a sample size of 15 subjects per group provided 80% power to detect differences of 5 degrees of ROM between the preintervention and postintervention measurements as well as between the 2 groups of interest at an alpha level of .05 with a 2-tailed test.