Effect of dialect density and modified scoring (Marencin et al., 2024)
Purpose: We investigated and compared the outcomes from two standardized, norm-referenced screening assessments of language (i.e., Clinical Evaluation of Language Fundamentals Preschool–Second Edition [CELFP-2], Diagnostic Evaluation of Language Variation–Screening Test [DELV-ST]) with African American preschoolers whose spoken dialect differed from that of General American English (GAE). We (a) described preschoolers’ performance on the CELFP-2 Core Language Index (CLI) and its subtests with consideration of degree of dialect variation (DVAR) observed, (b) investigated how the application of dialect-sensitive scoring modifications to the expressive morphology and syntax Word Structure (WS) subtest affected CELFP-2 CLI scores, and (c) evaluated the screening classification agreement rates between the DELV-ST and the CELFP-2 CLI.
Method: African American preschoolers (N = 284) completed the CELFP-2 CLI subtests (i.e., Sentence Structure, WS, Expressive Vocabulary) and the DELV-ST. Density of spoken dialect use was estimated with the DELV-ST Part I Language Variation Status, and percentage of DVAR was calculated. The CELFP-2 WS subtest was scored with and without dialect-sensitive scoring modifications.
Results: Planned comparisons of CELFP-2 CLI performance indicated statistically significant differences in performance based on DELV-ST–determined degree of language variation groupings. Scoring modifications applied to the WS subtest increased subtest scaled scores and CLI composite standard scores. However, preschoolers who demonstrated strong variation from GAE continued to demonstrate significantly lower performance than preschoolers who demonstrated little to no language variation. Affected-status agreement rates between assessments (modified and unmodified CELFP-2 CLI scores and DELV-ST Part II Diagnostic Risk Status) were extremely low.
Conclusions: The application of dialect-specific scoring modifications to standardized, norm-referenced assessments of language must be simultaneously viewed through the lenses of equity, practicality, and psychometry. The results of our multistage study reiterate the need for reliable methods of identifying risk for developmental language disorder within children who speak American English dialects other than GAE.
Supplemental Material S1. Histogram: unmodified fall and spring CELFP-2 CLI standard score by DELV-ST language variation status. CELFP-2 = Clinical Evaluation of Language Fundamentals Preschool–Second Edition; CLI = Core Language Index; DELV-ST = Diagnostic Evaluation of Language Variation–Screening Test; GAE = General American English.
Supplemental Material S2. Bland-Altman plots for unmodified and modified CELFP-2 CLI composite scores. 95% confidence interval indicated by the top and bottom dashed lines; mean difference between fall and spring time points indicated by the central dashed line.
Supplemental Material S3. Bland-Altman plots for unmodified and modified WS subtest scores. 95% confidence interval indicated by the top and bottom dashed lines; mean difference between fall and spring time points indicated by the central dashed line.
Supplemental Material S4. Bland-Altman plots for SS and EV subtest scores. 95% confidence interval indicated by the top and bottom dashed lines; mean difference between fall and spring time points indicated by the central dashed line.
Supplemental Material S5. DELV-ST diagnostic risk status by DELV-ST language variation status.
Supplemental Material S6. CELFP-2 risk for developmental language disorder (-1.0 SD from the mean) in fall with spring unmodified CLI.
Supplemental Material S7. CELFP-2 risk for development language disorder (-1.0 SD from the mean) fall with spring modified CLI.
Marencin, N. C., Edwards, A. A., & Terry, N. P. (2024). African American preschoolers’ performance on norm-referenced language assessments: Examining the effect of dialect density and the use of scoring modifications. Language, Speech, and Hearing Services in Schools, 55(3), 918–937. https://doi.org/10.1044/2024_LSHSS-23-00134
Funding
This study was supported, in part, by funding from the United Way of Greater Atlanta, the Research on the Challenges of Acquiring Language and Literacy Initiative at Georgia State University, and the U.S. Department of Education Office of Special Education and Rehabilitative Services (Grant H325D190037).
History
Usage metrics
