This article began as a response to the article "Action research as narrative: five principles for validation" by Heikkinen, Huttunen and Syrjala, which appears in this issue of "Educational Action Research". In so doing it addresses the question "How can we tell whether an action research study is good?" by arguing that validity is a construct that can be used to evaluate the quality of qualitative research studies, including action research, and that, because of the moral and political aspects of action research, we are compelled to ask whether the results of our inquiries are valid. The paper also argues that we can continue to use validity in this way if we reject the extreme views of naive realism and radical constructivism. This argument and its conclusions are then used to critique the principles that Heikkinen "et al." suggest for quality in narrative forms of action research. The paper ends with additional suggestions for increasing the validity of action research studies.