Article Critique

  In research, calculatin" rel="nofollow">ing power is essential. The larger a sample size, the more powerful a study may be. It can be helpful to calculate the min" rel="nofollow">inimum sample size needed for basic statistics—such as the t-test—without the aid of software. Sample size calculation is based on the statistical testin" rel="nofollow">ing method used. Remember, sample size denotes the number of variables examin" rel="nofollow">ined in" rel="nofollow">in data collection. You use the software G*Power located in" rel="nofollow">in the Learnin" rel="nofollow">ing Resources to do your calculations. Calculatin" rel="nofollow">ing power based on sample size, rather than calculatin" rel="nofollow">ing sample size based on power, is useful when usin" rel="nofollow">ing secondary data where the sample size is already fixed. Calculate the min" rel="nofollow">inimum sample size you will need (usin" rel="nofollow">ing manual calculations) to conduct a t test to determin" rel="nofollow">ine if the difference in" rel="nofollow">in age in" rel="nofollow">in people with and without hypertension is significant. Use the standard deviation for age calculated in" rel="nofollow">in week 1. Assume a 2-tailed test with alpha = .05 and Power = 80% and that the difference you expect to observe between the 2 groups is 5 years. (30 poin" rel="nofollow">ints) Usin" rel="nofollow">ing two different effect sizes in" rel="nofollow">in addition to the one used in" rel="nofollow">in part 1, perform three power analyses of the sample size computed in" rel="nofollow">in step 1 usin" rel="nofollow">ing G*Power. (Assume a 2-tailed in" rel="nofollow">independent sample t test with alpha set at .05) (30 poin" rel="nofollow">ints) Perform a power analysis usin" rel="nofollow">ing G*Power usin" rel="nofollow">ing the actual sample size presented in" rel="nofollow">in the dataset for week 1 (180) and an effect size of .30. What does this mean in" rel="nofollow">in terms of the study and the probability of experiencin" rel="nofollow">ing a type 2 error? (40 poin" rel="nofollow">ints)