ABSTRACT
Between-subjects ANOVA In this case, use the arithmetic mean. Thus if there were three groups with 15, 20 and 30 in each, then the sample size per group should be treated as:
n
=
+ +15 20 30 3
= 21.67, or 21 to the next lowest person
Using the technique of linear interpolation, described in Appendix XIV for probability tables, approximate power values and sample sizes can be found where these are not contained in the tables given in the present appendix. Cohen (1988) gives a much wider range of tabled values.
