Examining the structure of dialect and language skills among early elementary African American children

Examining the structure of dialect and language skills among early elementary African American children

First Author: Brandy Gatlin-Nash -- University of California, Irvine
Additional authors/chairs: 
Nicole Patton Terry; Lakeisha Johnson
Abstract / Summary: 

Purpose: The purpose of this study was to analyze relations among two approaches for examining dialect use among a large sample of African American children in first and second grades. A second purpose was to evaluate these methods in relation to performance on a measure of language ability and to examine the factor structure of dialect and language.

Method: The sample included 200 children in Grades 1 (mean age = 6.8) and 2 (mean age = 7.6). Children completed two spoken dialect measures and a battery of language assessments, including three subtests of the Test of Language Development (Newcomer & Hammill, 2008). Correlational analyses were conducted using SPSS, and we used these results to inform models for confirmatory factor analyses, which were carried out using Mplus.

Results: Correlations among dialect measures were significant and moderate (r = .37 – .59), as were correlations among the subtests of the TOLD (r = .44 – .58). However, correlations among dialect measures and language measures were significant, moderate, and negative (r = -.24 – -.60). These analyses suggested a bi-factor model. Confirmatory factor analyses with Dialect and Language as separate factors indicated good model-data fit with a large and negative correlation between the two latent factors (r = -.90).

Conclusions: Findings suggest moderately similar dialect use among different methods for assessing dialect. In addition, negative bivariate correlations among dialect frequency and language skills, as well as a negative correlation between Dialect and Language, indicate adverse relations between frequency of dialect use and performance on the language measures. However, further analyses reveal potential confounding of dialect use and performance on the language subtests, particularly morphological completion and syntactic understanding. Finally, findings provide implications for measuring dialect use as a related, but separate construct from language ability.