The 2012 NSSME used a stratified, two-stage probability sample of schools and science and mathematics teachers in grades K–12 in the United States. In the first stage, 2,000 elementary and secondary schools were selected within sampling strata with probability proportional to size. In the second stage, 10,000 science and mathematics teachers were sampled at predetermined rates to ensure the required number of teachers for domain estimates, such as for particular regions or types of communities.
The target population for the teacher sample consisted of teachers in eligible schools who teach science and/or mathematics. The sampling frame for the teacher sample in each stratum (e.g., high schools) was constructed using lists of teachers and their teaching assignments. Oversampling of teachers of advanced science and mathematics was done in order to provide separate results for the various subgroups.