Based upon the belief that examinee-by-item interaction should be conceptualized as true score variability rather than as a result of errors of measurement, Lu (1971) proposed a modification of Hoyt's analysis of variance reliability procedure. Lu claims that his approach is free from the "impurities" and "entanglements" inherent in Hoyt's procedure since with nonadditive data his approach supposedly yields an error MS which estimates just σ∈2, whereas the residual MS of Hoyt's procedure would estimate σ∈ 2 + σαβ2. Via a computer simulation study, it is shown that Lu's approach does not separate interaction from error.
Get full access to this article
View all access options for this article.
References
1.
Gaebelein, J. W., Soderquist, D. H., and Powers, W. A.A note on variance explained in the mixed analysis of variance model . Psychological Bulletin, 1976, 83, 1110-1112.
2.
Hoyt, C.Test reliability estimated by analysis of variance. Psychometrika, 1941, 6, 153-160.
3.
Hoyt, C. J. and Stunkard, C. L. Estimation of test reliability for unrestricted item scoring methods. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 1951, 12, 756-758.
4.
Hoyt, C. J. and Krishnaiah, P. R.Estimation of test reliability by analysis of variance . Journal of Experimental Education, 1960 , 28, 257-259.
5.
Lu, K. H.Statistical control of "impurity" in the estimation of test reliability. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 1971, 31, 641-655.
6.
Searle, S. R.Linear models. New York: John Wiley and Sons, 1971 .
