Relying on just one model to say that heterogeneity doesn’t really make any meaningful difference is not how you do “science”, that’s how you do bad science, which is not the same thing. If epidemiologists are going to opine on policy, which to be clear I hope they do because they’re still the people who know the most about this, their advice should obviously not rely on any particular model. This is not what people in the climate modeling community, who know a lot about the interface between modeling and policy, have been doing. Obviously, I’m not expecting epidemiologists to put together something like the IPCC report in just a few months, which takes years and a lot of resources. However, I am expecting them not to misrepresent how much uncertainty there is about “herd immunity”, which unfortunately I think many of them have done.
It’s okay if Bergstrom and Dean personally oppose the “herd immunity” strategy and want to make a case against it. They are scientists, but they’re also citizens, who are entitled to defending their views. But since their credentials mean that many people are going to take very seriously anything they say on the issue, they have a responsibility not only to accurately represent the state of knowledge in their field and in particular how much uncertainty there is, but also to clearly distinguish between scientific claims they make because it’s the consensus in their field, scientific claims they make because, although it’s disputed in their field, they find them compelling and finally claims that are not scientific but political. When Dean says that pursuing the “herd immunity” strategy is “insane”, she is not making a scientific claim, she is making a political statement. Again, she is entitled to this political opinion, but she should make clear that it’s what it is, because otherwise a lot of people won’t understand it and will assume that it’s what “science” says.
I dated Lemoine's post because I want it to be clear he is intervening in unfolding debates in real time. I start with four disclaimers: (i) I am pretty confident that he and I disagree on a lot of political issues; (ii) in my one interaction with him -- when I asked for the data on one of his pieces -- he was unfailingly prompt and forthcoming; (iii) his post is argued forcefully, even polemically, because quite a few claims and policies put forward in the public domain by experts seem to be based on flimsy evidence while the stake are high and the consequences on other people's lives grave. (iv) I am genuinely unsure about what policy ought to be pursued. I want to develop his themes, because I think, inspired by Kofi Bright and Bradley, trained philosophers of science can help the public and even some scientists think about how to think and talk about the issues in public. As it happens I'll disagree in tone and emphasis, but I think his way of putting things deserve a wide hearing. (So, go read his post.)
In the final paragraph (quoted above) Lemoine slides from the duties of scientists to the duties of a sub-class of scientists who are especially prominent ("who are followed by a lot of people on social media and who have access to prestigious outlets such as the New York Times to air their opinion.") Together with Merel Lefevere, I call the latter "aggregators," anticipated by Michael Polanyi (1941), who calls them"influentials." As Eric Winsberg and I noted, some aggregators work for the government, or industry, and so they may well serve more masters. Aggregators have special responsibilities both to the scientific community they represent and to the public they inform and shape.
Before I get to that, one aside that creates a complication. In general, and optimally, aggregators have high standing within the scientific community. Of course, in times of crisis, scientists who write/communicate well or who are especially good may well generate a public following even though they lack high standing in their own scientific community. Even short of crisis this can happen. For example, some heterodox economists manage to be quite influential while being somewhat marginal in the field. I don't think that is a bad thing, and what I say also applies to them.
In general, I agree with the thrust of Lemoine's concerns, but unlike him I think there are a class of important utterances an aggregator can make that are not easily distinguished as 'political' or 'scientific' (even though there will be some claims that are clearly political and some that are clearly not political at all). I don't say that for postmodern reasons (involving denial of facts and objectivity) or because I wish to undermine the public status of science. Rather, the way many human/policy sciences work, they have internalized some important or fundamental commitment into their models and that commitment becomes completely uncontroversial within the models and their standard applications. So, for example, in medicine there are all kinds of non-trivial normative commitments that health is a good thing, and that there is a kind of proper (statistical) functioning that should be pursued. In economists certain things are taken to be good in the aggregate: more income is better than less, less unemployment is better than high unemployment, efficiency is...etc. Many scientific models aggregate in ways that can have non-trivial politically salient consequences and yet are utterly routine within a science. In most circumstances that's unproblematic, and practices develop to prevent the the downsides of modeling assumptions to cause trouble outside their proper domain.
So, evaluating aggregators dividing their statements into political and scientific is unfruitful. Before we treat them as political or partisan advocates, they and the rest of us need to become clear on what socially salient ends they pursue in modeling and, indeed, whether this is standard or not. This is one reason why Winsberg and I advocated transparency about modeling trade-offs.+ So, for example, modeling with, say, a homogeneous population and averages can be a way to make policy seem (even be) neutral and the science unbiased. But when the experts and some in the public cotton on to the fact that the averages disguise radically different outcome patterns, then working with homogenous agents and averages may well appear as highly politicized even when it is the normal thing to do in a scientific community. So, when policy economists work with family averages or ignore distributional effects, they may be promoting outcomes that can seem grounded in scientific consensus, and yet have non-trivial social consequences that some may view unfavorably.
In the circumstances of a pandemic, working with averages and homogeneous agents, may well (unintentionally) mask that pursuing some policies may foreseeable entail life or death, disability, or bankruptcy (etc.) for some groups or regions (etc). In such cases people affected seem to have a right to know that policy will be pursued that may impact them more greatly than others. It is far more important to hear about this than the uncertainty involved in any policy or debates over which model is more advanced.* Once we it put like that we can immediately discern that aggregators who work as policy advisors may well face adverse incentives that will nudge them into presenting policy in relatively anodyne ways sometimes even for politically sensible reasons (e.g., to prevent panic, to buy time to have some agencies prepare for the worst, to get policy done, etc.).
So, while I fully agree with Lemoine that one of the roles of aggregator is "to give citizens and decision-makers the tools they need to make informed decisions" that's not the only role and it may not be the most vital role. Sometimes it's morally more important to alert some groups of citizens (and their advocates) that their interests are being disproportionately harmed or affected while being in no position to be informed decision makers. I don't say this because I am a skeptical about the possibility of informed citizenry; there is a good evidence that in a crisis, when people are afraid—a natural response during a pandemic—, they become more politically attentive and seek out high quality information. Rather, and to put it like this is to face up to nature of political reality: in a crisis the vast majority of citizens are in no sense decision makers, they/we are impacted by decisions. And so, if we tell aggregators to act as if the polity (already) is a Habermasian deliberative democracy, we may (as Iris Marion Young taught) end up harming the most vulnerable.
It should be clear that in the previous paragraph I am privileging a version of what we may dub the most affected principle combined with a commitment to protecting the most vulnerable. You may wish to disagree with this!
Often aggregators are not quite aware that the ends they have scientifically internalized, and perhaps are communicating to the public as consensus (legitimately so), may have disproportionate impact on some groups. So, I have argued with Lefevere it is duty of aggregators, unlike the members of a scientific community, to investigate the possibility of such downstream effects. This is why, unlike Lemoine, I often think the reporting of scientific consensus (while welcome) about the family of models standardly used in a field is a lower priority than finding ways to help aggregators, policy advisors, and aggregators from adjoining scientists to learn about the effects of policy along dimensions that are not being modeled by that community at all. Sometimes that only becomes clear when people are allowed to ask tough questions in the public domain. I think synthetic philosophers can play a role in generating and facilitating such conversations, but about that some other time more.
*Recall Lemoine's "only to accurately represent the state of knowledge in their field and in particular how much uncertainty there is, but also to clearly distinguish between scientific claims they make because it’s the consensus in their field, scientific claims they make because, although it’s disputed in their field, they find them compelling and finally claims that are not scientific but political."
+Notice, that's not saying that transparency is the most important thing in public science. Sometimes a lot of transparency is also a device to prevent people from understanding what is truly significant.
"Non-trivial normative commitments that health is a good thing" are kind of expected of public health practitioners and medical researchers...
Posted by: David Duffy | 05/08/2020 at 01:02 AM