In research on teachers' beliefs, a distinction is often made between what teachers state ("professed beliefs") and what is reflected in teachers' practices ("attributed beliefs"). Researchers claim to have found both consistencies and inconsistencies between professed and attributed beliefs. In this paper, methods and research designs typically used in studies of teachers' beliefs are examined. It is asserted that, in some cases, the perceived discrepancy between professed and attributed beliefs may actually be an artifact of the methods used to collect and analyze relevant data and the particular conceptualizations of beliefs implicit in the research designs. In particular, the apparent dichotomy can be the result of a lack of "shared understanding" between teachers and researchers of the meaning of terms used to describe beliefs and practices. In addition, it is asserted that it is inappropriate to classify any belief as entirely professed since researchers make various attributions to teachers through choices about data collection, theory, analysis of data, and presentation of findings. Moreover, the emphasis on classifying beliefs in this manner may be inhibiting researchers from developing a more comprehensive understanding of teachers' beliefs. Traditional and alternative methods are described, a data example is provided to illustrate the claims, and implications for future research are discussed.