This paper examines whether a literacy intervention involving generic and content area literacy components can improve both achievement on a standardized reading test and the attainment of secondary school qualifications, and whether the intervention can be implemented by teachers in their regular classroom settings. We report on a design-based approach for whole-school improvement, the Learning Schools Model (LSM), which was implemented in seven schools with low secondary school qualification rates. The LSM's core premises are that instructional practices need to be developed from evidence about teaching and learning in specific contexts and that professional learning communities need to fine-tune their instructional practices through collaborative analysis of data. The study employed a quasi-experimental design within a design-based approach and included classroom observations and teacher and student surveys. Reading achievement postintervention was statistically significantly higher than the projected achievement levels had the intervention not occurred. Effect sizes for tracked cohorts were Cohen's d = 0.50 and 0.62. The hierarchical linear model of reading achievement, which included students' attitudes toward reading, accounted for about 95% of the total variance. The attainment of secondary school qualifications (measured by odds ratios) increased significantly compared with school attainment prior to the intervention. The rates of attaining these qualifications were faster than national rates. The results suggest that both generic and content area literacy instruction are required and that a strong foundation in generic literacy should be maintained. However, the appropriate blend of literacy instruction is determined by a profiling of teaching and learning needs.