No document available.
Abstract :
[en] The diagnostics of specific learning disorder in mathematics comes along with a thorough assessment of individual strengths and weaknesses, particularly in mathematics. Most diagnostic tools are based on language to present instructions and test contents, and hence test language proficiency may impact student’s test performance. This is especially challenging in multilingual settings, such as Luxembourg, where most students do not speak the language of mathematics instruction (usually corresponding to the test language) at home. However, even though currently used tests generally show good psychometric quality, they were mostly developed for native or proficient speakers, which challenges an adequate and valid outcome in a linguistically heterogenous environment. To optimize the diagnostic process in a multilingual education context, we have developed a standardized test battery for diagnostic purposes in mathematics for third-grade children. It consists of 19 tasks assessing basic numerical and arithmetic skills, in accordance with diagnostic guidelines and neurocognitive models of number processing. Following a pilot study (N = 211), normative data was collected in elementary schools in Luxembourg (N = 696), and special emphasis was put on the language aspect during test development and data analysis. In only five subtests, children who did not speak the language of math instruction at home were significantly outperformed by their native peers, which could generally be explained by their test language skills. More concretely, besides general vocabulary skills, content-specific (i.e., mathematical) language also significantly contributed to this relation, extending previous research findings in younger children to elementary-school students. These results highlight the importance of considering both children’s language background and test language proficiency in the diagnostics of specific learning disorder in mathematics, to enable a fairer diagnosis and, consequently, adequate support measures.