-1-

STATISTICS AND PROBABILITIES

Prepared

By

Dr. Hassan Salah M. Desouky

Ph. D. in Mathematical Statistics

Assoc. Professor of Applied Statistics

2025
-2-
 -3-

Preface

This book comprises two parts; the first is Statistics and

the second is Probabilities.

The first part devotes with the following subjects:

* Types of Data.

* Data Graph.

* Measures of Tendency.

* Measures of Dispersion.

* Correlation and Regression.

The second part devotes with the following subjects:

* Probability Theory.

* Probability Distribution.

* Estimation Theory.

* Tests of Statistical Hypotheses.

* Analysis of Variance.
-4-
 -5-

Contents

Chapter subject Page

one An Introduction 9
two Presenting Statistical Data 19
three Measures of Central Tendency 35
four Measures of Dispersion and 71
Skewness
five Correlation and Regression 101
six Probability Theory 117
seven Some Discrete Distributions 135
eight Probability Density Functions 157
nine Some Continuous Distributions 171
ten Normal Distribution 179
eleven Estimation Theory 197
twelve Chi Square Test 209
thirteen Analysis of Variance 225
-6-
 -7-

Firstly Statistics

This part devotes the following subjects:

* Types of Data.

* Data Graph.

* Measures of Tendency.

* Measures of Dispersion.

* Correlation and Regression.

-8-
 -9-

CHAPTER 1

An introduction
- 10 -
 - 11 -

CHAPTER 1

An introduction

Statistics is defined as the science of collecting,

organizing presenting, analysing, and interpreting
numerical date
That is statistics is a way to get information from data.
In other words, statistics is a set of techniques for
collecting, organizing, summarizing, presenting, and
interpreting information for drawing inferences in the
face of uncertainty. Statistics is divided into two basic
areas: descriptive statistics, and inferential statistics.

Descriptive Statistics:
Descriptive statistics is a set of procedures for
organizing and summarizing information in order to
make them readable and comprehensible.

Inferential Statistics:
It is used to make generalizations and extract
conclusions about the population. In other words,
statistical inference is the process of making an estimate,
prediction, or decision about population based on
sample data. Since populations are always large and the
investigation of each member will be expensive, it is
easier and cheaper to draw a sample from the
population under consideration and deduce conclusions
 - 12 -

or make estimates about the population. However, such

conclusions and estimates are not always going to be
correct. For this reason, we build into the statistical
inference a measure of reliability called the confidence
level or the significance level.

Statistical Analysis:
When we use statistics to make a decision, we carry
out a statistical analysis. Statistical analysis can divided
into four phases:
1. Study design,
2. Data collection,
3. Data analysis, and
4. Interpretation of the conclusions of the analysis in
order to take a convenient action.
The observations or measurements you collect from
an investigation or survey are referred to as data, the
data set, or the database. A single observation or
measurement is called a datum or data point. Data may
be either quantitative (numerical) or qualitative
(categorical):

Biostatistics:
Biostatistics is the science dealing with application of
statistics to the medical, health and biological fields.
 - 13 -

Population:
A population is the entire set of observations or
measurements under study. That is it is the set of all
items of interest in a statistical problem. In other words,
a population is a collection of all the elements about
which we want to draw conclusions.

Sample:
A sample is a set of observations selected from the
population, and is therefore only a part of entire
population. That it is a collection of the elements we are
studying.

Parameter:
A summary numerical result calculated from census
data is called parameter. Statisticians assign Greek
letters to parameters. For example,  (mu) is symbol
used for the population average,  (pi) is the symbol
used for the population proportion and 2 is the symbol
used for the population variance. Briefly, we can say that
the parameter is a descriptive measure of a population.

Statistic:
A summary numerical result calculated from sample
data is called a statistic. X (x–bar) is the symbol used
for the sample average. P is the symbol used for the
sample proportion, and s2 is the symbol used for the
sample variance. Inferential statistics provides the tools
for drawing inferences about a parameter from a
 - 14 -

statistic. On occasion, the inferences drawn from

samples are incorrect, but inferential statistics allows we
both to asses and to manage such risks.

Variable:
There are several; ways to classify the variables.
Variables can be divided into qualitative variables and
quantitative variables. On the other hand, the
quantitative variables can be divided into two type: the
discrete variables and the continuous variables.

Qualitative Variable:
The variable which can't be measured numerically is
called a qualitative variable. That is we can't assign
numerical values for the qualitative variable. In other
words, the qualitative variables are categorical
observations. The presence or absence of pain, name of
patient, sex and religion all are examples of qualitative
variables.

Quantitative Variable:
The variable can be numerically measured is called
quantitative variable. Ages, heights, blood pressure,
heart rates and temperatures all are examples of
quantitative variables.

Discrete Variable:
The variable which characterized by integer numbers
is called a discrete variable. That is there is a gap
between any two values of discrete variable. The
 - 15 -

number of tentacles on a hydra, the number of ray –

florets on a daisy, the number of animals caught in a
trap, the number daily admissions to NUB Hospital all are
examples of discrete variables.

Continuous Variable:
The variable which characterized by any value in
given interval is called a continuous variable. That is
there are no gaps between the values of continuous
variable. Variables such as height, weight and
temperature are examples of continuous variable.

Levels of Data:
Data can may be classified as nominal, ordinal, ratio,
and interval.

Nominal Data:
Nominal data name or denote differences in kind.
Nominal data are also referred to as categorical data. For
example, name, sex, religion, presence of pain and
occupation. Numerical data can nominal such as the
serial number on a patient sheet. Nominal numerical
data are names. Consequently, you should not perform
any calculations with the nominal numbers themselves.

Ordinal Data:
Ordinal data note differences in degree and can be
ordered or ranked Good / better / best and small /
medium / large are examples of ordinal data. Ordinal
 - 16 -

data contain more information than do nominal and

consequently you can rank the observations.

Ratio Data:
Permit the measurement of the exact difference
between any according to whether your data are
nominal, ordinal or ratio.
 - 17 -

Exercises 1
1. Consider the following record of a patient admitted
to an hospital today.
Name:
Age:
Gender:
Occupation:
City:
Marital:
status:
Number of children:
Weight:
Height:
Are you in pain?
Se verity of pain:
Systolic blood pressure:
Temperature:
Pulse:
Blood group:
Determine the type of data in each item.
2. In your own works, define and give an example of each
of the following statistical terms.
 - 18 -

(a) Population.
(b) Sample.
(c) Parameter.
(d) Statistics.
(e) Statistical inference.
3- Provide two examples each of qualitative data and
quantitative data.
4- Discuss the difference between qualitative and
quantitative data.
 - 19 -

CHAPTER 2

PRESENTING STATISTICAL DATA

- 20 -
 - 21 -

CHAPTER 2

PRESENTING STATISTICAL DATA

This chapter presents some procedures for

summarizing the various data types in tables and graphs.
Here, we will study how construct tables and graphs for
organizing and summarizing the data.
Frequency Distributions:
The frequency distribution is a simple, effective
method of organizing and presenting numerical data so
that one can get an overall picture of where
measurements are concentrated and how spread out
they are. To show the frequency table is constructed,
let's consider the ages in years of 50 patient admitted to
an hospital:

29 69 61 63 42 47 41 36 45 49
51 60 30 34 43 37 45 55 43 35
53 57 36 42 31 48 28 46 45 39
49 38 21 33 49 38 44 27 31 50
27 43 22 23 47 59 24 16 49 48
 - 22 -

To construct the frequency table that represents the

distribution of these ages, let's use the following steps:
1. Calculate the range of the data.
Range = largest value – lowest value
= 69 - 16 = 53
2. Divide the range into equal intervals: each of then has
lower limit and upper limit, Where the number of
these intervals can be obtained as follows:
Number of intervals = 1 + 3.3 log n
Where n is the number of observation. That is,
Number of intervals
= 1 + 3.3 log 50
= 1 + 3.3 × 1.69897
= 1 + 5.6066
= 6.6066  7

3. Determine the width of each interval as follows:

Width of each interval = Range ÷ number of intervals

= 53 ÷ 7 = 6.57  8
 - 23 -

4. Determine the lower limit and upper limit for each

interval as follows:
1st class 16 - 24
2nd class 24 - 32
3rd class 32 - 40
4th class 40 - 48
5th class 48 - 56
6th class 56 - 64
7th class 64 - 72

5. Record the data using tally marks.

Classes Tally Marks Frequency
16 - 24 //// 4
24 – 32 //// /// 8
32 - 40 //// //// 9
40 - 48 //// //// /// 13
48 - 56 //// //// 10
56 - 64 //// 5
64 - 72 / 1
Total 50
 - 24 -

Notes:
(a) For each: (say the first as an example)
16 is called the lower limit.
24 is called the upper limit.

(b) Class width = upper limit – lower limit

= 24 -16 = 8

(c) Midpoint = (upper limit + lower limit) ÷ 2

= (24 + 16) ÷ 2 = 20

Histogram:
The information in a frequency distribution is often
graphed. The common way of presenting a frequency
distribution graphically is the histogram, which is
constructed by marking off the class limit along the
horizontal axis and erecting over each class interval a
rectangle whose height equals the frequency of data
class. The histogram corresponding to the frequency
distribution in the last table is shown in the following
figure:
 - 25 -

Histogram of Patients' Ages

30

20

10

Std. Dev = 12.16

Mean = 41.3
0 N = 50.00
20.0 30.0 40.0 50.0 60.0 70.0

Histogram

Histogram of C1 N = 50

Midpoint Count

20.00 4 ****

28.00 8 ********

36.00 9 *********

44.00 13 *************

52.00 10 **********

60.00 5 *****

68.00 1 *
 - 26 -

Frequency Polygon:
Another common way of presenting a frequency
distribution graphically is the frequency polygon. A
frequency polygon is obtained by plotting the frequency
of each class above the midpoint of that class and then
joining the points with straight lines. The polygon is
usually closed by considering one additional class (with
zero frequency) at each end of the distribution and then
extending a straight line to the midpoint of each of these
classes. Frequency polygons are useful for obtaining a
general idea of the shape of the distribution.

Frequency Polygon
14
Number of Patients

12
10
8
Series1
6
4
2
0
0 50 100
Ages

Stem and Leaf Display:

A statistician named John Tukey introduced a method
organizing quantitative data called the stem and leaf
 - 27 -

display. This display, which may be viewed as an

alternative to the histogram, is most useful in
preliminary analysis. In particular, it provides a useful
first step in constructing a frequency distribution and
histogram. The histogram remains the best display to
use in formal presentations.
To organize the ages of the patients in the last table,
we define the stem as the tenths and the leaf as the units
as shown below:

Stem Leaf
1 6
2 97123847
3 806431786159
4 9322397781546535998
5 137950
6 9013

Note:
If the data contain decimal points, define the stem as
the digits to left of the decimal and the leaf as the digit
to right of the decimal.
 - 28 -

Pie Chart:
A pie chart is simply a circle subdivided into a number
of slices (or sectors) that represent the various
categories. It should be drawn so that size each slice is
proportional to the percentage corresponding to that
category.
Example:
Given the following table that shows blood group of 20
patients:

Group Number of patients

A 50
B 25
AB 40
O 85
Total 200

Display these data using pie chart

Solution:
(1) The percentage for each group
Number of patient in each group
= Total number of patieents
 - 29 -

50
Percentage for blood group A = = 25%
200

25
Percentage for blood group B = = 12.5%
200

40
Percentage for blood group AB = = 20%
200

85
Percentage for blood group O = = 42.5 %
200

(2) The angle corresponding to each blood group Sector

= 360o × percentage
The angle of blood group A sector
= 360 × 0.25 = 90o
The angle of blood group B sector
= 360 × 0.125 = 45o
The angle of blood group A B sector
= 360 × 0.20 = 72o
The angle of blood group O sector
= 360 × 0.425 = 153o
 - 30 -

Number of Patients

A; 50;
25%
O; 85; A
42% B
AB
B; 25;
O
13%

AB; 40;
20%
 - 31 -

EXERCIES 2

1. The Orange county agriculture extension agent

reviewed the rainfall records for the past two and half
years and found the following monthly rainfall, in
inches:

0.12 1.28 1.38 0.50 0.78

1.53 1.28 0.84 2.49 1.85
0.47 1.24 0.22 0.59 2.43
1.57 2.49 1.38 2.55 1.39
2.53 1.39 1.57 1.84 1.49
1.53 1.20 0.25 0.53 2.52

(a) Construct a frequency table for these data using 6

equal intervals of width 0.50 each ( e.g.: 0.0 – 0.50,
0.50 – 1.00,...,etc ).
(b) Construct a histogram for this distribution.
(c) Plot the polygon.
 - 32 -

2. The grades on a statistics exam are as follows:

86 75 66 77 66 64 73 91 65 59
78 61 86 61 58 70 77 80 58 94
55 62 79 83 54 52 45 82 48 67

(a) Construct a stem and leaf display for these data.

(b) Construct a frequency distribution for these data
using six class intervals.
3. Construct a stem and leaf display the following set of
data that shows the measurers of the blood alcohol
concentrations of 15 drivers involved in fatal
accidents:

0.27 0.17 0.17 0.16 0.13

0.24 0.29 0.14 0.16 0.12

0.16 0.21 0.17 0.18 0.24

 - 33 -

4. The following is the number of infected persons

with hepatitis in the four seasons of the year 1999.

Season The Number

of infectives
Summer 240
Autumn 75
Winter 65
Spring 120
Total 500

(a) Display these data using a pie chart.

(b) Use Excel to graph the pie chart.
- 34 -
 - 35 -

CHAPTER 3

MEASURES OF CENTRAL TENDENCY

- 36 -
 - 37 -

CHAPTER 3
MEASURES OF CENTRAL TENDENCY

The measure of central tendency (or location) of a set

of data is the value at the middle of the values of this set.
There are many types of measures of central location
such as arithmetic mean (average). Weighted arithmetic
mean, geometric mean, harmonic mean, mode and
median. Here, we will consider only three ; the
arithmetic mean, the median and the mode.

Arithmetic Mean:
The most commonly used measure of location is the
arithmetic mean (or mean). It is obtained by adding all
values of the set of data and dividing by the number of
these values. In other words, the mean is the sum of
values divided by the number of data points. The mean
of a set of measurements is defined as follows:

Firstly: Ungrouped Data

If we a set of data x1 , x2 , ... , xn , the arithmetic of this set

, which is denoted by x can be obtained as

 Sum of measurements
x 
umber of measurements
 - 38 -

In symbols.
n


x
i 1
i
x1  x2  ...  xn
X 
n n

Where:

X called " x – bar" is the mean
xi is the ith data point.

x i is the sum of all the data points.

n is the sample size

Example 1:
The following is a set of measure height of 8 children
(in cm).

130 100 90 115 105 120 80 70

1. What is the average height of this set?

2. Prove that the sum of variations of the values about
its mean is equal to zero.
 - 39 -

Solution:


x x x

130 130- 101.25= 28.75

100 100- 101.25= - 1.25
90 90- 101.25= - 11.25
115 115- 101.25= 13.75
105 105- 101.25= 3.75
120 120- 101.25= 18.75
80 80- 101.25= -21.25
70 70- 101.25= -31.25
810 0

1. The arithmetic Mean

n


X
i 1
i
810
X    101.25 cm
n 8

2. From the second column, we note that


(X  X )  0
Example 2:
The following is a set of observation represent the
waiting time (in minutes) of 6 patients at a surgery:
 - 40 -

5, 10, 12, 17, 20, 4.

Compute the arithmetic mean, and prove the sum of all

deviations of the values about its mean is equal to zero.
Solution:

X
 x  68 11.33 minutes
n 6

X xx
5 5-11.33= - 6.33
10 10 – 11.33 = - 1.33
12 12 – 11.33 = 0.67
17 17 – 11.33 = 5.67
20 20 – 11.33 = 8.67
4 4 – 11.33 = - 7.33
68 0.02  0

Secondly: Grouped Data

Example 3 below explains how to compute the
arithmetic mean from the grouped data in two method,
the direct method and the short-cut method.
Example 3:
Find the arithmetic mean for the following frequency
distribution:
 - 41 -

Class 16- 24- 32- 40- 48- 56- 64-72

Frequency 4 8 9 13 10 5 1

Solution:
To calculate the arithmetic mean, we use two methods
as follows.
1. Direct Method

Classes f x f x
16- 4 20 80
24- 8 28 224
32- 9 36 324
40- 13 (44) 572
a
48- 10 52 520
56- 5 60 300
64 - 72 1 68 68
Total 50 2088
 - 42 -

The Arithmetic mean


X 
 f x  2088  41.76
f 50

2. Short-cut Method

Classes f x d d' f d'

16- 4 20 24 -3 -12
24- 8 28 -16 -2 -16
32- 9 36 -8 -1 -9
40- 13 (44) a 0 0 0
48- 10 52 8 1 10
56- 5 60 16 2 10
64 - 72 1 68 24 3 3
Total 50 -14

Notes:

lower lim it  upperlim it

* Midpoint x 
2

* Length l  upper lim it  lower lim it

 - 43 -

*d xa

d
*d  a
'

The arithmetic mean


X a
fd '

l
f
  14
X  44   8  41.76
50

Properties of Mean:

1. Uniqueness: The mean of a set of data is unique

value.
2. Simplicity : The mean is easy to understand and
compute.
3. The mean is affected by the extreme values
because it use all observations in the given set of
data.
4. The sum of variations of the values about its mean
is equal to zero.

Median:
The median is the middle value when the data
arranged in order. Consequently the median is greater
than or equal to half of the observations and less than or
 - 44 -

equal the other half. There are two circumstances in

which we use the median or middle observation instead
of the mean. The first when we have ordinal,
nonnumeric observations, such as excellent, good, ok,
poor and terrible. Obviously we cannot compute a mean
for no numerical data. The other circumstance is when
there are extreme values on one side of the distribution.
When the data have extreme lows or extreme heights,
the median is a more representative measure. When we
want to compute the median of any set of data, we face
two cases.
Firstly: Ungrouped Data
1. The case of odd number:
When the number of a set of data is odd, the median
is the middle value when all values are arranged in order.
Example 4:
Given the following set of data:
10, 12, 15, 9, 13, 17, 30
Find the median of this distribution.
Solution:
We find the median in two steps:
1. Put the observations in ascending order (lowest
first to highest last).
2. Obtain the position (or rank) of the median.
 - 45 -

n 1
Position of median 
2

Where n is the number of observations, that is

7 1
Position of median  4
2

Therefore, the fourth observation in the order is the

value of the median.
Median = 13
II. The case of even number:
When the number of a set of data is even, there is no
single middle value, but two middle values. In this case,
the median is the average of the two middle values.
Example 5:
Find the median for the set of data :
10, 12, 15, 9, 13, 17, 30, 2
Solution:
1. Put the observation in its ascending order.
2, 9, 10, 12, 13, 15, 17, 30
2. Obtain the positions of the two middle values.

n 8
The position of the first value =  4
2 2

The position of the second value =4+1=5

 - 46 -

Therefore, the median is the average of the fourth and

the fifth values that is

12  13 25
Median =   12.5
2 2

Secondly: Grouped Data

In this section, we deal with how to compute the median

from the frequency distribution. The median can be
computed algebraic or graphic.
Example 6:
In respect to the frequency distribution in example (3),
find the median.

Classes 16- 24- 32- 40- 48- 56- 64-72

Frequency 4 8 9 13 10 5 1

Solution:
* Using Ascending Cumulative Frequency
The median of this frequency distribution may be found
carrying out the following steps:
1.Construct the ascending cumulative frequency
distribution table of the origin distribution.
 - 47 -

Ascending cumulative frequency distribution table

Upper limits CF
Less than 24 4
Less than 32 12
Less than 40 21  f1

Less than 48 34  f2

Less than 56 44
Less than 64 49
Less than 72 50

2. Find the position of median

Position of median 
f 
50
 25
2 2

3. Determine the following:

* Median class = (40 – 48)
* Lower limit of median class (Mo) = 40
* Width of median class ( l ) = 8
* Cumulative frequency preceding the position of
median f1  21 =
 - 48 -

* Cumulative frequency post the position of median

f 2  34

4. Apply the following formula to find the median.

Pos.  f1
Median  M 0   l=
f 2  f1

25  21
 40   8  42.46
34  21

* Using Descending Cumulative Frequency

1.Construct the ascending cumulative frequency
distribution table of the origin distribution.
Descending cumulative frequency distribution table
Lower limits CF
16 or more 50
24 or more 46
32 or more 38
40 or more 29  f2

48 or more 16  f1

56 or more 6
64 or more 1
 - 49 -

2. Find the position of median

Position of median  f 
50
 25
2 2

3. Determine the following:

* Median class = (40 – 48)
* Upper limit of median class (Mo) = 48
* Width (length) of median class ( l ) = 8
* Cumulative frequency preceding the position of
median f1  21

* Cumulative frequency post the position of median

f 2  34

4. Apply the following formula to find the median.

Pos.  f1
Median  M u   l=
f2  f 1

25  16
 48   8  42.46
29  16

Median graphically
To find the median graphically, we carry out the
following:
 - 50 -

1. Construct the ascending (or descending) cumulative

frequency distribution.
2. Find the position of the median

position of median 
f
2

3. Graph the ascending (or descending) cumulative

frequency curve.
4. Draw a horizontal line from the position of median to
intersect the curve, then drop a perpendicular line on
horizontal axis to obtain the value of median.
Hint: Try to graph the curve.

Example 7:
Given the following frequency distribution:

Class 16- 24- 32- 40- 48- 56- 64-72

Frequency 4 8 9 13 10 5 1

Find the median graphically.

 - 51 -

Solution:
Ascending cumulative frequency distribution table

Upper limits CF
Less than 24 4
Less than 32 12
Less than 40 21
Less than 48 34
Less than 56 44
Less than 64 49
Less than 72 50

Position of median  f 
50
 25
2 2
 - 52 -

CF

50 *
*
40 *

30 *

20 *

10 * Classes
*

24 32 40 48 56 64 72

Median 42,46 =

Properties of Median:
1. Uniqueness: The median of a set of data is unique
value.
2. Simplicity: The median is easy to obtain.
3. The median is not affected by extreme values.
4. The median does not use all information available, and
therefore it is usually less efficient than the mean
because it wastes information.
 - 53 -

Quartiles
Firstly: Ungrouped Data
Example 8:
Given the following set of data:

11 9 14 8 15 19 21 24 17 30 27
Find
(a) First quartile.
(B) Third quartile.
(C) Quartile deviation.

Solution:
(a) First Quartile (Q1)
Ascending order of data

8 9 (11) 14 15 17 19 21 (24) 27 30

n  1 12
Position of Q1   3
4 4

The first quartile is the third observation, that is

Q1 = 11
(b) Third Quartile
3 (n  1) 36
Position of Q3   9
4 4
 - 54 -

The third quartile is the ninth observation, that is

Q3 = 24
Quartile deviation = Q3 – Q1 = 24 – 9 = 15

Secondly: Grouped Data

Example 9:
The following is the frequency distribution of the profit
(in millions LE) of 100 companies in some year.

Profit 5- 10 - 15 - 20 - 25 - 30 - 35 - 40
Number of 4 11 21 37 17 7 3
companies

Compute:
(a) The first quartile.
(b) The third quartile.
(c) Quartile Deviation.

Solution:
(a) The first quartile
1. Ascending cumulative frequency distribution table
 - 55 -

Upper limits CF
Less than 10 4
Less than 15 15  f1

Less than 20 36  f2

Less than 25 73
Less than 30 90
Less than 35 97
Less than 40 100

2. Find the position of the first quartile

Position of Q1 
f 
100
 25
4 4

3. Determine the following:

* Class of Q1 = (15 – 20)
* Lower limit of Q1- class (Mo) = 15
* Width of Q1- class ( l ) = 5
* Cumulative frequency preceding the position of Q1
f1  15

* Cumulative frequency post the position of Q1

f 2  20

4. Apply the following formula to find Q1.

 - 56 -

Pos.  f1
Q1  M 0   l=
f 2  f1

25  15
 15   5  17.381
36  15

(b) The third quartile

1. Ascending cumulative frequency distribution table

Upper limits CF
Less than 10 4
Less than 15 15
Less than 20 36
Less than 25 73  f1

Less than 30 90  f2

Less than 35 97
Less than 40 100

2. Find the position of the third quartile

3f 300
Position of Q3    75
4 4

3. Determine the following:

 - 57 -

* Class of Q3 = (25 – 30)

* Lower limit of Q3- class (Mo) = 25
* Width of Q3- class ( l ) = 5
* Cumulative frequency preceding the position of Q3
f1  73

* Cumulative frequency post the position of Q3

f 2  90

4. Apply the following formula to find Q3.

Pos.  f1
Q3  M 0   l=
f 2  f1

75  73
 25   5  25.590
90  73

(c) Quartile deviation = Q3 –Q1

= 25.590 - 17.381= 8.21

* Using Descending Cumulative Frequency

1.Construct the Descending cumulative frequency
distribution table of the origin distribution.
 - 58 -

Descending cumulative frequency distribution table

Lower limits CF
5 or more 100
10 or more 96
15 or more 85
20 or more 64
25 or more 27  f2

30 or more 10  f1

35 or more 3

2. Find the position of Q1

Position of Q1  f 
100
 25
4 4

3. Determine the following:

* Median class = (20 – 25)
* Upper limit of median class (Mo) = 25
* Width (length)of median class ( l ) =5
* Cumulative frequency preceding the position of
median f1  27
 - 59 -

* Cumulative frequency post the position of median

f 2  64

4. Apply the following formula to find the median.

Pos.  f1
Median  M u   l=
f2  f 1

50  27
 25   5  21.892
64  27

Mode:
The mode of measurements is the value that occurs
most frequency. If all measurements are different, then
the set has no mode. A set of data may have more then
one mode. If a set of data has two mode, it is called
bimodal set. In addition, if the set of data has more than
two modes, it is called a multimodal set
Firstly: Ungrouped Data
The mode of an ungrouped set data is most frequently
value
Example 10:
A set of data:
15, 19 , 21, 19, 16, 14
has one mode (19). It is called unimodal set of data.
 - 60 -

Example 11:
A set of data:
13, 12, 18, 17, 12, 18, 12
has one mode (12). Note that 12 occurs more than the
value of 18.

Example 12:
A set of data:
20, 12, 21, 20, 25, 28, 25
has two modes (20 and 25). It is called bimodal set of
data because 20 and 25 occur with equal frequency.

Example 13:
A set of data:
20, 18, 15, 12, 42, 30
has no mode because all values are different.

Properties of Mode:
1. It is not unique.
2. It is easy to obtain
 - 61 -

Secondly: Grouped Data

There are some methods to compute the mode from the
frequency distributions. In this section, we will use the
Pearson difference method

Example 14:
Find The mode for the frequency distribution in
example (3).
Solution:
1. Determine the modal class (the class opposite the
largest frequency).
2. Determine the preceding and next classes, and put the
data in a table like, the following one:

Class Frequency

32 - 40 9

40 - 48 13 Modal class

48 - 56 10

3. Calculate the following differences of frequencies.

1 = 13 – 9 = 4
2 = 13 – 10 = 3
 - 62 -

4. Apply the following formula to find the mode.

4
Median  40   8
4  3

= 40 + 4.57 = 44.57
Example 15:
Mode in case of unequal classes frequency distributions
If the frequency distribution has unequal classes, it is
required to modify the frequency to proportion to
lengths of classes, where
frequencyof the class
Modified frequency
lenthof the class

Example 16:
Given the following frequency distribution:

Classes 0- 5- 12 - 20 - 30 - 45 - 50
Frequency 3 8 16 11 7 5

Find the mode of this distribution.

Solution:
We note that this frequency distribution is an unequal
classes, so the frequencies must be modified before
computing the mode as follows:
 - 63 -

classes frequency The length of Modified

the class frequency
0- 3 5 0.6
5- 8 7 1.143
12 - 16 8 2
20 - 11 10 1.1
30 - 7 15 0.467
45 - 50 5 5 1
Total 50 50

classes Modified
f

5 - 12 1.143
12 - 20 2 Modal class
20 - 30 1.1

1 = 2 – 1.143 = 0.857
2 = 2 – 1.1 = 0.9
0.857
Median  12   8  15.9
0.857  0.9
 - 64 -

Mode graphically
The mode can be graphically obtained.
This will be done through lecture.
Notes:
1. When the distribution has extremes on one side, the
median is the preferred measure of central location.
2. When the data are ranked, the median is preferred.
3. The mode is preferred in case of qualitative data, for
example, suppose the severity of pain o patients seen in
a pain clinic are classified as : no pain / mild moderate /
severe / very severe, the mode can describe the most
frequent degree of severity of pain.
4. The mode is the only measure of location which can
be used for variables measured at nominal scale such as
sex.

Relationships Among Mean, Median & Mode:

The relationship among the three measures of
ventral location can be observed from the smoothed
relative frequency polygons in figures from (1) to (3).
If the distribution is symmetrical and unimodal, the
three measures coincide, as in figure (1).
If a distribution is not symmetrical, it is said to be
skewed. The distribution in figure (2) is skewed to the
right, or positively skewed, since it has a long tail
 - 65 -

extending off to the right. In this case, mean is greater

than median and the median is greater than mode.
The distribution in figure (3) is skewed to the left, or
negatively skewed, since it has a long tail to the left but
a short tail to the right. In this case, mean is less than
median and median is less than mode.
1. Symmetric Distribution:
 - 66 -

2. Distribution skewed to the right (positively skewed):

3. Distribution skewed to the left (positively skewed):

 - 67 -

EXERCISES 3

1. The number of days for which each of 15 office

workers was absent during a one – month period is a
follows :
0 , 1 , 1, 3 , 0 , 0 , 2 , 5 , 0 , 1 , 1 , 2 , 0 , 1, 1
Compute the mean, the median, and the mode of the
number of days absent.
What are the characteristics of a set of data for which
the mean, median and mode are identical ?

2. Given a set of qualitative (categorical) data, what

measure of central location is always appropriate ?
State an example.

3. The ages of the employees of a fast – food outlet

are as follows:
19 , 19, 65, 20, 21, 18, 20
(a) Compute the mean, the median, and the mode
of the ages.
(b) How would these three measures of central
location be affected if the oldest employee
retired?
 - 68 -

4. Sporting competition that use judges scores to

determine a competitor's performance often drop
the lowest and highest scores before computing the
mean score, in order to diminish the effect of
extreme values on the mean. Suppose competitors A
and B receive the following scores :
A : 6.0, 7.0, 7.25, 7.25, 7.5, 7.5, 7.5
B : 7.0, 7.0, 7.0, 7.25, 7.5, 7.5, 8.5
(a) Compare the performance of competitors A and B
based on the mean of their scores both before and after
dropping the two extreme scores.
(b) Repeat part (a) using the median instead of the
mean.

5. Twenty families were asked how many cars they

owned. Their responses are summarized in the following
table.

Number of Cars Number of Families

0 3

1 10

2 4

3 2

4 1
 - 69 -

Find:
(a) The Arithmetic mean
(b) The Median.
(c) The Mode.

6. The following table represents frequency distribution

of weights of 100 patients.

Weight 30- 40- 50- 60- 70- 80- 90-

Number of 3 11 18 35 19 12 2
patients

Find :
(d) The Arithmetic mean
(e) The Median.
(f) The Mode.
- 70 -
 - 71 -

CHAPTER 4

MEASURES OF DISPERSION &

SKEWNESS
- 72 -
 - 73 -

CHAPTER 4
MEASURES OF DISPERSION & SKEWNESS

The measures of dispersion or variability are used to

measure the degree of variation in a set of data.
Statisticians use the term dispersion to describe the
degree to which a set of values differs from the values
mean. There are some types of measures of dispersion
such as the range, the mean deviation, the variance, the
standard deviation and the coefficient of variation.

Range:
The first and simplest measure of dispersion is the
range. The range of a set of measurements is the
numerical difference between the largest and smallest
measurements.
Firstly: Ungrouped Data

Range = Largest value – Smallest value

Example 1:
For the set of data:
20, 12 , 15, 17, 19
 - 74 -

Compute the range.

Solution:
Arrange the data in its ascending order:
12, 15, 17, 19, 20
Smallest value = 12
Largest value = 20
Range = largest value – smallest value
Range = 20 - 12 = 8
Secondly: Grouped Data
The range for the grouped data can be obtained as
the difference between the midpoint of the last class and
the midpoint of the first class.
Since the range is interested in the smallest and the
largest values and most of the information is ignored,
then it is seldom used analytically as a measure of
variations.

Mean Deviation:
The mean deviation or the average deviation takes
into account each observation and thus is superior
measure of the variability of the data than is either the
range. We can calculate the mean deviation using the
formula.
 - 75 -

Firstly: Ungrouped Data

The mean deviation can be obtained for ungrouped
data using the following formula:

n 

i 1
xi  x
Mean Deviation 
n

Where : xi  x is called the absolute value of the

deviations from the mean.

Example 2:
Compute the mean deviation of the following set of
observations:

10 12 18 20 25
Solution:
 
x xx xi  x

10 -7 7
12 -5 5
18 1 1
20 3 3
25 8 8
85 24
 - 76 -


x
 x  85  17
n 5
n 

i 1
xi  x
24
Mean Deviation    4.8
n 5
This means that on average, the data points differ from
their mean by 4.8
Note:
as another method to obtain the mean deviation, we
can use
n 
 xi  x
n


i 1
xi  median
i 1
instead of .
n n

Secondly: Grouped Data

The following example explains how to compute the
mean deviation from the grouped data.
Example3:
Given the following frequency distribution.

Class 10 - 20 - 30 - 40 - 50 - 60 - 70
frequency 7 11 19 15 9 4
 - 77 -

It is required to compute the mean deviation.

Solution:


class f x f x
xxf

10 - 7 15 105 161.56
20 - 11 25 275 143.88
30 - 19 35 665 58.52
40 - 15 45 675 103.80
50 - 9 55 495 152.28
60 - 70 4 65 260 107.68
Total 65 2475 727.72


x
 f x  2475  38.08
f 65
n 

i 1
xi  x f
727.72
Mean Deviation    11.20
f 65

Variance and Standard Deviation:

Variance is one of the two most widely accepted
measures of the variability of a set of data (the other is
standard deviation). Variance and standard deviation
 - 78 -

take into account all data in a set and are of fundamental

importance in statistical inference.

Firstly: Ungrouped Data

For ungrouped data, the variance of set of
measurements is the mean of the squared deviation of
each observation in the set from the mean. We can
compute the variance of a set of data by using the
formula:

S2 
 ( x  x) 2
n
and the standard deviation is the positive squared root
of the variance, that's,

S 
 ( x  x) 2
n

We note that this formula to compute the variance

depends on the arithmetic mean of the set of data. There
is another formula to compute the variance depends
directly on the values of observations. That is,
 - 79 -

x x
2 2

S 2
  
n  n 
 
Example 3:
The following is the set of height – measurement for
six students:
150, 165, 165, 180, 190, 170
Find the mean, variance and standard deviation.

Solution:

 
x
xx ( x  x) 2

150 - 20 400

165 -5 25

165 -5 25

180 10 100

190 20 400

170 0 0

1020 0 950
 - 80 -

The mean:


x
 x  1020  170
n 6

The variance:


S2 
 (x  x )2 
950
 158.33
n 6
The standard deviation:


S 
 (x  x ) 2


950
 158.33  12.58
n 6

Another Solution:

x x2

150 22500

165 27225

165 27225

180 32400

190 36100

170 28900

1020 174350
 - 81 -

x x
2 2 2
  174350  1020 
S 2
      158.33
n  n  6  6 
 

Note that these deviations from the mean will tend to be

large if measures are very spread and small they are not.

Example 4:
Given the following set of data:

15 18 20 14 17 21 22 24 30 19

It is required to compute:
(a) The arithmetic mean.
(b) The variance.
(c) The standard deviation.
 - 82 -

Solution:
 
x xx ( x  x) 2

15 -5 25
18 -2 4
20 0 0
14 -6 36
17 -3 9
21 1 1
22 2 4
24 4 16
30 10 100
19 -1 1
200 0 196

The mean:

x
 x  200  20
n 10

The variance:


S2 
 ( x  x) 2 
196
 19.6
n 10
 - 83 -

The standard deviation:



S 
 (x  x) 2

196
 4.43
n 10

Secondly: Grouped Data

If the data presented in the form of frequency
distribution, the variance and standard deviation can be
obtained in some method.
The First Method
In this method, the variance and standard deviation can
be computed by definition. That is,
The Variance


S 2

 (x  x ) 2
f
f

The Standard Deviation



S 
 (x  x )2 f
f
The Second Method
By using the midpoints of the classes.
 - 84 -

The Variance

x f   xf
2
2

S 2
  
f f




The Standard Deviation

x f   xf
2
2

S   
f f




The Third Method

A short-Cut Method

The Variance
 d '2 f  d ' f 
 
2

S  l 
2 2
  

 f  f
 

 

The Standard Deviation

 d '2 f  d ' f 
2  
2
 
S  l   

 f  f
 

 
 - 85 -

Example 5:
The following table presents the frequency distribution
of the weekly wages for 100 workers in some company:

Wage 100 - 200 - 300 - 400 - 500 - 600 - 700

Number of 9 16 29 21 15 10
workers

Compute the variance and the standard deviation in

three methods.
Solution:
The First Method

classes f x fx
(x  x )2 f

100 - 9 150 1350 549081

200 - 16 250 4000 345744
300 - 29 350 10150 64061
400 - 21 450 9450 58989
500 - 15 550 8250 351135
600 - 700 10 650 6500 640090
total 100 39700 2009100
 - 86 -

The Arithmetic Mean


x
 xf 
39700
 397 LE
f 100

The Variance


S2 
 (x  x )2 f 
2009100
 20091 LE
f 100

The Standard Deviation



S
 (x  x ) 2
f
 20091  141.74 LE
f

The second Method

classes f x fx fx 2

100 - 9 150 1350 202500

200 - 16 250 4000 1000000
300 - 29 350 10150 3552500
400 - 21 450 9450 4252500
500 - 15 550 8250 4537500
600 - 700 10 650 6500 4225000
total 100 39700 17770000
 - 87 -

The Variance

x f   xf
2
2

S 2
  
f f



2
17770000  39700
    20091 LE
100  100 

The Standard Deviation

x f   xf
2
2

S     141.74 LE
f f




Third Method

classes f x d d' fd ' fd '

2

100 - 9 150 - 200 -2 - 18 36

200 - 16 250 - 100 -1 - 16 16
300 - 29 (350) a 0 0 0 0
400 - 21 450 100 1 21 21
500 - 15 550 200 2 30 60
600 - 700 10 650 300 3 30 90
total 100 47 223
 - 88 -

The Variance
 d '2 f  d ' f 
2  
2
 
S l 
2
 

 f  f



 

 223  47  2 
 10000     20091 LE
 100  100  
 

The Standard Deviation

 d '2 f  d ' f 
 
2

S  l 
2
   
 20091  141.74 LE
 f  f
 

 

Notes:
*d  x  a
*d '  d / l

Coefficient of Variation:
The coefficient of variation of a set of measurements
is the standard deviation of the measurements divided
by their mean. This measure is a relative one and it
allows us to compare the relative variability of the two
data sets, because it adjusts for differences in the
magnitudes of the means of the data sets. For example
we may wish to know, for a certain population whether
serum cholesterol levels, measured in mg par 100 ml, are
more variable than body weight, measure in pounds. To
 - 89 -

obtain the coefficient of variation which denoted CV, we

can use the following formula:

S
CV  
(100)
x

Example 6:
Compute the coefficient of variation for the data of the
example 4.
Solution:

We have x  20 and S  4.43

S 4.43
CV  
(100)  (100)  22.15 %
x 20

Measures of Skewness:
Two characteristic of a distribution have been
studied namely, measures of central tendency and
dispersion. Another characteristic which be measured is
skewness. As we previously mentioned, the distribution
which is symmetrical has no skewness. That is the
skewness is zero. Karl Pearson developed the coefficient
of skewness (SK) to measure the amount and direction
of skewness.

SK = 3 ( mean  median)

 - 90 -

Where:
Sk : Coefficient of skewness.
 : Standard deviation.

and
-3  SK  3

Example 7:
The following is the number of patients admitted
University Hospital through a week:

10, 5, 2, 15, 9, 20, 23

Find:
(a) The arithmetic mean.
(b) The standard deviation.
(c) The median.
(d) Pearson's coefficient of skewness.

Solution:
To calculate the mean and the standard deviation,
construct a table like the following:
 - 91 -

Xi Xi - X X i X
2

10 -2 4
5 -7 49
2 -10 100
15 3 9
9 -3 9
20 8 64
23 11 121
84 0 356

a. The arithmetic mean:

 Xi 84 = 12 patients
X  
n 7
b. The standard deviation:

 X i  X 
2
356
  
2 7

 50.857 = 7.13

= 7 patents approximately
 - 92 -

(c) The median:

(i) Put the measurement in its ascending order:
2, 5, 9, 10, 15, 20, 23

(ii) Position of median = n 1

2
7 1 8
=  4
2 2
Then the fourth measurement is the value of median,
that is:
Median = 10 patients

(d) Pearson's coefficient of skewnees:

SK = 3 ( mean  median)

= 3 (12 10 )  0.857
7

From the shape of the figure of these data, we can

deduce that the distribution of number of patients
admitted University Hospital through a week, is skewed
to right.
The preceding formulas are used to calculate the
measures of dispersion and skewness in case of
ungrouped data. On the other hand, the following
formulas are used to find these formulas in case of
grouped data, that is:
 - 93 -

1. Range (R) = midpoint of last class – midpoint of first

value.
 f Xi  X
2. Mean deviation (MD) =
f
 f X i  X 
2

3. Variance 2 =
f
2
 f X i2   f X i 
=   
f  f 

 f X i  X 
2

4. Standard deviation () =

f

2
 f X i2   f X i 
=   
f  f 


1. Coefficient of variation (CV)  X (100)

3 ( mean  median)
2. Coefficient of skewness (SK) =


Example 8:

The following table represents the distribution of the

number of decayed missing or filled teeth of 100
children.

Number of DMF 0- 2- 4- 6- 8- 10– 12

Number of children 8 19 26 21 17 9
 - 94 -

Find:
(a) The range (R).
(b) The mean deviation (MD).
(c) The variance (2).
(d) The standard deviation ().
(e) The coefficient of variation (CV).
(f) The coefficient of skewness (SK).
Solution:
a. Range (R) = midpoint of last class – midpoint of first
class
= 11 – 1 = 10
b. Mean Deviation (MD):

Number Number of x fx X X f X X
of DMF children (f)

0- 8 1 8 1 – 5.94 = 4.94 39.52

2- 19 3 57 3 – 5.94 = 2.94 55.86

4- 26 5 130 5 – 5.94 = 0.94 24.44

6- 21 7 147 7 – 5.94 = 1.06 22.26

8- 17 9 153 9 – 5.94 = 3.06 52.02

10 - 12 9 11 99 11– 5.94 = 5.06 45.54

Total 100 594 239.64

 - 95 -

 f X 594
X    5.94
f 100

 f X X 239.64
MD    2.396
f 100
The variance:

Number Number of x fx f x2
of DMF children (f)
0- 8 1 8 8
2- 19 3 57 171
4- 26 5 130 650
6- 21 7 147 1029
8- 17 9 153 1377
10 - 12 9 11 99 1089
Total 100 594 4324

2
 f X i2   f X i 
 2
  
f  f 
2
432 4  594 
  
100  100 

= 43.24 – 35.28 = 7.96

c. The standard deviation the positive square root of

variance, that is:
 - 96 -

2
 f X i2   f X i 
    
f  f 
 7.96  2.82

d. The coefficient of variation:


CV = (100)
X
2.82
= (100) = 47.47%
5.94

e. The coefficient of skewness:

Firstly, we must obtain the median by constructing the

cumulative frequency distribution.
Origin frequency distribution Cumulative freq. Dist.
Class f Upper limits CF
0- 8 Less than 2 8
2- 19 Less than 4 27
4- 26 Less than 6 53
6- 21 Less than 8 74
8- 17 Less than 10 91
10 - 12 9 Less than 12 100
Total 100
 - 97 -

f 100
Position of Median    50
2 2

Pos  CF
Median  M   w
f

50  27
4 2
26

23
4 2
26

= 4 + 1.77 = 5.77

Now, we can calculate the (SK) as follows:

3 ( mean  median)
SK 

3 ( 5.94  5.77 )

2.82
3 ( 0.17 )
  0.18
2.82
 - 98 -

EXERCISES 4

1. The following is the set of data in which the blood

alcohol concentrations of 15 drivers in fatal accidents
are measured :
0.24, 027, 0.17, 0.17, 0.61,
0.13, 0.29, 0.24, 0.14, 0.16,
0.12, 0.16, 0.21, 0.17, 0.18
Find:
(a) The arithmetic mean.
(b) The standard deviation.
(c) The mode.
(d) The median.
(e) The coefficient of variation.
(f) Pearson's coefficient of skewness.

2. Is it possible for a standard deviation to be negative?

Explain.

3. Compute the range, variance, and standard deviation

for the following sample of data :
5, 7, 12, 14, 15, 15, 17, 20, 21, 24
 - 99 -

4. Calculate the mean and variance for the following

sample of data :
3, -2, -4, 1, 0, -1, 2

5. The number of hours a student spent studying over

the past seven days was recorded as follows :
2, 5, 6, 1, 4, 0, 3
Compute:
(a) The range.
(b) The mean.
(c) The variance.
(d) The standard deviation.

6. Sporting competitions that use judges scores to

determine a competitor's performance often drop
the lowest and highest scores before computing the
mean score, in order to diminish the effect of
extreme values on the mean. Suppose competitors A
and B receive the following scores :
A : 6.0, 7.0, 7.25, 7.25, 7.5, 7.5, 7.5
B : 7.0, 7.0, 7.0, 7.25, 7.5, 7.5, 8.0
Compare the variability of A and B using a convenient
measure of dispersion' measures.
 - 100 -

7. Given the following frequency distribution of ages in

years of 70 patients,

Age 0- 10- 20- 30- 40- 50- 60-70

Number of patients 3 7 29 19 8 3 1

Find:
(e) The range.
(f) The mean deviation.
(g) The variance.
(h) The standard deviation.
(i) The coefficient of variation.
(j) Pearson's coefficient of skewness.
 - 101 -

Chapter 5
Correlation & Regression
- 102 -
 - 103 -

Chapter 5
Correlation & Regression
Firstly: Correlation
Coefficient of Linear Correlation
The coefficient of linear correlation measures the type
and degree of the relation between two variables; one
of them is called the independent variable ( say: x ) and
the other is called the dependent variable ( say: y )
where the value of y depends on the value of x.
The relationship between two variables x and y can be
obtained by


n  x y   x   y 
 n  x   x   n  y   y  
2 2 2 2

Where 1     1
* If   1 , there is a perfect positive relation between x
and y .
* If    1 , there is a perfect negative relation between
x and y .

* If   0 , there is no relation between x and y .

 - 104 -

Example1:
The income x and consumption y of six persons (in
thousands pounds) are presented in the following table:

Income 4 6 7 9 11 13
Consumption 2 3 5 8 9 10
Compute coefficient of linear correlation between x and
y . comment on the result.

Solution:
The coefficient of linear correlation between x and y

x y x y x2 y2

4 2 8 16 4
6 3 18 36 9
7 5 35 49 25
9 8 72 81 64
11 9 99 121 81
13 10 130 169 100
50 37 362 472 283


 x y   x   y 
n
 n  x   x   n  y   y  
2 2 2 2
 - 105 -

6 (362)  (50) (37)


[ 6 (472)  (50) 2 ] [ 6 (283)  (37) 2 ]

2172  1850

[ 2832 2500] [1698  1369]

322
  0.97
[ 332] [329]

There is a strong positive relationship between x and y.

Secondly: Simple Linear Regression

The regression is defined a statistical method that used

to obtain the linear relation which describes the relation
between two variables or more. This relation is called
Regression Equation. The regression equation is a
relation between two variables; the independent
variable (say X) and dependent variable (say Y), where
the value of dependent variable Y depends on the value
of independent variable X. For example, the weight of a
person depends on his tall, and the value of sales
depends on the cost of advertisement.

The regression equation takes the form

Y a  b X

Where

Y is the dependent variable.

X is the independent variable.

 - 106 -

a is the cutting part from vertical axis by the straight

line which represents the regression equation.

b is called the coefficient of regression Y/X (or the slope

of regression line)

To estimate the regression equation for a sample of

data, we firstly must estimate the two parameters a
and b using the formulas.

 n  x y  ( x ) ( y )
b 
n  x 2  ( x ) 2


and 
a
 y  bx
n

Scatter diagram graphically describes the relation

between the variables x and y.

)1( ‫شكل االنتشار‬

14
12
10
8
‫ص‬

‫شكل االنتشار‬
6
4
2
0
0 5 10 15 20
‫س‬
 - 107 -

The scatter diagram (1) shows that there is a positive

relation between x and y, and b = positive value

)2( ‫شكل االنتشار‬

16
14
12
10
‫ص‬

8 ‫شكل االنتشار‬
6
4
2
0
0 5 10 15 20
‫س‬

The scatter diagram (2) shows that there is a negative

relation between x and y, and b = negative value

)3( ‫شكل االنتشار‬

30

25

20
‫ص‬

15 ‫شكل االنتشار‬

10

0
0 10 20 30
‫س‬
 - 108 -

The scatter diagram (3) shows that there is no relation

between x and y, and b = 0
Example 2:
The following table represents the cost of advertisement
and the corresponding volume of sales for some
commodity (in a million pounds):
Advertisement cost 1 2 3 4 5 6 4
Volume of sales 3 5 8 7 9 11 10

a. Graph the scatter diagram for these data.

b. Estimate the regression equation for these data.
c. Estimate the volume of sales when the advertisement
cost is equal to L.E. 250000
Solution:
a. The scatter diagram

‫العالقة بين تكلفة اإلعالن وحجم المبيعات‬

12
10
‫حجم المبيعات‬

8
6 ‫شكل االنتشار‬
4
2
0
0 2 4 6 8
‫تكلفة اإلعالن‬
 - 109 -

The scatter diagram shows that there is a positive relation between the
advertisement cost and the volume of sales, and the value of regression
coefficient b will be a positive value as we will see.

b. Estimation of the two parameters a and b.

x2
x y xy y2
1 3 3 1 9
2 5 10 4 25
3 8 24 9 64
4 7 28 16 49
5 9 45 25 81
6 11 66 36 121
4 10 40 16 100
25 53 216 107 449

n=7  x  25  y  53
 x y  216 x 2
107 y 2
 449

 n  x y  ( x ) ( y )
b 
n  x 2  ( x ) 2

 7 (216)  (25) (53)

b   1.508
7 (107)  (25) 2


a
 y  bx
n
 - 110 -

 53  1.508(25)
a  2.186
7

Therefore, the estimated regression equation is given as


y  2.186  1.508 x

What is the meaning of b  1.508 ?

That means that every million pound is spent on

advertisement leads to an increase 1.508 million pounds
in sales in average.
c. When x = L.E. 250000

y  2.186  1.508 (0.25)  2.563 million pounds

Example 3:

Compute the sum of the squares of the deviations of the original values
about the estimated values ( sum of squares of error).


e 2   ( y  y) 2

Solution:

y  2.186  1.508 x


y1  2.186  1.508(1)  3.694

y2  2.186  1.508(2)  5.202

y3  2.186  1.508(3)  6.710


y4  2.186  1.508(4)  8.218

y5  2.186  1.508(5)  9.726
 - 111 -


y6  2.186  1.508(6)  11.234

y7  2.186  1.508(4)  8.218

  
x y y ( y  y) ( y  y) 2
1 3 3.694 -- 0.694 0.481636
2 5 5.202 -- 0.202 0.040804
3 8 6.710 1.290 1.6641
4 7 8.218 1.218- 1.483524
5 9 9.726 -- 0.726 0.527076
6 11 11.234 -- 0.234 0.054756
4 10 8.218 1.782 3.175524
25 53 53.002 - 0.002 7.42742

Therefore, the sum of squares of error is 7.42742 and it is as little as

possible.

Standard Error of Estimate

The standard error of estimate can be obtained by one
of the two following formulas,

Sy/x 
 ( y  y) 2

n2

 
Or Sy/x 
y 2
a  y b  x y
n2

For the last example, using the first formula of standard

error of estimate, this yields,
 - 112 -


Sy/x 
 ( y  y) 2


7.42742
 1.2188
n2 72

using the second formula of standard error of estimate,

this yields,

 
Sy/x 
y 2
a  y b  x y
n2

449 2.186(53) 1.508(216)

Sy/x   1.2177
72
 - 113 -

Second Part
Probabilities
* Probability Theory.
* Some Discrete Distributions.
* Probability Density Functions.
* Some Continuous Distribution.
* Estimation Theory.
Chi Square Test.
* Analysis of Variance.
- 114 -
 - 115 -

CHAPTER 6

PROBABILITY THEORY
- 116 -
 - 117 -

CHAPTER 6

PROBABILITY THEORY

Probability plays a special role in all our lives, because

we use it to measure uncertainty. We are continually
faced with decisions that lead to uncertain outcomes,
and we depend on probability to help us choose our
course of action.
A probability is a numerical value that measures the
uncertainty that a particular event will occur. The
probability for an event ordinarily represents the
proportion of times under identical circumstances that
the event can expected to occur.
In other words the probability is the likelihood or
chance that a particular outcome will occur. Probabilities
are usually expressed as decimals or factions and range
in value from zero to one. For example, the probability
of tossing a fair coin and obtaining a tail is 0.5. The
probability of rolling a balanced die and getting a 2 is 1/6.
Probabilities can also be expressed as percentages, as in
" a 70% chance of rain " . Now, let's define some terms.

Experiment:
An experiment is the process that gives rise to an
Outcome. Rolling a die and tossing a coin are examples
of an experiment.
 - 118 -

Outcome:
An outcome is a single possible result experiment
Obtaining a 2 is an outcome when rolling a die.

Sample Space:
The sample space denoted by S is the set of all
possible outcomes of an experiment. The outcome of 1,
2, 3, 4, 5, or 6 dots showing are the sample space when
rolling a die.

Event:
An event is a collection of one or more specific
outcomes. Obtaining an even number of dots when
rolling a die is an example of an event. In this example,
the event comprises three outcomes.
The probability of an event A is equal to the sum of
the probabilities assigned to the simple events contained
in A. In other words, the probability of an event, A
denoted by P(A) can be obtained using the following law.

Number of outcomes in event A

P (A) =
Total number of outcomes in sample space
n( A)

n( S )
Example 1:
Tossing a balanced coin, write the sample space of
this experiment.
 - 119 -

Solution:
The sample space: S = {H, T}

Example 2:
Tossing a balanced coin twice, write the sample space
of this experiment and find the probability of getting:
(a) Exactly two heads.
(b) No heads.
(c) Exactly one tail.
Solution:
The sample space:
S = {(H, H), (H,T), (T,H), (T,T)}
Suppose that:
A = Event of getting two heads.
B = Event of getting no heads.
C = Event of getting one tail
(a) A = { (H,H) }
n( A) 1
P( A)  
n( S ) 4

(b) B = {(T,T)}

n( B ) 1
P( B)  
n( S ) 4
 - 120 -

(c) C = {(H,T) , (T,H)}

n(C ) 2 1
P(C )   
n( S ) 4 2

Example 3:
Tossing a balanced coin three times, write the sample
space of this experiment and find the probability of:
(i) Getting exactly one head.
(ii) Getting at least one head.
(iii) Getting at most one head.
Solution:
The sample space:

1st toss 2nd toss 3rd toss Sample space

H H H (H, H, H)
H H T (H, H, T)
H T H (H, T, H)
H T T (H, T, T)
T H H (T, H, H)
T H T (T, H, T)
T T H (T, T, H)
T T T (T, T, T)
 - 121 -

S = {(H,H,H), (H,H,T), (H,T,H), (H, T, T), (T,H,H),

(T,H,T), (T,T,H), (T,T,T)}
( i ) A = event of getting exactly one head.
A = { (H,T,T), (T,H,T), (T,T,H) }

n( A) 3
P( A)  
n( S ) 8

(ii) B = event of getting at least one head.

B = {(H,H,H), (H,H,T),(H,T,H),(H,T,T), (T,H,H),
(T,H,T), (T,T,H), (T,T,T) }

n( B ) 7
P( B)  
n( S ) 8

(iii) C = event of getting at most one head.

C = {(H, T, T), (T,T,H), (T,H,T) , (T,T,H)}

n(C ) 4 1
P(C )   
n( S ) 8 2

Addition rule for mutually exclusive events:

The two events A and B said to be mutually exclusive
events of they cannot occur at the same time. That is if
A and B are two mutually exclusive events then.
P (A and B) = 0
And therefore P ( A or B ) = P (A) + P (B)
On the other hand, if A and B are not mutually exclusive
events, then.
 - 122 -

P ( A and B )  0
Andtherefore
P(A or B) = P (A) + P (B) – P (A and B)

Conditional Probability:
Let A and B two events such that P (B)  0. The
conditional probability that A occurs, given that B has
occurred, is

P( A  B)
P( A / B) 
P( B)

Independent Events:
Two events are statistically independent if the
occurrence of either event has effect on the probability
of the occurrence of the other. That is two events A and
B are said to be independent if
P (A/B) = P (A) or P (B/A) = P (B)
Otherwise, the events A and B are said to be
dependent.

Multiplication Rule for Independent Events:

P (A and B ) = P (A) . P (B)
For any two independent events A and B.
 - 123 -

Complement Rule :
If A' is the event complementary to A, than
P (A') = 1 – P (A)
Example 4:
Let A and B be any two events where P (A) = 0.4 and
(B) = 0.7. If you know that P (A  B) = 0.24, find :
(a) P (A')
(b) P (A or B)
(c) Are A and B independent events or not?
(d) Are A and B mutually exclusive events or not?
Solution:
(a) P (A') = 1 – P(A)
= 1 – 0.4 = 0.6
(b) (A or B) = P (A) + P(B) – P (A  B)
= 0.40 + 0.70 – 0.24
= 1.10 – 0.24 = 0.86
(c) P (A) = 0.4 , P(B) = 0.7
P (A) – P (B) = 0.4 × 0.7 = 0.28
P (A  B) = 0.24
Since P (A  B)  P (A) – P (B) , A and B are not
independent.
(d) P (A  B) = 0.24
 - 124 -

That is probability of intersection between A and B is not

equal to zero, so the events A and B are not mutually
exclusive events.
Some Rules:
If A and B are events in a specific sample space, we can
deduce that.
(1) 0  P (A)  1
(2) P (A') = 1 – P (A)
(3) P (A or B) = P (A) + P (B) – P (A and B)
Where
P (A') = P (not A) is the symbol for the probability that
the event A does not occur.
Notice that:
* P (A or B) = P (A  B)
* P (A and B) = P (A  B)
Example 5:
The following is the Contingency table showing the
breakdown of AIDS, by AIDS status and test result for
10000 persons.

Test result
Positive Negative Total
Aids status
Infected 49 1 50
Not infected 199 9751 9950
Total 248 9752 10000
 - 125 -

Find:
(a) P (infected).
(b) P ( Not infected).
(c) P (infected and positive).
(d) P (infected and negative).
(e) P (infected or positive).
(f) P (infected or negative).
(g) P (infected / positive).
(h) P (Negative / Not infected).

Solution:

(a) P (infected) = Number of in fected people

Grand total number
= 50 5
  0.005
10000 1000

Number of not inf ected person

(b) P (Not infected) =
Grand total number

9950 995
  0.995
10000 1000
49
(c) P (Infected and positive) = 10000 = 0.0049

1
(d) P (Infected or positive) = = 0.0001
10000
 - 126 -

(e) P (Infected or Positive) = P (Infected) + P (Positive) -

- P (infected and positive)

= 50 248 49 50  248  49
  
10000 10000 10000 10000
= 249 = 0.0249
10000

(f) P (Positive or Negative) = P(Positive) +

P(Negative)

248 9752 10000

   1
10000 10000 10000

Notice that the events positive and negative are

mutually exclusive events, that is:
P ( Positive and Negative ) = 0

(g) P (Infected / Positive) = P ( inf ected and positive)

PPositive
49
49
 10000 
248 248
10000

(h) P (Positive and Negative) =

P ( Negative  Not inf ected )

P ( Not inf ected )
 - 127 -

9751 / 10000 9751

 
9950 / 10000 9950

Probability Distribution of Random Variable:

A random variable is a quantitative variable that
varies in a chance or random fashion. The random
variables may be divided into two types; the discrete
random variables and the continuous random variables.

Discrete random variable :

A discrete random variable is a variable that is limited
to specific points along a scale of values, the specific
value of which varies in a chance or random fashion. The
number of patients admitted to 6 Oct. University
Hospital is an example of discrete random variable.

Continuous random variable:

A continuous random variable is a variable that may
assume any value along a scale, the specific value of
which varies in a chance or random fashion. The amount
of rainfall in a year, and the value of the annual losses
covered by an insurance com[any are examples of
continuous random variables.
 - 128 -

Discrete Probability distribution:

The probability distribution of a discrete random
variable is a table, a graph, or a formula that yield the
probability of wach possible value of the discrete
variable.
If a random variable X can take values xi, than the
following must be true.
1 – 0  P (xi)  1 for all xi
2 -  P (xi) = 1

Example 6:
A dentist has 50 children less than 12 years old in his
clinic. He counted the number of decayed, missing or
filled teeth each child has and listed them in a table like
the following.
The probability distribution of number of DMF teeth per
child will be:
 - 129 -

Frequency of
xi Occurrence of P (xi)
xi
0 1 1/15
1 4 4/50
2 6 6/50
3 4 4/50
4 9 9/50
5 10 10/50
6 7 7/50
7 4 4/50
8 2 21/50
9 2 2/50
10 1 1/50
Total 50 1

Where xi = the number of DMF per child. If we randomly

select a child, what is the probability that:
(a) This child will have exactly 5 DMF teeth?
(b) This child will have 5 or 7 DMF teeth?
(c) This child will have 2 DMF at teeth?
(d) This child will have 8 DMF at teeth?
Solution:
10
(a) p ( x  5) 
50
(b) P (x = 5 or x = 7) = P (x = 5) + P ( x = 7)
(c) P (x  2) = 1 – P (x  2)
= 1 – P (x = 1 or x = 0)
= 1 – {P (x = 1) + P (x = 0)
 - 130 -

4 1 5 45 9
(d) 1     1  
 50 50  50 50 10

(e) P (x  8) = 1 – {P (x  8)
= 1 – P (x = 9 or x = 10)
= 1 – {P(x = 9) + P (x = 10)
2 1 3 47
1     1 
 50 50  50 50

Expected Value:
Given a discrete random variable X with values xi that
occur with probabilities P (xi) the expected value og X
denoted by E (X) is:
E (X) =  xi P (xi)
Laws of expected value:
If X and Y are random variables and C is any constant,
the following identities hold.
1. E (C) = C
2. E (CX) = CE (X)
3. E (X + Y) = E (X) + E (Y)
E (X - Y) = E (X) - E (Y)
4. E (X Y) = E (X) E (Y),
If X and be a independent random variables.
 - 131 -

Variance:
Let X be a discrete random variable with possible
values xi, that occur with probabilities P (xi), and let E (X)
= , the variance of X, denoted by 2 is defined to be :

 
 2  E  X   2  E X 2    2
  X i2 P  X i    2

The standard deviation of a random variable X, denoted

by , is the positive square root of the variance of X.

Laws of Variance:
If X and Y are random variables and C is constant, the
following identities hold:
1. V (C) = 0
2. V (CX) = C2 V (X)
3. V(X + C) = V (X)
4. V (X + Y) = V (X) + V(Y)
V (X – Y) = V (Y) + V (Y).
If X and Y are independent.
 - 132 -

EXERCISES 6
1. If you roll two balanced dice, what is the probability
of obtaining :
(a) A total of 3 dots ?
(b) A total of 6 dots ?
(c) A total of 10 dots ?

2. given P (A) = 0.8, P (A  B) = 0.48 and P (A  B) 0.92,

find p (B)

3. The result of flipping two fair coins is observed.

(a) Define the sample space.
(b) Find the probability of observing one heads and one
tails.
(c) Find the probability of observing at least one head.

4. An ordinary deck of playing cards has 13 cards of each

suit, suppose a card is selected at random from the
deck.
(a) What is the probability that card selected is an ice?
(b) What is the probability that card selected is a spade
king?
(c) What is the probability of selecting red queen?
 - 133 -

(d) What is the probability of selecting spade or black

card?

5. If P (A) = 0.6, P (B) = 0.5 and P (A  B) = 0.9, find P (A

 B). Are A and B independent or not?

6. Let A and B be two mutually exclusive events for

which P (A) = 0.25 and P (B) = 0.6, find P (A  B) and
P (A  B).

7. Let x represent the number of patients under 18 year

old in a clinic.

x 0 1 2 3 4 5
P (x) .50 .21 .19 .09 .02 .01

(a) Express the probability distribution of x in

graphical form.
(b) Comment on the symmetry or skewness of the
distribution.
(c) What is the probability of a clinic's having more
than two patients ?

8. Let x be a random variable with the following

probability distribution.
 - 134 -

x -10 -5 0 5 10
P (x) .01 .02 .02 .02 .03

Find the mean and standard deviation of x .

Find the mean and standard deviation of 2x.
Find the mean and standard deviation of 2x + 5 .
 - 135 -

CHPTER 7

Some Discrete Probability

Distributions
- 136 -
 - 137 -

CHPTER 7
Some Discrete Probability Distributions

Binomial Distribution
The binomial distribution is one of the most important
discrete distributions where many of statistical
phenomena behave as the binomial distribution. Any
random experiment consists of n independent trials and
each trial has two outcomes; one is a success event which
is fixed from trial to another, and the other is a failure
event, this experiment has a behavior of binomial
distribution.

Conditions of Binomial Distribution Application

1. The random experiment repeats n independent trials.
2. Each trial has two outcomes; a success outcome with
probability p, and the other is a failure outcome with
probability (1 – p).

Some Binomial Experiments

1. Tossing a balanced coin.
2. Rolling a balanced die.
3. Drawing a ball from a box with replacement.
4. Investigation of production units.
 - 138 -

Binomial Distribution Function

P( X  r )  C r p r (1  p) n  r
n

n!
 p r (1  p ) n  r
r! (n  r )!

where r  0 , 1, 2 , . . ., n and 0  p 1

Example 1:
A balanced coin was tossed three consecutive times, if the
random variable X is defined as the number of heads
appeared in the outcomes, find the probability
distribution of X using:
(a) The sample space of this experiment.
(b) The binomial distribution function.
(c) Graph the probability distribution of X .

Solution:
(a) The Probability Distribution of X Using the Sample
Space
 - 139 -

First toss Second toss Third toss the number of heads

H H H 3

H H T 2

H T H 2

H T T 1

T H H 2

T H T 1

T T H 1

T T T 0

The probability distribution of the random variable X

will be

X 0 1 2 3
P (X ) 18 38 38 18

We note that  P(X )  1

(b) The Probability Distribution of X Using the binomial

function
n!
P( X  r )  p r (1  p ) n  r , r  0 ,1, 2 , 3
r! (n  r )!
 - 140 -

0 3
3!  1   1  1
P( X  0)      
0! 3!  2   2  8
2
3!  1   1  3
P( X  1)      
1! 2!  2   2  8
2
3!  1   1  3
P( X  2)      
2! 1!  2   2  8
3 0
3!  1   1  1
P( X  3)      
3! 0!  2   2  8

(c) The graph of the probability distribution

P (X)

3/8

1/8

X
0 1 2 3
 - 141 -

The Expected Value of Binomial Distribution

If the random variable X is distributed as a binomial

with (n , p) and has the function
n!
P ( X  r)  p r (1  p ) n  r , r  0 , 1, 2 , ... , n
r! (n  r )!

(a) The expected value of X denoted E ( X ) will be

E (X ) n p
Proof:
n
E (X )   x P ( x)
X 0

n
n!
 
x0
x
x ! (n  x)!
p x (1  p) n  x

n
(n  1) !
n 
x 1 ( x 1) ! (n  x)!
p x (1  p ) n  x

n
(n  1) !
n p x 1 ( x 1) ! (n  x)!
p x  1 (1  p ) n  x

n
 n  1 x 1
n p    p (1  p) n  x
x 1  x  1
 - 142 -

Putting y  x 1  x  y  1 , this yields

n 1
 n  1 y
E (X )  n p    p (1  p) n  y  1
y 1  y 

n p  p  (1  p)n  1  n p

Lemma:
n
 n x n  x
a  b n
    b a
x0  x

(b) The variance of X denoted V ( X ) will be

V ( X )  n p (1  p )

Proof:

V ( X )  E ( X  )2  E ( X 2 )   2 (1)

E ( X 2 )  E X ( X  1)  X 

 E X ( X  1)  E ( X ) (2)
n
E X ( X  1)   x ( x  1) P ( x)
x0

n
n!
  x ( x  1)
x0 x ! (n  x) !
p x (1  p) n  x

n
(n  2) !
 n (n  1) 
x2 ( x  2) ! ( n  x ) !
p x (1  p ) n  x
 - 143 -

n
(n  2) !
 n (n  1) p 2  p x  2 (1  p ) n  x
x2 ( x  2) ! (n  x) !

Putting y  x  2  x  y  2 , this yields

 n  2 y
E X ( X  1)  n (n  1) p 2   p (1  p ) n  y  2
 y 

 n (n  1) p 2  p  (1  p)
n2

 n (n  1) p 2
From (2), we have

E ( X 2 )  n (n  1) p 2  n p
 n p [ (n  1) p  1 ]
From (1), we have

V (X )  E (X 2 )   2

 n (n  1) p 2  n p  (n p) 2

 n p [( n  1) p  1  n p]
 n p [ n p  p  1  n p]

 n p (1  p)
Since the standard deviation is the positive square root
of the variance, then

Std ( X )  n p (1  p)
 - 144 -

Note: The expected value of the random variable

referred also to the mean,  , of the random variable.

Example 2:
In some factory, if the percentage of defective units was
15 %. If a sample of five units was selected from this
factory production, compute:
(a) The probability that this sample contains exactly a
defective unit.
(b) The probability that this sample has no defective units.
(c) The probability that this sample contains at least a
defective unit.
(d) The probability that this sample contains at most a
defective unit.
(e) The expected value and the standard deviation of the
number of defective units.
Solution:
p  0.15 1  p  0.85 n  5 units

(a) The probability of exactly a defective unit.

5!
P ( x  1)  (0.15) (0.85) 4  0.392
1! 4 !

(b) The probability of no defective unit.

 - 145 -

5!
P ( x  0)  (0.15) 0 (0.85) 5  0.444
0 ! 5!

(c) The probability of at least a defective unit.

P ( X  1)  1  P ( X  1)

 1  P ( X  0)

 1  0.444  0.556

(d) The probability of at most a defective unit.

P ( X  1)  P ( X  0)  P ( X  1)

 0.444  0.392  0.836

(e) The expected value and standard deviation

  E ( X )  n p  5  0.15  0.75

Std ( X )  n p (1  p)  5  0.15  0.85  0.798

Example 3:
An urn contains five white balls, three red balls, if
consecutive three balls were withdrew with replacement,
compute:
(a) The probability of getting exactly one white ball.
(b) The probability of getting no white ball.
© The probability of getting at least two white balls.
(d) The mean, variance, standard deviation and
coefficient of variation of selected white balls.
 - 146 -

Solution:
It is known that the withdrawing of balls considered
independent trials, therefore the probability of getting a
white ball is fixed from trial to another. That is, the
binomial distribution function will be applied to compute
the required.
5 3
n3 p 1 p 
8 8
(a) The probability of getting exactly one white ball.
2
3!  5   3 
P ( X  1)       0.264
1! 2!  8   8 

(b) The probability of getting no white ball.

0 3
3!  5   3 
P ( X  0)       0.053
0! 3!  8   8 

© The probability of getting at least two white balls.

P ( X  2)  1  P ( X  2)

 1  P ( X  0)  P ( X  1)
 1   0.053  0.264  0.683
Another Solution:
P ( X  2)  P ( X  2)  P ( X  3)

Try to complete!
 - 147 -

(d) The mean, variance, standard deviation and

coefficient of variation.

* The mean
5
  E (X )  n p  3   1.875
8
* The variance
5 3
V ( X )  n p (1  p )  3    5.625
8 8
* The standard deviation

Std ( X )  5.625  2.372

* The coefficient of variation
Std ( X ) 2.372
CV  (100)  (100)  126.51%
E (X ) 1.875

Bernulli Distribution
The Bernulli distribution is a special case of binomial
distribution when the number of independent trials
equals one. That is, if the random variable X is
distributed as Bernulli, the probability distribution of X
will be
P ( X )  p x (1  p)1  x , X  0, 1

Expected Value of Bernulli Distribution

 - 148 -

1
  E (X )  
x 0
x P (x ) 

1
 
x 0
x p x (1  p)1  x  p

Variance of Bernulli Distribution

1
V (X )  
x 0
x 2 P ( x)   2

1
E (X 2 ) 
x 0
 x 2 p x (1  p )1  x  p

V ( X )  p  p 2  p (1  p)

Poisson Distribution
The Poisson distribution is one of important discrete
distributions where it expresses many of practical
phenomena in our life that are described with rareness and
independence, such as
(a) The number of mistakes in some book.
(b) The number of accidents on some day in a high way.
(c) The number of patients who attain to some clinic on
some day.
(d) The number of cars that arrive some service station at
an hour.
(e) The number of calls attain to academy switch.
 - 149 -

Poisson Distribution Function

If the random variable X expresses rare event with
probability p , ( p  0 ) and the number of independent
trials is very large ( n  ) , then X is distributed as
Poisson distribution and has a function given by
e   x
P (X )  , x  0 , 1, 2 , . . .
x!
,  n p  0
, e  2.71828

This probabilistic function satisfies the following two

conditions:

1. P ( X )  0


2. 
x0
P ( x)  1

Proof:

e   x 
x

x0 x!
 e 
x0 x!

2 3
 e   [1      ... ]
2! 3!

 e e  1
 - 150 -

The Expected Value of Poisson Distribution


E ( X )   x P ( x)
x0


e   x
 
x0
x
x!

x
 e 

x0
x
x!

 x 1
 e 
x 1 ( x  1) !

  e e  
That is the mean of Poisson distribution equals the
parameter of the distribution.

Variance of Poisson Distribution

V ( X )  E ( X 2 )  E ( X )
2


E (X ) 
2
x
x0
2
P ( x)


e   x
 
x0
x2

x!
 - 151 -


x
e 

x0
( x  x  x)
2

x!

x
e 
 x ( x 1)  x 
x0 x!

 x 
x 
  x ( x 1) 

e  x 
x  0 x! x0 x! 

 
x  2 
 x 1 
 e    2  ( x  2) !
  
( x 1) !
 x2 x 1


 e   2 e    e  
 e   e  ( 2   )

 2  
The variance of Poisson distribution will be
V (X )  2      

That is the variance and mean of Poisson distribution

equals the distribution parameter.

Example 4:
If the average number of accidents on some high way was
4 accidents within two weeks, compute the probability of
occurrence:
(a) An accident within a week.
 - 152 -

(b) No accident within a week.

(c) At most two accidents within a week.
(d) At least two accidents within a week.
Solution:
Since the accidents that occur on the high way considered
rare accidents if they were compared with the number of
cars that transit this way, so the number of these accidents
is a random variable distributed as Poisson distribution.
The average number of accidents in a week
4
   2 / week
2

e2 2
(a) P ( x  1)   0.271
1!

e  2 20
(b) P ( x  0)   0.135
0!

(c) P ( x  2)  P ( x  0)  P ( x 1)  P ( x  2)

P ( x  0)  0.135

P ( x 1)  0.271
e 2 22
P ( x  2)   0.271
2!

P ( x  2)  0.135  0.271 0.271  0.677

(d) P ( x  2)  1  P ( x  2)

 1  P ( x  0)  P ( x  1)
 - 153 -

 1  0.135  0.271  0.594

Example 5:
The patients arrive at some clinic at a rate of 12 patients
per day, compute the probability of arriving:
(a) three patients within an hour.
(b) five patients within 3 hours.
(c) four patients within half a day.
Solution:
(a) the average number of patients within an hour,
12
  0 .5
24

e  0.5 (0.5) 3
P ( x  3)   0.01264
3!

(b) the average number of patients within 3 hours,

12
  3  1.5
24

e  1.5 (1.5) 5
P ( x  5)   0.01420
5!

(c) the average number of patients within 12 hours,

12
  12  6
24

e  6 (6) 4
P ( x  4)   0.13392
4!
 - 154 -

Example 6:
During the production investigation in some factory, it is
shown that every 250 units contain 3 defective units. A
sample of 25 units were randomly selected, compute the
probability of getting:
(a) no defective units in the sample.
(b) a defective unit in the sample.
(c) two defective units in the sample.
(d) at least three defective units in the sample.
(e) more than one defective units.

Solution:
The number of defective units is a random variable
distributed as binomial distribution, where

The sample size n  25

3
Probability of success p  0.012
250
Probability of failure 1  p  1  0.012  0.998

But if we use the binomial function we will face some

difficulties in calculations, so we will use Poisson
distribution as an Approximation of Binomial
Distribution under three conditions, they are:
1.   n p  25  0.012  3  5

2. p  0.05
 - 155 -

3. n  25
(a) The probability of no defective units
e  3 (3) 0
P ( x  0)   0.0498
0!

(b) The probability of getting a defective unit

e  3 (3)
P ( x  1)   0.1494
1!

(c) The probability of getting 2 defective units

e  3 (3) 2
P ( x  2)   0.2240
2!
(d) The probability of getting at least 3 defective units
P ( x  3)  1  P ( x  3)

 1  P ( x  0)  P ( x 1)  P ( x  2)

 1  0.0498  0.1494  0.2240

 1  0.4232  0.5768
(e) The probability of getting more than a defective unit
P ( x  1)  1  P ( x  1)

 1  P ( x  0)  P ( x 1) 
 1  0.0498  0.1494 

 1  0.1992  0.8008
- 156 -
 - 157 -

Chapter 8

Probability Density Function

- 158 -
 - 159 -

Chapter 8

Probability Density Function

Continuous Random Variable

A continuous random variable is a variable that may

assume any value along a scale, the specific value of
which varies in a chance or random fashion. The amount
of rainfall in a year, and the value of the annual losses
covered by an insurance company are examples of
continuous random variables. In other words, a random
variable is continuous if it takes any value in some
interval; (e. g.: x  [0,1] ). If is a continuous random
variable, then a function y  f (x) is called a probability
density function for X if and only if it has the properties:
1. f ( x)  o for all x


2. 

f ( x) dx 1

and we can define that

b
3. P(a  X  b)  a f ( x) dx

Notes:

1. a probability density function is used to compute the

probability between any values in the interval on that the
continuous random variable is defined.
 - 160 -

2. the area under the curve of a probability density

function between the lower limit and upper limit of the
interval on that the continuous random variable is defined
is equal to one.

3. the probability that a continuous random variable takes

a certain value in its interval is equal to zero; because the
pdf is used to compute the area between two values in its
interval.

Example 1:

Given the following function:

1 if 0 x 1
f ( x)  { 0 otherwise

a. prove that f (x ) is a density function.

b. find the mean, variance and standard deviation.

Solution:

a. 1. f (0)  1  0

f (1)  1  0

1
2. 
0
f ( x) dx  [ x ]10 1  0 1
 - 161 -

Since this function satisfies the two conditions, it’s a pdf.

b. The Mean

1 1 1 1
  E ( X )   x. f ( x) dx  [ x 2 ]10  (1 0 ) 
0 2 2 2

The Variance

 2  E ( X 2 )  ( E ( X ))2
2
1 1
  x f ( x) dx   
2
0
2

1 1 1 1 1
 [ x 3 ]10   (1  0)  
3 4 3 4 12

The Standard Deviation

1
  0.29
12

Example 2:

3 x2 if 0 x 1
f ( x)  { 0 otherwise
 - 162 -

a. prove that f (x ) is a density function.

b. find the mean, variance and standard deviation.

c. compute P(0.2  X  0.5)

d. compute P ( X  0.7)

Solution:

a. 1. f (0)  3(0) 2  0

f (1)  3(1) 2  3  0

1 1
2. 0
f ( x) dx   3x 2 dx  [ x 3 ]10 1  0 1
0

Since this function satisfies the two conditions, it’s pdf.

b. The Mean

1 1 1 3 3 3
  E ( X )   x. f ( x) dx   x.3 x 2 dx  3 x 3 dx  [ x 4 ]10  (1  0 ) 
0 0 0 4 4 4

The Variance

 2  E ( X 2 )  ( E ( X ))2
 - 163 -

2
1 3 1 9 1 9
  x f ( x) dx      x 2 .3x 2 dx 
2
 3 x 4 dx  
0
4 0 16 0 16
3 9 3 9 3
 [ x 5 ]10    
5 16 5 16 80

The Standard Deviation

3
  0.19
80
0.5 0.5
c. P(0.2  X  0.5)  0.2 f ( x)dx  30.2 x 2 dx  [ x 3 ]00..52 

 (0.5) 3  (0.2) 3  0.125  0.008  0.117

 1 
d. P( X  0.7)  0.7 f ( x)dx  0.7 f ( x)dx  1 f ( x)dx 
1
 3 x 2 dx  0  [ x 3 ]10.7  (1) 3  (0.7) 3  1  0.343  0.657
0.7

Example 3:

Is the following function pdf, if not, convert it to pdf.,

x if 0  x  1
f ( x)  
0 otherwise

then compute:

a. the mean, variance and standard deviation.

b. P( X  0.4) .
 - 164 -

Solution:

First condition:

f ( 0)  0

f (1)  1  0

Second condition:

1 1 1 1 1
 f ( x)dx   xdx  2 [ x ]  (1  0)   1
2 1
0
0 0 2 2

The function does not satisfy the second condition, this

function is not pdf. To convert this function into pdf, the
following steps are considered:

1. multiplying the function by a constant (say: a , a  0 )

2. integrating on the interval of the function and equal the

result with one to find the value of constant a.

3. multiplying the function by the value of a obtained in

step 2, the function is converted to pdf.

This yields,

1
a  x dx  1
1
0
af ( x) dx  1  0
 - 165 -

a 2 1 a
[ x ]0  1  (1  0)  1
2 2

a
1  a2
2

The pdf of the continuous random variable will be

2 x if 0  x  1
f ( x)  
0 otherwise

a. The mean

1 1 2 3 1 2 2
E ( X )   xf ( x) dx  2  x 2 dx  [ x ]0  (1  0) 
0 0 3 3 3

The variance

1 1 1 4 1 1 1
E ( X 2 )   x 2 f ( x) dx  2  x 3 dx  [ x ]0  (1  0) 
0 0 2 2 2

 2  E ( X 2 )  ( E ( X ))2

1 2 1 4 1
2  ( )2   
2 3 2 9 18

The standard deviation

1
  0.24
18

0.4 0 0.4
b. P( X  0.4)   f ( x)dx   f ( x)dx  0 f ( x)dx
 - 166 -

0.4 0.4
0   f ( x)dx  2  xdx  [ x 2 ]00.4  (0.4) 2  (0) 2  0.16
0 0

Example 4:

Given the following probability density function:

a e  x if x  0
f ( x)  
0 otherwise

a. Find the value of a.

b. Compute the mean, variance and standard deviation.

c. P (0.2  X  0.7) .

d. P (10  X  1).

Solution:

Since f (x ) is a pdf, this yields

 

0
f ( x)dx  1  
0
a e  x dx  1


a  e  x dx  1  a[ e  x ]0  1
0

a (e   e 0 )  1  a (0  1)  1  a 1

Now, the pdf will be

 - 167 -

 e x if x  0
f ( x)  
0 otherwise

b. The mean

  E ( X )   x f ( x)dx


 
 
0
x e  x dx   xe x    e  x dx
0


 [ xe x ]0   e  x dx
0

 [ xe x ]0  [e x ]0

 0  (0  1)  1

The variance

E( X 2 )  
x 2 f ( x)dx

 
 
0
x 2e  x dx  x 2 
0
e  x dx   2 x(e  x )dx


 x 2 [e  x ]0   2 x(e  x )dx
0

 
 x 2 [e  x ]0  2  xe x   (e  x )dx
0


 x 2[e x ]0  2 [ xe x ]0  [e x ]0 
 - 168 -

Try to complete!

0.7 0.7
c. P(0.2  X  0.7)   0.2
f ( x)dx  0.2
e  x dx

[e x ]00..72  (e0.7  e0.2 )  0.4966 0.8187  0.3221

1
d. P( 10  X  1)   10
f ( x)dx

0 1
  10
f ( x)dx   f ( x)dx
0

1
 0   e  x dx  [e  x ]10  e 1  e 0
0

 0.3679  1  0.6321
Example 5:

Given the following probability density function:

a e 3 x if x  0
f ( x)  
0 otherwise

a. Find the value of a.

b. Compute the mean, variance and standard deviation.

c. P(0.3  X  0.5) .
 - 169 -

d. P(1  X  1).

Solution:

Since f (x ) is a pdf, this yields

 
0
f ( x)dx  1  0
a e 3 x dx  1

 a
a  e 3 x dx  1  [e 3 x ]0  1
0 3
a
a (e   e0 )  1  ( 0  1)  1  a 3
3

Now, the pdf will be

3 e 3 x if x  0
f ( x)  
0 otherwise

Cumulative Distribution Function (CDF)

If x is a continuous random variable whose a probability

density function f (x ) , the cumulative distribution

function of x which denote F (x ) is defined as

x
F ( x)   f ( x)dx


The cumulative distribution function of x gives the

cumulative probability from the lower limit of its interval
 - 170 -

till a certain point in this interval. That is, it gives the

cumulative probability for all values of x smaller than
this certain point.

Example 7:

Given a probability density function of a continuous

random variable x

1 if x  0
f ( x)  
0 otherwise

Find the cumulative distribution function of x and grave

it.

Solution:

x
F ( x)  0
dx  x
 - 171 -

CHAPTER 9

Some Continuous Distributions

- 172 -
 - 173 -

CHAPTER 9

Some Continuous Distribution

Uniform Distribution

The uniform density function of the random variable x is

given by

 1
 if 0  x  b
f ( x)   b  a
0
 otherwise

f (x)

1/ ( b – a)

x
a b
 - 174 -

This function satisfies the conditions of the density

function as follows:

1. f ( x)  0


2. 

f ( x)  1

b
3. P (a  x  b)  
a
f ( x) dx

Example 1:

Prove that the uniform function is a density function, and

compute P (1  x  2) .

Solution:

1 1
1. f (a )   0 and f (b)  0
ba ba

b 1 x b
2.  dx  ]a 
a ba ba
b a ba
   1
ba ba ba

2 1 x 2 2 1 1
p (1  x  2)   dx  ]1   
1 ba ba ba ba ba
 - 175 -

The Expected Value of Uniform Variable

The expected value of a uniform random variable x is

obtained as

b
E ( x )  x  f ( x ) dx
a

b 1 x
x
b

a ba
dx  a ba
dx

x2 b b2 a2
  
2 (b  a ) a 2 (b  a ) 2 (b  a )

b2  a2 (b  a ) ( b  a ) (b  a )
  
2 (b  a ) 2 (b  a ) 2

That is

ab
  E ( x) 
2

The Variance and Standard Deviation

The variance and standard deviation of a uniform random

variable are obtained as

b
E ( x 2 )  x 2  f ( x) dx
a
 - 176 -

b 1 b x2
x  dx  
2
dx
a ba a ba

x3 b b3 a3
  
3 (b  a ) a 3 (b  a ) 3 (b  a )

b3  a3 (b  a ) ( b 2  a b  a 2 )
 
3 (b  a ) 3 (b  a )

b2  a b  a2

3

By definition, the variance is given by

Var( x)  E( x 2 )  E ( x)
2

b2  a b  a2  a b
2

  
3  2 

b2  a b  a2 a2  2 a b  b2
 
3 4

4 (b 2  a b  a 2 )  3 (a 2  2 a b  b 2 )

12

a2  2 a b  b2 ( a  b) 2
 
12 12
 - 177 -

That is

( a  b) 2
Var ( x) 
12

and the standard deviation is given by

( a  b) 2
Std ( x ) 
12

Example 2:

If the random variable x is distributed as a uniform (2, 5)

Exponential Distribution

The probability density function of exponential random

variable X is given by

 e   x if x  0
f ( x)  
0 otherwise

Cumulative Distribution Function (cdf)

x x
F ( x)  

f ( x) dx    e   x d x


x
  e x  1  e x


That is, the cumulative distribution function is given by

 - 178 -

x 1  e   x x0
F ( x)  

f ( x) dx  
0 x 0

Mean

1
  E (X ) 


Proof:

Moment Generating Function


 (t )  E [ e t x ]  t
 t

d2 2
E (X ) 
2
 (t ) t 0 
d t2 2

Variance

V ( X )  E ( X 2 )  E ( X ) 
2 1
2
 - 179 -

CHAPTER 10

NORMAL DISTRIBUTION
- 180 -
 - 181 -

CHAPTER 10

NORMAL DISTRIBUTION

The normal distribution is the most important. The

graph of the normal distribution is the familiar
symmetrical, bell–shaped curve shown in the figure
below.
One reason for the importance of the normal
distribution is that it usefully models or describes the
distributions of numerous random variables that arise in
practice, such as the heights or weights of a group of
people, the grades of a class of students, and the
measurement errors that arise in the performance of an
experiment. In examples such as these, the observed
measurements tend to cluster in a symmetrical fashion
about the central value, giving rise to a bell–shaped
distribution curve.
A second reason for the normal distribution's
importance is that this distribution provides a useful
approximation to many other distributions including
discrete ones such as the binomial distribution.
Finally, the normal distribution is the cornerstone
distribution of statistical inference, representing the
distribution of the possible estimates of a population
parameter that may arise from different samples.
 - 182 -

The following are some important characteristics of the

normal distribution.
1. It is symmetrical about its mean.  where the curve on
either side of  is mirror image of the other side.
2. The mean, the median, and the mode are all equal.
3. The total area under the curve above the X-axis is one
square unit. This characteristic follows from the fact that
the normal distribution is a probability distribution.
Because of the symmetry already mentioned, 50 percent
of the area is the right of a perpendicular erected at the
mean and 50 percent is to the left.
4. If we erect perpendiculars a distance of one standard
deviation from the mean in both directions, the area
enclosed by these perpendiculars, the X-axis and curve
will be approximately 68.26 percent of the total area. If
we extend these lateral boundaries a distance of two
standard deviations on either side of the mean
approximately 95.44 percent of the area will be enclosed,
 - 183 -

and extending them distance of three standard deviations

will cause approximately 99.74 percent of the total area
to be enclosed.
5. The normal distribution is completely determined by
the parameter  and  . In other words, a different normal
distribution is specified for each different value of  and
.
Standard normal distribution :
There is a limitless number of normal distributions
because there is an infinite number of possible pairs of
means and standard deviations. We find normal
probability by using a probability table. Fortunately, we
need only one probability table when we standardize. If
we have a random variable X that normally distributed
with mean  and standard deviation  , then.

Z= X  is a normal with  = 0 ,  = 1

This standardization converts any normal distribution
into a normal distribution with a mean of 0 and a standard
deviation of 1 . Therefore the standard normal distribution
is a normal distribution with  = 0 ,  = 1 .
Note :
If x is normal distributed with mean  and standard
deviation  , then :

Z= X   is normal distributed with mean of 0 and

/ n
standard deviation of 1.
 - 184 -

Example 1:
Given the standard normal distribution, find the area
under the curve, above the X-axis between 0 and Z = 2 .

Solution:
It will be helpful to draw a picture of the standard
normal distribution and shade the desired area as in the
figure below. If we locate Z = 2 in table (1) and read the
corresponding entry in the body of the table, we find the
desired area to be 0.4772 . We may interpret this area in
several ways. We may interpret it as the probability that
a Z picked random from the population of Z's will have a
value between 0 and 2. We may also interpret it as the
relative frequency of occurrence (or proportion) of values
of Z between 0 and 2, or we may that 47.72 percent of the
Z's have a value between 0 and 2.

0 2

Example 2:
What is the probability that a Z picked from the
population of Z's will have between -2.55 and +2.55?
 - 185 -

Solution:
The figure shows the area desired. Table (1) gives
across until this area to be 0.4946. If we look at the picture
we draw, we see that this is only part of the area desired,
and if we recall that the normal distribution is symmetric
about the mean, we realize that we have found exactly one
half of the desired area. To obtain the total desired we
double .4946 to get 09892 .

- 2.55 0
2.55

Example 3:
What proportion of Z values are between -2.74 and 1.53?
Solution:
The figure shows the area desired. We find in table (1)
that the area between 0 and 1.53 is 0.4370 . Because of
symmetry we know that the area between 0 and -2.74 is
the same as the area between 0 and +2.74 which from the
table, we find is 0.4969 . To obtain the total area desired
we add these two areas to get 0.4370 + 0.4969 = 0.9339.
 - 186 -

- 2.74 0 1.53

Example 4:
Given the standard normal distribution find P (Z  2.71).
Solution :
The area desired is shown in the figure below. We
obtain the area to the right of Z = 2.71 by subtracting
the area between 0 and 2.71 from 0.5 Thus
P (Z  2.71) = 0.5 – P (0  Z  2.71)
= 0.5 – 0.4966 = 0.0034

0 2.71
 - 187 -

Example 5:
Given the standard normal distribution, find P (0.84  Z
 2.45)

Solution:
The area we are looking for is the figure below. We
first obtain the area between 0 and 2.45, and from that
subtract the area between 0 and 0.84 . In other words.
P (0.84  Z  2.45) = P (0  Z  2.45) – P (0  Z  0.84)
= 0.4929 – 0.2995 = 0.1934

0 0.84 2.45

Application:
Although its importance in the field of statistics is
indisputable, one should realize that the normal
distribution is not a law that is adhered to be all
measurable characteristics occurring in nature. It is true,
however, that many of these characteristics are
 - 188 -

approximately normally distributed. Human stature and

human intelligence are frequently cited as examples.
Whenever it is known that a random variable is
approximately normally distributed or when, in the
absence of complete knowledge, it is considered
reasonable to make this assumption, the statistician is
aided tremendously in his or her efforts to solve practical
problems relative to this variable.
There are several other reasons why the normal
distribution is so important in statistics, and these will be
considered in due time. For now, let us se how we may
answer simple probability questions about random
variables when we know, or are willing to assume, that
they are at least, approximately normally distributed.

Example 6:
A physical therapist feels that scores on a certain
manual dexterity test are approximately normally
distributed with a mean of 10 and a standard deviation of
2.5 . If a randomly selected individual takes that, what is
the probability that he or she will make a score of 15 or
better?
Solution:
First let us draw a picture of the distribution and shade
the area corresponding to the probability of interest. This
has been in the figure below. If our distribution was the
standard normal distribution with a mean of 0 and
standard deviation of 1, we could make use of table (1)
and find the probability with little effort Fortunately, it is
 - 189 -

possible for any normal distribution to be transformed

easily to the standard normal. What we do is transform all
values of X to corresponding values of Z. This means that
the mean of X must become 0, the mean of Z. In the
figures, both distribution are shown.
We must determine what value of: say Z0 corresponds
to an x of 15. This is done by the following formula:

X 
Z 


Which transforms any value of x in any normal

distribution to the corresponding value of Z in the
standard normal distribution. For the present example we
have.

15  10 5
Z   2
25 2 .5

The value of Z0 we seek, then, is 2.

Let us examine these relationships more closely. It is
seen that distance from the mean, 10, to the X-value of
interest, 15, is 15 – 10 = 5 , which is a distance of 2
standard deviations. When we transform x values to Z
values, the distance of the Z value of interest from its
mean, 0, is equal to the distance of the corresponding x
value from its mean, 10, in standard deviation units. We
have seen that this latter distance is 2 standard deviation .
In the Z distribution a standard deviation is equal to one,
and consequently the point on the Z scale located a
distance of 2 standard deviations from 0 is Z = 2, the
result obtained by employing the formula. By consulting.
 - 190 -

Table (1), we find the area to the right of Z = 2 is 0.0228.

We may summarize the above discussion as follows:

0 2

 15  10 
P ( X  15)  P  Z  
 2 .5 

 P ( Z  2)  0.5  P (0  Z  2)

To answer our original question, we say that probability

that random chosen individual taking the test will make a
score of 15 or better is 0.0228.

Example 7:
Suppose it is known that the weights of a certain group
of individuals are approximately normal distributed with
a mean of 140 pounds and a standard deviation of 25
pounds. What is the probability that a person picked at
random from this group will weigh between 100 and 170
pounds.
 - 191 -

Solution:
The figure below shows the distribution of weights and
the Z distribution to which we transform the original
values to determine the desired probabilities. We find the
Z value corresponding to an x of 100 by:

100=-140
1.6
Z
25

Similarly x = 170 we have

170  140
Z  1.2
25

- 1.6 0 1.2

From Table (1), we find the area between 0 and -1.6 to be

0.4452, and the area between 0 and 1.2 to be 0.3849. The
area desired is the sum of 0.4452 + 0.3849 = 0.8301
To summarize
 100  140 170  140 
P (100  x  170) = P  Z 
 25 25 
= P (01.6  Z  1.2)
 - 192 -

= P (-1.6  Z  0) + P (0  Z  1.2)
= 0.4452 + 0.3849 = 0.8301
The probability asked for in our original question, then ,
is 0.8301 .
 - 193 -

EXERCIES 10

Given the standard normal distribution find:

1. The area under the curve between
Z = 0 and Z = 1.43
2. The probability that a Z picked at random will have a
value between Z = -2.87 and Z = 2.64
a. P (Z  0.55).
b. P (Z  -0.55).
c. P (Z  2.33).
d. P (Z  2.33).
e. P (-1096  Z  1.96).
f. P (-258  Z  1.58)
g. P (-1.65  Z  1.65)
h. P (Z = 0.74)

3. Given a normally distributed population with a mean

of 100 and a standard deviation of 20, find the following
probabilities based on a sample of size 16:
(a) P ( X  100)
(b) P (96  x  108)
(c) P ( X  110)
 - 194 -

X= 50,  = 16, n = 64, find

2
4. Given :
(a) P 45  X  55)
(b) P (  53)
X
(c) P ( X  47)
(d) P (49  X  56)

5. Suppose it is known that the hourly wages of a certain

type of hospital employee are approximately normally
distributed with a mean and standard deviation of L.E
54.50 and L.E 5.50 respectively. If a random sample of
size 16 is selected from this population, find the
probability that the mean hourly wage for the sample will
be:
(a) Greater than L.E 54.25 .
(b) Between LE. 54.25 and LE. 55.
(c) Greater than LE. 54.80 .
(d) Less than LE. 54.20 .

6. It has been found that, following a period of training

the mean time required for certain handicapped persons
to perform a particular task is 25 seconds with a standard
deviation of 5 seconds. Assuming a normal distribution
of times, find the probability that a sample of 25
individuals will yield a mean :
(a) Of 26 seconds or more.
 - 195 -

(b) Between 24 and 27 seconds.

(c) 26 seconds or less.
(d) Greater than 22 seconds.

7. If the uric acid values in normal adult males are

approximately normal distributed with a mean and
standard deviation of 5.7 and 1 mgs percent respectively,
find the probability that a sample of size 9 will yield a
mean :
(a) Greater than 6 .
(b) Between 5 and 6 .
(c) less than 5.2 .

8. For a certain large segment of the population, for a

particular year, suppose the mean number of days of
disability is 5.4 with a standard deviation of 2.8 days.
Find the probability that a random sample of size 49 from
this population will have a mean:
(a) Greater than 6 days.
(b) Between 4 and 6 days.
(c) Between 4 1/2 and 5 1/2 days.
- 196 -
 - 197 -

CHAPTER 11

ESTIMATION Theory
- 198 -
 - 199 -

CHAPTER 11

ESTIMATION Theory

The objective of estimation is to determine the

approximate value of a population parameter on the
basis of a sample statistic. An estimator of a population
parameter is a random variable that is function of the
sample data. An estimate is X the calculation of a
specific value of this random variable. We know that the
sample mean X is an estimator of the population mean
, and that the sample variance S2 is an estimator of the
population variance 2 . After the sample has been
drawn and the statistics have been calculated, the value
of X becomes the estimate of  and the value of S2
becomes the estimate of 2 .
We can use sample data to estimate a population
parameter in two ways. First, we can compute the value
of the estimator and consider the value as the estimate
of the parameter. Such an estimator is called a point
estimator. In drawing inferences about a population, we
expect that a large sample will produce more accurate
results, because it contains more information than
smaller sample does. Point estimators don't have the
capacity to reflect the effects of larger sample sixes. The
second way of estimating a population parameter is use
an interval estimator.
 - 200 -

Estimation:
The estimation is a technique by which, we can drawn
inferences about a population.

Estimator:

An estimator of a population parameter is a random

variable that is a function of the sample data, such as  X
 X  X 
2 n
and
n

Estimate:
An estimate is the calculation of a specific value of
random variable.

Point Estimator:
A point estimator draws inferences about a population
by estimating the value of an unknown parameter using
a single value or point.

Interval Estimator:
An interval estimator draws inferences about a
population by estimating the value of an unknown
parameter using an interval that is likely to include the
value of the population parameter.
 - 201 -

Estimating the population mean when the population

variance is known.
We now describe how an interval estimator produced
from a sampling distribution Suppose we have a
population with mean  and variance 2. The population
mean  is assumed to be unknown and our task is to
estimate its value. As we just mentioned, the technique
of estimation involves drawing a random sample of size
and calculating the sample mean. We know that is
normally distributed. This means the variable.
X 
Z 
 / n
is standard normally distributed. When a sample is
drawn, the value of calculated from the sample is such
that

X 
Z 
 / n

can be anywhere a long the horizontal axis from -  to

+  . We know that the probability that z falls between
– Za/2 and + Za/2 is equal to 1- . This can be expressed as

 X  
1  P   Z a / 2   Z a / 2 
 / n 
   
1    P   Z a / 2  X    Za/2 
 n n
 - 202 -

   
1   P  X  Z a / 2    X  Za/2 
 n n

This equation says that, with repeated sampling from

this population , the proportion of value of X for which
 
 falls between X  Z  / 2 and X  Z  / 2 is equal to
n n
1- . This interval is called the confidence interval
estimator of  . A shortcut method of representing it is:

X  Z / 2
n

In other words, the confidence interval estimator of 

can be summarized as:


 X  Z / 2
n

 The probability 1- is called the confidence

level.

 X  Z / 2 is called the lower confidence
n
limit (LCL).


 X  Z / 2 is called the upper confidence
n
limit (UCL)

From the confidence level 1- we determine , we

determine , /2 and Za/2 as in the following table.
 - 203 -

Confidence Significance
Level 1- level Za/2
 /2
0.90 0.10 0.05 1.645
0.95 0.05 0.025 1.960
0.99 0.01 0.005 2.576

Example 1:
Ina random sample of size n = 100 from a population
whose variance is 2 = 18. We calculated X = 30. Find
the 90% confidence interval estimate the population
mean .

Solution:
X = 30 =9
n = 100 Za/2 = 1.645
The 90% confidence interval estimate will be as follows.

   
 X  Z a / 2    X  Za/2 
 n n
9 9
X  1.645    X  1.645
100 100

X  1.4805   X  1.4805
 - 204 -

When we substitute X = 30 in the interval estimator of

the population mean , we will obtain the confidence
interval estimate of  as
30 – 1.4805    30 + 1.4805
28.5195    31.4805
Where:
LCL = 28.5195
UCL = 31.4850

Example 2:
The following observations were drawn from a
normal population whose variance is 100 :
140, 170, 150, 180, 130, 120, 100
Determine the 95% confidence interval of the
population mean.
Solution:

 X i 140 170 150 180 130 120 100

X  
n 7
= 990 = 141.429
7
X = 141.429 n =7
 = 10 Za/2 = 1.96
The 95% confidence interval estimator of the population
mean  will be .
 - 205 -

   
 X  Z a / 2    X  Za/2 
 n n
10 10
X  1.96    X  1.645
7 7

X  7.4081    X  7.4081

Substitute X = 141.429, we get the 95% confidence

interval estimate of 
141.429 – 7.4081    141.429 + 7.4081
134.021    148.837
Where:
LCL = 134.021
UCL = 148.837
 - 206 -

EXERCISES 11

1. In a random sample of 400 observations from

population whose variance is 2 = 100, we calculated
X =75. Find the 95% confidence interval estimate of
the population mean .

2. Suppose a random sample of fine observations is

taken from a normal population whose standard
deviation of 5. The results are 8, 15, 12, 6, 7 find the
99% confidence interval estimate of the population
mean.

3. A random sample of 400 observations from a

population whose standard deviation is 90 produced
X = 1500. Find the 90% confidence interval estimate
of  .

4. Given the following information, determine the 95%

confidence interval estimate of  X. n = 1000 ,  = 40,
= 125.

5. The following observations were drawn from a normal

population whose variance is 100 : 12, 8, 22, 15, 30,
 - 207 -

6, 39, 48 . Determine the 90% confidence interval of

the population mean.

6. In a random sample of students in a large university,

a dean found that the mean weekly time devoted to
homework was 14.3 hours, if we assume that
homework time is normally distributed, with a
standard deviation of 4 hours, find the 99%
confidence interval estimate of the weekly time spent
doing homework for all the university's students.

7. As part of a project to develop better lawn fertilizers,

a research chemist wanted to know the mean weekly
growth rate of Kentucky Bluegrass. A sample of 250
blades of grass was measured, and the amount of
growth in one week was recorded. The mean was
calculated to be 0.76 inches. Estimate with 99%
confidence the mean weekly amount growth,
assuming that Kentucky Bluegrass growth is normally
distributed with standard deviation 0.13 inches.
8. Suppose it is known that the hourly wages of a certain
type of hospital employee are approximately
normally distributed with standard deviation of LE. 50
. If a randomX sample of size 16 is selected from this
population . We calculated = LE 300. Determine the
95% confidence interval of the population mean  .
 - 208 -

9. If the uric acid values in normal adult males are

approximately normally distributed with a variance of
one mgs percent, A random sample of size 9 was
drawn and we calculated X = 51. Estimate the 99%
confidence interval of the population mean  .
 - 209 -

CHAPTER 12

CHI-SQUARE TESTS
- 210 -
 - 211 -

CHAPTER 12

CHI-SQUARE TESTS

The chi-square tests are used to determine whether:

a. two random variables are independent,
b. two or more population proportions are equal, and
c. an observed distribution and a hypothesized
distribution are equal.
The chi- square test can be outlined as follows:
1. Establish the null and alternative hypotheses (Ho and
H1, respectively)
2. Draw he random sample (or samples)
3. Calculate the expected frequencies.
4. Calculate the chi-square statistic.

(Oi  Ei ) 2
 2 
Ei

5. Set alpha and the determine degrees of freedom,

and then look up the criticalvalue
2

of chi-square;
(tabulated  ) consulting Table (2).
2

Where : Oi is the ith observed frequency.

E1 is the ith expected frequency
6. Reject H0 when the calculated  2  the tabulated  
2

and otherwise accept H0.

 - 212 -

Example 1:
Consider a subset of the data from a case-control
study on smoking and lung cancer.

Smoker Nonsmoker Total

Lung cancer 60 3 63
Without lung 32 11 43
cancer
Total 92 14 106

Test to see are smoking and lung cancer independent

variables or not using level of significance of 0.05 .
Solution:
The hypotheses:
H0 : The smoking and lung cancer are independent .
H1 : The smoking and lung cancer are not independent .
The expected frequencies:

(ith row total)  ( jth column total)

Eij 
Grand total

63 92
E11   54.68
106

6314
E12   8.32
106
 - 213 -

43 92
E 21   37.32
106

43  14
E 22   5.68
106

Now, we have the observed and expected frequencies;

they are:

Observed Frequency Expected Frequency

O11 = 60 E11 = 54.68

O12 = 3 E12 = 8.32

O21 = 32 E21 = 37.32

O22 = 11 E22 = 5.68

Total = 106 Total = 106

The chi-square statistic:

r c (Oij  Eij ) 2
 
2
i 1 j 1 Eij

r = number of roes, r = 1, 2
c = number of columns, c = 1, 2
 - 214 -

2 2 (Oij  Eij ) 2
 
2
i 1 j 1 Eij

(60  54.68) 2 (3  8.32) 2 (32  37.32) 2 (11 5.68) 2

   
54.68 8.32 37.32 5.68
= 0.52 + 3.40 + 0.76 + 4.98 = 9.66

The critical value of chi-square:

To look up the critical value of chi-square, it must
have two parameters; they are the significance level ; 
; and the degrees of freedom . In this example
 = 0.05
Degrees of freedom = (r – 1) (c – 1) = 1

 (21,0.05)  3.84 (tabulated  2 )

Statistical decision:
Since the calculated  2 is greater than tabulated  2 ,
then we reject H0 .
Clinical decision:
On the basis of these data, we may deduce that there
is an association between the smoking and lung cancer.
 - 215 -

Example 2:
The following data pertain to a cohort of 609 healthy
individuals between the ages of 40 and 76. The cohort
was followed for 7 years, after which new cases of
coronary heart disease (CHD) were identified. The level
of circulating catecholamine (CAT) is the exposure
variable of interest, The following table present the
results.

High CAT Low CAT Total

CHD 27 44 71
No CHD 95 443 538
Total 122 487 609

Test to see whether the CHD and CAT are

independent or not using level of significance of 0.05 .
Solution:
The hypotheses:
H0 = The CHD and the CAT are independent.
H1 = The CHD and the CAT are not independent.
The expected frequencies:

(ith row total)  ( jth column total)

Eij 
Grand total

71  122
E11   14.22
609
 - 216 -

71  487
E12   56.78
609

548  122
E 21   107.78
609

538  487
E 22   430.22
609

Now , we have the observed and expected frequencies :

O11 = 27 E11 = 14.22

O12 = 44 E12 = 56.78
O21 = 95 E21 = 107.78
O22 = 443 E22 = 430.22

Total 609 609

The chi-square statistic (calculated  2 :

2 2 (Oij  Eij ) 2
 
2
i 1 j 1 Eij

(27 14.22) 2 (44  56.78) 2 (95  107.78) 2 (443  430.22

   
14.22 56.78 107.78 430.22

= 11.49 + 2.88 + 1.52 + 0.38 = 16.27

 - 217 -

The critical value of chi-square (tabulated  2 ) :

12,0.05  3.84

Statistical decision:
Since the calculated  2  tabulated  2 , then we
reject H0 .
Clinical decision:
On the basis of these data, we may extract the CHD
and CAT are related variables.

Example 3:
A certain drug is claimed to be effective in curing
colds. In an experiment on 164 people with cold, half of
them were given the drug and half of them were given
sugar pills. The patients reactions to the treatment are
recorded in the following table. Test the hypothesis that
the drug and the sugar pills yield similar reactions with 
= 0.01.

Helped Harmed No Total

effect
Drug 52 10 20 82
Sugar 44 12 26 82
Total 96 22 46 164
 - 218 -

Solution:
The hypotheses:
H0 = The drug and the sugar pills yield similar
reactions.
H1 = The drug and the sugar pills don't yield similar
reactions.

The expected frequencies:

(ith row total)  ( jth column total)

Eij 
Grand total

82  96
E11   48
164

82  22
E12   11
164

82  46
E13   23
164

82  96
E 21   48
164

82  22
E 22   11
164

82  46
E 23   23
164
 - 219 -

Now , we have the observed and expected frequencies :

Observed Frequency Expected Frequency

O11 52 E11 48
O12 10 E12 11
O13 20 E13 23
O21 44 E21 48
O22 12 E22 11
O23 26 E23 23

Total 164 Total 164

The chi-square statistic (calculated  2 ) :

2 2 (Oij  Eij ) 2
 
2
i 1 j 1 Eij

(52  48) 2 (10 11) 2 (20  23) 2 (44  48) 2 (12 11) 2 (26  23) 2
     
48 11 23 48 11 23

= 0.33 + 0.09 + 0.39 + 0.33 + 0.09 + 0.39 = 1.62

The critical value of chi-square:

 = 0.01
Degrees of freedom = (r – 1) (c – 1)
 - 220 -

= (2 -1) (3 – 1) = 2
 22,0.01  9.21 (tabulated  2 )

The decision:
Since the calculated X2 is smaller than the tabulated
X2 , we accept H0. That is the drug and sugar pills yield
similar reactions.
Example 4:
In a breeding experiment of flowers, the results were
recorded as follows:

flowers Red Yellow Pink White Total

Number 149 211 191 119+ 670

If we know that for large enough numbers, it is expected

that the colour of the flowers are distributed as 1 : 2 : 1.
Test that the observed and expected distributions are
the same using  = 0.05
Solution:
The hypotheses:
H0 : The observed distributions and the expected
distribution are equal.
H1 : The observed distributions and the expected
distribution are not equal.
 - 221 -

The expected frequencies:

The expected frequencies have been obtained by
dividing the total number of flowers according to the
ratio predicted by theory follows :
1
The expected number of red flowers = 670  5 = 134

The expected number of yellow flowers = 670  2 = 268

5
The expected number of pink flowers = 670  1 = 134
5
1
The expected number of whites flowers = 670  5 = 134
Now, the observed and expected number of flowers,
and calculation of  2 can be summarized in the following
table.

Expected
Flower Observed (O – (O – (O  E ) 2
number
colour number E) E)2 E E
(E)
Red 149 134 15 225 1.68
Yellow 211 268 -57 3249 12.12
Pink 191 134 57 3249 24.25
White 119 134 -15 225 1.68
Total 670 670 39.73
The summation of the last column represents the chi-
square statistic. That is,
The calculated X2 = 39.73
The critical value of chi-square with  = 0.05 degrees
of freedom will be:
 - 222 -

(tabulated  2 )
 2
3, 0.05  7081

The decision :
Since the calculated  2 is greater than tabulated  2 ,
we reject the null hypothesis; H0 . That is the observed
distribution and expected distribution are not equal.

Example 5 :
The number of field mice caaptured at five traps
during a period of 3 months were recorded as follows :

Traps A B C D E Total
Number 23 7 25 19 21 95

Uaing  = 0.01, test the claim that an equal number of

mice will be caught at every trap.
Solution:
H0 : The observed distribution and the expected
distribution are equal.
H1:The observed distribution and the expected
distribution are not equal.
The expected number of mice captured at each trap
95
  19 mice
5
 - 223 -

And therefore the observed and expected number of

mice, and calculation of  2 can be summarized in the
following table.

Expected
Observed (O  E ) 2
Trap number (O – E) (O –E)2
number (O) E
(E)
A 23 19 4 16 0.84
B 7 19 -12 144 7.58
C 25 19 6 36 1.89
D 19 19 0 0 0.00
E 21 19 2 4 0.21
Total 95 95 10.52

The calculated  2 = 10.52

The critical value of chi-square with  = 0.01and 4

degrees of freedom (tabulated  2 ) :

 32,0.05  7081

The decision :
Since the calculated  2 is smaller than tabulated  2 ,
we accept the null hypothesis; H0 . That is the observed
distribution and the expected distribution are equal.
 - 224 -

EXERCIES 12
1. A management behavior analyst has been
studying the relationship between male/female
supervisory structures in the work place and the
level of employees' job satisfaction. The results of a
recent survey are shown in the table below.
Conduct a test with  = 0.05 to determine whether
the level of job satisfaction depends on the
boss/employee gender relationship.
Level of Boss / Employee
Satisfaction F/M F/F M/M M/F
Satisfied 21 25 59 71
Neutral 39 49 50 38
Dissatisfied 31 48 10 11

2. Two batches of barley seed were to be tested to

determine the effect of a certain heat treatment
upon the viability of the seed. Batch A was the
untreated ; control batch, and Batch B was given the
heat treatment. The results are set out in the
following table.
Viable Not viable Total
Batch A 64 16 80
Batch B 34 46 80
Total 98 62 160
Test using  = 0.01 to see are the effect of treatment
and viability of the seed independent.
 - 225 -

CHAPTER 13

ANALYSIS OF VARIANCE
- 226 -
 - 227 -

CHAPTER 13

ANALYSIS OF VARIANCE

Analysis of variance (ANOVA) is a statistical

procedure for teting the hypothesized equality of two or
more population means. The name of the technique is
derived from the way in which the calculations are
performed. That is, the technique analyzed the variance
of the data to determine whether we can infer that the
population means differ.
The process of performing the analysis of variance is
facilitated by the notation in the following table.
Independent Samples from K Population (Treatments)
Treatment
1 2 K
X11 X12 -- X1k
X12 X22 -- X2k
. . -- .
. . -- .
. . -- .
Xn,1 Xn,2 -- Xn,k
Sample size N1 N2 -- --
Sample mean X1 X2 -- Xk
 - 228 -

Xij = ith observation of the jth sample

nj = number of observations in the sample drawn from
the jth population .

X
i 1
ij

X 1  mean of the jth sample

n1
k n

  X
j 1 i 1
ij

X  grand mean of all the observations 

n

Where n = n1 + n2 + …… + nk and k is the number of

populations. Notice that we allow the sample sixes to be
different.
The variable x is called the response variable, and its
value are called responses. The unit that we measure is
called an experimental unit. That is, an experimental unit
is the unit of material to which one application of a
treatment is applied; the treatment is the procedure
whose effect is to be measured and compared with
other treatments. The experimental unit may be an
animal, 10 birds in a pen, a half-leaf, and so on . The
treatment may be a standard ration, a spraying
schedule, a temperature – humidity combination, etc.
When the effect of a treatment is measured, it is
measured on a sampling unit.
There are two ways to do analysis of variance are :
1. The first is with respect to means.
2. The second is with respect to summation .
 - 229 -

Here, we will use the first approach to do analysis of

variance.

Partitioning the total variation:

The total variation observed in all the data is the sum
of the squared deviations of each observation from the
grand partitioning the total variation into its explained
and unexplained components. That is,
2
 

k nj
Total variation = 
j 1 i 1
 ( x1  x 
 
2
k
  
Variations due to treatments = 
j 1
n j  (x j  x 
 
 

k nj

  ( xij  x j 
2

Variations due to treatments = j 1 i 1  

The total variations is referred to as total sum of

squares and denoted by SSTOT, the variations due to
treatments are named the variations between
treatments (explained component) and denoted by SST,
and the variations due to error are called the variations
within the treatment (explained component) and
denoted by SSE.

SSTOT = SST + SS

 

2 2 k nj
 
   
nj

k k

   ( xij  x j 
2
 ( x1  x  n j  (x j  x 
j 1 i 1   = j 1   + j 1 i 1  

Where :
 - 230 -

SSTOT = Total sum of squares.

SST = Treatment sum of squares.
SSE = Error sum of squares.
Proof :
2
 

k nj
SSTOT = 
j 1 i 1
 1 

( x  x

2
=  
  

k nj


j 1 i 1
 ( xij  x j  x j  x 
 
2
 
nj
k 
   
= 
j 1 i 1
 x
 ij

 x j  
 
 x j  x 


k nj
  2
  
  
    2

= 
j 1 i 1
 xij  x j   2  xij  x j   x j  x    x j  x  
       

nj nj    k   
k
 
 k
 
    

j 1 i 1
 xij  x j   2   x ji  x j    x j  x    n j  x j  x 
  j 1   j 1   j 1  
   

2
nj 2  
k
 
 
k

= 
j 1 i 1
 xij  x j    n j  x j  x 
  j 1  
 

 
 
k

Notcie that : 
j 1
 x j  x  =0
 
The test of the equality of two or more population
means can be cut lined as follows :

1. Establish the null and the alternative hypotheses:

 - 231 -

H0 : 1 = 2 = …… = k
H1 : 1 = 2 = …… = k
2. Drew independent random samples from populations
and calculate the sample means.
X1 , X2 , …….. Xk
3. Set alpha and determine the degrees of freedom
of both numberator and denominstor as (k-1) and
(n-k) respectively, where k is the number of means
being tested, and n is the total number of
observations,
Fka1,nk than look up the critical value of F,
consulting Table (3);
4. Calculate :
(a) The treatment sum of squares ; SST.
(b) The error sum of squares ; SSE.
(c) Mean squares of treatments; MST.
Where : SST
MST = k  1
(d) Mean squares of Error, MSE
Where : SSE
MSE = n  k
5. Calculate the F statistic .
F = MST
MSE
6. Reject H0 when F  F a
k 1 , n  k

Accept H0 when F  Fka1 , nk

 - 232 -

Now, we can summarize the preceding steps in the

following table which is called ANOVA table.

Source of Degrees Sums of Mean F-

Variation of squares squares statistic
freedom
SST
Treatments K -1 SST MST = k  1
MST
SSE F=
Error n -k SSE MSE = n  k MSE

Total n-1 SSTOT

Example 1:
Five methods of teaching were used on of
homogenous groups, at the end of the semester, we
have the scores in the final examination as follows.
Treatments
A B C D E
30 30 40 54 55
35 40 45 40 45
25 45 55 35 60
40 45 35 40 50

Test to see are there any significant differences

between methods used with  = 0.05 .
 - 233 -

Solution :
30  35  25  40 130
X1    32.5
4 4

30  40  45  45 160
X2    40
4 4

40  45  55  35 175
X3    43.75
4 4

45  40  35  40 160
X4    40
4 4

55  45  60  50 210
X5    52.5
4 4

 32.5  40  43.75  40  52.5

X   208.75  41.75
5

Treatment sum of squares:

SST = 5 n  X  X 
j  j 
j 1  
= 4  (32.5 – 41.75)2 + (40 – 41.75)2 + (43.75 -
)41.75)2 + (40- 41.75)2 + (52.5 – 41.75)2 

= 4 85.56 + 3.06 + 4.00 + 3.06 + 115.56

= 4 (211.24) = 844.97
Error sum of squares :
 - 234 -

SSE = 5 4
 

   X ij  X 
j 1 i 1  j 

= (30 – 32.5)2 + (35 – 32.5)2 + (25 – 32.5)2 + (40

– 32.5)2  + (30 - 40)2 + (40 – 40)2 + (45 – 40)2
+(45 – 40)2  + (40 – 43.75)2 + (45 -43.75)2 +
(35 – 43.75)2  + (45 – 40)2 + (40 – 40)2 + (35
– 40)2 + (40 – 40)2  + (55 – 52.5)2 + (45 –
52.5)2 + (60 – 52.5)2 + (50 – 52.5)2 
= 6.25 + 6.25 + 56.25 + 56.25 + 100 + 0 + 25 +
25 + 14.06 + 1.56 + 156.56 + 76.56 + 25 +
0 + 25 + 0 + 6.25 + 56.25 + 56.25 + 6.25
= 125 + 150 + 248.74 + 50 + 125 + = 698.74
Now, we can construct the ANOVA table of this
experiment as follows :

Source Degrees Sums of Mean F-statistic

of of squares squares
Variation freedom
844.97
Treatme 4 SST = 844.97 MST =
4
nts = 211.24
211.24
F=
15 SSE = 698.74 MSE = 698.75 46.58
Error 15 = 4.53
= 46.58

Total 19 SSTOT=1543.71
Calculated F = 4.53
Critical value of F (tabulated F)
 - 235 -

.05 = 4.08
F40,15
Decision
Since the calculated F is greater than the tabulated F, we
reject H0 . That is the responses of the treatments are
different ( 1  2  3  4  5 )
 - 236 -

EXERCISES 13

1. The following data give the yields of wheat on some

experimental plots of ground corresponding to
different sulfur treatments for the control of rust. The
treatments consisted of dusting before rains, dusting
after rains, dusting once each week, and no dusting.
Test to see if there are significant differences in yields
due to the dusting methods using  = 0.05 (level of
significance) .

Dusting method
1 2 3 4
5.3 4.4 8.4 7.4
3.7 5.1 6.0 4.3
14.3 5.4 4.9 3.5
6.5 12.1 9.5 3.8

2. An estrogen assay of several solutions which had been

subjected to an in vitro inactivation techniques is
made. The uterine weight of the mouse was used as
a measure of estrogenic activity. Uterine weights in
milligrams of four mice for each of control and six
different solutions are shown in the following table.
 - 237 -

Control I II III IV V VI
89.8 84.4 64.4 75.2 88.4 56.4 65.6
93.8 116.0 79.8 62.8 90.2 83.2 79.4
88.4 84.0 88.0 62.4 73.2 90.4 65.4
112.6 68.6 69.4 73.8 87.8 85.6 70.2
Test to see if there are significant differences in
weights due to the different solutions using level of
significance of 0.01 .

3. Given the following data:

Treatments
1 2 3 4
11 13 21 10
4 9 18 4
6 14 15 19
21 36 54 33

Using  = 0.05 , test the hypotheses:

H0 : 1 = 2 = 3 = 4
Vs
H0 : 1  2  3  4
Hint : SSTOT = 358, SST = 186, SSE ] 172

4. The following data represent the scores made by 24

typists during an experiment to determine whether
 - 238 -

there were any differences between three different

brands of typewriters. The 24 typists were split into
three equal groups by random selection, with each
group of 8 then assigned to one of the brands of
typewriters.
Brands
I II III
44 40 54
39 37 50
33 28 40
56 53 55
43 38 45
56 51 66
47 45 49
58 60 65
Are there significant differences in scores due to the
brands ? (  = 0.05 ).

5- Three groups of cows each of four observations

feeded with different kinds to see the effect of the
growth rate. After the experiment, the results are
recorded as follows.

Treatments
Feed I II III
240 185 115
210 195 125
195 210 135
220 200 130
 - 239 -

What can you say about the effect of these 3 kinds of

food using  = 0.01 ?

6.
Treatments
I II III
12 10 9
11 10 7
6 4 3
7 8 5
Test 1 = 2 = 3 using  = 0.01
 - 240 -

General Exercises
1. Tossing a balanced coin consecutive tree times, the
probability of getting no heads ……………..
2. Tossing a balanced coin consecutive ten times, the
number of outcomes in its sample space ….........
3. Rolling a balanced die, the probability of getting a
prime odd number is equal to……………
4. Rolling a balanced die, the probability of getting no
prime even number is equal to……………
5. Rolling a balanced die twice, the probability of
getting a sum of prime even number is …………
6. Rolling a balanced die twice, the probability of
getting a sum of prime even number is ………….
7. The number of outcomes if rolling a balanced die
consecutive three times is…………..
8. Rolling a balanced die twice, the probability of
getting a sum of prime number divisible 3 …………..
9. The quotient of the number of elements in an event
space and the number of elements in a sample
space…..
10. The mean of a distribution is 14 and the standard
deviation is 5. What is the value of the coefficient of
variation?
11. The sum of deviations about mean is always…..
 - 241 -

12. The middle value of an ordered array of number is

the ………
13. The standard deviation of a population is 9 , the
population variance is……..
14. The mean deviation of the values 8 ,7 , 10 , 7 is……
15. The standard deviation of 4, 4, 4, 4, 4, 4 is……
16. The lowest value of variance is……

17. Given the following probability distribution of

random variable X, it is required to answer the
following questions.
x -1 0 2
P(x) 1/3 1/3 1/3

a. The standard deviation of X is…….

b. The expected value of X is…….
C, The expected value of X2 is…….
d. The variance of X is………
e. The coefficient of variation of X is………..