CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU

CHAPTER 3 MEASURES OF DISPERSION

Though the measures of central tendency are useful but they alone are not sufficient to explain the
data exactly. We must find how representative the average is. For e.g., if average depth of a river is
1 metre, we must test reliability of this average before entering into the river. If the variation is
small, an average becomes meaningful otherwise it has limited significance.
Dispersion (also known as Scatter, Spread or Variation) measures extent to which the items vary
from some central value. Since these measures give an average of the differences of various items
from an average, they are also called ‘Averages of the Second Order’.

PROPERTIES OF A GOOD MEASURE OF DISPERSION

 Simple to Understand & calculate  Capable of further Algebraic Treatment
 Rigidly Defined  Sampling Stability
 Based on all items  Not Unduly Affected by Extreme Items

ABSOLUTE AND RELATIVE MEASURE

 Absolute Measures of Dispersion: It is the measure of dispersion which is expressed in the
same statistical unit in which the original data are given such as Kilograms, Tonnes, Kilometers,
Rupees etc. These measures are suitable for comparing the variability in two distributions having
variables expressed in the same units and of the same average size. These measures are not
suitable for comparing the variability in two distributions having variables expressed in different
units.

ABSOLUTE measures of Dispersion

Based on SELECTED items Based on ALL items

Range Inter- Mean Deviation

Quartile Range Standard
Deviation
 Relative Measure of Dispersion: It is the RATIO of a measure of absolute dispersion to an
appropriate average or the selected items of the data. It can be used for comparison of two or more
sets of data. It is also called as Coefficient of Dispersion as it is a pure number & is independent
of the unit of measurement.

RELATIVE measures of Disperson

Based on SELECTED items Based on ALL items

Coefficient of Range Coefficient of Mean

Dev.
Coefficient of Quartile
Dev. Coefficient of Variation

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.1

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU

RANGE: Range is the simplest but a very rough measure of dispersion. Its value depends upon the
extreme items and not on all the items. It is denoted by ‘R’.

Measures of Range Formula

ABSOLUTE MEASURE
 Individual or Discrete Series R = L – S where, R = Range, L = Largest Value, S =
Smallest Value
 For Continuous Series R = UL – Ls Where, R = Range, UL = U.C.B. of Largest
Class, Ls = L.C.B. of Smallest Class
RELATIVE MEASURE
 Individual or Discrete Series Coefficient of Range =
 For Continuous Series Coefficient of Range =
Note: Frequencies are not at all considered for calculating the range and coefficient of range.

PROPERTIES OF RANGE
 If all items are equal, then range is zero.
 It remains UNAFFECTED due to change of ORIGIN, but is AFFECTED by change of SCALE in same
ratio. Thus if there is change of origin of ‘a’ & change of scale of ‘b’. (Refer Page no. 5.22)
i.e. if Y = a + bX, then, 𝑅 = |𝑏| × 𝑅

MERITS OF RANGE: The merits of range are as follows:

 It is easy to calculate & understand  It facilitates to study price variations
 It facilitates statistical quality control  It facilitates the weather forecasts

LIMITATIONS OF RANGE: The limitations of range are as follows:

 Not based on all items of the series  Not applicable to open end classes
 Not suitable for mathematical treatment  Highly affected by sampling fluctuations
 Extremely affected by extreme values

1) From the following data, calculate Range and Coefficient of Range:

Roll No. 1 2 3 4 5 6
Marks 5 15 25 35 45 55

2) From the following data, calculate Range and Coefficient of Range:

Marks 5 15 25 35 45 55
No. of students 10 20 30 50 40 30

3) From the following data, calculate Range and Coefficient of Range:

Marks 0-10 10-20 20-30 30-40 40-50 50-60
No. of students 10 20 30 50 40 30

INTER QUARTILE RANGE

It represents the difference between the upper quartile (Q3) and the lower quartile (Q1)
Inter-Quartile Range = Q3  Q1

QUARTILE DEVIATION: (SEMI-INTER QUARTILE RANGE)

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.2

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU
Quartile Deviation is half the difference between upper quartile (Q3) & lower quartile (Q1). It is a
positional measure & does not exhibit any scatter around an average. Hence it is also called as a
measure of partition. It indicates the average amount by which the two quartiles differ from the
median. In symmetrical distribution the two quartiles are equidistant from the median [i.e.
Median – Q1 = Q3 – Median]. As a result Median,  Q.D. covers exactly 50% of the items. In certain
cases the central 50% of the items are typical. Q.D. is a suitable measure in such cases.

Measures of Quartile Deviation Formula

𝑄 −𝑄
ABSOLUTE MEASURE OF Q.D. 𝑄. 𝐷. =
2
𝑄 −𝑄
RELATIVE MEASURE OF Q.D. 𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑜𝑓 𝑄. 𝐷. =
𝑄 +𝑄

PROPERTIES OF QUARTILE DEVIATION

 In case of symmetrical distribution, Median =
 If Q.D. is small it indicates a Small Variation (or High Uniformity) among the central 50% items of
the distribution and vice versa.
 Q.D. remains UNAFFECTED due to change of ORIGIN, but is AFFECTED by change of SCALE in
same ratio. Thus if there is change of origin of ‘a’ & change of scale of ‘b’.
i.e. if Y = a + bX, then, 𝑄. 𝐷.( ) = |𝑏| × 𝑄. 𝐷.( )
 Coefficient of Q.D. remains UNAFFECTED due to change of SCALE, but is AFFECTED by change
of ORIGIN (Not to the same extent) (Refer Page no. 5.22)

MERITS OF QUARTILE DEVIATION

 It is easy to understand & calculate  It can be used in Open-ended series
 Useful for study of middle 50% items  It is least affected by extreme items

LIMITATIONS OF QUARTILE DEVIATION

 Not suited to algebraic treatment  Not based on all the items of series
 Much affected by sampling fluctuations

4) From the following data, calculate Quartile Deviation & Coefficient of Quartile Deviation:
Roll No. 1 2 3 4 5 6
Marks 25 55 5 45 15 35

5) From the following data, calculate Quartile Deviation & Coefficient of Quartile Deviation:
Marks 5 15 25 35 45 55
No. of students 10 20 30 50 40 30

6) From the following data, calculate Quartile Deviation & Coefficient of Quartile Deviation:
Marks 0-10 10-20 20-30 30-40 40-50 50-60
No. of students 10 20 30 50 40 30

PERCENTILE RANGE AND PERCENTILE DEVIATION

𝑷𝟗𝟎 𝑷𝟏𝟎 𝑷𝟗𝟎 𝑷𝟏𝟎
Percentile Range = P90 – P10 𝑷. 𝑫. = 𝟐
𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑜𝑓 𝑷. 𝑫. = 𝑷𝟗𝟎 𝑷𝟏𝟎

7) From the following data, calculate Percentile Deviation & Coefficient of Percentile Deviation:
Marks 0-10 10-20 20-30 30-40 40-50 50-60
No. of students 10 20 30 50 40 30

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.3

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU
MEAN DEVIATION (AVERAGE DEVIATION)
Mean Deviation is based on all items of a series. It is the A.M. of absolute deviations (i.e. ignoring
signs) of all items of a series from any measure of central tendency, i.e. mean, median or
sometimes even mode. Since mode cannot always be determined, mean deviation is computed
from mean or the median. Median is preferred as the sum of absolute deviations taken from the
median is the minimum. But mean is generally used due to its wide applications. This measure is
found useful than others for forecasting of business cycles.

MEASURES OF MEAN DEVIATION

Measures of Mean Deviation may be absolute or relative.
ABSOLUTE MEASURE M.D. about Mean M.D. about Median M.D. about Mode
Individual Series Σ|𝑋 − 𝑋| Σ|𝑋 − 𝑀| Σ|𝑋 − 𝑍|
𝑀. 𝐷. = 𝑀. 𝐷. = 𝑀. 𝐷. =
𝑁 𝑁 𝑁
Discrete Series Σ𝑓|𝑋 − 𝑋 | Σ𝑓|𝑋 − 𝑀| Σ𝑓|𝑋 − 𝑍|
𝑀. 𝐷. = 𝑀. 𝐷. = 𝑀. 𝐷. =
𝑁 𝑁 𝑁
Continuous Series Σ𝑓|𝑚 − 𝑋| Σ𝑓|𝑚 − 𝑀| Σ𝑓|𝑚 − 𝑍|
𝑀. 𝐷. = 𝑀. 𝐷. = 𝑀. 𝐷. =
𝑁 𝑁 𝑁

RELATIVE Coeff. Of M.D. about Coeff. Of M.D. about Coeff. Of M.D. about
MEASURES MEAN MEDIAN MODE
Individual, Discrete 𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑒𝑎𝑛 𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑒𝑑𝑖𝑎𝑛 𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑜𝑑𝑒
& Continuous Series 𝑀𝑒𝑎𝑛 𝑀𝑒𝑑𝑖𝑎𝑛 𝑀𝑜𝑑𝑒

PROPERTIES OF MEAN DEVIATION

 If all items are equal, then M.D. is zero
 M.D. takes minimum values when deviations are taken from the median.
 M.D. remains UNAFFECTED due to change of ORIGIN, but is AFFECTED by change of SCALE in
same ratio. Thus if there is change of origin of ‘a’ & change of scale of ‘b’,
i.e. if Y = a + bX, then, 𝑀. 𝐷.( ) = |𝑏| × 𝑀. 𝐷.( ) (Refer Page no. 5.22)
 Coefficient of M.D. remains UNAFFECTED due to change of SCALE, but is AFFECTED by change
of ORIGIN (Not to the same extent). (Refer Page No. 5.22)

MERITS OF MEAN DEVIATION

 Easy to understand & calculate  Is rigidly defined
 Based on all items of the series  Gives weight to items as per their size
 Less affected by extreme values than S.D.  Can be computed from any average
 Less affected by sampling fluctuations  Useful when detailed analysis not required

LIMITATIONS OF MEAN DEVIATION

 Not reliable as ignores (+) & () signs  Cannot be computed for open-end classes
 Not suitable for algebraic treatment

8) From the following data calculate Mean Deviation & Coefficient of Mean Deviation about Mean
Roll no. 1 2 3 4 5 6
Marks 25 55 5 45 15 35

9) From the following data calculate Mean Deviation & Coefficient of Mean Deviation about Mean
Marks 5 15 25 35 45 55
No. of students 10 20 30 50 40 30

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.4

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU
STANDARD DEVIATION (Root Mean Square Deviation, Mean Error or Mean Square Error)
Karl Pearson introduced the concept of Standard Deviation in 1893 as an improvement over Mean
Deviation which was criticized severely for ignoring signs of deviations. In this method the
drawback of ignoring signs is overcome by taking squares of deviations.
It is defined as positive square root of arithmetic mean of the squares of deviations of all items
of the distribution from arithmetic mean. It is denoted by σ (sigma). It measures the absolute
dispersion of a distribution. It is useful in judging the representativeness of the mean.

VARIANCE
It is the arithmetic mean of squares of deviations of all items of a series from arithmetic mean. It is
square of standard deviation or standard deviation is positive square root of variance.
Thus, 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = 𝜎 Or 𝜎 = √𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒

MEASURES OF STANDARD DEVIATION

Measures of Standard Deviation may be absolute or relative.
ABSOLUTE
Individual Series Discrete Series Continuous Series
MEASURES
Direct Method – 1 Σ𝑥 Σ𝑓𝑥 Σ𝑓𝑚
(based on values) 𝜎= −𝑋 𝜎= −𝑋 𝜎= −𝑋
𝑁 𝑁 𝑁

Σ𝑥 Σ𝑓𝑥 Σ𝑓𝑥
Direct Method – 2 𝜎= 𝜎= 𝜎=
(Actual mean) 𝑁 𝑁 𝑁
𝑥 = (𝑋 − 𝑋 ) 𝑥 = (𝑋 − 𝑋) 𝑥 = (𝑚 − 𝑋)

Short-cut Method Σ𝑑 Σ𝑑 Σ𝑓𝑑 Σ𝑓𝑑 Σ𝑓𝑑 Σ𝑓𝑑

𝜎= − 𝜎= − 𝜎= −
(Assumed Mean) 𝑁 𝑁 𝑁 𝑁 𝑁 𝑁
𝑑 = (𝑋 − 𝐴) 𝑑 = (𝑋 − 𝐴) 𝑑 = (𝑚 − 𝐴)

Step Deviation Σ𝑑" Σ𝑑 " Σ𝑓𝑑 " Σ𝑓𝑑" Σ𝑓𝑑" Σ𝑓𝑑 "
− ×𝑐 − ×𝑐 − ×𝑐
Method (Common 𝑁 𝑁 𝑁 𝑁 𝑁 𝑁
Factor) 𝑥−𝐴 𝑥−𝐴 𝑚−𝐴
𝑑" = 𝑑" = 𝑑" =
𝑐 𝑐 𝑐

RELATIVE MEASURES
Coeff. of 𝜎 They remain UNAFFECTED due to change of SCALE, but are
𝐶. 𝑉. = × 100
Variation 𝑋 affected by change of origin (not to the same extent) (Refer
Coeff. of S. D. 𝜎 Page no. 5.22)
𝐶𝑜𝑒𝑓𝑓. 𝑜𝑓 𝑆. 𝐷. =
𝑋

DISTINCTION BETWEEN MEAN DEVIATION AND STANDARD DEVIATION

Basis of Distinction Mean Deviation Standard Deviation
Algebraic Signs Actual  signs are ignored. Actual  signs are not ignored.
Deviations from A.M., median or mode. Only A.M.
Properties Lacks Mathematical Properties Is Mathematically sound.

PROPERTIES OF STANDARD DEVIATION

 If all items are equal, then S.D. is zero.
 S.D. remains UNAFFECTED due to change of ORIGIN, but is AFFECTED by change of SCALE in
same ratio. If there is change of origin of ‘a’ & change of scale of ‘b’.
i.e. if Y = a + bX, then, 𝑆. 𝐷.( ) = |𝑏| × 𝑆. 𝐷.( ) (Refer Page no. 5.22)

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.5

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU

 Empirical Relationship: In a normal distribution the empirical relationship between Quartile

Deviation (Q.D.), Mean Deviation (M.D.), Standard Deviation (S.D.) & Range is:
2 5 1
𝑄. 𝐷. ≅ 𝑆. 𝐷. ≅ 0.6745 𝑆. 𝐷. ≅ 𝑀. 𝐷. ≅ 𝑅𝑎𝑛𝑔𝑒
3 6 9
4 6 2
𝑀. 𝐷. ≅ 𝑆. 𝐷. ≅ 0.7979 𝑆. 𝐷. ≅ 𝑄. 𝐷. ≅ 𝑅𝑎𝑛𝑔𝑒
5 5 15
3 5 1
𝑆. 𝐷. ≅ 𝑄. 𝐷. ≅ 𝑀. 𝐷. ≅ 𝑅𝑎𝑛𝑔𝑒
2 4 6
15
𝑅𝑎𝑛𝑔𝑒 ≅ 9 𝑄. 𝐷. ≅ 𝑀. 𝐷. ≅ 6 𝑆. 𝐷.
2
Q.D. : M.D. : S.D.: Range = 10 : 12 : 15 : 90

 Fixed Area Relationship for a symmetrical distribution – The fixed area relationship is:
Values Mean  1 σ Mean  2 σ Mean  3 σ
Coverage 68.27% 95.45% 99.73%

Thus, Range = L – S = (Mean + 3 σ) – (Mean – 3 σ) = 6 σ

 Coverage of items: In a normal distribution mean  1 σ covers 68.27% of the items whereas
only 50% values are covered by Q.D. and 57% by mean deviation.

 Minimum sum of squares: Sum of squares of deviations of items from A.M. is minimum.

 For ‘n’ natural numbers, standard deviation is: 𝜎 = (𝑁 − 1)

𝑹𝒂𝒏𝒈𝒆
 For only 2 values, standard deviation is 𝝈 =
𝟐
 For only 1 value, standard deviation is 𝝈 = 𝟎

 If frequency of all items is equal, the frequencies can be ignored for calculating S.D.

 Combined Standard Deviation (𝝈𝟏𝟐 ): Combined Standard Deviation of two groups is:
𝑵𝟏 𝝈𝟐𝟏 𝑵𝟐 𝝈𝟐𝟐 𝑵𝟏 𝒅𝟐𝟏 𝑵𝟐 𝒅𝟐𝟐
𝝈𝟏𝟐 = 𝑵𝟏 𝑵𝟏
; Where 𝒅𝟏 = 𝑿𝟏 − 𝑿𝟏𝟐 ; 𝒅𝟐 = 𝑿𝟐 − 𝑿𝟏𝟐

MERTIS OF STANDARD DEVIATION

 Rigidly defined & based on all items  Coefficient of Variation is based on it
 Amenable to further algebraic treatment  Basis for finding Skewness, Correlation etc
 Least affected by sampling fluctuations  Removes drawback of ignoring + &  Signs
 S.D. of two groups can be combined

LIMITATIONS OF STANDARD DEVIATION

 Not useful for comparison (Absolute)  Unduly affected by extreme values
 Difficult to compute & understand  Cannot be computed for open-end classes

% of Coeff. of Variation Interpretation

Lower Group is more consistent/ uniform/ stable/ homogeneous.
Higher Group is less consistent/ uniform/ stable/ homogeneous.

10) From the following data, calculate the Standard Deviation:

Roll No. 1 2 3 4 5 6
Marks 5 15 25 35 45 55

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.6

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU

11) From the following data calculate Standard Deviation, Coefficient of Variation and Variance:
Marks 5 15 25 35 45 55
No. of students 10 20 30 50 40 30

12) Find Coefficient of Variation if Median is 23, Mode 29, Variance 100.

13) The following table gives height of boys & girls. Find (i) Standard deviation of the heights of
boys and girls taken together, (ii) Whose heights are more variable?
Particulars Boys Girls
Number 400 100
Average height 68 inches 65 inches
Variance 9 4

GRAPHIC METHOD OF DISPERSION (LORENZ CURVE)

Lorenz curve is a graphic method of dispersion devised by Dr. Max O. Lorenz who used this
technique in measuring the distribution of wealth & income. Statements like ‘top 10% of the
people of a country earn 50% of national income’ give an idea about income disparities. It is useful
in comparing variability of two or more distributions.

The steps involved in constructing the Lorenz curve are:

 The values of variable and frequencies are converted into cumulative totals.
 Percentages are calculated for each cumulative figure taking the total to be hundred.
 Percentages of cumulated values of variable & percentages of cumulated frequencies are plotted
on graph with a smooth free-hand curve. This curve is known as Lorenz curve.
 The percentage can be 0 to 100 on both axes starting from the origin. But mostly the percentages
on the X axis are shown as 100 to 0 starting with 100 at the origin. In the former case the Lorenz
curve is upward rising to the right whereas in the latter case it is downward falling to the right.
 A straight line joining the co-ordinates (0, 0) with (100, 100) is drawn in the former case. In the
latter case a straight line joining 100 of Y axis and 0 of X axis is drawn. This line is called as Line of
Equal Distribution. This line would imply situations like top 20% people earn 20% of total income
and top 60% earn 60% of the total income.
 The greater the distance between the curve & line of equal distribution, the greater is the
dispersion and vice versa.

TRUE OR FALSE STATEMENTS

1) Range is the best measure of dispersion.

2) Quartile deviation is more suitable in case of open-end distribution.
3) Mean deviation can never be negative.
4) Mean Deviation is minimum, when deviations are taken about arithmetic mean.
5) Mean absolute deviation from median and mean can be 5 and 4 respectively.
6) If each observation is increased or decreased by 10, Range, Quartile Deviation and Standard
Deviation will be increased or decreased by 10.
7) If each observation is multiplied or divided by 10, Range, Quartile Deviation and Standard
Deviation will be multiplied or divided by 10.
8) If each observation is increased or decreased by 10, Coefficient of Quartile Deviation and
Coefficient of Variation will also be changed.
9) If each observation is multiplied or divided by 10, Coefficient of Quartile Deviation and
Coefficient of Variation will be changed.

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.7

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU
10) If all observations of an individual series are same, Range, Quartile Deviation and Standard
Deviation will be 1.
11) If all observations of an individual series are same, Coefficient of Quartile Deviation and
Coefficient of Variation will be 100%.
12) Standard Deviation is smaller than Mean Deviation about mean.
13) Root-Mean-Square-Deviation is minimum when deviations are taken about median.
14) Standard Deviation of two observations is equal to their arithmetic average.
15) Standard Deviation of one observation is 1.
16) The unit of variance is same as that of observations.
17) The variance of a sample of a variable is 10.
18) The variance of first 13 natural numbers is 13.
19) Coefficient of variation is expressed in the same units as the original data.
20) For comparing the variability of two or more distributions whose means differ widely, the
appropriate measure of dispersion is Standard Deviation.
21) Two samples A and B have the same standard deviation but the mean of A is greater than that
of B. The coefficient of variation of A will be greater than that of B.

1) Coefficient of variation 2) True & 3) True 4) Median

5) 4 and 5 6) Unchanged 7) True
8) True 9) Unchanged 10) 0
11) 0 12) Larger 13) Arithmetic Mean
14) Half their range 15)0 16) Square of
17) Cannot be negative 18) 14 19) Has no unit
20) Coefficient of variation 21) Lesser
18) Variance of 1st 13 natural nos. = = = = 14

ANSWERS
TRUE – 2, 3, 7, 8. FALSE - Remaining

EFFECT OF CHANGE OF ORIGIN/ SCALE ON MEASURES OF DISPERSION

Measures Change of ORIGIN Change of SCALE Change of BOTH

X X+5 X–3 4X X/2 2X + 7 50 – 2X
12 17 9 48 6 31 26
13 18 10 52 6.5 33 24
14 19 11 56 7 35 22
15 20 12 60 7.5 37 20
16 21 13 64 8 39 18
17 22 14 68 8.5 41 16
18 23 15 72 9 43 14
Range 6 6 6 24 3 12 12
Cft. of Range 0.2 0.15 0.25 0.2 0.2 0.16 0.3
Q.D. 2 2 2 8 1 4 4
Cft. of Q.D. 0.13 0.1 0.16 0.13 0.13 0.108 0.2
M.D. 1.71 1.71 1.71 6.85 0.85 3.42 3.42
Cft. of M.D. 0.11 0.08 0.14 0.11 0.11 0.09 0.17
S.D. 2 2 2 8 1 4 4
C.V. 13.33 10 16.66 13.33 13.33 10.81 20

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.8

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU

FORMULAE OF MEASURES OF DISPERSION

Measures Individual Series Discrete Series Continuous Series

RANGE L–S L–S U L – LS
𝐿−𝑆 𝐿−𝑆 𝑈 −𝐿
Coefficient of Range
𝐿+𝑆 𝐿+𝑆 𝑈 +𝐿
𝑄 −𝑄 𝑄 −𝑄 𝑄 −𝑄
Quartile Deviation
2 2 2
𝑄 −𝑄 𝑄 −𝑄 𝑄 −𝑄
Coefficient of Q.D.
𝑄 +𝑄 𝑄 +𝑄 𝑄 +𝑄
Σ|𝐷| Σ𝑓|𝐷| Σ𝑓|𝐷|
Mean Deviation 𝑀. 𝐷. = 𝑀. 𝐷. = 𝑀. 𝐷. =
𝑁 𝑁 𝑁
𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑒𝑎𝑛 𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑒𝑑𝑖𝑎𝑛 𝑀. 𝐷. 𝑎𝑏𝑜𝑢𝑡 𝑀𝑜𝑑𝑒
Coefficient of M.D.
𝑀𝑒𝑎𝑛 𝑀𝑒𝑑𝑖𝑎𝑛 𝑀𝑜𝑑𝑒
Σ𝑋 Σ𝑓𝑋 Σ𝑓𝑚
Standard Deviation 𝜎= −𝑋 𝜎= −𝑋 𝜎= −𝑋
𝑁 𝑁 𝑁

Σ𝑥 Σ𝑓𝑥 Σ𝑓𝑥
Standard Deviation 𝜎= 𝜎= 𝜎=
𝑁 𝑁 𝑁

For any 2 values: 𝜎 = For ‘n’ natural nos., 𝜎 = (𝑁 − 1)

𝑁 𝜎 +𝑁 𝜎 +𝑁 𝑑 +𝑁 𝑑
Coefficient of Variation: 𝐶. 𝑉. = × 100 𝜎 =
𝑁 +𝑁

PROPERTIES OF MEASURES OF DISPERSION

Quartile
Mean Deviation
Particulars Range (R) Deviation Standard Deviation (σ)
(M.D.)
(Q.D.)
Difference Half of A.M. of absolute
Square root of A.M. of
between difference deviations of all
Meaning squares of deviations of
largest & between upper items from an
all items from A.M.
smallest item. & lower quartile average.
Relative Coefficient of Coefficient of Coefficient of
Coefficient of Variation
Measures Range Q.D. M.D.
ABSOLUTE measures have SAME UNIT as that of the series & RELATIVE
Unit of Measure
measures have NO UNIT.
Based on all
No No Yes Yes
items
Extreme Affected less
Yes Least affected Affected more than M.D.
values? than S.D.
Open-end
Not Suitable Suitable Not Suitable Not Suitable
Distribution
Combined
No No No Yes
Dispersion?
If all items If all items are equal, ALL the measures of dispersion (absolute & relative)
equal (say k) will be ZERO.

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.9

CA SURAJ SATIJA MEASURES OF DISPERSION SSGURU
If any/ all items Any measure of dispersion is NEVER NEGATIVE. The lowest value of ANY
negative measure of dispersion is ZERO.
Sampling Affected more
Yes (Most) Yes Least affected
Fluctuations? than S.D.
Signs of
Not Applicable Not Applicable Ignored Taken into consideration
Deviations?
Taken into consideration (can be ignored only if all items have
Frequencies? Ignored
equal frequency)
Independent of ABSOLUTE measures are AFFECTED and RELATIVE measures are AFFECTED
Origin? by change of ORIGIN.
Independent of ABSOLUTE measures are AFFECTED (same ratio) and RELATIVE are
Scale? UNAFFECTED by change of SCALE.
Mathematical 𝑄. 𝐷.( ) 𝑀. 𝐷.( )
𝑅 = |𝑏| × 𝑅 𝑆. 𝐷.( ) = |𝑏| × 𝑆. 𝐷.( )
Property = |𝑏| × 𝑄. 𝐷.( ) = |𝑏| × 𝑀. 𝐷.( )
Value is Deviations from
N.A. N.A. Deviations from A.M.
Minimum if median
Range = 9 Q.D. M.D. = 6/5 Q.D. 𝑋 ± 1𝜎, 𝑋 ± 2𝜎, 𝑋 ± 3𝜎
In a M = (Q3 + Q1)/2
Range = 7.5 M.D. = 4/5 σ covers 68.27%, 95.45%,
Symmetrical Q.D. = 5/6 M.D.
M.D. QD : MD: SD = 10 99.73% items. Σ = 3/2
distribution Q.D. = 2/3 σ
Range = 6 σ : 12 : 15 Q.D. = 5/4 M.D. = 1/6 R

MEASURES OF DISPERSION BY CA SURAJ SATIJA 3.10