Tekstdekking in Groep 5

Research on primary school children shows that the differences between L1-Dutch children on the one hand, and L2-Dutch children on the other, tend to become greater, which is a process in which the children's proficiency in Dutch plays a major role. In this paper it is shown that word knowledge is the main source of this Matthew-effect. The level of difficulty of ten different school texts is investigated with respect to the relative number of unknown words (lemmas) in these texts for DFL and DSL-speakers, who have an absolute receptive vocabulary size in Dutch of 7.000 and 4.000 words respectively. On the basis of a word frequency list in primary education, covering percentages (lemmas and types) are calculated for both groups, for ten texts on various school subjects (five on reading, five on science). The mean (lemma) coverage of the reading text was 89% for DFL-, and 84% for DSL-children; for the science texts these percentages were 85% for DFL-, and 78% for DSL-speakers. Comparing these outcomes with what is known about understanding texts and learning words from texts, it is shown that for the DFL-children, only two science texts were quite difficult. For the DSL-children, however, all the science texts, and two of the five reading texts were too difficult. They could only marginally understand what the text was about, and learning new words from the text was almost impossible, due to the high percentage of unknown words.