Answer:
B) Reduce selection bias in subject allocation to treatment
Explanation:
Typically, researching the entire population is economically and technically unfeasible. For this, there are statistical approaches that aim to analyze the population through a random sample that will represent the population in a statistically coherent way.
 The method is to extract a subset of individuals from a population so that all individuals are equally likely to be chosen. This will avoid selection bias, a problem that can be caused when the sample is chosen.
For example, imagine the US election polls. Some states are historically Republican and others Democratic. If a poll is made by choosing voters only in the state of California, the result is likely to be a huge advantage for Democrats, a clear selection bias. However, we know that this is not the reality. An electoral poll must be done randomly between states to form a faithful representation of reality.
