A model for multiple choice test-taking behavior is proposed which is different from those presently used for item sampling theory. A new theory is developed, which includes a new concept to facilitate comprehension of item sampling theory, a "number-known" score distribution. A major advantage of the model is that it accommodates data from…

A reanalysis of Shaycroft's matrix of intercorrelations of 10 test variables plus 4 random variables is discussed. Three different procedures were used in the reanalysis: (1) Image Component Analysis, (2) Uniqueness Rescaling Factor Analysis, and (3) Alpha Factor Analysis. The results of these analyses are presented in tables. It is concluded from…

CERTAIN SPECIFIC ASPECTS OF HYPOTHESIS TESTS USED FOR ANALYSIS OF RESULTS IN RANDOMIZED EXPERIMENTS WERE STUDIED--(1) THE DEVELOPMENT OF THE THEORETICAL FACTOR, THAT OF PROVIDING INFORMATION ON STATISTICAL TESTS FOR CERTAIN EXPERIMENTAL DESIGNS AND (2) THE DEVELOPMENT OF THE APPLIED ELEMENT, THAT OF SUPPLYING THE EXPERIMENTER WITH MACHINERY FOR…

The univariate and multivariate models for the analysis of covariance are compared for the case where an experimental design contains between and within subject factors, one dependent variable, and one observation per subject. (Author/JKS)

Marascuilo's and Levin's use of the term "nesting" to describe a statistical model, and application of the Scheffe tests on restricted sets of contrasts, is opposed. Changing the model during analysis of a complex set of data is both necessary and desirable. (Author/CP)

A new method of measuring item bias based on the latent class model proposed by the author is suggested. A test for item bias is also suggested that is based on standard asymptotic results. (Author/DWH)

Two alternatives to the problems of conducting planned and post hoc comparisons in tests of concordance and discordance for G groups of judges are examined. The two models are illustrated using existing data. (Author/JKS)

Using LISREL, the only types of constraints allowed are fixing parameters at a constant value and constraining parameters to be equal. In this paper, two new concepts ("phantom" and "imaginary" latent variables) are introduced which allow fairly general equality and inequality constraints on factor loadings and structural model…

Investigates implications for finite and known item populations for classical test theory, and the alpha coefficient among items in paper-and-pencil testing. (Author/AG)

Typically, nonlinear models such as those used in the analysis of covariance structures, are not globally identifiable. Investigations of local identifiability must either yield a mapping onto the entire parameter space, or be confined to points of special interest such as the maximum likelihood point. (Author/JKS)