When different individuals operating in different contexts implement similar innovations, they approach the various parts of the innovation in different ways. When assessed according to the research procedures used to study innovation implementation at the Research and Development Center for Teacher Education at the University of Texas at Austin, the patterns formed by the combination of these modified parts into new, tailored versions of the innovation are called innovation configurations. This report describes the five-step process for developing the checklists with which innovation configurations can be assessed and evaluated, then delineates some of the conceptual and technical problems associated with using the process. The five data collecting steps include identifying the components of the innovation, verifying the use of these components and establishing the range of variations in that use, refining the checklist developed in the first two steps, collecting data, and analyzing the data. Discussed here are the following problems associated with the procedure: establishing the reliability of the checklists, developing suitable data aggregation methods, and assessing the degree of change represented by the use of a particular version of an innovation. Examples of data types and data presentation methods are included. (PGD)