Measure of Dispersion

and Location

John Irish G. Lira, PhD

Dispersion

ü The difference between the actual value

and the average value.

Measures of Dispersion Measures of Location

Range Quartiles
Average Deviation Deciles
Variance Percentiles
Standard Deviation Midhinge
Interquartile Range
Quartile Deviation
Range

ü The difference of the highest value and the

lowest value in the data set.

Example: The daily rates of a sample of eight

employees at GMS Inc. are ₧550, ₧420, ₧560, ₧500,
₧700, ₧670, ₧860, ₧480. Find the range.

Solution:

420, 480, 500, 550, 560, 670, 700, 860

HV = ₧860 and LV = ₧420

Range = HV – LV = 860 – 420 = 440

Average Deviation

ü It is the absolute difference between the element

and a given point.

3 2 1 0 1 2 3
Mean
Average Deviation for Ungrouped Data

Individual
value Sample
mean

AD =
å /X-µ/
N
Average Population
Deviation

Sample AD: AD =
å /X-X/
n
Average Deviation for Grouped Data

Frequency
Individual
value Sample
mean

AD =
å f /X-X/
N
Average Population
Deviation

Sample AD: AD =
å f /X-X/
n
Example 1 for Average Deviation

The daily rates of a sample of eight employees at

GMS Inc. are ₧550, ₧420, ₧560, ₧500, ₧700, ₧670,
₧860, ₧480. Find the average deviation.

Solution:
Compute for the mean

X=
å X 550 + 420 + 560 + 500 + 700 + 670 + 860 + 480
=
n 8
4,740
= = 592.50
8
Solution: AD Ungrouped Data

550 – 592.5

X X-X /X-X/
550 –42.5 42.5
420 –172.5 172.5
560 –32.5 32.5
500 –92.5 92.5
700 107.5 107.5
670 77.5 77.5
860 267.5 267.5
480 –112.5 112.5
å X = 4,740 å (X - X ) = 0 å / X - X / = 905

AD =
å / X - X / 905
= = 113.125 » 113.13
n 8
Example: AD Grouped Data

The data below shows the frequency distribution of

the amounts of electric consumption of a typical
household in Batangas City for the month of January
2009. Find the average deviation.

Amount of Number of
Electric Bill Families
700 – 849 2
850 – 999 9
1,000 – 1,149 15
1,150 – 1,299 9
1,300 – 1,499 5
Solution: AD Grouped Data

Midpoint

Class Limits f X fX
700 – 849 2 774.5 1,549.00
850 – 999 9 924.5 8,320.50

Add
1,000 – 1,149 15 1,074.5 16,117.50
1,150 – 1,299 9 1,224.5 11,020.50
1,300 – 1,499 5 1,374.5 6,872.50
40 å fX = 43,880

X=
å fX 43,880
= = 1,097
n 40
Solution (continuation)

774.5 – 1,097 f / X - X /Mean

Class Limits f X X-X /X-X/ f /X - X/

700 – 849 2 774.5 –322.5 322.5 645.0
850 – 999 9 924.5 –172.5 172.5 1,552.5
1,000 – 1,149 15 1,074.5 –22.5 22.5 337.5
1,150 – 1,299 9 1,224.5 127.5 127.5 1,147.5
1,300 – 1,499 5 1,374.5 277.5 277.5 1,387.5
40 å f / X - X / = 5,070

AD =
å f / X - X / 5,070
= = 126.75
n 40
Standard Deviation

ü It is a statistical term that provides a good

indication of volatility.
Measure
of risk
ü It measures how widely values are
dispersed from the average.

ü It is calculated as the square root of

variance.
Range Rule of Thumb

ü It is used to approximate or to give a rough

estimate of the standard deviation

Standard range
Deviation s»
4

Example: The daily rates of a sample of eight

employees at GMS Inc. are ₧550, ₧420, ₧560, ₧500,
₧700, ₧670, ₧860, ₧480. Find the range.

Solution:
range HV - LV 860 - 420 440
s» = = = = 110
4 4 4 4
Variance

ü It is a mathematical expectation of the

average squared deviations from the mean.

Sample Variance:
For ungrouped data For grouped data

s 2
=
å (X - X ) 2
s2 =
å f ( X - X ) 2

n-1 n-1
or or
å X 2
-
( å X ) 2
( å fX) 2

s =
2 n å fX 2
-
n
n-1 s =
2

n-1
Variance

For ungrouped data For grouped data

s=
å (X - X ) 2
s=
å f ( X - X ) 2

n-1 n-1
or or
( å X) 2

å
( å fX) 2

åX - n -
2
2 fX
s= n
s=
n-1 n-1
Population Variance and SD

Individual
Sample Variance value

Sample Sample mean

Variance
s2 =
å ( X - X ) 2

n-1
Sample
Population
Sample Standard Deviation
Sample SD

s=
å ( X - X ) 2

n-1
Example: SD & Variance Ungrouped Data

The daily rates of a sample of eight employees at

GMS Inc. are ₧550, ₧420, ₧560, ₧500, ₧700, ₧670,
₧860, ₧480. Find the average deviation.

Solution:
Compute for the mean

X=
å X 550 + 420 + 560 + 500 + 700 + 670 + 860 + 480
=
n 8
4,740
= = 592.50
8
Solution 1: SD & Variance Ungrouped Data

X X-X ( X - X )2
550 –42.5 1,806.25
550 – 592.5
420 –172.5 29,756.25
560 –32.5 1,056.25
500 –92.5 8,556.25
700 107.5 11,556.25
670 77.5 6,006.25
860 267.5 71,556.25
480 –112.5 12,656.25
å X = 4,740 å (X - X ) = 0 å / X - X / = 142,950
s2 =
å ( X - X ) 2

=
142,950
= 20,421.43
n-1 8-1

s=
å ( X - X ) 2

=
142,950
= 20,421.43 = 142.90
n-1 8-1
Solution 2: SD & Variance Ungrouped Data

X X2
å X 2
-
( å X ) 2

550 302,500 s =
2 n
420 176,400 n-1
560 313,600 ( 4,740)2
2,951,400 -
500 250,000 = 8
700 490,000 8-1
670 448,900 2,951,400 - 2,808,450
860 739,600 =
480 230,400 7
= 20,421.43
å X = 4,740 å = 2,951,400
X 2

å X 2
-
( å X ) 2

s= n = 20,421.43 = 142.90
n-1
Example: Variance & SD Grouped Data

Using the example provided in Chapter 2 on SJS

Travel Agency. Determine the variance and standard
of the frequency distribution on the ages of 50 people
taking travel tours.

Class Limits Frequency

18 – 26 3
27 – 35 5
36 – 44 9
45 – 53 14
54 – 62 11
63 – 71 6
72 – 80 2
Solution 1:

Class Limits f X fX
18 – 26 3 22 66
27 – 35 5 31 155
36 – 44 9 40 360
45 – 53 14 49 686
54 – 62 11 58 638
63 – 71 6 67 402
72 – 80 2 76 152
Total 50 å fX =2,459

X=
å fX 2,459
= = 49.18
n 50
Solution 1 (continuation)

Class Limits f X X-X ( X - X )2 f ( X - X )2

18 – 26 3 22 –27.18 738.7534 2,216.2572
27 – 35 5 31 –18.18 330.5142 1,652.5620
36 – 44 9 40 –8.18 84.2724 758.4516
45 – 53 14 49 –0.18 0.0324 0.4536
54 – 62 11 58 8.82 77.7924 855.7164
63 – 71 6 67 17.82 317.5524 1,905.3144
72 – 80 2 76 26.82 719.3124 1,438.6248
Total 50 å f (X - X ) 2
= 8,827.3800

Variance: s =
2 å f ( X - X ) 2

=
8,827.38
= 180.15
n-1 50 - 1

Standard Deviation: s = å f (X - X ) 2

= 180.15 = 13.42
n-1
Solution 2:

Class Limits f X fX fX2

18 – 26 3 22 66 66
27 – 35 5 31 155 155
36 – 44 9 40 360 360
45 – 53 14 49 686 686
54 – 62 11 58 638 638
63 – 71 6 67 402 402
72 – 80 2 76 152 152
Total 50 å fX =2,459 å fX 2
=129,761

(å fX)2 (2,459)2
å fX -
2

n
129,761 -
50 129,761 - 120,933.92
s =
2
= = = 180.15
n-1 50 - 1 49

(
å fX2 - nå fX) 2

s= = 180.15 = 13.42
n-1
Population Variance and SD
Individual
Population Variance value

Population Population mean

Variance
s =
2 å ( X - µ ) 2

N
Population

Population Standard Deviation

Population
SD
s=
å ( X - µ ) 2

N
Example for Variance & SD

The monthly incomes of the five research directors of

Recoletos schools are: ₧55,000, ₧59,500, ₧62,500,
₧57,000, and ₧61,000. Find the variance and standard
deviation.
Solution:
Compute for the mean

µ=
å X 55,000 + 59,500 + 62,500 + 57,000 + 61,000
=
N 5
295,000
= = 59,000
5
Solution for Variance & SD

X X-µ ( X - µ) 2
55,000 –4,000 16,000,000
59,500 500 250,000
62,500 3,500 12,250,000
57,000 –2,000 4,000,000
61,000 2,000 4,000,000
å X = 295,000 å (X - µ) = 0 å (X - µ) = 36,500,000
2

s =
2 å ( X - µ ) 2

=
36,500,000
= 730,000
N 5

s=
å ( X - µ ) 2

= 730,000 = 2,701.85
N
Quartiles, Deciles & Percentiles

Ungrouped Grouped
æ kN ö
k(N + 1) ç - cf ÷
Quartiles Qk = Q k = LB + ç 4 ÷(i)
4 ç f ÷
ç ÷
è ø
æ kN ö
k(N + 1) ç - cf ÷
Deciles Dk = Dk = LB + ç 10 ÷(i)
10 ç f ÷
ç ÷
è ø
æ kN ö
k(N + 1) ç - cf ÷
Percentiles Pk = Pk = LB + ç 100 ÷(i)
100 ç f ÷
ç ÷
è ø
Example for Quartiles
Find the 1st, 2nd, and 3rd quartiles of the ages of 9
middle-management employees of a certain
company. The ages are 53, 45, 59, 48, 54, 46, 51, 58,
and 55.

Solution:
1(N + 1) 1(9 + 1) 10
Q1 = = = = 2.5
4 4 4

2(N + 1) = 2(9 + 1) = 2(10) = 5

Q2 =
4 4 4

3(N + 1) 3(9 + 1) 3(10)

Q3 = = = = 7.5
4 4 4
Example for Quartiles

45, 46, 48, 51, 53, 54, 55, 58, 59

2.5th 5th 7.5th

46 + 48 94 55 + 58 113
Q1 = = = 47 Q3 = = = 56.5
2 2 2 2

Therefore, Q1 = 47, Q2 = 53, and Q3 = 56.5.

Example for Qk, Dk, Pk

Using the example provided in on SJS Travel Agency.

Determine the Q1, Q2, Q3, D7, P22 of the frequency
distribution on the ages of 50 people taking travel
tours.

Class Limits Frequency

18 – 26 3
27 – 35 5
36 – 44 9
45 – 53 14
54 – 62 11
63 – 71 6
72 – 80 2
Solution for Q1
N 50
Q 1(Ranked Value) = = = 12.5
4 4
Class Limits f cf cf
LB = 36 – 0.5 18 – 26 3 3
= 35.5 27 – 35 5 8
36 – 44 9 17 Q1 Class
45 – 53 14 31
54 – 62 11 42
f
63 – 71 6 48
72 – 80 2 50
Total 50
æN ö æ 50 ö
ç - cf ÷ ç - 8 ÷
Q 1 = LB + ç 4 ÷(i) = 35.5 + ç 4 ÷(9) = 40
ç f ÷ ç 9 ÷
ç ÷ ç ÷
è ø è ø
Solution for Q2
2N 2(50)
Q 2 (Ranked Value) = = = 25
4 4
Class Limits f cf cf
18 – 26 3 3
LB = 45 – 0.5 27 – 35 5 8
= 44.5
36 – 44 9 17
45 – 53 14 31 Q2 Class
54 – 62 11 42
63 – 71 6 48 f
72 – 80 2 50
Total 50
æ 2N ö æ 2(50) ö
ç - cf ÷ ç - 17 ÷
Q 2 = LB + ç 4 ÷(i=) 44.5 + ç 4 ÷(9) = 49.64
ç f ÷ ç 14 ÷
ç ÷ ç ÷
è ø è ø
Solution for Q3
3N 3(50)
Q 3 (Ranked Value) = = = 37.5
4 4
Class Limits f cf
cf
18 – 26 3 3
27 – 35 5 8
LB = 54 – 0.5 36 – 44 9 17
= 53.5
45 – 53 14 31
54 – 62 11 42 Q3 Class
63 – 71 6 48
72 – 80 2 50 f

Total 50
æ 3N ö æ 3(50) ö
ç - cf ÷ ç - 31 ÷
Q 3 = LB + ç 4 ÷(i)= 53.5 + ç 4 ÷(9) = 58.82
ç f ÷ ç 11 ÷
ç ÷ ç ÷
è ø è ø
Solution for D7
7N 7(50)
D7 (Ranked Value) = = = 35
10 10
Class Limits f cf
18 – 26 3 3 cf
27 – 35 5 8
LB = 54 – 0.5 36 – 44 9 17
= 53.5
45 – 53 14 31
54 – 62 11 42 D7 Class
63 – 71 6 48
72 – 80 2 50 f
Total 50

æ 7N ö æ 7(50) ö
ç - cf ÷ ç - 31 ÷
D7 = LB + ç 10 ÷(i) = 53.5 + ç 10 ÷(9) = 56.77
ç f ÷ ç 11 ÷
ç ÷ ç ÷
è ø è ø
Solution for P22
22N 22(50)
P22 (Ranked Value) = = = 11
100 100
Class Limits f cf cf

LB = 36 – 0.5 18 – 26 3 3
= 35.5 27 – 35 5 8
36 – 44 9 17 P22 Class
45 – 53 14 31
54 – 62 11 42
f
63 – 71 6 48
72 – 80 2 50
Total 50

æ 22N ö æ 22(50) ö
ç - cf ÷ ç - 8 ÷
P22 = LB + ç 100 ÷(i) = 35.5 + ç 100 ÷(9) = 38.5
ç f ÷ ç 9 ÷
ç ÷ ç ÷
è ø è ø
Midhinge

It is the mean of the first (Q1) and third (Q3)

quartiles in the data set. It is used to
overcome potential problems introduced by
extreme values (or outliers) in the data set.

Q1 + Q3
Midhinge =
2
Example for Midhinge

Find the midhinge of the ages of 9 middle-

management employees of a certain company. The
ages are 53, 45, 59, 48, 54, 46, 51, 58, and 55.

Solution:

Recall that Q1 = 47 and Q3 = 56.5.

Q 1 + Q 3 47 + 56.5 103.5
Midhinge = = = = 51.75
2 2 2

The midhinge age is 51.75.

Interquartile Range (IQR)

A is a measure of statistical dispersion, being equal

to the difference between the third and first
quartiles.

It is also called the midspread or middle fifty.

Interquartile Range (IQR) = Q3 – Q1

Example for Interquartile Range

Find the interquartile range of the ages of 9 middle-

management employees of a certain company. The
ages are 53, 45, 59, 48, 54, 46, 51, 58, and 55.

Solution:

Recall that Q1 = 47 and Q3 = 56.5.

Interquartile Range (IQR) = Q3 – Q1 = 56.5 – 47 = 9.5

The IQR is 9.5.

Quartile Deviation (QD)

It is slightly better measure of absolute dispersion

than the range.

It ignores the observation on the tails.

The difference samples from a population &
calculate their quartile deviations, their values are
quite likely to be sufficiently different (called
sampling fluctuation).

It is calculated from the sample data does not help us

to draw any conclusion (inference) about the quartile
deviation in the population.
Q3 - Q1
Quartile Deviation (QD) =
2
Example for Quartile Deviation

Find the quartile deviation of the ages of 9 middle-

management employees of a certain company. The
ages are 53, 45, 59, 48, 54, 46, 51, 58, and 55.

Solution:

Recall that Q1 = 47 and Q3 = 56.5.

Q3 - Q1
Quartile Deviation (QD) =
2
56.5 - 47 9.5
= = = 4.75
2 2
The QD is 4.75.
Coefficient of Variation (CV)

It is used when one is interested to compare

standard deviations of two different units,
coefficient of variations can be applied.

s
Sample Mean: CV = (100% )
X

s
Population Mean: CV = (100% )
µ
Example 1 for Coefficient of Variations
The average age of the engineers at VSAS Pipeline
Corporation is 33 years, with a standard deviation of
3; the average monthly salary of the engineers is
₧45,000, with standard deviation of ₧3,150.
Determine the coefficient of variations of age and
salary.
Solution:
s 3
CV = (100% ) = (100%) = 10% Age
X 33
s
CV = (100% ) = 3,150 (100%) = 7% Salary
X 45,000

Since the coefficient of variation is larger for age, the

ages are more variable than the salary.
Example 2 for Coefficient of Variations

The mean of commissions of Educational Insurance

over 1-year period is ₧12,500, and the standard
deviation is ₧1,350. The mean of the number of sales
is 60, and the variance is 56.25. Compare the
variations of the two.

Solution:
s
CV = (100% ) = 1,350 (100%) = 10.8% Commission
X 12,500
s 56.25
CV = (100% ) = (100%) = 12.5% Sales
X 60

Since the coefficient of variation is larger for sales,

the sales are more variable than the commissions.
Chebychev’s Theorem

For ay set of observations, the proportion of

the values that lie within k standard deviations
of the mean is at least 1- 12 , where k is any
constant greater than 1. k
Example for Chebychev’s Theorem

The mean price of laptop computer is

₧25,500 and the standard deviation is₧2,500.
Find the price range for which at least 88.89%
of the laptop will sell.
Example 1 for Chebychev’s Theorem

Chebychev’s theorem states that 88.89%, of

the data values will fall within 3 standards of
the mean.
25,500 + 3(2,500) = 25,500 + 7,500 = 33,000

25,500 – 3(2,500) = 25,500 – 7,500 = 18,000

Therefore, at least 88.89% of all laptop sold

will have a price range from ₧18,000 and
₧33,000
Example for Chebychev’s Theorem

A survey conducted by Commission on

Higher Education (CHED) found that the
mean amount of training allowance for
department heads of colleges and
universities was ₧25,000. The standard
deviation was ₧1,500. Using Chebychev’s
theorem, find the minimum percentage of the
data values that will fall between ₧22,000 and
₧28,000.
Example 2 for Chebychev’s Theorem

Subtract the mean from the larger value.

₧28,000 – ₧25,000 = ₧3,000

Divide the difference by the sd to obtain k

3,000
k= = 2.0
1,500

Use Chebychev’s theorem.

1 1 1
1- 2 = 1- 2
= 1- = 1- 0.25 = 0.75
k 2.0 4
Therefore, at least 75% of the data value will fall
between ₧22,000 and ₧28,000.
Empirical Rule

68%
95%
99.7%
Kurtosis

Statistical measure used to describe the

distribution of observed data around the
mean. It measures the relative peakedness or
flatness of a distribution (as compared to the
normal distribution, which shows a kurtosis
of zero)

ìé n(n + 1) ù é æ X i - X ö ù üï
4
3(n - 1)2
ï n
kurt = íê ú êå çç ÷÷ ú ý -
ïîë (n - 1)(n - 2)(n - 3) û êë i=1 è s ø úû ïþ (n - 2)(n - 3)
Three Types of Kurtosis

Leptokurtic are distributions where values clustered

heavily or pile up in the center. (k > 0)

Mesokurtic are intermediate distribution w/c are

neither too peaked nor too flat. (k = 0).

Platykurtic are flat distributions with values more

evenly distributed about the center with broad
humps and shot tails. (k < 0)
Three Types of Kurtosis

Leptokurtic

Mesokurtic

Platykurtic
Example for Kurtosis

Example: The daily rates of a sample of eight

employees at GMS Inc. are ₧550, ₧420, ₧560, ₧500,
₧700, ₧670, ₧860, ₧480. Find the range.

Solution:
Compute for the mean

X=
å X 550 + 420 + 560 + 500 + 700 + 670 + 860 + 480
=
n 8
4,740
= = 592.50
8
Solution…

X X-X ( X - X )2
550 –42.5 1,806.25
420 –172.5 29,756.25
560 –32.5 1,056.25
500 –92.5 8,556.25
700 107.5 11,556.25
670 77.5 6,006.25
860 267.5 71,556.25
480 –112.5 12,656.25
å X = 4,740 å (X - X ) = 0 å / X - X / = 142,950

s=
å ( X - X ) 2

=
142,950
= 20,421.43 = 142.90
n-1 8-1
Solution…
4
æX-Xö
X X-X ( X - X )2 çç ÷÷
è s ø
550 –42.5 1,806.25 0.0078
420 –172.5 29,756.25 2.1232
560 –32.5 1,056.25 0.0027
500 –92.5 8,556.25 0.1755
700 107.5 11,556.25 0.3202
670 77.5 6,006.25 0.0865
860 267.5 71,556.25 12.2779
480 –112.5 12,656.25 0.3841
4,740 0 142,950 15.3779
Solution…

ìé n(n + 1) ù é æ X i - X ö ù üï
4
3(n - 1)2
ï n
kurt = íê ú êå çç ÷÷ ú ý -
ïîë (n - 1)(n - 2)(n - 3) û êë i=1 è s ø úû ïþ (n - 2)(n - 3)

ìé 8(8 + 1) ù ü 3(8 - 1)2

= íê ú [15.3779]ý - = 0.3724
îë (8 - 1)(8 - 2)(8 - 3) û þ (8 - 2)(8 - 3)

The distribution on the daily rates of employee is

somewhat leptokurtic.
Coefficient of Skewness

Measures the general shape of the distribution or the

lack of symmetry of a distribution.

Ranges from –3 to +3.

It relates the difference between the mean and the

median to the standard deviation.

The direction of the long tail of the distribution

points the direction of the skewness.
Types of Distribution

Symmetrical. Data values are evenly distributed.

The distribution is unimodal.
The mean, median, and mode are similar
& are at the center of the distribution.

4 5

4 0

3 5

3 0

2 5

2 0

1 5

1 0

1 2 3 4 5 6 7 8 9

Mean = Median = Mode

Types of Distribution

Positively Skewed (or Right-Skewed).

Most of the values in the data fall to the left of the
mean and group at the lower end of the
distribution.
The tail is to the right.
The mean is to the right of the median, and the
mode is to the left of the median
4 5

4 0

3 5

3 0

2 5

2 0

1 5

1 0

1 2 3 4 5 6 7 8 9

Mode Median Mean

Types of Distribution

Negatively Skewed (or Left-Skewed).

The mass of the data values fall to the right of the
mean and group at the upper end of the
distribution.
The tail to the left.
The mean is to the left of the median, and the
mode is to the right of the median.
4 5

4 0

3 5

3 0

2 5

2 0

1 5

1 0

1 2 3 4 5 6 7 8 9

Mean Median Mode

Computation for Skewness

Pearson’s Coefficient of Skewness

3( X - median)
sk =
s

Software Coefficient of Skewness

n é æX-Xö ù
3

sk = êå çç ÷÷ ú
(n - 1)(n - 2) ê è s ø ú
ë û
Example for Pearson’s Coefficient of Skewness

Example: A motorcycle dealership pays its

salesperson a salary plus a commission on sales. The
mean monthly commission is ₧8,800, the median
₧9,000, and the standard deviation ₧1,200. Determine
the coefficient of skewness. Comment on the shape
of distribution.
Solution:

3( X - median) 3(8,800 - 9,000) 3(-200)

sk = = = = -0.50
s 1,200 1,200

There is a slight negative skewness in the distribution

of commission on sales.
Example for Software Coefficient of Skewness

Example: The daily rates of a sample of eight

employees at GMS Inc. are ₧550, ₧420, ₧560, ₧500,
₧700, ₧670, ₧860, ₧480. Find the range.

Solution:
Compute for the mean

X=
å X 550 + 420 + 560 + 500 + 700 + 670 + 860 + 480
=
n 8
4,740
= = 592.50
8
Solution…

X X-X ( X - X )2
550 –42.5 1,806.25
420 –172.5 29,756.25
560 –32.5 1,056.25
500 –92.5 8,556.25
700 107.5 11,556.25
670 77.5 6,006.25
860 267.5 71,556.25
480 –112.5 12,656.25
å X = 4,740 å (X - X ) = 0 å / X - X / = 142,950

s=
å ( X - X ) 2

=
142,950
= 20,421.43 = 142.90
n-1 8-1
Solution….
3
æX-Xö
X X-X ( X - X )2 çç ÷÷
è s ø

550 –42.5 1,806.25 –0.03

420 –172.5 29,756.25 –1.76
560 –32.5 1,056.25 –0.01
500 –92.5 8,556.25 –0.27
700 107.5 11,556.25 0.43
670 77.5 6,006.25 0.16
860 267.5 71,556.25 6.56
480 –112.5 12,656.25 –0.49
4,740 0 142,950 4.59
Solution….

n é æX-Xö ù
3

sk = êå çç ÷÷ ú
(n - 1)(n - 2) ê è s ø ú
ë û

8
= ( 4.59) = 0.19047619( 4.59) = 0.87
(8 - 1)(8 - 2)

The daily rates of employees are somewhat

positively skewed.
Outliers

A data set should be checked for extremely high or

extremely low values. These values are called
outliers. Outliers can strongly affect the mean and
standard deviation of a variable. One method in
determining the outliers is when a data value in a
data set is less Q1 – 1.5(IQR) or greater than Q3 +
1.5(IQR)

Outliers can strongly affect the mean and standard

deviation of a variable.

One method in determining the outliers is when a data

value in a data set is less Q1 – 1.5(IQR) or greater than
Q3 + 1.5(IQR).
Example for Outliers

Check the data set of the ages of 9 middle-

management employees of a certain company for
outliers. The ages are 53, 45, 59, 48, 54, 46, 51, 58,
and 55.
Solution:
Remember that Q1 = 47, Q3 = 56.5, and IQR = 9.5

Q1 – 1.5(IQR) = 47 – 1.5(9.5) = 47 – 14.25 = 32.75

Q3 + 1.5(IQR) = 56.5 + 1.5(9.5) = 56.5 + 14.25 = 70.75

Since, there are no values outside this interval;

hence, there are no outliers in the data set.
Boxplot (Box-and-Whisker plot)

Introduced by John Tukey in 1970’s..

It gives the following information:

Ø If the median is near the center of the box, the

distribution is approximately symmetric.

Ø If the median falls to the right of the center of the

box, the distribution is negatively skewed.

Ø If the median falls to the left of the center of the

box, the distribution is positively skewed.

Ø If the lines are about the same length, the

distribution is approximately symmetric.
Boxplot (Box-and-Whisker plot)

Ø If the left line is larger than the right line, the

distribution is negatively skewed

Ø If the right line is larger than the left line, the

distribution is positively skewed.
Figure: Boxplot

Xlowest Xhighest

Q1 Q2 = Median Q3

0 10 20 30 40 50 60
Example for Boxplot

Construct a boxplot for the data set of the ages of 9

middle-management employees of a certain
company. The ages are 53, 45, 59, 48, 54, 46, 51, 58,
and 55. What can we say about the distribution of the
data set

Solution:
Recall that Q1 = 47, Median = 53, and Q3 = 56.5

Lowest Value = 45

Highest Value = 59
Solution for Boxplot

Figure: Employees Age

The data set of the distribution in negatively-skewed,

since the median falls to the right of the center of the
box
Effects of Changing Units

Let X1, X2, X3, …. Xn be n observations. If we added

a constant k to each observation or multiplied each
observation by a number (h, h ¹ 0), then

Added Constant
Constant Multiplier

Summary Yi = Xi + k Yi = h Xi
Measure
Range Unaffected Range (Y) = h Range (X)
Standard Unaffected s (Y) = /h/ s (X)
Deviation
Quartiles Qi (Y) = Qi (X) + k Qi (Y) = h Qi (X)
IQR Unaffected IQR (Y) = h IQR (X)
Example 1

A business professor gave a test to a set of students. The

test had 40 questions, each worth 2 points. The summary
for the applicants’ scores on the test are as follows:

Summary Statistics
Range 22
Standard deviation 7.33
First quartile 64
Third quartile 72
Interquartile range 14

After grading the test, the professor realized that, because

he had a typographical error in question no. 4, no students
was able to answer the question. So he decided to adjust
the students’ scores by adding 2 points to each one. What
will be the range, standard deviation, Q1, Q3 & IQR?
Solution

Summary Statistics
Range 22 (unaffected)
Standard deviation 7.33 (unaffected)
First quartile 64 + 2 = 66
Third quartile 72 + 2 = 74
Interquartile range 14 (unaffected)
Example 2

The summary statistics for the monthly contribution

of employees on cooperative are as follows:

Summary Statistics
Range 700
Standard deviation 222.28
First quartile 1,100
Third quartile 1,400
Interquartile range 300

This month, employees voted to increase the

contribution by 5 percent (5%), in order to support
their new project. What will be the mean and median
for the new increased of contributions.
Solution

Summary Statistics
Range 1.05(700) = 7.35
Standard deviation 1.05(222.28) = 233.39
First quartile 1.05(1,100) = 1.155
Third quartile 1.05(1,400) = 1,470
Interquartile range 1.05(300) = 315
Success doesn’t come to You, You go to it.”
– Ben Francia