A decision tree is a tool that serves as a guide through a sequence of choices. In the tree below, choices along the decision path have implications for the type of impact evaluation that will be most appropriate. The tree provides a warning to those who elect to proceed against the advice the diagram offers, such as choosing to use an experimental design even thought the size of an evaluation's treatment and control groups would be small and thus tests of differences between groups might be statistically underpowered and inconclusive.