Question

You are a researcher working for an educational research firm. Your firm was asked to conduct a satisfaction survey among the parents of students attending the School District public schools. Your boss has asked you to develop a sampling design for this survey.

For this survey, you are asked to select a representative sample of parents. When you answer the questions, keep the following in mind:

• Assume that you have unlimited authority to access all the information you may need and remember that your task is to select the most representative sample of the parent population.
• There are 10,000 households with children at public schools in the school district.
• One parent will be interviewed for each household in the sample. (Remember that multiple children from the same household may attend public schools, and there may be more than one parent in each household. Regardless of these variations, only one parent will be interviewed from each household selected for the sample.)

1. Describe your population. Elaborate on the specifics of your population.

2. Describe your sampling frame. Be specific in your descriptions. Where would you find the information to define it? Remember that you have full authority to access all the information available, (You just need to know whom to ask and what to ask for.)

3. If you use a simple random sampling design, what is the minimum sample size you will need to determine the percentage (proportion) of those parents who are satisfied with the public schools at the 4% accuracy level, with the 95% confidence level? Explain how you determined your sample size.

4. If you use a multistage cluster sampling design, would your minimum sample size be larger or smaller than the previous simple random sampling design? Why?

5. If you use a proportionate stratified sampling design, would your minimum sample size be larger or smaller than the simple random sampling design. Why?

Answer 1

Part A and B:

There are 10,000 household with children studying at public schools. As only one parent can be selected from every house hold, we will select the random sampling technique as the population size is very huge. It will be very difficult to visit every household of the locality. Here, depending on the required sample size, households are selected in random. In case of this sampling technique, every household has the equal probability of being selected. So the households are selected unless and until the required size is obtained.

Part C:

Part D and E:

Yes, the sample size will be small in both the cases. In the case of Clustered Sampling - Multi Staged, the design effect comes into action. A design effect is a type of an adjustment made in order to find a survey sample size.

Here, the calculation will be like:

End of the Solution...