Sampling

 Population
Sample
Sampling
Techniques of sampling
Population
 In Statistics, the term ‘Population or Universe’ refers to
the totality or aggregate of cases (or items) under
investigation.
 In other words, a population is the entire group of items
or members or individuals of interest in a study.
 A population may be finite or infinite.

Sample
 Some units selected from the population in known as
sample.
 In other words, a sample is a small representative part of
a population.
 Population: Set of entire items
under consideration
Sample : Set of items chosen
for study.

Sampling :
 The process of selecting some units from the population in order
to draw conclusion about the population in known as sampling.
 Sampling is a process of choosing a representative portion of a
population.
 Sampling may be defined as the selection of some units or parts of
an aggregate or totality in such a way that some of these parts or
units can represent the totality under study.
Objectives of Sampling:
 As a sample study includes only few elements which
uses minimum resources such as minimum time,
minimum budget, less administrative work etc. to
make conclusion about entire population under study.
 The major objectives of sampling technique are as
follows:
(1) To obtain maximum information about the
population under consideration by examine number
picked up in the sample.
(2) To ascertain confidence interval to estimate the
population parameter.
(3) To test the significance of the population parameter
at given level of significance.
 Techniques of sampling
[There are various methods of sampling. The selection of appropriate
method depends on: - the objective and scope of the enquiry
- nature of population
- size of the sample
- available resources, etc.
Sampling methods can be classified into two broad categories.]

1. Random / Probability Sampling 2. Non-random / non-probability Sampling

1. Random / Probability Sampling
 Random sampling is that method of drawing a
sample from a population so that all possible
samples of fixed size (n) have the same probability
of being selected.
 Probability sampling have the following
characteristics:
• Each sample unit has an equal chance of being
selected.
• Sampling units have some known probability of
samples of entering the sample.
• Probability of selection of a unit is proportional to
the sample size.
 Types of Random Sampling
a) Simple Random Sampling
b) Stratified Random Sampling
c) Systematic Random Sampling
d) Cluster Sampling
e) Multi – stage Sampling
Simple Random Sampling
 In Simple Random Sampling, each unit in the
population has equal chance of being selected.
 The selection of an item depends on chance and not
on the discretion or judgment of the investigator.
 This method is also called “Equal Probability
Sampling”.
 The probability of selecting a unit in first draw =1/N ,
where N = population size.
 This method is suitable for small homogenous
population.
 Sampling can be done either “Sampling With
Replacement” or “Sampling Without Replacement”.
 A simple random sampling is generally selected in two ways:
(i) Lottery method
(ii) Use of Random Number Table.
Merits: (1) this method is simple and convenience to use.
(2) It is free from the personal bias or prejudice of the
investigator.
(3) It takes less time and labour.
(4) As this method is based on the theory of probability, sampling
error can also be calculated.
Demerits: (1) For this method, a complete list of the population is
required but which is not always possible.
(2) The method becomes costly and time consuming when the
field of inquiry is very large and scattered.
(3) The sample selected may not be the representative of the
population in some cases.
Systematic Random Sampling
 In this sampling, only the first unit is selected at random and the
remaining units are automatically selected at fixed equal intervals
from one another.
 This method may be successfully used when the complete and up-
to-date list of sampling units is available.
 The items are first arranged in numerical, alphabetical or some other
orders. Then they are numbered serially.
 If the population size be N and sample size be n, then the sample
interval (k) = N/n
 The first item between 1 and K is selected at random by Lottery
method and every Kth item after that is selected in succession.
 E.g. K = 100/10 = 10
 Select the first item from the numbers 1 to 10 and complete the
sample by taking every 10th item after it.
 Suppose the number selected first be 5, then the sample will consist
of the items numbered 5, 15, 25, 35, 45, 55, 65, 75, 85, and 95.
Stratified Random Sampling
 For large and heterogeneous population, stratified
sampling is the best design.
 Under stratified random sampling, the population is
divided into various homogeneous groups or strata on the
basis of certain characteristics.
 The various strata are of different sizes and non-
overlapping.
 The characteristics within the strata are homogeneous and
between strata are heterogeneous.
 Then a simple random sampling technique is applied to
select a sample from each stratum. These samples are
then combined to form a single sample of the universe.
 The sample sizes in various strata may be proportionate or
disproportionate.
 S1 S2 S3
1000 1400 1600
Proportionate sample (10% of each stratum)
= 100 + 140 + 160 = 400
Disproportionate sample (150 of each stratum)
= 150 + 150 + 150 = 450
Cluster Sampling
 This sampling method is also appropriate if the
population characteristics are not homogeneous.
 In this method, the population is divided into sub-
population (group), called cluster in such a way
that the characteristics within the cluster are
heterogeneous and between the clusters is
homogeneous.
 The number of sampling units in each cluster
should be approximately same.
 Then a few clusters are selected as a and each and
every element in the cluster is studied.
Difference between stratified random sampling and
cluster random sampling is that:
 Stratified random sampling is used when each
group (a strata) has small variation within itself
but there is wide variation between the groups
(strata).
 Whereas the cluster sampling is used when there
is considerable variation with each group (cluster)
but the groups (clusters) are essentially similar
(homogeneous) to each other.
 In cluster sampling, clusters themselves are
selected and all the units in the selected clusters
are enumerated.
Multi-stage Sampling (Multi-phase Sampling):
 Multi-stage sampling is a further development of
the principle of cluster sampling.
 This sampling is done in various stages.
 At the first stage, the universe is divided into large
sample units and the sample is selected at random
from them.
 At the second stage, the samples selected at the
first stage are divided again into smaller sample
units and from which a further random sample is
taken.
 Sampling in any other stages may be done in the
same way till we get the required sample.
Example:
 For survey of the living standards of families in the
villages of Nepal,
 A sample of Anchals may be taken first.
 From each of these selected Anchals, a sample of
districts is taken.
 Then a few villages are selected from each of the
selected districts.
 Finally, some families are selected from each of
the selected villages.
 2. Non-random / non-probability Sampling
• If the selection of the individual from the
population to form a sample is not depending
on chance but depends on the judgment or
convenience of the investigator , the sampling
is said to be non-random sampling.
• In other words, non-random sampling is a
procedure of selecting a sample without the
use of probability or randomization. It is based
on convenience, judgment etc.
 Types of Non-random Sampling
a) Judgment or Purposive Sampling
b) Convenience Sampling
c) Quota Sampling
Judgment or Purposive Sampling
 In this method, the choice of items in the sample
depends entirely upon the judgment of the
investigator.
 The investigator may select a sample which shall
yield results favourable to his/her point of view.
 E.g. if economic conditions of people living living
in a Anchal are to be studied, a few towns and
villages may be purposively selected for intensive
study on the principle that they can be
representative of the entire Anchal.
Convenience Sampling:
 In this method, the investigator selects the
samples on the basis of the convenience of the
investigator.
 This is also known as ‘chunk sampling’ where a
chunk or part of the population is selected
without using any probability law.
 However, the results obtained by convenience
sampling method can hardly said to be
representative of the population parameters.
Quota Sampling:
 A sampling method in which researchers are given
quotas to be filled from the different strata and
within the pre-assigned quotas, the process of
drawing the required samples from these strata
by judgment sampling is called quota sampling.
 This is a restricted type of purposive or judgment
sampling.
 Sampling units may be fixed according to some
specified characteristics such as income group,
sex, occupation, political or religious affiliation
etc.
e.g. : a researcher wants to study the attitude of
patients towards the nursing services in
government hospitals. He may fix a quota of 200
patients and go on collecting data from first 200
patients in government hospitals. The sample of
200 patients thus selected is called quota sample.
Sampling Distribution:
 The Sampling Distribution is the probability distribution of sample
statistics computed from all possible samples drawn from the
population under study.
 Sampling distribution plays very important role in statistics as it
provides a major simplification in the statistical inference.

Sampling Distribution of Sample Mean:

 The sampling distribution of mean is the probability distribution
of sample means computed from all possible samples drawn from
the population under study.
Example :
 If we draw all possible samples of size n = 2 from a population
having observation 2 , 4 , 6 , 8 , 10 without replacement, then
 Total no. of possible samples ( k ) = NCn = 5C2 = 10
 The 10 possible samples are:
Sample Sample Sample Sample Frequency Relative Probability
No. values Mean ( X ) Mean ( X ) (f) Frequency
(X)
1 (2,4) 3 3 1 1/10 0.1
2 (2,6) 4 4 1 1/10 0.1
3 (2,8) 5
5 2 2/10 0.2
4 ( 2 , 10 ) 6
5 (4,6) 5 6 2 2/10 0.2
6 (4,8) 6 7 2 2/10 0.2
7 ( 4 , 10 ) 7
8 1 1/10 0.1
8 (6,8) 7
9 ( 6 , 10 ) 8 9 1 1/10 0.1
10 ( 8 , 10 ) 9
10 1.0

 This is called the sampling distribution of sample means.

 Similarly, the probability distribution of the sample variances is
called the sampling distribution of the sample variances.
 Similarly, the probability distribution of the sample proportions of
all possible samples is called the sampling distribution of the
proportion.

Standard Error ( S.E. ) :

 The standard deviation of the sampling distribution of
sample statistic is known as its standard error (S. E.) of
the statistic.
(A) Standard Error of Mean :
 The standard deviation of the sampling distribution of sample
mean is known as its standard error (S. E.) of the sample mean.
 It is denoted by S.E.( X ) or σX
(B) Standard Error of Proportion :
 The standard deviation of the sampling distribution of sample
proportion is known as its standard error (S. E.) of the sample
proportion.
 It is denoted by S.E.( p ) or σp

Formulas for Standard Error ( S.E. ):

(1) If the random sample is drawn with replacement from large
population:
S.E.( X ) = σ σ = Population S.D.

n n = Sample size
S.E.( p ) = PQ P = Population Proportion
n Q=1–P
p = sample proportion
(2) If the random sample is drawn without replacement from a finite
population of size ( N ), then
S.E.( X ) = σ x N - n
n N-1
S.E.( p ) = PQ x N - n
n N-1

Where, N-n is called finite correction factor or

N–1 finite population multiplier

(3) Standard Error of difference between two sample means is given

by –
S.E.( X1 – X2 ) = σ12 + σ22
n1 n2
(4) Standard Error of difference between two sample proportion is
given by –
S.E.( p1 – p2 ) = P1 Q1 + P2 Q2
n1 n2
Q. 1 : An accounting firm has 5 branch offices in the 5 areas of the city. These offices
employ 3 , 4 , 5 , 6 and 7 employees respectively.
(i) Write down all possible samples of size two without replacement.
(ii) Show the mean of the sampling distribution of the sample mean is equal to
population mean.
(iii) Obtain the standard error of the sampling distribution of the sample mean.
Sol. : Here, Population data ( Y ) = 3 , 4 , 5 , 6 , 7
Population size ( N ) = 5 Sample size ( n) = 2
Population mean ( μ ) = ∑Y / N
= ( 3+4+5+6+7)/5 = 5
Population s.d. ( σ ) = ∑( Y – μ )2/N

= [(3 – 5)2 + (4 – 5)2 + (5 – 5)2 + (6 – 5)2 + (7 – 5)2 ]/5

= (4 + 1 + 0 + 1 + 4)/5 = 2
(i) Total no. of possible samples drawing without replacement ( K ) = NCn = 5C2 = 10
The 10 possible samples are : (3,4) , (3,5) , (3,6) , (3,7) , (4,5) , (4,6) , (4,7) , (5,6)
(5,7) and (6,7)
(ii) Calculation of sampling distribution of mean ( X )
Sample No. Sample values ( X ) Sample Mean ( X )
1 (3,4) 3.5
2 (3,5) 4
3 ( 3 , 6) 4.5
4 ( 3 , 7) 5
5 ( 4 , 5) 4.5
6 (4,6) 5
7 (4,7) 5.5
8 (5,6) 5.5
9 (5,7) 6
10 (6,7) 6.5
K = 10 ∑ X = 50

We have , X = ∑ X / k = 50/10 = 5
Since X = μ = 5.
Hence, the mean of the sampling distribution of mean is equal to the
population mean.

(iii) The standard error of the sampling distribution of the sample mean is given by-

S.E.( X ) = σ x N-n
n N-1

= 2 x 5–2
2 5–1
= 3/4
= 0.87