9722723 Hoffer Current efforts to improve mathematics and science education in American schools are strongly influenced by research based on national surveys of schools and their students. Highquality data on curriculum content and instructional practice are increasingly seen as crucial to understanding the variability in student educational outcomes. This need may well be greatest at the elementary school level, where national data are relatively scarce. The proposed study seeks to improve the understanding of elementary schooling and outcome differences, by assessing the validity of data collected from early elementary teachers on what they teach and how they teach it. The U.S. Department of Education's National Center for Education Statistics is about to embark on an unprecedented data collection effort, titled the Early Childhood Longitudinal Study (ECLS). The ECLS will begin in fall 1998 with a nationally-representative sample of 23,000 kindergartners drawn from over 1,000 elementary schools. These students will be followed through the fifth grade, and data will be collected from their teachers each year. The teacher questionnaires include extensive sets of items asking about curriculum and instruction, but little is known about the validity of these measures: To what extent do they measure the phenomena of actual interest? As the largest study ever conducted of early elementary education in the United States, the ECLS data and data-collection instruments will be drawn upon by other studies for many years to come, and will greatly influence how curriculum and instruction are conceived and studied. Clearly, there are great benefits to making the instruments as good as possible. To address the issues of validity, the proposed study will administer field-test versions of the ECLS teacher questionnaires to a sample of 60 kindergarten, first-grade, and second-grade teachers in 12 schools. These teachers will then complete daily logs of how they used their time to teach mathematics and scie nce. The project will collect extensive archival records of the texts and materials used in the classes, and will code this information to make it comparable with the questionnaire items. The daily log and archive-derived data will then be systematically compared with the questionnaire data to identify points of agreement and disagreement. If the study is conducted in the spring ofthe 1996-97 school year, cognitive achievement test data for random samples of children in the teachers' classes will be available, and will be correlated with the curriculum and instruction variables collected from the teachers. The analytic efforts will be followed by debriefing sessions and focus groups with the participating teachers as a means of further checking the researchers' interpretations of the results. The study will produce technical reports documenting the entire project, as well as journal articles focussing on the general methodological and substantive implications of the study. These documents will be written in close consultation with leading authorities in research on mathematics and science education.
