An empirical sampling study investigated six procedures for testing differences between means in the presence of unequal n's and variances. Support was obtained for previous research which found t robust to heterogeneous variances only when n's are equal and of moderate size. The procedure which emerged as providing the best control over Type I…

A general method of devising stepwise multiple testing procedures with fixed experimentwise error is applied to the problem of non-parametric randomized block design with ordered alternatives. In addition, the method is applied to other models with ordered alternatives. (Author)

An alternate procedure is presented for testing the trial-by-subject interaction in the Shine Combined analysis of variance test. This new procedure is designed to get around the independence and distributional problems of the orginal F test. (Author)

DSYSTM is an interactive computer program designed to perform a linear-graph analysis of contingency table data. The program is especially suitable for the analysis of change models. (Author/JKS)

In this paper, the author gives a review of the literature on complex multivariate distributions. Some new results on these distributions are also given. Finally, the author discusses the applications of the complex multivariate distributions in the area of the inference on multiple time series. (Author)

Recent results of Bloomfield and Watson and Knott are used to derive a class of union-intersection tests for sphericity from likelihood ratio tests of independence of two sets of variates. It is shown that the ordinary likelihood ratio test for sphericity has a natural union-intersection interpretation. (Author/RC)

Previous arguments against the use of Analysis of Variance (ANOVA) to determine test bias are extended. Explains why ANOVA may obscure test bias when it exists and also create a false impression of bias when it does not exist. (RB)

This paper provides the computational formulas necessary for testing the significance of the difference between mean values of two bivariate normal populations. (Author)

It is suggested that the binomial theorem be used to compute the probability that a given number of Type I errors would occur when a group of null hypotheses are true and that this result be used as the level of significance for the test of an overall hypothesis. (Author/BW)

A review of recent volumes of the "Research Quarterly" reveals that few of the studies employing multiple dependent measures have attended to the appropriate analysis. (JD)

Highlights some of the major questions facing the measurement community in deciding how to conceptualize validity in the revised "Standards for Educational and Psychological Testing" of the American Association for Educational Research, the American Psychological Association, and the National Council on Measurement in Education. Focuses…

An indispensable part of any curriculum design in an educational setting is the analysis of the needs of the learners involved in the context. The needs can be addressed from different perspectives. Among them, the learners' needs in terms of their perceptions toward what constitute learning/teaching and testing processes are of prominent values.…