Abstract: This paper is the first of a series that will compare estimates of error that arises when state assessments are linked to the National Assessment of Educational Progress (NAEP). Different forms of linkage are discussed. Comparisons are made between whole-sample regression, repeated half-sample replication, bootstrap, and jackknife estimates of the proportion of variance explained by the linkage function, as well as the standard error of linkage. Data from four states that participated in a study to link their state assessments to the 1998 (State) NAEP fourth and eighth grade reading assessments suggest that each of the methods produces comparable estimates of these error quantities when schools are treated as the primary sampling unit. (Contains 4 tables, 2 figures, and 13 references.) (Author/SLD)

Title: Estimating Error in State-to-NAEP Linkages, Part I: Mean Plausible Value Distributions from a Simple Model.
Author: Arenson, Ethan
Note: Paper presented at the Annual Meeting of the American Educational Research Association (New Orleans, LA, April 24-28, 2000).
Page Length: 38
Publication Year: 2000
Document Type: Research Report (143); Conference Paper (150)
Target Audience: Practitioners and Teachers
ERIC Identifier: ED442854
Clearinghouse Identifier: TM031274
You may be able to order this document.

Descriptors: Elementary Secondary Education; * Equated Scores; Error of Measurement; * Estimation [Mathematics]; National Competency Tests; Reading Tests; Regression [Statistics]; State Programs; * Test Results; Testing Programs

Identifiers: Bootstrap Methods; Jackknifing Technique; *Linking Metrics; *National Assessment of Educational Progress