David J. Krus

INTERPRETATION OF CANONICAL ANALYSIS: ROTATED VS, UNROTATED SOLUTIONS

Orthogonal rotation of canonical variates is shown to preserve the major properties of the canonical solution and may increase its interpretability.

Demonstration of properties of a suppressor variable.

A general logical model of properties of suppressor variables is proposed. Consistent exploration of possible manifestations of suppressor variables within this theoretical framework accounts for extant classifications of suppressor variables into the classical, net, and cooperative categories and suggests existence of new subcategories, not detected previously. The discussed model leads to consistent identification and classification of suppressor variables and facilitates computer simulation.

Order Analysis: An Inferential Model of Dimensional Analysis and Scaling.

Order analysis was proposed as a method for description of formal structures in multidimensional space. Its algorithm was derived using a combination of psychological measurement theory, formal logic theory, information theory and graph theory concepts. It suggests isomorphism between algebraic, geometric, logical, and cognitive structures. The model also provides for adjustment of its sensitivity to random variation with an elective degree of statistical confidence.

The Problem of Negative Reliabilities.

Negative coefficients of reliability, sometimes returned by the standard formulae for estimation of the internal-consistency reliability, are neither theoretically nor numerically correct. Alternate strategies for test development in this special case are suggested.

Using Order Analytic Instructional Hierarchies of Mnemonics to Facilitate Learning Chinese and Japanese Kanji Characters

Abstract The effects of applying an order-analytical, multidimensional scaling method for determining hierarchies among a list of Chinese and Japanese kanji characters and mnemonics were explored. The authors used those hierarchies to design a computer-assisted, instructional program for learning the characters. The participants were presented a series of kanji characters in a hierarchical or random order, with and without mnemonics. Recall and recognition tests were administered immediately after the presentation and 48 hr later. Characters that were arranged in a hierarchical sequence or those presented with mnemonics were better learned and remembered than those that were presented in a random order or without mnemonics. The interaction of sequence and mnemonics, however, did not yield significant effects for learning. The results suggest that techniques such as order analysis that scale and organize variables according to their variance contribution and dominance can be useful for designing effective ...

Changing Values over the Last Half-Century: The Story of Thurstone's Crime Scales

The present article is a follow-up of Thurstones study, published 50 yr. ago, concerning the perceived seriousness of crimes. Comparisons were also made with Coombs follow-up study from the mid-sixties. The continuing trend toward the judgment of sex offenses as less serious and offenses against the person as more serious was observed. The obtained scale also showed heightened sensitivity of subjects to offenses such as burglary, libel, counterfeiting, and forgery, reflecting the prevailing public opinion during the political trials of the mid-seventies.

Lost: McCall's T Scores: Why?:

A description of a FORTRAN program for linear and area transformations of test scores with optional generation of symmetric tables of areas under the standard normal curve is presented. Also included is a historic note on the origin of the T scale along with a discussion of the relative merits of area versus linear scale transformations.

Normal Scaling of the Unidimensional Dominance Matrices: The Domain Referenced Model.

Described is an alternative algorithm for Thurstones pair comparisons method. This algorithm provides an improved capacity for handling indeterminate proportions of 1.00 and 0.00 as well as untruncated estimates of population scale values.

Empirical Comparison of Factor and Order Analysis on Prestructured and Random Data.

Results of empirical comparisons of an inferential model of order analysis with factor analytic models were reported for two sets of data. On the prestructured data set (Thurstones box data) both order and factor analytic models returned its dimensions of length, width and height. However, on the random data set (replication of the Armstrong and Soelberg study), only the factor analytic models indicated the presence of nonexistent dimensionality.

Rotation in Canonical Variate Analysis

Description of a computer program performing rotated canonical variate analysis was presented. The program was written as compatible with the IBM Scientific Subroutines Package. Examples of rotated and unrotated solutions of a sample problem were given, together with their interpretations illustrating the advantages of the rotated solution.