The framework and results of a study exploring the effects of alternative assessment on mathematics teaching are reported. The instrument for the observation of alternative assessment in the classroom, based on the processes defined by P. L. Peterson (1988), is described along with a discussion of the methodology and the data implications. Eighteen elementary school and secondary school teachers were observed over the 1991-92 school year, as part of the Innovative Mathematics Assessment and Teachers' Classroom Practice project. In addition, results are presented from the collection of informal data, such as teacher interviews, comments, and written summaries describing assessment strategies that worked or did not work in the classes. Study data, along with data from the teacher observation reports, provide evidence that the use of alternative assessments helps mathematics teachers change their teaching strategies, incorporating the use of an emphasis on meaning and understanding, encouragement of student autonomy and persistence, and the direct teaching of higher order cognitive strategies. Six tables present study findings, and three figures illustrate the discussion. (SLD)