Because record sampling treats each record equally, the amount to be sampled is 1 for all samples. Also, larger items are no more likely to be selected than smaller ones. This example has the following characteristics.
Line 2 is selected because the random start is 2. Because the interval is 5, every 5th record after that is selected; in this case, records 7 and 12.
The records are grouped into sets of 5 records. We assume that the following records were chosen:
• | The first record in the first group |
• | The fourth record in the second group |
• | The second record in the third group |
• | The first record in the fourth group |
This method has generated one more selection than the fixed-interval method, due to the chance selection of the first record in the last group.
We assume that the three random values between 1 and 16 are 2, 5 and 14.