Which of the following is not a condition for constructing a confidence interval to estimate the difference between two population proportions? A- The observed number of successes and failures for both samples must be at least 10.

B- When samples are taken without replacement, each population must be at least 10 times as large as its corresponding sample.

C- The samples must be independent of each other.

D- The data must come from populations with approximately normal distributions.

E- The samples must be selected randomly.