ecocentric perspective (recognising humans as one species among many, all of which have intrinsic value), it has been largely a bad thing, resulting in losses and declines of many species, for the benefit of a few. Whether one takes an anthropocentric or an ecocentric ethical stance will clearly have an affect on what we define as “good”. If ecologists differ from wider society in being more open to the possibility that we have ethical obligations to other species, and not just to other humans, then maybe we should ask if they have a point. Public opinion polls would be a poor way to settle ethical questions.

    A third bias is that we tend to distinguish greater variation within anthropogenic landscapes than less modified landscapes. We often map “native vegetation” as a uniform shade of green on maps, while distinguishing cropland, pasture, mines, urban areas and so on as different categories. I would contend that in many cases, there is a far greater heterogeneity, at various scales, than we acknowledge within those areas of native vegetation, with great spatial and temporal variation in plant species, vegetation height and type, flooding and fire regimes, dead wood, soil characteristics, etc etc. We are biased against perceiving these variations and tend to see a landscape that has been partially modified by people as more heterogeneous. That might be true in some cases and for some variables, but is likely a perceptual error in other cases for other variables. For a forest species, two areas of “forest” can be more different from each other than a vegetable garden is from a carpark. Again, this is a bias in the opposite direction from those Mark has highlighted.

    That was a longer comment than I had envisaged, but hopefully of interest. I find the question of psychological biases a fascinating and important one.

    * My impression is that Mark is using biodiversity to refer to the number of species, interactions, etc. in a given spatial extent, such as a landscape. I, and think many conservationists, tend to focus above all on the number of species, interactions etc. at a global level. I suspect this subtle distinction already explains a lot.

    • Ben – thanks for taking the time to write such thoughtful comments and pushing the conversation forward. Much appreciated.

      Clearly what constitutes “good” or “bad” is in the eye of the beholder. You are most concerned about global extinctions, and so for you it would be very bad if an endemic island plant goes extinct. Other people hold different values. Some are quite unbothered that that extinction. Others might care more about the island maintaining ecosystems that are productive and diverse and aesthetically pleasing, regardless of whether the species are globally rare or common or where they come from. Local ecosystem function is mentioned very frequently as a conservation goal, and if we accept that humans are present and have modified environmental conditions, then under that selective pressure an assemblage of widespread generalists may well have higher function (pick your measure) than the historical assemblage of endemic specialists (adapted to conditions that no longer exist). This literature very often argues that more species (higher richness) is “better”. In short, one person might care most about global extinctions and find questions of landscape-scale diversity uninteresting and irrelevant, and someone else might think that global extinctions tell us almost nothing about the functioning of local or regional ecosystems, the latter of which is the “real” concern.

      “we are interpreting our results through the lens of what it tells us about biodiversity change globally”. I think that’s fine, but as a minimum, I would say that this should be stated explicitly when it is happening. This is not any kind of universally agreed upon lens through which normative judgements about changes in nature are to be made, but it can easily appear as such if some species in a particular place are just declared to have no value. Using a different normative lens (even an ecological one), those excluded species might be seen to have great value. And since scientists are making the judgement (and they do so often), the normative stance can end up with the appearance of itself being based on science, when it is nothing of the kind. That concerns me.

      “Whether one takes an anthropocentric or an ecocentric ethical stance will clearly have an affect on what we define as “good””. Absolutely! But I think these have gotten seriously muddled. To my eye, a major thrust of arguments justifying conservation over the past couple of decades is that an ecocentric stance *is* an anthropocentric stance. It is easy to find documents – government, NGO, academic – that shift smoothly from “global extinctions are accelerating” to “this will cause compromised ecosystem services in your backyard”. In other words, protecting biodiversity for whatever reason (e.g., a belief in intrinsic value) is what we need to safeguard human well-being (anthropocentric view). And so it appears that science can be used to convert people to our normative stance (or at least to prompt them to support actions that align with our normative stance), which leads to the core concern of my post and article: when undeclared values influence what we present as “facts”.

      • Hi Mark, thanks for your response. I agree with you that the more we can surface hidden assumptions and values, the better. I liked the metric used by Barlow et al. because it seemed to reflect well what I, and I think many others, care about, but (perhaps because of space limitations) the basis for it was not explained as fully as would be desirable.

        I remain unconvinced, however, that species richness is a good metric to use for local ecosystem function. The problem here is that ecosystem function is rarely defined. What function(s), and for whom? One might assume that a dramatic change in richness would have an effect on function, but as we know, those effects are not always predictable and rarely linear. If we are really interested in function, whether is carbon storage, aesthetic value for recreation, or drought resistance, then it would seem far preferable to measure it more directly (wood density and biomass, visitor perceptions, vegetation recovery after drought, etc). Not that the relationships between species richness and different functions aren’t interesting, but there are situations where such functions are driven more by the presence/absence of certain species than by richness per se. Species richness is a worthwhile initial summary statistic, but my suggestion would be that we shouldn’t rely much on it when making decisions.

      • We agree. If one is interested in this or that ecosystem function or service (or a “bundle”), then one can measure it and target it directly. Thanks again for engaging in the discussion…

  10. Pingback: Friday links: Covid-19 vs. BES journals, Charles Darwin board game, and more | Dynamic Ecology

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.