19 Non-probability Sampling: Principles and Procedures

Soumyajit Patra

1. Objective

This module you will teach about the importance of non-probability sampling in social science research. At the end you will find some digital resources and a bibliography for further study.

2. Introduction

Data collection is the most important part of any research and the success of a research work depends on the accuracy of data. It is expensive and time consuming; it demands a well-thought planning too. In social science research both types of data – primary and secondary – are important. Secondary data are collected mainly from the documentary sources such as office records, police records, census reports, reports of the Planning Commission, records of the municipalities or corporations etc. But in most of the researches data are collected directly from the respondents and then they are supplemented by the secondary data. However, it is impossible to collect data from all potential informants considering the time-cost-labour components that such a large scale study incurs. Time, labour and cost of a study proportionately increase with the increase in the number of respondents from whom a researcher collects data. If we want to know, following Bourdieu, for example, how the process of socialization, by forming one’s habitus, plays a role in generating criminal propensities among the slum dwellers of Kolkata, it is impracticable to obtain data from each and every adult male and female slum dwellers of the city. So, we select a representative group from the population or universe to study the socialization process the respondents underwent and to observe their present behavioural pattern. This representative group is called sample. And the aggregate of individuals or units from which the sample is drawn is known as population or universe. Often the researchers want to know about the population characteristics by collecting relevant data from the sample. No doubt the ideal way to have knowledge of the population is to conduct a study on each and every member of the population. Sample is the short cut way to understand the population characteristics. In social science researches, particularly if the research is a quantitative one, the researchers generally draw sample from a population in order to examine the features of the population or universe. However, for the example given above non-probability sampling would be the best choice for the researchers because of the difficulties in getting a source list (complete list of the members of the population) in such kind of study. We shall get back to this point later.

3. Learning Outcome

This Module will help you understand some basic concepts related to sampling and the principles and procedures of non-probability sampling.

4. Sampling

According to Payne and Payne (2005: 200), ‘sampling is the process of selecting a sub-set of people or social phenomena to be studied, from the larger “universe” to which they belong.’ In the words of Bloor and Wood (2006: 153), ‘a sample is representative of the population from which it is selected if the characteristics of the sample approximate to the characteristics in the population’. This representativeness of the sample is very important, particularly if it is a probability sample, because it is presumed that the results obtained from the sample can be used to describe the population or universe as such. The individuals selected for a sample are called sampling units. In other words, a sample consists of sample units. When the research is conducted on the entire universe, i.e. when information is collected from each and every individual of the population it is called census.

According to Bryman (2012: 187), census is

‘the enumeration of an entire population. Thus, if data are collected in relation to all units in a population, rather than in relation to a sample of units of that population, the data are treated as census data’.

There are different types of sample and the researcher has to pay sincere attention in selecting the appropriate one. Otherwise, the findings of the study will be misleading. Prior knowledge of the various characteristics of the population is essential in many cases for the selection of the right sampling design.

In many non-probability samplings, like the probability sampling, the researchers prepare a source list (the complete list of the units of the population is known as source list or sampling frame) before the selection of the final sample. But sometimes, as in the example given in the ‘Introduction’, it is difficult to prepare a source list. The researchers in such cases have to resort to some kind of non-probability sampling like snow ball sampling to select the respondents for their study. Researchers face difficulties to prepare a source list if the population, for example, is mobile. In some situations the researchers cannot identify the actual units of the population. If, for example, we want to collect a sample from the student communities who have tendencies to commit suicide, it would not be very easy to locate the right persons and to collect a sample from them. In all these cases, non-probability sampling would be the best choice.

4.1. Probability Sampling and Non-Probability Sampling

Sampling can be of two types – probability and non-probability. According to Das (2004: 61), ‘the chance of being included in the sample is commonly known as probability.’ Non-probability sampling, on the contrary, does not follow the rule of probability. Bryman (2012) points out that non-probability sample is a sample which is not selected using a random selection method. That means in case of non-probability sampling some units in the population are more likely to be selected than others. In the words of Babbie (ibid.: 182), ‘any technique in which samples are selected in some way not suggested by probability theory’ may be called non-probability sampling. In some social science researches probability sampling does not seem feasible. In those cases, non-probability sampling is preferred. For example, if we want to study homelessness it is impossible to collect the list of such people. In this case non-probability sampling would be appropriate (ibid.). In non-probability sampling, there is no way to ensure that each item of the population has a chance of being included in the sample. The selection here depends, to a large extent, on the researcher and therefore the representativeness of the sample cannot be guaranteed in most of the cases.

There is little controversy among the sociologists about the basic features of a sample. A sample should possess at least two features: it should represent the population and it should be unbiased. In non-probability sampling, although the selection of the sampling units depends on the decision of the researchers themselves, they try to be impartial in such selection. Sociologists, in non-probability sampling, do not fix the sample size beforehand. One of the major advantages of non-probability sampling is that it is flexible as the researchers can adopt whatever they think suitable for their research at any point of time. It can be employed in such cases where it is impossible to prepare a source list. So it needs less time and cost. It has a disadvantage too. Sampling error cannot be calculated from a non-probability sample. As the selection of sample units depends to a large extent on the researchers, these should be properly qualified to ensure the reliability of his/her findings. Some sociologists argue that quantitative-qualitative dichotomy should be abandoned. A better way to understand the reality is to combine the two (Bryman 1988).

5. Sampling in Quantitative and Qualitative Research

Sampling techniques vary with the nature of research. In quantitative research generally the researcher wants to focus on the quantitative aspect of social life through the collection and analysis of some numerate statistical data like average age of the population, average income, dropout rate, etc. For this purpose, s/he wants to draw a truly representative sample from a large population and tries to understand the population parameter through sample statistic. According to Neuman (2007), probability sampling is most appropriate for quantitative research because probability sampling produces more accurate result expressed in terms of numerate data than the non-probability sampling and sampling error can be calculated.

Qualitative research, on the other hand, focuses on the peculiar features of social life, or on the meanings created and transformed in course of inter-human interactions, or sometimes on the inter-subjective feelings and emotions. These demand proper and in-depth understanding of the social reality that simple numerals cannot express. Neuman (ibid.: 141) writes:

Qualitative researchers’ concern is to find cases that will enhance what the researchers learn about the process of social life in a specific context. For this reason, qualitative researchers tend to collect a second type of sampling: non-probability sampling.

Self-Check Exercise -1

1. What is population?

The aggregate of individuals or units from which a sample is drawn is known as population. It is also called universe. In some cases, the researchers cannot precisely demarcate the population and resort to non-probability sampling.

2. What is census?

Complete enumeration is called census. In census, all the units of population are covered, that means, data are collected from each and every member of the population.

3. What is sample?

Sample is the representative of the population. When a researcher selects some units from the population or universe following some standardized procedures as representative of the population, this group is called sample. Sample reflects the characteristics of the universe.

4. What is non-probability sampling?

In Non-probability sampling each member of the population do not have a chance of being selected in the sample. This sampling is not based on the theory of probability. In it, the selection of the sample units depends on the researcher’s objectives and interests.

5. What do you mean by qualitative research?

Qualitative research is a kind of research that focuses on the qualitative aspect of social life, particularly on the meanings social actors create through their interactions in concrete situations. According to Bryman (1988), this kind of research emphasizes on participant observation, unstructured and in-depth interviewing. Case study method is also important for qualitative research. Non-probability sampling is used in most of the qualitative researches.

6. Types of Non-Probability Sampling

You know that in case of non-probability sampling, the chance or probability of a unit of being selected in the sample cannot be statistically calculated. The selection of sample here depends on the researchers, who, according to the objectives of their study, select the units of their sample. But in doing so, they suspend their values, beliefs and preferences to make it sure that the sample is unbiased. Researchers’ objectivity and value neutrality are equally important in non-probability sampling as they are in probability sampling.

There are different types of non-probability sampling. The researchers adopt any one of them according to the purpose and nature of their study. However, they should be very cautious in selecting the type of sampling because any wrong decision may jeopardize their research project. Before taking the final decision regarding the type of non-probability sample to be adopted for a particular study, the researcher should gather sufficient knowledge about the characteristics of the universe or population on which s/he would conduct the research. According to Neuman (2006), in non-probability sampling, the researchers, who mainly collect qualitative data, rarely draw a representative sample from a huge population. Rather they want to concentrate on small group of people and do not fix the sample size in advance. Neuman (ibid.) calls non-probability sampling a qualitative sampling and probability sampling a quantitative sampling.

However, when the population is highly homogeneous, non-probability sampling can safely be employed. Suppose, you are conducting a study on the beliefs and practices of the adult tribal women related to their own religion. You have decided to study 100 tribal women for this. You can prepare a source list (remember source list is not essential for non-probability sampling) and then select 100 women as per your convenience and objectives of your study.

6.1 Purposive or Judgmental Sampling

This is a kind of non-probability sampling in which the researcher selects the units of the sample not by random method of selection but on the basis of his or her own choice. In purposive sampling, the researcher should have some knowledge of the population in order to select the sample units purposively. Payne and Payne (2005: 210) wrote:

People and events are deliberately selected because they are interesting or suitable, rather than being representative. This ‘purposive sampling’ picks its sub-set for a particular, non-statistical purpose…. For example, we deliberately select Key Informants because they are not typical: they know more about the community or organisation than other people.

According to Majumdar (2005: 204), ‘the major guiding principle in purposive sampling is to select units with characteristics that have matching parallels in the universe. This matching can only be done on the basis of knowledge already available.’ In purposive sampling the rule of equal probability for all units in  the population of being selected in the sample does not work. The selection of sample units, instead, totally depends on the decision of the researcher. The representativeness of the sample, therefore, depends on the expertise and experience of the researcher. However, in some cases where the population is mobile or the units of the population cannot be properly identified, this type of sampling is preferred. In fact, some researchers think that it is impossible to generalize the findings of any social science research even if the sampling follows rigorous random method. So, instead of striving for generalization and producing some abstract law-like propositions, they try to understand empathetically the peculiar features of social life exhibited in a particular social context.

The significance of purposive sampling is that it gives the researchers an extreme autonomy to select the respondents/events for their study. So the goal of purposive sampling is to sample events/respondents in a ‘strategic way’, so that they can help to find out the answers of the research questions. Often the researchers want to sample in order to ensure that there is a good deal of variety in the resulting sample, so that sample units differ from each other in terms of key characteristics relevant to the research question. For instance, in a study on women workers in the informal sector, the researcher went for the purposive sampling because of a lack of any specific list of workers employed. This sampling frame allowed the researcher to select respondents from various sources and also on the basis of their marital status, namely unmarried, married, widow, divorcee or deserted respondents. A random sampling was not possible in that case (Choudhuri 2014).

Let us take another example here. Suppose you are interested to conduct a study on the pattern of friendship among the rural boys and girls. You can purposively select your sample from the population you have targeted. In such cases, it is difficult to prepare a source list and then select the sample following random method. So keeping in mind the objectives of the research, you can select the boys and girls from the villages, selected for the study, who can serve your purpose by providing necessary information. You can select equal number of boys and girls belonging to specific age groups from specific localities.

Bryman (2012) has differentiated between a convenience sampling and purposive sampling. In case of convenience sampling the researchers select the units of sample from among the available respondents. So the hazard of locating the potential respondents or units can be averted. Say, for example, a teacher is interested to know about the impact of advertisements shown on the television on the school children. S/he can conveniently select students as his/her sample from among the students, who are readily available. But this is not the case in purposive sampling. In purposive sampling the researcher does not bother much about the availability or accessibility of the respondents, rather the thrust is on the chance of getting information. As it is a non-probability sampling approach, purposive sampling does not allow the researcher to generalize (ibid.: 418). You know that sampling error also cannot be calculated from such a sample. Purposive sampling technique is particularly weak from the statistical point of view since ‘there is no way to calculate the limits of permissible error… the practical consequence is that the sample varies in unknown ways from the universe’ (Goode and Hatt 1982: 230-31).

6.2 Snowball Sampling

In snowball sampling the researcher locates a few persons of the population and then seeks their help to find out other potential respondents. Babbie (2004: 184) has defined snowball sampling in the following way:

In snowball sampling, the researcher collects data on the few members of the target population he or she can locate, then asks those individuals to provide the information needed to locate other members of that population whom they happen to know. “Snowball” refers to the process of accumulation as each located subject suggests other subject.

Snowball sampling is used in those cases where the units of the population are unknown or difficult to locate. Here also the representativeness of the sample cannot be assured. In fact, snowball sampling is almost exclusively used in qualitative researches which bother least for straightforward generalizations and rigorous scientism. The researchers do not predetermine the sample size as well. Snowballing ends when the researcher is sure that there are no more respondents to be selected or further increase in sample size will not provide additional information. Henn et al (2006) opine that snowball sampling is a form of purposive sampling, where the researchers want is to obtain a pool of respondents that, they think, is appropriate for the study. For obvious reasons, the selection of sample units largely depends on the decisions of the researchers.

Let us suggest an example here. Suppose the researcher wants to study the nature of relationships the pickpockets or gangsters maintain with the wider society. A major chunk of information, for this study, should have to be collected from the pickpockets of a selected city. However, you can understand the difficulty for the researcher to identify the pickpockets to collect the relevant data. What the researcher can do is that she or he can locate one pickpocket/gangster and then convince him/her to help in finding out others who also indulge in similar type of activities. Then again the researcher can collect information of some others from these respondents. In this way, the sample is collected through a cumulative process. According to Henn (2006), snowball sampling is effective for those groups of people who are difficult to locate or contact like the disabled persons, political activists, business elites, etc.

Atkinson and Flint (2003) have pointed out some distinct advantages of snowball sampling despite the fact that the sample drawn in this method is unlikely to be the representative of the population, even if the population or universe can be identified. However, they have noted the following points:

• It enables access to previously hidden populations.
• Snowball sampling is cost-effective.
• It has been shown to be capable of producing internationally comparable data.
• Snowball sample can be used to examine changes in the phenomenon understudy over time.
• It can produce in-depth results.

6.3 Quota Sampling

According to Babbie (2004: 184), quota sampling is a kind of ‘non-probability sampling in which units are selected into a sample on the basis of pre-specified characteristics, so that the total sample will have the same distribution of characteristics assumed to exist in the population being studied.’ In quota sampling, the researcher, after defining the population from which the sample is to be drawn, classifies the population on the basis of certain characteristics which s/he thinks relevant to his/her study, like stratified random sampling. Then the researcher calculates the proportion of the population falling into each category. In doing so, the researcher may take the help of some previous studies. The researcher now gives a quota (number of respondents to be selected) for each stratum, according to their proportion in the population and selects the sample units. In this selection random method is not followed. The selection finally depends on the decision of the researcher, like the purposive sampling. Bryman (2012: 203) writes:

The aim of quota sampling is to produce a sample that reflects a population in terms of the relative proportions of people in different categories, such as gender, ethnicity, age groups, socio-economic groups, and region of residence, and in combinations of these categories. However, unlike a stratified sample, the sampling of individuals is not carried out randomly, since the final selection of people is left to the interviewer.

Berg (2001) states that a ‘quota sample begins with a kind of matrix or table that creates cells or stratum. The quota sampling strategy then uses a non-probability method to fill these cells’.

Let us suppose an example here. A researcher wants to draw a quota sample from among the students of an Engineering college. There are, say, four departments and 500 students altogether. The researcher wants to draw a sample of 100. The calculations for quota to be specified for each category are given below.

Departments	Sex	Total No. of	Proportion in population	Size of the sub-sample*
		students		(Quota)
IT	M	70	70/500 = .14	100 x .14 = 14
	F	60	60/500 = .12	100 x. 12 = 12
Civil	M	85	85/500 = .17	100 x .17 = 17
	F	50	50/500 = .1	100 x .1 = 10
Electronics	M	50	50/500 = .1	100 x .1 = 10
	F	50	50/500 = .1	100 x .1 = 10
Mechanical	M	75	75/500 = .15	100 x .15 = 15
	F	60	60/500 = .12	100 x .12 = 12
Total		500  (Size of the population)	1	100 (Size of the sample)

Now the students (sample units) are not selected following random method, rather the researcher decides who will be his/her respondents. Thus, at the final stage of selection, the decision of the researcher is ultimate. When the population is heterogeneous, quota sampling is effective because the representativeness of the sample is ensured in this type of sampling. But, in practice, sociologists rarely employ quota sampling (Bryman 2012).

6.4 Accidental Sampling or Convenience Sampling

Accidental sampling depends on the availability of the respondents. In the words of Das (2004: 62), ‘accidental sampling refers to a method of selecting respondents who happen to meet the researcher and are willing to be interviewed.’ In this type of sampling the selection of the sample units neither follows the rule of probability, nor depends on the decision of the researcher. It is sometimes called convenience sampling as well. Davidson (2006: 196) writes, Convenience samples are also known as accidental or opportunity samples. The problem with all of these types of samples is that there is no evidence that they are representative of the populations to which the researchers wish to generalize.

For example, you are interested to conduct a study on passers-by who violate traffic rules at the zebra crossings. You can select the respondents instantly by observing who are violating the rules at a particular zebra crossing. You may collect data, if they allow at all, on the spot. This may help you unveil their psychological conditions at the time of the commission of the mistakes as well. Their contact detail can also be collected to talk with them later. Although this sample cannot be called a representative one, in the true sense, it is convenient and time saving. According to Berg (2001: 33), ‘under certain circumstances this strategy is an excellent means of obtaining preliminary information about some research question quickly and inexpensively’.

6.5 Theoretical Sampling

According to Bryman (2012: 420), ‘what distinguishes theoretical sampling from other sampling approaches is the emphasis on the selection of cases and units with reference to the quest for the generation of a theoretical understanding’. In theoretical sampling, those cases or persons are selected which are of particular interest to the study and which challenge or modify or support a theory thereby making the theory more definitive and useful. B. G. Glaser and A. L. Strauss first propounded the idea of theoretical sampling. Sample selection and data collection, in theoretical sampling, are done at many stages as the emerging theory demands. For example, suppose a researcher wants to study the problem of role conflict of the married working women and its associated coping strategies. S/he can select purposively a sample of say 200, from a city and collect some relevant data from the sample units. Now s/he can select a sub-sample of say 10 respondents from that sample in order to have an in-depth study. The researcher may again collect a sample from the population to support or substantiate the emerging theory, if necessary. So sampling (and data collection) here is an ongoing process. It continues till the point of what is called theoretical saturation. Theoretical saturation means the researcher continues sampling and data collection till he or she is sure that no new data will come up that can substantiate a particular category (a group of concepts having some commonalities are denoted by a category). At this point the category is saturated with data.

Self-Check Exercise – 2

1. What is purposive sampling?

Purposive sampling is a kind of non-probability sampling in which the researcher selects the units of the sample not by random method of selection but on the basis of his/her own choice. The researcher keeps in mind the objectives of the research and picks the suitable respondents or events up for the inclusion in his/her sample. Purposive sample is also called judgemental sampling.

2. What do you mean by snowball sampling?

Snowball sampling is used in those cases where the units of the population are unknown or difficult to locate. In snowball sampling, the researcher locates a few persons of the population and then seeks their help to find out other respondents.

3. Define quota sampling.

In quota sampling, the researcher, after defining the population from which the sample is to be drawn, classifies the population on the basis of certain characteristics which s/he thinks relevant to his/her study, like stratified random sampling. Then the researcher calculates the proportion of the population falling into each category and finally selects the sample units purposively.

4. What is convenience sample?

Convenience sampling depends on the availability of the respondents. In this type of sampling, the selection of the sample units neither follows the rule of probability nor depends on the decision of the researcher.

5. What is theoretical sampling?

In theoretical sampling, those cases or persons are selected which are of particular interest to the study and which challenge or modify or support a theory thereby making the theory more definitive and useful. Sample here is collected at different stages of research as per the demand of the emerging theory. B. G Glaser and A. L. Strauss first propounded the idea of theoretical sampling.

7. Summary

With the growing popularity of qualitative researches, the importance of non-probability sampling is increasing in social science research. In non-probability sampling the researchers generally do not bother much about the representativeness of their sample as they seldom try to generalize their findings numerically. Some social science researchers also oppose the quantitative-qualitative controversy. They want a flexible method combining the two ‘so that respective strengths might be reaped’ (Bryman 1988: 126). However, as in case of non-probability sampling, the decision of the researchers is very important in the selection of sample, they always try to suspend their personal choices and values as far as possible when they select the sample units. The sample size here is also not a matter of much concern, which is important in probability sampling. Non-probability sampling is flexible in its orientation, so the expertise of the researcher becomes crucial for its success to generate fresh insights and new knowledge.

Bibliography

• Atkinson, R. And Flint, J. “Sampling, Snowball: accessing Hidden and Hard-to-Reach Populations”, in The A-Z of Social Research edited by R. L. Miller and J. D. Brewer. London: sage Publications, 2003.
• Babbie,  E. The Practice of Social Research. Australia: Thomson Wadsworth, 2004.
• Berg, B.L. Qualitative Research Methods for the Social Science. Boston: Allyn and Bacon, 2001.
• Bloor, M. and Wood, F. Key words in Qualitative Methods. London: Sage Publications, 2006.
• Bryman, A.  Quantity and Quality in Social Research. London: Routledge, 1988.
• ……               Social Research Methods. Oxford: Oxford University Press, 2012.
• Choudhuri, Tanima. Women Workers in the Urban Informal Sector: The Case of Workers in Rice Mills and Nursing Homes in Burdwan. Ph. D Thesis. Department of Sociology. Vidyasagar University, 2014.
• Das, D. K. L. Practice of Social Research. Jaipur: Rawat Publications, 2004.
• Davidson, J. “Non-probability (Non-random) Sampling”, in The Sage Dictionary of Social Research Methods, edited by V. Jupp. London: Sage Publications, 2006.
• Goode, W and Hatt, P.K. Methods in Social Research. Auckland: McGraw-Hill Book Company, 1981.
• Henn, M.  et. al.  A Short Introduction to Social Research, London: Sage Publications, 2006.
• Majumdar, P. K.  Research Methods in Social Science. New Delhi: Viva Books Pvt. Ltd., 2005.
• Neuman, L. W. Basics of Social Research: Qualitative and Quantitative Approaches. Boston: Pearson Education Inc., 2007.
• Payne, G. and Payne, J. Key Concepts in Social Research. London: Sage Publications, 2005.