A stratified two-stage cluster sample design was chosen for the Texas Assessment Project and is described in this document. The sample was designed so as to achieve a chosen level of precision in item analysis data within each of the 20 geographic regions chosen for analysis. Four sampling frames were developed: 9-year-olds; sixth graders; 13-year-olds; and 17-year-olds; or eleventh graders. Within each of the four sampling frames, stratification of schools was carried out on the basis of region, district type, district annual per-pupil expenditure, and expected number of students. Within each sampling frame, the schools in each of the 20 regions were cross-classified into 30 cells by five levels of district type and six levels of district annual per-pupil expenditure. It was also considered important to have both blacks and Mexican-Americans represented approximately in proportion to their proportions in the Texas student population. Large and small schools were sampled separately. In regions with a high proportion of black students, a larger number of schools were selected. The sampling method divides the frame in such a way that a near proportional number of students is selected from each final stratum. Finally, within these constraints, schools were selected random by computer. Within schools, students were selected randomly. Characteristics of the resulting samples are presented. (Author/CTM)