Chapter 3

Measure
of
location and dispersion

UNGROUPED/RAW DATA

BP Mdakane 1
 Introduction

This chapter is set to show how to analyze raw/ungrouped data.

The aim of these analyses is to get a value that shows the location and
dispersion on the scale of the dataset.
• The Arithmetic mean/Average (𝑥̅ )

Sum of all observations

Number of observations

• The Mode (Mo)

 The most repeated value(s) in the dataset.

• Example: Given the following dataset:

2 5 8 -3 5 2 6 5 -4 9 2
Mean:
∑ 𝑥 2 + 5 + 8 + −3 + 5 + 2 + 6 + 5 + −4 + 9 + 2 37
𝑥̅ = = = = 3.36
𝑛 11 11

Mode: 2 and 5

BP Mdakane 2
 • Percentiles
 Measure of position in the dataset.

 The percentile (𝑃 ), is the value that measures 𝑖% of the data.(0 ≤ 𝑖 ≤ 100).

• Commonly used percentiles: 𝑄 ,𝑀 ,𝑄

𝑄 =𝑃 first quartile/lower quartile

𝑀 =𝑃 median

𝑄 =𝑃 third quartile/upper quartile

• Computing percentiles:
1. Rearrange the dataset from lowest to highest values.

( )
2. Calculate the position using .

3. Find the observation in the rearranged data using the position.

Example: Suppose the dataset represent the mileage(km) of different cars:

6, 47, 49, 15, 42, 41, 7, 39, 43, 40, 36, 56

Find 𝑄 , 𝑀 , 𝑄

1. Ordered Data Set: 6, 7, 15, 36, 39, 40, 41, 42, 43, 47, 49, 56

BP Mdakane 3
 1) 𝑄 = 𝑃
( ) ( )
Position = = = 3.25

𝑄 = 15 + 0.25 36 − 15 = 20.25

2) 𝑀 = 𝑃
( )
Position = = 6.50

𝑃 = 40 + 0.5 41 − 40 = 40.5

3) Try it yourself!!! (Answer: 46)

• Finding the 20 percentile:

( ) ( )
Position = = = 2.60

𝑃 = 7 + 0.60 15 − 7 = 11.8

Measure of dispersion:
• Interquartile range (IQR)
𝐼𝑄𝑅 = 𝑄 − 𝑄
Example: Calculate the interquartile range of the data:
6, 47, 49, 15, 42, 41, 7, 39, 43, 40, 36, 56
𝑄 = 46
𝑄 = 20.25

𝐼𝑄𝑅 = 𝑄 − 𝑄 = 46 − 20.25 = 25.75

BP Mdakane 4
 • Variance (𝑠 ) and standard deviation (𝑠).

Alternative:
∑
∑ ∑ ̅
 

∑
∑ ∑ ̅
 

Example: Using the mileage data from previous:

6, 47, 49, 15, 42, 41, 7, 39, 43, 40, 36, 56

𝑥 = 6 + 47 + 49 + ⋯ + 36 + 56 = 421

𝑥 = 6 + 47 + 49 + ⋯ + 36 + 56 = 17767

421
17767 −
𝑠 = 12 = 272.45 Variance
12 − 1

421
17767 − 12 Standard
𝑠= = 16.51 Deviation
12 − 1

BP Mdakane 5
 Coefficient of variation (CV)
 It is used to compare the variation of datasets.

𝐶𝑉 = ̅ × 100 expressed in %

Example:
Data 1: 2 5 8 -3 5 2 6 5 -4 9 2
Data 2: 6 47 49 15 42 41 7 39 43 40 36 56

𝑠 4.11
𝐶𝑉 = × 100 = × 100 = 122.32%
𝑥̅ 3.36
𝑠 16.51
𝐶𝑉 = × 100 = × 100 = 47.06%
𝑥̅ 35.08

Interpretation: Dataset 2 is more efficient than Dataset 1.

GROUPED DATA

BP Mdakane 6
 Introduction
• This chapter is set to show how to analyze grouped data.
• The aim of these analyses is to get a value that shows the location and dispersion
on the scale of the data.

• The Arithmetic mean (𝒙) Sum of product

mid-points and
frequencies of each
∑ 𝑥𝑓 class/row
𝑥̅ =
𝑛
Sum of frequencies

Example: The following data represents the number of telephone calls received for two days at a
municipal call centre. The data was measured per hour.

Calculating the mean (𝒙):

𝑓 𝑥 = 3 3.5 + 4 6.5 + ⋯ + 2 21.5 = 597

∑ 𝑓𝑥 597
𝑥̅ = = = 12.44
𝑛 48

BP Mdakane 7
 The Mode (Mo):
How to estimate the mode using the histogram:

Mode =± 12.3

Percentiles: Measure of position.

𝑖×𝑛 𝑖×𝑛
(𝑈 − 𝐿 ) −𝐹 𝑐 −𝐹
100 100
𝑃 =𝐿 + =𝐿 +
𝑓 𝑓
𝑈 - upper class limit of percentile class.
𝐿 - lower class limit of percentile class.
𝑓 - frequency of percentile class
𝑛 - sample size
𝐹 - Cumulative frequency of previous percentile class.
𝑐 - class width.
Calculating the percentiles:
Step 1: Find the percentile class (row) using the cumulative frequency:
𝑖×𝑛
Position: 𝐹 ≥
100
Step 2: Write down all the information of the unknowns.

×
Step 3: Substitute in: 𝑃 = 𝐿 +

BP Mdakane 8
 Example: Suppose we are given the following frequency distribution table.

 Determine 𝑃
Step 1: Find the percentile class (row) using the cumulative frequency column:
𝑖 × 𝑛 30 × 48
𝐹 ≥ = ≥ 14.4
100 100
The row that contains ≥ 14.4 observations is after 7 in the cumulative frequency – class -> [8 - 11)
Step 2: Write down all the information of the unknowns: 𝑟 = 3; 𝑈 = 11; 𝐿 = 8; 𝑐 = 3; 𝑓 = 11; 𝐹 = 7
×
.
Step 3: Substitute in 𝑃 : 𝑃 =𝐿 + =8+ = 10.02

Calculate the Median

×
Step 1: 𝐹 = ≥ 24
Row that contains ≥ 24 observations is after 18 in the cumulative frequency – class -> [11 - 14)

Step 2: The class limit 11 – 14 is in the 4th row 𝑟 = 4

𝐿 = 11; 𝑐 = 3; 𝑓 = 13; 𝐹 = 18

Step 3: Substitute in: 𝑃 = 𝐿 +

50 × 48
3 − 18
100
𝑀 =𝑃 = 11 + = 12.38
13

BP Mdakane 9
 Estimating the percentiles using the Ogive:

𝑷𝟓𝟎= 𝑴𝒆 = ±𝟏𝟐. 𝟒𝟒
𝑷𝟑𝟎 = ±𝟏𝟎. 𝟎𝟐

The Variance (𝒔𝟐 ) and Standard deviation (𝒔):

∑ 𝑓𝑥
∑ 𝑓𝑥 − ∑ 𝑓 𝑥 − 𝑛𝑥̅
𝑠 = 𝑛 𝑠 = Variance
𝑛−1 𝑛−1

∑ ∑ 𝑓 𝑥 − 𝑛𝑥̅ Standard
𝑓𝑥 𝑠=
∑ 𝑓𝑥 − deviation
𝑠= 𝑛 𝑛−1
𝑛−1
Class limits Frequency (f) F Midpoint (x) 𝒇𝒙 𝒇𝒙𝟐
[2 – 5) 3 3 3.5 10.5 36.75
[5 – 8) 4 7 6.5 26 169
[8 – 11) 11 18 9.5 104.5 992.75
[11 – 14) 13 31 12.5 162.5 2031.25
[14 - 17) 9 40 15.5 139.5 2162.25
[17 – 20) 6 46 18.5 111 2053.5
[20 – 23) 2 48 21.5 43 924.5
Total n=48 597 8370

BP Mdakane 10
 Calculating the variance and standard deviation:
• Variance:
∑ 𝑓𝑥 597
∑ 𝑓𝑥 − 8370 −
𝑠 = 𝑛 = 48 = 20.10
𝑛−1 48 − 1

• Standard deviation:
597
8370 −
𝑠= 48 = 20.10 = 4.48
48 − 1
Calculating coefficient of variation:

𝑠 × 100 (4.48)(100)
𝐶𝑉 = = = 36.01%
𝑥̅ 12.44

Interquartile range (IQR)

𝐼𝑄𝑅 = 𝑄 − 𝑄

Example: Calculate the interquartile range of the data:

𝑄 = 15.67

𝑄 = 9.36

𝐼𝑄𝑅 = 𝑄 − 𝑄 = 15.67 − 9.36 = 6.31

BP Mdakane 11