CHAPTER 6
MEASURES OF RELATIVE
DISPERSION AND MEASURES
OF SHAPE
MEASURES OF RELATIVE
DISPERSION
This is used when we wish to
compare the spread of one
distribution relative to the other
distribution. The most commonly
used measures of relative
dispersion are the following:
1. COEFFICIENT OF VARIATION (CV)
this is usually expressed in percent. It is
the ratio of the standard deviation to the
mean.
Formula:
CV= SD X 100
M
2. COEFFICIENT OF QUARTILE DEVIATION(CQD)
� this involves the use of the third and first
quartile of the distribution.
Formula:
CQD= Q3-Q1 X 100
Q3+Q1
3. THE STANDARD SCORE (Z-SCORE)
strictly speaking the standard score is not a
measure of relative dispersion however; it is
useful in comparing two values from two
different distributions especially when these
two distributions differ with respect to the
mean and standard deviation. The standard
score measures how many standard
deviations an observation is above or below
the mean.
Formula:
Z-Score = X – M
SD
Example
Below are the distributions of scores in the
unit test in Statistics and Physics class.
STATISTICS PHYSICS
Scores F Scores F
95-99 4 82-90 2
90-94 6 73-81 1
85-89 8 64-72 11
80-84 9 55-63 7
75-79 2 46-54 4
70-74 1 37-45 5
N=30 N=30
a. Find the CV and CQD in the two subjects
and compare which of the two distribution
exhibit homogeneity?
b. Supposing a student got a score of 80 in
Statistics and 65 in Physics , in subject did
he perform better?
Solution:
For a question A
Get the mean and standard deviation of the scores
in Statistics to compare for the value of the
coefficient of variation.
Using the previous formula and step we’ve learned
in getting mean and standard deviation of
grouped data, the mean of statistics scores is
86.67 and the standard deviation of grouped data,
the mean of statistics scores is 86.67 and the
standard deviation is 6.30
Computing for the coefficient of variation
CVs= 6.30 X 100%
86.67
CVs = 7.27%
To compute for the CQD, we need to solve for
the value of Q3 and Q1 . Again using the
formula in finding the Quartiles of grouped
data, Q1 = 82 and Q3 = 91.58
CQDs = 91.58 – 52 X 100%
91.58+ 82
= 9.58 x 100%
173.58
CQDs = 5.52%
For Physics, the mean of the scores is 60.5 and
the standard deviation is equal to 12.33.
Getting the value of CV
CVp = 12.33 X 100%
60.5
CVp = 20.38%
Using the previous formula in solving the Q3
and Q1 of grouped data, we obtained a value
of 68.82 and 51.12 respectively.
Therefore:
CQDp= 68.82-51.12 X 100%
68.82+51.12
= 17.7 X 100%
119.94
CQDp= 14.76%
Notice that the CV and CQD in Physics are
higher as compared to Statistics, so we can say
that scores of the students in Physics are more
scattered. Since the CV and CQD in Statistics
subject are smaller, we can say that the
performance of the students here are less
variable, more homogenous or more
consistent.
Solution for question B
In order to answer the question for B, we have to
convert the scores obtained in the two subjects into
standard score or Z-score as follows:
Z stat = X-M
SD
= 80-8.67
6.30
Z stat = 1.06
Z phy = X-M
SD
= 65 – 60.5
12.33
Z phy= 0.36
Since the Z score in Physics is higher than in
Statistics, we can conclude that the student
performanced better in Physics than in
Statistics. This is because in Statistics the score
of the student is 1.06 standard deviation units
below the mean while his Z-score in Physics is
0.36 standard deviation units above the mean.
MEASURES OF SKEWNESS
Skewness (SK) refers to the symmetry or
departure from symmetry of a distribution. A
normal distribution is bell-graphed, however,
if a distribution is symmetrical it is said to be
skewed.
Types of Skewness
1. Positively Skewed Distribution or Skewed to
the Right.
this occurs if the mean is greater than the
median. The bulk of the distribution is on the
left and the tail of the distribution is longer to
the right.
To illustrate:
2. Negatively Skewed Distribution or Skewed to the
Left
the distribution is negatively skewed if more
concentration of values above the mean is
observed and the distribution tapers more to the
left than to the right.
To illustrate:
Formula
Sk= 3 (Mean – Median)
Standard deviation
Interpretation:
If: Sk = 0, normal distribution
Sk > 0, positively skewed
Sk < 0, negatively skewed
Kurtosis (Ku) refers to the peakedness or flatness Of
a distribution.
A distribution is said to be:
Mesokurtic if the bell-shaped distribution is not
flat or so peaked in height.
Leptokurtic if it is more peaked than the normal
distribution.
PLatykurtic if it is more flat than the normal
distribution.
To illustrate:
Formulas
a. Ungrouped data
Kurtosis (Ku) = T x – M 4
Nsd4
Where: x= raw data
M= mean
n= sample size
SD= standard deviation
b. Grouped data
Ku= T F(Mpt-M)4
NSD4
Where: F = frequency
Mpt= midpoint
M = mean
N= total frequency
SD= standard deviation
Interpretation: if : Ku= 3, normal distribution-
mesokurtic
Ku>3, platykurtic
Ku<3, leptokurtic
Example
Using the data on example of the Z-score (score in
Statistics), find the Sk and Ku. Interpret your answers.
Scores F CF Mpt (Mpt-M)4 F(Mpt-M)4
95-99 4 30 97 11386.79 45547.16
90-94 6 26 92 807.06 4842.36
84-89 8 20 87 0.01 0.08
80-84 9 12 82 475.63 4280.67
75-79 2 3 77 8743.91 17487.82
70-74 1 1 72 46314.87 46314.87
N=30 118472.96
From the previous computation we got: M= 86.87 and
SD = 6.30. Computing the median using the formula
for grouped data we got, Md = 86.38
Solution:
Sk= 3 (M – Md)
SD
Sk= 3 (86.67-86.38)
6.30
Sk= 0.14 (positively skewed)
Ku= T {F(X-M)4}
NSD4
= 118472.96
30(6.30)4
= 118472.96
47,258.88
Ku= 2.51 (platykurtic)
