Andrew Jaffe introduces the difficulties we face when trying to take a scientific view because largely we are dependent on induction: predicting the future from what has previously been observed. He explores what probability is, the two key ways of looking at it (frequentist and Bayesian) and how scientists use (or misuse it) to work out the implications of their experiments for hypotheses. This is then expanded into looking at the nature of scientific models and the philosophy of science before heading out to entropy, quantum randomness and attempting to achieve meaningful cosmology with its potential dearth of evidence.
The topic might sound a little dry, but in fact Jaffe does it with good humour and a very readable style. For example, he uses measuring his daughter's height by making marks on the wall, as many of us have, as a guide to the way that imprecision in experimentation needs an understanding of probability and statistics if we are to be accurate in presenting what we do and don't know.
I have one small moan and one proviso. The moan is that it's a bit slow to start. The first two chapters contain some necessary information, but do it at a slow pace. The proviso is that while most of the book is readable without much prior knowledge, the chapters on probability and Bayesianism don't break things down enough. They're fine if you already know a bit, and introduce some points I haven't seen before, but may be hard going otherwise. (I'd recommend reading both my Dice World for general probability and Tom Chivers' Everything is Predictable for Bayes first if you aren't familiar with the topic.)
This doesn't in any way undermine the fact that this is probably the best book I've ever read on what the scientific method really is, how scientists build models, how they deal with randomness, uncertainty and the pitfalls of induction. Engaging stuff.
Hardback:   |
Review by Brian Clegg - See all Brian's online articles or subscribe to a weekly email free here
Comments
Post a Comment