The most recent reauthorizations of No Child Left Behind and the Individuals with Disabilities Education Act served to usher in an age of results and accountability within American education. States were charged with developing more rigorous systems to specifically address areas such as critical academic skill proficiency, empirically validated instruction and intervention, and overall student performance as measured on annual statewide achievement tests. Educational practice has shown that foundational math ability can be easily assessed through student performance on Curriculum-Based Measurements of Math Computational Fluency (CBM-M). Research on the application of CBM-M's predictive validity across specific academic math abilities as measured by state standardized tests is currently limited. In addition, little research is available on the differential effects of ethnic subgroups and gender in this area. This study investigated the effectiveness of using CBM-M measures to predict achievement on high stakes tests, as well as whether or not there are significant differential effects of ethnic subgroups and gender. Study participants included 358 students across six elementary schools in a large suburban school district in Arizona that utilizes the Response to Intervention (RTI) model. Participants' CBM-M scores from the first through third grade years and their third grade standardized achievement test scores were collected. Pearson product-moment and Spearman correlations were used to determine how well CBM-M scores and specific math skills are related. The predictive validity of CBM-M scores from the third-grade school year was also assessed to determine whether the fall, winter, or spring screening was most related to third-grade high-stakes test scores.