|
Carrying out a survey always has the aim to collect data about a certain population. In the easiest
Case, the population is small and, therefore, it is possible to question all relevant persons to get a
Complete picture of their opinions on a certain topic
|
|
As group size increases, the difficulties to question everyone are increasing too (organizing
The survey, costs of the survey, etc.). Although web surveys offer a very good cost ratio for large
Surveys, it becomes obvious that at a certain stage, only a sample of people can be questioned.
|
|
To gain a representative sample of customer opinions it is not necessary to speak to all customers.
|
|
Sampling allows companies to gain information from a small number of customers who will represent
the total universe of their customer base, therefore ensuring reduced costs and shorter timescales
for research.
|
|
Sample sizes are worked out on a mathematical basis using square roots. Therefore the sample size
needed to produce results within 5 percent accuracy (when the universe size is known) for a range
of universe sizes is, for example:
|
| Universe Number |
Sample Size |
| 500 |
222 |
| 1000 |
206 |
| 5000 |
370 |
| 10000 |
305 |
| 50000 |
397 |
|
|
Caution
|
|
The smaller the sample size the higher the probability of a range of error being present in research results.
|