This paper describes quantitative tests which show the sophisticated-guessing model to be better than the criterion bias model in its ability to predict the outcome of certain kinds of recognition experiments. (Author/RK)

Study compares the Rasch and the Guttman models of measurement and thus adds to the description of the characteristics of Rasch's logistic model. Such knowledge is of importance in making decisions as to which model and which statistics should be used in evaluations of tests. (Author/CB)

In this comment a recent attempt by Samuel A. Livingston to develop a theory of reliability for criterion-referenced measures is critiqued. For Livingston's rejoinder see TM 500 560. (Authors/MB)

The unique conceptual framework and language of generalizability theory are presented. While this chapter is relevant to any area in which generalizability theory is applicable, it emphasizes evaluation research, and most examples come from that area. (Author/PN)

Outlines four limitations inherent in using a minimum competency test to determine eligibility for a high school diploma and, in some jurisdictions, to determine advance to the next grade. Discusses the appropriate uses of testing in an overall program.(MP)

Studied the effects of experience in administration and scoring of the Wechsler Intelligence Scale for Children (Revised) on types of examiner errors. Results showed total score and administrative error rates dropped significantly with experience and feedback, but response scoring errors, mathematical errors, and IQ errors were not reduced…

Presents the results of a study on data collection instruments, specifically questionnaires, used by libraries in evaluating resource use and institutional effectiveness. A number of causes for inconsistent survey results are identified and problems with imprecision in survey question terminology are discussed. (JL)

Lord's, Millman's and Saupe's methods of approximating the standard error of measurement are reviewed. Through an empirical demonstration involving 200 university classroom tests, all three approximations are shown to be biased. (Author/JKS)

Two estimation procedures for the Rasch Model of test analysis are reviewed in detail, particularly with respect to new developments that make the more statistically rigorous conditional maximum likelihood estimation practical for use with longish tests. (Author/JKS)

In two treatment-repeated measurements designs, the ratio between the unbiased variance of the differences and twice the variance of the errors of measurement can be used to test for interaction of subjects and treatments. The use of this statistic is illustrated. (Author/CP)

Multiple-choice tests composed entirely of two-, three-, or four-choice items were investigated. Results indicated that number of alternatives per item was inversely related to item difficulty, but directly related to item discrimination. Reliability and standard error of measurement of three-choice item tests was equivalent or superior.…

Test-retest correlations can lead to biased reliability estimates when there is instability of true scores and/or when measurement errors are correlated. Using three administrations of the Test of Standard Written English and essay ratings, an analysis is demonstrated which separates true score instability and correlated errors. (Author/BW)

This paper elaborates on the work of Cardinet, and others, by clarifying some points regarding calculations, specifically with reference to existing computer programs, and by presenting illustrative examples of the calculation and interpretation of several generalizability coefficients from a complex six-facet (factor) design. (Author/RL)