The ideal situation is to have a list of all sampling units to extract the sample from. This allows for a representative sample that will eventually lead towards a unbiased estimates. So the ideal situation would be to obtain an update list of all sampling units. However, this is not always possible, and other options are possible.
The best alternative option is area sampling: the total area under investigation is divided into small sub-areas which are sampled at random. Each of the chosen sub-areas is then used to select sampling units from. Therefore, a complete listing will be needed for each of the selected sub-areas, but not for the whole population. This will allow for probability sampling and unbiased estimates.
In the other end, the worse available option is to use snowball sampling: sampled units recruit other participants. It is used where potential participants are hard to find. It’s called snowball sampling because (in theory) once you have the ball rolling, it picks up more “snow” along the way and becomes larger and larger. Snowball sampling is a non-probability sampling method and yields biased estimates.
An intermediate option is to use key informants to obtain a list of sampling units, the most accurate possible.
From Alex:
The ideal situation is to have a list of all sampling units to extract the sample from. This allows for a representative sample that will eventually lead towards a unbiased estimates. So the ideal situation would be to obtain an update list of all sampling units. However, this is not always possible, and other options are possible. The best alternative option is area sampling: the total area under investigation is divided into small sub-areas which are sampled at random. Each of the chosen sub-areas is then used to select sampling units from. Therefore, a complete listing will be needed for each of the selected sub-areas, but not for the whole population. This will allow for probability sampling and unbiased estimates. In the other end, the worse available option is to use snowball sampling: sampled units recruit other participants. It is used where potential participants are hard to find. It’s called snowball sampling because (in theory) once you have the ball rolling, it picks up more “snow” along the way and becomes larger and larger. Snowball sampling is a non-probability sampling method and yields biased estimates. An intermediate option is to use key informants to obtain a list of sampling units, the most accurate possible.