This study used a hierarchical linear model (HLM) to estimate direct and indirect effects of instructional practices recommended by the National Science Education Standards on individual achievement. Three pedagogical reforms—namely, providing more opportunities for laboratory inquiry, increasing emphasis on critical thinking, and reducing the amount of teacher-centered instruction—were expected to account for variability in school mean achievement and explain why gender, racial-ethnic status, and socioeconomic status have more influence on achievement of students in some schools than in others. Results suggest that whereas the instructional policies recommended by the authors of the Standards may be associated with higher achievement overall, they are equally likely to have the unintended consequence of contributing to greater achievement gaps among students with different demographic profiles. Theoretical expectations about the impact of instructional practices on academic excellence and equity require further evaluation. © 1999 John Wiley & Sons, Inc. J Res Sci Teach 36: 1110–1126, 1999