Statistics (S1101)

By
Abbas MOURAD
 S1100
❑ Chapter 1 : Basic Statistical Concepts
➢ Some definitions

❑ Chapter 2 : Basic Statistical Concepts

➢ Graphical representation of data : Bar chart, Strip chart, Histogram, Boxplot.
➢Numercial representaion of data: Mean, Median, mode, variance, quantiles,
skweness and kurtosis

 Chapter 3: Two-dimensional Descriptive Statistics

➢ Data Representation, Coefficient of correlation and simple linear regression
❑ Chapter 4: Combinatorial analysis & Probability
➢ Permutations, Arrangements and Combinations
➢Basic Concepts of Probability, Conditional probability and Bayes Theorem
2
 Chapter 1:
Basic Statistical Concepts
 What is Statistics?
Data Why?
1. Collecting Data
Analysis
e.g., Survey
2. Presenting Data
e.g., Charts & Tables
Decision-
3. Characterizing Data
Making
e.g., Mean, Median

© 1984-1994 T/Maker Co.

4
 What is Statistics?
❑ Statistics: The science of collecting, describing, and
interpreting data.

Two areas of statistics:

Statistical Methods

Descriptive Statistics
Inferential Statistics

➢ Descriptive Statistics: collection, presentation, and

description of sample data.
➢ Inferential Statistics: making decisions and drawing
conclusions about populations.

5
 Descriptive Statistics

1. Involves
• Collecting Data $
• Presenting Data 50
• Characterizing Data
25
2. Purpose
• Describe Data 0
Q1 Q2 Q3 Q4

Mean= 30.5 S2 = 113

6
 Inferential Statistics
1. Involves
• Estimation
• Hypothesis
Testing
2. Purpose
Population?
• Make decisions about
population characteristics

7
 Population & Sample
Sample 4
Size : n Sample 3
Size : n

Sample 1
Size : n
Population
Sample 2
Size : N Size : n

❑ Understand that statistics can be used to gain information about a

population by examining a sample of the population ; generalizations
about a population from a sample are valid only if the sample is
representative of that population.

8
 Simple Random Sample
 Every subset of a specified size n from the
population has an equal chance of being selected

9
 Types of Data
Types of
Data

Quantitative Qualitative
Data Data

Quantitative data are measurements that are

recorded on a naturally occurring numerical scale.

Qualitative data are measurements that cannot be

measured on a natural numerical scale; they can only
be classified into one of a group of categories.

10
 Type of variables
❑ Example: The table below represents employees of a
communications company.
Number of Age Gender Number of children Level of satisfaction
employees
1 25 M 0 Satisfied
2 35 M 2 Very Satisfied
3 24 F 0 Satisfied
4 52 M 3 Unsatisfied
…. …. …. …. ….
99 48 F 2 Very Satisfied
100 26 F 1 Satisfied

▪ Population: Employees of a communications company

▪ Sample: 100 employees of a communications company
▪Individual: 1 employee
▪ Variables: Age, Gender, Number of children and Level of satisfaction

11
 Descriptive Statistics
❑ Two kinds of variables:
➢ Qualitative, or Categorical, Variable: A variable that categorizes
or describes an element of a population.

➢ Quantitative, or Numerical, Variable: A variable that quantifies

an element of a population.

Nominal (Gender with 2 levels: M & F)

Qualitative
Ordinal (Level of satisfaction with 5 levels)

Variable
Discrete (Number of children)
Quantitative
Continuous (Age)

12
 Descreptive Statistics
 Nominal Variable: A qualitative variable that categorizes (or
describes, or names) an element of a population.
 Ordinal Variable: A qualitative variable that incorporates an
ordered position, or ranking.
 Discrete Variable: A quantitative variable that can assume a
countable number of values. Intuitively, a discrete variable can
assume values corresponding to isolated points along a line
interval. That is, there is a gap between any two values.
 Continuous Variable: A quantitative variable that can assume
an uncountable number of values. Intuitively, a continuous
variable can assume any value along a line interval, including
every possible value between any two values
13
 Chapter 2:
Univariate Statistical Analysis
 Basic Notations
k

k k

Xi ni fi pi Ni Ni
[0-10[ 2 2/10 = 0.2 0.2×100 = 20% + 2 8 +2 =10
[10-20[ 4 0.4 40% 2+4 = 6 4+4 =8
[20-30[ 3 0.3 30% 6+3 = 9 1+3 =4
[30-40[ 1 0.1 10% 9+1= 10 1
Total n =10 1 100% ------- -------

15
 Descriptive Statistics
 Descriptive Statistics are Used by Researchers to
Report on Populations and Samples

 Types of descriptive statistics:

✓ Organize Data (Qualitative Variable)
▪ Tables
▪Graphs
✓Summarize Data (Quantitative Variable)
▪Central Tendency
▪Variation

16
 Representation of qualitative variable
✓Tables
Frequency Distributions
✓Qualitative Variable : (Percentage)
✓Graphs
▪ Measurement scale: Nominal Bar Chart or Pie Chart
✓ Measures of Central
Gender
Tendency
❑ Male Mode
❑ Female
✓Tables
▪ Measurement scale: Ordinal Frequency Distributions
What is your level of satisfaction with (Percentage) and cumulative
public transport in Lebanon? frequency
❑ Very Satisfied ✓Graphs
Bar Chart or Pie Chart
❑Satisfied
✓ Measures of Central
❑ Neither satisfied/unsatisfied Tendency
❑ Unsatisfied Mode and Median
❑ Very Unsatisfied

17
 Descriptive Statistics : Qualitative variable
Example : Variable Gender : Bar Chart and/or Pie Chart
Gender ni fi pi Angle = fi × 360°
Male 33 0.236 23.6% 84.96°
Female 107 0.764 76.4% 275.04°
Total n =140 100 100% 360°

Bar Chart Pie Chart

 Descriptive Statistics: Quantitative Discrete variable
✓Quantitative Discrete Variable : ✓Measures of Central Tendency:
Mode, median, mean, Variance, ….
✓Graphs : Strip Chart
Number of ni fi Ni
persons
1 5 0.1 5
2 9 0.18 14
3 15 0.3 29
4 10 0.2 39
5 6 0.12 45
6 3 0.06 48
7 0 0 48
8 2 0.04 50
Total 50 1 …

19
 Descriptive Statistics: Quantitative Discrete variable
✓Quantitative Discrete Variable : Cumulative diagram or cumulative distribution function

Number of ni fi Ni
persons
1 5 0.1 5
2 9 0.18 14
3 15 0.3 29
4 10 0.2 39
5 6 0.12 45
6 3 0.06 48
7 0 0 48
8 2 0.04 50
Total 50 1 …

20
Quantitative continuous Variable
Histogram
&
Ogive
 Descriptive Statistics: Quantitative Cont. variable
Quantitative Continuous Variable : Measures of Central Tendency:
Mode, median, mean, Variance, ….
Graphs : Histogram

Same width : Based on ni

Histogram
Different widths : Based on di

22
 Descriptive Statistics: Quantitative Cont. variable
Quantitative Continuous Variable : Histogram (Same width)

Frequency
Classes ni Width: ai Polygon
[2-6[ 60 6-2 = 4
[6-10[ 80 10-6= 4
[10-14[ 130 14-10= 4
[14-18[ 100 18-14= 4
[18-22[ 30 22-18= 4
Total 400 …..

23
 Descriptive Statistics: Quantitative Cont. variable
Quantitative Continuous Variable : Histogram (Different widths)

Classe ni Width : di = ni/ai

s ai
[2-3[ 60 3-2 = 1 60/1 = 60
[3-4[ 80 4-3= 1 80/ 1 = 80
[4-8[ 130 8-4= 4 130/4 = 32.5
[8-16[ 100 8 100/8 = 12.5
[16-22[ 30 6 30/6 = 5
Total 400 ….. …..

The Frequency polygon in case of different widths is based on

the greatest common divisor (gcd) of ai :
In our example: gcd(ai ) = 1

24
 Descriptive Statistics: Quantitative Cont. variable

The Frequency polygon in case of different widths is based on

the greatest common divisor (gcd) of ai :
In our example: gcd(ai ) = 1

25
 Descriptive Statistics: Quantitative Cont. variable

Frequency
Polygon

The Frequency polygon in case of different widths is based on

the greatest common divisor (gcd) of ai :
In our example: gcd(ai ) = 1

26
 Descriptive Statistics: Quantitative Cont. variable
Quantitative Continuous Variable : Cumulative Frequencies Curve (Ogive)

Classes ni Ni
[2-3[ 60 60
[3-4[ 80 140
[4-8[ 130 270
[8-16[ 100 370
[16-22[ 30 400
Total 400 …..

27
The Numerical Characteristics
Measures of Central Tendency : The Mean
 Measures of Central Tendency : The Mean
Non grouped data:

Example: 2, 4, 5, 1, 3 =>

 Grouped data: The distribution is given as a frequency table

Example:

Xi ni
2 1
3 5
4 2
5 2
Total 10

29
 Measures of Central Tendency : The Mean
 Grouped data by classes : the data are grouped into classes, ci is the class midpoint, the mean is

Classes ni Center of
class: ci
[2-6[ 60 6+2 /2= 4
[6-10[ 80 10+6/2= 8
[10-14[ 130 14+10/2= 12
[14-18[ 100 14+18/2= 16
[18-22[ 30 22+18/2= 20
Total 400 …..

30
The Numerical Characteristics
Measures of Central Tendency : The Mode
 Measures of Central Tendency : The Mode
 The mode is the most repeated value in a distribution

 Non grouped data

Example 1: 2, 2, 4, 5, 8, 12, 2, 1, 2 => Mode = 2
Example 2: 2, 1, 4, 5, 8, 12, 2, 1, 1 => Mode = 1 & 2
Example 3: 2, 4, 6, 8, 1, 10, 22, 45 => Mode doesn't exist

 Grouped data
Xi ni
2 1
 The most repeated value i.e. the value with highest frequency
3 5
 4 2 => Mode = 3
5 2
Total 10

32
 Measures of Central Tendency : The Mode
 The mode is the most repeated value in a distribution
 Grouped data : Graphically
Strip Chart
Xi ni
6
2 1 5
3 5 4
4 2 3

ni
5 2 2

Total 10 1
0
0 1 2 3 4 5 6

Xi

Mode = 3

33
 Measures of Central Tendency : The Mode
 Grouped data by classes : In the case of grouped data by classes, the mode is affected by the
width (same or different)
 Same Width:

 Different Widths:

Example (Same Width)

Classes ni Width (ai)
 Step 1: Modal Class: The class with highest frequency
[2-6[ 60 4 => Modal class = [10-14[
[6-10[ 80 4
[10-14[ 130 4 Step 2 :
[14-18[ 100 4
[18-22[ 30 4
Total 400
34
 Measures of Central Tendency : The Mode
Example (Same Width) : Graphically

Classes ni Width (ai)

[2-6[ 60 4
[6-10[ 80 4
[10-14[ 130 4
[14-18[ 100 4
[18-22[ 30 4
Total 400

Mode ≈12.5
35
 Measures of Central Tendency : The Mode
Example (Different Widths) Step 1: Modal Class: The class with highest density
=> Modal class = [3-4[
Classe ni Width : di = ni/ai
s ai
Step 2 :
[2-3[ 60 3-2 = 1 60/1 = 60
[3-4[ 80 4-3= 1 80/ 1 = 80
[4-8[ 130 8-4= 4 130/4 = 32.5
[8-16[ 100 8 100/8 = 12.5
[16-22[ 30 6 30/6 = 5
Total 400 ….. …..

 Graphically

Mode ≈3.3
The Numerical Characteristics
Measures of Central Tendency : The Median
 Measures of Central Tendency : The Median
The median is the middle value in the distribution when the numbers
have been arranged into increasing order
Case 1- Ungrouped data
Step 1 : Order observations from smallest to largest : x(1) ≤ x(2) ≤ … ≤ x(2)
 Step 2 : The median is

Example 1 : n is odd (n= 13) Example 2 : n is even (n=8)

1232363445443 5 6 8 10 9 9 4 7

Step1 (ordering): 1 2 2 3 3 3 3 4 4 4 4 5 6 Step1 (ordering): 4 5 6 7 8 9 9 10

Step 2: n is even =>
Step 2: n is odd => The integer part of [13/2 = 6.5] Me = (x(4) + x(5))/2= (7+ 8)/2 = 7.5
is 6 => Me = x(6+1) = x(7) = 3 (the 7th value of ordered (the median is the value between the 4th and 5th
data) observation of ordered data)

38
 Measures of Central Tendency : The Median
Case 2- Grouped data
Step 1 : Calculate the increasing cumulative frequency : Ni
 Step 2 : Calculate n/2 :
 If n/2 belongs the increasing cumulative frequency, we take the value
of X associated with n/2
 Elsewhere, we take the value of X associated the increasing
cumulative frequency immediately greater than n/2

Example 3 : Data in the right Table

Step1: Calculate Ni

Step 2: n/2 = 50/2 = 25 =>

Me = (x(25) + x(26) )/2 = (3+3) /2 = 3

39
 Measures of Central Tendency : The Median
 Median graphically – Example 3

50

40

30
 n/2 = 50 /2 = 25

20

10

Median = 3

40
 Measures of Central Tendency : The Median
Example 4 : Data in the right Table xi ni Ni
Step1: Calculate Ni 10 10 10
Step 2: n/2 = 24/2 = 12 => 11 2 12
Me = (x(12) + x(13) )/2 = (11+ 12) /2 = 11.5
12 8 20
13 4 24
Total 24

25

20

15
n/2 = 24/2 = 12
10

10 11 11.5 12 13

41
 Measures of Central Tendency : The Median
Case 2- Grouped data by classes – Continuous Variable
Step 1 : Calculate the increasing cumulative frequency (Ni)
 Find the Median Class: Calculate n/2 : If n/2 belongs the increasing cumulative frequency, we take
the value of X associated with n/2
 Elsewhere, we take the value of X associated the increasing cumulative frequency immediately
greater than n/2

 Step 2:

Classe ni Ni Ni
s
Example 5 : Data in the right Table
[10-20[ 10 10 80
Step1: Find the median class
[20-30[ 25 35 70
n/2 = 80/2 = 40 => Me ∊ [30-40[
[30-40[ 15 50 45
 Step 2 :
[40-50[ 15 65 30
[50-60[ 15 80 15
Total 80

14
 Measures of Central Tendency : The Median
 Median graphically – Grouped data by classes

Ni

Median
15
 Measures of Central Tendency : Properties
 Properties of the mean:
 The extreme values, called outliers, in the data set have an effect on the
mean. (When outliers are present, the mean is “pulled” in the direction of
the outlier)
 The mean obtained after grouping data into classes differs from the mean
obtained from ungrouped data

 The median:
 Unlike the mean, extreme values (outliers) have no effect on the median
 Like the mean and the mode, the median obtained after grouping data
into classes differs from the mean obtained from ungrouped data

 Properties of the mode:

 is not affected by extreme values
 is very sensitive to variations in class width (amplitude)
 may not be unique ; may not exist
 Like the mean, the mode obtained after grouping data into classes differs
from the mode obtained from ungrouped data
 The Numerical Characteristics
Measures of dispersion:
The range, the variance and the Inter-
quartile range
 Measures of dispersion

 Range

Variance and Standard deviation

Coefficient of Variation

Quantiles (Quartiles, Deciles and percentiles)

46
 Measures of dispersion : The Range
 Difference between the maximum and the minimum of the set of the
data : Range = Highest value – Lowest value = X(n) – X(1)

 The larger the range, the more dispersion in a set of data

 Affected by the extreme values

Example 2
Example 1 : Classes ni
1232363445443 [10-20[ 10
[20-30[ 25
Range = 6 – 1 = 5
[30-40[ 15
[40-50[ 15
[50-60[ 15
Total 80
Range = 60 – 10 = 50

47
 Measures of dispersion : The Variance
 Case 1- Non Grouped data

 Case 2- Grouped data

 Case 3- Grouped data into classes

48
 Measures of dispersion : The Standard deviation

 Properties of variance and standard deviation

 The higher the variance (standard deviation), the more spread out the data:
The values become more dispersed

 95% observation belong to

 If the variance is equal to zero, then the values are equal

 The variance (resp. standard deviation) is affected by the extreme values

 The variance obtained after grouping data into classes differs from the
variance of data before grouping The
dispersion
around the
mean

49
 Measures of dispersion : Coefficient of variation

 Is a measurement without unit and defined by

 Expressed as a percentage

 The higher the coefficient of variation, the greater the level of

dispersion

 It allows for comparison between distributions of values whose scales

of measurement are not comparable

Example :

10 individuals
Age expressed in years
 Salary expressed in dollars

Age is more
dispersed than
the salary

50
 Measures of dispersion : Quartiles

 Values that divide a ranked data into several equal parts

 Quartiles : divide a distribution into four equal parts : Q1, Q2, Q3

 Deciles : divide a distribution into 10 equal parts

 Percentiles : Percentiles are quantiles that divide a distribution

into 100 equal parts

51
 Measures of dispersion : Quartiles
 Quartiles: Divide the distribution into 4 equal parts. There are 3
quartiles: Q1 (25% accumulated) , Q2 = Me (50% accumulated), Q3 (75%
accumulated).

 Case 1- Non Grouped data

52
 Measures of dispersion : Quartiles
 Case 1- Non Grouped data

Example 1 : n is odd (n= 13) Example 2 : n is even (n=8)

1232363445443 5 6 8 10 9 9 4 7

Step1 (ordering): 1 2 2 3 3 3 3 4 4 4 4 5 6 Step1 (ordering): 4 5 6 7 8 9 9 10

Step 2:
Step 2:  Q1: n/4 = 8/4 = 2 => Q1 = (x(2) + x(3))/2=
 Q1: n/4 = 13/4 = 3.25 => The integer part = (5+6))/2=5.5
of 3.25 is 3 => Q1 = x(3+1) = x(4) = 3 (the 4th  n/2 = 4 => Q2 = Me = (x(4) + x(5))/2= (7+
value of ordered data) 8)/2 = 7.5
 Q2 : 2n/4 = 13/2 = 6.5 => The integer part  3n/2 = 6 => Q3= (x(6) + x(7))/2= (9+9)/2 = 9
of 6.5 is 6 => Q2 =Me = x(6+1) = x(7) = 3 (the
7th value of ordered data)
 Q3 : 3n/4 = 3(13)/4 = 9.75=> The integer
part of 9.75 is 9 => Q3 = x(9+1) = x(10) = 4

53
 Measures of dispersion : Quartiles
 Case 2- Grouped data

Step 1 : Calculate the increasing cumulative frequency : Ni

 Step 2 : Calculate nα (α = ¼; ½; ¾)
 If nα belongs the increasing cumulative frequency, we take the value
of X associated with nα
 Elsewhere, we take the value of X associated the increasing
cumulative frequency immediately greater than nα

Example 3 : Data in the right Table

Step1: Calculate Ni

Step 2:
 Q1: n/4 = 50/4 = 12.5 => Q1 = (x(12) + x(13) )/2 = (2+2)/2 = 2
 Q2: n/2 = 50/2 = 25 => Q2 = (x(25) + x(26) )/2 = (3+3)/2 = 3
 Q3: 3n/4 = 150/4 = 37.5 => Q3 = (x(37) + x(38) )/2 = (4+4)/2 =
4

54
 Measures of dispersion : The Quartiles
 Quartiles graphically – Example 3

50

40
 3n/4 = 37.5
30
 n/2 = 50 /2 = 25

 n/4 = 12.5 20

10

Q1 Q2 Q3

55
 Measures of dispersion : The Quartiles
Case 2- Grouped data by classes – Continuous Variable
Step 1 : Calculate the increasing cumulative frequency (Ni)
 Find the Quartiles Class: Calculate αn : If αn belongs the increasing cumulative frequency, we take
the value of X associated with αn
 Elsewhere, we take the value of X associated the increasing cumulative frequency immediately
greater than αn

 Step 2:

Classe ni Ni Ni
s
Example 5 : Data in the right Table
[10-20[ 10 10 80
Step1: Find the 1st Quartile class
[20-30[ 25 35 70
n/4 = 80/4 = 20 => Q1 ∊ [20-30[
[30-40[ 15 50 45
 Step 2 : Q1 = 20 + [(20-10)/25] × 10 = 24
[40-50[ 15 65 30
[50-60[ 15 80 15
Total 80

14
 Measures of dispersion : The Quartiles
 Quartiles graphically – Grouped data by classes

Ni
Ni

n/4 = 20

Q1

15
 Measures of dispersion : Interquartile Range
 Interquartile range : IQR = Q3 – Q1

 Interquartile interval : [Q3; Q1]

 The dispersion around the median

IQR (50%)

58
 Grouped data, Boxplot, Skewness,
Kurtosis, Tranformation of variables,
Means and Variances of two Groups
 Grouped data into classes
 Given a set of raw or ungrouped data, how would you group that data
into suitable classes that are easy to work with and at the same time
meaningful?

 The first step is to determine how many classes you want to have

 The second step is to determine the class interval (width)

60
 Grouped data: Example
 We observed the salary of 12 employees and we obtained :
315; 423; 300; 548; 374; 452; 463; 339; 324; 432; 315; 450
 Population: Employees
Sample: 12 Employees
Individual: 1 employee
 Variable: Salary
Type: Quantitative continuous
Salary ni
 Number of classes: k ≥ ln(12)/ln(2) [300-362[ 5

=> k≥ 3.6 => k =4 [362-424[ 2

 Width: a ≥ (548 – 300) /4 = 62 [424-486[ 4

[486-548] 1

Total 12

61
 Boxplot
 A Boxplot is a standardized way of displaying the distribution of data
based on a five number summary (“minimum”, first quartile (Q1), median,
third quartile (Q3), and “maximum”).
 It can tell you about your outliers and what their values are
 It can also tell you if your data is symmetrical, how tightly your data is
grouped, and if and how your data is skewed.

 Outliers or extreme values:

 Minimum: L = Q1-1.5 × IQR
 Maximum: U = Q3+1.5 × IQR
62
 Boxplot: Example
 Suppose that the dataset consists of these hypothetical test scores:
5 39 75 79 85 90 91 93 93 98
 Q1 = 75, Q2 = 88, Q3 = 92. IQR = 93 - 75 = 18.
 Outliers or extreme values:
 Minimum: L = Q1-1.5 × IQR = 75 – 1.5 × (18) = 48
 Maximum: U = Q3+1.5 × IQR = = 92+1.5 × (18) = 119

5 39 75 79 85 90 91 93 93 98
48 119
2 outliers
(left side)

63
 Skeweness
 The asymmetry (skewness) of a distribution of a variable X is
measured using:
 the Fisher asymmetry coefficient:

 Pearson coefficient

 Yule Coefficient

64
 Kurtosis
The kurtosis of a distribution indicates how much the distribution is
flat. In other words, it indicates whether a distribution has a peack or not.
The kurtosis is measured using Fisher coefficient defined by

65
 Means and Variances of two Groups
 We assume that the n observations are divided into two groups:
Group GA and Group GB

Group A Group B

Sample Size nA nB
Mean
Variance VB(X)
VA(X)

 The total or the agregated mean

is a weighted average of the two
means groups. This is written as  The agregated variance is
follows

66
 Means and Variances of two Groups: Example
 Student in a class are divided into two groups A and B. The mean of
the age of 12 students in the group A is 23 years old. While the mean of
the age of 15 students in the group B is 25 years old. Compute the mean
of the age of all students of this class.

Group A Group B

Sample Size 12 15
Mean 23 25

67
 Transformation of variables
 A change of a variable X is a transformation of this variable into a
new variable Y by multiplying X by a constant and/or adding a constant to
the variable X.

yi = axi + b

68