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)

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…

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)

Besides the ubiquitous Pearson product-moment r, there are a number of other measures of relationship that are attenuated by errors of measurement and for which the relationship between true measures can be estimated. Among these are the correlation ratio (eta squared), Kelley's unbiased correlation ratio (epsilon squared), Hays' omega squared,…

Statistics are an essential tool for making proper judgement decisions. It is concerned with probability distribution models, testing of hypotheses, significance tests and other means of determining the correctness of deductions and the most likely outcome of decisions. Measures of central tendency include the mean, median and mode. A second…

THE PURPOSE OF THIS RESEARCH WAS TO STUDY THE EFFECT OF ERROR OF MEASUREMENT UPON THE POWER OF STATISTICAL TESTS. ATTENTION WAS FOCUSED ON THE F-TEST OF THE SINGLE FACTOR ANALYSIS OF VARIANCE. FORMULAS WERE DERIVED TO SHOW THE RELATIONSHIP BETWEEN THE NONCENTRALITY PARAMETERS FOR ANALYSES USING TRUE SCORES AND THOSE USING OBSERVED SCORES. THE…

Livingston's work is a careful analysis of what occurs when one pools two populations with different means, but similar variances and reliability coefficients. However, his work fails to advance reliability theory for the special case of criterion-referenced testing. See ED 042 802 for Livingston's paper. (MS)

This paper, a revision of the original document, "Correcting Partial, Multiple, and Canonical Correlations for Attenuation" (see TM 000 535), presents the formula for correcting coefficients of partial correlation for attenuation due to errors of measurement. In addition, the correction for attenuation formulas for multiple and cannonical…

The correction for attenuation formulas for partial, multiple, and canonical correlation coefficients are discussed and the effects of measurement errors on these statistics are explored. The notation is standardized and the derivation extended where appropriate. It is shown that as the reliabilities of the predictors become more disparate, the…

The generalizability theory approach to the estimation of reliability is outlined, and a design is developed in which systematic variations in behavior over differing situations are separated from random fluctuation. Three coefficients of reliability are proposed. (CK)

The various costs of Type I and Type II errors of inference from data are discussed. Six methods for minimizing each error type are presented, which may be employed even after data collection for Type I and which minimizes Type II errors by a study design and analytical means combination. (Author/CM)

Rules, procedures, and algorithms intended to aid researchers and practitioners in the application of generalizability theory to a broad range of measurement problems are presented. Two examples of measurement research are G studies, which examine the dependability of some general measurement procedure; and D studies, which provide the data for…

The general linear model was described, and the influence that measurement errors have on model parameters was discussed. In particular, the assumptions of classical true-score theory were used to develop algebraic relationships between the squared multiple correlations coefficient and the regression coefficients in the infallible and fallible…