The absence of the invariant item ordering (IIO) property in two nonparametric polytomous item response theory (IRT) models is discussed, and two nonparametric models are discussed that imply an IIO. Only two parametric polytomous IRT models are found to imply an IIO. A method is proposed to investigate whether an IIO is implied with empirical…

Describes similarities and differences between additive conjoint measurement and the Rasch model, and formalizes some new nonparametric item response models that are, in a sense, probabilistic measurement theory models. Applies these new models to published and simulated data. (SLD)

Compared the analysis of variance (ANOVA) "F" and several nonparametric competitors for two-way designs for empirical alpha and power through simulation. Results suggest the ANOVA "F" suffers from conservative alpha and power for the mixed normal distribution, but is generally recommended. (Author/SLD)

Describes relationships among 20 polytomous item response theory (IRT) models (parametric and nonparametric) and 8 measurement properties relevant to polytomous IRT. Provides three tables to assist in the choice of an appropriate model and demonstrates the use of the models in test construction. (SLD)

This study looked at criteria important to dental accreditation consultants when evaluating programs that utilized some distance technology. Attitudinal and demographic data were collected to determine any relationships between these profiles and how a consultant would apply predetermined evaluative criteria. Descriptive statistics, internal…

Is the use of an online course delivery format, when compared with the more traditional face-to-face format, good or bad in the context of university education? Those who subscribe to the no-significant-difference perspective argue that online delivery is good, because it allows students with time and geographic distance constraints to obtain the…

In many practical settings, essentially the same differential item functioning (DIF) procedures have been in use since the late 1980s. Since then, examinee populations have become more heterogeneous, and tests have included more polytomously scored items. This paper summarizes and classifies new DIF methods and procedures that have appeared since…

Robustness and power of parametric, semi-parametric, and nonparametric tests of between-group discordance were compared in this simulation study. The empirical Type I error rates and power of nine tests were compared. When data were sampled from the any differences especially favor young women in single-sex Catholic secondary schools, and whether…

Predictive Sample Reuse is an alternative to parameter estimation for solving a variety of statistical problems. The technique is described and its relevance to criterion referenced measurement discussed. An item analysis procedure which does not require multiple test administrations or variability in test or item scores and which adheres to the…

Formulas and a FORTRAN program for computing Kendall's Tau as well as a generalized Spearman rho coefficient from ordered contingency tables are described. (Author)

Three different nonparametric tests for scale--the Siegel-Tukey (S-T), the Mood (M), and the Normal Scores (NS)--are compared in order to contrast varying methods of scale test development and usage. Procedures for developing the three scale tests are discussed, and two examples of the use of each test in solving the same problem are given. From…

This note described a FORTRAN IV, CDC, computer program for the canonical analysis of a two-way contingency table. (Author)

The Shine-Bower single subject analysis of variance is extended from being applicable to only an even number of trials to being applicable to any number of trials. (Author/JKS)

Uses Monte Carlo methods to assess the relative power of the paired samples t test and Wilcoxon's signed-ranks test under 10 population shapes. Concludes that, insofar as these two statistics are concerned, the often-repeated claim that parametric tests are more powerful than nonparametric tests is not justified. (Author/CB)

Real world data often fail to meet the underlying assumption of population normality. The Rank Transformation (RT) procedure has been recommended as an alternative to the parametric factorial analysis of Covariance (ANCOVA). The purpose of this study was to compare the Type I error and power properties of the RT ANCOVA to the parametric procedures…