When a series of statistical tests are employed in the same experiment, the researcher is obligated to consider the error rate (probability of making a type I error) in terms of the experiment, not merely in terms of each comparison. A table is provided for the rapid determination of experimentwise alpha level when a number of independent statistical tests are employed in the same experiment. Suggested applications and the rationale for this procedure are supplied.
Get full access to this article
View all access options for this article.
References
1.
Dayton, C. M. and Schafer, W. D.Extended tables of t and chi square for Bonferroni tests with unequal error allocation. Journal of the American Statistical Association, 1973, 68, 78-83.
2.
Kirk, R. E.Experimental design: Procedures for the behavioral sciences. Belmont, California: Brooks/Cole, 1968.
3.
Ryan, T. A.Multiple comparisons in psychological research. Psychological Bulletin, 1959, 56, 26-47.
4.
Ryan, T. A.The experiment as the unit for computing rates of error. Psychological Bulletin, 1962, 59, 301-305.
5.
Sakoda, J. M., Cohen, B. H., and Beall, G.Tests of significance for a series of statistical tests. Psychological Bulletin, 1954, 51, 172-175.
6.
Udolf, R. and Raymond, B.Multivariate analysis of variance and confidence levels . American Psychologist, 1973, 28, 1139-1140.
