Measures of Variation/ Dispersion

Ms. G.D Manyeagae

Ms. G.D Manyeagae Measures of Variation/ Dispersion 1 / 21

Introduction

A measure of dispersion indicates the extent to which a distribution is

stretched or squeezed.
It is a nonnegative real number that increases as the data become more
diverse.
We define various measures of dispersion, as illustrated in the sequel.
1 Range
This is the simplest measure of dispersion. The range of a set of data is the
difference between the largest value and the smallest value. That is, for an
increasingly ordered set of observations x1 , x2 , ..., xn ;
Range=xn − x1
When two datasets have the same mean but different ranges, there is wider
dispersion in the dataset with a larger range

Ms. G.D Manyeagae Measures of Variation/ Dispersion 2 / 21

Example

The height, in inches, of the 5 starting players on each of 2 men’s college

basketball teams are as follows:
Team I: 72, 73, 76, 76, 78 and
Team II: 67, 72, 76, 76, 84.
a Find the range for each team.
b Which team has more variability in the heights of players?
Solution
Notice that the two teams have the same mean heights (75 inches), the same
median heights (76 inches) and the same modal heights (76 inches). Now
Team I: Range=78 − 72 = 6 inches and
Team II: Range=84 − 67 = 17 inches.
Therefore Team II has more variability in the heights of players since it has a
larger range than Team I.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 3 / 21

Interquartile Range,QD, Mean and Median deviation

2 The interquartile range (IQR) is defined as the difference between the

75th and 25th percentiles, or between the upper and lower quartiles, ie
IQR = Q3 − Q1
3 The semi-interquartile range (SIQR), also called the quartile
deviation (QD), is one-half the difference between the first and third
quartiles, ie
SIQR = QD = 12 (Q3 − Q1 )

Ms. G.D Manyeagae Measures of Variation/ Dispersion 4 / 21

Example

Consider the wheat production (in Kg) of 20 acres given below:

1120, 1240, 1320, 1040, 1080, 1200, 1440, 1360, 1680, 1730, 1785, 1342,
1960,1880, 1755, 1720, 1600, 1470, 1750, and 1885.
Given that Q1 = 1260 and Q3 = 1753.75
Compute:
i The Interquartile range and
ii quartile deviation
Solution
i IQR = Q3 − Q1 = 1753.75 − 1260 = 493.75
Q3 −Q1 493.75
ii QD = 2 = 2 = 246.875

Ms. G.D Manyeagae Measures of Variation/ Dispersion 5 / 21

Mean and Median Deviation
The common shortcoming of the range, IQR, and the QD is that they only use
two values from the entire data set.
A better measure which is computed by using all the observations in a data set
is the mean absolute deviation and median absolute deviation.
3 Mean deviation:
We define
Pn Mean deviation as;
|X −X̄ |
MD= i=1 n i
4 Median deviation:
The median deviation
Pn is given as;
|X −median|
Median deviation= i=1 i n
Example
Consider the following set of data set:
2, 2, 3, 4, 14
Compute:
i Mean deviation
ii median deviation

Ms. G.D Manyeagae Measures of Variation/ Dispersion 6 / 21

Solution

2+2+3+4+14 25
Clearly the median=3 and the mean, X̄ = 5 = 5 =5

Table: Absolute Deviation

Pn Pn
i=1 |Xi −X̄ | i=1 |Xi −median|
n n
2(|2−5|)+|3−5|+|4−5|+|14−5| 2(|2−3|)+|3−3|+|4−3|+|14−3|
5 5
3.6 2.8

Ms. G.D Manyeagae Measures of Variation/ Dispersion 7 / 21

Mean and Median absolute deviation for grouped data

For grouped
Pn
data say with k groups, the Mean deviation is given by;
i=1 fi ×(|Xi −X̄ |)
MD= n
Where xi represent the class mark for the i th class
the frequency of the i th class, and
fi is P
k
n = i=1 fi
Similarly the
PnMedian deviation is given by; Median
f ×(|X −median|)
deviation= i=1 i ni
Where xi , fi and n are defined as before.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 8 / 21

Example
Suppose a farmer grew 50 baby carrots using special soil. The farmer digs them
up and measures their length (to the nearest millimetres) and then groups the
results as follows:

Table: Length of baby carrots Produced by a farmer using special soil

Length No of carrots
150-154 5
155-159 2
160-164 6
165-169 8
170-174 9
175-179 11
180-184 6
185-189 3
50

We want to calculate the the mean deviation and median deviation.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 9 / 21

Solution

Table: Length of baby carrots Produced by a farmer using special soil

X No of carrots xi |xi − x̄| |xi − x̄|fi |Xi − median| |Xi − median|fi

150-154 5 152 18.6 93 19.722 98.61
155-159 2 157 -3.6 27.2 14.722 29.444
160-164 6 162 8.6 51.6 9.722 58.332
165-169 8 167 3.6 28.8 4.722 37.776
170-174 9 172 1.4 12.6 0.278 2.502
175-179 11 177 6.4 70.4 5.278 58.058
180-184 6 182 11.4 68.4 10.278 61.668
185-189 3 187 16.4 49.2 15.278 45.833
50 355.2 392.223

Then MD= |xi −x̄|f

n
i
= 355.2
50 = 7.104, and
|Xi −median|fi
Median deviation= n = 392.233
50 = 7.844, where the median = 171.722.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 10 / 21

Variance & Standard deviation

5 Variance
To find the indicator of the overall variation, we take the average of the
squared deviations.
If the deviations are squared and averaged over all the observations, this
yields a measure called variance.
The variance is useful for comparing the variability of two or more separate
data sets that have approximately the same mean.
The data set with the smaller(smallest) value of the variance has less (the
least) variability of measurements around the mean.
The square root of the variance is called the standard deviation.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 11 / 21

Variance and Standard deviation for ungrouped data

Let X1 , X2 , ..., XN be a set of observations for a population of size N. Then

the variance
PN
is given by;
2
2 i=1 (Xi −µ)
σ = N
If we have
Pn
a sample of size n, then we define the variance as;
(x −x̄)2
S 2 = i=1n−1i , which ca be writen as
Pn
xi2 −nx̄ 2
S2 = i=1
n−1
Example
A hen lays eight eggs. Each egg was weighed and the weight, in kilograms,
recorded as follows: 60g, 56g, 61g, 68g, 51g, 53g, 69g, 54g.
Compute:
i The variance
ii The standard deviation

Ms. G.D Manyeagae Measures of Variation/ Dispersion 12 / 21

Solution
P8
Xi 472
First calculate the mean, X̄ = i=1
n = 8 = 59

Table: Weight of eggs in grams

x (xi − x̄) (xi − x̄)2 xi 2

60 1 1 3600
56 -3 9 3136
61 2 4 3721
68 9 81 4624
51 -8 64 2601
53 -6 36 2809
69 10 100 4761
54 -5 25 2916
472 0 320 28168
Pn 2
i=1 (xi −x̄) 320
Thus variance= n−1 = 7 = 45.714 OR alternatively
Pn 2 2
i=1 xi −nx̄ 28168−8(592 ) 320
n−1 = 7 = 7 = 45.714

Ms. G.D Manyeagae Measures of Variation/ Dispersion 13 / 21

Solution...

√ √
Then the standard deviation is given as Var = 44.714 = 6.761
Notice
Pn that the sum of deviation is always =o, ie
i=1 (X I − X̄ ) = 0
Variance & Standard deviation for grouped Data
As mentioned before, for grouped data the observations in a class are
represented by the class mark, xi , with k number of groups (classes)
Then the variance
Pk can be calculated as follows:
(x −x̄)2 fi
Var(x) = i=1 ni which can be writen as
Pk
xi2 fi −nx̄ 2
Var(x) = i=1
n

Ms. G.D Manyeagae Measures of Variation/ Dispersion 14 / 21

Example

Use the above example to compute variance and standard deviation.

Ms. G.D Manyeagae Measures of Variation/ Dispersion 15 / 21

Solution

Table: Length of baby carrots Produced by a farmer using special soil

x No of carrots xi fi xi (xi − x̄) (xi − x̄)2 fi xi 2 fi

150-154 5 152 760 -18.6 1729.8 115520
155-159 2 157 314 -13.6 369.92 49298
160-164 6 162 972 -8.6 443.76 157464
165-169 8 167 1336 -3.6 103.68 223112
170-174 9 172 1548 1.4 17.64 266256
175-179 11 177 1947 6.4 450.56 344619
180-184 6 182 1092 11.4 779.76 198744
185-189 3 187 561 16.4 806.88 104907
50 8530 4702 1459920

Ms. G.D Manyeagae Measures of Variation/ Dispersion 16 / 21

Solution...

ThenPusing results from the table

k
xf
x̄ = i=1n i i = 8530
50 = 170.6 and
Pk 2
i=1 (xi −x̄) fi 4702
Var(x) = n = 50 = 94.04 OR
Pk 2 2
i=1 xi fi −nx̄ 1459920−50(170.62 )
Var(x) = n = 50 = 4702
50 = 94.04, and hence
The standard
p deviation√is given as;
SD(x) = Var (x) = 94.04 = 9.697

Ms. G.D Manyeagae Measures of Variation/ Dispersion 17 / 21

Coefficient of Variation
Coefficient of Variation compares variability in amongst datasets
expressed in different units of measurement
The coefficient of variation (CV) is defined as the ratio of the standard
deviation to the mean. It is often expressed as a percentage. Thus;
CV = standarddeviation
Mean × 100
⇒ CV = Sx̄x × 100
Example
A researcher is comparing two multiple-choice tests with different conditions. In
the first test, a typical multiple-choice test is administered. In the second test,
alternative choices (i.e. incorrect answers) are randomly assigned to test takers.
The results from the two tests are recoded in the table below.

Table: Multiple choice Tests

Regular test Randomized Answer

Mean 59.9 44.8
SD 10.2 12.7
cv 17.03 28.35

Ms. G.D Manyeagae Measures of Variation/ Dispersion 18 / 21

Example...

Regular test: CV = 17.03 and Randomized answers: CV = 28.35

These results indicate that there is more variability in the scores for the
Randomized answers than those for the Regular test.
Choosing the Appropriate Measure of variability
If the data is symmetric with no serious outliers use the range and standard
deviation
If the data is skewed and or have serious outliers , use the IQR
When comparing variation across two data sets, use the coefficient of
variation

Ms. G.D Manyeagae Measures of Variation/ Dispersion 19 / 21

Exercise 1

The following data represent the scores of 9 students in a test out of 15 marks.
11, 8, 13, 6, 12, 7, 10, 9, 11
Compute:
i Mean deviation
ii Median deviation
iii The variance
iv The standard deviation

Ms. G.D Manyeagae Measures of Variation/ Dispersion 20 / 21

Exercise 2

Use the data below to find the standard deviation.

Table: Table 1

class limits frequency

0-9 4
10-19 9
20-29 6
30-39 4
40-49 2
25

Ms. G.D Manyeagae Measures of Variation/ Dispersion 21 / 21