More About Sample Size

If you want the full discussion of sample size -- Click Here

There are several additional features that help the user explore different sample size possibilities. The most general aid in exploring different sample sizes is the ability to retain on disk the results of a number of sampling calculations. If you look at the first example of the Sample Size Menu shown above you will see that the second line from the bottom is:

To record this display type r.

If you select this item you will be asked for a file name and the next time you see the Sample Size Menu this line will read:

This display is being recorded in Size Results.  To change this type r.

From this point until you use r to change the setting or you quit the program every time the Sample Size Menu, or a sample size explanation is displayed on the screen the same information will be written to a disk file called "Size Results." This file can later be opened and read by any word processing program (or even a spreadsheet although this will look strange.) With this command you can later print out the results and look at all the choices that you have considered.

A second option that helps investigate different sample sizes is the ability of the program to let you specify a particular sample size and have the machine calculate what the precision would be for that size sample. When you chose the Sampling Menu option: "p -- Precision," You will see the following message:

Please enter a value for precision.
Enter the figure as a whole number with no decimal point or percent sign.
Enter zero if you want me to calculate the precision for a given sample size.
Press the return key when you have entered precision.

If you enter a zero, the Sample Size Menu will look like this:

                              SAMPLE SIZE
Type the letter to set or change the values used in calculating a sample size

p -- Precision--how close the sample value should be to the true value
                I will calculate precision for a specific sample size.
c -- Confidence--how certain that the observation is within the
                given precision.
                you have chosen 90%
v -- Variability--what percent of the sample will be in the most
                important category?
                you have chosen 50%
s -- Sample size -- to be used to compute a precision value.
                you have not yet chosen a sample size.

This display is being recorded in Size Results.  To change this type r.
To quit this menu just type return.

There is now a menu choice "s -- Sample size" that lets you specify a sample size for which the precision will be calculated. As soon as you pick a Sample Size the Precision part of the menu will change to display a calculated value for precision.

A final aid in experimenting with different sample sizes appears then you use the "e -- Explain" item that appears in the Sample Size Menu after you have given all the information required to select a sample size. In the example at the end of the last section last lines from the explain message said:

Do you want to correct the sample size for the population size?   
Enter yes or no

If at this point you answer "yes" you will be asked to estimate the size of the population from which you are sampling. After you enter a number the response takes the form:

Correcting for a population of 4000
the sample size is 254

Note that correcting for population size is only done one time. If you change any of the Sample Size Menu values the new sample size will be calculated without correcting for population size. You must tell the program to make the correction every time you want to see the information.


this page is at