Lab: Exploring Sampling

In this lab we will be studying sampling using the software SAMP. Turn on the computer, go into windows and click on the SAMP icon. Read the introduction and then go to the procedures menu.

What is the name of the town where the research is being done?
How many people live there?
What is the study about?

We will use samples to estimate the population mean.

Research questions:
What sampling methods are most accurate? What sampling methods are most cost effective? What, if any, trade off is there between cost and accuracy?

For this lab we will assume a sample of 1% of the population or about 110 people. We will take one sample of each of several different types. Go to the procedures menu and select the sample type. Then, read about that sample type and select 1 sample from each (follow the list below). When you are done with each sample, click on the SAVE button. When you have completed one SET go to SUMMARY and print your results.

When you read about each type of sample, make sure to write down what the sampling frame for each method and what steps are involved in each method.

Set 1.

• A simple random sample;
• A stratified sample with 1% selected from each group;
• A cluster sample with 5 areas selected;
• A quota sample with 55 males, 55 females, and 37, 37, and 36 from each age group

Set 2

• A simple random sample
• A cluster sample with 2 areas selected.
• A stratified sample with 1% selected from each group (yes, the same as above)
• A quota sample with 55 males, 55 females, and 37, 37, and 36, respectively, from each age group

Set 3

• A simple random sample
• A cluster sample with 10 areas selected.
• A stratified sample with 1% selected from each group (yes, the same as above)
• A quota sample with 55 males, 55 females, and 37, 37, and 36 from each age group

Look at the print out. What is the true value of the population mean?

How to think about costs:

Samp will tell you the cost per interview, the cost per attempted interview, the cost of drawing the sample, and the about of driving that would have to be done along with the cost per mile. Use these to estimate the costs. We will do this in Excel.

• Click here to open a spreadsheet with labels

We will work together, but you should start on your own if you are ready.

Start by entering the sample means for each of your 12 samples.

Then enter the true population mean in each column (copy this from your output; you can copy and paste to put it in each column).

Then, in cell b4 calculate the error by subtracting the sample mean from the population mean. You do this by typing in the fomula =b3-b2. Copy and paste this formula to all of the columns in row 4.

• Which kind of sample gave the best estimates (smallest errors)?
• Which gave the worst (biggest errors)?
• What patterns do you notice?

For the remainder of the lab work on completing the spread sheet.

Start entering the information that you have on your printouts for each of the samples. Items in red are available on your printout. (Look on the web page to see colors) Items in blue were discussed in the introduction to each sampling method and use information given in the printout. Fill the red and blue items in first. Note that you can use copy and paste for information that is the same across the different sample types.

Items in black need to be calculated using Excel and the information in the red and blue cells. Try to determine the formulas for these on your own. Your instructor will assist you with translating the formulas into Excel. The best strategy is to do all of the formulas for one column; you will be able to copy and paste many of them to the other columns.

To print your table: Use your mouse to highlight the area you typed in. Click on File, Print and select "selected area," print.

• Looking at the various cost factors involved in each method, which is the least expensive?
• Which method is the most expensive?
• How would you balance cost and accuracy in choosing a sampling design?