• Post author:

## What is Non Probability sampling? Highlight its types.

Non Probability sampling unlike probability sampling does not offer all the subjects equal chances of being selected. True random sampling is always difficult to perform because the researchers are bound by time, money and other constraints which inevitably leads to them using another sampling technique called non probability sampling. The central characteristic of non probability sampling is that the researcher selects his subjects based on his subjective judgment and not using random methods. Accessibility of the subjects become a primary factor affecting whether a subject will be selected or not. There is a downside to non probability sampling due to which many researchers consider it an inferior method as compared to probability sampling. Since an unknown proportion of the population was not sampled, the sampled subset may or may not be entirely representative of the entire population. It is why to generalise such results to the entire population is not possible. However, despite this shortcoming, there are both theoretical and practical reasons behind the use of non probability sampling.

Sometimes to carry a research using a probability sampling technique can be very difficult or entirely impossible because of the type of population being hidden or hard to reach. Many times a list of the population does not exist and so snowball sampling technique offers a solution for such situations. In case of exploratory research too, non probability sampling may prove quite useful which  is because the researcher needs to find out if a problem or issue really exists quickly and within minimum costs. While non probability sampling often gives rise to a biased sample, but if a problem does not exist in your biased sample then it is unlikely to exist in a non biased sample. Otherwise the same process would get more time consuming and expensive. When one is considering the use of non probability sampling, he must check if his choice of research strategy fits with the non probability sampling. Even if the non probability sampling model fits within the research strategy, it is important to select the right non probability sampling techniques.

The types of non-probability sampling techniques are laid out in detail below:

– Quota sampling

– Convenience sampling

– Purposive sampling

– Self-selection sampling

– Snowball sampling

### 1. What is Quota sampling?

It is a form of non probability sampling technique where the aim is to ensure proportional representation of subjects from each strata depending upon the trait being considered for the basis of quota. For example the basis of quota has been selected as year in the college. In that case to ensure proportional selection, the researcher must select equal students from each year. Suppose his sample size is 200, then he must sleet 50 from each year from first to fourth.  Now think of another case where you are doing a comparison of the career goals of male and female students in a college. To get a proportionate sample would be possible only if there were equal number of male and female students in the college. Let’s say there are 10000 students out of which 6500 are male and 3500 females. To get a proportionate sample which is the aim of quota sampling, you must select 65% males and 35% females in your sample size. So, if your sample size is 200, there would have to be 130 males and 70 females in the group.

In case you are unable to carry out a probability sampling and still want the sample to be as representative of the population as possible, Quota sampling can be useful. One can also see it as the non probability sampling counterpart of stratified random sampling.

It is easier and quicker to carry out than stratified random sampling because of no need of a sampling frame or strict use of random sampling techniques.

While quota sampling improves the representation of specific strata within the sample, it also ensures no overrepresentation.

Comparison of groups/strata becomes easier due to stratification of samples in quota sampling.

Sampling error cannot be determined because the random selection of sampling units has not been done. Sampling bias usually results because election of units being carried out on the basis of ease of access and cost considerations. Problems of generalisations result because it is not possible to derive statistical inferences from the sample to the population.

There must be clear stratification.

If the sampling needs are extended, it will also extend the need for stratification and result in higher costs and time being taken. For example if you are interested in how career goals change among the male and female students with each passing year, you will need to create four more strata based on each year in college apart from the first two.

How to create a quota sample:

Choose the basis for stratification and divide the population.

Calculate quota for each group. As we did earlier, we can include 65% males and 35% females for proportionate representation.

Ensure that the quota for each group is met and continue to invite cases till then.

### 2. What is Convenience sampling?

Convenience sampling is perhaps the most common of all sampling techniques which is because it is the easiest to perform. It is because samples are selected on the basis of accessibility. This is quite the opposite of probability sampling.  Subjects are selected on the basis of ease of recruitment in convenience sampling. It is not just the easiest and cheapest but also the least time consuming of all the sampling techniques. Suppose you want to research the shopping preferences of retail customers of a specific retail brand like Walmart, you can stand at the entrance of a Walmart store and talk to your subjects. If your sample size is 100 customers, it would not be difficult to come across 100 customers willing to be a part of the research. Thousands of customers visit the store each day but you only need to sample 100 from them conveniently.  As such possibly there is no other sampling technique which is less time consuming or less costly.

– There being no strict requirements regarding sample collection, convenience sampling is the easiest one to carry out.

– Compared to probability sampling techniques, Convenience sampling is not just the easiest but also the cheapest. So you can achieve your sample size without investing much time or money.

– Convenience sampling can help you gather useful data and information that is otherwise not possible using probability sampling techniques.

– Major biases can occur in the case of convenience sampling. For example whether you are studying the shopping choices of retail customers or job motivators for employees of large organization, the chances of both over and under representation exist.  You do not know why some students decided to participate in the survey and others did not and how career choices have changed over years. Similar multiple biases can occur in case of convenience sampling.

– Generalisations cannot be made form sample to the population which is because you do not have a definite sampling frame and the sample is not selected at random which means there is an inherent bias and the sample is not very likely to be representative of the entire population.

### 3. What is Purposive sampling?

Purposive sampling is also known as judgmental or subjective sampling where you have a specific purpose in mind when you are selecting samples. The sections are made based on the judgement of the researcher. Usually the samples are quite small in size and the researcher believes some subjects would be more fit for investigation than the others. Purposive sampling mainly focuses on specific traits of a population that are of interest and can help you answer your research question.

– There are a variety of purposive sampling techniques that can be used across various qualitative research designs.

– Different purposive sampling techniques offer different advantages.

– Purposive sampling offers a wide variety of non probability sampling techniques that can be particularly useful for multistage qualitative research.

– The chances of researcher bias are higher with purposive sampling. However, its purposive component can be a disadvantage only in cases where judgments are based on weak or unclear criteria.

– Due to the subjective nature of unit selection, the representativeness of sample is doubtful.

Types of purposive sampling:-

Maximum variation sampling: –

Also known as heterogeneous sampling, this technique is used to capture wide ranging perspectives related to the perspectives you are interested in studying. It captures wide variations ranging from typical conditions to extremest conditions.

Homogeneous sampling:

This sampling technique intends to achieve a homogeneous sample where each unit carries the same trait.  It is the opposite of the heterogeneous sampling technique.

Typical case sampling:

This technique is used when researcher is interested in the normality of units and because they are normal/typical.

Extreme case sampling:

This sampling technique is used in cases whereto focus is on special or unusual cases like notable successes and failures. this is the opposite of typical cases.

Critical case sampling:

This technique proves particularly useful in exploratory qualitative research with limited resources or research where a single casey prove decisive in explaining the phenomenon of interest.

Total population sampling:

This sampling technique is used when you chose to examine an entire population with  a particular set of characteristics. It is done in case where the entire population of units with desired traits is small.

Expert sampling:

This sampling technique is used when there is a need for experts with expertise in particular areas for the purpose of research. It can happen in various cases like during the exploratory phase of qualitative research or when a particular expertise s the basis of the research being conducted. It is mainly useful in areas where there is lack of empirical evidence or when the results of the research may take too long to be uncovered.

### 4. What is Self-selection sampling?

Self section sampling as the name indicates allows the subjects to select themselves. It allows them to choose to participate in the research. This sampling method can be used with a wide range of research designs and research methods. For example researchers may put n online form for students from a university to participate in the research program. Similarly the scientists needing to conduct research involving human subjects must invite volunteers. So a key characteristic of self selection sampling is that researchers do not approach the subjects but that they take part in the research of their own accord. people also have various reasons to participate insect studies including interest and inclination as well as strong support for a specific cause.

#### Advantages of self selection sampling:

– No need to approach sampling units or individuals since people volunteer of their own accord.

– Willingness and commitment of volunteers helps gain better insists into the phenomenon being studied.

#### Disadvantages of self selection sampling:

– Self selection can also give rise to sampling bias since such samples will carry an inherent bias towards the topic.

– This may lead to the sample not being representative of the population.

Creating a self selection sample involves two steps: inviting participants and checking if they match the criteria; accepting or rejecting them.

### 5 What is Snowball sampling?

Snowball sampling is used in cases where the sample population is difficult to reach or hidden. For example in case of drug addicts, AIDS patients, domestic violence victims, prostitutes and similar other stigmatised or socially marginalised populations meeting the sample size can be very difficult. In such cases non probability sampling can be used to gain access to such populations. In case of snowball sampling, the researcher tries to find one or two such units and then uses them to find more such units till the sample size is met.

– In above mentioned cases finding an obvious list of sampling units and population may not be possible. One cannot find a list of drug addicts or prostitutes.

– Such people would generally not come forward for fear of social stigma. Since snowball sampling uses recruits from sample population to recruit more sample units, it makes it possible to reach such people easily.

– Moreover, the secret nature of particular subgroups also makes it difficult to access them.

– Possible sampling error cannot be determined and it is not possible to make statistical inferences based on the sample population.

Sources:

https://dissertation.laerd.com/snowball-sampling.php

https://dissertation.laerd.com/purposive-sampling.php

https://dissertation.laerd.com/convenience-sampling.php

https://dissertation.laerd.com/quota-sampling.php

https://dissertation.laerd.com/self-selection-sampling.php

https://explorable.com/non-probability-sampling?gid=1578

Research Methods for Business Students By Mark Saunders

### Abhijeet Pratap

Abhijeet has been blogging on educational topics and business research since 2016. He graduated with a Hons. in English literature from BRABU and an MBA from the Asia-Pacific Institute of Management, New Delhi. He likes to blog and share his knowledge and research in business management, marketing, literature and other areas with his readers.