Purposive Sampling and its Types

When choosing the method of sampling, we usually try to make sure that the samples represent the general population. But there are some cases where we might need an inquiry on certain items that fulfill specific criteria. This is where purposive sampling is useful.

In purposive sampling, the items are selected in such manner that each of them are rich in information about the parameters that we are trying to study in the population.

purposive sampling figure

In the above figure, a pollster wants to conduct a survey about the latest fashion trends in his city. The pollster assumes that the information he is seeking can be best found from the people with age group 16-30. So, he selects only people from that age range (shown in darker shade) as samples. The main target for this sampling is not to represent the whole population but to get sufficient information on fashion conscious people.

Purposive sampling focuses on the theoretical aspects of the data, explores the characteristics of the items selected to form an opinion on them. Hence, purposive sampling focuses on qualitative research.

Types of Purposive Sampling

Purposive Sampling Types
Politically Important Cases Sampling
Critical Case Sampling
Snowball or Chain Sampling
Criterion Sampling
Convenience Sampling
Expert Opinion Sampling
Combination Sampling
Opportunistic Sampling
Operational Construct Sampling
Confirm or Disconfirm Sampling

Politically Important Cases

In politically important cases sampling, some cases or issues with higher importance are focused while some sensitive or undesirable cases are eliminated. When the inclusion of a sample affects the notion of the overall poll to a great extent, such samples are excluded.

For example, a government is conducting a poll regarding the fear of terrorism in common people in a particular region. The view of the people who have been under direct threat from terrorists or directly affected by a terrorist attack will be too strong and might bias the analysis.

In such cases, some views which seem to be more influenced, are excluded and the view with apparently lesser influence will be taken for further analysis.

Critical Cases Sampling

Sometimes, we can make logical generalizations by taking a handful of special cases because they provide us enough information about the general population.

For example, a mathematics teacher wants to know if the 200 students of his class understood the lecture he gave on Probability Sampling. he will chose 5 of the most brilliant students and 5 out of the weakest ones and ask them some questions related to the topic.

Here is a list of probable results and the recommended conclusions for them:

Probable Results Recommended Conclusion
Everybody gave correct answers The class was quite effective
Nobody gave correct answers The class was ineffective
Brilliant students gave correct answer, but others could not The lecturer did not illustrate the details well
A mix of brilliant students and weak students gave correct answer Some students might have been distracted

Critical Case Sampling assumes that “If it is true for this one case, it is likely to be true of all other cases”. In the example, if even the weakest students can answer the questions, we can generalize that everyone must have understood the lecture. Similarly, if even the most brilliant students couldn’t answer, we can reach the conclusion that the lecture was ineffective. In this way, by choosing samples rich in certain information, we got information about the general population that is the main advantage of purposive sampling.

Snowball or Chain Sampling

In purposive sampling, we are always in search of samples that are rich in certain information. In snowball or chain sampling, the data is collected in such a way that helps in selection of next samples too. We form a chain of samples and since each selection helps in the selection of next sample through the snowball effect, it is called snowball or chain sampling.

On some occasions, the people we select can help us or refer us to other people who might be rich in the information we are looking for. For example, if you are conducting a survey regarding the popularity of a tourist destination, we first start by finding a few tourists. Once we find them, we can find more by asking them about other tourists who they know have visited the destination.

In this way, finding one sample helps us find another to form a chain of samples that are rich in information that the surveyor needs.

Criterion Sampling

This method selects those samples that satisfy certain specified criteria that help us to separate out the information rich samples.

For example, if the tourism ministry wants to focus their efforts on the needs of the regular tourists, they can survey tourists who have visited a particular place at least thrice and have been there at least in two different seasons. Such tourists can tell us more about their experience about the place.

Criterion Sampling helps us to get more information from smaller number of samples.This can save us time and money because the information collected from samples are collected applying multiple criteria can be useful to multiple purposes too.

Convenience Sampling

When the surveyor chooses the samples based on his/her convenience like geographical proximity or ease of contact with the samples, it is called convenience sampling. This saves time and money because the surveyor can quickly find samples at low cost.

For example, when you are asked to survey about the habit of drinking alcohol in your town, you will probably not go and roam all around the town asking people about their drinking habits. What you can do is meet a few households around you and ask them about those who drink alcohol.

It might not exactly reflect the position or characteristics of the entire town but it still gives information about some major characteristics at a very low cost . It increases the convenience of the surveyor and reduces the cost associated with selecting the samples . If we had selected samples at random, contacting them, convincing them to take part in the survey and collecting data would have been time-taking and costly processes.

Expert Opinion Sampling

The opinions of experts in any field are considered to be more reliable. Expert opinion sampling is a method that gives preferences to the experts of the field under study when choosing the sample from the population and their views are regarded as authoritative while doing the research study. As the opinion of experts can be more convincing to others, this can make the users of the data more confident on the conclusion drawn from the samples.

For Example, If we are to survey about the best building material for a specific region, we may choose the views of those structural engineers who have at least 10 years of experience and received at least one national level award.

By choosing opinions of experts, we can increase the validity of the research and save time and cost because the experts can give us more reliable and accurate information, are easy to separate out based on criteria(e.g. we can survey PhDs in the field of study) and can give us the information we need quicker due to their experience.

Combination Sampling

The process of triangulation or combination of two or more methods of sampling to get more effective and more specific results is called Combination or Mixed Purposeful Sampling. The idea is that if we apply different methods of sampling and get the same results upon analysis of both sets of samples, we can be more confident that our results are accurate.

For example, when we are to survey about the cinematic quality of a movie, we may want to choose those viewers who have done at least graduation in film making, direction or any other faculties related with films ( Expert Sampling) and have watched the movies at least twice (Criterion).

After we get the results by taking opinions of experts and from those who have watched the same movie at least twice, we analyze the results. If we get similar results from both set of samples, we can conclude that the result is accurate.

Opportunistic Sampling

This method takes samples as they appear in their natural state. Elements are chosen as samples if they satisfy certain criteria and are available at the time at which the observation is being conducted.

For example, a student doing research on customer satisfaction choosing her friends who are at canteen as the samples.

Theory-Based or Operational Construct

Sometimes a person wants to study about a Theoretical Construct like anxiety. Since it is very difficult to quantify anxiety, the researcher operationalizes it in terms of social stress. i.e. he tries to study anxiety by relating it with social stress. The people who might be facing anxiety can now be traced easily. We just have to find those who are going through social stress because of their condition For Example, someone who has recently become homeless, physically disabled people etc.

In this way, even theoretical topics can be researched by choosing an appropriate operational construct and conducting the sampling process based on it.

Confirm or Disconfirm

After the sampling process, the researcher may want to further analyze the emerged patterns. The researcher then picks some other information rich samples to get confirmed about the previous results. This is mostly used for drug tests. After we conduct the first test, the second is tested to Confirm or Disconfirm the results of the first.

Such process taken to confirm or disconfirm a theory, principle or a conclusion of a sampling process is called Confirm or Disconfirm Sampling. If the results from the Confirm or Disconfirm Sampling suggests the same as the previously found results, it confirms the results. If the results suggest another or opposite suggestions, this will disconfirm the previous results and another theory or principle will be drawn.

For example, a survey was conducted about the views of public on economic issues. The survey concluded that the people believe that the world is about to adopt socialism in near future. To confirm or disconfirm this result, we can conduct another survey.

From the above discussion, we can infer that Purposive Sampling can be very much useful for the qualitative researches and other similar purposes to deduce a theory or develop some principles and theories.