Critical features important for transforming large-scale educational assessment into effective policy research are examined. These include a developmental orientation to time-ordered data, empirically-grounded construct interpretations of measures and relationships, speculative inquiry into the role of diverse contributing factors, and appraisal of alternative perspectives on both the questions and the findings. (22pp.)