MAT2001-Statistics for Engineers

MOMENTS, MEASURES OF SKEWNESS AND KURTOSIS

MOMENTS:

The word “Moment” is a concept in the Quantum Mechanics of physical sciences. It

refers to the measures of force with respect to its tendency to produce rotation. The strength of
this tendency to produce rotation depends upon the amount of force and the distance from the
origin of the point at which the force is exercised. In statistics, the concept of “Moments”
describes the characteristics of the given frequency distribution such as arithmetic mean,
variance skewness and kurtosis It is used in general sense, that the class frequencies are the
forces and the deviations of the observations from either mean or any arbitrary value are the
distances. The moment contribution is measured by the sum of products of the class frequencies
and the corresponding deviation from either mean or any arbitrary value, raised to an appropriate
power.

Since arithmetic mean may be considered as the centre of gravity of the distribution, the first
moment about the mean will be zero. The second moment about the mean equals to the variance
of the distribution.

Moments are of two types.

(i)Central Moments: The moments obtained by using the deviations of the observations from
their arithmetic mean are called ‘Central Moments’ or “Moments about the arithmetic mean”.
The central moments of different orders may be 1 ,  2 , 3 ,  4 etc.,

(ii)Raw Moments: The moments computed by using the deviations of the observations from any
arbitrary point ‘A’ are called Raw Moments. These are usually denoted by 1 , 2 , 3 ,  4 etc.,
1 1 3 1

Central Moments or Moments About Mean:

We shall present the central moments of two types namely, (i) raw data or ungrouped data
and (ii) frequency distribution in the following table.

[Link] Rayalu,[Link],Ph.D.,
 Ungrouped Data Frequency Distribution
First Central Moment :

1 
 X  X  0 1 
 f  X X 0
n N
Second Central Moment
 X  X   f  X X
2 2

2    2  variance 2  2
n N
Third Central Moment
 X  X   f  X X
3 3

3  3 
n N
Fourth Central Moment:
 f  X X
4

 X  X  4
4

4  N
n

Here, X  X X 
 fX ,
n N
N= f  Total frequency

In general for a frequency distribution the rth central moment of a

variable X is defined as
 f X X  fX
r

r  , where X 
N N
r = 0, 1, 2, 3, 4, …..
 f  X X
0

For r = 0, we have,   
f  1;
0
N N

r = 1 , we have    f X X
0
1
N
[ The algebraic sum of deviations of observations from mean is zero]
 f  X X
2

and for r = 2, we have     2  Variance

2
N

Raw Moments or Moments about any arbitrary point ‘A’

For a frequency distribution the rth raw moment of a variable X about any arbitrary point
‘A’ is defined as
 f  X  A
r

r 
1
, r = 0, 1, 2, 3, 4, …..
N
where N   f  Total frequency

[Link] Rayalu,[Link],Ph.D.,
  f  X  A f
0

For r = 0, we have,  01   1
N N
For r = 1, we have 11   f  X  A

 fX 
A f

 fX  A.
N N N N
Or X  A   1
1

Here, 11 is the first raw moment about any arbitrary point ‘A’. The raw moments of
different order for ungrouped data and frequency distribution are given in the following
table;
Moments about any arbitrary point ‘A’:

Ungrouped Data Frequency Distribution

First raw Moment about A:

11   11  
 X  A f  X  A
n N
Second raw Moment about A:
  X  A  f  X  A
2 2

2 
1
 1
2  2
n N
Third raw Moment about A:
  X  A   X  A
3 3

3 
1
 1
3 
n N
Fourth raw Moment about A:
  X  A   X  A
4 4

4 
1
 1
4 
n N

Moments About origin:

The rth order raw moment about origin can be obtained by the substitution A = 0 in the definition
of rth raw moment about any arbitrary point ‘A’.

i.e.,  r 
1  f  X  0
r

or  r 
1  f Xr
, r = 0, 1, 2, 3, 4, …..
N N
Raw Moments about origin
Ungrouped Data Frequency Distribution
First Moment about origin:

11    Mean 11  

X f X
 Mean
n N
Second Moment about origin:
 
2 2
X f X
2 
1
2 
1

n N
Third Moment about origin:

[Link] Rayalu,[Link],Ph.D.,
 X f X
3 3

 1
3   1
3 
n N
Fourth Moment about origin:
 f X
4 4
X
4 
1
 1
4 
n N

CONCEPT OF SKEWNESS (ASYMMETRY)

Skewness refers to ‘Lack of symmetry’ in the data. For a symmetric distribution mean,
median and mode are equal.

If Sk = 0, then the frequency curve is symmetric or a normal curve.

If A.M. > Mode (or) A.M. > Median, then the skewness is said to be positive and the curve
is positively skewed. In this case Sk > 0.
If A.M. < Mode (or) A.M. < Median, then the skewness is said to be negative and the curve
is a negatively skewed curve. In this case Sk < 0.

Symmetric distribution:

- 
Positively skewed distribution

Negatively skewed distribution

- 

[Link] Rayalu,[Link],Ph.D.,
 1. Bowley’s Coefficient of skewness:

Concept of Kurtosis

The concept of Kurtosis was introduced by Prof. Karl Pearson in 1905. It refers to the
shape of the frequency curve of a frequency distribution. A measure of Kurtosis together

- 
with the measures of central tendency, dispersion and skewness describe completely the
characteristics of the frequency distribution. Kurtosis provides an idea about the
peakedness or convexity of the frequency curve. This concept has been frequently used in
biological research studies rather than sociological studies.
There are mainly three types of Kurtic curves namely :
(i)Leptokurtic curve
(ii) Mesokurtic curve or normal curve and
(iii) Platykurtic curve

The three types of kurtic curves are shown below

Curve (1) exhibits more peakedness than the two curves. It is called leptokurticcurve.
Curve (2), which is neither peaked nor flat is called mesolurtic curve or normal curve.
Curve (3) is more flatter than the normal curve. It is called platy kurtic curve

[Link] Rayalu,[Link],Ph.D.,
Solved problems on moments

Problem: Calculate the first four moments about the mean for the following wage distribution:
Daily wages (Rs) 30-40 40-50 50-60 60-70 70-80 80-90
No. of workers 2 5 8 20 8 7

Solution:
Frequency f Mid value x x A
Class d fd fd2 fd3 fd4
C
30-40 2 35 -3 -6 18 -54 162
40-50 5 45 -2 -10 20 -40 80
50-60 8 55 -1 -8 8 -8 8
60-70 20 65 = A 0 0 0 0 0
70-80 8 75 1 8 8 8 8
80-90 7 85 2 14 28 28 112
Total 50 - - -2 82 -38 370

From the table, we have, N=50,  fd =-2;  fd =82;  fd =38,  fd =370

2 3 4

Since, the moments are not independent of change of scale, first four raw moments are
given by,

1'     C    10  0.40
 fd   2 
 N   50 

 2'   
 fd 2  2  82 
 C    100  164
 N   50 
  fd 3   38 
1  
'
 C3   1000  760
 N   50 

1'   
 fd 4  4  370 
C   10000  74000
 N   50 

The first four central moments about arithmetic mean are given by:
2
1  0;  2   21  1'
3
3  31  3 21 1'  2 1'
2 4
 4   14  4311'  6  211'  31'
 1  0
 2  164  (0.40) 2  163.84
3  760  3(164)(0.40)  2(0.40)3  563.328

[Link] Rayalu,[Link],Ph.D.,
  4  74000  4(760)(0.40)  6(164)(0.40) 2  3(0.40) 4  75941.363

Since mean is not independent of change of origin it is given by

X  A  1'  65  ( 0.40)  64.6
Since, variance is independent of change of origin but not scale, it is given by
 2   2  163.84

Problem: The first three moments of a distribution about the value 2 of a variable are 1, 16 and
–40 respectively. Obtain the first three moments about mean and origin. Also find mean
and standard deviation.

Given, A=2, moments about ‘A’ are : 1  1,  2  16and 3  40

' ' '
Solution :
Moments about arithmetic mean are:
1  0
2
 2   21   2'  16  (1) 2  15
3
3  31  321 1'  21'  40  3(16)(1)  2(1)3
 40  48  2  86
The mean and variance are given by, mean = X  A  1'  2  (1)  3
Variance =  2   2  15
Moments about origin X=0
First moment about origin = Mean = 1' =3
Second moment about origin  21 =  2  1'  15  (3) 2  24
2

Third moment about origin 3 = 3  3m2 1'  1'

1 3

 (86)  3(15)(3)  (3)3

 86  135  27
 31  76

Problem: The first four moments of a distribution about the value 5 of a variable are 2, 20, 40
and 50. Show that the mean is 7, variance 16, 3 = -64 and  4 =162. Find moments
about the point x=2. Also find 1 and 2.

Solution:- Given, A = 5 and moments about A are:

1'  2, 21  20, 31  40and  14  50
Mean and variance are given by:
Mean = X  A  1'  5  2  7
2
Variance =  2   21   2'  20  (2) 2  16
Central moments are given by:
1  0
 2  16

[Link] Rayalu,[Link],Ph.D.,
 3
3  31  3 21 1'  2 1'  40  3(20)(2)  (2)3
= 40 – 120 + 16
 3  64
2 4
 4   14  4311'  6 21 1'  31'
 50  4(40)(2)  6(20)(2) 2  3(2) 4
 50  320  480  48
  4  162

Moments about the point x = 2:

Since X  A  1' , we have 1'  X  A

Putting A = 2, the moments about X=2 are given by
1  X  A  7  2  5
'

2
 12   2   2'  16  (5) 2  16  25  41
3
31  3  3 2 1'  1'
 64  3(16)(5)  (5)3
 64  240  125
 31  301
2 4
 14  4  43 1'  6 2 1'  1'
 162  4(64)(5)  6(16)(5) 2  (5) 4
 162  1280  2400  625
  4  1907
1

[Link] Rayalu,[Link],Ph.D.,