Sunteți pe pagina 1din 18

- MEASURES OF DISPERSION I -

Structure
14.0 abjectives
14.1 Introduction
14.2 What is Dispersion?
14.3 Significance of Measuring Dispersion
14.4 Properties of a Good Measure of Dispersion
14.5 Absolute and Relative Measures of Dispersion
14:6 Measures of Dispersion
. 14.7 Range
14.8 Quqrtile ~ e v i a h o n
14.9 Mean Deviation
14.10 Let Us Sum U p
14.11 Key Words and Symbols
14.12 .Answers to Check Your Progress
14.13 Terminal QuestionsExercises

After studying this unit, you should be able to :


9 explain the concept of dispersion and significance of measuring it,
e differentiate between absolute and relative measures of variation,
e compute several measures of dispersion such as the range, quartile deviation and
mean deviation for different types of data, and
decide the use of appropriate measures under different situations.
- L

14.1 INTRODUCTION
In Un~ts10 to. 13 you have studied about different measures of central tendency. But
central tendency alone is not sufficient to analyse the data. For more meaningful
analysis of the data. it is necessary to study disrjersion i.e., the spread of the data or
the extent ti, which items deviate from central tendency . In this unit, you will study
the meaning -and significance of disperdon. You will also learn in detail about the
three measures of dispersions viz., r a k e , quartile deviation and'mean deviation.

14.2 WHAT IS DISPERSION?


The word, dispersion is used to denote the degree of heterogeneity in the data. It is
an important characteristic indicating the extent to which observations vary among .
themselves. A measure of dispersion dekcribes the spread or scattering of individual
values around thc central valuc. To understand the concept of dispersion clearly.
study Illustration 1 carefully.

h l l y Sales of Thm MnTerent Firms (in Rs.)

Firm A Flrm B Flrm C

CI(l.IWK) 62,500 5 1 ,(100


M.UK) f10.Oal 32.MW)
clo.kw, 52.250 22,1100
CI(l,O(K) 56.5(K) lR,(WKl
Measures of Dispersion and Since the average sales of firms A, B , and C are the same, we are likely t o conclude
Skewncs
that the three distributions of the sales are similar. But you should note that the
variations in the sales are different from firm to firm. Dailysales are the same for all
the days in the case of Firm A whereas there is some variation in the daily sales o i
Firm B and greater amount of variation for Firm C. Here, although these three data
sets have the same mean, they differ in terms of scatter of items. Therefore, different
sets of data may have the same measure of central tendency, but may differ greatly
in terms of spread or scatter of the items i.e. dispersion.
T h e word dispersion can be interpreted in another sense also. When all items of the
data are not equal to central tendency, then the various items differ from central
tendency by a certain amount. Dispersion gives, on an average, by how much amount
items differ from central tendency. You may note that in the case of Firm B,
deviations of individual sales from the mean sale (i.e., 60,000) are much smaller than
the deviations of Firm C. This implies that the average of the deviations from the
mcan sales will be smaller for Firm B compared to Firm C. In other words. Firm B
has smaller dispersion than Firm C.

14.3 SIGNIFICANCE OF MEASURING DISPERSION -


Measures of variations (dispersion) are calculated to serve the following purposes :
1 Measuring variability determines the reliability of an avarage by pointing out to
what extent the average is representative of the entire data. In Illustration 1
discussed earlier, mean sales Rs. 60,000 is the perfect representative of salcs for
different days tor Firm A. In case of Firm B , the variation is low as the mean
sale is quite close to sales figures of different days. Therefore, in this case, the
mcan can be considered as representative of the sales for each day. But in case
of Firm C the variation in individual figures is very large so the average of
Rs. 60,000 is hardly a representative of all high' and low figures such as
Rs. 2,10,000 and Rs. 18,000. ,
2 Measures of dispersion enable comparisons of two or more distributions with
regard to their variability.
3 Another purpose of measuring variability is to determine the nature and cause I
of variation in order to control the variation itself.
4 Measuring variability facilitates the use of other statistical measures like
correlation, regression, statistical inference, ctc.

14.4 PROPERTIES OF A GOOD MEASURE OF


DISPERSION ..
As you know, a measure of dispersion is the average of the deviations of items from
its mean i.e., it is an average of second order. Hence, it should also possess all the
qualities of a good measure of an average. According to Yule and Kcndall the
qualities of a good measure of dispersion are as follows :
1 Statistical measures are used even by layman. So complicated definitions and
calculations are not desirable. It should be simple to understand and easy to.
calculate.
2 It should be rigidly defined. For the same data, all the methodsshould produce
the same answer. Different methods of computation leading to different answers
is not proper.
3 It should be based on all items. Wherl it is based on all items, it will produce a
more representative value. Thus, good measure of dispersion should be based
on the entire data.
4 It should be amenable t o further algebric treatment. This means combining
groups, calculations of missing values, adjustment for wrong entries, etc., which
are possible.without the knowledge of actual values of all items. Such treatment
6 '
should be possible with 2 good measure of dispersion also.
Measures of I)ispersio~ls.l
; It should have sampling stability. I t means that the average difference t ; ; , v e ~ n
[he values obtained from the sample and the corresponding values from the
population should be the least. If it is so far a measure of dispersion, it is the
best Yleasurc.
6 It should not be unduly affected by the extreme items. Extreme items, many
times, are not true representatives of the data. So their presence should not affect
a
the 'calculation to large extent.
This list is not a complete-list of the properties of a good measure of dispersion.
But these are the most important characteristics which a good measure of
dispersion should possess.

14.5 ABSOLUTE AND RELATIVE MEASURES OF


DISPERSION
- -

The measure of dispersion which are expressed in terms of the original units of data
areatermedas Absolute Measures. For example, in Illustration 1 discussed earlier, the
daiiy sales of the Firm R range between Rs. 52,250 to Rs. 68,250. So the spread of
the data is of the order Rs. 68,25&52,250 or Rs. 16,000. This is the absolute measure
of the spread of the sales. Such measures expressed in units of data are not suitable
for comparing the vari'ability of the distributions or series expressed in different units
of measurement. Relative Measures of dispersion, on the other hand, are obtained as
ratios or percentages. Therefore, relative measures are pure numbers independent of
the unit of measurement. A measure of relative dispersion is the ratio of a measure
of absolute dispersion to an appropriate average or the selected items of the data.
Hence, it is also known as Coeficier~tof Dispersion. For example, in lllustration 1
discussed earlier, if one expresses the spread Rs. 16,000 as the ratio of average sales
Rs. hO.00 i.e., 16,000160,000 i t becomes a relative measure. 'This value is a simple
number and has no specific units of measurement with it. Similarly, the spread
Rs. 16.000 could also be expressed as the ratio of sum of two extreme sales i.c.,
1h ,000 This will also give a relative measure of the spread of the sales.
+
52.250 68.150'
Sometimes, even when data are in the same units, the comparison of variation by
absoldte measure of variation is not worth comparing. A variation of one kilometer
(1,00,000 ern) in measuring distance from Delhi to Bombay is hardly of any
significance. But a variation of 10 cni in measuring a piece of cloth of 1.40 meters is
of very great significance. So wl~cncvcrcomparisons of vi~riabilityin two scts of data
are donc, it is always donc in terms of relative measures.
- - - - -

Check Your Progress A


1 What is thc meaning of the term Dispersion'?

........................................................................................................
Differentiate bctwccn absolute measures and rclative mcasures of dispersion.

........................................................................................................
State whether the following statements are True or.False.
i) Variation mcans only the spread of itcms of a data set.
ii) The only purpose of measuring dispcrsion is to assess the reliability o i an
average,
iii) Variability itself may be some timcs uscd in dccision making.
iv) Comparison of two sets of data can only be done by relntivc measures of
dispersion.
Measures of Dispersion and v) All absolute measures of dispersion are pure numbers. ,
Skewness
vi) A good measure of dispersion should preferably use all the items of the data.

14.6 MEASURES OF DISPERSION


The following measures of absolute dispersion are in common use :
1 Based on selected items of the data
i) Range - spread for entire data
ii) Inter Quartile Range - spread for middle 50% data. More commonly
Quartile Deviation is used in its place, which is half of inter quartile range.
2 Based on all items of the data
i) Mean Deviation - mean of the absolute deviations from central tendency.
ii) Standard Deviation or Root Mean Square Deviation about arithmetic mean
3 A Graphic Method - Lorenz Curve.
The relative measures of dispersion corresponding to the measures of'absolute
dispersion are :

Absolute Measures of Dispersion Relative Measures of Dispersion

i) Range Coefficient of Range


ii) Quartile Deviation Coefficient of Quartile Deviation
iii) Mean Deviation Coefficient of Mean Deviation
iv) Standard Deviation Coefficient of Standard Deviation

Coefficient of standard deviation when expressed in percentages is called coefficient


of variation.
Study Figure 14.1 carefully for classification of measures of dispersion. You will study
Range, Quartile Deviation and Mean Deviation later in this unit, and Standard
Deviation and Lorenz Curve in Unit 15.
Figure 14.1

Classification of Measures of Dispersion

I Measures of Dispersion
I
.Measures Based on Measures Based on
Selected Items All Items of Data

I
- L - - - - - - - - - - - I - - A - 1
"
I
Measure of Absolute Quartlle Mean Standard
Range
Devlallon Deviation Deviation 1
Dispersion

- -- - - - . - - - - - -- _ -_ -- - -. -- - -- -I -
a

a ,, - - - - - L

I
Coefiiclent Coefficient
Measure of
of Quartile
Coefficient
of Mean of Slanderd I
Relative
Deviallon Deviation Devl@tion
Dlsperslon I
- - - - , _ __ _ I _ - - - - - - - --- 1
Measures of Dispersions-1
14.7 RANGE
The range is defined as the difference between the highest (numerically largest) value
and the lowest (numerically smallest) valu'e in a set of data.
Thus, Range = X,,, - Xmi"
Where, X,,, = highest value,' Xmin= lowest value.
From Illustrati~n1 discussed earlier, consider the daily sales data for the three firms
and compute the range.
For Firm A, Range = 60,000 - 60,000 = 0
- For Firm B, Range = 68,250 - 52,250 = 16,000
For Firm G , Range = 2,10,000 - 18,000 = 1,92,000
The interpretation of the value of range is very simple. In this illustration, the
variation is zero in case of daily sales for Firm A, the variation is small in casegof
Firm B, and the variation is very large in case of Firm C .
For grouped data, the range may be approximated as the difference between the
upperlimit of the largest class and the lower limit of the smallest class. The relative
measure corresponding to range, called the coefficient of range, is obtained by
expressing range as the Latio of sum of two extreme items. In this case ratio is not
expressed in k r m s of average, as the range does not depend on average. It relates
only to two selected itetns of the data. So the coefficient of range is defined as :

Coefficient of Range = XnI;Ix - 'mln


Xmix + Xmin

Study Illustration 2 carefully and understand the procedure involved in the


computation of the coefficient of range. .

Calculate the coefficient of range from the following data :

Snles(Rs.in Lakhs) No. of Dnys

30 - 40 12
40 - SO 18
50 - 60 20
60 - 70 19
70 - 80 13
80 - 90 8

Solution
Range = Xnlilx - Xnun
= 90 - 30

= 60.

Coefficient of Rangc = X l ~ -; Xnlln


~ ~ ~
Xm;lx + Xrilin

The range is very easy to calculate and it gives us some idea about the variability of
. the data:S'ince only t k o extreme va!ues are used for computing range, it is a crude
rne.isure of variation.
-

;\hasurcs or Dispersion and The concept of range is extenc;vely used in statistical quality control. Range is helpful
Skcw~~ca~
in studying variations in the, -ices af shares, debentures and agricultural commodities
which are very sensitive to price changes. The range is a good indicator for weather
forecast.

14.8 OUARTILE DEVIATION


Quartile deviation is defined as half the difference between the upper and lower
qualities. You have already studied the methods of computing Quartiles in Unit 11.
Quartile Deviation = Q3 - Ql
2
Where Q, is the first quartile and Q3 is the third quartile.
A s the difference between Q, and Q3 is the distance between the two quartiles, this
may be called Inter Quartile Range and half of this, Semi-Inter Quartile Range is
called Quartile Deviation.
Quartile Deviation (QD) is dependent on the two quartilcs, and docs pot take into
account the variability of the largest 25% and the smallest 25% of observations. It
is, therefore, unaffected by extreme values. Another advantage of quartile deviation
is that it is the only measure of variability which can be used for open-end
distribution. The main limitation of quartile deviation is that it does not depend on
the magnitudes of all observations. It is based on the middle 50% of the observations.
The relative measure of dispersion based on quartile deviation is called coefficient of
quartile deviation. The coefficient of quartile deviation is defined as :
Coefficient of Quartile Deviation = Q3 - Q ,
Q3 + Ql
This is because quartiles are also two selected items of the data and they have nothing
t o do with the average of the data. Study the following Illustrations carefully, you
will understand the procedure involved in the calculation of Quartile Deviation.
illustration 3
Calculate quartile deviation and its cocfficiqnt from the following data :
Weight(in Kgs): 60 61 62 63 65 70 75 80
No. of Workers : 1 3 5 7 10 3 1 1
Solution
Computation of Quartile Deviation and its Coeflicient
- -
Weight in Kgs, Frequency Cumulative Frequency

60 1 1
61 3 4
62 5 9
63 - 7 I6
65 10 26
70 3 29
75 1 30
80 1 31=n

Q, = Size ofq1 (n f 1) o r 8th observation


= 62 Kgs. (because 8th observation falls in this category)
+
Q3 = Size o f 1 (n 1) o r 24th observation
4
I
= 65 Kgs. (because 24th observation falls in this category)

.'. Quartile Deviation - QS - Q 1


- I
2 .
= ----
65 - 62
2 I

= 1.5 Kgs.
Measures of Disporslons-I
Coefficient of Quartile Deviation = QJ - Q1
Q3 + Q,

Illustration 4
Calculate semi-interquartile range and its coefficient from the following data :
Marks : 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90
No. of
Students : 11 18 25 28 30 33 22 15 22

Solution
To compute quartile deviation, we need the values of the first quartile and the third
quartile which can be obtained from the following table :
Marks Frequency (f) Cumulative
Frequency (c.f.)

N observations i.e., 204 = 51 observations, below it. So Q1 lies in the 20-30


Q I has -
4 4
class.

Where 1 = lower limit of quartile class


c = cumulated frequency preceding the quartile class
f = simple frequency of the quartile class
i = class-interval of quartile class

Q3
3N observations i.e., 3 x
has- B = 153 observations, below it. So Qg lies in
4 4
the 60-70 class.
~eeruresof Dispersion .ad Semi-inter Quartile Range or Qllartile Deviation is given by
Skewnew
Q 3 - Q1
Q.D. = - 7

= 17.42 marks.
The relative measure corresponding to quartile deviation, called the coefficient of
quartile deviation, it calculated as follows :

Coefficient of Q.D. = Q3 - QI
Q3 + Q,

Illustration 5
Compute an appropriate measure of dispersion for the data given below:
' Monthly Expenditure (Rs.) No. of Families

Below 850 12
850 - 900 16
900 - 950 39
950 - 1,000 ' 56
. 1,000- 1,050 62
1,050 - 1,100 75

1,150 and above 10

Sution
Since the frequency distributiongas open-end class,'quartile deviation will be the
most appropriate measure of dispersion,
Monthly Expenditure (Rs.) No. of Families ~ u m u l a t l Frequency
~e

Below 850 . 12 12
850 - 900 16 . 28
900 - 950 39 67
950 - 1,000 56 123 '

, 1,000 - 1,050 62 185


1,050 - 1,100 75 260

1 ,I 50 and above 10 300 = n

' 1

Ql h a s43 observations i.e.,300


4 = 75 observations, below it. So Q, lies in the I

class 950 - 1,000. II


1

= Rs. 957,,14.
1
Q, has- 3N obsekationsi.e., = 225 observations, below it. So Q3 lies in I
4 4
12 ,the class 1,050 1,100.-
.'. Q.D. =
a; - a,
2

= Rs.59.76.
. Check Your Progress B
- 1 ~ i s t i n ~ u i sbetwein
h the absolute and relative measures of dispersion.
t .
.................................................................
'

.........................................

2 . Define quattile 'deviation.


................................................................................................ ;a,.......

...................................................
! .
i.,...'................1,...........1.....~............

3 Distinguish between range and the coefficient of range,'


..........................................................................................................

4 Compute the range and quartile deviation for the following data onthe number
of patients treated at the Hospital emergency room per day.
Measures of Dhpersion and
Skewness
- 5 Compute range, quartile deviation and related coefficients from the following
data:
Size : 5-7 8-10 11-13 14-16 17-19
Frequency : 14 24 38 20 4

6 State whether the following stateme-ntsare True or False :


i) Range ignores the distribution of values between the two extremes.
ii) Range can be calculated for an open-end data.
'
iii) Quartile deviation is sarne.as the inter percentile range of P,, PY5.
iv) Quartile deviation is a best measure of dispersion for open-end data.
- v) Relative measure of Quartile Deviation is based on Median. .I.'

.14.9 MEAN DEVIATION


As you know, one of the characteristics of an ideal measure of dispersion is that it
should be based on all items. From this point of view, range and quartile deviations
are not ideal as they are not based on all the observations of the data. But, the
measure of mean (or average) deviation is ideal in this sense as this measure is based
on all observations in the given data set. This measure is computed as the arithmetic
mean.of the absolute deviations of the individual observations from the average of
the given data. The average which is frequently used in computing the mean deviation
is mean or median, though sometimes mode can also be used. Absolute deviations
means the deviations are treated as positive regardless of the actual sign. Given the
observations XI,XZ,....,X,, in order to find Mean Deviation about average A, we
first obtain the deviations (XI- A), (X2 - A), ..., (X, 1- A). Some of these
deviations may be positive and some negative. If we write Ixi- AI to denote the
positive value of (Xi - A), whatever be the actual sign, the sum of these absolute
+
1

+ +
deviations is (XI- A1 IX2- A1 .......... IX, - kl = 2 I X - AI. The arithmetic
mean of these absolute deviations is the Mean Deviation.
I

1 1 lX - A1
'.'. MeanDeviation(ab0utA) =- I
n
i) When A is the arithmetic mean (x), , .
I

Mean Deviation (about x)=L


n 1 IX - XI

ii) When A is the median (Me) ,


1 1 IX - Me[
Mean Deviation about M, =-
n
.. I

iii) When A is mode (M,)


1
Mean Deviation about M, = - 1 /X- M,I
n
I
I
If the data given is in frequency dktribution form, taking anologL from caiculation of
arithrnetic'mean, mean deviation (M.D.) about:x =- 1 1flX1- XI. Similarly, mean I
n I

n =- 1flX - M~.I. Mean deviation about M, =-


deviation about Me 1 1 X flX - M,(. t
>I4 n
&,*$ I
Usually, the name of average from which deviations are taken is always mentioned Measures of Dispersions-l

with mean deviation. But in someiexts, when deviations are taken from insteadx,
of mean deviation about X only 'Mean Deviation' has been stated as its name. An
important property of Mean Deviation is that it has the minimum value when
deviations are taken from median, i.e., Mean Deviation about median is the least.
The relative measure corresponding to the mean deviation, called the coefficient of
mean deviation, is obtained by dividing mean deviation by the particular average used
in computing the mean deviati~n.Thus, if mean deviation has been cornpu'ted from
median, the coefficient of mean deviation shall be obtained by dividing the mean
deviation by the median. Coefficient of M.D. about M, = M.D. Median.
Median
Similarly, a coefficient of M. D, about =
(M.D. about
-
x)
X
' Mean deviation is based on all observations and hence takes into account the variability
of each of the items in the data set. However, the practice of neglecting signs and taking
absolute deviations makes it difficult to be treated algebrically.
Although the average deviation is a good measure of variability, its use is limited. If one
desires only to measure and compare variability among several sets of data, the average
deviation may be used. The concept of mean deviation will become clear, if you study
the following illustrations carefully.

Illustration 6 .
Calculate the mean deviation of the following values about the Median :

Solution
Since there are ten observations which in an even number, the median is the average of
the two middlemost observations, when arranged in order of magnitude

Calculations for Mean Deviation

Mean Deviation about Median = 1 2: IX - M,]


'

-
- o-
: X 272
= 27.2.
Illustration 7
Find'the Mean Deviation about Mean of the followihg series ':
X : 10 11 12 13 14 Tdtal
Frequency 3 12 18 12 3 . 48'
Measures of Dispersion and Solution
Skewness
Calculation of Mean Deviation about a

1
Mean Deviation (about Mean) = - C f.,lX -
n .

Illustration 8
From the folloking groupkd data relating to the sales of 100 Companies, find out
Coefficient of Mean Deviation by using mean ( X ) .

Sales (Rs. '000) No. of Companies

Solution

-
T o construct average deviation, we have to construct the following table .:
Sales Mid No. of fX IX -TI flX - X
I
(Rs.'000) Values (X) Cornpahies (0 i.e., IX - 71 1

40- 50 45 5 225 26 130


50- 60 55 15 825 16 240
60- 70 65 25 1 ,625 6 1 SO
70- 80 75 30 2.250 4 120
80- 90 85 20 1.700 14 . 280
90- 100 95 5 475 24 120

Total n = 100 \'fX = 7.100 \f IX = 1.040

n 1 fJX-
Mean Deviation (about Mean) =L

- 1
-- x 1.040 = 10.40 or Rs. 10.4 thousands.
100

Coefficient of Mean Deviation = Mean Deviation about 7

X
I Illustration 9 Measures of Dispersions-I

The following is the age-distribution of 80 LIC Policy holders insured through an


agent. Calculate the coefficient of mean deviation from the median.
Age Group (in Years) Frequency

16 - 20 8
21 - 25 15
26 - 30
I
13
I
I 31-35 20
36 - 40 11
41 -45 7
46 - 50 3
51-55 . 2

Solution
Calculation of Mead Deviatlon from Medlnn
--
Age-Group Frequency Cumulative . Class -
IX Mdl fJX - Mdl
(In Years) (0 Frequency Mid-Point -
i.e., lX 31.51
(CC) (MI

I " 41-45 7 74 , 43 11.5 80.5


-
46 50 3 ' 77 48 16.5 49.5
I
51-55 2 79 53 21.5 43.0
I 56 - 60 1 80 58 26.5 26.5

Total n = 80 If IX - M,I = 582.0

Median h a s d or 40 observations below it. So it lies in the class of 31-35 or


2 .
30.5 - 35.5 (in terms of real limits).

,1 --
N c
1 Median = 1 + -
2 xi
I f

= 30.5 + 20
x 5 = 31.5 years.

I
Mean Deviation (about Median) =I. Xf IX - Mdl
I n
I -
--1 x 582 = 7.275 years.
I 80
1 ,
1 M.D. about Md
Coefficient of Mean Deviation about median =
Md

I 14h0 LET US SUMUP, ,

I Dispersion represents the Spread or the scatterness of the data. It is also used to
/ denote the average of deviation of items fr.om some measure of central tendency,
I Dispersion is calculated to assess the reliability of an average or to compare variability
of two or more data or to control the variation itself. A good measure of dispersion
should be based on all observat;~ns,should easily be calculated, least affected by
sampling fluctuations and amenable to further algebric treatment. Relative
measures of dispersion are computed to compare variability in two or more sets of
data. They are obtained by expressing absolute measures of dispersion as the ratio
of the appropriate average or the sum of two selected items of the data.
The various measures of dispersion in common use are range, quartile deviation,
mean deviation and standard deviation. Range is defined as the difference between
the highest and the lowest items of the data. It gives the spread of entire data.
Quartile deviation is half the difference between Q1 and Q, and is based on middle
50% items only. Mean deiiation is the arithmetic mean of the absolute deviations of
items from a measure of central tendency, which could be mean or median or some]
times even mode.

Quartile deviation is a suitable measure for open-end data. Range is usef;l when
extreme items are important such as in quality control, price study or meteorological
data. As mean deviation is based on all items, in most of the cases it is a better
representative of the variability of the data than the other two measures.

14.11 KEY WORDS AND SYMBOLS


Inter Quartile Range : A measure of dispersion which considers the spread in the
middle 50% ..It is (Q3 - Q1) of the data.
Mean Deviation : The arithmetic mean of the absolute deviations from the mean
median or the mode.
Quartile Deviation :One-half the distance between the first and the third quartiles.
Range : The difference between the largest and the smallest value in a set of data.
' ,---

List of Symbols
In! addition to the list of symbols given in Block 3, following are the additional
symbols used in connection with.measures of dispersion given in this Unit.
Mean Deviation I M.D., 6, A.D.
Mean Deviation about Arithmetic Mean : M.D. 6 x,AD
(x),
Mean Deviation about Median : M D Md, 6 Md, A D Md
Mean Deviation about Mode : M D (M,), 6 M,, 62, AD M,
Quartile Deviation : Q.D, Q
'.
Range ': R

14.12 ANSWERS TO CHECK YOUR PR,OGRESS


A 3 i) False ii) False iii) True iv) True v) False vi) True

5 kange = 14, Coefficient of Range = 0.58, Q.D. = 2.25,


Coefficient of Q.D. = 0.101
6 i) True ii) False iii) False iv) True v) False

14.13 TERMINAL OUES~~ONS/EXERCISES

1 What is the mean deviation? Review its advantages and disadvantages.

. I
Measures of Dispersiuns-I

1 Calculate quartile deviation and mean deviation about for the following data :
Age (in Years) : 20 30 40 50 60 70 80
No.ofMembers: 3 61 132 153 140 51 3

1
4
2
(Answer : Q.D. =.10, M.D. = 9.52)
A frequency distribution for the duration of 20 long distance telephqne calls are
shown below :
Calls Duratioa . Frequency

4 but less than 8 ,

&butless than 1 2 ,
1
12 but less than 16
16 but less than 20 2
20 but less than 24 1
24 but less than 28 1
I
I
Total 20
II
Compute the mean, median and quartile deviation.
i (Answer : Mean = 12.8, Median = 12.6, Q.D. = 3.3)
I
I 3 Calculate the mean deviation about Median and coefficient of mean deviation
I
from the following data :
Sales (Rs. ' 0 0 ) No. of Companies

Less than 20 3
Less than 30 9
Less than 40 20
tl
Less than 50 23
I
i
Less than 60 25
Il
(Answer :'M.D. about M, = 8.9, Coefficient of M.D. about M, = 0.29)
1
I
4 A. survey of domestic consumption of electricity gave the following distribution
of the units consumed. Compute the quartile deviation and its coefficient.
I
1

i No. of Units No. of Consumers


,
J1
t
Below - ,200 9
200 - 400 18
400 - 600 27
I 600 - 80() 32
I ROO - 1,000 . 45

I
I .ooO - 1,200 , 38
I
I 1,200 - 1,4(0 20
1
1.400 and above 11

I/ (Answer : Q.D. = 520.6, Coefficient of Q.D. = 0.317)

1 I
I
5 .

'
.
Calculate the mean deviation about the mean and median from ihe following
Zta:
Class Interval : 0-9 10-19 20-29 30-39 40-49 50-59
Frequency : 15 36 53 42 17 2
'j
Y , ; (Answer:M.D.~=9.10,M.D.MC=9.08)
4
I ' 6 Calculate the mean deviation about Mode and its coefficient for the following
l data :
No. or Detects per Item Frequency

0- 5 18
-
40 45 80
45 - 50 50
'
(Answer : M.D.M, = 9.02, Coefficient M.D.M, = 0.338)
7) Compute the mean deviation3and its coefficient for the following data :
No. of Sham Applkd lor No.of Applicants
50-100 2,500
100 - 150 1,500
150-200 1,300
200 - 250 1,100
250 - 300 900 '

300- 350 750


-
350 400 675
400-450 - 525
450- 500 450

(Answer : M.D.Me= 102.13, Coefficient of M.D. Me= 0.011)

Note : These questions and exercises will help you to understand the unit better.
Try to write answers for thern..But do not submit your answers to the University,
These are for your practice only.
UNIT MEASURES OF DISPERSION II -
1 Structure

11
15.0 Objectiv~,~
15.1 Introduction
15.2 Standard Deviation
15.2.1 Meaning
I 15.2.2 Computation

r 15.2.3
15.2.4
Properties
Merits and Limitations

I
15.3 Coefficient of Variation
15.4 Some Illustrations
15.5 Lorenz Curve
15.6 Comparison of Measures of Dispersion
15:7 Let Us Sum U p
15.8 Key Words and Symbols
15.9 Answers to Check Your Progress
15.10 Terminal Questions/Exercises

1; After studying this unit, you should be able to


e define and compute standard deviation and coefficient of variation for different
kinds of data,
e explain the merits and limitations of standard deviation,
e draw Lorenz Curve and graphically determine the extent of unequalities of items,

li e compare different measures of dispersion and use them at appropriate places.

15.1 INTRODUCTION
I In Unit 14 you have learnt about three measures of dispersion viz., range, quartile
deviation and mean deviation. The first two are based on two selected items of the
I
data and the third measure is computed by using the values of each and every item.
But in the calculation of mean deviation the negative signs of the deviations of items
from a central tendency are ignored. As we ignore the signs which arise during
1
calculations, mean deviation suffers from certain limitations. There is another measure
of dispersion viz., standard deviation, which takes care of the problem of signs. In
this unit you will learn about the method of computing standard deviation and its
coefficient for different kinds of data, their merits, limitations and uses. You will also
learn the Lorenz Curve which is the graphical method of finding the dispersion.

( i
15.2 STANDARD DEVIATION
As discussed earlier, while computing the mean deviation we ignore the negative
signs of the deviations of the items from the central tendency. This is because in
dispersion we are interested only in knowing how much, on an average, items deviate
from central tendency irrespective of the fact that items are less than or more than
central tendency. This ignoring of signs which arise during calculations, introduces
some limitations on the measure. A mathematical solution for ignoring signs is
squaring. As the square of any negative item becomes positive, a new measure of
dispersion is defined in which deviations are first squared (to ignore the signs) and
then averaged out. f i e value so obtained gives the average of the squafes of the
deviations and not of deviations directly. So, finally a square root of this value is
extracted. Thus the result obtained will give an indirect average of deviations. As this
measure is calculated by finding square root of the mean of the squares of the
deviations of items from central tendency, it is called Root Mean Square Deviation.
i
Like mean deviation, root mean square deviation can also be calculated by 21

S-ar putea să vă placă și