Use the following article to help you answer the following questions: McCrum-Gar

Question

Use the following article to help you answer the following questions: McCrum-Gardner, E. (2010). Sample size and power calculations made simple. International Journal of Therapy and Rehabilitation, 17, 10-14.

Note: m is the ratio of subjects in the experimental group and control group. Use a ratio of 1.

a) Suppose you were conducting a two independent groups t-test to investigate the effect of a new weight training program on physical strength. Previous studies have found a minimum clinically important difference of 30 and a pooled standard deviation of 20. Using an alpha of .05 and power of .80, how many subjects would you need in each of the experimental and control groups _______ ? Be careful how you enter your answer. For example, if you needed 100 subjects in the experimental group and 100 subjects in the control group, your answer would be 100, not 200.

b) Suppose you were conducting a two independent groups t-test to investigate the effect of a new intervention designed to improve math scores. A previous study found a difference of 5 and a pooled standard deviation of 10. Using an alpha of .05 and power of .80, how many subjects would you need in each of the experimental and control groups _______ ? Be careful how you enter your answer. For example, if you needed 100 subjects in the experimental group and 100 subjects in the control group, your answer would be 100, not 200.

c) Using the same example as the previous question (part b), what effect size did you use in your sample size calculations? Please round your answer to the nearest 2 decimal places.

d) Suppose you were conducting a two independent groups t-test to investigate the effect of a new weight training program on physical strength. Previous studies have found a minimum clinically important difference of 55 and a pooled standard deviation of 10. Using an alpha of .01 and power of .80, how many subjects would you need in each of the experimental and control groups _______ ? Be careful how you enter your answer. For example, if you needed 100 subjects in the experimental group and 100 subjects in the control group, your answer would be 100, not 200.

e) Suppose that you planned to conduct a paired t-test to investigate the efficacy of a weight loss program. Previous research has shown a mean loss of 40 pounds in the population with a standard deviation of 10. How many pairs of subjects will be needed in order to achieve a power of .90 if we use an alpha of .05 __________?

Explanation / Answer

a. The sample size is calculated as follows:

n=(zalpha/2+zbeta)^2*2*sigma^2/d^2, where, z is critical z value at alpha/2 (alpha=0.05), that is 1.96, zbeta is z critical at power of 0.80 (beta=0.20), that is 0.84, sigma is population standard deviation and d is estimated difference.

=(1.96+0.84)^2*2*20^2/30^2

=7

b. Use the same formula, by substituting d=5, sigma=10.

n=63

c. Effect size=d/spooled=5/10=0.5

d. Use the same formula for computing sample size, and change the value of d=55, sigma=10, alpha=0.01.

n=1

Related Questions

Navigate

• Browse (All)
• Browse U
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.