Activity 9.3
Using SPSS to Select a Random Sample

SPSS allows you to select random subsamples from existing sets of data. This facility can be used to help you understand the nature of sampling error.

1. Open SPSS and then open the "COUNTRIESdepr" data file (download / help downloading).
2. Run a frequency distribution for the variable freestat (Freedom Status) and record the percentage of countries that are categorized as "free." (See Exercise 8.4 if you need to review how to run a frequency distribution.)
3. Click on "Data," then on "Select Cases."
4. In the "Select Cases" box select "Random sample of cases," then click on "Sample."
5. In the "Select Cases: Random Sample" box under sample size select "Approximately," type in 20%, and then click "Continue." In the "Select Cases" box click "OK."
6. Run a frequency distribution for your selected sample for "freestat" and record the percentage of countries that are categorized as "free."
7. Calculate the total sample error for your sample by calculating the difference between the percentage categorized as "free" for the entire data set and the percentage categorized as "free" for your random sample.
8. Repeat the process of sampling four more times and calculate total sample error for each sample.
9. Draw a chart similar to Figure 9.5 to demonstrate total sampling error. Plot the percentage of countries for the entire sample categorized as "Free" as the population parameter. Indicate the statistical estimate of the percentage of nations that are "Free" for each of your five samples.