In sampling, units are the things that make up the population.
Units can be people, cases (e.g., organisations, institutions, countries, etc.), pieces of data, and so forth.
In this respect, there are two aspects of this example that illustrate when total population sampling may be appropriate: Due to the very small sample sizes and the uncommon characteristics of populations that make up a total population sample, researchers generally look at these samples in-depth using qualitative research methods.
The examples of total population sampling below attempt to highlight two of the characteristics of total population samples, discussed above: (a) the fact that the population size is very small; and (b) the fact that the population shares an uncommon characteristic(s).
As discussed earlier in this article, units are the things that make up the population.
These units can be people, cases (e.g., organisations, institutions, countries, etc.), pieces of data, and so forth.
As with probability sampling techniques that require the researcher to get a list of the population (i.e., the sampling frame) from which a sample is selected, total population sampling also requires the researcher to get such a list.
However, as can be learnt from probability sampling, being able to get hold of such a population list can be very time consuming and challenging. It may also be difficult to build a list if the population is geographically dispersed or requires the permission of a gatekeeper not only to get the list, but also to contact members on the list.
If the list of the population is incomplete or a large (or even small) proportion of members choose not to take part in the research, the ability of the total population sample to allow the researcher to make analytical generalisations can be severely compromised.
To learn more about other purposive sampling techniques, see the article: Purposive sampling.