There have been many calls, in this pandemic, to “listen to the science” or “listen to the experts”, which is certainly the right thing to do, but we shouldn't forget that the choice of WHICH (category of) experts to listen to matters immensely. •1/12
The seemingly obvious choice in a pandemic is to turn to physicians and public health specialists, specifically, to epidemiologists. And of course we should, but we must remember in doing so that they will have their inherent biases. •2/12
Epidemiologists are specially qualified to make assessments of the medical effects of epidemics. If we care about other things, like psychological, social or economical effects, they are no longer the obvious choice, for that we should turn to specialists in other fields. •3/12
So if we turn primarily to epidemiologists for advice on dealing with the pandemic, we are making an implicit policy decision, namely to prioritize medical effects over other effects. If this is a deliberate choice, it's defensible, but we should beware of hidden biases. •4/12
By this I mean that epidemiologists will tend to advise in favor of policy decisions which minimize the number of deaths directly or indirectly caused by the disease, because that's their job and specialty, — and disregard deaths by other causes, inc. these very decisions. •5/12
A better way to make decisions would be to ask experts not for advice but for their best estimate of what will happen if we do this-or-that: then we can compare the plausible effects of various policy choices across all domains and try to maximise some utility function. •6/12
Sadly, I feel that epidemiologists (and to a lesser extent, economists) have been given far too much limelight at the expense of psychologists, sociologists and other experts who might have a very different view of what the costs and benefits of various decisions might be. •7/12
And this is true not only in making overall policy choices, but also in making forecasts. Epidemiological models that I've seen make incredibly simplistic assumptions about social behaviors which, it seems to me, need to be reviewed by social scientists. •8/12
One glaring such assumption is that people will behave more or less the same way throughout the pandemic, or rather, that behavior changes happen only in response to public policy decisions (lockdowns, mask mandates, etc.). This seems very wrong to me! •9/12
(One spectacular example of this is the infamous June 2020 ‘Nature’ paper by Neil Ferguson's team, “Estimating the non-pharmaceutical interventions…”, nature.com/articles/s4158… — they ASSUME that R changes ONLY under public intervention in order to assess the effects.) •10/12
And the problem of such assumptions, if wrong, isn't just that they will give bad conclusions, but also that the conclusions will suffer from systematic bias, of the exact same kind as the one you get by listening to epidemiologists for public policy decisions. •11/12
Now there's no such thing as specialists of everything (or of maximising total happiness). But the very least we can do is remember that epidemiologists are not the only experts with a word in this, — and that their word carries an implicit bias. •12/12
• • •
Missing some Tweet in this thread? You can try to
force a refresh
