Improving the Efficacy of First Grade Reading Screening: An Investigation of Word Identification Fluency With Other Early Literacy Indicators

This study investigated the accuracy of several early literacy measures that have been used in research and practice for first grade reading screening. A set of measures, Word Identification Fluency (WIF), Letter Naming Fluency, Phoneme Segmentation Fluency, and Nonsense Word Fluency, were administered as screening measures with 138 first grade students in the fall of the school year. Reading skills were assessed at the end of first grade using Oral Reading Fluency, the Test of Word Reading Efficiency (Sight Word Efficiency and Phonemic Decoding Efficiency), reading Maze, and a latent variable composite. Analyses compared the accuracy at which each screening measure, as well as combinations of screening measures, classified students according to successful or unsuccessful reading outcomes at the end of the year. Receiver operating characteristic curve analyses indicated that when compared with other single screening measures, WIF tended to demonstrate the strongest overall classification accuracy. With levels of sensitivity held constant at .90, combining screening measures with WIF using predicted probabilities from logistic regression analyses resulted in modest improvements in accuracy by reducing the number of false positives. More measures were not always better, however, as models of two or three measures were as accurate as models consisting of all measures. Overall results provided support for WIF as a starting point for first grade reading screening, whereby additional steps might be taken to reduce false positives.