COHENS d EFFECT SIZE r 80 POWER 90 POWER 10 063 1571 2102 20

COHEN\'S d                EFFECT SIZE r                 .80 POWER                 .90 POWER

.10                               .063                                    1,571                             2,102

.20 ( small )                 .100 ( small )                      393                               526

.30                                .148                                    175                               234

.40                               .196                                      99                                132

.50 ( medium)             .243 ( medium )                   64                                 85

.60                               .287                                     45                                 59

.70                               .330                                    33                                 44

.80 ( large )                 .371 ( large )                       26                                34

.90                             .447                                      17                                  22

1.00                             .514                                     12                                 16

You are designing a study to be evaluated with a two-sample nondirectional independent t test at = .05. You want .90 power. You expect Cohens\'s d to be about .40. How many subjects will you need per group?

Solution

Look at the row corresponding to Cohens\' d = 0.40 and look at the value corresponding to 0.90 power. These values correspond to 0.05 nondirectional alpha. The sample size is given the value at the intersection of row Cohens\' d = 0.40 and power = 0.90. Therefore, the sample size is 132.