The members of LaCPI also included postdoc Rikke Vang Christensen and eighteen students. We Running head: INFERENCES AND VOCABULARY IN READING COMPREHENSION 2 would like to thank all the members of LaCPI as well as the children, parents, and schools who participated in the study. AbstractWe examined the relationship between inference making, vocabulary knowledge, and verbal working memory on children's reading comprehension in 62 6 th graders (aged 12). The effect of vocabulary knowledge on reading comprehension was predicted to be partly mediated by inference making for two reasons: Inference making often taps the semantic relations among words, and the precise word meanings in texts are selected by readers on the basis of context. All independent variables were significantly and moderately correlated with reading comprehension. In support of our prediction, the link between vocabulary knowledge and reading comprehension was significantly mediated by inference making even when verbal working memory was controlled. An alternative mediation hypothesis (vocabulary as a mediator of the effect of inference making on comprehension) was not supported by the data. The study replicates and extends the findings of earlier work (Cromley &

Dyslexia is hard to diagnose in a second language. Poor performance on a test of reading may be caused by poor language proficiency in the second language or by limited schooling rather than by poor reading ability per se. This confound was supported in a study of 88 adult second language learners and 65 native language speakers. The incidence of dyslexia in the second language learners varied widely depending on the measure of reading. In order to reduce language and schooling confounds, a dynamic test of acquisition of basic decoding ability was developed. In the dynamic test, participants are taught three novel letters and to synthesise the letter sounds into new words. Results from the study indicated that the dynamic test provided results in accordance with the current IDA definition of dyslexia, while significantly reducing the influence second language vocabulary and amount of schooling. With the dynamic measure, the same cut-off point between dyslexic and non-dyslexic performance appeared valid in both native language speakers and second language learners.

No abstract

Tekster kan ses som instrukser til læseren om at aktivere og kombinere baggrundsviden til et forestillingsindhold (en mental model). Sådanne kombinationer kræver inferenser, dvs. tankeslutninger. Denne artikel drejer sig om to forsøg med undervisning i inferenser i læsning. Det første forsøg involverede 236 11-årige norske elever. Det indførte grafiske modeller til at illustrere, hvilke oplysninger der kommer fra teksten, og hvilke der kommer fra læserens baggrundsviden. Den positive effektaf forsøgsundervisningen på læseforståelsen var klokkeklar. Et andet forsøg omfattede 245 12-13-årige danske elever. Her var effekten langt mindre tydelig. Én vaesentlig forskel var, at den norske undersøgelse drejede sig om fagtekster modsat den mere skønlitterært prægede danske undersøgelse. Inferenser kan være en særlig udfordring i fagtekster, for fagtekster traekker i særlig høj grad på baggrundsviden, der ikke er almenmenneskeligt erfaret. Det er viden, som ikke sidder på rygmarveneller følger automatisk med indlevelse, men som skal mobiliseres aktivt. Og det kan direkte undervisning bidrage til.

No abstract