The target population of our primary research might be very big, therefore businesses decide to sample their target population. Samples should be representative: they should have the same characteristics as the target population. This means that we want the customers in our sample to have similar opinions on our product as all the customers that buy that product. There are different types of sampling businesses can use:
Random sampling: gives each member of the target population equal chances of being chosen.
Stratified sampling: also a random sampling method. However, before the samples are drawn, the sample population is divided into groups (called strata) based on the previous knowledge about the target population. Once those groups are set, researchers choose customers at random that fit in those strata.
Quota sampling: researchers divide the population into groups that share similar characteristics (e.g., age, gender) and then have pre-set number of people in each group they have to interview.
Cluster sampling: involves separating the target population into ‘clusters’ usually in different geographic areas. A random sample is then taken from each cluster.
Snowball sampling: used in special occasions, when firms are looking for specialists in a particular field they want to employ. This is based on pre-existing network of these specialists that the firm approaches and asks to work for them.
| Advantages | Disadvantages | |
| Random | ✓ Representative | ✓ Time-consuming and costly. ✓ Assumes that all the members of the target population are homogeneous. |
| Stratified | ✓ Representative. ✓ Each subgroup provide quality results. |
✓ Time-consuming and costly. |
| Quota | ✓ Cheap and not time-consuming. | ✓ Not representative. |
| Cluster | ✓ Useful when results are needed quickly. ✓ Useful for results about how the opinions vary geographically. |
✓ Representativeness of the sample is assumed. ✓ Assumes that all the members of the target population are homogeneous. |
| Snowball | ✓ Firms will get the results they are looking for. | ✓ Not representative. |
