This packet will introduce and define the following topics:
There are also examples of each located after the powerpoint, and scenarios which will ask you which sampling method is most appropriate.
Source: Greene
Here are some scenarios for you to see how well you can identify which sampling design are appropriate. Solutions are posted below.
We are looking at how students get to school in high school. We choose one school to take our sample from. There are four grades in high school (freshman, sophomore, junior, and senior). We would like roughly 100 students for our sample. Which sampling method(s) would be appropriate, and how would you execute it(them)?
A land surveyor is interested in species of animals at a state park. The park is naturally split up into five areas, all of which are very similar in size and landscape. Which sampling method(s) would be appropriate, and how would you execute it(them)?
A CEO of a large company would like to sample some of his employees and see how much they like working for him. He has a list of all employees, and they are numbered 1-1000. He would like 25 employees for his sample. Which sampling method(s) would be appropriate, and how would you execute it(them)?
Source: Greene
We are looking at how students get to school in high school. We choose one school to take our sample from. There are four grades in high school (freshman, sophomore, junior, and senior). We would like roughly 100 students for our sample. Which sampling method(s) would be appropriate, and how would you execute it(them)?
Since there will be a big difference between how freshman and seniors get to school (seniors will have cars, freshman will most likely take the bus), it would be appropriate to get some students from each grade. In this case, a stratified sample would be appropriate. The four grades would be our stratas, and since we need 100 students, we could randomly sample 25 students from each grade to comprise our 100 students.
A land surveyor is interested in species of animals at a state park. The park is naturally split up into five areas, all of which are very similar in size and landscape. Which sampling method(s) would be appropriate, and how would you execute it(them)?
The key here is that all five areas are very similar. In reality, if we can only go to one or two areas instead of all areas in our sampling frame, we will save a lot of time and money. This would call for a cluster sample. The five areas are our clusters, and since they are similar, going to one will be representative of going to all five. We will randomly select one cluster and sample all species in that cluster.
A CEO of a large company would like to sample some of his employees and see how much they like working for him. He has a list of all employees, and they are numbered 1-1000. He would like 25 employees for his sample. Which sampling method(s) would be appropriate, and how would you execute it(them)?
Since we already have a list of people, and they are in numerical order, either a simple random sample or systematic random sample would be appropriate.
For a simple random sample, we could simply generate 25 random numbers between 1 and 1000, and use those employees in our sample.
For a systematic sample, we would take our population size divided by our sample size. Our population is the 1000 employees, and the desired sample size is 25. So 1000/25 = 40. This is our m. Next, we will randomly select a number between 1 and 25. Suppose we get 12. We would then sample the 12th employee on the list, then the 12+40 = 52nd employee, then the 12+2(40) = 92nd employee, and so on until we have all 25 employees. All the numbers for the sample chosen in this way are listed below:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
12 52 92 132 172 212 252 292 332 372 412 452 492 532 572 612 652 692 732 772 812 852 892 932 972.
Source: Greene