Measures of Dispersion or Variability

The common measures used to determine variability or dispersion are the range,
the variance, the standard deviation and the coefficient of variation. The higher the
computed value of dispersion, the more disperse or variable the distribution of the
observations or cases.

Measures of Dispersion or Variability for Ungrouped Data

Range

The range is a quick measure of dispersion or variability. The range(R) is the difference
between the highest value and lowest value of a set of observations.
R = Highest value- Lowest value

Example. The following are the estimated number of hours in which 10 students spent in
the library per day. Find the range
2.5; 3.0; 3.75; 1.0; 1.5; 3.0; 4.0; 1.75; 0.5; 2.0
R = 4.0 – 0.5 = 3.5 hrs

Variance
The variance is the average of the squared deviations from the mean.

∑ ∑ ∑ ̅
Sample Variance or

∑ ∑
Population Variance or

where - is the sample variance - is the population variance

̅ - is the sample mean - is the population mean
- is the respective observation or case
n- is the number of observations or cases in the sample
N- is the number of observations or cases in the population

Standard Deviation

The standard deviation is generally the most useful measure of dispersion. It is

simply obtained by getting the square root of the variance.

∑ ∑
Sample Standard Deviation √ = √
Population Standard Deviation √∑ = √

where
- is the sample standard deviation - is the population standard deviation
- is the sample variance - is the population variance
̅ - is the sample mean - is the population mean
- is the respective observation or case
n - is the number of observations or cases in the sample
N - is the number of observations or cases in the population

Example. The different weights of packages in kilograms carried by 6 delivery men of a

courier company at a particular time are 3.2; 4.5; 2.0; 9.2; 5.5 and 7.6
The mean of the weights is ̅
Following the table
Weights Deviation (xi -̅) (xi -̅)2
(xi)
3.2 3.2 – 5.33 = -2.13 4.54

4.5 4.5 – 5.33 = -0.83 0.69

2.0 2.0 – 5.33 = -3.33 11.1

9.2 9.2 – 5.33 = 3.87 15

5.5 5.5 – 5.33 = 0.17 0.03

7.6 7.6 – 5.33 = 2.27 5.15

∑ ̅

∑ ̅
The variance is
And the standard deviation of the weights carried by the delivery men is
√ √
The weights of packages are 2.7 standard deviations from the mean weight.
Coefficient of Variation

The coefficient of variation (CV) is a measure of dispersion, expressed in

percentage, and it is used to compare variability of set of data with different units. The CV
is the quotient of the standard deviation and the mean multiplied by 100.

Coefficient of Variation for the population

Coefficient of Variation for the sample ̅

Example. The average of the scores in Math 10 prelim exam of section A students is 85 with
a standard deviation of 2.5 while the averagescoresof section B students is 90 with a
standard deviation of 4.5. Determine which section have scores that are more variable
using coefficient of variation.

Section A :

Section B :

Thus, section A students have more variable scores in Math 10 prelim exam.

Other measures of Relative Position

Percentiles. Percentiles are measures that divide an ordered set of observations or data
into 100 equal parts. The measures are denoted P1; P2; P3; …; P99. For example, the 1st
percentile (P1) means that 1% of the observations falls below the value of P1; the 30th
percentile (P30) means that 30% of the observations fall below P30; and the 98th percentile
means that 98% of the observations fall below P98.

For ungrouped data

First the observations must be arranged in order from lowest to highest value and
the following sample formula will be used.Just follow the pattern when you want to
determine other positions.

P8 is the value of observation (x) in the position 0.08(n+1)

P25 is the value of observation (x) in the position 0.25(n+1)
P50 is the value of observation (x) in the position 0.50(n+1)
P90 is the value of observation (x) in the position 0.90(n+1)

For grouped data

Formula: [ ]
where - is the desired relative position (i.e. P10; P25; P90)
L - is the lower class boundary of the class where (Pn)th value belongs.
F - is the cumulative frequency of the class interval immediately before
the class where (Pn)th value belongs.
f- the frequency of the classwhere (Pn)th value belongs
c - is the class size or class width

Quartiles. Quartiles are measures that divide an ordered set of observations or data into
4 equal parts. The measures are denoted Q1; Q2; and Q3. For example, the 1st quartile (Q1)
means that 25% of the observations falls below the value of Q1; the 2ndquartile (Q2) means
that 50% of the observations fall below Q2 ; and the 3rd quartile means that 75% of the
observation fall below Q3.

For ungrouped data

First the observations must be arranged in order from lowest to highest value and
the following sample formula will be used. Just follow the pattern when you want to
determine other positions.

Q1 is the value of observation (x) in the position 0.25(n+1)

Q2 is the value of observation (x) in the position 0.50(n+1)
Q3 is the value of observation (x) in the position 0.75 (n+1)

For grouped data

Formula: [ ]

where - is the desired relative position (i.e. Q1; Q2; Q3)

L - is the lower class boundary of the class where (Pn)th value belongs.
F - is the cumulative frequency of the class interval immediately before
the class where (Pn)th value belongs.
f - the frequency of the classwhere (Pn)th value belongs
c - is the class size or class width

Deciles. Deciles are measures that divide an ordered set of observations or data into 10
equal parts. The measures are denoted D1; D2;D3;…D9. For example, the 1st decile (D1)
means that 10% of the observations falls below the value of D1; the 2nddecile (D2) means
that 20% of the observations fall below the value of D2 ; and the 3rddecile means that 30%
of the observation fall below D3.

For ungrouped data

First the observations must be arranged in order from lowest to highest value and
the following sample formula will be used. Just follow the pattern when you want to
determine other positions.
D1 is the value of observation (x) in the position 0.10(n+1)
D3 is the value of observation (x) in the position 0.30(n+1)
D5 is the value of observation (x) in the position 0.50(n+1)
D7 is the value of observation (x) in the position 0.70(n+1)

For grouped data

Formula: [ ]

where - is the desired relative position (i.e. D1; D2; D3; D4)
L - is the lower class boundary of the class where (Pn)th value belongs.
F - is the cumulative frequency of the class interval immediately before
the class where (Pn)th value belongs.
f - the frequency of the classwhere (Pn)th value belongs
c - is the class size or class width

Example:

For Ungrouped data

The following are scores of 13 students in an exam in history: 70, 85, 77, 94, 90,
79,78, 75, 82, 72, 93, 74, 86.
You have to arrange the data set or scores in increasing order: 70, 72, 74, 75, 77,78, 79, 82,
85, 86, 90, 93, 94.

1. Calculate the 50th percentile or P50: The position of x in the 50th percentile or
7
From the arranged data set, the score corresponding to the 7th position is 79.
So
If the student’s score is in the th percentile, you can now say that 50% of the scores fall
below 79 or the student is higher than 50% of the students who took the history exam.

2. Calculate the 90th percentile or P90: The position of x in

. Note that the result is not a whole number. From the
arranged data set, take first the score on the 12th position which is 93 and include the score
in the next higher position (the 13th), that is 94, in the computation as shown below:
The 12.6th position is
So

3. Calculate the 3rd quartile or Q3 : The position of x in

. Note that the result is not a whole number. From the arranged
data set, take first the score on the 10th position which is 86 and include the score in the
next higher position (the 11th), that is 90, in the computation as shown below:
The 10.5th position is
So

4. Calculate the 1st decile or D1 : The position of x in

 . Note that the result is not a whole number. From the arranged
data set, take first the score on the 1st position which is 70 and include the score in the next
higher position (the 2nd), that is 72, in the computation as shown below:
The 1.4th position is
So

For grouped data

Frequency distribution of students’ scores in mathematics exam.

Class Interval Class Class Cumulative
Frequency boundary frequency
(Scores)
(f)
43 – 46 1 42.5 – 46.5 50
39 - 42 3 38.5 – 42.5 49
35 - 38 4 34.5 – 38.5 46
31 - 34 5 30.5 – 34.5 42
27 - 30 10 26.5 – 30.5 37
23 - 26 9 22.5 – 26.5 27
19 - 22 7 18.5 – 22.5 18
15 - 18 6 14.5 – 18.5 11
11 - 14 3 10.5 – 14.5 5
7 - 10 2 6.5 – 10.5 2
n = 50

1. Find the 30th percentile or P30

Note: Pnis 30% of 50, that is = (0.30)(50) = 15
The class where Pnth value = 15 is reached using the cumulative frequency is
the class 19 -22.
[ ]

[ ]

means that 30% of the scores fall below 21 or the

student’s score is in the th percentile and he/she is higher than
 30% of the other examinees in the exam. However this does not
necessarily mean that the student got 30 on the examination.

2. Find the 2nd quartile or Q2

Note: Pn is 50% of 50, that is = (0.50)(50) = 25
The class where Pnth value = 25 is reached using the cumulative frequency is the
class 23 -26.
[ ]

means that 50% of the scores fall below 26 or the

student is higher than 50% of the other examinees in the
mathematics exam.

3. Find the 7th decile or D7

Note: Pn is 70% of 50, that is = (0.70)(50) = 35
The class where Pnth value = 35 is reached using the cumulative frequency is the
class 27 -30.
[ ]

means that 70% of the scores fall below 30.