Network


Latest external collaboration on country level. Dive into details by clicking on the dots.

Hotspot


Dive into the research topics where James D. Church is active.

Publication


Featured researches published by James D. Church.


Journal of Verbal Learning and Verbal Behavior | 1976

Comments on Clark's “The language-as-fixed-effect fallacy”

Edward L. Wike; James D. Church

Clarks arguments for treating language materials as random rather than fixed effects are examined, and the problems with random effects designs and approximate statistical tests (quasiF-ratios) are reviewed. In view of the difficulties with Clarks recommended procedures and the present lack of knowledge regarding approximate tests, it is suggested that researchers use fixed factors, which are better understood statistically, and seek nonstatistical generality by means of replication.


Bulletin of the psychonomic society | 1976

The robustness of homogeneity of variance tests for asymmetric distributions: A Monte Carlo study

James D. Church; Edward L. Wike

Robustness properties of three equality of variances tests, the jackknife procedure, the k-sample Box-Andersen test, and Levene’s z test, were investigated and compared in a Monte Carlo study employing small samples from underlying gamma distributions and gamma mixtures. Overall, the jackknife test yielded rejection proportions usually closer to nominal levels under departures from normality, but the Box-Andersen test performed relatively well when departures from normality were not extreme. Four recommendations regarding testing variances were offered.


Bulletin of the psychonomic society | 1979

A Monte Carlo study of nonparametric multiple-comparison tests for a two-way layout

James D. Church; Edward L. Wike

A Monte Carlo study of two overall tests, the Friedman and Doksum, and five nonparametric multiple-comparison tests, the McDonald-Thompson, Nemenyi, Wilcoxon, Rhyne-Steel, and sign test, was done to determine the tests’ Type I error rates in a two-way layout with k = 3, 5, and 7 treatments and n = 8, 11, and 15 blocks. It was found that (1) the Friedman test was superior to the Doksum test as an overall test, and (2) the Wilcoxon signed-ranks test, when protected by a significant Friedman test, was the best pairwise multiple-comparison procedure. In a second Monte Carlo study, a modified sign test, termed the stepped-down sign test, was found to be superior to the Wilcoxon test as a protected test; it is recommended for pairwise comparisons in a two-way layout.


Bulletin of the psychonomic society | 1977

Analysis of variance methods for the design and analysis of Monte Carlo statistical studies

Edward L. Wire; James D. Church

It was proposed that the data from Monte Carlo statistical investigations be subjected to analysis of variance methods rather than the conventional techniques of tabling, graphing, and inspecting the data. Two examples in which analysis of variance methods were applied to published Monte Carlo studies were presented. It was suggested that balanced factorial designs should be used whenever possible in Monte Carlo studies so that analysis of variance methods would be directly applicable. Finally, three advantages of analysis of variance methods over conventional techniques were described and analysis of variance methods were contrasted with a possible alternative analytical method, multiple regression analysis.


Bulletin of the psychonomic society | 1982

Nonrobustness in F tests: 1. A replication and extension of Bradley’s study

Edward L. Wire; James D. Church

Portions of Bradley’s Monte Carlo study on the effects of violations of assumptions on the F test were replicated. The empirical ps agreed well with those observed in the original study. Reciprocal, loge, and square-root transformations failed to reduce the differences between the empirical ps and nominal as. An analysis of the relative importance of the factors in the Monte Carlo design suggested other statistical procedures that might be applied to ameliorate the difficulties posed by Bradley’s results.


Bulletin of the psychonomic society | 1978

A Monte Carlo investigation of four nonparametric multiple-comparison tests for k independent groups

Edward L. Wike; James D. Church

A Monte Carlo study of four nonparametric multiple-comparison tests (t), the Wilcoxon, Steel, Nemenyi, and Ryan tests, was done to determine their Type I error rates when overall Kruskal-Wallis H tests were and were not significant. A total of 13,000 simulated experiments, N, were performed in a design with k = 3, 5, 7, and 9 groups and n = 7, 10, and 15 replicates per group at a =.05 and and a =.01. It was found that: (1) most of the variance in the error rates was accounted for by k, t, and k × t; (2) when the Wilcoxon test was used for pairwise comparisons after a significant H test, it was the best of the four procedures. The latter finding is concordant with recent studies of parametric multiple-comparison tests.


Bulletin of the psychonomic society | 1982

Nonrobustness in F tests: 2. Further extensions of Bradley’s study

Edward L. Wire; James D. Church

In further extensions of Bradley’s work on the effects of violations of assumptions in analysis of variance, random samples of k=3 and 4 and n=8 were drawn from 16 population combinations. The obtained values were subjected to four statistical procedures: the trimmed-means Ft test, the Kruskal-Wallis H test, Welch’s W test, and the Brown-Forsythe F* test. The FT test performed more poorly (had greater discrepancies between empirical ps and nominal as) than the regular F test. The F* test was the best procedure, but, like the F test, it had excessive ps with population combinations that had a starting L-shaped population and heterogeneous variances. It was concluded that none of the four statistical procedures wholly resolved the problems posed by Bradley’s results.


Psychological Reports | 1980

TWO MONTE CARLO STUDIES OF SILVERSTEIN'S NONPARAMETRIC MULTIPLE COMPARISON TESTS

James D. Church; Edward L. Wike

Two Monte Carlo studies were done to find the Type 1 error rates for Silversteins nonparametric pairwise multiple-comparison tests for a one- and two-way layout with ranked data. Silversteins tests had excellent experimentwise error rates but did not do as well as Wilcoxons tests and the stepped-down sign test when pairwise comparisons were performed after significant over-all tests. Silversteins tests were shown to be equivalent to recently proposed tests by Levy.


Perceptual and Motor Skills | 1977

Further Comments on Nonparametric Multiple-Comparison Tests

Edward L. Wike; James D. Church

A Monte Carlo study of four nonparametric multiple-comparison tests (the Ryan, Nemenyi, Steel, and Wilcoxon) for k independent groups indicated that the Wilcoxon rank-sum test performed optimally when it was used after a significant over-all Kruskal-Wallis H test.


Psychological Reports | 1981

SILVERSTEIN'S NONPARAMETRIC MANY-ONE TEST FOR A TWO-WAY DESIGN: A MONTE CARLO STUDY

James D. Church; Edward L. Wike

A Monte Carlo study was done to find the Type I error rates for three nonparametric procedures for making k – 1 many-one comparisons in a two-way design. The tests were the Silverstein and Steel many-one tests and the two-sample step-down sign test. k = 3, 5, 7, and 10 treatments were crossed with n = 8, 11, and 15 blocks with 1000 simulations per k, n combination. The Silverstein test had the best experimentwise error rates and is recommended for many-one comparisons in a two-way design.

Collaboration


Dive into the James D. Church's collaboration.

Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar

Lester Lipsky

University of Connecticut

View shared research outputs
Researchain Logo
Decentralizing Knowledge