The general use of the Spearman-Brown formula for calculating the reliability of parallel tests with different lengths is reviewed. The importance of the assumption that the component tests be parallel is noted and the property that parallel tests must be non-negatively correlated is derived. (Author)

Rules are given for finding average variances and covariances from mean squares, and the reverse. These serve the same purpose as recursion formulas provided by Stanley and Wiley (1962), but a knowledge of matrix algebra is not needed. A four-factor design is used to illustrate their application. (Author/RC)

A two-factor analysis of variance with multiple measurements on one factor was conducted among the 40 items of the Vocabulary test of the Test of English as a Foreign Language (TOEFL) for six language groups. All sources of variance were found significant beyond the one per cent level. (Author/RC)

A generalization of the G index of agreement is presented, allowing for its application in multiple ratings. The generalized coefficient is intended for use in the psychiatric clinic. An illustrative example is provided, using hypothetical data. (Author)

Underlying factor structure of 10 concepts rated by 67 students in an introductory advertising course was examined. Dimensions of relevance were found to conform to results reported by Menges and Trumpeter (1971) suggesting the potential for a university of basic dimensions of educational relevance across diverse fields of learning. (Author/RC)

A procedure for item analysis using distance clustering is described. Items are grouped according to the predominant factors measured, regardless of what they are. The procedure provides an efficient method of treating unanswered items. (Author/RC)

Sixty subjects, 30 volunteers and 30 nonvolunteers, participated in an experiment examining pretest sensitization effects. Two interpretations of these effects--the commitment hypothesis and the demand interpretation--are explored. (Author)