Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank

You

Chapter 4
Measures of Variation/Dispersion

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

Institute of Public Health

Departement of Epidemiology and Biostatistics

April 3, 2019

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 1 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Schedule

1 Introduction

2 Objective of Measuring Variation

3 Types of measure of variation

4 Variance and Standard Deviation

Population and sample variance
Special properties of standard deviation /variance

5 Coefficient of variation (CV)

Standard Scores (Z-Scores)
Moments
Skewness
kurtosis

6 Thank You

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 2 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Introduction

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 3 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Introduction

Example
Consider the following two sets of scores:Set 1: 40, 50, 60, 60, 40, 50Set 2: 0,
100, 25, 75, 80, 20

Alert block

Both these sets have the same mean (50),

but the second set is a lot more widely dispersed ("scattered") than the first....

Figure: Example images from the data set

Alemakef Wagnew(Bsc and MPH in EPi-BIO)
UoG Chapter 4 April 3, 2019 4 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Measure of variation/dispersion

The scatter or spread of items of a distribution is known as dispersion or variation.

In other words the degree to which numerical data tend to spread about an
average value is called dispersion or variation of the data.

Measures of dispersion are statistical measures which provide ways of measuring

the extent in which data are dispersed or spread out.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 5 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Objective of Measuring Variation

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 6 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Objective of measuring variation

To determine the reliability of an average by pointing out as how far an average is

representative of the entire data.

To determine the nature and cause of variation in order to control the variation
itself.

Enable comparison of two or more distribution with regard to their variability.

Measuring variability is of great importance to other statistical analysis. E.g., it is

the basis of statistical quality control

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 7 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

A good measure of variation

It should be easy to compute and understand.

It should be based on all observations.

It should be Uniquely defined

It should be capable of further algebraic treatment.

It should be as little as affected by extreme values

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 8 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Types of measure of variation

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 9 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Types of measure of variation

Absolute measure:

Range
Quartile deviation
Mean deviation
Variance
Standard deviation

Relative measures:

Relative range
Coefficient of quartile deviation
Coefficient of mean deviation
Coefficient of variation
Standard scores

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 10 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

The range

Several measures of dispersion are available. We will discuss the common ones below.
The Range:
The difference between the largest (maximum) and smallest (minimum) values.

Range = Maximum − Minimum (1)

 For frequency distributed data, the range is:

The difference between the upper class boundary of the last class and the lower
class boundary of the first class.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 11 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Measure of dispersion

Given the following frequency

distribution. Find the range
Measure of variation-dispersionn Class frequency
52.5-63.5 6
Find the Range of 54.5, 55.0, 55.7, 63.5-74.5 12
51.8, 54.2, 52.4 Solution: 74.5-85.5 25
range(R) = 55.7- 51.8 = 3.9cm 85.5-96.5 18
96.5-107.5 14
107.5-118.5 5

Solution:Range =UCBl − LCBf = 118.5 − 52.5 = 66

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 12 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Measure of variation

The relative range:

l −s
RR = relative range = (2)
l +s
we adjust similarly for Grouped data
Quartile deviation (QD):
The range expresses the extreme variability of observations of a variable. is half
of the inter quartile range.

InterquartileRange = Q3 − Q1
Interquartilerange
QD =
2
Q3 − Q1
QD =
2

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 13 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Measure of variation

Coefficient of quartile deviation (CQD):

It gives the average amount by which the two quartiles differ from the median

Q3 − Q1
CQD = (3)
Q3 + Q1

Mean Deviation(M.D):
The average deviation measures the scatter of the individual observations around a
central value usually the mean or the median of a distribution.
The mean deviation is defined as the arithmetic mean of positive deviations of each
observation from either the mean or the median of a distribution.
If the deviations are taken from the mean then it is called mean deviation about the
mean. On the other hand, if the deviations are taken from the median we call it mean
deviation about the median.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 14 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

mean deviation about the mean

The mean Deviation (M.D) is the arithmetic mean of the absolute deviations of the
values from the mean.
It is the “average absolute deviation of the values from the mean”.
Pn
i=1 | Xi − x̄ |
Mean deviation = (4)
n
Note that:while dealing with population values, it is adjusted accordingly
Mean Deviations for Grouped data (discrete or continuous)
Pn
i=1 fi | Xi − x̄ |
Mean deviation = (5)
n

Where m = number of classes and xi = class mark of thei th class, n = number of observation

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 15 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

mean deviation

Mean deviation about the median ( MD())

ungrouped data:
Pn
i=1 | Xi − x̂ |
MD(X̂ ) = (6)
n
grouped Frequency Distribution:
Pk
i=1 fi | Xi − x̂ |
MD(X̂ ) = (7)
n

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 16 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Example

The weights of a sample of six students from a class (in kilograms) is measured
as: 53, 56, 57, 59, 63 and 66. Find the mean deviation about the mean and the
mean deviation from the median.
solution:First find the mean and the median. The mean is 59 kg and the median
is 58 kg. Then take the deviations of each observation from these averages as
shown below
weight Xi Ad from mean| xi − x̄ | AD from median| xi − x̂ |
53 6 5
56 3 2
57 2 1
59 0 1
63 4 5
66 2 8
Total 22 22

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 17 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

example continued

mean deviation a bout the mean:

Pn
i=1 fi | Xi − x̄ |
MD(X̄ ) =
n
22
= MD(x̄) =
6
= 3.67
mean deviation about median:

Pk
| Xi − x̂ |
i=1 fi
MD(X̂ ) =
n
22
MD(x̂) = = 3.67
6

Example 4.4: Calculate the mean deviation from the mean and median for the following
Class interval 1-5 6-10 11-15 16-20
data.
Frequency 4 1 2 3
Alemakef Wagnew(Bsc and MPH in EPi-BIO)
UoG Chapter 4 April 3, 2019 18 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

solution:

3 ∗ 4 + 8 ∗ 1 + 13 ∗ 2 + 18 ∗ 3
Mean = (8)
10

CL Xi fi fi | xi − 10 |
1-5 3 4 28
6-10 8 1 2
11-15 13 2 6
16-20 18 3 24
Total 10 60
Therefore MD around the mean :
60
MD =
10
=6

Exercise: Mean Deviation about the Median

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 19 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Coefficients of Mean Deviation(C.M.D)

C.M.D = M.D/Average about which deviations are taken

Coefficient of mean deviation about the

MD X̄
CMD(X̄ ) =
X̄
Coefficient of mean deviation about the median=

MD X̂
CMD(X̂ ) =
X̂

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 20 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Example

Example 4.5
Find the coefficient of mean deviation about the mean and mean deviation about
the median for the weights of six students in example above.
Solution:
Coefficient of mean deviation about the mean

MD X̄
CMD(X̄ ) =
X̄
3.67kg
=
59kg
= 0.0622

Coefficient of mean deviation about the median

MD X̂
CMD(X̂ ) =
X̂
3.67kg
=
58kg
= 0.0632

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 21 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Variance and Standard Deviation

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 22 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Variance and Standard Deviation

The variance and standard deviation are the most superior and widely used
measures of dispersion

Both measures the average dispersion of the observations around the mean.

The variance is defined as the average of the squared deviation from the mean.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 23 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Population and sample variance

population and sample variance

PN
i=1 (xi − µ)
population variance = σ 2 = i = 1, 2, 3.......N
N
for the case of frequency distribution
Pk
i=1 fi (xi − µ)
σ2 = (9)
N
Where i=1,2,3..K and xi areclassmarks
Sample variance:the sample variance is dented by S2 andisgivenby

Pk
− X̄ )
i=1 fi (xi
sample variance = S 2 = , i = 1, 2, 3...n
n−1

Pk
− X̄ )
i=1 fi (xi
for the case of frequency distribution = S 2 =
n−1
X
n= fi
where i = 1, 2, 3..K and xi are class marks

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 24 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Population and sample variance

Standard Deviation

The positive square root of the variance is called standard deviation. Therefore
sP
p (xi − µ)2
population standard deviation σ 2 = (10)
N

sP
p (xi − X̄ )2
sample standard deviation = S2 =
n−1

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 25 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Population and sample variance

Example

24, 25, 29,29,30,31 Find variance and standard deviation ?

Solution:
Value value-mean(xi − X̄ ) difference Difference square(Xi − X̄ )2
24 24-28 -4 16
25 25-28 -3 9
29 29-28 1 1
29 29-28 1 1
30 30-28 2 4
31 31-28 3 9
total 0 40
Pk
− X̄ )
i=1 fi (xi
S2 = , i = 1, 2, 3...n
n−1
40 40
= , =
6−1 5
√
S 2 = 8, standard deviation, S = 8 = 2.83

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 26 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Population and sample variance

Exercise

Find the variance and standard deviation of the following sample data
i 5, 17, 12, 10,8
ii The data is given in the form of frequency distribution.

class Frequency
40-44 7
45-49 10
50-54 22
55-59 15
60-64 12
65-69 6
70-74 3

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 27 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Special properties of standard deviation /variance

Special properties of standard deviation /variance

The main drawback of variance => unit is squared and this is difficult to interpret.

Variance gives weight to extreme values than those near to the mean value. This
is because the difference is squared.

Variance will be zero for distributions with equal magnitude. The greater the
difference in the values, the greater the variance and vise versa.
sP sP
n 2 n 2
i=1 (Xi − X̄ ) i=1 (Xi − A)
< , X̄ 6= A
n−1 n−1
If the standard deviation of X1 , X2 , . . . Xn is S,
Then the standard deviation 0f:
X1 + K , X2 + k .......Xn + k will alsoS
KX1 , KX2 , ........KXn wouldbe |k |S
a+KX1 , a + KX2 , ........a + KXn would be |k |S

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 28 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Special properties of standard deviation /variance

special properties of standard deviation/variance

If a sample of n1 observationshasavarianceS12
and a sample of n2 observationshaveavarianceofS22
then the combined variance called the pooled variance (S2p )isgivenby :

(n1 − 1)S12 + (n2 − 1)S22

Sp2 = (11)
n1 + n2 − 2

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 29 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Coefficient of variation (CV)

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 30 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Coefficient of variation (CV)

In situations where either two series have different units of measurements, or their
means differ sufficiently in size, the CV should be used as a measure of
dispersion.

standard deviation
coefficient of variation(CV ) = ∗ 100%
mean
S
CV = ∗ 100%for sample and
X̄
σ
CV = ∗ 100%for population
µ

In spite of the fact that the C.V. is broadly applied, its disadvantage is that it’s not
useful when the mean is negative or zero or very close to zero.

Interpretation of the coefficient of variation: the distribution having less CV is

said to be less variable or more consistent

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 31 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

EXAMPLE

Example 4.7:For the garment length data mean = 53.6 and standard deviation = 1.46cm,
so that the coefficient of deviation is
Solution:
S
CV= ∗ 100%
X̄
1.46
CV= ∗ 100%,
53.6
CV=2.72%
Example 4.8 Suppose that the mean weight of a group of students is 165 pounds with
a S.D of 8 pounds. If the height of the same group of students has a mean of 60 inches
with a S.D of 3 inches, compare the variability in weight and height measurements.
Solution:
8lb
for weight CV= ∗ 100% = 4.85%
165lb
and for height:

3 in
CV= ∗ 100% = 5%
60 in

=> The height data is more variable/less consistent than the weight data.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 32 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Standard Scores (Z-Scores)

Standard Scores (Z-Scores)

Are not measures of relative dispersion, but one of the applications of standard
deviation.

We define the standard score as:

X − X̄ X −µ
Z = or Z = (12)
S σ

Tells us how many standard deviations a value lies above (if positive) or below (if
negative) the mean.

Standard score gives the deviations from the mean in units of standard deviation

It is used to compare two observations coming from different groups.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 33 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Standard Scores (Z-Scores)

Standard z-score

Questions:Two third year Medical laboratory sections were given introduction to bio-
statistics examinations. The following information was given.
Value Section1 section2
mean 78 90
Standard deviation 6 5
Student A from section1 scored 90 and student B from section2 scored 95.
Relatively speaking who performed better ?

X1 − X¯1 90 − 78
Z1 = = =2
S1 6
X2 − X¯2 95 − 90
Z2 = = =1
S2 5

Student A performed better relative to his section because the score of student A is two
standard deviation above the mean score of his section while, the score of student B is
only one standard deviation above the mean score of his section.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 34 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Moments

Moments

The rth moment about the mean (the r th central moment) defined as :

(Xi − X̄ )r
P
Mr = , r = 0, 1, 2.. (13)
n
for continuous grouped data it is given by:

fi (Xi − X̄ )r
P
Mr = , where Xi are class marks (14)
n
Example:Find the first three central moments of the numbers 2, 3 and 7
Solution first find the mean:
2+3+7
mean== =4
3

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 35 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Skewness

Skewness

Skewness is the degree of asymmetry or departure from symmetry of a

distribution.
A skewed frequency distribution is one that is not symmetrical.
Skewness is concerned with the shape of the curve not size
If the frequency curve (smoothed frequency polygon) of a distribution has a longer
tail to the right of the central maximum than to the left, the distribution is said to be
skewed to the right or said to have positive skewness. If it has a longer tail to the
left of the central maximum than to the right, it is said to be skewed to the left or
said to have negative skewness.
For moderately skewed distribution, the following relation holds among the three
commonly used measures of central tendency.

Mean − Mode = 3(Mean − Median) (15)

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 36 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Skewness

Skewness

Remarks:

In a positively skewed distribution, smaller observations are more frequent than

larger observations. i.e. the majority of the observations have a value below an
average.
In a negatively skewed distribution, smaller observations are less frequent than
larger observations. i.e. the majority of the observations have a value above an
average.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 37 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Skewness

Skewness

Example. Suppose the mean, the mode, and the standard deviation of a certain distri-
bution are 32, 30.5 and 10 respectively. What is the shape of the curve representing the
distribution?
Solution:

mean − mode
Sk =
standard deviation
32 − 30.5
= = 0.15
10

The distribution is positively skewed

Measures of Skewness
The Karl Pearson’s Coefficient of Skewness (SK):

mean − mode 3(mean − median)

Sk = , Sk =
standard deviation standard deviation

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 38 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Skewness

skewness

If SK = 0, then the distribution is symmetrical.

If SK > 0, then the distribution is positively skewed.

If SK < 0, then the distribution is negatively skewed

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 39 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

kurtosis

Kurtosis

Kurtosis is the degree of peakedness of a distribution, usually taken relative to a

normal distribution.

When the curve of a distribution is relatively:

flatter than normal it is known as platykurtic and

the distribution is more peaked than normal, it is called leptokuric.

The normal distribution which is not very high peaked or flat topped is called
mesokurtic.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 40 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

kurtosis

kurtosis

The moment coefficient of skewness ( β2 )

m4
β2 = (16)
(m2 )2
If B2 =3, then the distribution is mesokurtic.

If B2 > 3, then the distribution is leptokurtic.

If B2 < 3, then the distribution is platykurtic.

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 41 / 42
 Introduction Objective of Measuring Variation Types of measure of variation Variance and Standard Deviation Coefficient of variation (CV) Thank You

Thank You

Alemakef Wagnew(Bsc and MPH in EPi-BIO)

UoG Chapter 4 April 3, 2019 42 / 42