This paper presents a contribution to the sampling theory of a set of homogeneous tests which differ only in length, test length being regarded as an essential test parameter. Observed variance-covariance matrices of such measurements are taken to follow a Wishart distribution. The familiar true score-and-error concept of classical test theory is employed. Upon formulation of the basic model it is shown that in a combination of such tests forming a “total” test, the singal-to-noise ratio of the components is additive and that the inverse of the population variance-covariance matrix of the component measures has all of its off-diagonal elements equal, regardless of distributional assumptions. This fact facilitates the subsequent derivation of a statistical sampling theory, there being at most m + 1 free parameters when m is the number of component tests. In developing the theory, the cases of known and unknown test lengths are treated separately. For both cases maximum-likelihood estimators of the relevant parameters are derived. It is argued that the resulting formulas will remain reasonable even if the distributional assumptions are too narrow. Under these assumptions, however, maximum-likelihood ratio tests of the validity of the model and of hypotheses concerning reliability and standard error of measurement of the total test are given. It is shown in each case that the maximum-likelihood equations possess precisely one acceptable solution under rather natural conditions. Application of the methods can be effected without the use of a computer. Two numerical examples are appended by way of illustration.

When items cannot be answered correctly by guessing, certain two-stage testing procedures are about as effective over the ability range of interest as the “best” up-and-down procedures studied previously. When answers can be guessed correctly 20 percent of the time, no two-stage procedure is found to match the “best” up-and-down procedures over this ability range. Feet-on-the-desk designs for two-stage procedures may produce poor results.

Suppose D is a data matrix for N persons and n variables, and \documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\dot D$$\end{document} is the matrix obtained from D by expressing the variables in deviation-score form. It is shown that if D has rank r, \documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\dot D$$\end{document} will always have rank (r − 1) if r = N < n, otherwise it will generally have rank r. If \documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\dot D$$\end{document} has rank s, D will always have rank s if s = n, but if s<n it will generally have rank (s + 1). Thus two cases can arise, Case A in which D has rank one greater than \documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\dot D$$\end{document}, and Case B in which D has rank equal to \documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\dot D$$\end{document}. Implications of this distinction for analysis of cross products versus analysis of covariances are briefly indicated.

A simple normal approximation is supplied to test the equality of two independent Chi-square variables. The approximation appears to be valid for degrees of freedom as small as two. Thus no new table of critical values is needed. The standard unit normal distribution tables are sufficient.

The present note illustrates the application of Lancaster & Hamdan's [1964] polychoric series method for estimating the correlation coefficient in contingency tables. A simple format for the calculations involved, using a desk calculator, is suggested and hence applied to a specific 3 × 3 contingency table.

A general one-way analysis of variance components with unequal replication numbers is used to provide unbiased estimates of the true and error score variance of classical test theory. The inadequacy of the ANOVA theory is noted and the foundations for a Bayesian approach are detailed. The choice of prior distribution is discussed and a justification for the Tiao-Tan prior is found in the particular context of the “n-split” technique. The posterior distributions of reliability, error score variance, observed score variance and true score variance are presented with some extensions of the original work of Tiao and Tan. Special attention is given to simple approximations that are available in important cases and also to the problems that arise when the ANOVA estimate of true score variance is negative. Bayesian methods derived by Box and Tiao and by Lindley are studied numerically in relation to the problem of estimating true score. Each is found to be useful and the advantages and disadvantages of each are discussed and related to the classical test-theoretic methods. Finally, some general relationships between Bayesian inference and classical test theory are discussed.

The problem of whether a precise connection exists between stochastic processes of the type considered in mathematical learning theory and the Guttman simplex is investigated. The approach is to consider a class of models that characterize the sequential properties of discrete data, and to derive a set of conditions which a model must satisfy in order to generate inter-trial correlations rij with the ‘perfect simplex’ property: rik = rijrjk for all trials i < j < k. It is shown that the Chapman-Kolomogorov Equations provide a necessary and sufficient condition for this property to hold. It follows that a process which is Markovian in the errors and successes will have this property. It is also shown that, if certain reliability assumptions are introduced, then the all-or-none model has the simplex property if the appropriate correlation coefficients are corrected for attenuation.

Recent studies of sequential decision making performance have utilized optimal dynamic programming solutions as the criteria for evaluating the quality of human decisions. The difficulties incumbent in increasing the dimensionality of dynamic programming formulations effectively prohibit their extension beyond the case of a single decision variable. In the following article an alternative optimization formulation for a two-decision-variable case is presented, followed by the development of a corresponding one-decision-variable problem. In addition, related methodological problems limiting the comparability of results among previous studies are presented. It is suggested that the proposed problem formulation and its accompanying interpretation for the decision maker largely eliminate these previous methodological problems. Furthermore, the problem formulation permits quantification of the relative weights that should optimally be placed on future decision periods in the multiple-stage decision problem.