A researcher wants to determine the success rate of a driver's education program conducted in high schools in a particular
state. Which method would assure random selection of a sample from the population?
The researcher should select one high school in the state and survey all of the students in that school who are
enrolled in the training program.
The researcher should randomly select one high school in the state and survey a random batch of students who are
enrolled in the training program.
The researcher should randomly select a city in the state and survey all students in that city who are enrolled in the
training program.
The researcher should randomly select students from among all the students in the state who are enrolled in the
training program.

Choice D)

The researcher should randomly select students from among all the students in the state who are enrolled in the training program.

Further explanation:

I'm going to go through all four answer choices to say whether they are the answer or not the answer.

Choice A is not a good idea because focusing on one high school only is only representative of just that one high school. The other high schools do not get represented. Things get worse if this selected high school is very unique and different from any other average high school (eg: the income is higher than average). We eliminate choice A due to this bias.Choice B is a similar story to choice A. We need to pick more than one high school. Choice B is worse than choice A because we are selecting a smaller sample, which may skew things even more. We eliminate choice B as a potential answer.Choice C is on the right track since we are selecting more high schools to work with, but there are other cities in the state. We need a more inclusive picture. Choice C is crossed off the list.Choice D is the only thing left. It is the answer because of that and also because we are finally focusing on every student in the state who are enrolled in the driving program (as the population was defined earlier at the top of the problem). Of course we don't have the time/resources to ask every single student, which is why a random sample is needed.

So that's why choice D is the final answer.