Are Scientists Really (sometimes) This Dumb?

Short Answer : yes

Evidence? : erm, lived experience (as modern woke jargon would put it)

I’m not a great scientist by any stretch of the imagination, but I’ve had my moments. I’ve been able to make a few half-decent contributions to our knowledge about quantum mechanics and implications thereof. Nothing spectacular, but also not entirely useless either.

There have been, however, far too many times in my life where I think I would have lost a round of Jeopardy to a plate of blancmange.

Sometimes, me being described as merely “dumb” would have represented a step up. I shake my head, look in the mirror, and try to “do better”

My last piece looked at the ideas of Dr Jessica Esquivel who puts forward the (dumb) argument that because quantum particles behave in a way that’s “weird” this means the universe is “queer”. To be fair, this kind of thing has happened ever since the early days of quantum mechanics and even some great physicists turned it all into some kind of quantum woo by occasionally indulging in flights of (Eastern) mystical fancy.

And, yes, I have read The Dancing Wu Li Masters, which is a great example of literal Wu.

Then we’ve had the recent example of the physicist Sean Carroll, who is an extremely smart guy. Most of the time. But he’s been arguing that “science” has shown that sex isn’t binary by appealing to the very small percentage of people with DSD (Differences in Sexual Development). It’s a dumb argument and, of course, “science” has shown no such thing. There are only two sexes in humans - even if our biology (very rarely in this instance) sometimes goes a bit wonky.

This week someone called Michael Osborne has managed to catapult himself into the ranks of “really smart people who occasionally say something really dumb”.

Assuming this account is genuine, Osborne is a professor of machine learning at Oxford. Whilst the h-index is not a perfect measure of academic ‘quality’, Osborne’s h-index is 40 (according to google scholar). This is a very respectable h-index indeed. There would have to be a quite special set of circumstances for someone with this h-index to be a ‘dud’. So I think we’re on very safe grounds if we assume Prof Osborne to be pretty damned good at what he does.

This tweet of his, however, completely baffles me.

Is it meant to be parody? Is it a troll? Is he being serious?

Again assuming that the account is genuine, this is not just some ‘rando’ on the WWW (Web of Woke Wankery), but someone who has reached the pinnacle of his scientific field. Someone who can, legitimately, wear the mantle of “expert” without the ™ symbol.

He uses a personal anecdote to make the case that masks “work”.

You what?

I could, with equal evidential value, say that since covid appeared I have been wearing a St Christopher medallion¹ and I haven’t caught covid yet, so it works. Trust me. I’m a scientist.

What’s even more baffling is that he caught covid back in 2020 and rather than ascribe the “miracle” of not catching covid again to, oh I don’t know, that mystical thing known as immunity, he ascribes it to his mask. It’s not clear from his tweet whether his family masked up too, but the implication from the tweet is that he’s talking about the (very) alleged protective value of his mask, and not masks as source control.

Epictetus makes the point rather beautifully in his² reply.

Epictetus uses the notion of conditional probability to make his point. The idea here is that we have a variable, say C, which is binary. C = 1 means we have covid, C = 0 means we don’t and so we look at P(C|X) where X is the set of all factors that might be relevant.

So, X represents a set of things which might look like
X = {natural immunity, acquired immunity, masks, screens, social distancing, lockdown, vaccination, . . .} and so on

We can pick out one of those things, say masks, and ask
What is the probability that I catch covid given I wear a mask?
And we would (symbolically) write this as P(C|mask)

Epictetus is using conditional probability in a more subtle way than this - although I think he’s got things a bit the wrong way round. In my view, the correct question to ask is the following

Given that I didn’t get covid (C = 0), what is the probability that masking was the explanation?

It’s clear from the data³ that

P(mask as explanation| C = 0 ) << P(other explanation| C = 0 )

In other words, the probability that some other explanation is true, rather than masking, is much higher.

So Epictetus, I think, is making a great argument, but he’s got the analysis a bit muddled (in my view). Smart guy making a (possibly⁴) silly, but easy to make, mistake.

Osborne, however, is making a dumb argument - and everything is wrong about it. Smart guy being dumb.

I’ve often remarked on the mystery of why so many academics appeared to fall hook, line and sinker for the ‘official’ narrative in the face of what, to me, appeared to be stacks of counter-evidence. And not just any old “counter-evidence”, but seriously in-your-face, water is wet, kind of counter-evidence.

Maybe they were just enjoying, as I have so many times, that special moment of dumb privilege.

1. This is almost true. I’ve been wearing my dad’s old St Christopher medallion for nearly 2 years. Nothing to do with covid, of course, but it’s a nice way of remembering him.

2. Epictetus may be a her. No idea. I will adopt the old-fashioned and misogynistic convention of using the masculine pronoun. If you don’t hear from me for a few weeks, it will be because I’ve been incarcerated for this egregious act of misgendering.

3. We knew this, for example, even early on from the data from the Diamond Princess - not everybody caught covid (about 20% did, if I recall correctly). This 20% figure was also confirmed in a large study of asymptomatic infection in households, published in JAMA, in which the secondary transmission rate with symptomatic infection was found to be about 20%

4. I might be wrong - probability is the bane of my existence 😂