Literaturnachweis - Detailanzeige
| Autor/in | Palaszewski, Bo |
|---|---|
| Titel | Multidimensional Models for Matrix-Sampled Reading Tasks. |
| Quelle | (1997), (22 Seiten)
PDF als Volltext kostenfreie Datei |
| Sprache | englisch |
| Dokumenttyp | gedruckt; online; Monographie |
| Schlagwörter | Adults; Foreign Countries; International Education; International Studies; Item Response Theory; Literacy; Matrices; Multivariate Analysis; Reading; Sampling; Scaling; Test Items; Sweden; International Adult Literacy Survey |
| Abstract | The International Adult Literacy Survey (IALS) was conducted in 1994 in seven European countries with the aim of producing literacy profiles to be used in individual national analyses as well as in international comparisons. It was assumed that the test items could be divided into three unidimensional scales corresponding to the aspects "Prose" (ability to read and comprehend text), "Doc" (document, comprehension of texts with charts and figures), and "Num" (numeracy, the ability to make calculations from text information). The assumption of unidimensionality is fundamental in the item response theory (IRT) estimation technique used to rescale the results of test subjects into a common scale. This paper explores this assumption for IALS results using models for multivariate data. It is shown that it is possible to identify a general dimension, measuring a general literacy ability, as well as more specific dimensions. Data from the Swedish IALS for 2,919 subjects were analyzed with a structural modeling approach that coped with the structural missing data introduced by the balanced incomplete block design. The final nested-factor model fitted on the passage scores obtained a reasonably good fit. The general literacy dimension in the Swedish test, defined as "Lit," did explain most of the variability of the data. The general dimension did not differ across gender, but the numerical dimensions exhibited different patterns for males and females. The analysis supports the existence of several latent dimensions in addition to the three originally postulated. The existence of block factors, probably related to speededness, is not surprising. Results indicate discernible passage effects that need not be a serious reliability problem, but that do merit further study. Two appendixes present standardized estimates and an analysis based on item levels. (Contains 1 figure, 8 tables, and 13 references.) (SLD) |
| Erfasst von | ERIC (Education Resources Information Center), Washington, DC |
Literaturbeschaffung und Bestandsnachweise in Bibliotheken prüfen
Permalink als QR-Code
Inhalt auf sozialen Plattformen teilen (nur vorhanden, wenn Javascript eingeschaltet ist)
Teile diese Seite:
