Table 9.
Numbers of participants required for various designs when d = .4 and power is increased to 90%. The latter decreases the chances of not finding an effect present in the population.
| d = .4, power = .9, p < .05 | d = .4, power = .9, BF > 10 | |
|---|---|---|
| 1 variable between-groups | ||
| • 2 levels | 264 | 480 |
| • 2 levels, null hypothesis | 1084 | 3600 |
| • 3 levels (I = II > III) | 570 | 840 |
| • 3 levels (I > II > III) | 2160 | 3450 |
| 1 variable within-groups | ||
| • 2 levels | 70 | 130 |
| • 2 levels, null hypothesis | 271 | 1800 |
| • 3 levels (I = II > III) | 100 | 150 |
| • 3 levels (I > II > III) | 360 | 610 |
| Correlation | 260 | 460 |
| 2 × 2 repeated measures | ||
| • Main effect one variable | 35 | 65 |
| • Interaction (d v. 0) | 145 | 270 |
| 2 × 2 split-plot | ||
| • Main effect between | ||
| ◦ r = .5 | 200 | 360 |
| ◦ r = .9 | 250 | 450 |
| • Main effect repeated-measure | 70 | 130 |
| • Interaction (d v. 0) | 300 | 540 |
| • ANCOVA | ||
| ◦ rrep_measure = .5 | 210 | 360 |
| ◦ rrep_measure = .9 | 260 | 460 |