Friday, January 29, 2021

There's No Such Thing as "Following the Science"

Ezra Klein quotes a Harvard epidemiologist's criticism of the FDA for blocking rapid at-home Covid tests: "They are inadvertently killing people by not following the science."

I agree with the spirit of the criticism (and was heartened to read that Biden’s surgeon general nominee agrees that the FDA has been "too conservative"), but it's worth clarifying that the FDA's failure here is fundamentally ethical, not scientific.

It's a popular rhetorical move, to present one's preferred policies as being backed by the authority of science.  It immediately puts one's critics on the back foot: who are they to question science, after all?  But it's also misleading.  Science doesn't recommend policies for us to follow, for the simple reason that science merely tells us what is the case, and cannot by itself answer normative questions about what ought to be done.

Whether we realize it or not, we use normative bridging principles to cross the is/ought gap.  If some such principle is implicitly presupposed in a context, it might then seem as though the scientific claim alone suffices to yield a policy recommendation.  Opponents of the policy may then try to undermine our scientific knowledge in order to muddy the waters (cf. climate and covid "skeptics").  Perhaps such silliness could be decried as a failure to "follow the science". But such a framing risks reinforcing the mistaken impression that the science alone determines what should be done.  It's often worth making explicit the underlying normative bridging principles, not least because these are often debatable and worthy of critical scrutiny.

Difficult policy decisions tend to involve trade-offs, whether between different people's interests, different kinds of values, or different levels of risk.  This is what makes them difficult.  Accurate scientific data is crucial for getting clear on what the prospects for various policy options look like: the range of possible outcomes and their probabilities.  But that still leaves open the crucial normative work of evaluating those prospects, or making the difficult trade-offs.

Of course, most of the time when people say "follow the science," what they really mean is, "follow the scientists."  But scientists have no expertise in evaluating trade-offs.  They aren't experts in ethical or rational decision-making.  Again, their expertise merely concerns the descriptive facts, providing the essential inputs to rational decision-making, but not what to do with those inputs.

If you blindly defer to doctors and scientists, the resulting policies will be distorted by whatever implicit normative bridging principles they happen to unreflectively hold.  These are likely to be unduly conservative (since most people suffer from a wide range of conservative biases).  They may oppose challenge trials and other utilitarian policies as "too risky" for the participants, not because they have a more accurate conception of what the risks actually are, but because they lack moral understanding of when risks of that magnitude can be justified.

Don't get me wrong: I like science, and I like scientists.  We should defer to them on empirical questions: what the facts are, and what the risks are.  But they do not have the expertise to evaluate either those facts, or those risks.  If you want to trust and follow the experts, you need more than just doctors and scientists at the table.


  1. There's an important distinction I think the author overlooks. Yes, it's true that policy requires bridging premises to connect the empirical and the normative. In such cases, so long as normative disagreement is reasonable, there can be reasonable disagreement about policy.

    However, there's another class where people make policy recommendations based on empirical claims that contradict a consensus of relevant experts. This has nothing to do with normativity. In these cases, following the science means proposing policies that assume empirical claims that conform with a consensus of relevant experts. In so far as policy proposals depend on empirical claims that contradict a consensus of relevant experts, those policy proposals are *not* following the science--regardless of the normative commitments that underlie them.
    My two bits, anyway!

    1. Yes, as I wrote: empirically misinformed views like climate- and covid denialism "could be decried as a failure to "follow the science". But such a framing risks reinforcing the mistaken impression that the science alone determines what should be done."

    2. This comment has been removed by the author.

    3. I just deleted and reposted my comment below for better exposition

  2. Hello prof. Chappell. If i understand correctly what you wrote, you are casting as "covid-denialism" or "silliness" even views by scientists who hold views re Covid that are not accepted by the majority of relevant experts. Do i understand you correctly?

    1. Just to make my worry a bit more informative: if i understand you correctly, then Dr Anders Tegnell's assessment that generalized mandating of masks does not result in a net "saving of lives" (scare quotes to indicate my agreement with your caveats re the expression "saving lives" (1) ) counts as silly Covid denialism. Another similar worry of mine: Prof John Ioannidis's view that lockdowns do not result in a net saving of lives should count as Covid denialism, if what you are claiming is that mere departure from the scientific consensus counts as Covid denialim (again, i may be misinterpreting your thesis). I oppose this conclusion, I believe Prof Ioannidis re vlockdowns and stay non-committal on the usefulness of generalized mask use, but even if I were persuaded by the scientific consensus I would still oppose the appropriateness of characterizing the views of these 2 scientists as Covid denialism. The term “Covid denialism” carries harsh connotations, no less because it reminds us of climate-change denial, a very serious threat to the well-being of humanity, orders of magnitude bigger than the Covid threat, and, additionally, a denialism whose motivating forces are obviously (I speak with no concrete evidence) emanating from what is called the “Oil lobby”. There is no such lobby worth the name funding “Covid denialism”, and if there is I am open to hear. So I think it is greatly unfair to those scientists who are sceptical regarding the empirical facts surrounding Covid and those who are persuaded by them to be dismissed as silly, or as Covid deniers, as if we are disputing the existence of the virus.


    2. Just to clarify an unfortunate expression of mine:it is Climate-change that i consider to be a very serious threat to the well being of humanity, not its denial. I do believe we can overcome the threat without shutting down bad-faith deniers.

    3. Hi Dionissis, I agree, and definitely don't mean to rule out reasonable scientific disagreements about the efficacy of different interventions. By "Covid denialist" I meant the kooky conspiracy theorists who literally deny that Covid is real, or assert that Bill Gates has slipped brainwashing devices into vaccines, or Trumpian claims that the virus would soon magically disappear, etc.

    4. I can see that my response to Amitabha was misleading, in that they instead suggested that "following the science means proposing policies that assume empirical claims that conform with a consensus of relevant experts." I don't think that scientists who hold minority views are necessarily failing to "follow the science" -- that obviously depends on the substantive quality of their scientific reasoning and evidence.

      But Amitabha might instead be read as claiming that laypeople (meaning: anyone who hasn't read the relevant studies and/or isn't capable of assessing the scientific merits for themselves) should generally defer to the expert consensus rather than cherry-picking a maverick expert whose views they happen to like more. And that would seem right to me, as a general rule of thumb. (But I wouldn't apply the term 'denialist' in such cases, I agree that that is too harsh, and best reserved for the real kooks.)

    5. PART 1

      Hello again Prof Chappell.
      Yes, it was your response to Prof Palmer that guided my interpretation of your view. Anyway, I wanted to respond with some questions of mine as soon as you posted your reply regarding what is the rational thing for laypeople to do with respect to deference to expert consensus in the presence of a dissenting minority of prima facie credible scientists, but I realized how theoretically impoverished my thinking is on epistemological matters. I haven’t read anything on the epistemology of testimony, or the epistemology of disagreement, or rational decision-making, so I found myself incapable of making analytically-philosophical sense when I was trying to put my thoughts on paper, I couldn’t even “collect” my numerous thoughts –this, I am guessing, must have been due to my lack of relevant epistemological concepts (suggestions for introductory papers or books welcome, in case a philosopher is reading this comment of mine. And, as luck would have it, Prof Palmer, who made the first comment on this thread, has tons of material on their site that is relevant to another interest I developed recently, namely philosophically thinking about certain issues relevant to health (1) ). I should add that I belong to the class of laypeople that are “not capable of assessing the scientific merits for themselves” because sadly I am statistically illiterate (I have only taken Stats 1 & Stats 2 at College, yes I know what the area under the normal curve represents and what conditional probability is, but when it comes to evaluating study designs I hope there are scientists or statistically literate journalists who can inform me, I cannot do it myself) so I am one of those who you Prof Chappell sound like saying that they have (pro tanto?) epistemic duty to defer to the scientific consensus, never mind the presence of a dissenting minority of scientists. More specifically, here is how I understand what I take you to be saying in your alternative interpretation of Prof Palmer’s comment:

      If the Agent tentatively rejects the scientific consensus, favoring instead a minority view of other scientists, then she is acting irrationally if she herself does not have the capacity to epistemically evaluate the merits of the conflicting theories.

    6. PART 2

      Now, I can think of counterexamples from the real-life situation we are facing with Covid: if the Agent has good reason not to trust that the experts who shaped the scientific consensus really intend to disseminate the whole truth and nothing but the truth when they utter sentences regarding Covid, then the Agent is relieved from her duty to follow the scientific consensus. I should add here that such defective scientific testimony (i.e. scientists knowingly uttering false or misleading sentences) doesn’t need to be motivated by nefarious designs; they may be motivated by a sincere wish to promote what they take to be the common good. This view has been clearly expressed (and normatively endorsed) by your interlocutor at Daily Nous, commenter with the screen name Avalonian ( judging from the quality of his comments he is a Prof of philosophy):

      Commenter Avalonian had said at Daily Nous in his discussion with Prof Chappell (emphasis added):

      “As a related aside re:(6): the CDC and the WHO have to worry enormously about the perlocutionary effects of what they say given the way people actually are. Saying that there is “no evidence” for something might be a way of encouraging caution in the populace, and not a literal report of what is strictly believed by the relevant scientists. This, again, is the sort of real-world constraint these people [CDC, WHO] are under which is ignored by armchair analysis: if much of the populace is strongly inclined to think “ah, I’m sure this isn’t that bad”, anything you can do to move *them* towards true belief is a good thing by the lights of social policy.”

      Note to irregular readers of this blog: Prof Chappell responded to commenter Avalonian not just at Daily Nous but also here:

    7. PART 3

      I should add in passing that what commenter Avalonian considers (in light of his hypothesis that the current populations were inclined to see Covid as a non-serious threat) to be the best way for the medical authorities to behave (i.e. to move people towards good behaviours even via inaccurate scientific statements, so as to make people adhere in effect to a hypothetically true belief that Covid is a very serious threat (2) ) could easily backfire, because if some laypeople (hypothetically)realize that the medical authorities are indeed not saying the truth when they speak, then these laypeople have perfectly good reason to dismiss the scientific claims of the scientific consensus expressed by the medical authorities (3), even if the laypeople believe that the authorities have the best intentions and are (the authorities) even acting under the most wise public health plan – a wisdom that supposedly dictates the utterance of misleading scientific statements.
      All I am trying to say, Prof Chappell, is that if Avalonian’s views are resembling the medical establishment’s practices, then I am not irrational in believing the dissenting minority’s views re the IFR of the virus, instead of the consensus IFR. After all, ex hypothesi, the medical estasblishment may be playing fast & loose with accuracy in making statements, in order to promote the common good. If I believed this, and if a I also believed that the medical establishment’s implemented good intentions are indeed promoting the common good, then I would be irrational to oppose the utterance of the sentences (the speech act), because I too want the common good to be promoted, but I would still be rational in secretly disbelieving the content of the propositions expressed in the relevant speech acts – it would be rational to believe the dissenting minority, but I should stay silent for the sake of the promotion of the common good that ex hypothesi is better served by the uttering of inaccurate scientific statements by the medical establishment. Practically untestable (and irrelevant for the purposes of our discussion) empirical claim of mine: uttering only truths in the context of public health is the best policy in the long run – barring situations where humanity faces extinction, or something that grave.

    8. PART 4

      In real life, Dr Fauci stated publicly fairly recently at the New York Times that at least one time he wasn’t telling the whole truth because the public was not “ready to hear what he [Dr Fauci] really thinks”. A similar real-life example that, in my mind, warrants disbelief in the content of some propositions expressed by the medical authorities is the UK policy of using “hard-hitting emotional messaging” in order to “increase the perceived level of personal threat” from Covid in the eyes of complacent people. Why do I say this? After all, the strategy of casting the risks in an intonated way, so as to make the “complacent” people see the magnitude of risk as the UK authorities sincerely believe it to be, is not disseminating any false descriptive proposition, it only disseminates a normative proposition to the effect “a risk of numerical magnitude x warrants this or that emotional and behavioral reaction”. But I am strongly inclined to believe that authorities who don’t hesitate to bang their hand on the table in order to elicit the appropriate emotional response of the complacent, while there are alternatives to banging their hand on the table(4), are likely to use descriptive falsities too in their quest to make the complacent comply. And anyway, the warranted attitude towards the risk of Covid that the authorities have been trying to instill seems to be one that not even the authorities themselves considered warranted on a personal level: UK’s famous Prof Ferguson, whose model led us to lockdowns, violated the lockdowns himself – and, even worse, in his defense he made statements tacitly presupposing the truth that natural infection confers immunity for some time, which is true but it was a truth that was cast as uncertain at the time by the medical establishment even though it was a very natural assumption based on the behavior of similar past viruses. In other words, either Prof Ferguson, contra his public statements, did not consider the risk of Covid as warranting serious fear, or if he did consider it serious then he refrained from telling the public that they need not be afraid once they have caught the virus and recovered. In the US, I lost the count how many politicians (who have epistemic access to top level medical information) violated their own injunctions re masks or lockdowns – meaning that they themselves did not consider Covid such a serious threat as to warrant the level of mask-wearing they suggested to the public. Am I irrational to believe that the content of the statements of the dissenting minority of the scientists who didn’t make that much of mask use are closer to the truth than the statements of the scientific consensus?

      Let me repeat that one needs not doubt the goodness of the intentions of the medical establishment, I actually do believe that some of them were definitely moved by the de dicto intention to promote the common good (I have serious disagreement though as to whether they promoted it de re, though of course from their point of view their intention seemed de re. But let’s stipulate for the sake of argument that what they achieved in terms of public health was indeed the best possible outcomes under the circumstances). What I would like to claim (defeasibly, and I am open to hear conflicting views and allow myself to “take them in”) is the claim that I am not irrational in believing the dissenting minority of scientists, partly for the reasons I stated (forfeiture of the medical authorities’ credibility in utterances re Covid). And that this is so in spite of my regrettable statistical illiteracy. But I will await Prof Chappell’s take on this claim, if he finds the time.

    9. PART 5

      Prof Chappell, the thought occurred to me, while I was writing, that on the subject of the handling of Covid you are standing in a notional middle ground while a collective agent similar to the ancient Procrustes (5) is pulling you on one side towards Conservatism in Public Health policy (that’s the part of the collective agent represented by commenter Avalonian) while the other part of the collective agent (me) is pulling you towards the other side, hardcore anarchy. 😊 I invite the readers to disregard appearances and see me as the middle ground, and Prof Chappell as a little Conservative.(6) 😊 😊

    10. FOOTNOTES (1)-(3)


      2) i am not denying that Covid is a serious threat. I deny that it is as serious as to justify the (inneffective, some scientists say) lockdowns even while there is the alternative of shielding only those who are afraid, without incurring the lockdown harms. (6) And obviously I deny that Covid is the biggest threat that humanity has faced since WW2, as many political leaders and others said. Those who make the claim disregard the annual deaths from smoking, or from cardiovascular diseases, or from cancers. We have seen from the Swedish example that non-coercive relatively laissez-faire policies a la Sweden do not lead to a number of Covid deaths exceeding deaths from cardios. Has anyone batted an eyelid upon learning how many people on average die in the US each year from cardiovascular diseases? Come and think of it, why did the medical authorities leave unexploited the opportunity of trying to instill in people good dietary habits? One would have thought that a policy that aims at “saving lives at any cost” would have gladly endorsed costless policies that can save lots of lives: there are millions of preventable deaths each year due to unhealthy nutritional habits, why didn’t the medical authorities try to exploit people’s fears of Covid by systematic messages that lots of fruits and veggies arm your immune system and help you, in the long run, against the Covid threat? The omission has been spotted by experts like Dr David Katz, and Prof Francois Balloux (the former disagrees with the Swedish response, I don’t know about the latter):
      Dr Katz:

      Katz thinks there should be a national health-promotion campaign to encourage people to improve their diets and lifestyle before they become infected.

      The opportunity to attenuate the pandemic toll by means of a national health promotion effort cannot be overstated, yet with rare exception has barely been stated at all. Now is not a moment too soon to fix that.

      (i have the exact tweet, but I can’t find it now, he clearly stated that in some or other high-level meeting on Covid he tried to put in a word for healthy nutrition but this particular view of his was ignored)


      Harvard Professor Martin Kulldorff stating the obvious, namely that the medical authorities shot their credibility in the foot.

      Prof Francois Balloux on the lack of wisdom to allow truth-telling be sacrificed for a common good that will end up not being the common good.

    11. FOOTNOTES (4)-(7)

      4) an ultraminimal sketch of my proposal is in footnote (6)


      6) so what would I propose? Something like this:

      It is similar to the Barrington Declaration. My own proposal (to which I had arrived independently long before we heard about the Barrington Declaration –which I happily support, anything is preferable to the current status quo—and long before Mr Sayers suggested his Traffic Lane plan, shields not only the vulnerable, but also those who ask to be shielded (i.e. shields also non-vulnerables who declare they are afraid of Covid and want to be shielded). Public space is to be temporally segmented for use by only one of each of 3 tribes, one tribe at a time. The tribes are: those who ask to be shielded and do not belong to an intergenerational family (Red Identities), those intergenerational families that collectively ask to be protected (Blue IDs, the alternative option being to belong collectively to the last tribe, the Green IDs), and those like me (Green IDs) who don’t want to be protected, and are ceding our priority for a hospital bed in case we catch Covid and the hospitals are overwhelmed and there is a triage situation. All that people in my tribe are asking is to be left free to go on with our lives (using the public space at the time that is allocated for us) without ever coming in touch with those who want to be shielded (ergo, we pose zero risk to them). But in exchange we want our freedom to hug and kiss and caress all the consenting members of our Green tribe. I am in effect saying that the lives of the other tribes should matter more than my tribe’s lives in the deliberations of the decision maker of the authorities in case there is a need to choose between the one or the other: please don’t mourn our deaths if –contrary to all the indications re the IFR of Covid-- it turns out that we were indeed reckless. All lives are inherently equal, i.e. deserving of equal concern, but when some lives are telling you “please do not value our lives in the way you want, let us value our lives in the way we want”, I believe the rest of society is under no obligation to try to “save” us from our purportedly reckless ways. All we owe you is to shield you. We do not owe to you to let you discharge what you think is your duty to protect us. We don’t want to be protected by you, we will take care of ourselves. In the eyes of the medical authorities my tribe, whose members I guess include Lord Sumption (7), should count for less, because we ourselves are asking the authorities to count us for less, in the sense that if there is a triage situation we consent that priority for saving be given to the other IDs (Red or Blue) and not to us (Green IDs). Dr Devi Shridhar’s life should take priority over Lord Sumption’s life in case there is a triage situation where both caught Covid. That’s another way of reading Lord Sumption’s statement regarding which lives matter more, a statement that led Dr Shridhar to attack him – I am not implying Lord Sumption meant it the way I am reinterpreting him. Those like me who voluntarily assume the risk of dying from Covid lose their priority for being protected. I think it is the lack of this particular moral intuition that led to the current predicament we as a society face with Covid. We didn’t err on the side of caution, we erred on the side of Paternalism – I find it a bit terrifying that the younger among our top level brightest are so drawn to Paternalism.


    12. Hi Dionissis, I'm sympathetic to much of what you say. In particular, you're right that general norms of deference can only apply to sincere assertions of belief, so if you've good reason to think that some particular public pronouncement is insincere or misrepresenting the actual scientific consensus, that would undermine the grounds for deferring to its contents.

      I also agree that it's bad strategy for public health experts to undermine their credibility by lying or misleading the public. (Their early disparagement of masks would seem another example of this.)

      That said, I don't think that wholesale disregard of mainstream views on the pandemic would be justified on these grounds. For example, it would make no sense for public health experts to manipulate us into supporting lockdowns if they didn't regard the latter as an effective tool for slowing the spread of the virus. So it seems pretty safe to assume that the expert consensus really does regard lockdowns as efficacious -- in which case we layfolk are in no position to dispute that. That doesn't commit us to supporting lockdowns all things considered, however, because (as I emphasize in my post) we needn't defer to public health experts' views about what risks and tradeoffs are warranted. (They aren't experts on those topics.)

      Your suggestion for temporally segmenting public spaces during a pandemic sounds very sensible to me.

    13. This comment has been removed by the author.

    14. This comment has been removed by the author.

    15. I am reposting the comment i deleted, there were messed up sentences in it

    16. Prof Chappell yes, I do not doubt that the medical establishment sincerely believes that the lockdowns slow the spread. If I gave the impression that I am doubting every single claim made by the medical establishment I regret it. I thought I had cautioned the reader by stating in PART 4 that “A similar real-life example that, in my mind, warrants disbelief in the content of some propositions expressed by the medical authorities”/emphasis added. If i misled you on lockdowns in particular, I hope my subsequent sentences here clarify that I was not doubting that lockdowns slow the spread, or that I believe that the medical establishment was insincere in claiming lockdowns do slow it. That lockdowns slow the spread via reducing transmission from person to person is so common sense that it is close to a necessary truth (1) – if no person comes in touch with another person, how can the virus jump from one to another? But if the virus cannot find enough hosts because we are under lockdown, then the spread will slow down eventually. In fact, no lockdown-skeptical scientist that I follow, and certainly none from the leading scientists of the lockdown-skeptical opposition, has claimed that lockdowns do not slow the spread. On the contrary, they take for granted that lockdowns slow the spread and their criticisms were directed either to the long run futility of slowing the spread in this way (a belief they held initially on account of their then-held view that the populations could not be fully vaccinated within at least 18 months, and the corresponding conclusion that lockdowns are merely delaying the inevitable in-the-long-run deaths (2) ), or on the broader thesis that the lockdowns cause so much collateral damage that whatever gains in lives/life years the slowing of the spread may offer (gains eventuated either by saving people who would have died due to hospital-overwhelming if lockdowns were not instituted, or by shielding all people until vaccines arrived, some of which would otherwise have died if exposed but who will be saved by the vaccines) it is not rational to sacrifice a greater number of lives/life years of people who will die because of the lockdowns in order to save now ,by slowing the spread via lockdowns, a smaller number of lives/life years. Characteristic quotes of every single of those scientists such as Prof John Ioannidis, the first lockdown sceptic who showed up, and who supported the initial lockdown until we gathered enough evidence, and opposed lockdowns after he came to the conclusion that there was evidence that the IFR of the virus did not justify blanket lockdowns (he suggested shielding the vulnerables), Prof Battacharrya, one of the 3 architects of the Barrington Declaration, Dr Giesecke, one of the two architects of the Swedish response to the virus) forthcoming when I find a little time. So nobody from the mainstream heavy-weight sceptics base their skepticism on the view that lockdowns do not slow the spread.

      1 It is not a necessary truth of course, I am just exaggerating the obviousness of the truth that lockdowns slow the spread. The lack of necessity in the proposition can be shown by imagining a situation where every single person is isolated alone from safe distance from each other person and yet there is no slowing of transmission because a plane is spraying each one from above with virus particles, as a hypothetical variant of the ultra-silly so-called chemtrail conspiracy theory would have it (2). Therefore, since we can imagine a situation where we have perfect lockdown and yet can imagine the absence of slowing of the spread, we can conclude that the proposition lockdowns slow the spread is not a necessary tuth.
      This is the most famous conspiracy theory in Greece


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.