Specific language impairment (SLI) in bilingual Malaysian children is grossly underidentified for two reasons. First, there is an absence of locally-developed norm-referenced language assessment tools. The challenge in developing a local assessment tool comes from the inadequate knowledge of children’s English as it develops in the bilingual environment. Second, the characteristics of Malaysian English, a non-Standard form of English, are often confused with the features of SLI in monolingual children learning Standard English. To date, the literature has no information on SLI in bilingual children learning a non-Standard English because research mainly focuses on bilingual Standard English and monolingual non-Standard English. Spontaneous language sample is the recommended language assessment tool for bilingual Malaysian children because it provides quantitative and qualitative information for language development and assessment in communities with complex language environment. Phase 1 of the current study investigated the developmental sensitivity of four language sample measures (LSMs), including mean length of utterances (MLU), lexical diversity (D), the Index of Productive Syntax (IPSyn), and frequency of code-switching (FCS), by examining the relationship between these LSMs and age in 52 bilingual Chinese-English Malaysian children with normal language development (NLD), aged between 3;06 and 6;09. Analyses showed a highly significant linear relationship with age in D (r=.536) and IPSyn (r=.451), moderately significant linear relationship with MLU (r=.364), but not in FCS. The findings suggested that MLU, D, and IPSyn were developmentally sensitive to non-Standard English in the bilingual Malaysian children studied. Phase 2 compared the same measures obtained from nine children with SLI and their age-matched controls from the NLD group in Phase 1. The SLI group had significantly lower MLU and the IPSyn scores than the NLD group. Discriminant function analysis indicated that the IPSyn alone classified children into SLI and NLD groups at 77.8% classification accuracy. The four misclassified children, two in the SLI and two in the NLD group, were aged below 4;06 years. MLU and the IPSyn showed potential as SLI markers for bilingual Malaysian children. Two types of error patterns were observed from the SLI group. The first pattern included verb and the copula -be omissions which were also observed in younger children from the NLD group. The second pattern was omission of prepositions which was not observed in younger NLD group, reflected linguistic transfer from the first language. Future research in this area is suggested to a) extend the age range of the study to include older children to further verify the diagnostic potential of LSMs, b) consider longitudinal research design so that language development over time can be described and c) include non-linguistic measures given reports of these measures as potential markers of SLI. As an initial study on bilingual children learning non-Standard English, the current study provided empirical data for charting language development and also suggested potential markers for SLI. The study can serve as the basic framework not only for further research on other bilingual non-Standard English-speaking groups but also for children with language impairment secondary to developmental disorders in the same language group. / published_or_final_version / Speech and Hearing Sciences / Master / Master of Philosophy