Simple Random Sampling

UNIT 2 SIMPLE RANDOM SAMPLING

Structure
2.1 Introduction
Objectives
2.2 Methods of Selection of a Sample
Lottery Method
Random Number Method
Computer Random Number Generation Method
2.3 Properties of Simple Random Sampling
Merits and Demerits of Simple Random Sampling
2.4 Simple Random Sampling of Attributes
2.5 Sample Size for Specific Precision
2.6 Summary
2.7 Solutions/Answers

2.1 INTRODUCTION
Simple random sampling refers to the sampling technique in which each and
every item of the population is having an equal chance of being included in
the sample. The selection is thus free from any personal bias because the
investigator does not make any preference in the choice of items. Since
selection of items in the sample, depends entirely on chance, this method is
also known as the method of probability sampling.
Random sampling is sometimes referred to a representative sampling. If the
sample is chosen at random and the size of the sample is sufficiently large, it
will represent all groups in the population. An element of randomness is
necessary to be introduced in the final selection of the item. If that is not
introduced, bias is likely to enter and make the sample unrepresentative.
Methods of selection of a simple random sample are explained in Section 2.2.
In Section 2.3 the properties of simple random sampling are described. Simple
random sampling of attributes is introduced in Section 2.4 whereas in Section
2.5 the sample size determination for specific precision is described briefly.
Objectives
After studying this unit, you would be able to
 describe the simple random sampling;
 explain the method of SRSWR and SRSWOR;
 explain and derive the properties of simple random sampling;
 calculate the variance of the estimate of the sample mean;
 describe the SRS for attribute and its properties; and
 describe and draw a appropriate sample size for specific precision.

25
Statistical Techniques
2.2 METHODS OF SELECTION OF A SAMPLE
The random sample obtained by a method of selection in which every item
has an equal chance to be selected in the sample. The random sample depends
not only on method of selection but also on the sample size and nature of
population.
Simple Random Sampling
If a sample of n units is selected randomly from a population of size N, this
method is known as simple random sampling. As the name suggest, simple
random sampling is a method in which the required number of elements /units
are selected simply by random method from the target population. One can
select a simple random sample by either of these two methods with
replacement method and without replacement method.

Simple Random Sampling with Replacement (SRSWR)

When simple random samples are selected in the way that units which has
been selected as sample unit is remixed or replaced in the population before
the selection of the next unit in the sample then the method is known as
simple random sampling with replacement.

Simple Random Sampling without Replacement (SRSWOR)

When simple random sample are selected in the way that a unit is selected as
sample unit is not mixed or replaced in the population before the selection of
the next unit. This method is known as simple random sampling without
replacement i.e. once a unit is selected in the sample will never be selected
again in the sample.
Some procedures are simple for small population and is not so for the large
population. Proper care has to be taken to ensure that selected sample is
random. Random sample can be obtained by any of the following methods:

1. By Lottery method;
2. By ‘Mechanical Randomization’ or ‘Random Numbers’ method; and
3. By Computer Random Number Generation method.
2.2.1 Lottery Method
This is very popular method of selecting a random sample under which all
items of the population are numbered or named on separate slips of paper.
These slips of paper should be of identical size, color and shape. These slips
are then folded and mixed up in a container or box or drum. A blind fold
selection is then made of the number of slips required to constitute the
desired size of sample. The selection of items thus depends entirely on
chance. For example if we want to select n candidates out of N. We assign
the numbers 1 to N. One number to each candidate and write these numbers
(1 to N) on N slips which are made as homogeneous as possible. These slips
are then put in bag and thoroughly shuffled and then n slips are drawn one by
one. Then the n candidates corresponding to numbers on the slip drawn will
constitute a random sample.
If we draw a slip and note down the number written on the slip and then again
replace the slip in the bag and then draw the next. These number of slips
constitute a sample of required size is called a sample of SRSWR. If we do
not replace the first slip which has already been drawn in the bag for the
26 subsequent draws then it is called SRSWOR.
The above method is very popular in lottery draws where a decision about Simple Random Sampling
prizes is to be made. However, while adopting lottery method, it is absolutely
essential to see that the slips are as homogeneous as possible in terms of size,
shape and color, otherwise there is a lot of possibility of personal prejudice
and bias affects in the result.
2.2.2 Random Number Method
The lottery method, discussed above, become quite cumbersome to use if the
size of the population is very large. An alternative method of random selection
is that of using the table of random numbers. The most practical and
inexpensive method of selecting a random sample consists in the use of
‘Random number tables’ which have been so constructed that each of the
digits 0, 1, 2, … , 9 appears with approximately the same frequency and
independently.
If we have to select a sample from a population of size N (≤ 99) then the
numbers can be combined two by two to give pairs from 00 to 99. The method
of drawing the random sample consists in the following steps:
1. Identify the N units in the population with the numbers from 1 to N;
2. Select at random, any page of the random number tables and pick up the
numbers in any row or column or diagonal at random and discart the
number which is greater than N; and
3. The population units corresponding to the numbers selected in step-2,
constitute the random sample.
The sample may be selected with replacement or without replacement. In the
case of sampling with replacement, a number occurring more than once is
accepted. A unit is repeated as many times as a random number occurs. But in
the case of sampling without replacement, if a number in random number
table or remainder occurs more than once is omitted at any sequent stage. In
the above selection procedure numbering of units from 00 onwards and
making use of remainders has an advantage, as no random numbers is being
wasted during the selection procedure. This saves time and labor. For example
a population consists of 20 units and a sample of size 5 is to be selected from
this population. Since 20 is a two digit figure, unit are numbered as 00, 01…
19. Five random numbers are obtained from a two digit random number table.
They are given as 85, 63, 52, 34, and 46. On dividing 85 by 20 the remainder
is 5, hence select the unit on serial no. 5. Similarly, dividing 63, 52, 34 and 46
by 20, the respectively remainder are 3, 12, 14 and 6. Hence selected units are
at serial numbers 05, 03, 12, 14 and 06. These selected units constitute the
sample.
A number of random number tables are available such as:
1. Tippet’s Random Number Tables
These tables consist of 10,400 four digits numbers, giving in all 10,400 
4 i.e. 41,600 digits.

2. Fisher and Yate’s Random Number Tables

These tables consist of 15000 digits arranged in two digits numbers.

3. Kendall and Babington Smith’s Random Tables

These tables consist of 1,00,000 digits grouped into 25,000 sets of 4 digits
random numbers.
27
Statistical Techniques

4. Rand Corporation Random Number Tables

These tables consist of one million random digits consisting of 2,00,000
random numbers of 5 digits each.
2.2.3 Computer Random Number Generation Method
The main disadvantage of random number method is that if we want to draw a
sample of large size from the target population then it will take a long time to
draw random numbers from the random number table. Thus to save time and
energy in generation of large numbers of a random numbers one can opt
computer random numbers generation method. Many kinds of methods have
been used for generation of random number from computer but we shall not
discuss all of them here and describe only linear congruential generation
method.

Linear Congruential Generation Method

Linear congruential method can take many different forms but the most
commonly used form is defined by
zi = (azi−1+ c) mod m for i =1, 2, …
where, a, c and zi are to be in the range (0, 1, 2,…, m-1) and integers

a - Multiplier Integer

c - Shift or Increment Integer

m - Modulus
Here, (n) mod m means remainder term when n is divided by m.

Suppose we want to draw a sample of size 4 from the population of size 8.

Take m = 8, a = 5, c =7 and z0 = 4 the resulting sequence of random numbers
is calculated as

z1 = (5  4 + 7) mod 8 = 3

z2 = (5  3 + 7) mod 8 = 6

z3 = (5  6 + 7) mod 8 = 5

z4 = (5  5 + 7) mod 8 = 0

For above calculation we may also make a computer program which is

beyond this course. We are, therefore, not going to discuss it here.

2.3 PROPERTIES OF SIMPLE RANDOM

SAMPLING
Terminologies
N = Population size
n = Sample size
Xi = Value of the character under study for the ith unit in the
population
28
 xi = Value of the character under study for the ith unit in the sample Simple Random Sampling

1 N
X  X i  Population mean
N i1

1 n
x  x i  Sample mean
n i1

1 N
S2   X i  X 2  Population mean square
N  1 i1

1 n
s2   x i  x 2  Sample mean square
n  1 i1

1 N
σ2   X i  X 2  Population variance
N i1

Theorem 1: Prove that the probability of selecting a specified unit of the

population at any given draw is equal to the probability of its being selected at
the first draw.

Proof: In simple random sampling method an equal probability of selection is

assigned to each unit of the population at the first draw.
Thus, in SRS from a population of N units, the probability of drawing any unit
1
at the first draw is , the probability of drawing any unit in the second draw
N
from among the available N  1 is 1 N  1 and so on.

Let, Er be the event that any specified unit is selected at the rth draw.
P (Er) = Prob.{A specific unit is not selected at any one of previous
(r-1) draws and then selected at the rthdraw}
r 1
  P ( It is not selected at i th draw)
i1

 P (It is selected at rth draw that is not selected at the

previous (r-1) draw)
r 1
 1  1
P (Er) =  1  N  i  1   N  r  1
i 1  
r 1 
N i  1
P (Er) =   
i 1  N   i  1  N   r  1

P (Er) =
N  1  N  2  N  3  ... N  r  1  1
N N  1 N  2 N  r  2 N  r  1
1
P (Er) =
N
That means
1
P (Er) = = P( E1)
N
29
Statistical Techniques
Theorem 2: The probability that a specified unit is selected in the sample of
n
size n is
N
Proof: Since a specified unit can be selected in the sample of size n in n
mutually exclusive ways, viz. it can be selected in the sample at the rth draw (r
=1, 2,…, n) and since the probability that it is selected at rth draw is
1
P (Er) = ; r 1, 2, 3, ..., n
N
Therefore, the probability that a specified unit is included in the sample would
be the sum of the probabilities of inclusion in the sample at 1st draw, 2nd draw,
…, nth draw. Thus, by addition theorem of probability, we get
n n
1 n
P (  Er ) = N 
r 1 r 1 N

Theorem 3: The possible numbers of sample of size n from a population of

size N if sampling is done with replacement is Nn.

Proof: The first unit can be drawn from N units in N ways. Similarly, second
unit can also be drawn in N ways because the first selected unit again mixed
with the population. So on up to the selection of nth unit.

Thus, the total number of ways are

N
C1. N C1. N C1... N C1 (n times)

n
  C1   N
N n

Theorem 4: In SRSWOR the sample mean x is an unbiased estimator of

population mean X .

Proof: We have
1 n  1 N 
E  x   E   xi   E  n  a i Xi 
 n i1   i1 
where, ai = 1 if ith unit is included in the sample
0 if ith unit is not included in the sample

Since, ai takes only two values 1 and 0

E a i  =1.P a i 1  0.P (a i  0)

= 1.P (ith unit is included in a sample of size n) + 0.P (ith

unit is not included in the sample)

n  n n
= 1.  0 . 1   
N  N N
Hence,

1 N n 1 N
E( x )  
n i1 N
Xi   Xi  X
N i1
30
Theorem 5: In SRSWR, the sample mean x is an unbiased estimator of Simple Random Sampling
population mean X .

Proof: We have
1 n 
E x   E   x i 
 n i 1 
1 n
  E(x i )
n i1
1 n
 X
n i1
1
 n. X = X
n

Theorem 6: In SRSWOR, the sample mean square is an unbiased estimate of

the population mean square, i.e.
 
E s 2  S2
Proof: We have
1 n 
 x i  x  
2
s2 =
n  1  i1 
2
1  n 2 1 n  
  x i    x i  
n  1  i1 n  i1  

1  n 2 1 n 2 1 n 
  xi   xi   x i x j 
n  1  i1 n i 1 n i  j1 

1  1  n 2 1 n 
  1    x i   x i x j 
n  1  n  i 1 n i  j1 

1 n  1 n x 2  1 n x x

n 1
  i n (n  1) i
n i1  j1
i j

1 n 2 1 n
  i n  n  1 i
n i 1
x 
 j1
xi x j

1  n 2 1  n 
E s 2  E  x i   E   xi x j 
n  i 1  nn  1 i  j1  … (1)
We have
 n  N 
E  x 12   E  a i Xi 2 
 i 1   i1 
N
  E a i  Xi2
i 1 … (2)
th
where, ai = 1 if i unit is included in the sample
0 if ith unit is not included in the sample … (3)
31
Statistical Techniques
Therefore,

n  n N
E  x i2    X i2
 i 1  N i 1 … (4)
 n   N 
and E   x i x j   E   a i a jX i X j 
i  j1  i  j1 
 N 
   E a i a j X i X j 
i j1  … (5)
where, ai and aj are defined in equation (3)
Therefore,
E ai a j   1.P  a i a j 1  0.P  a i a j  0 
 P  a i 1   a j 1 

a 
 P  a i  1 .P  j 1 
 a i  1 
n (n  1)

N ( N  1) … (6)
Because
n
E  a i  1  P i th unit is included in the sample  
N
and
a j  1   jth unit is included in the sample given 
P P
 a i  1 th
 that i unit is included the sample



n 1

N 1
Substituting in equation (5), we get
 n  N
n n  1
E   xi x j   Xi X j
i  j1  i j1 N  1
N
… (7)
Substituting from equations (4) and (7) in equation (1), we get
2 1 N 2 1 N
E s    Xi   Xi X j
N i1 N( N  1) i  j1

1  N 1 N 2 1 N 
  
N  1  N i 1
Xi   Xi X j 
N i  j1 
1  1 N 2 1 N 
  1    Xi   Xi X j 
N  1  N  i 1 N i  j1 
1  N
2 1 N
2
N 
  X i    X i   X i X j 
N  1  i 1 N  i 1 i  j1 
2
1 N 2 1 N  
  X i    X i  
N  1  i 1 N  i 1  
32
 1  N 2  Simple Random Sampling
   X i  NX 2   S 2
N  1  i 1 
E (s2) = S2

Theorem 7: In SRSWOR, the variance of the sample mean is given by

1 1 
Var  x      S2
n N
Proof: We have
2
Var  x   E  x  E  x  
2
 E  x   X2
… (8)
Now
2
1 n 
 2
E x  E  x i 
 n i 1 
1  n 2 n 
 2
E  ix   xi x j 
n  i1 i  j1 
n n
1     
 2 E   x i2   E  x j  
n   i 1   i1   … (9)
From equation (4), we have
n  n n
E  x 2i    x i2
 i1  N i 1
N N

 X i  X    X i  NX
2 2 2
But
i 1 i 1
N N
  X i2   X i  X   NX 2
2

i 1 i1
N
  X i2  N  1S2  NX 2
i 1
Therefore,
 n   N  1  2 2
E   x i2   n  S X 
 i1   N   … (10)
Also from equation (7), we have
 n  n n  1 N
E   x i x j    XiX j
 i  j1  N N  1 i j1
2
n  n  1  N  N
2

   Xi    X i 
N  N  1  i 1  i 1

n  n  1  2 2 2 2

 N X   N  1 S  N X
N  N  1  
n  n  1  N  N  1 X 2   N  1 S2 
  
N  N  1
 2 S2 
 n ( n  1)  X  
 N
… (11)
33
Statistical Techniques
Substituting from equations (10) and (11) in equation (9), we get
1  1 2 2  1  2 S 2
E  x2    1   S  X   n  X  N 
 1 
n  N      
1 2  1 2 1 1 1  1
 X   1   X   1   S2   1   S2
n  n n N N n
1 1 
 X 2     S2
n N ... (12)
Substituting from equation (12) in equation (8), we get
2 1 1  2
Var  x   X     S2  X
n N
1 1 
Var  x      S2
n N

Theorem 8: In SRSWR, variance of sample mean is given by

Var x  
N  1 S2
nN
Proof: We have
1 n
x   xi
n i1
1 n 
Again, Var  
x  Var   xi 
 n i1 
1  n 
 2 Var   x i 
n  i1 
n
1
 2  Var ( x i )
n i 1
Since in case of SRSWR each observation is independent, therefore
1 n
Var x   2  σ 2
n i 1
1 1
 2 n. 2   2
n n … (13)
But
N  2   N  1S2
N 1 2
2  S
N … (14)
Substituting from equation (14) in equation (13) we get
Var x  
N  1 S2
nN
Theorem 9: The variance of the sample mean is more in SRSWR in
comparison to its variance in SRSWOR, i .e.

VarSRSWR  x   VarSRSWOR  x 

Proof: We have
N 1 2
VarSRSWR  x   S
nN
34
 Nn 2 Simple Random Sampling
and VarSRSWOR x   S
nN
Therefore,
VarSRSWR x   VarSRSWOR x  
N  1 S2 
N  n  S 2
nN nN
1 2
 S N  1   N  n ) 
Nn
1 2
 S n  1
Nn
 n 1  2
   S  0
 nN 
That implies VarSRSWR x   VarSRSWOR x 
That means variance of the sample mean is more in SRSWR as compared
with its variance in the case of SRSWOR. In other words SRSWOR provides
a more efficient estimate of sample mean relative to SRSWR.
2.3.1 Merits and Demerits of Simple Random Sampling
Merits
Simple random sampling has the following merits:
1. In simple random sampling each unit of the population has equal chance
to be included in the sample; and
2. Efficiency of the estimates can be found out in simple random sampling
because all the estimates are calculated by using the probability theory.

Demerits
Despite merits, simple random sampling has some demerits too viz.
1. An up-to-date frame of population is required in simple random sampling;
2. Some administrative inconvenience arises in simple random sampling if
some of the units are spreaded in a wide area. So collecting information
from these related units may be problem; and
3. SRS required larger sample size than any other sampling for a fix level of
precision.

Example 1: A population have 7 units 1, 2, 3, 4, 5, 6, 7. Write down all

possible samples of size 2 (without replacement) which can be drawn from the
given population and verify that sample mean is an unbiased estimate of the
population mean. Also calculate its sample variance and verify that

VarSRSWR x   VarSRSWOR x 

Solution: We have
X = 1, 2, 3, 4, 5, 6, 7
1 2  3  4  5  6  7
X 4
7
1 N
S2   X i  X 2
N  1 i1

35
Statistical Techniques
1
  9  4  1  0  1  4  9 
6
28
  4.666
6
All possible samples of size 2 are as follows:

Sample Sample values Sample Mean xX x  X  2

No. x
1 (1,2) 1.5 -2.5 6.25
2 (1,3) 2.0 -2.0 4.00
3 (1,4) 2.5 -1.5 2.25
4 (1,5) 3.0 -1.0 1.00
5 (1,6) 3.5 -0.5 0.25
6 (1,7) 4.0 0 0
7 (2,3) 2.5 -1.5 2.25
8 (2,4) 3.0 -1.0 1.00
9 (2,5) 3.5 0.25
-0.5
10 (2,6) 4.0 0
0
11 (2,7) 4.5 0.25
+0.5
12 (3,4) 3.5 0.25
-0.5
13 (3,5) 4.0 0
0
14 (3,6) 4.5 0.25
+0.5
15 (3,7) 5.0 1.00
+1.0
16 (4,5) 4.5 0.25
+0.5
17 (4,6) 5.0 1.00
+1.0
18 (4,7) 5.5 2.25
+1.5
19 (5,6) 5.5 2.25
+1.5
20 (5,7) 6.0 4.00
+2.0
21 (6,7) 6.5 6.25
+2.5

Total 84.0 35.00

From the table, we have

k

 x  X   35.00
2
x i  84.0 and i
i 1

NC
n
x i
84
E x   i1
N
  4X
Cn 21
N
cn
1 2 1
Var  x   N
Cn
 x
i 1
i X  
21
 35.00 1.667

N 1 2 6
2  S   4.667  4.0008
N 7
36
Verification: In SRSWOR the variance of sample mean is given by Simple Random Sampling

Nn 2 72
Var x   S  4.667  1.667
Nn 7 2

In SRSWR the variance of sample mean is given by

 2 4.0008
Var x     2.0004
n 2

Hence, Var x SRSWR  Var x  SRSWOR

E1) Draw all possible samples of size 2 from the population {2, 3, 4} and
verify that E x   X also find variance.

E2) How many random samples of size 5 can be drawn from a population of
size 10 if sampling is done with replacement?

E3) From a population of 50 units, a random sample of size 10 is drawn

without replacement. From the sample following result are obtained.
n n
2
 x i  48 ,
i 1
x
i 1
i  x   36

Calculate the sample mean and its variance.

E4) Draw all possible samples of size 2 from the population {8, 12, 16} and
verify that
E x   X
and find variance of estimate of the population mean.
E5) From a population of size N=100, a random sample of size 10 is drawn
without replacement. From the sample following results are obtained.
n

 x
i 1
i  x   45
2

Calculate the variance of sample mean.

2.4 SIMPLE RANDOM SAMPLING OF

ATTRIBUTES
A qualitative characteristic which cannot be measured numerically is known
as an attribute i.e. honesty, intelligence, beauty, etc. In many situations, it is
not possible to measure the characteristic under study but possible to classify
the population into various classes according to the attributes under study. For
example, we can divide a population of a colony into two classes only say
literate and illiterate with respect to attribute literacy. Hence the units in the
population can be distributed in these two classes accordingly as it possesses
or does not possess the given attribute. After taking a sample of size n, we
may be interested in estimating the total number of proportion of the defined
attribute.
37
Statistical Techniques
Notations and Terminologies

Let us suppose that a population having N units X1, X2, …, XN is classified

into k mutually disjoint and exhaustive classes. Then

 = The proportion of units processing the given attribute in population A/N

’ = The proportion of units not processing the given attribute in population

A’/N = 1-

Let us consider SRSWOR sample of size n. From this population if ‘a’ is the
number of units in a sample possessing the given attribute then

p = proportion of sampled units possessing the given attribute = a / n

q = proportion of sampled units not possessing the given attribute = a’ /n

Let Xi be the ith unit of the population, where i = 1, 2, …, N.

Then, Xi = 1 if ith unit possesses the given attribute

= 0 if it does not possess the given attribute

Similarly, xi denote ith unit in the sample

Then, xi = 1, if ith sampled unit possesses the given attribute

xi = 0, if ith sampled unit does not possess the given attribute

N

The X
i 1
i = A, the number of units in the population possessing the given
attribute.
n

and  x i = a, the number of sampled units possessing the given attributes.

i 1

N
1 A
Thus, X 
N

i 1
Xi 
N
 

1 n a
and x  
n i1
xi 
n
p

Similarly,
N
2
X
i 1
i  A  Nπ

n
2
and x
i 1
i  a  np

1 N 1 N 2 
S2   X i  X 2
   X i  NX 2 
N  1 i 1 N  1  i 1 
1 N  (1  )
  N   N 2  
38 N 1 N 1
Similarly, Simple Random Sampling

1 n 2 1 n 2 
s2    x i  x      x i  nx 2 
n  1  i1  n  1  i1 

1 npq

n 1
 
np  np 2 
n 1

Theorem 10: Sample proportion p is an unbiased estimate of population

proportion π, i.e. E (p) = 

Proof: We have
1 n a
x  
n i1
xi 
n
p

1 N A
X  
N i 1
Xi 
N
 π

We know that in simple random sampling the sample mean provides on

unbiased estimate of the population mean
E x   X
Therefore E(p) = 

N  n  (1  )
Theorem 11: In SRSWOR, show that Var (p) = .
N 1 n
Proof: We have, Var (p) = Var x 
Nn 2
= S
nN
N  n N. (1  )
 .
n. N N 1
N  n  (1  )
= .
N 1 n

2.5 SAMPLE SIZE FOR SPECIFIC PRECISION

A very first problem faced by a statistician in any sample survey is to
determine the sample size so that the population parameters may be estimated
with a specified precision. The degree of precision can be determined in terms
of
1. The level of significance in the estimate; and
2. The confidence interval within which this estimate lie with respect to
given level of significance.

Let us consider the parameter X the population mean of the population of

size N. We know that x sample mean based on n units is unbiased estimate of
X . Let the difference between estimate value x and the population mean X is
d and level of confidence is (1-  ), then the sample size is determined by the
equation.

39
Statistical Techniques
 
P x  X  d 1  α … (15)

or 
P x  X d    ... (16)

where,  is very small preassinged probability and is nown as the level of

significance..

If n is sufficiently large and we consider SRSWOR, then the statistic

x  E x  xX xX
   
SE x  Var x  1 1
S 
n N … (17)

where, Z is a standard normal variate.

Accordingly, if we take  = 0.05, then we have

P Z  1.96  0.05

 
 xX 
P  1.96  0.05
 1 1 
S n  N 
 

 1 1
P  x  X  1.96  S    0.05
 n N

Comparing with equation (16), we get

1 1
d  1.96  S 
n N

d2 1 1
2 2
 
S 1.96  n N

2
NS2 1.96 3.84 NS2
n 
Nd 2  S2 1.96 
2
3.84S2  Nd 2
… (18)

This formula gives the sample size in SRSWOR for estimating population
mean with confidence level 95 % and margin of error d, provided n is large.

Similarly, if n is small the statistics z follows the student’s t distribution with

(n-1) degree of freedom is given by

xX
t 
1 1 
S   
n N
40
If t  is the critical value of t for (n−1) df and at  level of significance then n Simple Random Sampling

is given by the equation

 1 1 
P  x  X S  . t   
 n N  … (19)

Comparing with equation (16) we get

1 1 
d S    .t
n N

d2 1 1
2 2
 
S t n N

NS 2 t 2 S2 t 2
n 
Nd 2  S 2 t 2  S 2 t 2 
d2  
 N 

2.6 SUMMARY
In this unit, we have discussed:
1. The simple random sampling;
2. The method of SRSWR and SRSWOR;
3. The properties of simple random sampling;
4. Method of finding the variance of the estimate of the sample mean;
5. The simple random sampling for attribute and its properties; and
6. The sample size determination for specific precision.

2.7 SOLUTIONS / ANSWERS

E1) In SRSWOR the number of sample is
N
C n  3C2  3
The samples with their means are as follows:

Sr. No Sample Mean x 

1 2,3 2.5
2 2,4 3
3 3,4 3.5
3

x
i 1
i 9

41
Statistical Techniques N
Cn
1
E x   N x i
Cn i 1

9
 3
3
1 N
X  Xi
N i1
1
 2  3  4 
3
9
 3
3
Therefore,
E x   X
Nn 2
Again V x   S
N.n

1 n 2
S2  
N  1 i 1
X  X
i

1

2

2  32  3  32  4  32 
=1

Therefore,
32
Var x   1
3 2
1
  0.166
6

E2) The first unit can be drawn from 10 units in 10C1 = 10 ways. Since
sampling is done with replacement so the second unit can be drawn in
10
C1 = 10 ways … so on upto the selection of 5th Unit. Thus the total
ways are 10.10.10.10.10 = 105 ways.

E3) We have
n

x
i 1
i  48

1 n 1
x 
n i1
x i   48  4.8
10
So
1 n
s2   x i  x 2
n  1 i1
1
s 2   36  4
9
which is the estimate value of S2.
42
 Therefore, Simple Random Sampling

N  n 2 50  10
Variance x   S  4
Nn 50  10
16
  0.32
50
E4) In SRSWOR the number of samples is NCn = 3C2 = 3 and samples with
their means are

S. No. Sample Mean

x i 
1 (8, 12) 10
2 (8, 16) 12
3 (12, 16) 14
Total 36
3

 x  36 ,
i 1
i

Therefore,
N
Cn
1
E x   N x i
Cn i1

1 3 
  xi 

3  i 1 
1
  36  12
3
8  12  16
Again X  12
3
Therefore,
E x   X
Again estimator of population mean is sample mean and so its
variance
1 1 
Var x      S2
n N
where,
1 N
S2   X i  X 2
N  1 i1
1

3 1

8  122  12  122  16  122 
1
 16  0  16  16
2
Therefore,
1 1
Varx      16
 2 3

43
Statistical Techniques
3 2 16
 16 
 6  6
8
  2.66
3
E5) We have,
1 1 
Var x      S2
n N
and
n
1
s2   x i  x 2
( n  1) i 1
1
  45  5
9
Which is the estimate value of S2.
Therefore,
1 1 
Varx     5
 10 100 
9
  5  0.45
100

44