23 Probable Sampling Technique II

R. Saratha

epgp books

 

 

 

Introduction:

 

There are few other techniques of sampling, besides those that were discussed in the previous chapter. These techniques are classified under Probable techniques II. These are significant methods used in research studies. This chapter describes the nature of Stratified random sampling including their advantages and limitations, cluster sampling and multi-stage sampling and the applications of these techniques.

2. Learning Objectives:

 

At the end of the session you will be able to:

 

Ø  Know in detail about stratified random sampling, its advantages and disadvantages.

Ø  Understand what is cluster sampling, its principles and merits & demerits.

Ø   Comprehend what is multistage sampling, its advantages and limitations.

Ø   Be aware of the various applications of the techniques in research studies.

 

3.STRATIFIED RANDOM SAMPLING

 

Stratification means dividing the units of the population into groups according to geographical, sociological or economic characteristics. Stratified random sampling is best suitable for a population which is heterogeneous in nature. Using this technique will facilitate more efficient and accurate results.

 

A stratified random sample is obtained by separating the population elements into non-overlapping group (strata), and then selecting a simple random sample from each stratum.

 

Let us take the case used in the simple random sampling narration and see how the same sampling can be done using the stratified random sampling procedure. Let us assume that the researcher uses simple random sampling procedure to select the 40 students out of the 298 students of the university which has UG course with 2 sections in each year and PG course with 1 section in each year. The tabulation below describes the class-wise distribution which was used in the simple random sampling.

Let us also assume that the students selected from each class using the simple random procedures may be as follows. It is theoretically possible that even with the use of random numbers; one class may find more students participating in the study than from other classes. The following may be the result of the simple random sampling in which the students of the UG I year Section 2 are more in number and the PG II year class has the least representation. One may also argue that the percentage of students participated in the study was as high as 25% in the case of UG II Year and PG I year classes whereas the lowest representation was from UG III Year which constituted only 7.5% of the sample.

The researcher may like to see that every class has more or less equal distribution in the sample and therefore, divide the sample into eight strata as below:

Though time consuming, the above procedure ensures that adequate numbers of students from each class participate in the study. The researcher then considers each of the 8 strata as a sub-population and use simple random technique to select the required number for the study. The strata 1 (UG I year – Section 1) has 45 students and the researcher may use the table of random numbers and select 5 students. As the sub-population in this case is manageable, the researcher may use a lot system too to select the required number of students.

 

In summary, dividing the population into various strata and applying simple random technique to select the sample to have representation in every stratum of the study is called Stratified random sampling. Again the researcher may select the sample within each stratum on the basis of percentage too. Take the following for example:

In the above example, the researcher satisfies the selection of sample according to strata and also determining number as per percentage representation too. Therefore, there are many ways of improving the sample selection procedures to ensure that the results obtained from the sample are easily generalized to the entire population.

 

Stratified random sampling is conducive for better statistical analysis as the researcher can use various groups and sub-groups to find out the differences between them using the data obtained. Many research studies involving more than one independent variable and desirous of using parametric statistics generally apply stratified random sampling provided they satisfy the norms for applying such statistical techniques.

 

3.1 ADVANTAGES OF STRATIFIED RANDOM SAMPLING:

  • Even a small number of units turn into a representative sample if the correct stratification is applied.
  • Every significant group is represented.
  • There is very less chances of bias.
  • Stratified random sampling saves time and cost as even a small sample size will produce better results.
  • Different degrees of accuracies can be obtained for different segments of population.

3.2 LIMITATIONS:

  • It is difficult to divide the population into homogeneous strata.
  • The selection of samples will not be representative if the strata are over-lapping or disproportionate.
  • There will be biased results if the stratification is faulty.
  • Errors due to bias cannot be compensated even if large samples are taken.

4.   CLUSTER SAMPLING

 

Often times, researchers like to know the opinions of specific groups and in such cases cluster sampling becomes handy. Industrial houses use cluster sampling to find out the potential market for their products. For example, say a television company wants to find out how many people generally use a particular brand of the Company’s TV to set up retail outlets. The retail outlets are established only when the company is sure of making profits through the sale of its products. The company may also like to know the number of people in each economic stratum who use that particular TV.

 

In a district or a city, the company may identify a representative cluster which includes more people belonging to the high income group, a cluster where more middle income group people live and another cluster where average income group individuals live average. Once these representative clusters are identified based on baseline information and after discussions with people about the representativeness of clusters, an in-depth study is undertaken to cover almost the entire cluster. Trained representatives who have the skill of eliciting maximum information usually meet the individuals of that cluster and gather data about the usage of the particular TV. In such meetings, other preferred items from that particular group are also listed which will help the company for placing the right product in each of the retail outlets. Market researches usually apply the cluster sampling and try to cover every individual in that cluster to make the results of the study meaningful. As this is time consuming and laborious, selection of the representative clusters becomes vital in this research. Collecting data from a non-representative cluster of the defined population domain will be in vain and therefore, utmost care is necessary in the pre-preparation stage of cluster sampling.

 

4.1 PRINCIPLES:

  • The clusters should be as small as possible so as to avoid the high cost and limitations of the survey.
  • The number of units selected from each cluster should be approximately the same.

   Many localized issues are also analyzed through cluster sampling. Problems faced by people of a particular region because of noise pollution, industrial wastages, rights issues, etc., are often studied well using cluster sampling. Surveys aimed at specific issues relevant for focused groups such as women, persons with disabilities, different professional categories, etc., may also use cluster sampling. Irrespective the sampling technique applied, whether it is simple random sampling or stratified random sampling or even cluster sampling, the researcher should try to ensure that the results obtained are generalizable to the entire population as it is mostly the centre aim of research. At times researchers pay too much attention on sampling technique only, which is of course vital, and may not control the impact of extraneous variables which may result in non-sampling errors. Therefore, selection of the sampling technique should go hand-in-hand with the effective control over the variables as well as the research design to get the most authentic results from the research.

 

4.2 ADVANTAGES:

  • Cluster sampling is a more practical and easier method of sampling.
  • It facilitates a significant cost gain.

4.3 LIMITATIONS:

  • Probability and representativeness of the sample will be affected if the clusters are too large.
  • If the selected units are not approximately same from all the clusters, the results will be biased.

5. MULTI-STAGE SAMPLING

 

Multi stage sampling technique is used when the population of the study is very large. It refers to sampling carried out in various stages.

 

There are instances when the researcher feels that a single sampling technique may not be adequate or suitable to gather data and may realise the compulsion to use different sampling techniques in the same research.

 

Take the example used in describing the stratified random sampling. After getting the required data, the researcher finds during the analysis that there is a particular trend emerging from specific sub groups. In case the mean score of UG students is higher than that of the PG students, the researcher may try to make an in-depth study of the PG group to find out why their mean score is lower. The researcher may then want to study the entire cluster and gather both qualitative as well as quantitative data for further analysis of the research results.

 

On such occasions, the researcher uses different types of sampling techniques at different stages of the same research and such a combination of sampling techniques is called multi-stage sampling. Some researchers use emergent design wherein the purpose is to explore the key issues that may have impact on the criterion and for this purpose the researcher may be using a focus group using cluster sampling techniques. At some stage, the researcher may not be interested in the endure group to verify a hypothesis and may use a representative sample from a particular strata of the cluster. Therefore, the researcher may be using cluster, stratified random sampling and simple random sampling, all in the same research. Historical research studies and more exploratory research studies which are emergent in nature generally use multi-stage sampling.

 

Another example: If the government wants to take a sample of 10,000 households residing in TamilNadu state. At the first stage, the state can be divided into the number districts, and then few districts can be selected randomly. At the second-stage, the chosen districts can be further sub-divided into the number of villages and then the sample of few villages can be taken at random. Now at the third-stage, the desired number of households can be selected from the villages chosen at the second stage. Thus, at each stage the size of the sample has become smaller and the research study has become more precise.

5.1 ADVANTAGES:

  • It is a very flexible method of sampling and it is simple to carry out.
  • It ensures administrative convenience by allowing the study to be concentrated yet cover a large area.
  • Multi- stage sampling is of greater significance in underdeveloped areas where the division of the population into sub-units is not possible.
  • This method is reliable and produces satisfactory results.
  • It can be accomplished with a considerable speed.

5.2 LIMITATIONS:

  • In certain cases this method is less efficient when compared to single stage sampling.
  • Errors are likely to be large in this method.
  • It involves considerable amount of listing like the first stage, second stage etc. which is time consuming

6. How does one decide which type of sampling to use?

 

Simple random sampling is the method which is used in most of the research. Whereas more complex techniques of sampling are required to be handled, so as to build an effective research resource for any institution, or even the society at large.

 

Stratified random sampling gives more precise information than simple random sampling for a given sample size. So, if information on all members of the population is available that divides them into strata that seem relevant, stratified sampling will usually be used.

 

If the population is large and enough resources are available, usually one will use multi-stage sampling. In such situations, usually stratified sampling will be done at some stages. Cluster sampling is extensively used in marketing researches.

 

7.  How do we analyze the results differently depending on the different type of sampling?

 

The main difference is in the computation of the estimates of the variance (or standard deviation). A very simple statement of the conclusion is that the variance of the estimator is smaller if it came from a stratified random sample than from simple random sample of the same size. Since small variance means more precise information from the sample, we see that this is consistent with stratified random sampling giving better estimators for a given sample size.

 

8.  APPLICATIONS OF THE PROBABLE SAMPLING TECHNIQUES: Historical Study:

 

Historical study looks at the present characteristics of events in the light of what happened in the past in order to make a projection for the future. In this research, the design itself demands different types of sampling techniques and the researcher may use the simple random sampling, stratified random sampling, cluster sampling, and multi-stage sampling in a judicious way to get the most accurate facts.

 

Ethnographic Study:

  • This is a type of research normally adopted by anthropologists
  • The researcher does not form specific hypothesis and by nature, the research design is emergent in nature. In an emergent design, the researcher may be compelled to use multi-stage sampling using the required sampling techniques.
  • Researcher documents every minute information that comes in the course of the study
  • Observation is the primary source of data collection
  • Qualitative methods of documentation used

As is evident from the previous sections of this chapter every sampling technique has its own merits and limitations and selection of the appropriate technique depends on the nature of the study. The techniques are only tools in a research and the most important entity in research is the researcher himself/herself. Following are usually the sources of a research problem:

  • Review of related literature
  • Discussion with professionals familiar with the subject
  • Discussion with the peer group researchers
  • Growing trends in the field
  • Intuition of the researcher
  • Replication studies
  • Follow-up of previous research on the subject

 

The insights of the researcher are more important than anything else to make the research successful. The insights coupled with a sound knowledge in research applications will enrich the research and guide the research in the right perspectives.

 

9.   Summary

 

In this chapter, we have discussed the nature of Stratified random sampling, cluster sampling and multi-stage sampling, their applications in research studies, advantages and limitations. In the section dealing with the stratified random sampling, we used different examples to explain how the researcher can ensure selection of sample from every stratum of the population and also ensure the selection of the right variables to make the results of the study generalizable to the population. To put it in a nutshell, dividing the population into various strata and applying simple random technique to select the sample to have representation in every stratum of the study is called Stratified random sampling. With this knowledge you can review the research designs of the existing research studies and see how the researcher has attempted to control the influence of extraneous variables in the selection of sample using stratified random sampling.

 

Market researches usually apply the cluster sampling and try to cover every individual in that cluster to make the results of the study meaningful. As this is time consuming and laborious, selection of the representative clusters becomes vital in this research. Though multistage sampling is a flexible method of sampling, the errors are likely to be large in this approach. This lesson clearly explained how the time, resources, urgency, focus, etc., are determining the selection of a particular sampling technique.

you can view video on Probable Sampling Technique II

Suggested References

  • C.R. Kothari (2004), Research Methodology, methods & techniques second edition, revised. New Delhi, India: New Age Publishing Company, P55-67
  • Ranjit Kumar (2011), Research Methodology a step-by-step guide for beginners, third edition, New Delhi, India, Sage Publications , P 175- 189
  • John W. Creswell & Vicki. L .Plano Clark (2006), Designing and conducting Mixed Methods Research, second edition, California, Sage Publications, P 195 & 196
  • Yin, R.K. (2016). Qualitative Research from Start to Finish, Second Edition. New York: The Guilford Press.
  • Santhosh Gupta(2001) Research Methodology and Statistical technique, , New Delhi, India , Deep& Deep publications ISBN 81-7100-501-2
  • G.R. Basotia &K.K. Sharma (2002) Research Methodology, Jaipur, India , Mangal Deep Publications, ISBN: 81-7594-090-5 P.Saravanavel (2007) Research Methodology, Allahabad, India, Kitab Mahal Publications, ISBN: 81-2225-0010-2
  • R.Panneerselvam(2004), New Delhi, India, Phi Learning Private Limited, ISBN: 978-81-203-2452-7
  • Welter R. Borg & Meredith D. Gall, Educational Research- An Introduction, fourth edition, New York & London, Longman Publications, ISBN: 0-582-28246-2