Samples and Populations

You will work with two other research partners.  With your research partners, select a research question to investigate that interests your group (i.e., Do you think blue or yellow is a more pleasant color? Are you in favor of school vouchers? Do you favor school uniforms? Note: These are being provided as examples and are not suggestions that you should feel obligated to use.  For this assignment, limit your research question to one which people can provide only one of two possible responses (i.e., yes/no, blue/yellow). Once you have settled on a research question, each of you will separately define a target population, select a sample from your target population using one of three sampling techniques, collect data, and report your results. Each member of your research team will use a different sampling technique...one person will do a simple random sample, one person will do a stratified sample, and one person will do a systematic sample.

In this project, you  will each apply one of three of sampling techniques: simple random, stratified, or systematic sampling. Under normal research conditions, we would select the best possible sampling procedure that fits our research question and situation. We would seldom use three separate sampling procedures. However, this is a practice exercise and I would like each of you to have experience using one of the three sampling procedures listed above.

Steps to follow:

1. Select a research question with your partners.

2. Decide among your group members who will do each of the three sample techniques described above. One person uses simple random, one person uses stratified, and one person uses systematic.
   
   
3. On your own, define your target population and count (or estimate) the size of your target population. Select a very different population from your partners. You may select teachers while your partners select students and parents. Note: Select a target population that you can reach. It is unrealistic to define your target population as UConn students...you could define it as students enrolled in a particular course fall semester.

4. Use the Determine Sample Size calculator to determine the number of subjects you need to sample in order for you to be 95% confident (confidence level) that your results are within a 10% margin of error (confidence interval). A confidence interval of 10% is lower than normal. I have set this lower limit to reduce the size of the sample you must collect. I have created instructions for  using the Determine Sample Size calculator.
   
   

5.  From your target population, select subjects by the sampling procedure you selected.
       -- Use the random number table on page 646 (4th ed.) or page 626 A-2 (5th ed.) of your text to select a simple random sample (if that were the type of sampling you selected).
       -- Select a stratified sample of your target population by gender (if you selected the stratified option). Note: While you stratify by gender when selecting your participants, combine their data when reporting the results. If the sample size calculator indicated that you needed 30 subjects and your population consisted of 60 males and 40 females, you would randomly select 18 males and 12 females, ask them your question, and then total their responses. I am not interested in the separate responses of the males and females. I am only interest in the total. The purpose here is to practice selecting a stratified sample.
       -- If you elected to use systematic sampling, you may need to round your selection slightly to obtain the correct number of subjects (e.g., if you were trying to select 5 people from a group of 29, you would probably select every 6th person-- but you would be short one-- so you might select the 5th person on your last selection --> 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 26 27 28 29)

6. Collect data from the sample you selected.

7. Use the Find Confidence Interval calculator to determine the actual confidence interval for your sample. In the sample below, the researcher sampled 49 people from a population of 100.  He or she wished to be 95% confident of her results. 20% of her subjects indicated they liked yellow better than blue.  The calculator shows that she has an 8.04 confidence interval.  In other words, he or she is 95% confident that between 11.94 and 28.04 percent of her population likes yellow better than blue. There are instructions on using the Find Confidence Interval calculator.
   
   

8. Report your findings for your sample. Compare your results with your partners' results.
   
   Please submit the following:
       -- The type of sampling procedure used, (1 point)
       -- a description of how you selected the subjects, (1 points)
       -- what percentage of your subjects favored each option, (1 point)
       -- the actual confidence interval (range i.e., between 12.5 and 22.5) with a 95% confidence level for the sample. (1 points)
       -- Finally, based on what your partners found, discuss in the assignment you submit how and why the results from your target population were similar or different from the results your partners obtained from their target populations (1 points).

Del Siegle, Ph.D.
Neag School of Education - University of Connecticut
del.siegle@uconn.edu
www.delsiegle.info