JCPSLP Vol 15 No 2 2013

Research

Bilingual language sample analysis: Considerations and technological advances John J. Heilmann and Marleen F. Westerveld

With the increasing cultural and linguistic diversity of speech-language pathologists’ caseloads, there is a pressing need for assess­ ments that enable accurate and authentic evaluations of the communication skills of children from diverse backgrounds. Language sample analysis (LSA) has many properties that make it an effective tool for the comprehensive evaluation of culturally and linguistically diverse (CALD) children’s expressive language skills. Using LSA allows the clinician to assess language skills within naturalistic discourse, and as such, is more suitable for CALD children than most decontextualised norm-referenced assessments. Furthermore, LSA provides rich descriptive data and can be used within a dynamic assessment protocol to assist with the accurate identification of CALD children with language impairments. The goal of this paper is to summarise the complex issues that arise when completing LSA with paediatric CALD clients and describe how technological advances in computerised LSA have improved the accuracy and efficiency of the process. T hroughout much of the world, speech-language pathologists’ (SLPs) caseloads are becoming more culturally and linguistically diverse. This is particularly evident in Australia, where more than one fifth of the population speaks more than one language (Australian Bureau of Statistics [ABS], 2010). While most children speak English as their primary language, a substantial percentage (about 12%) has a different dominant language (McLeod, 2011). In addition to the linguistic diversity, SLPs need to consider their clients’ concurrent cultural diversity; in the 2010 Census, Australians identified more than 270 different ancestral backgrounds (ABS, 2010). Cultural and linguistic influences on language assessment Even when clients have strong English skills, a mismatch between the SLP’s and client’s culture can impact clinical

services. Hand (2011) documented breakdowns in communication between a group of SLPs and their clients who were English-speaking and from diverse cultural backgrounds which resulted in poor reviews of the clinical services. It is imperative that clinicians provide sensitive and appropriate care for their culturally and linguistically diverse (CALD) clients (Speech Pathology Australia [SPA], 2009). One of the biggest challenges for SLPs working with CALD clients is effectively identifying children who have true disorders and distinguishing them from those who have communication differences based on their cultural or linguistic background. While the need for sensitive and accurate assessment is clear, it can be difficult to execute. CALD children have a greater likelihood of being over- or under-identified with a language impairment when compared to mainstream monolingual peers (Bedore & Peña, 2008). When assessing CALD children, SLPs need to consider a child’s relative proficiency across the dominant language (L1) and second language (L2). Bilingualism is a complex and dynamic phenomenon that is distinct from monolingual language acquisition (Paradis, Genesee, & Crago, 2011). Children can be fluent bilinguals with typically developing language skills (L1 = L2), have limited proficiency in their second language (L1 > L2), experience loss of their first language (L1 < L2), or have a true language impairment (both L1 and L2 are below expected levels; Kohnert, 2010). Direct assessment of both L1 and L2 is difficult given the lack of normative data available for most of the languages spoken in Australia. Assessing non-English languages presents a challenge for most Australian SLPs, who are predominantly mainstream monolingual English speakers and/or do not speak the language of their clients (Williams & McLeod, 2012). Professional associations, such as Speech Pathology Australia (2009), caution against the use of norm-referenced tests when working with CALD children. Most norm- referenced tests are laden with biases that discriminate against CALD populations; they do not account for the distinct language profiles of children learning multiple languages, often do not use CALD children in their norming samples, and frequently contain content and formatting that are unfamiliar to CALD children (White & Jin, 2011). Given these biases, CALD children who are proficient in English may still have significant difficulty with norm-referenced tests. For example, Hemsley, Holm, and Dodd (2006) found that bilingual 11-year-old Australian children who were fluent in English scored significantly lower than their

Keywords Language assessment Language sample analysis bilingualism

cultural diversity

This article has been peer- reviewed

John J. Heilmann (top) and

Marleen F. Westerveld

87

JCPSLP Volume 15, Number 2 2013

www.speechpathologyaustralia.org.au