Type I error rates and statistical power for the univariate F test and the James second-order test were estimated for the two-factor fixed-effects completely randomized design. Results reveal that the F test Type I error rate can exceed the nominal significance level when cell variances differ. (SLD)

Reference-reliability relates variability due to change and error. It indicates whether some suspected change can be reliably differentiated from random fluctuations. A means by which the process of change can be measured at different points in time is outlined, using empirical data. (TJH)

The Shine-Bower single subject ANOVA is extended to a multivariate case, with one example assuming between-variate dependencies among within-subject errors and the second assuming no between-variate dependencies among within-subject errors. Standard and simplified multivariate ANOVA procedures are used, respectively. (Author/CM)

Some results on how the Alexander-Govern heteroscedastic analysis of variance (ANOVA) procedure (R. Alexander and D. Govern, 1994) performs under nonnormality are presented. This method can provide poor control of Type I errors in some cases, and in some situations power decreases as differences among the means get large. (SLD)

The use of multiple comparisons in analysis of variance (ANOVA) is discussed. It is argued that experimentwise Type I error rate inflation can be serious and that its influences are often unnoticed in ANOVA applications. Both classical balanced omnibus and orthogonal planned contrast tests inflate experimentwise error to an identifiable maximum.…

Various methods of estimating main effects from ordinal data are presented and contrasted. Problems discussed include: (1) at what level to accumulate ordinal data into linear measures; (2) how to maintain scaling across analyses; and (3) the inevitable confounding of within cell variance with measurement error. An example shows three methods of…

Dependability of class means is analyzed by applying generalizability to a split-plot design with students nested within classes. Basic generalizability concepts are reviewed, and the derivation and interpretation of distinct generalizability concepts are discussed. Four generalizability coefficients are compared with each other and with the three…

Type I error rates and power were estimated for 10 tests of variance equality under various combinations of the following factors: similar and dissimilar distributional forms, equal and unequal means, and equal and unequal sample sizes. (TJH)

This book presents a comprehensive overview of univariate and multivariate generalizability theory, a psychometric model that provides a powerful approach to the analysis of errors of measurement through the use of random-effects and mixed-model analysis of variance. (SLD)

The N=1 analysis differs from a typical analysis of variance in that there is no within-cell error term. Thus interaction terms are used as estimates of error variance. If the interaction term in question represents a significant interaction, the F tests will be conservative. Tukey's test for nonadditivity will detect a common form of interaction.…

This note shows that, under conditions specified by Levin and Subkoviak (TM 503 420), it is not necessary to specify the reliabilities of observed scores when comparing completely randomized designs with randomized block designs. Certain errors in their illustrative example are also discussed. (Author/CTM)

Comments (TM 503 706) on an earlier article (TM 503 420) concerning the comparison of the completely randomized design and the randomized block design are acknowledged and appreciated. In addition, potentially misleading notions arising from these comments are addressed and clarified. (See also TM 503 708). (Author/CTM)

This note continues the discussion of earlier articles (TM 503 420, TM 503 706, and TM 503 707), comparing the completely randomized design with the randomized block design. (CTM)

Incidence sampling is a parsimonious method whereby a large number of examinees can be measured on many variables (such as test items) to assess group characteristics. Parameters used to describe an incidence sample are estimated using the theory of generalized symmetric means and generalizability theory. (Author/JKS)

The effects of under and overrotation on common factor loading stability under three levels of common variance and three levels or error are examined. Four representative factor matrices were selected. Results suggested that matrices which account for large amounts of common variance tend to have stable factor loadings. (Author/RL)