Determination of Power
The researcher wants to know what power to expect in a two-tailed t-test to determine the difference in assessment scores between two equal group, given beta = .05, effect size at d=.5 (medium), when only 35 participants per group are available. Based on these preconditions, power would be .54, which means there would be only a little over a 50/50 chance of statistical verification of the effect-not a design that should go forward without revision. An increase in sample size to 80 participants per group would increase power to a more reasonable level of .88. Note that the effect size of .5 is not small;an anticipated smaller effect size (e.g. d=.2) would result in considerably lower power (power at .13, only a 13% chance of finding significance).