The Construct Validity of a Performance-Based Assessment Program.Report as inadecuate

The Construct Validity of a Performance-Based Assessment Program. - Download this document for free, or read online. Document in PDF available to download.

The purpose of this study was to examine the construct validity of a performance assessment program, the Maryland School Performance Assessment Program (MSPAP). Based on analyses of the longitudinal associations of Grade 5 MSPAP data in 1996 with Grade 3 MSPAP data in 1994, the following hypothesis was examined: the unattentuated correlation or the group-mean correlation between two similar measures of the same content area is higher than its correlations with different content areas. This hypothesis was not supported. In addition, the results analyzed by structural equation modeling (SEM) of this longitudinal correlation matrix reveal that the SEM model specified by the MSPAP six latent traits was unable to capture the underlying information of this data. Extra factors, such as a general ability and an assessment method effect, may need to be considered for better fitting data. SEM was performed on the multitrait-multimethod correlation data, and the traits of Reading and Mathematics were assessed by MSPAP and the Comprehensive Test of Basic Skills (CTBS). The trait effects of MSPAP reading and CTBS mathematics application may be attenuated by the method effects of the performance-based assessment and the multiple-choice assessment, respectively. (Contains 6 figures, 7 tables, and 19 references.) (Author/SLD)

Descriptors: Construct Validity, Correlation, Elementary Education, Elementary School Students, Grade 3, Grade 5, Multiple Choice Tests, Performance Based Assessment, State Programs, Structural Equation Models, Testing Programs

Author: Li, Yuan H.; Ford, Valeria; Tompkins, Leroy J.


Related documents