One purpose of educational research is to provide information about the likely impact of interventions or treatments on policy-relevant populations of students. Randomized experiments are useful for estimating the causal effects of interventions on the students in schools that participate in the experiments. Unfortunately, the samples of schools…

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;…

The Texas Poll of Elementary School Teachers was a statewide telephone survey designed to accurately describe the current science teaching practices in Texas public elementary schools, and the extent to which these elementary teachers feel prepared for the task. This paper focused on the survey sampling procedures and questionnaire design used in…