ESTIMATING PROFILE RELIABILITY AND MAXIMALLY RELIABLE COMPOSITES
- 1 January 1974
- journal article
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 9 (1) , 85-104
- https://doi.org/10.1207/s15327906mbr0901_7
Abstract
Two indices of profile reliability (Conger and Lipshitz's canonical reliability, p., and Bock's P) are shown to be equivalent in terms of the individual independent canonical composites; however, because of different weighting procedures, they yield different overall indices of profile reliability. A common formula is provided from which both indices can be derived. Based on Bock's multivariate analysis of variance approach to reliability, sample estimates of profile reliability are derived for Conger and Lipshitz's pp. Because of the frequent lack of parallel forms for profiles, an alternative to the analysis of variance approach is discussed, along with its own inherent problems. In addition, it is shown how maximally reliable independent profile composites can be constructed using either the Bock or Conger and Lipshitz weightings. An example using the WISC for 7 1/2-year olds is provided to illustrate the various techniques.Keywords
This publication has 4 references indexed in Scilit:
- Canonical Reliability for Profiles and Test BatteriesPublished by American Psychological Association (APA) ,1971
- Statistical principles in experimental design.Published by American Psychological Association (APA) ,1962
- Assessing similarity between profiles.Psychological Bulletin, 1953
- Test Reliability Estimated by Analysis of VariancePsychometrika, 1941