The Impact of Direct Instruction on Elementary Students' Reading Achievement in an Urban School District

Abstract

This article reports the preliminary achievement outcomes of the first 4 years of direct instruction (DI) reading, implemented in 6 Baltimore elementary schools. On the primary measure of reading comprehension, members of the original kindergarten cohort were, on average, reading at grade level (49th percentile) by the end of 3rd grade. Members of the original 2nd-grade cohort were nearing grade level (40th percentile) by the end of 5th grade. However, students at control schools (where other curricula to improve reading achievement were being implemented) were achieving at the same level, so there were no significant differences between the outcomes for the 2 groups (controlling for demographics and pretest factors). Future research, based on a well-established implementation (rather than the problem-filled early years of implementation experienced by these 6 schools) might uncover significant effects that were not evident at the time this article was written. DI appears to be a viable option for raising student reading achievement, even if this study has not yet yielded evidence that DI performs significantly better than other reading curricula. The call for educators and policymakers to consider the results of research when selecting reform models for high poverty schools (e.g., Slavin & Fashola, 1998) is heeded, at least occasionally. As Muriel Berkeley described elsewhere in this special issue, the search for a curriculum with a research-proven track record led members of the Baltimore Curriculum Project to select DI, even before it was identified as 1 of only 3 school reform models with strong evidence for improved student achievement by the American Institutes for Research report on schoolwide reform models (Herman et al., 1999). This article extends the circle by analyzing the first outcomes of an experiment that was motivated by previous research results.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 34 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.