Charig et al. (1986) compared the success rates of two treatments (A vs B) for k
ID: 3336163 • Letter: C
Question
Charig et al. (1986) compared the success rates of two treatments (A vs B) for kidney stones. Treatment A includes all open procedures and Treatment B is percutaneous nephrolithotomy. Here are the data they analyzed Small stones A B Success 81 234 Failure 6 36 Large stones A B Success 192 55 Failure 71 25 The researchers wanted to determine which treatment was better. The odds ratio for the small stones table was 2.077 and for the large stones was 1.229 Suppose a conventional statistical hypothesis test (H: All odds ratios the same versus H: At least one odds ratio different than the others) was performed to check for homogeneous odds ratios across strata and the empirical significance level (a.k.a., P value) was 0.38. Suppose further that the researchers interpreted this to mean that the odds ratios across strata (stone sizes) were homogenous (i.e., the same). Why is this interpretation faulty? Reference Charig, Clive R., David R. Webb, Stephen Richard Payne, and John E. Wickham. "Comparison of treatment of renal calculi by open surgery, percutaneous nephrolithotomy, and extracorporeal shockwave lithotripsy." Br Med J (Clin Res Ed) 292, no. 6524 (1986): 879-882 O a. There is no way to (statistically) prove a null hypothesis O b. The Mantel-Haenszel approach does not require homogenous odds ratios ° C. The empirical significance level would need to be higher than 0.5 0 d. You cannot (statistically) prove a null hypothesis using conventional hypothesis testing O e. The researchers should have used the Woolf test.Explanation / Answer
Solution:
(d) is the correct answer.
Reason:
You cannot prove a null hypothesis using conventional hypothesis testing.
Cheers!!