Calculating the Statistical Power of the Univariate and the Multivariate Repeated Measures Analyses of Variance for the Single Group Case Under Various Conditions

Abstract

Repeated measures designs find frequent application in behavioral research. Researchers using such designs must choose between the univariate approach and the multivariate approach for a test statistic. The statistical power of these two approaches has been shown to be varyingly different depending upon the structure of the variance-covariance matrix relative to the pattern of mean differences. However, the guidelines for choosing the most powerful test for a given situation have not been well defined. A FORTRAN program is described which calculates the statistical power of the omnibus tests for the univariate and the multivariate approaches to repeated measures in the single group case. Program input, output, applications, and availability are discussed.

Get full access to this article

View all access options for this article.

References

Barcikowski, R. S. (1973). A computer program for calculating power when using the T2 statistic with selected designs. Educational and Psychological Measurement, 33, 723-726.

Barcikowski, R. S. and Holthouse, N. (1972). A computer program for calculating the power of the F tests in analysis of variance and covariance for specified alpha levels, sample sizes, and effect sizes. Educational and Psychological Measurement, 32, 169-172.

Barcikowski, R. S. and Robey, R. R. (1983). Univariate and multivariate repeated measures analysis through PROC's REG, GLM, and MATRIX. In SAS Institute (ed.), Statistical Analysis System Users Group International Proceedings. Cary, NC: SAS Institute.

Box, G. E. P. (1954). Some theorems on quadratic forms applied in the study of analysis of variance problems: II. Effect of inequality of variance and of correlation between errors in the two-way classification. Annals of Mathematical Statistics, 25, 484-494.

Carroll, R. M. , and Nordholm, L. A. (1975). Sampling characteristics of Kelley's ε2 and Hay's ω2. Educational and Psychological Measurement, 35, 541-554.

Cohen, J. (1977). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press.

Davidson, M. L. (1972). Univariate versus multivariate tests in repeated-measures experiments. Psychological Bulletin, 77, 446-452.

Edgington, E. S. (1974). A new tabulation of statistical procedures in APA journals. American Psychologist, 29, 25-26.

Greenhouse, S. W. and Geisser, S. (1959). On methods in analysis of profile data. Psychometrika, 24, 95-112.

10.

Huynh, H. and Feldt, L. S. (1970). Conditions under which mean square ratios in repeated-measurement designs have exact F-distributions. Journal of the American Statistical Association, 65, 1582-1589.

11.

Huynh, H. and Feldt, L. S. (1976). Estimation of the Box correction for degrees of freedom from sample data in randomized block and split-plot designs. Journal of Educational Statistics, 1, 69-82.

12.

Imhof, J. P. (1962). Testing the hypothesis of no fixed main-effects in Scheffé's mixed model. The Annals of Mathematical Statistics, 33, 1085-1095.

13.

The IMSI Library, Edition 9. (1982). International Mathematical and Statistical Libraries, Houston.

14.

Pearson, E. S. and Hartley, H. O. (1951). Charts of the power function for the analysis of variance tests derived from the non-central F distribution. Biometrika, 38, 112-130.

15.

Rich, C. E. (1983). Repeated measures designs. In R. S. Barcikowski , (Ed.), Computer packages and research design with annotations of input and output for BMDP, SAS, SPSS and SPSSX, Vols. 1-3. Washington, D. C.: University Press of America.

16.

Rouanet, H. and Lépine, D. (1970). Comparison between treatments in a repeated-measurement design: ANOVA and multivariate methods. British Journal of Mathematical and Statistical Psychology, 23, 147-163.

17.

Winer, B. J. (1971). Statistical principles in experimental design (2nd ed.). New York: McGraw-Hill.