Assessing the instructional level for mathematics: A comparison of methods

This study compared the mathematics performance of 434 second-, third-, fourth-, and fifth-grade students to previously reported fluency and accuracy criteria using three categories of performance (frustration, instructional, and mastery). Psychometric properties of the fluency and accuracy criteria were explored and new criteria for the instructional level were empirically derived. Two sets of mathematics probes were administered to the students and delayed alternate-form reliability coefficients were obtained from multiskill probes. Results suggested that fluency data were significantly more reliable than accuracy data. Slopes from the single-skill probes were used to categorize students as high responders, and the mean fluency scores from that group's multiskill probes were used to suggest an alternative instructional level of 14-31 digits correct per minute for second- and third-graders and 24-49 digits correct per minute for fourth- and fifth-graders. Implications for practice and suggestions for future research are discussed.