This is a guest post by Dene Bebbington.
I would add Reckoning With Risk by Gerd Gigerenzer (who’s been sensible on Covid-19) to the list.
In a 2007 study women were asked what age put them most at risk of getting breast cancer. Less than 1% gave the correct answer that women over 80 have the greatest risk, and many thought age wasn’t a factor. Yet age is the biggest risk factor for breast cancer and cancer as a whole, although not for some specific cancers.
Unfortunately, human brains are not naturally good at, and suited to, understanding statistics and relative risk. A raft of cognitive biases which may have been advantageous over evolutionary history damage our ability to think rationally about many subjects. Many aspects of the modern world requires the use of head more than gut, whereas the opposite was probably true in our ancestral past.
Perhaps the most pernicious cognitive bias at play in the Covid-19 pandemic is availability bias. This makes things which easily come to mind seem representative of reality when they’re anything but. What we can recall quickly and vividly is likely to be from the media’s saturation coverage of the pandemic. Typically the more sensational a subject the more time and effort the media will put into reporting it. Just think how a terrorist attack which kills a few people is reported compared to the thousands of deaths from accidents which go unreported unless they’re of a sensational nature. Most suicides go unnoticed unless it happens to be a “celebrity” or person in the public eye who has killed themself.
The media frenzy in reporting Covid-19 statistics, and endless stories of people becoming ill from it, does inform us to an extent, but thanks to availability bias it skews our understanding of relative risk. The annual flu is hardly ever reported, and that’s perhaps why there are idiots calling for indefinite lockdown to tackle Covid-19 even though they may unknowingly cause a flu death in their lifetime through socialising or poor hygiene.
Cognitive biases are so hardwired in us that some are almost impossible to overcome. However, being more aware of biases helps to identify their presence in the assertions and arguments of others.
The good news is that there are excellent books worth reading to help us assess relative risk and identify biases. You’ll learn more important things from reading those books than watching the TV news. Sadly, I suspect that the people most in fear of Covid-19 who are spreading their fear on social media are also the least likely to read relevant books – and the least able to understand them.
Below are my book recommendations. Please list yours in the comments.
The Norm Chronicles by Michael Blastland and David Spiegelhalter
This is a book that should be required reading at school since 16 years should be able to get the gist of it. The authors use a fictional average man called Norm, and a woman called Prudence, to explain chance and risk. They use the concept of a MicroMort (1 in a million chance of death) to put risks into context.
Risk by Dan Gardner
Subtitled “The Science and Politics of Fear”, this is another great look at relative risk and how people are often unable to make sound risk judgements. It’s not just the media who are at fault for exacerbating availability bias, we may believe that something we witnessed or experienced is likely to be a bigger future danger than it really is. Fear is a powerful emotion that clouds our thinking.
Thinking, Fast and Slow by Daniel Kahneman
The hardest read of the books on my list, this one unpacks cognitive biases and is sobering in what it reveals about our intellectual limitations. Kahneman uses the concept of type 1 (fast and intuitive) and type 2 (conscious and methodical requiring effort) thinking and shows how type 1 thinking leads us astray.
Future Babble by Dan Gardner
Essentially this is a popular presentation of Philip Tetlock’s findings about the poor record of experts’ predictions (read Tetlock’s academic book Expert Political Judgement for the detailed and nuanced results of his work). Gardner shows how our brains prefer certainty and that cognitive bias rationalises failed predictions.
Predictably Irrational by Dan Ariely
A more accessible and readable look at cognitive biases than Kahneman’s book. The experiments reported are insightful and further exemplify what a powerful yet strange system the human brain is.
17 thoughts on “Dene Bebbington: Pandemic reading recommendations”
Thinking fast and slow is unfortunately rather outdated as some of the material has got caught up in academic psychology’s “replication crisis”, particularly the stuff about priming. The ideas and work of Kahneman and Tversky ought to be far wider known among the general public but I don’t think TF&S, for all its success as a pop-science book in terms of readership, is presently the best place to go for it. It deserves a rewrite and a fresh edition really.
It is not obvious that the problem lies in our brains (with all their admitted cognitive biases) rather than with the information we feed into our brains. That is, the problem with the over-reaction to C-19 lies in the data that people have (or rather, the lack of data), not the inability of our brains to make appropriate risk assessments about it.
In the US, on average 7,800 people die on a normal day. That piece of data would give people some context to view breathless CNN reporting. Flu killed 85,000 people in the US in 2018 — and, to be brutally honest, most of us did not even notice. And flu is typically only the 8th largest cause of death in the US.
Give our brains context and data, and they will work fine. It is understandable that Jane Public listening to the BBC will be seriously misinformed. But it is disappointing that the likes of Boris Johnson and his advisers act as if they are equally uninformed. That is where the real problem lies, not in our brains.
The Gigerenzer book is triff and brill. Everybody should read it anyway, plague or no.
And MBE is right about Daniel Kahneman: in the old days we’d have said that much of his work has been “exploded”.
Kahneman has even acknowledged the irony of his mistake in including some of those studies in TF&S: https://replicationindex.com/2017/02/02/reconstruction-of-a-train-wreck-how-priming-research-went-of-the-rails/comment-page-1/#comment-1454
But it would be better if he gave the thing a pretty brutal rewrite, it still seems to shift a surprising number of copies for a book even whose author accepts it is deeply flawed.
Cognitive biases can be exacerbated by what information we take in, but they do seem to be a real issue in how our brains work. I did a quick search and found this paper looking at availability bias affecting medical diagnoses:
https://qualitysafety.bmj.com/content/early/2020/01/27/bmjqs-2019-010079
Thanks for posting this. It’s made me think I should also recommend Black Box Thinking by Matthew Syed which talks about learning from mistakes!
“learning from mistakes!”
In an ideal world, learning from other people’s mistakes.
Individuals may learn, but it is difficult for organizations to learn. It is common in industry for organizations to look back at major projects through processes known as Post-Project Appraisals, Retrospective Reviews, etc. Depending on the relationship between the people who were responsible for the project and those responsible for the review, these tend to become either fluff pieces praising the project team or vicious backstabbing exercises. Learning takes a back seat to intra-organization politics.
We can expect that when the Covid Panic has passed, there will be scholarly articles citing the millions of lives saved by the lockdowns and others citing the millions of unnecessary deaths, depending on the politics of the authors. Not much of this will help improve the response the next time.
Look back at the Great Depression of the 1930s where academics are still arguing about whether FDR’s New Deal helped or hurt. If we can’t reach consensus on something like that after over 8 decades and when every involved decision-maker is dead, we have to be realistic about our ability to learn lessons from the C-19 over-reaction.
Here is my suggestion;
How to Change Your Mind: What the New Science of Psychedelics Teaches Us About Consciousness, Dying
Michael Pollan
This was my first introduction into ‘thinking’. I found this after trying to understand why a project I worked on for a couple of years turned into such a debacle.
It’s a speech given by Charlie Munger, Warren Buffet’s 2IC.
The Psychology of Human Misjudgment, by Charlie Munger
https://fs.blog/great-talks/psychology-human-misjudgment/
Agree completely. When this started my first alarm was triggered by the breathless reporting of numbers that provided no context at all.
That was followed with
Exponential! *squawk*
Exponential! *squawk*
Exponential! *squawk*
Exponential! *squawk*
It was almost designed to generate panic.
I’ve been reading a tremendous amount of behavioural and clinical psychology recently. Tangentially related to the replication crisis, it strikes me just how much is thought to be incontrovertibly known on the basis of “win or lose a dollar”-type experiments with American college freshmen.
In addition, the way people treat each other in interactions is very dependent on the expected number of iterations of those interactions, something not adequately addressed in most psychology experiments. The iteration thing of course explains why lawyers and estate agents (apologies to any who are reading*) are so dreadful.
This is directly relevant to Covid because we are being led to believe our interaction with it is a one-off that justifies once-off measures.
*: Actually, not. Especially not German estate agents or German property litigation “specialists”.
Incidentally understanding of the small-study effect has almost 100% penetrance in my field of clinical research. It’s the main reason so many drugs fail at phase 3*, the first time you do studies fully powered for efficacy having seen some “promising signs”, which aren’t even usually statistically significant, in an unpowered study. Maybe we could help the psychologists with their study designs.
* and one of the reasons I come down so hard on crap like hydroxychloroquine.
“We can expect that when the Covid Panic has passed, there will be scholarly articles citing the millions of lives saved by the lockdowns and others citing the millions of unnecessary deaths”
This thing is so infectious that there isn’t really a way to stop most people in the world from getting it eventually. At least no way to do it without utterly destroying the economy. In the west, that would probably blast us back to 1950s standards of living plus cellphones. In Africa it would kill hundreds of millions.
This leaves us with “keep the ICU quiet” as the only rational defence of the lockdown, but we also know that around 90% of patients who need ICU for this die (for background, typically about 20% of patients ever admitted to ICU in normal times will die there).
As it’s also substantially less deadly than being claimed, we will also know, once almost everyone has had it, just how many people did die of it. And can calculate in hindsight whether the societal and economic devastation was worth something that’s about 2-3 times as deadly as ‘flu* (on the background of 2 years of mild flus).
I don’t believe the lockdown proponents will have many hiding places in a year from now, the time at which it will have burned itself out or vanished into the background of perennial respiratory bugs, but we will be suffering peak economic consequences. It will take a lot of spin for the politicians to extricate themselves from that one.
*: My expectation, not necessarily that of anyone else. We will see.
I am yet to be convinced that Boris’s advice is not an agenda.
I once had the misfortune to work on a project run under the Scrum method even though that was inappropriate and the Scrum Master was clueless in many ways. We had retrospectives, but the one lesson they wouldn’t countenance was that Scrum was not suitable for that project. It took months and eventually the involvement of other people until the Scrum Master’s eyes were opened and we moved to a Kanban style approach. I had many WTF and head-banging moments during those months.
Other times in IT we’d have lessons learnt sessions and still, for whatever reason, someone would slip back into a previous way of doing things.
Surely a real classic is Mackay: Extraordinary Popular Delusions and the Madness of Crowds. Free online here, including for Kindle
http://www.gutenberg.org/ebooks/24518
The section on Prophecies has description of the response to several epidemics in history
http://www.gutenberg.org/files/24518/24518-h/24518-h.htm#prophecies
For what it’s worth our experience of useless solicitors occurred only in England. Our experience in Scotland was far better – abler people, and a brisker, more competent legal culture. Less expensive too.
Whether a devolved government can ruin that culture remains to be seen.
Comments are closed.