ACQ Vol 13 no 3 2011

that the signs of language learning disorder are not missed. The crucial question is whether the child shows evidence of language difference or language disorder. Language impairment affects language learning capacity generally, not a specific language, so “a child with language impairment should demonstrate limited performance in both languages, not only in English” (Gutierrez-Clellan & Simon-Cerejeido, 2009, p. 239). The implication is clear: assessment of both languages is needed. This may not be possible, however. Standardised tests in the home language may not be available, and if they are available, will be difficult for the monolingual speech pathologist to administer. Variability in language experience means that standardised tests in English cannot be used with any degree of confidence, and it is likely that bilingual children will not score well on these measures (Fagundes, Haynes, Haak, & Moran, 1998). The literature provides a number of different approaches to assessment. Kohnert (2010) discusses assessment approaches for bilingual children under three headings – monolingual comparisons, bilingual comparisons, and within child comparisons. Assessment of either the first or second language against a normative group constitutes monolingual comparison. Kohnert also includes non-word repetition (NWR) tasks under this heading. Some research (Oetting & Cleveland, 2006; Rodekohr & Haynes, 2001) suggested NWR as a potentially non-biased method of assessment of language learning capacity in bilingual children. This suggestion was based on the premise that NWR is a processing-based, rather than a language- based task. However, research conducted by Kohnert and colleagues (for example Kohnert, Windsor, & Yim, 2006; Windsor, Kohnert, Lobitz, & Pham, 2010) has led to the suggestion that the use of NWR tasks in only one language may not be a clinical marker of language impairment in the case of bilingual children (Windsor et al., 2010). Language- based processing measures such as NWR are seen to reduce, but not eliminate bias when used monolingually. Bilingual comparisons look at the language performance of bilingual children with language impairment and that of other bilingual children. These comparisons have consistently shown that the children with language impairment differ from their bilingual peers. Comparing bilingual children is important for diagnosis, but Kohnert (2010) points out that there are still challenges inherent in the paucity of normed tests for many languages and the limited number of bilingual speech pathologists. Within child comparisons consider the child’s ability to learn language. Two main types (limited training [or fast mapping] tasks and dynamic assessment) are found in the literature. Dynamic assessment has most often been reported, and is used in domains other than speech and language. The approach is based on the work of Vygotsky, who suggested that learning takes place in interaction with more skilled others. A test- teach- retest paradigm is adopted, and a measure of modifiability is completed by the clinician (see, for example, Gutierrez-Clellan & Peña, 2001; Peña, 2000). Evidence suggests that children with language impairment, or those with weaker language, will be rated more poorly on their learning ability (modifiability) than those with typical, or stronger language (Peña et al., 2006; Peña, Iglesias, & Lidz, 2001; Ukrainetz, Harpel, Walsh, & Coyle, 2000). The clinician’s rating of modifiability has been shown to be a strong and accurate predictor of language ability (Peña et al., 2006).

Information from parents has also been shown to have value in identifying language disorder in bilingual children. Paradis, Emmerzael, and Duncan (2010) developed a non-culture specific questionnaire, the Alberta Language Development Questionnaire (ALDeQ) to tap into parent perception of children’s language development, and evaluated how well this differentiated language-impaired English language learners from typically developing English language learners. They found statistically significant differences between the two groups for total and section scores on the questionnaire, with large effect sizes. Specificity (96%) was better than sensitivity (66%). Similar results were found in a study which used the ALDeQ with English language learners in Perth, WA (see May & Williams, 2011). The current evidence on assessment of language in bilingual children indicates that it is a far from simple matter which will require consideration of information from multiple sources (Isaac, 2002; Langdon & Wiig, 2009). Lewis, Castilleja, Moore, and Rodriguez (2010) presented a framework for organising multiple sources of assessment information for school-aged bilingual children. This has been modified by the current author to include scope to record information which will allow it to be used with both preschool-aged and school-aged children (see Appendix). Judgements as to whether the evidence supports an interpretation of typical language learning processes, speech/language disability, or learning disability are recorded in the framework, and an overall judgement may be made on the basis of the weight of evidence. There are early indications in the literature that the future of assessment in this population may involve non-linguistic tasks. Kohnert, Windsor, and Ebert (2009) present evidence from a study which compared the performance of three groups of children (typically developing bilingual; typically developing monolingual English speakers; monolingual English speakers with primary language impairment [PLI]) on three types of task (perceptual-motor demands; non- linguistic demands; linguistic demands). All tasks were administered in English. The research aimed to identify points of similarity and difference, particularly in the performance of PLI and bilingual children. Their findings indicated that language-based tasks (such as non-word repetition) disadvantaged bilingual children compared to monolingual children. The non-linguistic tasks (visual detection, auditory pattern matching, mental rotation and visual form completion) were most successful in differentiating bilingual children from the typically developing monolingual children. Kohnert et al. (2009) concluded “it may be that performance on some set of non-linguistic processing tasks can be used to help identify children with PLI in a linguistically diverse population” (p. 109). If further research confirms these findings, it may be that our approach to assessment of bilingual children will be very The literature regarding speech and language intervention for bilingual children is less extensive than that addressing assessment, and high level evidence is scarce (Elin Thordardottir, 2010). The key issues are the advice that should be given to parents as to which language (or languages) to speak in the home, and the language (or languages) to be used in intervention. Parents may ask which language they should use at home, or may report that they have been advised not to different in the future. Intervention

108

ACQ Volume 13, Number 3 2011

ACQ uiring Knowledge in Speech, Language and Hearing