Problems inherent in relying solely on statistical significance testing as a means of data interpretation are reviewed. The biggest problem with statistical significance testing is that researchers have used the results of this testing to ascribe importance or meaning to their studies where such meaning often does not exist. Often researchers…

The original nine programs for semantic differential analysis have been condensed into three programs which have been further refined and augmented. They yield: (1) means, standard deviations, and standard errors for each subscale on each concept; (2) Evaluation, Potency, and Activity (EPA) means, standard deviations, and standard errors; (3)…

Population invariance is an important requirement of test equating. An equating function is said to be population invariant when the choice of (sub)population used to compute the equating function does not matter. In recent studies, the extent to which equating functions are population invariant is typically addressed in terms of practical…

Researchers increasingly recognize that significance tests are limited in their ability to inform scientific practice. Common errors in interpreting significance tests and three strategies for augmenting the interpretation of significance test results are illustrated. The first strategy for augmenting the interpretation of significance tests…

Use of "special" orthonormal mean contrasts and mean contrast variances can help educational researchers interpret a wide variety of repeated measures data. Most statistical packages allow educational researchers to test for differences across repeated measures using both the univariate mixed model F test and a multivariate test.…

Power is an indicator of the ability of a statistical analysis to detect a phenomenon that does in fact exist. The issue of power is crucial for social science research because sample size, effects, and relationships studied tend to be small and the power of a study relates directly to the size of the effect of interest and the sample size.…

A fundamental fact of the analysis of variance statistical procedure is that if the omnibus F test of an effect is significant, then there exists at least one contrast of that effect that will be significantly different from zero according to the S-method of Scheffe. The caveat to this rule is that the significant contrast(s) may not be of any…

Two types of statistical analyses of rating scale data are discussed. An example, with its accompanying mathematical calculations for each type, is presented; and the advantages and disadvantages of each method are compared. It is judged that the little-known and seldom-used Kolmogorov-Smirnov one-sample test should be reexamined because of its…

The implications of data from a review of ten years of the American Educational Research Journal (AERJ) indicating that random sampling is rare and that there is increased use of quasi-experimental designs lacking in random assignment are considered. It is suggested that tests of significance could be abandoned or at least placed in a subsidiary…

The problem of determing how many significant discriminant functions are present in a given data set for a one-way, fixed-effects multivariate analysis of variance design is studied using a Monte Carlo procedure. A variety of procedures, including the popular partitioned-U procedure, are compared with respect to their Type I error rates and power…

A correction of the standard F-ratio for unreliability of the dependent measure has recently been proposed by Winne; the rationale is analogous to that of correcting a correlation for attenuation. However, there are two problems associated with Winne's correction of which potential users should be aware. First, the corrected statistic, F*, has…

The known interval scale, referred to as the 7.8 scale, has been criticized as an invalid measuring instrument in the form of an attitude scale. It is the purpose of this paper to demonstrate that this scale can produce spuriously inflated correlation coefficients, high reliability, and false significance on statistical tests. The case will be…

Part and bi-partial canonical correlations were developed by extending the definitions of part and bi-partial correlation to sets of variates. These coefficients may be used to help researchers explore relationships which exist among several sets of normally distributed variates. (Author)