Page 95 - Design for Six Sigma for Service (Six SIGMA Operational Methods)

76 Chapter Four

There are several methods of probability sampling: random sampling, sys-
tematic sampling, stratified random sampling, and cluster sampling. We
discuss each of these probability sampling methods and nonprobability
sampling in detail.

Random Sampling
The best-known probability sampling method is random sampling. In the
random sampling method, each unit in the sampling frame is assigned a
distinct number. Then the units are chosen at random by a process that does
not favor certain numbers or certain patterns of numbers. The chosen units
will become the sample. A commonly used method to randomly choose
units from the sampling frame is the use of the table of random numbers.
Table 4.1 shows a portion of a table of random numbers.

Suppose there are 1000 people in the sampling frame and we want to select
a random sample of 30 people. Each person will be assigned a number
ranging from 000 to 999. Using Table 4.1, we can then arbitrarily select
three digits from the five digits given. For example, we can choose the last
three digits. In this case we will select the people with numbers 073, 849,
761, 622, 905, 276, 837, … ,033.

For large samples the use of a random number table will become tedious
and time-consuming, so computer-generated random numbers can be used
to select a random sample.

Systematic Sampling
Systematic sampling is an adaptation of the random sampling method. It is
used when the sampling frame is quite large and the sampling units cannot
be easily numbered. For example, if a sampling frame has 3,000,000 people

Table 4.1 A Portion of a Table of Random Numbers

1 2 3 4 5 6
1 77073 51849 15761 85622 38905 72276

2 20837 95047 50724 16922 04405 30858
3 37504 15645 36630 28216 10056 97628
4 40392 58557 60446 11553 60013 38037

5 53408 14205 33152 70651 17314 93033