In spite of lacking evidence for effects on cancer progression itself, an increasing number of cancer patients use various types of complementary and alternative medicine (CAM). There is disagreement between CAM practitioners, researchers and clinical oncologists, as to how evidence concerning effects of CAM can and should be produced, and how the existing evidence should be interpreted. This represents a considerable challenge for oncologists; both in terms of patient needs for an informed dialogue regarding CAM, and because some types of CAM may interact with standard treatments. There is a need for insight into which kinds of CAM may work, for whom they work, what the possible effects and side-effects are, and in what ways such effects may come about. The present article presents a framework for evaluating effects of CAM by suggesting a taxonomy of different levels of evidence related to different types of research questions and discussing the relevance of different research methodologies for different types of effects.