This paper outlines important considerations for program evaluators and school administrators when creating evaluation designs, disaggregating student outcome data, and interpreting and reporting achievement data for improved program implementation and informed classroom instruction to ensure treatment of English Learner (EL) students. Evaluation data from a large urban school district serving more than 25,000 EL students are presented to increase understanding and rigor of problems in serving EL students. The discussion focuses on narrowing the limitations of norm-referenced tests through a multiple measures model, overcoming the misguided treatment of EL students as a homogeneous monolithic student population, and the increased value of qualitative research data for program evaluations. The use of a multiple measures model to determine student achievement links student assessments more closely to instruction and curriculum, narrows the limitations of tests, and ensures that EL students are held to the same standards as non-EL students. Surveys of 390 high school students and 882 middle school students demonstrated how EL students felt about their educational experiences. Surveys completed by 102 parents of students in a two-way bilingual immersion program provided qualitative information about parent responses to their children's instruction. (Contains 12 tables, 6 figures, and 13 references.) (SLD)