Of particular interest to generalizability theory strategies is a relationship showing that integrated design variance components not involving attributes are identical with the average respective covariance components across attributes. An illustration is provided to demonstrate the relationships, strengths, and weaknesses of two strategies.…

The authors suggest that there may be a phenomenon of posttest sensitization which could affect the generalizability of experimental results in the way that pretest sensitization does. A quasi-experimental design is proposed to test for such a phenomenon. (JKS)

This article argues that the 1981 work of Carbno presented unwarranted conclusions because its design included an improper operationalization of the object of measurement, given the problems addressed, and because the sample sizes employed were too small. (Author/PN)

Using grade point average and graduation-nongraduation as two separate criteria and eight predictor variables selected from doctoral applicants' folders at an Arizona university, a canonical correlation of .624 for a sample of 65 graduate students shrank to -.095 for a cross validation sample of 21 students. (CTM)