Gregory KD, Oetting JB. Classification Accuracy of Teacher Ratings When Screening Nonmainstream English-Speaking Kindergartners for Language Impairment in the Rural South.
Lang Speech Hear Serv Sch 2019;
49:218-231. [PMID:
29621802 DOI:
10.1044/2017_lshss-17-0045]
[Citation(s) in RCA: 11] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/18/2017] [Accepted: 09/15/2017] [Indexed: 11/09/2022] Open
Abstract
Purpose
We compared teacher ratings as measured by the Teacher Rating of Oral Language and Literacy (TROLL; Dickinson, McCabe, & Sprague, 2001, 2003) and Children's Communication Checklist-Second Edition (CCC-2; Bishop, 2006) to 2 established screeners, the Part II of the Diagnostic Evaluation of Language Variation-Screening Test (DELV-ST-II; Seymour, Roeper, & de Villiers, 2003) and Dynamic Indicators of Basic Early Literacy Skills-Next (DIBELS; Good, Gruba, & Kaminski, 2009), and then examined whether teacher ratings alone or when combined with the DELV-ST-II or DIBELS accurately classify nonmainstream English-speaking kindergartners by their clinical status.
Method
Data came from 98 children who lived in the rural South; 47 spoke African American English, and 51 spoke Southern White English. Using the syntax subtest of the Diagnostic Evaluation of Language Variation-Norm Referenced (Seymour, Roeper, & de Villiers, 2005) as the reference standard, 43 were language impaired and 55 were typically developing. Analyses included analysis of variance, correlations, and discriminant function with sensitivity and specificity indices.
Results
The TROLL, CCC-2, DELV-ST-II, and DIBELS showed clinical status but not dialect effects, and they correlated with each other, the Diagnostic Evaluation of Language Variation-Norm Referenced, and other language measures. Classification accuracies of all 4 tools were too low for screening purposes; however, empirically derived cut scores improved the results, and a discriminant function selected the TROLL and DELV-ST-II as optimal for determining who should be referred for an evaluation, with the TROLL yielding the highest level of sensitivity (77%).
Conclusion
Findings support teacher ratings as measured by the TROLL when screening nonmainstream English-speaking kindergartners for language impairment in the rural South, while also calling for additional development and study of teacher rating tools and other screening instruments.
Supplemental Material
https://doi.org/10.23641/asha.6007712.
Collapse