MODELOS ESTATÍSTICOS E PONTOS DE CORTE PARA AVALIAÇÃO DE COMPETÊNCIAS COGNITIVAS E LINGUÍSTICAS DE IMIGRANTES.

The literature shows no validated instruments for measuring various components (cognitive and language) of proficiency of immigrant school population, with recent entry in a host country such as Portugal. On the other hand, the results of tests for the evaluation of Second language acuity are difficult to interpret. To examine the validity of diagnostic evidence of second-language proficiency, specifically for the linguistic dimensions (spelling, semantic, syntactic, lexical) and cognitive (verbal reasoning, lexical decision, rhythm prosodic), it is necessary determine cut-off points controlling the following variables: chronological age (7-17 years), age of immigration, nationality of the child, the parents' nationality, diversity of mother tongues and residence in the host country within < eight years. In the study that is underway, the statistical analysis will be made using discriminant analysis and shall be measured sensitivity and specificity (comparison between the hits and errors) through the model Receiver Operating Characteristics (ROC). It is expected that the groups differ in accordance with the selected variables and this distinction holds up the definition of distinct cut-off points whereas the level of difficulty of the subject, the type of evidence and the idiosyncrasies of our host school context. [ABSTRACT FROM AUTHOR]