The aim of this master's thesis was to evaluate the justification of equating results from the National Assessment of Knowledge for 9th-grade primary school students. The equating of results was based on an approach rooted in Item Response Theory, known as true score equating. Due to the nature of the data provided by the National Examination Center, methods based on the use of anchor items were applied. These are items that were repeated across different administrations of the assessment. To evaluate the appropriateness of the equating process, both systematic and random equating errors were analyzed. Systematic error was assessed through a qualitative analysis of the entire equating procedure based on established guidelines from the literature. Random error was evaluated using the average standard error of equating. The full equating procedure was possible for the subjects of Biology, History, and Physical Education. Due to serious violations of key assumptions, equating could not be conducted for Mathematics, Slovenian, English, and German. In subjects where equating was feasible, the average equating error remained below the set threshold in all cases. Nevertheless, certain aspects of the equating process did not fully comply with recommended guidelines. Based on these shortcomings, the thesis offers concrete proposals for improving the current equating procedure.
|
