Research on teacher thinking generally has failed to take into account teacher beliefs and understandings in both research methodology and in interpretation of data. Two models prevail in research on teacher thinking. The decision making model, which has undergone several modifications, tends to focus on explaining and understanding deliberate teacher activity. The second model for research on teacher thinking is the cognitive information process model, which focuses on how the teacher limits and structures the complex environment into what are called "problem spaces," which essentially are simplifications of environmental components. A review of research dealing with teacher thinking reveals that inadequate attention to the beliefs of both the teacher and the researcher can seriously weaken the inferences drawn. Problems arise when assumptions are made about teacher perceptions, when inadequate survey instruments are used, and when it is not recognized that teachers can have several widely varied and complex types of beliefs. An alternative methodology is being used to understand the various beliefs that drive teachers to plan and teach in different ways. The technique involves the identification, through interviews with teachers, of constructs or beliefs that form a grid that can be analyzed factorially. After discussion and analysis with the teacher, patterns of underlying factors emerge and lead to understanding of relationships and significant behaviors. This methodology could be used with lesson observation and stimulated recall to produce an effective research tool on teacher thinking. (FG)