This paper argues that the empirical evaluation of software engineering requires a wider range of strategies and data generation methods than has been used until now. It can learn from its sister discipline of information systems (IS). The paper summarises a range of strategies and methods that could be used. However, adoption of some of the additional strategies and methods would require software engineers to examine their assumptions about the nature of evidence and how it should be evaluated. The paper therefore explains the scientific, positivist paradigm and the interpretive paradigm. It argues for including qualitative methods and the interpretive paradigm in the empirical evaluation of software engineering. Finally the paper reflects on the implications for research and practice of having a range of available strategies and methods and two contrasting underlying philosophies.