Rules are presented for choosing error terms in complete factorial designs with equal replications. Since the paradigm from which they were derived can be generalized to designs with any number of factors, the same would appear to be true of these rules, if one accepts the Anderson-Bancroft position and follows the “never pool” procedure.
Get full access to this article
View all access options for this article.
References
1.
AndersonR. L., & BancroftT. A.Statistical theory in research.New York: McGraw-Hill, 1952.
2.
BinderA.The choice of an error term in analysis of variance designs. Psychometrika, 1955, 20, 29–50.
3.
CornfieldJ., & TukeyJ. W.Average values of mean squares in factorials. Ann. math. Statist., 1956, 27, 907–949.
4.
GaitoJ.Expected mean squares in analysis of variance techniques. Psychol. Rep., 1960, 7, 3–10.
5.
MoodA. M.Introduction to the theory of statistics.New York: McGraw-Hill, 1950.
