|1.||Concept of Validity. This section relates the discussion of validity begun in Section 1.2 to two major positions, one from test theory, one from quasi-experimental design. Neither position suffices for experimental analysis.|
|2.||Multivariate Analysis of Variance. Multivariate analysis of variance is a needed tool to handle multiple different measures on each subject. For common repeated measures designs, however, ɛ-adjusted Anova is generally preferable.|
|3.||Partial Analysis. Valuable simplification of multi-factor design can be obtained by using extrastatistical knowledge systems to strike out most of the high-way statistical interaction residuals.|
|4.||Pooling. Pooling interaction terms with error is sometimes appropriate when done a priori, but only in exceptional circumstances when done post hoc.|
|5.||Regression Artifact. The ubiquitous, insidious regression artifact needs to be understood by everyone; see especially the Exercises.|
|6.||Multiplication Model for Repeated Measures. With repeated measures variables, transformation of the data has a twofold potential benefit: Transformation can increase the size of the effect and also decrease the error term.|
|7.||General Linear Model. All most of us need to know about the general linear model in 1 ½ easy pages.|
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Book title: Empirical Direction in Design and Analysis. Contributors: Norman H. Anderson - Author. Publisher: Lawrence Erlbaum Associates. Place of publication: Mahwah, NJ. Publication year: 2001. Page number: 550.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.