Assessment of Speech Sound Disorders in Bilingual Latino Children: An Evidence-Based Procedure

Bilingual children are often misdiagnosed as having a speech impairment when they truly exhibit typical development. In addition, speech-language pathologists (SLP) are avoiding the assessment of bilingual children due to the lack of knowledge and resources necessary for best practice with this population. Misdiagnosis of speech sound impairments in this population has long-term emotional, financial, and educational costs. Misdiagnosis can be prevented if we acquire more information on what constitutes typical and impaired speech development in the bilingual population. The proposed research is designed as a continuation of Phases 1 and 2 of this research project to validate a model of speech sound production in bilinguals and to frame an assessment procedure for Latino preschoolers. Due to differences in speech production found in bilingual preschoolers (Fabiano-Smith & Goldstein, 2010a; 2010b; Fabiano-Smith & Barlow, 2010), an assessment procedure that evaluates the speech of bilingual preschoolers that considers both English and Spanish is necessary to prevent misdiagnosis of speech sound disabilities. The Specific Aims of the proposed research are to (1) validate PRIMIR 3, an extension of our theoretical framework for bilingual speech production; (2) revise our English and Spanish word lists using modern test theory methods, and (3) establish diagnostic sensitivity and specificity of our assessment procedure to determine replicability, fidelity, and validity. Children will produce sounds in single words and their productions will be phonetically transcribed. Child productions will be examined, in both English and Spanish and across both languages, on common diagnostic measures used for monolingual children. Using Network and Community Structure Analysis, we will identify phoneme connectivity based on previously and newly collected data on substitution patterns in both English and Spanish. Item Response Theory (IRT) will be used to identify a set of test items that are sensitive and specific to bilingual children with suspected speech sound disabilities. Finally, logistic regression models will be used to establish diagnostic sensitivity and specificity of traditional measures of phonological ability for bilinguals. Findings will lead to an evidence-based procedure for the identification of typical and disabled speech sound development in bilingual children and diagnostic guidelines for speech-language pathologists to use in the differential diagnosis of typical speech from speech disability. These findings will reduce diagnostic error in the identification of speech sound disabilities in this population, reducing health disparities for Latino children.