Assesseer leesbegripstoetse werklik leesbegrip?

Abstract

Opsomming. In hierdie artikel word die vraag gestel of leesbegripstoetse, wat volgens onderwysdepartementele voorskrifte opgestel is, werklik leesbegripsvlakke assesseer. Riglyne van die Nasionale Kurrikulum- en Assesseringsbeleidsverklaring (Suid- Afrika, 2011) is gebruik om twee gelyksoortige leesbegripstoetse (drie tekste elk), wat uit dieselfde tekstipes bestaan, maar handel oor verskillende temas, op te stel om Afrikaanssprekende graad 9-leerders, uit verskillende geografiese gebiede van ’n onderwysdistrik, se leesbegripsvlakke te bepaal. Die moeilikheidsgraad van die leestekste is met leesbaarheidsindekse bepaal en ’n 40-40-20%-verspreiding van kognitiewe vlakke is in die vraagstelling gevolg. Bo en behalwe die berekening van gemiddeldes, frekwensies en persentasies is die Cronbach Alphakoëffisiëntwaardes en die gemiddelde interitem-korrelasiewaardes vir elkeen van die drie tekste binne die twee tematies verskillende leesbegripstoetse vasgestel en leesbegripstoetse se vrae is volgens Bloom se kognitiewe vlakke vergelyk. Die hoofbevinding is dat leerders se leesbegripsvlakke nie akkuraat deur tipiese leesbegripstoetse geassesseer word nie, omdat hulle leesbegrippunte in ’n mindere of ’n meerdere mate op verskillende wyses gekontamineer word. Abstract. In this article the question is posed whether reading comprehension tests, compiled in accordance with department of education prescripts, actually assess reading comprehension levels. Guidelines of the National Curriculum and Assessment Policy Statement (South Africa, 2011) were used to compile two equivalent reading comprehension tests (three texts each), consisting of the same text types, on different themes, in order to determine the reading comprehension levels of Afrikaans-speaking grade 9 learners from different geographical areas of an education district. The degree of difficulty of the reading texts was determined by means of readability indexes and a 40-40-20%-distribution of cognitive levels was followed in the question statements. In addition to the calculation of averages, frequencies and percentages, the Cronbach Alpha coefficient values and the average interitem correlation values for each of the three texts within the two thematically different reading comprehension tests were determined, and the questions of the reading comprehension tests were compared in accordance with Bloom’s cognitive levels. The main finding is that readers’ reading comprehension levels are not accurately assessed by typical reading comprehension tests, seeing that their reading comprehension marks are contamininated to a lesser or a greater extent in various ways.