Probability Sampling Technique

Probability sampling technique helps the researcher to create sample from a population of interest for a particular study with the use of some form of random selection. This technique works on the fact that each participant in a population has an equal chance of being selected for the study.

Advantages of Probability sampling technique

  • Avoids/minimizes sampling bias in research

  • Ensures sample is true representative of the population

  • Provides significant statistical estimation for analysis purpose

Disadvantages of Probability sampling technique

  • Must know exactly how many sampling units needed to represent the target population

  • Can be expensive and time-consuming

Types of Probability sampling technique


1. Simple Random Sampling


This is the simplest probability sampling technique. In this type, each participant gets an equal opportunity of being selected as a sample.


For example, researcher has 24 individuals in which he wants to generalize his findings. For that he needs 6 individuals as sample. Now, the researcher can choose any 6 individuals out of 24 individuals with the help of lottery system, computer-generated list, or any other random method.

This technique can be useful when there is a very large population and it is difficult to identify each participant of the population.


Advantage of Simple random sampling

  • Each participant of the population has an equal chance of being selected as subject of the study

  • Easy to implement

Disadvantage of Simple random sampling

  • In large population, identifying each participant of the population becomes difficult which may result in leaving out many elements of the population that may be significant to the study

  • Required complete list of units of the population

2. Systematic Sampling

In this technique, participants are selected at regular intervals using a sampling frame. In this type, first unit of the sample is chosen randomly from the population and rest units will be taken at certain intervals or can say every Kth number (i.e., sampling interval) unit will be taken as sample.


Kth number can be calculated by dividing population size by sample size. Thus, if the population size (N) is 24 and sample size (n) is 6 then kth number should be 4.


This technique is useful when the given population is logically homogenous.


Advantage of Systematic sampling

  • Assures that population will be evenly sampled

  • More precise and easier to implement (compare to simple random sampling)

Disadvantage of Systematic sampling

  • If the list has periodic arrangement, then sample created through this list may not be a true representative of the population

  • Complete list of units of the population is required

3. Stratified Sampling

In this type, population is divided into two or more homogeneous groups according to one or more common attributes (e.g., gender, age, etc). These groups are known as “strata”. Now, any random sampling technique will be used to choose the sample unit from each stratum (strata is plural of stratum).


This technique is useful when researcher is interested in studying particular strata (groups with particular attributes) within the population (e.g., males vs. females). Stratified random sampling provides equal opportunity for selecting sample units from within a particular stratum of the population.


Stratified sampling can be of two types:


(A) Proportional stratified sampling


The size of sample selected from each stratum is proportional to the size of that stratum in the entire population. It means each stratum has the same sampling fraction.


Example - researcher creates two strata called stratum “A” and stratum “B” having 200 and 400 units in each stratum respectively. If the researcher chooses a sampling fraction of ½, it means 100 units (1/2 of 200 units) from stratum “A” and 200 units (1/2 of 400 units) from stratum “B” will be selected randomly regardless of the different sizes of strata.


(B) Disproportional stratified sampling

The size of sample selected from each stratum is disproportional to the size of that stratum in the entire population. It means different strata do not have same sampling fraction as each other.


Example - researcher creates two strata called stratum “A” and stratum “B” having 200 and 400 units in each stratum respectively. In disproportional stratified sampling, sampling fraction can be different for stratum” A” and stratum “B”. Suppose, stratum “A” has a sampling fraction of ½, it means 100 units (1/2 of 200 units) from stratum “A” will be taken out. Similarly, stratum “B” has a sampling fraction of ¼, it means 100 units (1/4 of 400 units) from stratum “B” will be taken out.


Advantage of Stratified sampling

  • Can capture certain groups of interest

  • Ensures representation of individuals across the entire population

  • High precision if variability within strata is smaller than variability between strata

Disadvantage of Stratified sampling

  • Requires accurate knowledge of population

  • Time consuming and expensive

  • Analysis can be complex and measuring sampling error can be difficult

4. Cluster Sampling

In cluster sampling, sub-group (cluster) of population is used as sampling unit instead of individuals. Clusters of participants that represent the population are identified and are used to create sample for the study. Then all units or randomly selected units within randomly sampled cluster is studied for the results. Cluster sampling can be single stage, two-stage or multi-stage.


The principal difference between cluster sampling and stratified sampling is that cluster is considered as sampling unit while in stratified sampling, certain attribute/individual of strata is considered for sample creation.


Cluster sampling is useful for area or geographical sampling. In this case, area can be divided into clusters and clusters as sample units can be selected for the study.


For example – researcher is interested in studying the academic performance of college students in a state. He classifies the population of the state in the groups (cities) called clusters and chooses randomly the clusters for the study.


Advantage of Cluster sampling

  • Sampling frame is not needed

  • Simple and easy to use

  • Less resource required

  • Very useful when population is large and spread over a large geographical area

Disadvantage of Cluster sampling

  • Statistically less efficient when cluster elements are homogeneous

  • Less likely to represent the whole population

  • Chances of higher sampling error

5. Multi-stage Sampling

Multi-stage sampling uses a combination of techniques. This technique combines the other sampling techniques (mentioned above) in the efficient manner to create sample for the study. These different combinations of sampling techniques are known as multi-stage sampling.


In this technique, large clusters of population are divided into smaller clusters/strata in multiple stages in order to make the research convenient.


For example, a researcher wants to calculate the percentage of the college going boys and girls who are punctual and who are late goers in a particular state. In order to make this research possible, colleges of the state can be considered as clusters. Now randomly selected clusters (colleges) can be used to form stratum of boys and stratum of girls. Post this, any simple random sampling technique can be used to choose the sample from two strata.


Advantage of Multi-stage sampling

  • Sampling frame is not required

  • Most feasible approach for large population

Disadvantage of Multi-stage sampling

  • Each stage of sampling can introduce error

  • Required more resources