Exploring boundary conditions of the listening comprehension-reading comprehension discrepancy index

Abstract

Conversations about the nature of dyslexia and how dyslexia impacts reading and listening comprehension get to the heart of classification and identification models of dyslexia. Recently, this conversation has been expanded to include efforts to estimate the prevalence of dyslexia in the population through the introduction of a discrepancy index of listening comprehension and reading comprehension. This discrepancy index was proposed to serve as a proxy for dyslexia when estimating its prevalence in the population. Individuals whose reading comprehension is considerably lower than their listening comprehension are thought to exhibit unexpected reading deficits. However, the index could underrepresent certain groups within the population. The current study explored this possibility using data from a sample of 4078 public-school students. We hypothesized that students from historically marginalized or otherwise disenfranchised groups (i.e., poor and minority groups) would be less likely to have a positive listening comprehension — reading comprehension (LC-RC) discrepancy index. Based on the results of multilevel linear mixed effect modeling, socioeconomic status (SES) contributed to differential performance on the discrepancy index when it was calculated using residual scores. Moreover, African American students were identified as having a reliably lower discrepancy index regardless of how it was calculated. It appears that this index, which only looks at the comprehension of language and not production, may, in fact, disadvantage students for whom oral language production differs from General American English (GAE). These outcomes suggest that this measure may lack the sensitivity to identify bidialectal students with dyslexia.