Depending on the school-year calendar in each participating country, data collection began in 1998 and continued through 2000. Sampling procedures required that lessons were randomly selected to be representative of eighth-grade mathematics and science lessons overall in each participating country. This includes eighth-grade mathematics and science lessons collected in Japan in 1995 as part of the prior video study. In each case, a teacher was videotaped for one complete, regular lesson (lessons devoted entirely to testing were not included). Lessons were selected across the year to try to capture the range of topics and activities that can take place throughout an entire school year. Schools were not permitted to substitute classes for the ones randomly selected. Finally, the sample from each country was reviewed for consistency by an international sampling expert. No substitutions of schools, teachers, students, or classes were allowed.