Statistics

MEASURES OF CENTRAL TENDENCY, DISPERSION

One of the most important objectives of statistical analysis is to get one single value that describes the characteristic
of the entire data. Such a value is called the central value or an average.
The following are the important types of averages:
1. Arithmetic Mean
2. Geometric Mean
3. Harmonic mean
4. Median
5. Mode
We consider these measures in three cases (i) Individual series (i.e. each individual observation is given) (ii)
discrete series (i.e the observations along with number of times a particular observation called the frequency is
given) (iii) continuous series (i.e. the class intervals along with their frequencies are given)
ARITHMETIC MEAN

(i) Individual Series : If x1 , x 2 .........xn are the values of the variable x , then the arithmetic mean usually
denoted by x is given by

x1  x2 .....  xn 1 n
x   xi
n n i 1

(ii) Discrete Series : If a variable takes values x1 , x 2 .........xn with corresponding frequencies

f1, f2 ….. fn then the arithmetic mean x is given by

n n
f1 x1  f 2 x2 .....  f n xn 1
x
f1  f 2 ....  f n

N

i 1
fi xi , where N   fi
i 1

(iii) Continuous Series : In case of a set of data with class intervals, we cannot find the exact value of the mean
because we do not know the exact values of the variables. We, therefore, try to obtain an approximate
value of the mean. The method of approximate is to replace all the observed values belonging to a class by
mid-value of the class. If x1, x2 … xn are the mid values of the class intervals having corresponding
frequencies f1, f2 … fn then we apply the same formula as in discrete series.
n n
1
x
N

i 1
fi xi , N   fi
i 1

PROPERTIES OF ARITHMETIC MEAN

(i) Sum of all the deviations from arithmetic mean is zero i.e.,
n

  x  x   0 (in case of individual series)

i 1
i

 f  x  x  0
i 1
i i (in case of discrete or continuous series)

(ii) If each observation is increased or decreased by a given constant K, the mean is also increased or decreased
by K

[1]
[2] Statistics

The property is also known as effect of change of origin. K can be taken to be any number. However, to
simplify the calculations, K should be taken as a value which is in the middle of the table.
(iii) Step Deviation Method or change of scale
If x1 , x2 .............xn are mid values of class intervals with corresponding frequencies f1, f2,…., fn then we
xi  A
may change the scale by taking di  , in this case.
h

1 
x  A  h
N
 f d  i i

A and h can be any numbers but if the lengths of class intervals are equal then h may be taken as width of
the class interval.
In particular if each observation is multiplied or divided by a constant, the mean is also multiplied or divided
by the same constant.
(iv) The sum of the squared deviation of the variate from their mean is minimum i.e., the quantity
2 2
  x  A
i or  fi  xi  A  is minimum when A  x

(v) If x i and x 2 be the means of two related groups having n1 and n2 items respectively then the combined

n1 x1  n2 x 2
mean x of both the groups is given by x 
n1  n2

WEIGHTED ARITHMETIC MEAN

If x1, x2, ....., xn are n vlaues of a variable X and w1 , w 2 ,....., w n denote respectively their weights, then their
weighted mean X w is given by
n

w1 x1  w 2 x 2  ....  w n x n w x
i 1
i i
Xw   n
w1  w 2  ......  w n
w
i 1
i

GEOMETRIC MEAN

In case of individual series x1 , x2 .............xn

G.M. = ( x1 , x2 .............xn )1/n

In case of discrete or continuous series
n
1/ N

G.M.  x x ....x f1 f 2
1 2 n
fn
 , where N   fi
i 1

HARMONIC MEAN
The harmonic mean is based on the reciprocals of the value of the variable

1 1 1 n 1
H.M. =   (Incase of Individual series)
1 1 1 1  or H n i 1 xi
   ....  
n  xi x2 xn 
Statistics [3]

1 1
and  n
(in case of discrete series or continuous series)
H 1 1
N

i 1
fi
xi
If x1, x2, …xn > 0 then it is known that A.M. > G.M. > H.M.
MEDIAN
It refers to the middle value in a distribution.
In case of individual series, in order to find median, arrange the data in ascending or descending order of magnitude.
In case of odd number of values
n 1 n n2
Median = size of th item. In case of even number of values Median = average of th and th
2 2 2
observation.
In case of discrete frequency distribution the median is obtained by considering the cumulative frequency (c.f.).
n
N
Find
2
, where N  
i 1
fi . Find the cumulative frequency (c.f.) just more than N/2. The corresponding value

of x is median.
In case of continuous distribution, the class corresponding to c.f. just more than N/2 is called the median class and
hN 
the median is obtained by Median = l   C
f 2 
Where l  the lower limit of the median class; f  the frequency of the median class; h  the width of the
n

median class; C  the c.f. of the class preceding to the median class and N   fi
i 1

MODE
The mode is that value in a series of observations which occurs with greatest frequency. In case of individual
series, for determining the mode, count the number of times the various values repeat themselves and the value
occurring the maximum number of times in the model value.
In case of discrete series, quite often mode can be determined just by inspection i.e. by looking to that value of
variable around which the items are most heavily concentrated.
In case of continuous series,
f1  f 0
Mode  l  h
2 f1  f 2  f 0
Where l  the lower limit of the modal class i.e. the class having maximum frequency; f1 = frequency of the
modal class; f0= frequency of the class preceding the modal class; f2 = frequency of the class succeeding the
modal class and h = width of the modal class.
When mode is ill-defined i.e. there are two or more values which occur with equal maximum frequency, the
mode can be computed by
Mode = 3 median - 2 mean
[4] Statistics

DISPERSION
Literally, dispersion means ‘scatteredness’. Dispersion measures the degree of scatteredness of the variable
about a central value. Different measures of dispersion are
1. Range
2. Mean-deviation
3. Quartile deviation
4. Standard deviation
1. Range
The range is the difference between the largest and smallest observation.
2. Mean-deviation
If x1 , x2 ............xn are n observations then mean deviation about a point A is given by

1
M.D. 
n
 | xi  A |
In case of discrete or continuous series
n
1
M.D. 
N
 f i | xi  A |, N   fi
i 1

M.D. is least when taken from the median

Standard Deviation

Variance  2 in case of individual series is given by

2
1 n 2 1 n 2 1 n 
2
   xi  x
n i 1
    xi    xi 
n i 1  n i 1 
If x1, x2, ……, xn occur with frequency f1, f2….fn respectively then
n 2
1 2 1 1 
 2 (variance) 
N
 x  x
i 1
i 
N
 2
f x    f i xi 
i i
n 
and standard deviation =+ var iance
There is no effect of change of origin on standard deviation
2
2
1 2 21  
 h 
x  fi d  
i  fi di  
 N N  

If there are two samples of sizes n1 and n2 with x1 and x2 as their means  1 and  2 their
standard deviations respectively, then the combined variance is given by

2
n1  12  d12   n2  22  d 22 
 
n1  n2
where d1  x1  x and d 2  x2  x , x being the combined mean.
Statistics [5]

COEFFICIENT OF VARIATION


  100
X

Properties of Variance :

1. Let x1 , x 2 , x 3 ,....., x n be n values of a variable X. If these values are changed to x1 + a, x2 + a, ...., xn + a,

where a  R , then the variance remains unchanged.
2. Let x1, x2, ...., xn values of a variable X and let ‘a’ be a non-zero real number. Then, the variance of the
observations ax1, ax2, ...., axn is a2 Var (X).
3. Let x1, x2, x3, .....,xn be n values of a variable X, and let xi = a + hui, i = 1,2, ......,n, where u1, u2,.....,un are the
values of variable U. Then, Var (X) = h2 Var (U), h  0