Practical Suggestions on Pooling Variances for F Tests of Treatment Effects1

Abstract
A two‐part rule is given as a practical guide in preliminary tests of significance for determining whether or not to pool two or more variances for use as “error term” in subsequent F tests for treatment effects. Four examples involving commonly occuring analysis of variance tables are given to illustrate the use of the rule.

This publication has 0 references indexed in Scilit: