Measures of Dispersion Quartile Deviation

Measures of dispersion – is a method of measuring the The Quartile deviation is also called the semi-quartile range.
degree by which numerical data or values tend to spread It is defined as the amount of dispersion in the middle of 50%
from or cluster about central point of average. of the values.

The most common measures of dispersion are the following: Q3−Q1

Q.D. =
1. The range 2
2. The Quartile Deviation Where:
3. The Average Deviation Q.D. = the quartile deviation
4. The Standard Deviation Q1 = the first Quartile
5. The Variance Q3 = the 3rd Quartile

The Range of UNGROUPED DATA Average Deviation for Ungrouped data:

1. For a Sample:
The range of a set of ungrouped data is the difference
between the highest and lowest values, in mathematical ε |x−x|
form. M.A.D. =
n
Range = Highest – Lowest Where:
M.A.D. = the population mean absolute deviation
Where: × = the individual value
R = the range x = the sample mean
H = Highest value n = the total number of items or observations
L = Lowest value
2. For a Population:
The Range for Grouped Data
The range for grouped data is generally defined as
the difference between the upper boundary of the highest M.A.D. =
∑ |x−u|
class and the lower boundary of the lowest class. N
Where:
u = the population mean
R = UB h - LBi
N = the total number of items or observations.
Example 1. Find the range in the following set of numbers:
Example; Find the mean absolute deviation for the
11, 14, 23, 18, 16, 20 and 19.
following scores: 22, 24, 26, 28, 20 and 32.
Solution:
Solution: Construct the following table.
R = H – L = 23 – 11 = 12 Scores Scores from Deviation Absolute Deviation

Example 2. Find the rage for the grouped data in the table
(X) from Mean (x - x ) |x−x|
22
below:
24
R = 47.5 – 10.5 = 37
26
28
Class Interval Frequency (f) Cum F (<) 30
9 – 11 1 32
12 – 14 2 ϵ x= Ɛ |x−x| =
15 – 17 0
18 – 20 3 Computations:
21 – 23 3
24 – 26 8 εx 162
27 – 29 13 x= = = 27
30 – 32 3
n 6
33 – 35 2
36 – 38 4
ε |x−x| 18
39 – 41 4 M.A.D. = = =3
44 – 42 4 n 6
45 – 47 3
N = 50
 Mean Absolute Deviation for Grouped Data

For a Sample:

εf |M −x| Standard Deviation of Ungrouped Data

M.A.D. =
n

√ ∑ [ x−x ]
2
S=
For a population:
n

M.A.D. =
∑ f |M −u| Where:
N S = standard deviation (sometimes, known as root mean
Where: square of a sample)
M.A.D. = mean absolute deviation X = observed value
f = class frequency x = sample mean
M = class mark n = total number of observed values
x = sample mean
u = population mean In modern approach to find the standard deviation of
n = total frequency (sample size) ungrouped data is to divide the sum of the squared
N = total frequency (population size) deviations by (n-1), that is

√ ∑ [ x−x ]
Solution: Find the midpoints or class marks from the following 2

table. S=
n−1
Class Mark (f) fM |M −x| f |M −x|
9 +11/2 = 10 1 However for the purpose of studying we will use the first
12+14/2 = 13 2 formula.
15+17/2 = 0
18+20/2 = 3 Standard deviation for a given population
21+23/2 = 3
24+26/2 = 8
27+29/2 = 13
30+32/2 = 3
33+35/2 = 2 Where:
36+38/2 = 4 δ = standard deviation for the population
39+41/2 = 4 = Greek small letter sigma.
44+42/2 = 4
45+47/2 = 3 Example:
N = 50 ƐfM = Ɛf |M −x| = Find the standard deviation of the following set of raw data.
Students’ Scores (X) [ x−x ] [ x−x ] 2 x²
Computations: 40
45
ε fM εf |M −x| 55
X= M.A.D. =
n n 65
70
75
80
85
90
95 2
Ʃ[ x−x ] =
2
Ʃx =

Another formula in solving for standard deviation of

ungrouped data is:
 31 3

√ [ ]
34 2
∑ x2 − ∑ x
2
S= 37 4
n n 40 4
43 4
46 3
2
N = 50 εf [ M −x ]

∑ - ∑
[ ]
2 2
2 fM fM
Standard Deviation for Grouped Data S =
n n
The standard deviation of a frequency distribution can be
computed by the following formula:

√ ∑ f [ M −x ]
2
S=
n Variance for Ungrouped Data

Where: For Sample

S = standard deviation of a sample
M = class mark
S
2
= ∑ [ x −x ]2
x = sample mean
n
f = class frequency
n =total frequency
In modern approach to find the variance for ungrouped data
is to divide the sum of the squared deviations by (n-1), that is
For easier computations, use the formula

∑
√
[ x −x ]2
∑f M
[∑ ]
2 2 2
fM S =
S= − n−1
n n
Another formula in solving for variance of ungrouped data is:
To compute the standard deviation of a population in a

[ ]
frequency distribution use any of the following:
∑ x2 - ∑ x
2
2
S =

√
n n
∑ f [ M −u ]
2
δ=
N Variance for a given population

Where:
δ = standard deviation for a population 2
δ =
∑ [ x −x ]2
f =class frequency N
M = class mark
m = population mean Variance for Grouped Data
N = total frequency
For Sample
Find the sample standard deviations
Class Mark (f) M-x [ M −x ] 2 f [ M −x ]
2
The variance of a frequency distribution can be computed by
10 1 the following formula:
13 2
16 0
19 3
S =
2 ∑ f [ M −x ] 2 or 2
S =S
22 3 n
25 8
28 13
For easier computations

[
∑ f M 2 - ∑ fM
]
2
2
Corporation x S
S = ABCD 134.4 2.6
n n
WXYZ 179.5 3.77
Z-corp 98.6 3.72
For population
a) Coefficient of variation of each stock
S
CV = x 100%
x
δ
2
= ∑ f [ M −u ]
2
2.6
N For: ABCD: CV = x 100% = 1.9345
134.4

3.77
Coefficient of Variation For: WXYZ: CV = x 100% = 2.1360
179.5
It is often difficult to use our standard deviation formula to
compare measurements from different populations. Due to
3.72
For: Z-corp: CV = x 100% = 3.7728
this fact, statisticians produced the coefficient of variation. 98.6
The coefficient of variation expresses the standard deviation
as a percentage of what is being measured relative to the b) Hint: If you take the point of view that the
sample or population mean. coefficient of variation represents volatility or the
level of activity of a stock, then you would use the
Dow Jones Industrial Average as a measure of
overall market activity. This would indicate that
For Sample Data stocks with a coefficient of variation below 2.40
were less active than the market in general, and a
S stock with a coefficient of variation above 2.40
CV = x 100% would be more active. Now, comment on each
x
stock:
Where:
x = the sample mean CV of ABCD 1.9345 is below 2.40 which means less active in
s = the standard deviation the market
CV = Coefficient of variation CV of WXY 2.1360 is below 2.40 which means less active in
the market.
For Population Data CV of Z-corp 3.7728 is above 2.4 which means more active in
the market.

δ Seat Work:
V= x 100%
μ 1. Given a sample of IQ scores given by 96, 104, 126,
Where: 134 and 140. Find the
δ = standard deviation a) M.A.D.,
μ = Population mean b) Variance,
c) Standard deviation and
Example: d) The coefficient of variance.

1. During April of 1999, the daily closing of the ABCD,

IQ Scores Scores from Deviation Absolute Deviation
WXY, and Z-corps, gave the following information:
a) For each stock, compute the coefficient of
(x) from Mean (x - x ) |x−x|
variation. 96
b) Comment on the results of each stock. 104
126
Solution: 134
140
Let: ϵ x= Ɛ |x−x| =
x = Mean closing values for April 1999 Solution:
S = Standard deviation of closing values for 1989
 εx
a) x =
n

ε |x−x|
M.A.D. =
n

IQ Scores (X) [ x−x ] [ x−x ] 2 x²

96
104
126
134
140

b) Variance
2
S =
∑ [ x −x ] 2

n−1

√ ∑ [ x−x ]
2
c) S =
n−1

S
d) CV = x 100%
x

2. The amount of time (in minutes) that a sample 3. Solve for the range, quartile deviation, average
of students spends watching television per day deviation, standard deviation, variance and
is given below. coefficient of variation
3a.
40 25 35 30 20 40 30 20 x f
40 10 30 20 10 5 20 50-59 4
60-69 13
(a) Construct a frequency distribution, a cumulative 70-79 14
frequency distribution and a relative frequency 80-89 15
distribution. Let the first class be 1-10.
90-99 23
(b) Compute the mean, median, mode.
100-109 12
(c) The standard deviation.
110-119 10
(c) The coefficient of variation.
120-129 10
(d) The 40th percentile.
Ʃx = n=
(e) The variance
(f) The interquartile range.
 3b.
4. Calculate the variance of the following samples
x F 2, 10, 6, 18, 14.
18-20 3
21-23 3
24-26 6
27-29 5
30-32 7
33-35 10
36-38 5
39-41 9
42-44 4
45-47 4
48-50 3
51-53 0
54-56 0
57-59 1
Ʃx = N=
5. Calculate the variance of the population
consisting of the integers 1, 2, 3, 4, 5 and 6.