UNIT 10 SAMPLING TECHNIQUES

Structure

10.0 Objectives
10.1 Introduction
10.2 Importance of Sampling
10.3 Types of Sampling Techniques
10.4 Probability based Sampling Techniques
10.4.1 Simple Random Sampling
10.4.2 Systematic Sampling
10.4.3 Stratified Sampling
10.4.4 Cluster Sampling
10.4.5 Multistage Sampling
10.5 Non-Probability based Sampling Techniques
10.5.1 Convenience Sampling
10.5.2 Purposive Sampling
10.5.3 Snowball Sampling
10.5.4 Self-Selection Sampling
10.6 Sample Size Determination
10.6.1 Determining Sample Size for Questions Involving Means
10.6.2 Determining Sample Size for Proportions
10.7 Sampling and Non-Sampling Errors
10.7.1 Random Sampling Errors
10.7.2 Non-Sampling Errors
10.7.3 Control of Errors
10.8 Let Us Sum Up
10.9 Key Words
10.10 References/Suggested Readings
10.11 Answers to Check Your Progress Exercises
10.12 Unit End Questions

1
10.0 OBJECTIVES
After studying the unit the student should be able to:
 explain the concept and importance of sampling;
 distinguish between probability-based and non-probability based sampling
techniques;
 decide on the use of appropriate sampling techniques for research;
 determine suitable sample size for study; and
 identify different types of sampling errors.

10.1 INTRODUCTION
To begin with this Unit, it is important to understand a few terms that we would be
using quite often during our discussion. These terms are:
i. Population
ii. Census
iii. Sampling
iv. Sampling element
A population is the aggregate of all the elements that share some common set of
characteristics and that comprise the universe for the purpose of the research problem.
Any research study aims to obtain information about the characteristics or parameters
of a population.
The collection of data from the entire population is referred to as a census study. A
census involves a complete listing of all the elements of a population. Collecting data
from the aggregate of all the elements (population) in case, the number of elements is
larger would sometimes render the researcher to incur huge costs and time.
A sample is a subset of a population and the process of selecting units (e.g., people,
organizations) from a population is referred to as sampling. In most cases, it is difficult
in terms of the high cost and time involved in collecting the information from the entire
population. With sampling, the researchers are able to collect the information from a
portion of the population, which otherwise was not possible due to the large size of the
population and other constraints like time and cost. Based on the sample findings, the
parameters for the population are inferred.
The results of the sample studies may be generalized for the population however care is
to be taken so that sampling errors are kept to the minimum. The concept of sampling
and population is shown in Fig 4.1.

2
 Fig. 4.1: Population, Sample, and sample elements

10.2 IMPORTANCE OF SAMPLING

It is possible to collect data from an entire population for a few research studies when it
is of a manageable size. However, if the population is too large then reaching all the
elements of the population is not feasible or sometimes not economical. In such
situations, we draw meaningful results by collecting data from a sample that represents
the entire population.
It is suggested to go for sampling as a valid alternative to a population in the following
situations:
i. Practicality to survey the entire population: It is impracticable for us to collect
data from the entire population in all the research studies. Hence, there would
be a need to select a sample. We might be able to obtain permission to collect
data from only two or three organizations. For example, we want to study the
satisfaction of the employees in all the fertilizer plants in India. From secondary
sources, we come to know that there are 16 fertilizer manufacturing plants in
India. However, when we proceed with our study, only six out of the identified
16 organizations may permit conducting the study.
ii. Budget constraints: Some research studies get constrained in reaching the
entire population due to a lack of budget for collecting the data. It is obviously
cheaper for you to collect data from 50 customers than from 500.
iii. Time constraints: if you want to complete your study on a tight schedule
then sampling saves time. As you have fewer data to enter, the results will be
available more quickly. Occasionally, to save time, questionnaires are used to
collect data from the entire population but only a sample of the data collected is
analyzed. Collecting data from a small sample size also means that you can
collect more detailed information. In addition, if you are employing people to

3
 collect the data, you can afford higher-quality staff. You also can devote more
time to obtaining data from difficult-to-reach cases.

Check Your Progress Exercise 1

Note: a) Use the space below for your answers.

b) Compare your answers with those given at the end of the unit.

1) Sampling is important in:

a) Data collection

b) Identifying, developing, and understanding new constructs

c) Developing scale measurements

d) All of the above

2) Sampling refers to the _____.

a) selection of a small number of elements from a larger defined target group

of elements

b) inclusion of every member of the defined target population into the research
study

c) an identifiable total set of elements of interest being investigated by a

researcher

d) complete set of elements identified for investigation

e) a person or object from the defined target population from which data is
sought

3) Sampling is often preferred to surveying a population because

a) Surveying a population can be very expensive

b) It can be physically impossible to find all the items in the population

c) Sample results can be very accurate

d) All of the above are reasons for sampling

4) Which of the following is a reason for sampling?

a) The destructive nature of certain tests

b) The physical impossibility of checking all the items in the population

c) The adequacy of sample results

4
 d) All of the above are reasons for sampling

10.3 TYPES OF SAMPLING TECHNIQUES

Sampling techniques are classified into two broad categories viz. probability samples
and non-probability samples.
In probability sampling, every element in the population has a known chance of
selection. The best-known probability sampling technique is the simple random sample,
in which each member of the population has an equal chance of being selected. In non-
probability sampling, the probability of selection from the population is unknown. The
non-probability sampling deploys an arbitrary and judgmental sample selection
approach.

Fig. 4.2: Types of Sampling Techniques

Source: Saunders, M., Lewis, P. and Thornhill, A. (2003), Research Methods for
Business Students, Third Edition, Pearson Education Ltd.
The following sections discuss the advantages and disadvantages of different sampling
techniques listed in the above figure.

10.4 PROBABILITY BASED SAMPLING TECHNIQUES

If you want to generalize on the basis of the sample to the population at large,
probability sampling is necessary. Probability or random sampling increases the
likelihood that the information collected is representative of the population. Probability
sampling generally is preferred over non-probability sampling because the risk of
incorrectly generalizing the population is known.

5
10.4.1 Simple Random Sampling
Simple random sampling is, in a sense, the prototype of population sampling. Suppose
that a teacher wants to select a sample of 25 students from a class size of 100 students.
The procedure would be very simple. The teacher will randomly select 25 students for
the study. Here every student would be having an equal chance of being selected during
the selection procedure. The scientific basis of drawing a sample based on random
numbers may also be used.
To summarise, a random sample is defined as follows:
 Selections are made from a specified and defined population (i.e., the sampling
frame is known).
 Each unit is selected with a known probability so that every unit in the
population has an equal chance of being selected.

10.4.2 Systematic Sampling

In most cases, we use systematic random sampling, which guarantees that units cannot
be sampled more than once. That is this technique avoids replacement. The procedure
includes selecting the sampling units based on a series beginning from a random
starting point. Consider a situation where the researcher intends to choose 10 students
from a class of 100. In order to choose these 10 students, the teacher selects the first
student at random from the first 10 roll nos. and then selects the other students with a
series of N+10 (where N is the roll no. of the first student selected). It means that the
students with the roll nos. 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100 will be selected, if
roll no. 10 is selected as the first student.

10.4.3 Stratified Sampling

Stratified sampling is a two-stage process. In this technique, the population is first
divided into sampling groups (strata) on some common characteristics, which are
mutually exclusive and exhaustive subsets. For example, the teacher may divide the
students into two groups based on their gender and then draw a sample of 5 students
from each group using a simple random or systematic random sampling technique.
The stratified sampling technique ensures the representation of all important sub-
populations in the sample and combines the use of simple random sampling with
potential gains in precision as the objective of this technique is to get homogenous
elements within each stratum.

10.4.4 Cluster Sampling

When a list of the sampling unit for the entire population is nonexistent or hard to
obtain, cluster sampling can be used for data collection. Cluster sampling is a method
of selecting sampling units in which the unit contains a cluster of elements. Ideally,
individual clusters in cluster samples should be as heterogeneous as possible. This is
the opposite of stratified samples where each stratum is homogenous. Cluster sampling

6
is used for a variety of purposes, particularly for large samples. For example, you could
group your data by type of manufacturing firm or geographical area.
This technique can be detailed in the following steps:
i. Determine and select the cluster grouping for your sampling frame.
ii. Number each of the clusters with a unique number. The first cluster is
numbered 0, the second 1, and so on.
iii. Select your sample using some form of random sampling as discussed earlier.
iv. Selecting clusters randomly makes cluster sampling a probability sampling
technique.
The teacher may form four clusters in the class based on the residential area of a city as
students residing in the north, south, west, and eastern part of the city and then select 5
students randomly from each cluster for the study. This is called a two-stage clustering
technique.

10.4.5 Multistage Sampling

Using a combination of two or more sampling techniques discussed above is referred to
as multistage area sampling. For example, a marketing manager for investigating taste
preferences for ready-to-eat food in the southern part of India might first choose states
within the southern part to ensure that the different areas are represented in the sample.
In the second step, cities within the selected states may be chosen. As a final step, the
manager may select blocks (or households) within the cities and then interview all the
blocks (or households) within the geographic area.

10.5 NON-PROBABILITY BASED SAMPLING TECHNIQUES

In non-probability sampling, the probability of selecting population elements is
unknown. But in a situation when a sampling frame is absent, one can easily go for
non-probability sampling methods to serve the objectives of the study. However, a
question may arise as to how closely these approximate representativeness. Additional
reasons for choosing non-probability over probability sampling are cost and time
factors.
The following techniques are discussed:

10.5.1 Convenience Sampling

Convenience samples are sometimes called accidental samples because the elements
included in the sample enter by accident. In other words, researchers based on their
convenience can decide and select the sampling units for their surveys or interviews.
A few examples of convenience sampling include (a) respondents who gather in a
temple (b) students in a classroom (c) mall intercept interviews without qualifying the
respondents for the study (d) tear-out questionnaires included in magazines and (e)
people in exhibitions or trade shows.

7
10.5.2 Purposive Sampling
A non-probability sample that conforms to certain criteria is called purposive sampling.
There are two major types within this type of sampling:
 Judgement sampling and
 Quota sampling
i. Judgement Sampling: The sample elements are chosen based on the judgement
that prevails in the researcher’s mind about the prospective individual. The
researcher may use his wisdom to conclude that a particular individual may be
representative of the population in which one is interested.
The distinguishing feature of judgment sampling is that the population elements are
purposively selected. Again, the selection is not based on whether samples are
representative or not, but rather because they can offer the contributions sought. For
example, companies often try out new product ideas for their employees. The
rationale is that one would expect the firm’s employees to be more favourably
disposed towards a new product idea than the public. If the product does not pass
this group, it does not have a prospect of success in the general market.
ii. Quota Sampling: This is often used to improve the representativeness of the study.
Quota sampling is a special case of a two-staged judgemental sampling technique,
wherein the first stage consists of developing quotas based on the different
categories existing in the population. The characteristics may include age, gender
and income level, etc. that are identified on the basis of the researcher’s judgement.
Further, the researcher will set quotas for the identified groups. For example, the
researcher may determine the categories of gender and further decides to study 40%
of men and 60% of women in a population.
In the second stage of the sampling, the researcher will select 40% of men and 60%
of women based on convenience or judgement as discussed in the convenience
sampling.

10.5.3 Snowball Sampling

Consider the example where Mr. Ramesh, a part-time research scholar is concerned
with the personal fitness schedule followed by Managing Directors (MD) of large
companies. As part of this, Ramesh needed to interview the MD. He arranged his first
interview with the managing director of his own company. Towards the end of the
interview, the MD asked Ramesh whether he can be helpful in some way. Grabbing the
opportunity Ramesh asked for references of additional MDs those who can participate
in the study and provide useful data. He got two references and reached his sample size
of 15 executives in the same manner by getting references from his MD participants.

10.5.4 Self-Selection Sampling

In self-selection sampling, the sample elements from the population are given a choice
of whether they would like to be part of the sample. You, therefore:

8
 i. Give wide publicity to your sampling requirements. This may be done by
advertising through appropriate media or by asking them to take part by sending
emails or other platforms.
ii. Distribute the questionnaires to those who respond and collect the data.
Sudha’s research was concerned with the amount of time spent browsing Facebook
profiles. Sudha decided to administer her questionnaire using the Internet and therefore
prepared her questionnaire using Google Docs and posted the link on her Facebook
page with a request to fill the same. She publicized her research on a range of other
bulletin boards, asking for volunteers to fill in a questionnaire. Those who volunteered
by clicking on a hyperlink were automatically taken to her online questionnaire.

Check Your Progress Exercise 2

Note: a) Use the space below for your answers.

b) Compare your answers with those given at the end of the unit.

1) A simple random sample is a sample selected from a population in such a

manner that all members of the population have an equal chance of being
selected (True/False)
..................................................................................................................
2) What is a simple random sample?
a) A sample in which each unit of a population has an equal probability of
inclusion in the sample
b) A sample in which a higher number of one population is selected
c) When the researcher has a certain quota which must be selected from
various social groups
d) None of the above
3) Systematic sampling can produce a very biased sample when:
a) The population is too heterogeneous.
b) There is no available sampling frame.
c) There is a structure to the sampling frame.
d) The population is too small.
4) Which of the following is not true about stratified random sampling?
a) It involves a random selection process from identified subgroups
b) Proportions of groups in the sample must always match their population
proportions
c) Disproportional stratified random sampling is especially helpful for getting
large enough subgroup samples when subgroup comparisons are to be done
d) Proportional stratified random sampling yields a representative sample

9
 a) Elements within a stratum should be as homogeneous as possible.
5) The purpose of stratified random sampling is to make certain that
a) Every member of the population has an equal chance of being selected for
the sample.
b) The sample proportionately represents individuals from different categories
of the population.
c) The participants chosen for the study are the ones most likely to react to the
treatment.
d) The sample is more representative of the target population than the
accessible population.
6) In which of the following situations would stratified sampling be
recommended?
a) The population has more males than females, and the researcher is studying
a topic related to gender.
b) There are four distinct ethnic groups in New Orleans, and the researcher
wants to be assured each group is represented equally in the sample.
c) A superintendent wants to be sure she selects teachers from elementary,
middle/junior high, and high schools in numbers that are equal to the
proportions of teachers in these types of schools in her district.
d) All of these.
7) Cluster sampling is often preferred to other random sampling strategies because
it:
a) Is more complicated than simple random sampling.
b) Allows you to easily identify subgroups in the population.
c) Equally represents all groups in the population.
d) Can be used when a sampling frame is not available.
8) In the cluster sampling method
a) Every unit in the cluster is chosen to be part of the sample if the cluster they
belong to has been selected.
b) The chance of a unit being chosen to be part of the sample depends on the
size of the cluster.
c) Every unit closely resembles another unit within the same cluster.
d) None of the above
9) What is ‘multi-stage’ sampling?
a) This is a form of random sampling in which a clustered group is used as
representative of the population, thus enabling the researcher to drill right
down to individual data sources within the cluster.
b) Multi-stage sampling is a kind of stratified sampling in which the samples
are investigated at different stages of the processes that are being critically
reviewed.

10
 c) Multi-stage sampling is a sampling that takes place in stages over a period
of time.
d) All the above
10) A convenience sample is best described as
a) Participants who are students
b) Participants the experimenter calls or polls on the street
c) Participants who the experimenter has easy access to
d) Experimenters who the participants have easy access to
11) Under which circumstances would judgemental sampling be appropriate?
a) Selection of households in a geographic area
b) Test markets to determine the potential of a new product
c) People on the street interviews
d) Selection of a sample of students to evaluate their experience at the
university
12) Which of the following is NOT a characteristic of quota sampling?
a) The random selection of units makes it possible to calculate the standard
error.
b) It is a relatively fast and cheap way of finding out about public opinions.
c) Those who are available to be surveyed in public places are unlikely to be a
representative sample.
d) The researcher chooses who to approach and so might bias the sample.
e) The sample is subdivided into groups from which the researcher must obtain
a certain number of responses.
13) Quota sampling can be based on:
a) Observation or questioning.
b) A sample frame and sampling units.
c) Group membership or participation.
d) Interviewer discretion or judgment.
14) Snowball sampling can help the researcher to:
a) Overcome the problem of not having an accessible sampling frame.
b) Collect data cost-effectively.
c) Theorise inductively in a qualitative study.
d) Access difficult or hidden populations.
e) All of the above.
15) Which of the following is not a benefit of snowball sampling?
a) It can be used when there is difficulty in creating a sampling frame.
b) It is always representative of the population.

11
 c) It can be used within a qualitative research strategy.
d) It can be used to reflect relationships between people by tracing
connections.

Activity 1:
1. Suppose that a researcher is going to conduct a study to determine the
characteristics of graduate school that are most stressful to students. Identify
two variables that a researcher might use to define strata before selecting a
sample of graduate students for the study.

2. A researcher conducted an experiment using the members of his “Introduction

to Research Methods” class. The purpose of the experiment was to determine
if providing students with instructor notes would increase achievement.
Identify the most likely target and accessible populations and the type of
sample.

3. A total of 60 patients in a psychiatric hospital are selected for a survey about

hospital treatment. Equal numbers (20) of patients are selected from three
different categories defined by the length of stay in the hospital. Selection
from these three categories is done in such a way that each individual in each
category has an equal chance of being selected. What type of sampling is
this? Explain

4. Make a proposal for understanding how the Farmers rate the ongoing

12
 Government schemes and programmes. Clearly give the objectives for your
study and the sampling plan.

10.6 SAMPLE SIZE DETERMINATION

Now that you have understood the different sampling techniques and are ready with the
way you are going to select the sample, the next question logically is how many to
sample. That is what should be the sample size for the study. Sample size determination
is an important element in any survey research, although it is a difficult one.
Methodology for calculating the sample size statistically is available but going into
detail for a clearer understanding of the concepts involved is outside the scope of this
unit. However, a concise methodology is presented in this section to describe the
necessary concepts related to sampling size determination. You will be required to read
good references on this topic. The main steps to determine the sample size are outlined
below for studies involving means and proportions separately.

10.6.1 Determining Sample Size for Questions Involving Means

This case deals with the situation when the study uses means as the analyzing tool. That
is when we used an interval or ratio scale. For example, a study was conducted to know
the average amount spent on the purchase of mobile phones. The steps are discussed
below:
i. Estimate the standard deviation of the population

13
A rule of thumb for estimating the value of the standard deviation is to expect it to be
about one-sixth of the range. The range is the difference between the largest and the
smallest observation value. That is if a researcher conducting a study on mobile phone
purchase, expected the price paid to range from Rs.1000 to Rs. 7000, a rule-of-thumb
estimate for the standard deviation would be Rs. 1000 ((Rs. 7000-Rs.1000)/6).
ii. Determine a confidence level
A confidence level is a percentage or decimal value that signifies how confident a
researcher can be about being correct; it states the long-run percentage of confidence
intervals that will include the true population means. A 95% confidence is assumed for
this example.
iii. Decide on the Magnitude of the error
The range of values above and below the sample statistic is called the margin of error.
In simple terms, it is the expected difference between the population finding and the
finding you get from your sample study.
iv. Determine the Sample Size
To estimate the mean of a particular population, the formula for sample size is:
𝑍𝑆 2
𝑛=( )
𝐸
Where,
Z = standardized value obtained from the normal distribution corresponding
to the selected confidence level.
S = sample standard deviation
E = acceptable positive/negative magnitude of the error.
Suppose a business researcher studying expenditures on mobile phones wishes to have
a 95 percent confidence level (Z = 1.96, from the normal distribution table) and a range
of error (E) of less than Rs. 200. If the estimate of the standard deviation is Rs. 2900
(known from similar studies in past), the sample size can be calculated as follows:
𝑍𝑆 2
𝑛=( )
𝐸
(1.96) (2900) 2
=( ) = 28.422 = 808
200
Hence the sample size required for the study is 808.
If a range of error (E) of Rs. 400 is acceptable, then the necessary sample size will be
reduced:
𝑍𝑆 2
𝑛=( )
𝐸

14
 (1.96) (2900) 2
= ( ) = 14.212 = 202
400
You can observe that increase in the value of error provides a less sample size.

10.6.2 Determining Sample Size for Proportions

This case deals with the situation when the study uses proportions as the analyzing tool.
That is when we used an interval or ratio scale. For example, a study was conducted to
know the proportion of households possessing two cars in a locality. The steps used are
similar to what we have used in sample size estimations for means with the following
formula:
𝑍 2 𝑝𝑞
𝑛=
𝐸2
Where
n = number of items in the sample
Z2 = square of the confidence level in standard error units
p = estimated proportion of successes
q =1 – p (estimated proportion of failures)
E2 = square of the maximum allowance for error between the true proportion and
the sample proportion.
Let us understand the application with the help of the following illustrative example:
Suppose a researcher believes that a simple random sample will show that 60 percent of
the population (p) recognizes the name of an automobile dealership. The researcher
wishes to estimate with 95 percent confidence (Z c = 1.96, from the normal distribution
table) such that the allowance for sampling error is not greater than 3.5 percentage
points (E). Substituting these values into the formula gives;
𝑛 = 1.962 (0.6)(0.4)/0.0352
0.922
𝑛 = 0.001225= 753

The estimated sample size is 753.

Check Your Progress Exercise 3

Note: a) Use the space below for your answers.

b) Compare your answers with those given at the end of the unit.

1) A researcher wants the estimate to be within ± Rs. 25 of the true population

value and he wishes to be 95 percent confident that it will contain the true
population means. Also, assume that early studies have demonstrated the

15
 standard deviation to be around Rs. 100. What would be the required sample
size?
............................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................

2) An alcohol company wants to conduct a survey to estimate the percentage of

preference for a specific brand within ±5 per cent points and that the company
wishes to be 95 per cent confident of the result. Also, assume that the
percentage of preference for the brand is believed to be 35 per cent. What
sample is appropriate in this case?
............................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................

10.7 SAMPLING AND NON-SAMPLING ERRORS

An agriculturist sampled a small number of fields growing rice in Maharashtra and
collected data on the type of fertilizer used and the soil. The agriculturist hoped to
measure the yield of the crop and expected this sample to be representative of the rice-
growing population. A good research design may not affect the outcome of this
research much and there is a possibility that the findings of the sample may be nearer to
the actual population parameter. However, there is a possibility of error called total
error in the findings of the sample when compared with the actual population
parameters. The total error comprises error sources in two broad categories:
1) Random sampling error
2) Systematic (non-sampling) error
For example, the annual average consumption of a product of a target population may
be 8 units per week, as determined from the database of the company, but a marketing
research study estimates it at 4 units per week based on a sample study.

10.7.1 Random Sampling Errors

The errors are caused due to imperfect representation of the population in the sample.
In statistical terms, it is defined as the variation between the true mean value for the
sample and the true mean value of the population.
The sampling method applied, and the sample size is the two principal sources of
sampling error. This is due to the fact that only a part of the population is covered in the
sample. The magnitude of the sampling error varies from one sampling method to the
other, even for the same sample size. For example, the sampling error associated with
simple random sampling will be greater than stratified random sampling if the

16
population is heterogeneous in nature. Intuitively, we know that the larger the sample
the more accurate the research. In fact, the sampling error varies with samples of
different sizes. Increasing the sample size decreases the sampling error.

10.7.2 Non-Sampling Errors

The non-sampling errors arise from faulty research design and mistakes in executing
research. There are many sources of non-sampling errors which may be broadly
classified as (a) respondent errors, and (b) administrative errors.
a) Respondent Errors: If the respondents co-operate and give the correct
information, the objectives of the researcher can be easily accomplished. However,
in practice, this may not happen. The respondents may either refuse to provide
information or even if he/she provides information it may be biased. If the
respondent fails to provide information, we call it a non-response error.
Response bias occurs when the respondent may not give the correct information
and try to mislead the investigator in a certain direction. The respondents may
consciously or unconsciously misrepresent the truth. For example, if the
investigator asks a question about the income of the respondent he/ she may not
give the correct information for obvious reasons. The response bias may also occur
because the interviewer’s presence influences respondents to give untrue or
modified answers.
b) Administrative Errors: Administrative errors are the errors that arise due to
improper administration of the research process. There are four types of
administrative errors.
These are as follows:
i. Sample Selection Error: For example, we may plan to use a systematic
sampling plan in a market research study for evaluating a product placement in
a movie and decide to interview every 5th customer coming out of a multiplex.
It may happen we are including all those patrons who may have not watched
that particular movie in which the product was placed. This may lead to an error
because of the unrepresentative sample selection.
ii. Investigator Error: When the investigator interviews the respondent, he/ she
may fail to record the information correctly or may fail to cross-check the
information provided by the respondent. Therefore, the error may arise due to
the way the investigator records the information.
iii. Investigator Cheating: Sometimes the investigator may try to fake the data
even without meeting the concerned respondents. There should be some
mechanism to crosscheck this type of faking by the investigator.
iv. Data Processing Error: Data editing, coding, and data entry into the computer
follow immediately after data collection. The errors can be minimized by
careful editing, coding, and entering the data into a computer.

17
10.7.3 Control of Errors
It is not possible to eliminate completely the sources of errors but we can make efforts
to minimize these sources of errors as much as possible. Some of the steps we should
use are:
a. Designing and executing a good questionnaire
b. Selection of appropriate sampling method
c. Adequate sample size
d. Employing trained investigators to collect the data; and
e. Avoid errors in editing, coding, and entering the data into the computer.

Check Your Progress Exercise 4

Note: a) Use the space below for your answers.

b) Compare your answers with those given at the end of the unit.

1) A researcher mailed surveys to 100,000 people but only 1,000 responded, resulting
in a response rate of 1%. This study is threatened by _____.
a) Nonresponse error
b) Response error
c) Sampling error
d) Measurement error
e) Coverage error
2) Which of the following is true regarding non-sampling errors?
a) They occur due to the normal variation in data that occurs as a result of each
unique sample.
b) Like sampling error, a non-sampling error can also be measured.
c) Nonsampling errors may not be measured but appropriate steps may be taken
to limit their occurrence.
d) Nonsampling errors may be calculated to four decimal places.
e) Nonsampling errors may be calculated except in cases where interviewers
cheat.
3. As the sample size increases, the sampling error:
a) Increases
b) Decreases

18
 c) Remain Constant
d) None of the above
4. The sampling errors arise due to:
a) The investigator’s bias
b) The data processing problem
c) The respondent’s bias
d) The sampling method applied
5. Which of the following is true?
a) The sampling error is unrelated to the sample size.
b) The sampling error becomes progressively larger with the increasing sample
size.
c) The sampling error becomes progressively smaller with the increasing sample
size.
d) The sampling error arises due to administrative errors.

10.8 LET US SUM UP

The sample size and the sampling technique to be used are determined by the
availability of resources, in particular budget and time constraints. Probability sampling
techniques are often more time-consuming than non-probability techniques. Non-
probability sampling techniques also provide the opportunity to select the sample
purposively and to reach difficult-to-identify members of the population. For many
research studies, it is needed to use a combination of different sampling techniques
leading to multistage sampling. However, it has to be understood that the selection of a
sampling plan is dependent on the ability to gain access to organizations and thus the
practically feasible plan should be selected.
The overall objective of drawing a good representative sample and selecting an
appropriate sample size is to minimize the total error, which is further classified as
random sampling errors and systematic non-sampling errors.

10.9 KEY WORDS

Cluster Sampling : Cluster sampling is a sampling technique used when
“natural” but relatively homogeneous groupings are
evident in a statistical population.
Convenience Sampling : A statistical method of drawing representative data by
selecting people because of the ease of their volunteering
or selecting units because of their availability or easy
access.

19
Judgement Sampling : In judgement sampling, the researcher or some other
“expert” uses his/her judgement in selecting the units
from the population for study based on the population’s
parameters.
Multistage Sampling : Multistage Sampling is a sampling strategy used when
conducting studies involving a very large population. The
entire population is divided into naturally-occurring
clusters and sub-clusters, from which the researcher
randomly selects the sample.
Purposive Sampling : Purposive sampling technique is a type of non-probability
sampling where the researcher consciously selects
particular elements or subjects for addition in a study to
make sure that the elements will have certain
characteristics pertinent to the study.
Quota Sampling : Quota sampling is a method for selecting survey
participants. In quota sampling, a population is first
segmented into mutually exclusive sub-groups, just as in
stratified sampling.
Simpl Random Sampling : A simple random sample is a sample of size n drawn
from a population of size N in such a way that every
possible sample of size n has the same chance of being
selected.
Sampling Error : Sampling Error is an error that occurs when using
samples to make inferences about the populations from
which they are drawn.
Sampling : Statistical method of obtaining representative data or
observations from a group.
Sampling Frame : A list of sampling units.
Self-Selection Sampling : Self-selection sampling is a type of non-probability
sampling technique. Non-probability sampling focuses on
sampling techniques that are based on the judgement of
the researcher.
Snow Ball Sampling : Snowball sampling is a non-probability sampling
technique that is used by researchers to identify potential
subjects in studies where subjects are hard to locate.
Stratified Sampling : A stratified sample is a probability sampling technique in
which the researcher divides the entire target population
into different subgroups or strata, and then randomly
selects the final subjects proportionally from the different
strata.

20
Systematic Sampling : A method of choosing a random sample from among a
larger population.

10.10 REFERENCES/SUGGESTED READINGS

1. Adams, J., Khan, H., Raeside, R. and White, D. (2007), Research Methods for
Graduate Business and Social Science Students, Response, Sage Publications,
New Delhi.
2. Anderson, D. R., Sweeney, D. J. and Williams, T. A. (2010), Fundamentals of
Business Statistics, Sixth Edition, South-Western, Cengage Learning Inc.
3. Black, K. (2004), Business Statistics – For Contemporary Decision Making,
Fourth Edition, John Wiley and Sons. Inc.
4. Hira, D. S. and Gupta. P.K. (2007), Operation Research, S Chand and
Company, New Delhi.
5. Kothari, C. R. (2004), Research Methodology: Methods and Techniques,
Daryaganj, Delhi, India, New Age International.
6. Malhotra, N and Birks D. (2007), Marketing Research- An Applied Research,
Third European Edition, Prentice Hall, Financial Times, England.
7. Saunders, M., Lewis, P. and Thornhill, A. (2003), Research Methods for
Business Students, Third Edition, Pearson Education Ltd.
8. Weathington, B, Cunningham, C, and Pittenger (2012), Understanding Basic
Research, John Wiley and Sons, New Jersy.
9. Zikmund, W. G., Babin, B. J., Carr, J. C. and Griffin, M. (2013), Business
Research Methods, Ninth Edition, Cengage Learning Higher Education.
Website:
 http://www.southalabama.edu/coe/bset/johnson/mcq/mc7.htm
 http://www.jsu.edu/depart/ccba/jthomas/downloads/497/497sampling.ppt

10.11 ANSWERS TO CHECK YOUR PROGRESS EXERCISES

Check Your Progress Exercise 1
1. d
2. a
3. d
4. d
[

21
Check Your Progress Exercise 2
1. True
2. a
3. c
4. b
5. b
6. d
7. d
8. a
9. a
10. c
11. b
12. a
13. c
14. e
15. b

Check Your Progress Exercise 3

1.
𝑍𝑆 2
𝑛=( )
𝐸
(1.96) (100) 2
=( ) = 64
25
2. 𝑛 = 1.962 (0.35)(0.65)/0.052 = 350

Check Your Progress Exercise 4

1. a
2. c
3. b
4. d
5. c

22
10.12 UNIT END QUESTIONS
1. Define the following terms
i. Population
ii. Sampling
iii. Census
iv. Sampling element
2. Discuss the importance of sampling.
3. What are the limitations to conduct a census survey?
4. How is random sampling done? Discuss the procedure.
5. Distinguish between Probability and Non- Probability-based Sampling
Techniques.
6. Discuss the different types of Sampling Techniques.
7. Despite the researcher’s judgment, judgment sampling is still a type of non-
probability sampling? Elaborate.
8. One of the strengths of snowball sampling is that it leads to a very
heterogeneous sample thus reducing bias. Discuss.
9. A self-selecting sample is usually highly representative of the target population.
Elaborate.
10. What do you mean by sample size determination?
11. How do determine the sample size for means?
12. How do determine the sample size for proportion?
13. Differentiate between sampling and non-sampling errors.

23