Question

 

Part 1. Suppose a test is performed to test H0: risk difference = 0 versus H1: risk difference ≠ 0, and the test rejects H0 at α = 0.05. Then, we can conclude that there is real significant evidence (α = 0.05) of a difference in proportions, _____________________ evidence that the risk difference is not 0, and _____________________ evidence that the relative risk and odds ratio are not _________________. Choose between significant, not significant, 0, 1.

Part 2. Contrast superiority vs. noninferiority vs. equivalence tests. Provide real world examples where you would use this using a new drug, a standard of care, a placebo, and a generic drug.

Part 3. For a one-factor ANOVA, we conduct one F-test. What is this testing?

Part 4. Contrast the per-comparison error rate with the familywise error rate.

Answer 1

Part-1:

Suppose a test is performed to test H0: risk difference = 0 versus H1: risk difference ≠ 0, and the test rejects H0 at α = 0.05. Then, we can conclude that there is real significant evidence (α = 0.05) of a difference in proportions, _______significant______________ evidence that the risk difference is not 0, and _______not significant______________ evidence that the relative risk and odds ratio are not ________1_________.

Part 2:

Superiority vs inferiority vs equivalence tests

Superiority tests are conducted to check whether the investigational product is better than the (standard treatment)

Inferiority tests are conducted to check whether the investigational product is worse than the (standard treatment)

Equivalence tests are conducted to check whether the investigational product is the same as the (standard treatment)

Example: The researcher wants to check whether the health of a group of people those consuming the new drug and those consuming the existing drug have any difference in the recovery pattern. Whether the new drug is better, worse, or the same as the existing drug.

Part 3:

We conduct the one-F test while performing Anova because it helps to determine the effect of variance by calculating their ratios. Without ratio it will be difficult to interpret the variances.

Part 4: Per-Comparison Error Rate and Family Wise Error Rate

The per-comparison error rate is the type 1 error rate of any significance test and family wise error rate is the probability that there might be one or more statistical test could have type 1 error rate