Whatever approach is taken, it is usually a good idea, as suggested above, to undertake some pilot work to try out both data collection and analysis techniques, and make sure that the research plan will provide the type of data needed. This allows the opportunity to revise the research design before too much time and effort have been committed to an approach that does not lead to the type of data needed. These issues are less acute in genuine action research, where the iterative process of cycles of action and evaluation in effect encompass pilot work as an integral part of the overall approach.