Business

Research Methods

Unit IV
Sampling Theory

❖Concept, Need and Importance of Sampling,

❖Types of Sampling Methods,
❖Sampling and Non-Sampling Errors,
❖Sample Design; Determinants of Sample Size,
❖Steps in Designing the Sample,
 Introduction

The concept of sampling has been introduced with a view to

making the research finding economical and accurate. It is a
process used in statistical analysis in which a predetermined
number of observations will be taken from a large population.
According to Davis S. Fox, “It is not possible to collect data
from every respondent relevant to our study but only from some
fractional part, the process is known as sampling.”
For example, to decide purchasing about milk, wheat, rice,
fruits, etc. The most important aim of sampling is to obtain
maximum information about the population under study with the
least uses of money, labour & time.
Basic Concepts of Sampling
 Difference between Census & Sampling
Census Sampling

1. A method of collecting information 1. A method of collecting information

about each item of the population. about a small part of population.
2. Data from census is reliable & accurate. 2. Data from sampling may have a margin
of error.
3.This method is more time consuming. 3. This method is less time taking.

4. In this method, there is zero sampling 4. In this method, there is a possibility of

variance, because the data used is drawn sampling variance, because the data used
from the entire population. is taken from only a small part of the
population
5. All the items relating to a universe 5. Only a few items are investigated.
(targeted population) are investigated.
6. This method is useful where units of 6. This method is useful where population
population are heterogeneous. units are homogeneous.
 Need and Importance of Sampling

We will discuss importance of sampling with the help of

following points;
1.Saves Time, Money & Effort:
2.When census is not possible or rational
3.More Effective:
4.Faster and Cheaper:
5.More Accurate:
6.Provides More Comprehensive Information:
Characteristics of a Good Sample
 Business
Research Methods

Unit IV
Sampling Theory

❖Concept, Need and Importance of Sampling,

❖Types of Sampling Methods,
❖Sampling and Non-Sampling Errors,
❖Sample Design; Determinants of Sample Size,
❖Steps in Designing the Sample,
 Types of Sampling Methods
Sample Designing is divided into two broad categories, i.e.
probability sampling & non-probability sampling;
 (A) Probability Sampling

Probability sampling is also known as ‘random sampling’ or ‘chance sampling’. In

this method each item of the universe has some chances of selection. We further
divide this method into following sub-parts;
1.Simple Random Sampling- If there is not much variation in the characteristics of the
members of a population, we can follow the method of simple random sampling. Here
each element in the population has equal chance of being included in a sample &
choices are independent of each other. This method eliminates the chance of bias.
There are two variants of simple random sampling, viz., simple random sampling with
replacement (SRSWR where total possible samples is Nn) and simple random
sampling without replacement (SRSWOR where total possible samples is NCn)
2.Systematic Sampling- one unit is selected at random from the universe & the other
units are at a specific interval from the selected unit. This method is used where
population is finite & arranged.
3.Stratified Random sampling- here; population is divided into strata or groups that are
mutually exclusive and include all items, then simple random sampling is chosen.
Symbolically; N1, N2, N3….N15 & N1 + N2 + N3....+ N15 = N, where population is
divided into 15 strata.
 Probability Sampling (cont…)

4) Cluster Sampling- In this method, the population is divided into some

recognisable sub-groups that are known as clusters. Now, we draw
clusters by using simple random, then all the units belonging to the
selected clusters constitute the sample. The clusters should be as small as
possible in size & number of units selected from each cluster should be
almost same. This method is economical but not as precise as random
sampling.
5) Multi-Stage Sampling- This is modified form of cluster sampling.

Mixed Sampling- Suppose, an institute has to send 5 students for

managerial training in a company during the summer vacation. Initially,
it may be shortlisted about 20 students who are considered to be suitable
for the training by applying its own discretion. Then from these 20
students, 5 students may finally be selected by random sampling.
 Business
Research Methods

Unit IV
Sampling Theory

❖Concept, Need and Importance of Sampling,

❖Types of Sampling Methods,
❖Sampling and Non-Sampling Errors,
❖Sample Design; Determinants of Sample Size,
❖Steps in Designing the Sample,
Introduction- This is inaccuracy in the process of collection, processing &
analysis of the data (sample). Broadly, we can divide errors in two categories;
(A) Sampling Errors, & (B) Non-Sampling Errors
 (A) Sampling Errors

This type of error originates at the time of collecting samples. One

of the important reason behind it is drawing conclusions about
population on the basis of a small sample. Sampling error may be
divided into two sub-categories;
A) Biased Sampling Error; when variation in result is because of
bias in selection, estimation, etc.
B) Unbiased Sampling Error- When we use simple random
sampling with some restrictions, but creates no bias in result.
Reasons for Sampling Errors-
1)Faulty Selection of the Sample-
2)Substitution-
3)Faulty Demarcation of Sample Units- when sample is taken after
wrongly demarcated sample of crop.
 Methods for Reducing Sampling Errors

We can reduce sampling errors with the help of two techniques, as

follows;
a)Increasing the Size of the Sample- It is obvious that if sample size
‘n’ will approach to population size ‘N’, there will be no sampling
error. Symbolically; …..
b)Stratification- As we know, when all the items in the population
are homogeneous, a simple random sample is likely to be
representative of the population. But, for heterogeneous items
simple random sample may not be representative of the population.
Therefore, we use stratification
 (B) Non-Sampling Errors

This type of error originates at the time of observation,

approximation & processing (but not in collection) of data. This
error may be due to wrong sampling planning, lack of trained &
qualified investigators, inaccuracy in responses collected due to
bias of researcher or respondent, errors in design of the survey &
errors in compilation.
Types;
1)Frame Error
2)Non-Response Error
3)Measurement Error
4)Data Processing Error
5)Data Analysis Error
 Causes for Non-Sampling Errors

Some of the factors that create non-sampling error are as

follows;
a)Errors due to Faulty Planning- when lack of trained investigators,
b)Response Errors-
c)Non-Response Error- when respondent is unable to complete the
information, do not want to answer certain questions, etc
d)Errors in Coverage-
e)Compiling Errors- The errors arise due to compilation e.g. error in
editing & coding of responses
 Methods to Reduce the Non-Sampling Errors

We can reduce non-sampling by following ways;

a)Careful selection of the time when survey is conducted,
b)Using an up-to-date & accurate sampling frame,
c)Planning for follow up of non-respondents,
d)Careful questionnaire design,
e)Providing complete training for investigators & processing staff,
 Business
Research Methods

Unit IV
Sampling Theory

❖Concept, Need and Importance of Sampling,

❖Types of Sampling Methods,
❖Sampling and Non-Sampling Errors,
❖Sample Design; Determinants of Sample Size,
❖Steps in Designing the Sample,
 Sample Size
The sample size is the number of observations that constitute it. This number is
denoted by ‘n’.
The objective of research determines level of accuracy in result & this level of
accuracy determines sample size. In this way sample size should be large….., but
with budget constraint.
Determinants of Sample Size-
In addition to objective of the study, some of the determinants are as below;
1)Size of the Population-
2)Degree of Accuracy Desired- The greater the degree of accuracy desired, the larger
should be the sample size. If untrained staff collects sample, there will be less
accuracy even after large sample size.
3)Resources Available-
4)Degree of Variability- More heterogeneous a population, larger the sample size
required to obtain a given level of precision.
5)Method of Sampling Adopted-
6)Nature of Respondents- When it is expected that a large number of respondents
will not cooperate & sent back the questionnaire, a large sample should be selected.
Numerical Example on Sample Size
Numerical Example on Sample Size (cont…)
Numerical Example on Sample Size (cont…)
Numerical Example on Sample Size (cont…)
 Business
Research Methods

Unit IV
Sampling Theory

❖Concept, Need and Importance of Sampling,

❖Types of Sampling Methods,
❖Sampling and Non-Sampling Errors,
❖Sample Design; Determinants of Sample Size,
❖Steps in Designing the Sample,
 Steps in Designing the Sample

Sampling plan should be designed in such a way that the resulting data
should be a representative sample, and all the required questions of the study
are answered. Some of the important steps taken in to designing sample are
as below;
1)Define the Population
2)Sample Frame
3)Specifying the sampling Units
4)Selection of Sample Design
5)Determination of Sample Size
6)Select the Sample
• Fun with standard
normal probabilities!

• Problem :

• you are 78 inches

(6’6”) tall & bet a
friend that you are
the tallest person on
campus. Campus
heights in inches are
~N (64, 10). What’s
the probability that
you’re wrong?
 t-table

• No longer can we assume

that the pop mean (μ) will be
within 1.96 std. deviations of
the sample mean in 95 out of
100 samples.

• The smaller the sample the

more std. deviations we can
expect μ can be from x-bar at
a given level of confidence.

• Degrees of freedom capture

the sample size, In our case=
n-1