The main TIMSS study was based on randomly chosen nationally representative samples of eighth-grade students in each country. The video study sample was designed to be a random subsample of approximately half of the TIMSS classrooms in Germany, Japan, and the United States. Some deviations from this plan occurred in the implementation of the project. In the U.S. more than 15 percent of initially chosen schools and teachers refused to participate in the study. These were replaced by similar TIMSS schools and teachers, resulting in a final sample of 81 classrooms. In Germany, also, more than 15 percent of initially-chosen schools and teachers refused to participate, and were replaced by other similar TIMSS schools and teachers, resulting in a final sample of 100 classrooms. In Japan, a random sample of 50 of the regular eighth-grade TIMSS schools were chosen for videotaping. Within each Japanese school, the videotaped classroom was usually not the classroom in which the TIMSS assessments took place. The classroom that was videotaped was chosen by the principal and officials at the Japanese Institute for Educational Research. With these caveats, the videotape sample can be considered representative of the instruction received by eighth-grade students in Germany, Japan, and the United States.