National Evaluation of Core Knowledge Sequence Implementation. Final Report.

12 schools implementing the Core Knowledge Sequence, with matched comparison schools; first and third graders tested

US elementary schools adopting the Core Knowledge Sequence as a whole-school reform, grades 4-6 focus with testing at grades 1 and 3

Three-year mixed-methods evaluation of the E. D. Hirsch Core Knowledge Sequence - the planned progression of specific topics in language arts, history, geography, maths, science and fine arts that is the ancestor of every knowledge-rich curriculum in this cluster. Design is longitudinal case study plus matched intervention-versus-comparison testing on a basic skills battery and a Core Knowledge test. Schools chose to adopt; comparison schools were matched, not randomised, so the estimate carries the whole adoption-selection problem (which schools choose a demanding whole-school curriculum, and why). The implementation findings are the durable contribution: after three years all 12 schools were still implementing, 10 of them authentically, and 9 had reached moderate or high implementation - i.e. a content-rich sequence is deliverable at scale by ordinary schools, which is an applicability fact even though it is not an impact estimate.

Core Knowledge schools posted three-year achievement gains in reading comprehension relative to matched comparison schools, with a consistent relationship between depth of implementation and test-score improvement. Teachers reported more interesting and content-rich instruction, greater curricular coherence and more collaboration. None of this is separable from the selection of schools that adopt a demanding whole-school reform, and the implementation-outcome correlation is a post-adoption contrast.

High for a causal reading. Schools self-selected into adoption and comparison schools were matched on observables only; family composition and school quality differences remain.

Superseded on the same curriculum by the Grissmer et al. (2023) kindergarten lottery, which supplies the randomisation this design lacks and reports ITT 0.24 / TOT 0.47 on state reading tests.