The purpose of this study is to evaluate the relationship of mathematics calculation rate (curriculum-based measurement of mathematics; CBM-M), reading rate (curriculum-based measurement of reading; CBM-R), and mathematics application and problem solving skills (mathematics screener) among students at four levels of proficiency on a statewide test. It was hypothesized that CBM-M provides insufficient information to make good screening decisions and that other measures with content more similar to that of large-scale tests of mathematics would function to improve screening. One hundred and seventy students in third grade from a rural elementary school in the Midwestern United States participated. Structural equation modeling was used to evaluate direct, mediator, and latent growth models. In general, CBM-R mediated the relationship between the mathematics ability screener and passing the state assessment, while CBM-M did not have any significant paths within these models. Results are discussed in terms of the utility of CBM-M and CBM-R procedures in screening for success on state test performance in mathematics.