Recently I’ve been part of more code reviews, and part of an increasing number of conversations about them. I’ve started to wonder: What impact do code reviews really have? What makes for an effective code review? I certainly feel like it results in improved code, and obviously that’s the opinion of the development community at large as well. However, what something feels like and what its effects really are can be quite different, so I looked for empirical studies of code review - and found some very interesting results.