Two Decades of Generalizable Evidence on U.S. Instruction from National Surveys

Abstract

Background Students’ instructional experiences—that is, their experiences working with subject matter during classroom instruction—are a major determinant of how they learn. Given the importance of classroom instruction, valid, generalizable evidence is needed by policymakers, researchers, and practitioners. Over the past two decades, a wealth of generalizable evidence on instruction has been generated by large-scale surveys administered to nationally-representative probability samples. But this vast body of research has not been systematically summarized. Purpose of the Study This article attempts to fill a gap in the research by describing evidence on instruction from all surveys conducted between 1987 and 2005 that measured instruction using nationally-representative samples. Our goal is to generate a portrait of the evidence from these surveys that identifies strengths and gaps in the literature and that summarizes what this research base says about the relationship between classroom instruction and student outcomes. Research Design Evidence on instruction was compiled and summarized in four steps: (1) all surveys conducted between 1987 and 2005 that measured instruction and were administered to nationally representative probability samples were identified, (2) manuscripts using data from these surveys were selected for review, (3) the dimensions of instruction addressed by each manuscript and other manuscript characteristics were coded, and (4) the methodology and findings of each manuscript were summarized. Findings More than half the studies used data more than a decade old; few studies examined instruction during important transition years such as sixth and ninth grade; and subject area emphasis was lopsided, with mathematics and science instruction receiving much greater attention than English/language arts and social studies. The summary also revealed a repeated finding of low-SES students receiving diminished learning opportunities than more affluent peers. We also found repeated evidence of a positive association with student achievement for six dimensions of instruction, and repeated evidence of a negative or null association with achievement for five dimensions of instruction. Conclusions More research is needed on disparities in the instructional experiences of low-and high-income students. More research is also needed on instruction at key transition points and on instruction in English/language arts and social studies. This review also suggests a need for studies that more rigorously test research questions about instruction using measures that more authentically reflect the complexities of instruction and that examine student achievement growth over longer periods of time.