Michael Blastland’s new book, The Hidden Half: How the World Conceals its Secrets, starts with a tale of marmokrebs. You might well ask. These are a species of crayfish, which appeared out of nowhere, and consist only of females who reproduce, solo, producing genetically identical offspring. But their parthenogenesis isn’t the surprising thing about them. It’s that when genetically identical marmokrebs are reared in carefully controlled identical laboratory environments, they turn out to be – completely different from each other, in physical size and markings and also behaviourally.
This sets the theme for the book, which is about the limits of knowledge, and hence the need for caution in acting on knowledge. Successive chapters look at the replication crisis in science (bound to hit economics eventually), the limitations of medical treatments, the way reasonable, intelligent people can draw opposite conclusions from the same set of undisputed facts – in short, it’s a mediation on expertise and its role in decisions. The message from the multiple examples is that expertise isn’t all experts claim it to be.
This couldn’t be more timely, of course, and the book is a terrific read with many examples from different domains of research and policy. It’s a model of clarity of exposition. But I shoudn’t give the impression that it has a Gove-ian mission (that’s Michael “We’ve had enough of experts” Gove). Expertise has limitations but rejection of expertise has more. So the book ends with some advice for plying expertise responsibly: experiment and adapt if need be; triangulate between different sources and types of evidence; don’t leap to conclusions; accept the uncertainty and be sure to explain it; avoid mechanical metaphors (there’s no such thing as a policy ‘lever’ for example.
All very sensible advice. Like all sensible advice, so hard to apply. Modern hyper-democratic politics hardly lends itself to policy experiments and their reversal if need be, or the communication of nuance. But The Hidden Half should make researchers – especially those aiming for causal inference – to be modest in the extreme about what we ‘know’ and to reflect carefully on any policy advice we offer. It’s an essential read, especially for those who don’t think they need to bother reading it. It’s out this week and Michael Blastland has a blog on the book.