A case-control study is being planned to see whether proximity to a nuclear power plant is associated with the development of certain cancers. 200 cases and 200 controls will be recruited and the distance from the nearest nuclear power plant will be dichotomized as close (< 3 miles) or far (≥ 3 miles). Use SAS to complete each problem below. (d) Since investigators find it very difficult to recruit cases, they decide to recruit three times as many controls as cases. How many cases and controls will they need to recruit in order to have 80% power? (Still assume a 2-sided, 5% level test, the same exposure probability among controls, and same OR.) If everything stays the same as above but the level of the test is reduced below 5%, what would happen to the total sample size computed first? (e) It is believed that about 4% of controls live close to the plant and the OR that investigators would be interested in detecting is 3.5. Assuming that a 2-sided, 5% level test will be conducted, what power will the test have for detecting a significant association between cancer and proximity to the power plant? If everything stays the same as above but the sample sizes are halved in each group, what would happen to the power computed first? (f) If everything stays the same as in part (e) but - without the sample sizes are halved in each group - the OR that investigators wish to detect is larger than 3.5, what would happen to the power in part (e)?

MATLAB: An Introduction with Applications

6th Edition

ISBN:9781119256830

Author:Amos Gilat

Publisher:Amos Gilat

Chapter1: Starting With Matlab

Section: Chapter Questions

Problem 1P

See similar textbooks

Related questions

Question

A case-control study is being planned to see whether proximity to a nuclear power plant is associated with the development of certain cancers. 200 cases and 200 controls will be recruited and the distance from the nearest nuclear power plant will be dichotomized as close (< 3 miles) or far (≥ 3 miles). Use SAS to complete each problem below. (d) Since investigators find it very difficult to recruit cases, they decide to recruit three times as many controls as cases. How many cases and controls will they need to recruit in order to have 80% power? (Still assume a 2-sided, 5% level test, the same exposure probability among controls, and same OR.) If everything stays the same as above but the level of the test is reduced below 5%, what would happen to the total sample size computed first? (e) It is believed that about 4% of controls live close to the plant and the OR that investigators would be interested in detecting is 3.5. Assuming that a 2-sided, 5% level test will be conducted, what power will the test have for detecting a significant association between cancer and proximity to the power plant? If everything stays the same as above but the sample sizes are halved in each group, what would happen to the power computed first? (f) If everything stays the same as in part (e) but - without the sample sizes are halved in each group - the OR that investigators wish to detect is larger than 3.5, what would happen to the power in part (e)?

Expert Solution