ERIC Number: ED397127
Record Type: RIE
Publication Date: 1995-Aug
Reference Count: N/A
A Note on the Relationship between Covariance Matrices and Consistently Estimated Variance Components. ACT Research Report Series 95-2.
Woodruff, David J.
The one observation per cell two-way items by examinees random effects analysis of variance (ANOVA) with all error components zero is considered. The estimated variance components are expressed as functions of the inter-item covariance matrix and the inter-examinee covariance matrix. These expressions show that under the random effects model if the inter-item and inter-examinee covariance matrices are unconstrained, then both the number of items and the number of examinees must approach infinity for the estimates of any of the variance components to be consistent. However, if these two covariance matrices are constrained so that each has homogeneous variances and covariances, then consistent estimates of the variance components can be obtained without having both the number of items and the number of examinees simultaneously approach infinity. (Contains seven references.) (Author/SLD)
ACT Research Report Series, P.O. Box 168, Iowa City, IA 52243.
Publication Type: Reports - Evaluative
Education Level: N/A
Authoring Institution: American Coll. Testing Program, Iowa City, IA.
Identifiers: Variance (Statistical)