Sample and Sampling
A population is the entire collection of all observations of the interest, for the research. After the selection of subjects or problems for the study, it is very costly and time-consuming for the study of the universe (entire population). Thus, to make it easier to study, a representative portion of the population IS selected for the study that is known as a sample. For example: If a researcher wants to study the performance of Nepalese commercial banks, he/she will not consider all banks for the study but take few banks which cover the entire characteristics of all commercial banks. Those few banks are known as samples.
The process of selecting the sample (individual, group, etc.) based on the nature and necessity of research is known as sampling. The sample must represent the population so that the findings of the research can be generalized in the population.
If a researcher studies the entire individual, area, and group then such study is known as a census study If the study is conducted selecting few representative samples from the population, then such study is known as a sampling study.
If the sample selected by the researcher represents the entire population and findings of the study considering to sample can be generalized to the entire population then such sample is considered as a representative sample. Sampling has made it possible to undertake research in various subjects because it reduces costs and time.
The following definitions give insights into sampling.
According to Cooper and Schindler, “Sampling is some elements Of population which help to draw conclusions about the entire population.“
According to U. Sekaran, “Sampling is a process of selecting sub-set of the population by the study of which a researcher would be able to draw conclusions that would be generalizable to the populations.“
According to P.V. Young, “A statistical sample is a miniature picture or cross-section of the entire group or aggregates from which sample is taken, the entire group from which sample is chosen is known as the population or universe.“
Reasons for Selecting Sample
There are several reasons for sampling. They are explained below:
1. Lowers cost:
Researchers should spend a large amount on census studies. If the researcher conducts the research by taking a sample, then the cost for collecting and analyzing data will be less.
2. Provides greater accuracy:
Sampling possesses the possibility of better interviewing through investigation of missing, wrong, or suspicious information, better supervision, and better processing than census study. Thus, there is likely to be greater accuracy.
3. Helps to greater speed of data collection:
There is less number of respondents and researcher or data collector. They can clearly give instructions to the respondents so that they can provide data quickly and the researcher also gets the data with a greater speed.
4. Inaccessible population:
Even though the population is well defined and countable, yet it is not possible to get information about every member. So, sampling is essential.
Factors affecting the Size of the Sample
The size of a sample depends upon a number of factors. The main factors that affect the selection of sample are given below:
1. Homogeneity/heterogeneity of universe:
When the population is comparatively homogeneous and the characteristics of the population are tentatively similar, smaller size of sample will be sufficient. If the character of the population is exactly the same then a single unit can be the sample. When the characteristics Of the population are varied in a wide range then the sample size should be large enough.
2. Number of classes proposed:
If we need to form a large number of classes in the population then obviously the sample size will be large because we need to take sample from all classes. If the size of the population is less then the researcher needs to formless classes so the sample size also will be smaller.
3. Nature of Study:
If an intensive study is to be made continuing for a long time, a large sample is unfit for the purpose because it requires large financial and other resources. Similarly, in the case of general study large number of cases can be taken but if the study is of a technical nature, a large number may become difficult to manage. So, small sample size may also be sufficient.
4. Practical consideration:
The availability of resources and time for undertaking research is also an important factor that determines the size of the sample. The limitations of the resources limit the size of the sample. Although these practical considerations affect significantly determining the size of the sample it should never be done at the cost of accuracy.
5. Standard of accuracy:
Generally, it is believed that a large sample gives better accuracy than a small size of the sample. A small but well-selected sample may give better results than a large and thoughtlessly selected sample. But the standard of accuracy or requirement of accuracy determines the sample size. If the required accuracy is high then a large sample size is considered better.
6. Nature of cases to be contacted:
If the cases are geographically scattered, a large sample is more suitable. On the other hand, if the cases related to the study are high then the size of the sample should be large. If the chances of refusal cases are high then the size of the sample should be large.
7. Type of sampling used:
If absolute random sampling is used, a larger sample is required. The random sample is reliable only when a large unit of samples are selected. On the other hand, if stratified sampling is used reliability can be achieved in the much smaller sample if stratification is proper.
Key Components:-
Sample and Sampling in research | Reasons for Selecting Sample | Factors Affecting the Size of the Sample | The Sampling process | Types of Sampling
The Sampling Process
Sampling is an important function of research. Right sampling helps to draw the right conclusions and such conclusions can only be applied in practice. Thus, a sample should not be selected in hunches but should be selected following a certain process. Generally, the following procedures are pursued while selecting a sample:
1. Define the population:
The population is the collection of whole units (people, object, etc.) that researchers are interested in knowing about them. Definition of the population depends on the subject and nature of research and availability of resources and time for research. Findings of the research should be implemented in the population; therefore, it is to be defined clearly and precisely. The researcher can select a sample after defining the population. The population should be defined in terms of elements, sampling units, and time. Defining a population incorrectly may render the results of the study meaningless or even misleading. If we are dealing with people, the population of individuals is typically defined in some combination of demographic (age, gender, income, etc.) geographic (Towns, village, etc), and behavioral ( introvert, extrovert, etc.) components.
2. Specify the sampling frame:
A sampling frame is the list of elements from which the sample is drawn. After defining the population, a researcher should get the full, accurate, and up-date list of all units of the population. A sampling frame can be a telephone directory, an employee roaster, voter list, or a list of all students attending a college. Thus, a perfect sampling frame is one in which every element of the population is represented.
3. Specify sampling unit:
A decision should be taken by the researcher concerning a sampling unit before selecting a sample. The sampling unit is the unit that represents every character of the population. The sampling is based on the sampling frame but the determination of the sampling unit depends on the subject and nature of research and research design. The sampling unit may be a geographical one such as state, district, village, etc., or a social unit such as family, club, school, etc, or an individual.
4. Determination of Sample Size (n):
Sample size refers to the number of items which are to be selected as sample from the population. Populations have qualitative and quantitative elements. Sample size should be determined in such a way so that it can capture all elements of the population and able to attain the goal of the research. The size of the sample should neither be excessively large nor too small but it should be optimal. An optimum size is that which. fulfills the requirements of efficiency, representativeness, reliability, and flexibility. Sample size can be determined in the following ways:
a. Using a sample size of a similar study:
A researcher can use the same sample size as used by the other researchers in similar studies. In this method, the researcher should review the literature for sample size and the size used by most of the researchers can be taken for the study.
b. Using published tables:
A number of tables are published by the expert after a long study under various conditions. The tables present the sample size that would be necessary for a given condition like precisions, confidence interval, and variability. The researcher can determine sample size under his/her condition.
c. Using statistical formulae:
A researcher can use various techniques (Formulae) of statistics to determine sample size. Different formulae are suggested in different conditions.
5. Preparation of plan for sampling:
The researcher should formulate a plan to make the work of sampling appropriate and well managed. The sampling plan provides guidelines to operationalize the sampling design and size. A sampling plan determines the decisions which are to be taken while selecting a sample and its use.
6. Select the sample:
It is the final step of sampling work. Selection Of sample requires a substantial amount of office and fieldwork, Selected sample should represent the population and use to attain the goal of research. Samples are to be selected based on either method of sampling like random, convenience, etc.
Types of Sampling
Proper sample selection is an important work in research, Appropriateness of sampling depends on the nature, goal, subject, and availability of resources and time. Thus, various sampling techniques are developed which are described below:
1. Probability Sampling
A sampling technique where every element in the population has an equal chance of being selected as a sample unit is known as probability sampling Selection of element depends on incident, In this method, the researcher also cannot estimate which element will be selected and can use his/her opinion in sampling, It is used when there is neceøsity Of generalizing findings of the research in a large population. There are various techniques for selecting samples based on probability. Some of the important probability sampling techniques are described below:
a. Simple random sampling:
sampling where every element in the population has an equal chance of being selected as a sample is known as simple random sampling. Under this technique, required samples are selected using the lottery method, number order method, random number test method, etc. but nowadays computerized lottery is used to select the sample. Simple random sampling is used when a sampling frame can be developed and researchers need to generalize the findings of the research in the population. For example, if a researcher wants to know the satisfaction level of banking employees in Nepal then he/she considers the employees of the banking sector as population and selects the number of samples using a computerized lottery system. Such sampling is considered as simple random sampling.
b. Systematic sampling:
Systematic sampling involves the random selection of the first item from the systematically ordered population and then the selection of sample items at every Kth interval. To select the sample items in systematic sampling, we need to calculate the sampling interval. The sampling interval is calculated as:
Sampling interval (K)
= Size of the population (N)/Size of the sample (n)
This is the simplest and most widely used method of drawing a sample. The interval (K) is fixed by dividing the population by sample size.
While applying this method, a researcher should take a random number between 1 to N which determines the first number for the sample. And adding or deducting to the internal value other sample items are selected. For example, if a researcher has framed 400 employees as population and intends to take 10% i.e. 40 persons as a sample then first he/she will select one item that is from 1 – 400. Suppose the first number selected is 5 then K-value is 10 (400/40 = 10). So, the sampling units are 5 + 10 = 15, 15 + 10 = 25, 25 + 10 = 35 and so on. Following procedures should be followed while applying this method.
- List the total number of units in the population.
- Decide the sample size (n).
- Calculate the sampling interval. Sampling interval =N/n
- Identify the random start of the first number selected from the population.
- Draw a sample by using a sampling interval.
c. Stratified sampling:
Strata refer to the overlapping homogenous groups in the population. Every group should be incorporated into the sample to represent the population. So, a sampling method which represents the samples in proportionate rate from the different group of population is known as stratified samplil$ If we want to represent a different section of the population in the study such as male and female, educated and uneducated or employed and unemployed, this method of sampling is suitable. In this sampling, the population is divided into sub-groups or strata and a proportionate sample is selected from each group or strata. It is used in research because it helps to increase the sample’s statistical efficiency, provide adequate data for analyzing the various subgroups, and enables the use of different research method in different groups.
It can be proportionate or disproportionate. Proportionate stratified sampling refers to the act of selecting samples from each group in the same proportion. If the selection of sample is made in less and more number when population of the sub-group is highly different such sampling technique is disproportionate sampling.
Stratified sampling can be made clear from the following example: If a manager wants to know the motivation of employees, he/she may select a sample as follows:
The processes that are to be followed while applying stratified sampling are given below:
- Determine the variables to use for stratification.
- Determine the proportions of the stratification variables in the population
- Select proportionate or disproportionate stratification based on information needs and risk.
- Divide the sampling frame into separate frames for each group
- Follow random or systematic procedures to draw the sample from each group.
d. Cluster sampling:
A cluster is a heterogeneous group that is in the population. Cluster sampling identifies clusters that are internally heterogeneous. Every cluster contains many elements into a single element so, it is considered a small population. Employees grouped in the branch office, customers at each supermarket branch are examples of clusters. Sampling, where a group is selected as a sample having all elements of the population, is known as cluster sampling. Generally, the random sampling method is used while selecting clusters as samples. Detail study of the selected clusters is essential to find the correct results. Clusters are heterogeneous within themselves and not like the homogeneous strata. Hence, the collection of data would be far easy as compared to other methods. It is suitable in the absence of a suitable sampling frame. Frames are needed for the selected cluster only and it reduces the cost of developing the frame as compared to another probability sampling. For example: If a manager wants to know the reason for the resignation of employees, he/she first develops the group on the basis of work level (position). He also prepares the list of employees of these groups. On the basis of random sampling, the cluster is selected as a sample. With the detailed study of such cluster, the researcher finds out the reasons for resignation.
2. Non-Probability Sampling
Sampling where there is no equal chance of selecting a sample to each unit and sampling is made based on pre-plan is known as nonprobability sampling. The findings of such sampling cannot be generalized because samples are selected with a specific purpose or separating the area in advance. This sampling is considered appropriate if the researcher needs to collect data with low cost and time and generalization of findings is not essential. There is a chance Of biasness in selecting a sample while using this method for sampling.
Some of the important non-probability sampling methods are given below:
a. Purposive or judgmental sampling:
A sampling method where samples are selected by the researcher based on his or her judgment is known as judgmental sampling. Those units or individuals are selected as a sample that can fulfill the purpose but it does not consider the convenience of the researcher. The researcher sets the bases and those units are selected as samples which can ensure those bases. The researcher should know every unit of the population and their features for applying this sampling method. Otherwise, a researcher cannot collect essential information and data. Thus, this sampling is generally used by experts.
If a researcher needs to get specific and specialized information this method is considered an appropriate method but it is a very difficult task that to find out the persons who have knowledge about the subject of research. For example, if a researcher wants to test the effectiveness of training, he/she selects as a sample to those employees who attend training and collects the information from those employees. This sampling is considered judgmental sampling.
b. Quota sampling:
A sampling method where the population is divided into different groups based on their nature, features, qualities, etc., and the sample is selected from each group at a certain rate is known as quota sampling. In this sampling, first of all, groups are formed on the basis of profession, level, caste, area, etc. and samples are selected based on the size of the population like some sample from more population and less sample from less population based on a certain rate. It is non-probability sampling so its findings cannot be generalized.
It is used widely because every society and work field has heterogeneous groups and for the study of those groups, this sampling method is used. To know the buying behavior of different groups of people, to know the attitude of the group of employees regarding the culture of their organization, this sampling method is considered an appropriate method. The following steps should follow to apply this sampling method:
- Classify the population into different classes based on demographic factors i.e. age, sex, income level, etc.
- Determine the number of samples to be selected from each class.
- Select the sample based on a pre-determined number from each class.
c. Convenience sampling:
The researcher selects the units as samples on the basis of his/her convenience is considered convenience sampling. The researcher selects those units that are available, nearby, and willing to participate or has a relationship as a sample. It is also known as accidental sampling because samples can be selected from anywhere else. Generally, this method is used when there is a high limitation of time and resources. For example: If any person wants to have research on the facilities provided to banking employees, then the researcher selects those banks as a sample where his/her relative works or nearby banks then it is convenience sampling. It is not a totally valid method but it is considered as the best method for pre-testing of questionnaires and descriptive research. A researcher can collect data quickly and at a low cost and time using the convenience sampling method.
d. Self-selecting sampling:
If the researcher gives information through media to the respondents and respondents provide information on the basis of information received through media then such sampling is known as a self-selecting sample. Those who provide information on the basis of information provided through media are considered as samples. The validity and reliability of this method is less. Generally, it is used to know the goodwill and evaluate the service provided by the organization.
e. Snowball sampling:
It is also known as reference sampling. If the population is infinite or not fixed, then the researcher selects one or few samples whose profile is fit to get the information and on the basis of recommendations of the previously selected samples, other samples are selected. Such sampling method is known as snowball sampling. If it is difficult to find out more samples then a few people or a single person is identified and other persons are identified based on the reference of previously selected persons. It is considered a snowball because in this sampling small sample unit forms a large sample as the small snowball forms a large ball. It is usually used by police to find out criminals, and It is also used to study the group activities, culture, and relationship of society, etc. The main problem of this sampling is to maintain relationships with the first person or sample unit. The following steps should be pursued to use this method:
- Finding out the first person from the population who can give information about the subject under study.
- Ask to refer to other persons who can give information about the subject under study.
- If a further sample is not found or sample formed a large number of person then close the sampling work.
Descriptions of Various types of samples are summarized below:
| Types of Sampling | Brief Description |
|---|---|
| A. Probability Sampling | – |
| a. Simple random sampling | All elements in the population are considered and each element has an equal chance of being chosen as the subject. |
| b. Systematic sampling | Every nth element in the population is chosen stating from a random point in the population frame. |
| c. Stratified sampling | Population is first divided into meaningful segment and samples are selected in proportionate or disproportionate way from each meaningful segment. |
| d. Cluster sampling | Groups that have heterogeneous members are first identified then some are chosen at random. Members in each of the randomly chosen groups are studied. |
| B. Non-Probability Sampling | – |
| a. Purposive or judgmental sampling | Subjects selected on the basis of their expertise in the subject investigated. |
| b. Quota sampling | Subjects are conveniently chosen from target groups accordingly to some predetermined number or quota. |
| c. Convenience sampling | The most easily accessible members are chosen as samples. |
| d. Self-selecting sampling | Selecting the sample based on the response provided spontaneously.. |
| e. Snowball sampling | Selecting few or single sample at first and obtaining other sample units on the basis of their reference. |
