I’ve been meaning to read this for a while, and happened to pick it up this week — just a week after doing a course on pharmacoepidemiology in the age of COVID, which involved a lot of discussion of how to evaluate papers, and responsible study design. I’m also studying biostatistics and epidemiology this year, of course, meaning that I understand more about statistics than I care to — which means, all in all, that this book slotted in admirably, though quite without meaning it.
The issue the book discusses is a serious one: through the current system of “publish or perish” and the way grants are awarded, tenure is granted, etc, poor science is becoming the way to do things. It’s more important to produce a positive result than to produce a correct one, and even good scientists are lead astray by the rush to publish impressive results from underpowered studies with small sample sizes and implausibly large effects. There are a fair number of innocent mistakes being made, along with the fraud, bias and hype, but it all adds up to a bit of a crisis. As students we’re taught all about how to recognise faulty studies and how to build good ones — but the scientific world we enter into, if we choose research, doesn’t build on those foundations.
The book is surprisingly readable, and I would recommend it to both laypeople and scientists. It offers some very good analyses of what can go wrong, and some suggestions for how we can fix that, move on, and create a better, more open scientific community. It’s possible that if you’re a scientist you’ll wince and recognise that you yourself fell prey to this — the temptation to report positive results and shelve the negative ones, perhaps — but the point isn’t that one should always have been perfect, just that we all have to hold our hands up and work to make things better.
Plus, for a layperson, you’ll gain a better understanding of when to be sceptical, and what the warning signs are.