Probability Sampling and its types
What is Probability sampling? What are its types?
Probability sampling is a sampling technique that allows each participant equal chances of of being selected in the process of sampling. The researcher uses methods of sampling that guarantees each subject equal opportunity of being selected. It is achieved by using the process of randomisation. This helps remove both systematic and sampling bias. This is the main advantage of random sampling that if done properly, the selected sample is representative of the entire population. Systematic bias is the difference between the results from the sample and the results derived from the population. The effect of randomisation is that either systematic bias is totally removed or remains minimal. Sampling bias also automatically gets removed when the subjects are selected randomly.
The types of Probability sampling are as follows:
- Simple Random Sampling (SRS)
- Stratified Sampling
- Cluster Sampling
- Systematic Sampling
- Multistage Sampling
1. Simple Random Sampling:
The simplest form of sampling is Simple Random Sampling. A researcher just has to ensure that he includes all the members of the population and then the required number of subjects are selected on a random basis. However, there are several methods to do it and one can do it manually as well as using a computer. Suppose you randomly select names using the lottery method and blindfolded or you let a computer software do it for you.
2. Stratified sampling:
This probability sampling technique first classifies the subjects into various groups based on various classifications. It is useful when it is sensible to classify the population into various groups called based on a factor which may influence the variable which is being measured. While the bigger group is called strata, the individual units are called stratum. When using stratified sampling, one must:
- Classify the population into groups.
- Get a simple random sample from each group.
- collect data on each sampling unit randomly taken from each group.
After having classified the subjects into various groups like age, gender or socioeconomic class, the researcher can select the final list of subjects from the different strata. It is also important to set different strata where there must be no overlapping. When researchers are interested in studying a particular subgroup within a population then they use stratified sampling. The reason that this technique of probability sampling is preferred over the simple random sampling is because it warrants more precise statistical results.
3. Cluster sampling:
Several times size of the population makes it impossible to perform simple random sampling. Suppose you have to perform a research on the entire population of US. It would not be possible to sample the entire population.
- Cluster sampling first identifies boundaries and in the case of US several types of boundaries can be identified. For example based on time zones or states.
- In cluster sampling, the researcher selects identified areas randomly and it is important that each area (US state or time zone) stands equal opportunity of being selected.
- Researcher can have all the subjects from the selected areas or he can randomly select them from the identified areas.
Cluster sampling is a lot different from stratified sampling. With cluster sampling, the researcher must:
- Divide the population into groups or clusters.
- he must obtain a simple random sample of so many clusters from all the possible clusters.
- he must obtain data on every sampling unit from each of the randomly selected clusters.
There are other differences between stratified and random sampling too. These clusters are not subsections as in stratified sampling but instead miniatures like a microcosm. Moreover, each of these clusters must be heterogeneous. Apart from that, the statistical analyses used in the case of cluster sampling are also more complex than the ones used in case of stratified sampling.
4. Systematic sampling:
Systematic sampling is similar to arithmetic progression; that is the difference between any two consecutive numbers always remains the same. Say for example you have a class of 100 pupils.
- You must start by selecting a number that is smaller than the total number of subjects. For example: 5
- You must select another integer which marks the difference between two consecutive individual subjects. e.g. 7.
- your subjects will be pupils numbered 5, 12, 19, 26, 33, 40, 47, 54… and so on.
However, the use of this technique does not offer any clear advantage.
5. Multistage sampling:
Multistage sampling uses a combination of more than one of the above highlighted methods of sampling. Many times using a single sampling method is either not suitable or not sufficient. In such a case more than one sampling methods are used in combination to conduct the research. Several times research is conducted in stages and each stage uses a different sampling technique. Complex researches whether conducted on field or inside the lab, mostly use multistage sampling.