Abstract
This article investigates the utility of various estimates of mathematics proficiency. The participants were 432 students in Grades 2 through 5. The delayed alternate form reliability of multiskill probes, retention probes, slopes of student growth, and trials to criterion were computed. The fluency probes were found to be both sufficiently reliable and significantly predicted proficiency on a group-administered accountability test. However, the trials to criterion and slope data were neither reliable nor significantly predictive. Finally, a receiver operating characteristic analysis was conducted to determine cut scores on the fluency probes for proficiency. Results found that 17 digits correct per minute for second- and third-grade students and 29 digits correct per minute for fourth and fifth graders demonstrated proficiency. These data resulted in overall correct classification estimates of .65 (second and third grades) and .73 (fourth and fifth grades).
Get full access to this article
View all access options for this article.
