Imagine a very small sample in which we observe k=3 cases with disease out of n=20 subjects. Should we estimate p with a confidence interval based on the normal approximation of the sampling distribution of proportions?
a. True
b. False