If you create Software there are a lot of things we are supposed to do: Write tests (first), work in pairs, use patterns and so on.
But do these things really work? Do they make us faster or the resulting software better? Very few people seem to care. Most people just tell you to do it this way or that way.
This book provides reliable answers. In each chapter a method, a common assumption gets challenged. Various studies get presented and the results discussed.
After reading this book you will have lots of information about what actually works and possibly even more important: Under what conditions these results apply.
I consider this a very valuable resource for directing your personal or organizational development efforts, no matter if you are a developer, a team lead or a manager of a software development team.
Each chapter is written by a different scientist, this has some important consequences:
1) From a scientific point of view the authors know what they are talking about.
2) Each chapter is written in a somewhat different style. In general the style is somewhat dry so it isn't exactly an easy reading. But it is far more digestible than the average scientific publication.
If you are interested in more details about a given topic each chapter ends in a long list of references for further research.