Revision Date: 2005-09-06
| Sampling is a complex subject. It can be divided into analytic and enumerative. In general, analytic sampling tries to predict what is going to happen (will the process stay the same, for example), and enumerative sampling tries to determine something about an existing population (what is the percent bad in the shipment just received, for example). You need to determine the purpose for taking the sample and what information will serve to accomplish the aim.
The main difference in taking the samples is that for a variable sample, measurements of a characteristic of interest are taken, and for an attribute sample, one counts the number of units having or not having specific properties (mostly good/bad or number of flaws). Generally, attribute samples are much larger than variable samples and to be useful, need to be very large, when the proportion of bad units (or flaws) is very small.
Gordon Constable, Ph.D. | |
