Predictive Validity of Early Literacy Measures for Korean English Language Learners in the United States

Abstract

This study examined the predictive validity of early literacy measures with first-grade Korean English language learners (ELLs) in the United States at varying levels of English proficiency. Participants were screened using Dynamic Indicators of Basic Early Literacy Skills (DIBELS) Phoneme Segmentation Fluency (PSF), DIBELS Nonsense Word Fluency (NWF), and Word Identification Fluency (WIF). Reading outcome measures included DIBELS Oral Reading Fluency (DORF) and the Woodcock Reading Mastery Tests (WRMT). Data were analyzed using correlation analyses, regression analyses, and receiver operating characteristic analyses. Results highlighted the role of WIF as a screening measure. Among all measures, WIF was the most correlated with reading outcomes for all groups and made a meaningful contribution in predicting reading outcomes after controlling for PSF and NWF. Specifically, WIF contributed an additional 30% to 42% of the variance in reading outcomes. There is also preliminary evidence that predictive validity varies by English proficiency level. Regarding predictive accuracy, sensitivity values for PSF and NWF were 73% or higher and specificity values were 68% or higher. An exploratory analysis to determine optimal cut-scores for the screeners was conducted. The optimal cut-scores for WIF yielded the highest levels of predictive accuracy with sensitivity and specificity values greater than 92%.