A Monte Carlo simulation study was conducted to determine the bootstrap correction formula yielding the most accurate confidence intervals for robust measures of association. Confidence intervals were generated via the percentile, adjusted, BC, and BC(a) bootstrap procedures and applied to the Winsorized, percentage bend, and Pearson correlation…

Meta-analytic techniques were used to summarize the statistical robustness literature on Type I error properties of alternatives to the one-way analysis of variance "F" test. The James (1951) and Welch (1951) tests performed best under violations of the variance homogeneity assumption, although their use is not always appropriate. (SLD)

Use of the method of reciprocal biweighted means (MBM) for dealing with the outlier problem in dual scaling compared favorably with other robust estimation procedures, such as the method of trimmed reciprocal averages. Four datasets from previous studies were used. Advantages of the MBM are discussed. (SLD)

Explores the robustness of conclusions from a statistical model against variations in model choice with an illustration from G. Box and G. Tiao (1973). Suggests that simultaneous consideration of a class of models for the same data is sometimes superior to analyzing the data under one model and demonstrates advantages to Adaptive Bayesian…

Civic competence is critical to the successful functioning of pluralistic democracies. Developing the knowledge, skills, and motivations for effective democratic participation is a national and global imperative that many higher education institutions have embraced through the teaching strategies of community-based learning and service-learning.…

Four-year colleges and universities submit faculty teaching load and instructional cost data annually to the Delaware Study of Instructional Costs and Productivity. While the Delaware Study currently adjusts the calculation of annual FTE students to account for the difference in annual student credit hours (SCH) earned by students at semester and…

This paper presents a new polychoric instrumental variable (PIV) estimator to use in structural equation models (SEMs) with categorical observed variables. The PIV estimator is a generalization of Bollen's (Psychometrika 61:109-121, 1996) 2SLS/IV estimator for continuous variables to categorical endogenous variables. We derive the PIV estimator…

This paper re-considers some of the evidence for low and declining social mobility in Britain, showing that one study based on a re-analysis of cohort figures appears to have had an impact on policy-makers out of all proportion to its scale and rigor. The study claimed to show that the income of parents and children were more closely related for…

Methods for the analysis of within-subjects effects in multivariate groups by trials repeated measures designs are considered in the presence of heteroscedasticity of the group variance-covariance matrices and multivariate nonnormality. Under a doubly multivariate model approach to hypothesis testing, within-subjects main and interaction effect…

Theoretical hypotheses generated from data analysis of a single sample should not be advanced until the replicability issue is treated. At least one of three questions usually arises when evaluating the invariance of results obtained from a canonical correlation analysis (CCA): (1) "Will an effect occur in subsequent studies?"; (2)…

This paper gives concise descriptions of a robust location statistic, the remedian of P. Rousseeuw and G. Bassett (1990) and a robust measure of dispersion, the "Sn" of P. Rousseeuw and C. Croux (1993). The use of Sn in least absolute errors regression (L1) is discussed, and BASIC programs for both statistics are provided. The remedian…

Two studies, both using Type 12 sampling, are presented in which the effects of violating the assumption of essential parallelism in setting confidence intervals are studied. Results indicate that as long as data manifest properties of essential parallelism, the two methods studied maintain precise Type I error control. (SLD)

Reviews methods for analyzing repeated measures data in addition to the conventional and corrected degrees of freedom univariate and multivariate solutions. Reviews the literature regarding recent procedures with respect to robustness, ability to handle missing data, and availability of software to obtain numerical results. (SLD)

The percentage bend correlation coefficient is introduced as one way to address the problem that the usual correlation coefficient is highly nonrobust. While this method is not a replacement for the usual test, it can offer advantages in terms of power and Type I errors. (SLD)

The phimax adjustment to phi (Pearson's product moment correlation applied to binary data) is explored. Phi/phimax is shown to be a measure of relationship apart from its affiliation with phi. The relations between phi/phimax and kappa and limitations of phi/phimax are discussed. (SLD)