Skewness And Kurtosis 165

Moments,

length of the class interval.

c isthe
where,corrections are known as Sheppard's corrections.
These
lustration: The moments calculated from a grouped frequency with
class interval 5 are given by m, =54.0, m, =100.5 and m, =
of thecorected
ength
moments
due to
Sheppard will be
8270The

(corected) = m, (uncorrected) 12 =54.05 = 54.0 = 2.1=51.9.

12
m,
(corrected) (uncorrected) = 100.5.
m, 7c
(corrected) (uncorrected) gm (uncorrected) + 240
7x54
= 7827.0 × 54.0 + = 7827.0 675.0 + 18.2.
2 240
= 7170.2

5.14. Worked-out Examples

data:6, 7, 10,
Example 1. Find the first three moments about 8 of the
9, 12.
Solution : We find the moments from the given table :

(x-8) (x- 8)
6 -2 4 -8
7 -1 1 -1
1
10 2 4

12 4 16 64

Total 26 64

No. of variate values (n) = 5.

Therefore,
L«-9 4 =0.8.
First moment about 8= 5 5

26
Second moment about 8= 5 5-=5.2

Third moment about 8= a-8

64
= 12.8.
5

Exampl
data e2,
:
2. Find the first three raw
5, 8, 9.
moment and central moments for
 166 Fundamental Treatise on'Probability and
Solution: First three raw moments are calculated from the Statistics
fol owing table:
2 4 8
5 25 125
8 64 512
81 729
Total 24 174 1374

Therefore,
m, = mean = ==6.
4 4

m, = _174 = 43.5.
4 4

1374E
, -= 343.5.

Now using the formulate (5.10),

m, =0.
m, =n, -(m,)=43.5 -6 = 43.5 36 =7.5.
m, = m,-3m,. m, + 2 (m,)= 343.5-3 x 43.5 x 6+ 2x 6
=-7.5.
Example 3. The first three moments of a distribution about the value 3 of
the variable are 2, 10and 30 respectively. Obtain the first three moments about
zero. Show also that the variance of the distribution is 6.
[L.C. WA, Jan. 1964]
Solution: If xdenotes the variable, then it is given E(-3)/n=2,
E(X-3)/n = 10 and 2(x-3)/n = 30 and we are calculate Ex/n, Lxln,
Exn and o 2 :
n

Now, L(*-3)/ n=2 gives,

Or, =2 or,

Or,

L«-3)? = 10 gives
L-6x+9) = 10
 skewness And Kurtoeie
MOments, 167
9n
= 10
Or, .-6x5+9= 10
or,

=31.

E(- 3)/n = 30 gives

Sr-3.*".3+3.x.9-27)
=30

n
-+27227n =30

-9x31+27x5-27 =30 Or,

n 20l
Thus, =201.
n un

Also, variance = =31-25 =6.

n

Example 4. Findthe first, the second and the third central moments of the
frequency distribution of expenditure (Rs. per months) given below :
Expenditure : 3-6 6-9 9-12 12-15 15-18 18-21 21-24 Total
No. of families: 28 292 -389 212 59 18 2 1000
[I.C.W.A. June, 1978]
Solution : to calculate the central moments, we first calculate the
corresponding raw moments with the help of the following table
X-13.5
Class Class frequency y=
3
interval mark
(«)
3-6 4.5 28 -3 -84 252 -756
6-9 292 -2 584 1168 2336
7.5
9-12 10.5 389 -1 389 389 -389
12- 15 13.5 212
15 - 18 59 1 59 59 59
16.5
36 72 144
18-21 19.5 18 2

|21-24 22.5 2 3 6 18 54
956 1958 3224
Total 1000

-956-=-0.956.
Thus, mí (y) = 1000
168 Fundamental Treatise on Probability and
m, (y) =.1958 =1.958.
Statistics
1000

m(y) = -3224
1000
=-3.224.

Hence, m, (1) = 3 x m' (y) =- 2.868

m, («) = 32 x m,' () = 17.622
m, (r) = 3 x m, () =-87.048.
Now, by (5. 10), the coresponding central moments are given by
m, = 0.
m, = m, - (m,') = 17.622 - (- 2.868)? = 9.44 Rs?
m, = m- 3m,m' + 2(m'
=-87.048 3 x 17.622 x (- 2.868) + 2(- 2.868)9
=-87.048 + 151.620 47.181 = 17.39 Rs. 3,
5. Find out the skewness and kurtosis of the series by the method of
ments.

Measurement 0-10 10-20 20-30 30-40

Frequency 1 3 4 2
[VU. (Pass) Math. 2002]
Solution:
Class Mid Frequencyxf- z(-$f\(-Ff(-I'f
interýalvalue (x)
0-10 5 1 5- 17 289 4913 83521
10-20 15 3 45 147 1029 7203
20-30 25 4 100 3 36 108 324
30-40 35 2 70 13 338 4394 57122
Total 10 220 810 1440 148170

f: 220
= 22 = m,
Lf 10

m, =
Ef(x-z)_810 = 81
10

m, =
Ef(x-z} 810 =- 144
10

m, = Ef(r-) 148170 = 14817

10
Moments, Skewness And Kurtosis 169

B.="l44) =0.039
Now, (81)'
14817
B,= =2.258
m (81'
Therefore, Skewness (Y,)=/B, = 0.198
Kurtosis (Y,)= b,- 3 = - 0.742,.
Tafreguency distribution, the first four central moments are :0.0, 2.5,
07and 18.75
respectively. Findthe skewness and kurtosis and hence explain
the
nature of the distribution. [VU. (Pass) Math, 2005]
Solution:

B, = m_(0.7) =0.031
m (2.5)
18.75
3
n (2.5)
Y=B, = 0:177
Y B2r3=0
So, skewness = 0-177, kurtosis = 0 and the distribution is positively skewed
and mesokurtic.
Example 7. Compute the coefficient of skewness based on quartiles from
the following data:
10 20 30 40 50 60
Age (under year) :
No. of Persons: 15 32 51 78 97 109
[B.U. (Pass) Stat. 2000]
Solution: To calculate the quartiles we consider in following table :
Class interval frequency CE. <
0-10 15 15
10-20 (32 15=) 17 32 ’ Q, class
20-30 (5132=) 19 51
30 40 (78 51=)27 78 ’ Q, class
40- 50 (97-78 =) 19 97 ’ Q, class
(109 97 =) 12 109
50- 60
Total N= 109

N 109 3N 3x109
n 109 = 81.75.
We see, 4 4
= 27.25, , 2
= 54.5, 4 4
 170 Fundamental Treatise on Probability and
So, the quartile classes will be as shown in the table. Statistics
N
--F
4
Now, Q, = First quartile = l + -XC

= 10 +
27.25-15,x10
17
= 17.206.

2
, = Second quartile = Median = l + -XC

54.25-15
= 30 + -x10
17
= 31.296.

4
, =Third quartile = l + -XC
f
81.75-78
= 40 + -x10
19
= 41.974.
To calculate coefficient of skewness based on quartiles, we use Bowle's
measure. So using (5.21),
Skewness = -20, +l_41.974-2x31.296 +17.206
,-2 41.974-17.206
=-0.14
Since coefficient is negative., so the distribution is skewed to the left.
Example 8. For a distribution the mean is 10, variance is 16, Y, is +1 and
B, is 4. Find third and fourth moments about the origin zero. (Y, and B, have
the standard meanings.) [B.U. (Hons.) Econ. 2002]
Solution : For the given distribution let, m,, m,, m,, m, denote tne
first, second, third and fourth central moment and m, m,, m, m be the
corresponding moments about the origin zero.
Now, m, = 10, m, = 16.
Also, Y,=1 i.e., B,= (Y, =(1}=1
But,
m
or, l=
m,
(16)
or, mi= 16 x 16x 16
or, m, = 64.