Diagnostic Elementary Reading Profile App
This novel app uses empirical classification schemes via latent mixture model & classification and regression tree analysis (CART) to classify students into profiles of readers based on their fluency performance in K-2 at the fall, winter, and spring. The empirically derived classification schemes (decision rules appended to this documentation) are generated based on the user input of a set of fluency scores. The purpose of the system is provide teachers, parents, school administrators and students a set of recommended practices for instruction based on empirical classifications. The current state of score profiling is such that the teacher is supposed to group students based on performance of one assessment, yet when the student is administered a group of assessments (n > 1), it is difficult to 1) reliability group students together, 2) group students in a manner that is valid, 3) make rapid sense of the relative strengths and weaknesses of student reading scores and 4) provide appropriate instruction and/or remediation based on the groupings. By using the Diagnostic Elementary Reading Profile app, students will be automatically sorted into empirically derived groupings at any given time-point during kindergarten through second grade. This will reduce assessment and work time for the teacher as the sorting and recommendations will occur automatically. The appended decision rules were normed on a set of 60,000 students that are nationally representative in terms of race/ethnicity, achievement, socio-economic status, and English language status. A major advantage is the app automatically classifies students into reliable and valid groups for instructional purposes in the K-2 classrooms. Present workarounds in the field are largely theoretically drive guesses without data-drive support. The novel feature is the use of mixture modeling along with classification and regression tree (CART) to empirically define the rules for classifying students into profiles.