Samar Colleges, Inc.
A. Mabini Ave., Catbalogan City,
Samar, Phils 6700 Tel:
+6355-5438381/8594
Website: samarcollege.edu.ph
Email: samarcollegesinc@samarcollege.edu.ph
samarcollegesinc@gmail.com
HANDOUT ON MEASURES OF VARIABILITY
Measures of variability provide us information on how the data are scattered and
dispersed for one to have adequate information relative to a set of distribution. Measures
of dispersion indicate the heterogeneity or homogeneity of the distribution.
Example: To groups of sample may have the same arithmetic mean but the other group may
be variable than the other. Consider the two sets of scores in Statistics examination:
Set A: 15 20 25 27 28 30 30 35 40
_
X = 27.78
Set B: 5 10 20 23 25 29 35 48 55
_
X = 27.78
Note that the two sets of data have the same arithmetic mean, that is, 27.78 yet the
dispersion of the scores for the two sets differed. Set A appeared to be more homogeneous
than Set B.
Range
Range is the simplest and easiest measure of variability. It is easily obtained and
readily understood. However, since its magnitude depends only on the extreme values of the
set of data, it is the most unreliable.
Range maybe classified into: Absolute Range (AR), Total Range (TR), Interquartile
Range, and Kelly Range (KR).
Absolute Range(AR) is obtained by taking the difference between the highest and
lowest scores in a set of data. Hence,
AR = H–L
Where:
H refers to the highest score or value
L refers to the lowest score or value
Total Range (TR) is obtained by computing the highest score minus the lowest score
plus 1. Hence,
TR = (H – L) + 1
Where:
H refers to the highest score or value
L refers to the lowest score or value
Kelly Range (KR) is obtained by taking the difference between the ninetieth percentile
and tenth percentile. Hence,
KR = P 90 – P10
Where:
P90 refers to the ninetieth percentile
P10 refers to the tenth percentile
KR refers to the Kelly Range
Using the grouped data/frequency distribution in Module 2: The value of N is 100, thus dividing it by
4 we get 25. Notice that for class interval 79 – 83 the <F is 22, thus this class interval cannot be
considered as the Q1 class. However, for the next class interval, 84 – 88 the <F is already 29.
Hence, the Q1 class is 84 – 88.
=============================================
Class Intervals Class Midpoint f <F
-----------------------------------------------------------------------------
139 – 143 141 3 100
134 – 138 136 5 97
P90 class 129 – 133 131 f = 6 92
124 – 128 126 5 86 F
119 – 123 121 6 81
114 – 118 116 6 75
109 – 113 111 7 69
104 – 108 106 10 62
99 – 103 101 8 52
94 – 98 96 6 44
89 – 93 91 9 38
84 – 88 86 7 29
79 – 83 81 6 22
P10 class 74 – 78 76 f = 7 16
69 – 73 71 9 9 F
-------
N = 100
Interquartile Range (QR) is obtained by taking the difference between the 3 rd quartile
and the 1st quartile.
QR = Q3 – Q1
Where:
Q3 refers to the 3rd quartile; Q1 refers to the 1st quartile
QR refers to Interquartile Range
=============================================
Class Intervals Class Midpoint f <F
-----------------------------------------------------------------------------
139 – 143 141 3 100
134 – 138 136 5 97
129 – 133 131 6 92
124 – 128 126 5 86
119 – 123 121 6 81
Q3 class 114 – 118 116 f = 6 75
109 – 113 111 7 69 F
104 – 108 106 10 62
99 – 103 101 8 52
94 – 98 96 6 44
89 – 93 91 9 38
Q1 class 84 – 88 86 f = 7 29
79 – 83 81 6 22 F
74 – 78 76 7 16
69 – 73 71 9 9
-------
N = 100
Quartile Deviation
Quartile Deviation (QD) is a measure of variability which divides the difference
between the 3rd and 1st quartile into two. Sometimes quartile deviation is called semi-
interquartile range. Hence:
Q 3 - Q1
QD = ----------------
2
where: Q3 refers to the 3rd quartile
Q1 refers to the 1st quartile
Average Deviation
The average or mean deviation (AD) is a measure of variability which is defined as the
arithmetic mean of the absolute deviations of the values from the mean. Hence,
For ungrouped data:
_
/X–X/ where: X refers to the individual values
AD = --------------- X refers to the arithmetic mean
n n refers to the total number
of cases
For grouped data:
_
f / Xm – X / where: Xm refers to the midpoint of the class
AD = --------------- X refers to the arithmetic mean
n n refers to the total number of cases
f refers to the class freq.
Example 1 (ungrouped): Consider the following scores (n = 9):
7 8 9 14 14 18 20 20 25
The following steps maybe done to compute AD:
_
Step 1: Compute the mean, X = 15
Step 2: Generate the following:
_ _
X X–X /X–X/
7 -8 8
8 -7 7
9 -6 6
14 -1 1
14 -1 1
18 3 3
20 5 5
20 5 5
25 10 10
_________
_
46 /X–X/
Step 3: Apply the formula:
_
/X–X/ 46
AD = --------------- = ---------- = 5.11
n 9
Example 2 (grouped): Consider the grouped data/frequency distribution in Module 2:
=============================================
Class Intervals Class Midpoint f fXm
-----------------------------------------------------------------------------
139 – 143 141 3 423
134 – 138 136 5 680
129 – 133 131 6 786
124 – 128 126 5 630
119 – 123 121 6 726
114 – 118 116 6 696
109 – 113 111 7 777
104 – 108 106 10 1,060
99 – 103 101 8 808
94 – 98 96 6 576
89 – 93 91 9 819
84 – 88 86 7 602
79 – 83 81 6 486
74 – 78 76 7 532
69 – 73 71 9 639
---------------
10,240
Step 1: Compute the mean
_ fXm 10,240
X = ---------- = ---------- = 102.40
N 100
Step 2: Generate the following:
_ _
Class Intervals Xm f / Xm – X/ f/Xm – X/
---------------------------------------------------------------------------------------------------------------------
139 – 143 141 3 38.60 115.80
134 – 138 136 5 33.60 168.00
129 – 133 131 6 28.60 171.60
124 – 128 126 5 23.60 118.00
119 – 123 121 6 18.60 111.60
114 – 118 116 6 13.60 81.60
109 – 113 111 7 8.60 60.20
104 – 108 106 10 3.60 36.00
99 – 103 101 8 1.40 11.20
94 – 98 96 6 6.40 38.40
89 – 93 91 9 11.40 102.60
84 – 88 86 7 16.40 114.80
79 – 83 81 6 21.40 128.40
74 – 78 76 7 26.40 184.80
69 – 73 71 9 31.40 282.60 (9 x 31.40)
------ ----------- _
100 1,725.60 = f / Xm – X /
(71 – 102.40 )
Step 3. Apply the formula:
_
f / Xm – X / where: Xm refers to the midpoint of the class
AD = --------------- X refers to the arithmetic mean
n n refers to the total number of cases
f refers to the class freq.
1,725.60
-------------- = 17.26
100
Variance and Standard Deviation
Variance (Var) is sometimes called the mean square and is highly related to the
standard deviation (SD). These two values are determined as follows:
For ungrouped data: _
(X – X)2
Var = ------------
n
where X refers to the scores/values
X refers to the mean
n refers to the total number of cases
For grouped data:
_
f(Xm – X)2
Var = ----------------
n
where _
X refers to the mean
Xm refers to the class mark or midpoints
n refers to the total number of cases
f refers to the class frequencies
On the other hand, the standard deviations are computed by merely extracting the
square roots of the formulas for the variance both for the grouped and ungrouped data.
Hence,
For ungrouped data:
_
(X – X)2
SD = ------------
n
where X refers to the scores/values
X refers to the mean
n refers to the total number of cases
For grouped data:
_
f(Xm – X)2
SD = ----------------
n
where Xm refers to the class mark or midpoints
X refers to the mean
n refers to the total number of cases
f refers to the class frequencies
Example: Consider the same set of data used in the computation of AD for
ungrouped data (n = 9), viz:
7 8 9 14 14 18 20 20 25
The following steps maybe done to compute the Var and SD:
_
Step 1: Compute the mean, X = 15
Step 2: Generate the following:
_ _
X X–X ( X – X )2
7 -8 64
8 -7 49
9 -6 36
14 -1 1
14 -1 1
18 3 9
20 5 25
20 5 25
25 10 100
_________
_
310 ( X – X )2
Step 3: Apply the formula:
_
( X – X )2 310
Var = --------------- = ---------- = 34.44
n 9
_
(X – X)2
SD = ------------
n
310
SD = ------------
9
SD = 34.44 = 5.87
Example 2 (grouped): Consider the grouped data/frequency distribution in Module 2:
=============================================
Class Intervals Class Midpoint f fXm
-----------------------------------------------------------------------------
139 – 143 141 3 423
134 – 138 136 5 680
129 – 133 131 6 786
124 – 128 126 5 630
119 – 123 121 6 726
114 – 118 116 6 696
109 – 113 111 7 777
104 – 108 106 10 1,060
99 – 103 101 8 808
94 – 98 96 6 576
89 – 93 91 9 819
84 – 88 86 7 602
79 – 83 81 6 486
74 – 78 76 7 532
69 – 73 71 9 639
---------------
10,240
Step 1: Compute the mean
_ fXm 10,240
X = ---------- = ---------- = 102.40
N 100
Step 2: Generate the following:
_ _
Class Intervals Xm f Xm – X (Xm – X)2 f(Xm – X)2
---------------------------------------------------------------------------------------------------------------------
139 – 143 141 3 38.60 1,489.96 4,469.88
134 – 138 136 5 33.60 1,128.96 5,644.80
129 – 133 131 6 28.60 817.96 4,907.76
124 – 128 126 5 23.60 556.96 2,784.80
119 – 123 121 6 18.60 345.96 2,075.75
114 – 118 116 6 13.60 184.96 1,109.76
109 – 113 111 7 8.60 73.96 517.72
104 – 108 106 10 3.60 12.96 129.60
99 – 103 101 8 - 1.40 1.96 15.68
94 – 98 96 6 - 6.40 40.96 245.76
89 – 93 91 9 - 11.40 129.96 1,169.64
84 – 88 86 7 - 16.40 268.96 1,882.72
79 – 83 81 6 - 21.40 457.96 2,747.76
74 – 78 76 7 - 26.40 696.96 4,878.72
69 – 73 71 9 - 31.40 985.96 8,873.64
------ ----------------- _
100 41,454.00 = f( Xm– X )2
(71 – 102.40 )
Step 3: Apply the formula:
_
f( X – X )2 41,454
Var = --------------- = ---------- = 414.54
n 100
_
f(X – X)2
SD = ------------
n
41,454
SD = ------------
100
SD = 414.54 = 20.36
Note: If one is computing the variances or standard deviations of data taken from
samples and not the entire population, the formulas will change slightly inasmuch as the
divisor will now be n – 1.
Coefficient of Variation
The coefficient of variation is another measure of dispersion which is derived using
the following formula:
SD
C.V. = -------- x 100
X
where: SD refers to the standard deviation
X refers to the mean
The value is expressed in percent (%) and the bigger the value, the more dispersed is
the distribution and the smaller the value, the less dispersed is the distribution.
Remember:
1. The range, among the measures of variability presented is the largest in a given set of
distribution.
2. The quartile deviation includes approximately the middle 50% of the distribution.
3. The average deviation includes approximately the middle 60% of the distribution.
4. The standard deviation includes approximately the middle 68 % of the distribution.
Exercise:
Given the frequency distribution you have
constructed in Exercise 2 (Module 2): Compute the
following:
1. Absolute Range, Total Range & Kelly Range
2. Average Deviation
3. Quartile Deviation
4. Variance; and
5. Standard Deviation
