Reason analysis is a long neglected method of data collection and analysis. This article describes the method and shows some fields of practical application. The methods and principles of reason analysis were first expounded by Lazarsfeld (1935). Reason analysis explores how respondents answer “why” questions during survey interviews. Typically, respondents are asked a simple question inquiring about the reasons that led them to make a specific decision or action. Data obtained in this way are often used to construct a simple classification of respondents. In reality, respondents often have many reasons for making a particular decision, but usually only mention one to an interviewer. Reason analysis contends that responses to ‘why’ questions are a combination of some or all reasons used by the respondent to formulate an answer to an interviewers’ question. Consequently, reason analysis constructs a “tree” of questions and an “accounting scheme” or model of the decision or action being studied. Using this framework, responses are grouped into classes and types according to their (dis)similarity. With the development of software tools it is now easy to estimate reason analysis models of survey response. One key advantage of reason analysis is that it facilitates developing a deeper understanding of the latent structure of groups; and hence allows a more precise estimation of individual level effects in studies of decision-making. As the demand for “structural estimation” models of decision making and action increase, it is likely the reason analysis will become a more influential methodological approach in the 21st century.