The Florida Comprehensive Assessment Test (FCAT) is Florida’s annual measure of student yearly academic progress with major implications for individual students and schools. The present study explored the inter-relationships among student background and reading performance indicators and the predictive utility of those variables as a means to identify students at risk for FCAT failure. The following indicators were incorporated into this study: student demographics (gender and socio-economic status), academic history (attendance rate and number of prior grade retentions), teacher-rated evaluation of student performance (student participation in Reading activities and Reading grade), and reading skill (decoding and oral reading fluency). Decoding and oral reading fluency were measured using the DIBELS Oral Reading Fluency (ORF) and Nonsense Word Fluency (NWF) subtests (Good and Kaminski, 2002b), three months prior to FCAT administration.
The student background and performance variables were first used in a multiple regression analysis to predict FCAT Reading achievement. Among these predictor variables, ORF, student’s third quarter Reading grade, and SES were found to be statistically significant predictors of Reading FCAT-SSS scores. The strong correlation found between ORF and FCAT achievement in this study was congruent with results in previous studies, whereas the addition of Reading grade and SES as significant predictors of Reading FCAT-SSS scores were new findings in this study. Next, it was of interest to determine whether any of the aforementioned
indicators also predicted ORF ability. Reading grade and NWF scores were both statistically significant predictors of ORF performance.
Lastly, this study examined the utility of two risk models of Reading failure that use cutpoint guidelines for student performance on the DIBELS ORF subtest to determine risk level. These cutpoint models were applied to this data set to determine the usefulness of these models in identifying students who may be at risk for reading difficulty. Both models explored were moderately effective in identifying at-risk students. However, a three-category risk model (low, moderate, high) was determined to be of greatest assistance to educators for efficiency and ease. In conclusion, this author identified limitations inherent in the present study and offered recommendations for future research directions.
