An analysis of measures used in assessment of innovation fidelity was studied. The purpose was to contribute to the development of valid and reliable methods of measuring critical elements of an innovation and identifying levels of implementation. Questionnaires were developed for high school introductory courses in English, calculus, and psychology. Controlling for student aptitude and teacher characteristics in the calculus and psychology courses, no significant relationships were found when the student was used as the unit of analysis. Scales used on the questionnaires had sufficient reliability (internal consistency and stability) to suggest that they measure recognized factors in these courses. The English student questionnaire was developed using the five categories: materials, structure, behavior, knowledge, and values. This is in contrast to the characteristics of the calculus and psychology courses which were based on activities found in most Personalized System of Instruction (PSI) course designs. After administration, several factor solutions were generated; these were not judged as good fits to the five logical categories. Recommendations are: (1) analyses of implementation be done using classrooms, not students, as the unit of analysis; (2) categories of study be based on course-specific design factors; and (3) multiple outcome measures be used as dependent variables. (Author/GK)