This study investigated the extent to which Hispanic newcomer ELLs in grades 4 and 5, who received math instruction in Spanish, improved their math scores in district and state tests as compared to ELLs who received math instruction in English only. Descriptive and inferential statistics such as frequency distribution, variability, and ttests informed the analysis. Cummins' (1981) Common Underlying Proficiency Theory served as the conceptual framework. Results indicate that the treatment group exhibited higher gains (7%) than the comparison group (< 3%). Those interested in closing achievement gaps may consider adding dual-language alternatives to their ESL programs.