Alternative theoretical approaches influencing the research design of the Beginning Teacher Evaluation Study are discussed. One line of research used an Experimental Teaching Unit (ETU) to create a semi-standardized situation for comparison of teachers. Though this approach introduces a uniformity which appears to increase comparability, it is at the cost of a somewhat unrepresentative sample of teacher behavior. Another line of research focused on naturally-occurring behavior, particularly instructional time. The design chosen for the Beginning Teacher Evaluation Study involved an "objectives-only" experimental teaching unit incorporated into the instructional time framework. Specific objectives in reading and mathematics were taught, but no materials were provided--only objectives. The effectiveness of 6-, 10-, and 40-day instructional programs were compared. The 40-day period was found to be most favorable in terms of consistent gains and variable amounts of actual instructional time. This method was also recommended because it provided a substantial sample of instruction, yet allowed for more than one sample during the school year. (Author/GDC)