Probability

Introduction
Probability is a measure of likelihood, the possibility or chance that an event will happen in
future.
It can be considered as a quantification of uncertainty.
Uncertainty may also be expressed as likelihood, chance or risk theory. It is a branch of
mathematics concerned with the concept and measurement of uncertainty.
Much in life is characterised by uncertainty in actual decision making.
Probability can only assume a value between 0 and 1 inclusive. The closer a probability is to zero
the more improbable that something will happen. The closer the probability is to one the more
likely it will happen.
Definitions of key terms
Random experiment results in one of a number of possible outcome e.g. tossing a coin
Outcome is the result of an experiment e.g. head up, gain, loss, etc. Specific outcomes are
known as events.
1 0 2 quantitative techniques
STUD Y TEX T
Trial- Each repetition of an experiment can be thought of as a trial which has an observable
outcome e.g. in tossing a coin, a single toss is a trial which has an outcome as either head or
tail
Sample space is the set of all possible outcomes in an experiment e.g. a single toss of a coin,
S=(H,T). The sample space can be finite or infinite. A finite sample space has a finite number of
possible outcomes e.g. in tossing a coin only 2 outcomes are possible.
An infinite sample space has an infinite number of possible outcomes e.g. time between arrival
of telephone calls and telephone exchange.
An Event of an experiment is a subset of a sample space e.g in tossing a coin twice S= (HH, HT,
TH, TT) HH is a subset of a sample space.
Mutually exclusive event - A set of events is said to be mutually exclusive if the occurrence of
any one of the events precludes the occurrence of other events i.e the occurrence of any one
event means none of the others can occur at the same time e.g. the events head and tail are
mutually exclusive
Collectively exclusive event - A set of events is said to be collectively exclusive if their union
accounts for all possible outcomes i.e. one of their events must occur when an experiment is
conducted.
Favourable events refers to the number of possible occurrences of a given event in an
experiment
e.g. if we pick a card from a deck of 52 cards the number favorable to a red card is 26, in tossing
a coin the number favourable to a head is one.
Independent events – Events are independent if the happening or non-happening of one has no
effect on the future happening of another event. E.g. in tossing two times of a coin, the outcome
of 1st toss does not affect 2nd toss.
Equally likely events – Events are equally likely if the happening of one is not favoured over
the
happening of others. In tossing a coin the tail and head are equally likely.
OTHER CONCEPTS
Unconditional and conditional probabilities – with unconditional probability we express
probability of an event as a ratio of favourable outcomes to the number of all possible
outcomes.
A conditional probability is the probability that a second event occurs if the first event has
already occurred.
Joint probability – joint probability gives the probability of the joint or simultaneous
occurrence
of two or more characteristics.
Marginal probability – is the sum of two or more joint probabilities taken over all values of one
or more variables. It is the probability that the results when we ignore one or more criteria of
classification when computing probability.
103
STUD Y TEX T
Probability
Industry context
Probability is used throughout business to evaluate financial risks and decision-making.
Every decision made by management carries some chance for failure, so probabiity
analysis is conducted formally.
In many natural processes, random variation conforms to a particular probability
distribution known as the normal distribution, which is the most commonly observed
probability distribution.
EXAM CONTEXT
The probability topic has been examined previously:
6/06, 12/05, 6/05, 12/04, 6/04, 6/03, 12/02, 6/02, 12/01, 12/00, 6/00
Laws of Probability
1. Rules of Addition
a) Special rule of addition
If two events A and B are mutually exclusive the probability of one or other occurring is equal to
the sum of their probability
P (A or B) = P (A) + P (B)
P (A or B or C) = P (A) + P (B) + P(C)
Illustration
An automatic plastic bag - a mixture of beans, broccoli and other vegetables, most of the
bags contain the correct weight but because of slight variations in the size of beans and other
vegetables. A package may be slightly under or overweight. A check of 4,000 packages of past
reveals the following:
Weight E vent N o of packages
Underweight A 100
Satisfactory B 3600
Overweight C 300
What is the probability that a particular package will be either underweight or overweight?
The two events are mutually exclusive
P (A or C) = P (A) + P (C)
P(A) =100/4000
P(C) =300/4000
P(A or C) =400/4000 = 1/10 =0.1
1 0 4 quantitative techniques
STUD Y TEX T
Note: The probability that a bag of mixed vegetable is selected underweight P(A) plus the
probability that it is not underweight P(NA) must logically be equal to one. This is referred to as
Complement rule.
b) General rule of Addition
The outcome of an experiment may not be mutually exclusive. This rule is used to combine
events that are not mutually exclusive.
P(A or B) = P(A) + P(B)- P(A and B)
Example
a) What is the probability that a card chosen at random from a pack of well shuffled deck will
either be a king or a heart
P(king or heart) = P(king) + P(heart) - P(king and heart)
4 + 13 + 1 = 16
52 52 52 52
b) R outine physical examination is conducted annually as part of the programme of a particular
organisation. It was discovered that 8% needed correcting shoes, 15% needed major dented
work and 3% needed for both correcting shoes and major dental work.
What’s the probability that an employee selected at random will either need correcting shoes or
major dental work?
2. Rule of Multiplication
a) Special Rule of multiplication
For two events A and B the probability that A and B will both occur is found by multiplying the
probability
P(A and B)=P(A) × P(B)
P(A and B and C) = P(A) × P(B) × P(C)
This rule is applicable for independent events i.e. the occurrence of one does not depend on the
occurrence of the other.
Example
i) T wo coins are tossed. What is the probability that both will land tails up
The two events are independent
P (T and T) = P (T) × P (T)
=1/2 × 1/2 = ¼ = 0.25
105
STUD Y TEX T
Probability
ii) From long experience firestone tyres have a 0.8 probability that their xB.70 will last 40,000
miles before it becomes bald and adjustments made. If you purchase four xB.70
What’s the probability that all tyres will last 40,000 miles?
What’s the probability that at least two will last 40,000 miles?
Solution
P(L and L and L and L) =P(L) × P(L) × P(L) × P(L)
= 0.8 × 0.8 × 0.8 × 0.8
= 0.4096
Probability of lasting, P(L) = 0.8; Probability of not lasting, P(N)= 0.2
Probability of at least two tyres lasting 40,000 miles is given by the following combinations:
b) General rules of Multiplication
It states that of two events A and B the joint probability that both events will happen is found by
multiplying the probability that A will happen by the conditional probability of event B
happening
P(A and B) = P(A) × P(B| )
Where P(B/A) stand for probability that B will occur given that A has already occurred.
Example
1. Assume that there are 10 rolls of film in a box 3 of which are defective. Two rolls are to be
selected one after another. What’s the probability of selecting a defective roll followed by
another defective roll?
P (D and D) = P (D) × P (D | D)
3/10 × 2/9 =6/90 =0.07
2. T hree effective toothbrushes were accidentally shipped to a chemist by Clean Brand Product
along with 17 non-effective ones
What’s the probability that of the first 2 toothbrushes sold one will be effective and the other one
will not.
Solution
P (One defective = P (D and N) or P (N and D)
= 2/20 x 17/19 + 17/20 x 3/19
Contingency Table
A two dimensional contingency table is formed by classifying two factors. One factor determines
the row categories and the other determines the column categories. Each element in the table
belongs to the two classifications known as a cell e.g. classifying subjects by gender (M and F)
and smoking status (Current/Former/Never). Such categories are said to be mutually exclusive
and collective. Exhaustive mutually exclusive means the categories include all possibilities and
so there is a category for everyone.
1 0 6 quantitative techniques
STUD Y TEX T
Table 3.1
Smoking Status
Gender
Current Former Never
Male a b c
Female d e f
The two factors here are smoking status and gender and the cells are a,b,c,d,e and f.
Example
A survey of executives dealt with their loyalty to the company. Out of the questions asked “If
you
were given an offer by another Company equal or slightly better than the present will you remain
in the Company or take another position. The responses were classified with their level of service
within the company.
Length of Service
Loyalty Less than 1
year
1-5 years 6-10 years More than
10 years
Total
Remain
Not Remain
10
25
30
15
5
10
75
30
120
80
200
What’s the probability of randomly selecting an executive who is loyal to the company (who will
remain) and who has been in the service in more than 10 years?
P(R and >10) = P(R) × P (>10|R)
= 120 × 75
200 120
= 75
200
What’s the probability of selecting at random an executive who will remain loyal to the company
and has less than 6 years of service
= 10 + 30
200
= 40
200
= 0.2
Tree Diagram (Probability Trees)
Is very useful for portraying conditional and joint probabilities. It is particularly useful for
analysing
business decisions where there are several stages of problems.
107
STUD Y TEX T
Probability
Bayes Theorem
In the 18th Century Rev Thomas Bayes, an English Presbyterian Minister, ordered this question,
“Does God really exists” Being interested in mathematics he attempted to develop a formula to
arrive at the probability that God really exists based on the evidence available on earth.
Later, Laplace refined Bayes work and gave it the name Bayes Theorem in a workable form, the
Bayes Theorem is
P(A/B) = P(A)P(B)
P(B)
= P(A1) × P(B/A1)
P(A) × P(B/A) + P(A2) × P(B/A2)
Note: P(PA and B) = P(A) × P(B/A)
Example 1
Suppose 5% of the population of a third world country has a disease peculiar to that country.
Let A1 refer to the event that a person has the disease and A2 refer to the event that a person
does not have the disease. Therefore, if a person is chosen at random, the probability that the
individual chosen will have a disease will be P(A1) = 0.05 thus P(A2) = 0.95.
These probabilities are called prior probabilities (they are assigned before any empirical data is
obtained). It’s the probability based on the present level of information. Assume further that
there
is a diagnostic technique to detect the disease but it is not accurate. Let B denote the event test
shows disease is present. Assume that historical evidence shows that if a person actually has the
disease the probability that the test will indicate presence of a disease is 0.9.
P (B|A1) =0.9
Assume that there is probability of 0.15 that a person does not actually have the disease but the
test indicates that the disease is present = (P (B/A2) = 0.15.
Let us randomly select a person from the country and perform a test. The test indicates that the
disease is present. What’s the probability that the person actually has the disease?
P(A1 | B) = P(A1) × P(B/A1)
P(A1) × P(B/A1) + P(A2) × P(B/A2)
Note: P(B) = P(A1) × P(B/A1) + P(A2) × P(B/A2)
Example 2
An electronic firm purchases its supplies from 4 different suppliers. A Ltd supplies 20%, B Ltd
30% C Ltd 25% and D Ltd 25%.
A Ltd tends to have the best quality. Only 3% of their supplies are defective. B Ltd supplies are
4% defective, C Ltd 7.0% and D Ltd 6.5% defective
Required:
a) What is the probability of selecting a defective item?
b) A defective supply was discovered in two shipments. What’s the probability that it came from
A Ltd
1 0 8 quantitative techniques
STUD Y TEX T
What’s the probability that the defective supply came from A ltd, C ltd and D ltd
Solution
a) What is the probability of selecting a defective item
P (D) = P( A and D) + P(B and D) + P(C and D)+P(D and D)
0.006+0.012+0.0175+0.01625
=0.05175
b) P(A/D) = P(A and D)
P(D)
= 0.006
0.05157
c) P(A/D) = P(A) × P(D/A)
= P(A) × P(D/A) + P(B) × P(D/B)) + P(C) × P(D/C)) + P(D)) × P(D/D))
P(D)
Principle of Counting
If the number of possible outcomes in an experiment is relatively easy to count and list all the
possible events e.g. there are six possible events resulting from a roll of a dice. If however,
there a large number of possible outcomes such as number of boys and girls for families with
10 children it would be tedious to list and count all the possibilities: we could have two boys and
eight girls, 1 boy and 9 girls etc. This situation is further complicated when we consider the order
of arrangements.
To facilitate counting, counting rules can be employed. These include multiplication rule,
permutation rule, and combination rule.
Multiplication Rule
If there are M ways of doing one thing and N ways of doing another thing, there are
M x N ways of doing both; that is, the total number of arrangements equals M×N. This can
be extended to more than two events e.g. for three events M, N and O the total number of
arrangements equal M × N × O.
Example I
An automobile dealer wants to advertise that for Kshs 500,000 you can buy a convertible, a
twodoor
or a four-door model with your choice as either a saloon or a station wagon. How many
different arrangements of models and car types can the dealer offer?
Solution
Convertible Saloon C onvertible S.W
Two-door saloon 2-door S.W
Four-door Saloon 4-door S.W
109
STUD Y TEX T
Probability
We can employ multiplication rule as a check when M- No of models
N - No of car types
Solution = 6
Example 2
The marketing manager of ABC Ltd would like to send a representative to the branches of a
company in Central Kenya. The company has three regions in Central Kenya, each with
5-branches. The sales representative will be based in one of the sub-branches. How many ways
can this be done?
M – No. of regions
N – No. of branches
M x N = 3x5 = 15
Solution =15 ways
Example 3
Pioneer manufactures 3 models of stereo receivers, 2 cassettes decks, 4 speakers and
3CDcarousels.
When the 4 types of compatible are sold together they form a system. How many
different systems can electronic firm offer?
2 x 4 x3=24syatems
Permutation Rule
The multiplication formula is applied in finding the number of possible arrangements for two or
more groups. The permutation formula is applied to find the number of possible arrangements
where there is only one group of objects.
Permutation is thus any arrangement of r objects from n possible objects. However this can be
calculated using the permutation formula:
nPr = n!
(n −
r)!
where; nPr = number of arrangements
n = number of items to be arranged
r = number of elements to be contained in the arrangement.
! = factorial e.g.; 4! = 4 × 3 × 2 × 1 = 24
Example 1
In how many ways can you arrange A, B and C?
Solution
In this case, note that the order in which the elements appear matters i.e ABC is different from
ACB.
Therefore
ABC BAC CA B
ACB BCA C BA
11 0 quantitative techniques
STUD Y TEX T
Answer: Six ways.
Using the permutation formula
nPr = n!
(n −
r)!
where; nPr = number of arrangements
n = number of items to be arranged
r = number of elements to be contained in the arrangement.
! = factorial e.g.; 4! = 4 × 3 × 2 × 1 = 24
In this example
3P3 = 3!
(3 −
3)!
= 3 × 2 × 1 = 6 ways
N ote that 0! = 1
ii) 3 electronic points are to be assembled in a plug-in unit for a TV set. The parts can be
assembled in any order. In how many different ways can the 3 parts be assembled?
All the objects (electronic parts) are to be assembled and therefore r=3. .
Solution
n =3, r =3
nPr = n! = 3!
(n −
r)! (3 − 3)!
=6
1
=6
Example 2
There are 10 numbers; 0 through to 9, which are to be used in code group of four to identify an
item of clothing e.g. code 1083 is to identify blue blouse, code 1030 is identify a pair of socks
and
so on. Repetition of numbers is not permitted i.e. same number can’t be used twice or more in a
total sequence e.g. 2256 or 2872
How many different code groups can be designed?
n = 10 r = 4
nPr = 5.040
Permutation allowing repetition
If repetitions are permitted the permutation formula nPr = nr e.g assume a 2-letter code where ‘a’
and ‘b’ are to be taken each at a time with repetitions such as aa allowed.
In this case n = 2, r = 2
The no. of permutations 2P2 = 22 = 4
111
STUD Y TEX T
Probability
Combination Rule
In determining the number of permutations of n-different things taken at a time, the order of the
items is important.
Combination determines the number of ways to choose r objects from group of n objects without
regard to order.
Combination Formula=
nCr = n!
(n −
r)!
Example
The sales department has been asked to design 42 colour codes for 42 different parts. A colour
code should consist of three colours and if 3 colours are used for one part, the same cannot be
used to identify a different part. Would 7 different colours be adequate to generate the 42 parts
colour codes?
Solution
n=7 r=3
7C3 = 35
35 colour codes would be inadequate. Thus 7 colours would not be adequate.
As an alternative to the use of three colour combination it has been suggested that only two
colours (r) be used for one colour code. Would 10 colours be adequate to colour code the 42
different parts?
n = 10 r = 2
10C2 = 45
This number would be adequate.
Probability Distributions
Statistical inference has the objective of making inferences or statement about a population
based on a sample selected from a population.
Probability distribution – Is a listing of all the outcomes of an experiment and the probability
associated with each outcome.
Example:
Suppose we are interested in determining the number of heads showing upon tossing a coin
three times.
11 2 quantitative techniques
STUD Y TEX T
Table 3.2: Enumerating the samples
Options 1st coin
toss
Second Third No. of
heads
Prob
1 T T T 0 1/8
2 T T H 1 1/8
3 T H T 1 1/8
4 T H H 2 1/8
5 H H H 3 1/8
6 H T H 2 1/8
7 H H T 2 1/8
8 H T T 1 1/8
x = outcome P(x) = Probability of outcome
0 1⁄8
1 2⁄8
2 3⁄8
3 1⁄8
Note:
Two important characteristics of probability distribution
The probability of a particular outcome must always be between zero and one.
The sum of probabilities of all mutually exclusive probabilities in a distribution is one
Exercise
Assuming you roll a die and observe the possible outcome of an experiment involving a six side
die;
Develop a probability distribution for this outcome
Portray the probability distribution graphically
Random variable
In any experiment of chance the outcomes occur randomly e.g. rolling a single dice is an
experiment in which any of the six possible outcomes can occur. Some experiments result in
outcomes that are quantitative such as shs, weight or number of children while others result in
outcomes that are qualitative e.g. colour, or religious preference.
A random variable is a quantity resulting from a random experiment that by chance can assume
different values, for example:
1. I f we count the number of students absent (the random variable) from class on a
Monday morning we can have 0, 1, 20, 30 etc.
2. I f we toss two coins and count the number of heads we could have 0,1 or 3 heads.
Since the exact number of heads resulting from this experiment is due to chance, it will
be called a random variable.
Other random variables might be the number of defective light bulbs produced during a week,
the
weight of members of DMS 201 regular class.
A random variable can either be discrete or continuous.
113
STUD Y TEX T
Probability
Discrete random variable
A discrete random variable can assume only a certain number of separate values of an interval;
for example, if there are 100 employees, the count of number absent on Friday can be 0, 2, and
100
In most cases a discrete random variable is usually a result of counting something.
Note: A discrete variable can assume fractional or decimal values in some cases. These values
must however be separated or countable.
Continuous random variable
A continuous random variable can assume any value in an interval. It is usually the result of
measuring something e.g. height of a person. It can assume one of an infinitely large number of
values within certain limitations.
E.g. the weight of a cloth can be 67, 67.2, 67.24kg, 67.241kg depending on the accuracy of
measuring instrument
Logically, if we organise a set of discrete random in a probability distribution the distribution is
called a discrete probability distribution.
How to determine mean, variance, standard deviation of probability distribution
The mean represents the central location of the data
The variance describes the spread in the data
In a similar way, a probability distribution is summarised by its mean and its variance
Mean (μ)
The mean is a typical value to represent a probability distribution. It is also the long-run average
value of the random variable. It is also referred to as expected value E(x).
E (x) can be computed as
E (x) = μ = Σx.P(x)
For a continuous variable
E (x) = μ = ∫ x.P(x)dx
Where P (x) is the probability of various outcomes of x
Variance and standard deviation
The mean does not describe the amount of spread (variation) of a distribution. The variance does
this.
s2 = Σ (x − μ)2P(x)
The standard deviation = square root of Variance
Example
Bill sells new cars for GM. Bill usually sells the largest number of cars on Saturday. He has
established the following probability distribution for the cars he expects to sell on a particular
Saturday.
∞
-∞
11 4 quantitative techniques
STUD Y TEX T
No. of
cars = x
Probability
P(x)
0 0.1
1 0.2
2 0.3
3 0.3
4 0.1
1. On a typical Saturday how many cars should Bill expect to sell?
2. What’s the variance and standard deviation of the distribution?
Solution
No. of
cars = x
Probability
P(x)
x.P(x) x- μ (x- μ)2 (x- μ)2 P(x)
0 0.1 0 -2.1 4.41 0.441
1 0.2 0.2 -1.1 1.21 0.242
2 0.3 0.6 -0.1 0.01 0.003
3 0.3 0.6 0.9 0.81 0.243
4 0.1 0.4 1.9 3.61 0.361
Total μ = 2.1 σ2 = 1.290
1) E(x) = Σx. P(x) = 2.1
2) Σ (x- μ)2 P(x) = 1.290
Standard deviation = √1.29 = 1.358 cars
How to interpret standard deviation
Suppose another salesperson Wendi also sold a mean of 2.1 cars on Saturday and standard
deviation of her sales was 1.9 cars. We would conclude that there is more variability of Saturday
sales of Wendi than Bill
Therefore Bill is doing better i.e he is more consistent and the number of cars he sells is not so
different from the mean as compared to Wendi.
Discrete Probability Distribution
There are 3 common Probability Distribution
Binomial Probability Distribution
Poisson Probability Distribution
Continuous Probability
Binomial probability distribution
This is an example of a discrete probability distribution. Here, there are only two possible
outcomes e.g. tossing a coin, the answer in a true /false question is either true or false. These
two outcomes are mutually exclusive.
115
STUD Y TEX T
Probability
More examples
A product is either classified as acceptable by the quality control department or not acceptable.
A sales call results in a customer either purchasing or not purchasing a product. Frequently we
classify the possible two outcomes as “success and failure” however, this classification does not
imply one outcome is good and the other is bad.
The information obtained in a binomial distribution is as a result of count i.e. we count the
number
of success in the total number Of trials e.g. if we flip a coin five times, we could count the
number
of times a head appears.
The probability of a success remains the same from one trail to another e.g. the probability that
you will give the first question of a true /false test correctly is one half (1/2). This is the 1st trial.
The probability that you will guess in the 2nd question (2nd trial) is also ½ and so on.
One trial in a binomial distribution is independent of another trial. In fact this is the same as
saying
that there is no rhythmic pattern with respect to the outcomes. As an example, the answers to the
true/false question are not arranged in any order e.g. FFFF TTTTTTTTT FFFF and so on
The number of trials is fixed at a certain value
Formula in binomial distribution
To construct a binomial Probability Distribution we must know
1. The number of trials
2. T he probability of success on each trial
P(r) = n! prqn − r
r!(n −
r)
= nCrprqn − r
Where n - Number of trials
r - Number of observed outcomes
p - Probability of success of each trial
q - Probability of failure = 1-p
Example
The answer to a true/false question is either true or false. Assume that;:
1. A n examination consists of four true/false questions
2. A student has no knowledge of subject matter. The chance (probability) that the student will
guess the correct answer to the first question is ½. Likewise the probability of guessing right
the remaining number of question is 0.5 for each question. What’s the probability that:
a) T he student will get none out of the four questions correct;
11 6 quantitative techniques
STUD Y TEX T
b) T he students will get exactly one out of the four correct?
n= 4
r=?
p = 0.5
q = 0.5
r=0
P(0) = 4!×0.5×0.54
0!(4 −
0)
= 0.0625
Poisson Probability Distribution
Fast Forward
The binomial probability distribution for probabilities of success P less than 0.05 can be
computed
but the calculations would be quite time consuming especially for a large n say 100 or more. The
distribution of probabilities would become more and more skewed as the probability of success
becomes smaller. Where the probability of success is very small and n is larger, we refer to it as
as Poisson probability distribution named after Simeon Poisson. The probability of a particular
event happening is quite small.
The poisson distribution is also a discrete probability distribution because it is formed by
counting
something. An important characteristic of such a distribution is that we can count only the
occurrences, we cannot count the non-occurrences. For example, we can only count the number
of telephone calls coming to a telephone exchange (switch board) each minute. We cannot count
the number of calls that we did not make. We can count accidents within a given time but we
cannot count the number of accidents that did not occur within a given period.
A poisson situation can be recognised by the following characteristics:
a) T he existence of events that
- occur randomly
- are rare
b) An interval of time, space or distance is defined within which events can occur
Assumptions
The probability of occurrence is the same everywhere in the interval
The probability of multiple occurrences at precisely same point is negligible
The occurrence rate is a random variance
Formula
p(x) = e−x|x
x!
Where x equals the arithmetic mean number of occurrences (success) in a particular interval. It’s
the specific value of the random variable.
e - is an exponential.
x - is the number of occurrences.
117
STUD Y TEX T
Probability
Continuous Probability Distribution
In continuous probability distribution, random variable can take any value within a given range
of
values by chance. The distribution of continuous random variable is characterised by a
probability
density function which represents all the possible outcomes in an experiment. The probability
density function can be a graph, an equation or a curve, which represents these possible values.
A probability of an experiment = 1.
Given a probability density function of a random variable x we can determine the probability that
x
has a value between two given points A and B. This probability is equal to the total area enclosed
by the curve of the function, two perpendicular lines erected at points A and B and the x-axis as
a proportion of the total area enclosed by the function curve and x-axis.
Daigram 3.3
y F (x)
ABx
There are several types of continuous probability distribution:
Normal distribution
The normal distribution curve is symmetrical and bell-shaped. Most observations in the
distribution are close to the mean with fewer observations further away. A normal distribution
can
be determined by the values of the mean and the standard deviation.
The normal curve is asymptotic to the x-axis i.e. its tails continuously approach the x-axis but
never quite touch it to infinity.
In a normal distribution the following observations are true:
68.26% of all observations lie within 1 standard deviation of the mean
95.44% of all observations lie within 2 standard deviations of the mean
99.73% of all observations lie within 3 standard deviations of the mean
Diagram 3.4
68.26%
95.44%
99.73%
-3δ -2δ -1δ μ δ 2δ 3δ
11 8 quantitative techniques
STUD Y TEX T
The normal distribution is a family of distribution e.g. standard normal distribution. Standard
normal distribution is a special case of the normal distribution. It has a mean of zero and a
standard deviation of 1 given any normal distribution. We may convert it to the standard normal
distribution by simply converting its μ to o and ó equal to 1. We convert each observation to the
standard normal variation as follows:
Z=c−m
σ
Where Z is the standardised random variable. Also known as standard normal deviate
orZ-score
x is the individual observed value of the random variable
μ is the mean
σ is standard deviation
By standardising any normally distributed random variable we can use the normal distribution to
estimate the area enclosed by the normal curve and true z-value. In the normal table, the values
of z only go to about z=3
If we have a list of observations of a normaly distributed random variable, what proportion of the
observation will be expected to be?
Example
If x is a continuous random variable with a mean of 50 and standard deviation of 2, what is the
probability that an observation picked at random is
1. Less than 52
2. Less than 46
3. Between 44 and 48
Solution
1. μ = 50
σ= 2
when x = 52
z = x-μ
σ
= 52 – 50
2
Calculate z = 1
= 0.3413
= 0.5 + 0.3413
= 0.8413
119
STUD Y TEX T
Probability
Less than 46
x = 46
x = 46-50 = 2 -4/2 = -2
Calculate 2 = -2
From table 2 = 0.4772
Less than 46
= 0.5 -0.4772
= 0.0228
2. Between 44 and 48
z = 44-50
2
=6/2 = -3
z = 0.49865
z = 48-50
2
= 2/2 = 1
z = 0.3413
3. Between 44 and 48
0.49865 – 0.3413 = 0.15735
Chapter Summary
A set of events is said to be mutually exclusive if the occurance of any one of the events
precludes
the occurrence of any of the other events e.g. when tossing a coin, the events are a head or a tail
these are said to be mutually exclusive since the occurrence of heads for instance implies that
tails cannot and has not occurred.
Binomial probability distribution is a set of probabilities for discrete events. Discrete events are
those
whose results or outcomes can be counted. Binomial probabilities are commonly encountered in
business situations e.g. in quality control activities when determining the probability of having a
certain number of defective items in a given consignment.
General rules of Multiplication – It states that of two events, A and B, the joint probability that
both events will happen is found by multiplying the probability that A will happen by the
conditional
probability of event B happening
P(A and B) = P(A) x P(B/A)
Where P(B/A) stand for probability that B will occur given that A has already occurred.
1 2 0 quantitative techniques
STUD Y TEX T
Chapter Quiz
1. T he mean does not describe the amount of spread (variation) of a distribution. The
………… does this.
2. T he probability of multiple occurrences at precisely same point is negligible
T rue
F alse
3. The ………………. curve is symmetrical and bell-shaped.
4. Which distribution is represented by the formula
n! prqn − r
r!(n −
r)
121
STUD Y TEX T
Probability
Answers to Chapter Quiz
1. Variance
2. True
3. Normal distribution
4. Binomial distribution
Questions from previous exams
June 2000: Question 7
i) Define probability as used in Quantitative Techniques. (3 marks)
ii) What is Bayes Theorem? Explain how Bayes Theorem can be utilised practically
(5 marks)
iii) KK accounting firm has noticed that of the companies it audits, 85% show no inventory
shortages, 10% show small inventory shortages and 5% show large inventory
shortages. KK firm has devised a new accounting test for which it believes the following
probabilities hold:
P (company will pass test/no shortage) = 0.90
P (company will pass test/small shortage) = 0.50
P (company will pass test/large shortage) = 0.20
Required:
(i) Determine the probability if a company being audited fails this test has large or small
inventory shortage. (7 marks)
(ii) If a company being audited passes this test, what is the probability of no inventory
shortage? (5 marks)
(Total: 20 marks)
December 2001 Question 4
a) The past records of Salama Industries indicate that about 4 out of 10 of the company’s
orders are for export. Further, their records indicate that 48 per cent of all orders are for
export in one particular financial quarter. They expect to satisfy about 80 orders in the
next financial quarter.
Required:
(i) Determine the probability that they will break their previous export record. (7 marks)
(ii) E xplain why you have used the approach you have chosen to solve part (i) above
(2 marks)
1 2 2 quantitative techniques
STUD Y TEX T
b) Gear Tyre Company has just developed a new steel-belted radial tyre that will be sold
through a national chain of discount stores. Because the tyre is a new product, the
company’s management believes that the mileage guarantee offered with the tyre will be
an important factor in the consumer acceptance of the product. Before finalising the tyre
mileage guarantee policy, the actual road test with the tyres shows that the mean tyre
mileage is 36,500 kilometres and the standard deviation is 5,000 kilometres. In addition,
the data collected indicate that a normal distribution is a reasonable assumption.
Required:
(i) Gear Tyre Company will distribute the tyres if 20 per cent of the tyres manufactured can
be expected to last more than 40,000 kilometres. Should the company distribute the
tyres? (4 marks)
(ii) The company will provide a discount on a new set of tyres if the mileage on the original
tyres does not exceed the mileage stated on the guarantee.
What should the guarantee mileage be if the company wants no more that 10% of the
tyres to be eligible for the discount. (4 marks)
c) Explain briefly some of the advantages of the standard normal distribution. (3 marks)
(Total: 20 marks)
June 2002: Question 3
a) State clearly what is meant by two events being statistically independent. (2 marks)
b) In a certain factory which employs 500 men, 20% of all employees have a minor accident
in a given year. Of these, 30% had safety instructions whereas 80% of all employees
had no safety instructions.
Required:
Find the probability of an employee being accident-free given that he had:
(i) N o safety instructions (5 marks)
(ii) S afety instructions (5 marks)
c) An electric utility company has found that the weekly number of occurrences of lightning
striking the transformers is a Poisson distribution with mean 0.4.
Required:
(i) The probability that no transformer will be struck in a week. (3 marks)
(ii) T he probability that at most two transformers will be struck in a week. (5 marks)
(Total: 20 marks)
June 2003: Question 7
a) The J.R Muchemi Computer Company is considering a plant expansion that will enable
the company to begin production of a new computer product. The company’s executive
director must determine whether to make the expansion a medium or large scale project.
Uncertainly exists in the demand for the new product, which for planning purposes
may be low, medium or high demand. The probability estimates for demand are 0.20,
0.50 and 0.30 respectively. The firm’s accountants have developed the following annual
profit (in thousands of shillings) forecast for the medium and large scale expansion
projects:
123
STUD Y TEX T
Probability
Medium
scale profit
Expansion
profitability
Large
scale profit
Expansion
profitability
Low 50 0.20 0 0.20
Demand Medium 150 0.50 100 0.50
High 200 0.30 300 0.30
Required:
(i) Which decision is preferred for the objective of maximising the expected profit
(4 marks)
(ii) Which decision is preferred for the objective of minimising the risk or uncertainty?
(4 marks)
From your answers in (i) and (ii) above, should the company go for medium scale expansion or
the large scale expansion? Explain. (2 marks)
b) A manufacturing firm based in Nairobi Kenya receives shipment of parts from two
different suppliers from UK and Japan. Currently, 65% of the parts purchased by the
company are from suppliers 1 (UK) and the remaining 35% are from supplier 2 (Japan).
The quality of the purchased parts varies with the source of supply. Historical data
suggest that the quality rating of the two suppliers are as shown below:
Percent of good parts Percent of bad
parts
Supplier 1 98 2
Supplier 2 95 5
The parts from the two suppliers are used in the firm’s manufacturing process and during the
processing, a machine breaks down because it attempts to process a bad part.
Required:
(i) Given the information that a part is bad, determine the probability that it came from
Supplier 1 and it came from Supplier 2. (5 marks)
(ii) S how the above information in a probability tree. (5 marks)
(Total: 20 marks)
December 2004: Question 7
a) Indicate which of the following statements you agree with and which you disagree with and
defend your opinion.
(i) When performing a Bayesian decision analysis, the prior probabilities must inevitably
be subjective probabilities (3 marks)
(ii) The Bayes action will be the same regardless of whether it is selected using expected
monetary pay-offs or expected utilities. (3 marks)
(iii) For a given decision situation, the maximin criterion, on one hand, and the maximax
criterion on, on the other, will always point towards different actions.
(3 marks)
1 2 4 quantitative techniques
STUD Y TEX T
c) In a flow-chart diagram, show the steps involved in a standard posterior analysis.
(6 marks)
With reference to probability theory, briefly but clearly, explain the statement “there is only
one thing certain and that is that nothing is certain” (5 marks)
(Total: 20 marks)
125
STUD Y TEX T
Probability
1 2 6 quantitative techniques
STUD Y TEX T
127
STUD Y TEX T
Sampling and Estimation
CHAPTER ONE
Sampling and
Estimation
STUSTUD YD YTE TEX TX T
FOUR
1 2 8 quantitative techniques
STUD Y TEX T
129
STUD Y TEX T
Sampling and Estimation
CHAPTER FOUR
Sampling and Estimation
Objectives
At the end of this chapter, you should be able to:
•• Explain why in many situations a sample is the only feasible way to learn something
about a population.
•• Explain the various methods of selecting a sample.
•• Distinguish between probability sampling and non-probability sampling.
•• Define and construct a sampling distribution of sample means.
•• Explain the Central Limit theorem and its importance in statistical inference.
•• Calculate confidence intervals for means and proportions.
•• Determine how large a sample should be for both means and proportions.
Fast Forward: Sampling is that part of statistical practice concerned with the selection of
individual
observations intended to yield some knowledge about a population of concern, especially for the
purposes of statistical inference.
Introduction
A population consists of all the items with which a particular study is concerned. A sample is a
much smaller number chosen from this population. The sample must be chosen randomly. The
data collected in the sample is used to draw inferences about the corresponding population
parameter.
The three types of distributions:
1. Population distribution
Population distribution is the distribution of the individual values of population. Its mean is
denoted by “μ”.
2. S ample distribution
I t is the distribution of the individual values of a single sample. Its mean is generally written
as ‘m.’ It is extremely unlikely that it will be the same as “μ”.
3 Distribution of sample means
A sample of size n is taken from the parent population and mean of the sample is calculated.
This is repeated for a number of samples so that a population of sample means is obtained.
This population approaches a normal distribution as n increases and is the distribution of
sample means.
Definition of key terms
1. Estimate – an approximate calculation of quantity or degree or worth; an estimate of what
it would