This study used sequential mixed-methods analyses to investigate the effectiveness of a research-based language arts curriculum for gifted third graders. Using analytic induction, researchers found that teachers' beliefs and expectations (time, sense of autonomy, expectations for students, professional expertise) influenced the degree to which they implemented the research intervention with fidelity to its design. Next, maximum variation sampling and quantitative analysis of student outcomes determined that postassessment achievement test scores are higher for students in classrooms with teachers who show high fidelity or adherence to the intervention. (Contains 6 tables and 1 note.)