• Susanne Kuratli Geeler
  • Urs Grob
  • Aiso Heinze
  • Miriam Leuchter
  • Anke Lindmeier
  • Franziska Vogt
  • Elisabeth Moser Opitz
An instrument used for longitudinal studies must fulfill the criteria of time-related measurement invariance in addition to the
general scientific criteria. This is challenging for the assessment of numerical competencies in kindergarten due to the high pace of development at that age. The aim of this analysis was to investigate whether the development of numerical competencies of kindergarten
children can be reliably and undistortedly measured with a test based on the TEDI-MATH item pool. The sample consisted of 894 kindergarten children from Switzerland (n = 523) and Germany (n = 371). Based on the Rasch model analysis, the instrument proved to be reliable over time after excluding 17 items with time-related differential item functions (DIF). A subsequent CFA showed, however, that the
factorial structure was theoretically consistent but not completely stable over time. In addition, the abridged instrument showed DIF at the level of the country subsamples. These differences could be caused by different kindergarten pedagogies; they also raise the question of to what extent assessments carried out as early as at kindergarten age depend on the context.
Translated title of the contributionThe longitudinal assessment of numerical competence of pre-schoolers:: Analysing measurement invariance with the test TEDI-MATH
Original languageGerman
JournalDiagnostica
Volume67
Issue number2
Pages (from-to)62-74
Number of pages13
ISSN0012-1924
DOIs
Publication statusPublished - 04.2021

    Research areas

  • longitudinal study, measurement invariance, Rasch analysis, kindergarten, numerical competence

ID: 1442126