EXAMING the DIMENSIONALITY of EARLI NUMERACY SKILL MEASURES

Open Access
Author:
Cheng, Weiyi
Graduate Program:
Educational Psychology
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
November 18, 2011
Committee Members:
  • Pui Wa Lei, Thesis Advisor
  • Hoi Kin Suen, Thesis Advisor
Keywords:
  • VALIDITY
  • EARLY ASSESSMENT
  • DIMENSIONALITY
Abstract:
ABSTRACT Test validity is a critical issue in test construction and evaluation. Evidence for construct validity can be gathered by examining the factor structure of a test (Messick, 1995). One assumption of many assessments is unidimensionality, which means that the test is constructed to represent a single attribute (Banerji, Smith, & Dedrick, 1997). The Early Arithmetic, Reading, and Learning Indicators (EARLI, Diperna, Morgan, & Lei, 2006) is a recently developed screening instrument that aims at providing practitioners with a psychometrically sound assessment to assess key academic skills and monitor growth of children at preschool age. The current study examined the dimensionality of the EARLI numeracy measures, which target key skills such as number identification, counting, and basic arithmetic. Considering the small sample size of the study, three programs, NOHARM, DIMTEST and DETECT were selected to provide a comprehensive picture of the underlying structure of the EARLI numeracy measures. Initial evidence of unidimensionality was obtained for some of the measures (Counting Aloud, Counting Object, and Measurement), which contributed to the support of their use in preschools. However, the dimensionality for two measures (Number Naming and Pattern Recognition), remained inconclusive due to discrepancy in results between programs. Further evidence concerning dimensionality is needed before any substantive conclusion can be made. Our literature review revealed the lack of empirical studies on dimensionality evaluation with small sample sizes. Some inconsistent results found in this study suggested that more research is needed to provide guidelines for dimensionality assessment with small samples.