CONTEXT EFFECT ON THE LINKING OF NAEP MARKET BASKET FORMS TO THE MAIN NAEP FORMS
Tay-Lim, S.H., Wang, M.M., Educational Testing Service, USA
The National Assessment of Educational Progress (NAEP) is the only regularly administrated, congressionally mandated national assessment program in the United States. The primary goal of NAEP is to measure changes in educational performance over time. The discussion of NAEP market basket began with the redesign effort in 1996. The purpose of market basket is to support faster and simpler reporting. Various options of market basket (MB) design have been considered, including the use of multiple short test forms, and summary reporting based on a large collection of. In NAEP 2000, trial MB short forms were assembled for the 4th grade math assessment and administered to a random subset of the NAEP sample. The MB booklets (forms) contain blocks of items that were drawn from the 1996 main NAEP blocks and the 1999 field tests. The items of these MB blocks do not appear in the same contexts as found in the main NAEP blocks. NAEP has learned some difficult lessons about the effects of making changes in item context and in other aspects of the assessment in the 1986 reading assessment, which has become known as the 1986 NAEP reading anomaly (Beaton & Zwick, 1990; Zwick, 1991). In order to evaluate the potential use of the MB forms for NAEP reporting, a study is design to address the issue of context effects on linking with the data collected in the NAEP 2000 MB forms.