N. 1 St.

Poisson Distribution
Introduction
In Binomial Distribution, n' is finite. So it is
number of times an event occurs and the number possible to count the
of times it does not
occur. Suppose an event is rare. For example, number of
road. Then it may be possible to count the accidents in a
occurs bu not possible to count the number ofnumber of times an event
times it does not occur.
This is because total number of trials of the
known. Poisson distribution isa distributionexperiment is not precisely
suitable to such situations.
Pefsson Distribution as a limiting case of Binomial
Poisson distribution may be obtained as alimiting Distribution
case of Binomial
distribution under the following conditions:
i) the number of trials is very large n ’ o )
Ai) the probability of success for each(ie,trial is very small (ie, p ’0)
Aii) 'np' is finite (say, 'm')
Therefore Poisson Distribution can be stated in the following
Let the probability for success in a single trial of a random manner.
ment is very small (ie p ’ 0) and the experiment be experi
repeated larger
number of times (ien’ o). Then the probability for x successes out of
these 'n' trials is given by P (x) = . This law is called Poisson
Law. X!

Meaning and Definition

A discrete random variable 'X' is said to follow Poisson
Distribu
tion if its probability density function is P(x) = e mm*
x! for xassuming
values 0, 1,2, ...
Poisson distribution may be expected in cases where the chances of
happening of any individual event is small. So the distribution is used to
describe the behaviour of rare events. Therefore it is called the 'law of.
improbable events'.
Putting x=0, 1, 2, 3, ....in the probability function,
e"m
P (x)= x! , we obtain the probabilities for 0, 1, 2, 3, . .
times the event to happen.
 N. 2

of the Poisson Distribution can

If we know 'm, allthe
terms
Poisson Distribution It is
be
obtained. 'm' is the parameter of
the the
mean of the poisson distribution.
Distribution
Uses (or importance) of Poisson used to explain the behaviour of the
Poisson Distribution can be probability of occurrence of
wherethe the
discrete
event is very smallvariables
random andIthetotal number of possible cases is sufficiently
applications in Queuino
large. As such, Poisson distribution has found Insurance, Physicg
Theory (waiting line problems) and in the fields of
Biology, Economics, Industry etc.
Practical situations where Poisson Distribution can be used
(1) to count the number of telephone calls arriving at atelephone switch
board in unit time (say, per minute)
(2) to count the number of customers arriving at the super market (say
per hour)
(8) tocount the number of defects per unit ofa manufactured product,
(in Statistical Quality Control)
(4) to count the number of radio-active disintegrations of a radio-ac
tive element per unit of time (in Physics)
(5) to count the number of bacterias per unit (in Biology)
(6) to count the number of defective materials say, pins,
blades etc., (in
a packing of manufactured goods by a
(7) to count the number of casualties concern)
ease such as heart attack in a year. (persons dying) due to a rare dis
(8) to count the number of
road.
accidents taking place in a day on a busy
Characteristics
1. Poisson
(or properties) of Poisson
Distribution
2 If'x Distribution is a discrete probability distribution.
follows aPoissonDistribution, then 'x' takes values 0, 1,2,...
to infinity
3 Poisson Distribution has a single
all the terms can be found out, parameter, m. When 'm' is known
4 Mean and variance of Poisson
5 Distribution are equal to m.
Poisson Distribution is a positively
Note 1:
skewed distribution.
x! =x (x-1)
(%-2)......1'
Eg: 3! =3x2x 1=6
Note 2:
em= Antilog (m xlog e) where loge =.4343
 0.1 St.

NormalDistribution
Introdution
As Binominal and Poisson distributions are the most useful theo
retical distributions of discrete random
the most useful theoritical distribution forvariables, Normal distributior. is
distribution is perhaps the most important continuous variables. Normal
and useful distribution in Sta
tistics. Many stastical data
can be displyed in the formconcerning business and economic
of Normal distribution. It is now problems
portant probability model in statistical analysis. most im
Normal distribution was first discovered by De-Moivre in
limiting form ofthe Binomial distribution. 1733 as
Fundamental importance of the Normal
lations seem to follow approximately a pattemdistribution is that may popu
of distribution as described
by Normal distribution. Numerous
phenomena such as the age
tion of a given species of animals, height of adult persons, the distribu
test scores of sch0ol children and many other intelligent
ered to be normaly distributed.
distributions are consid
It can be seen that quantities whose variation
depends on
causes is distributed acording to the Normal distribution. Normalrandom distri
bution is related to the distribution of errors, made by chance in experi
mental measurement.
Consider the following examples: Acompany manufactures screws
of a particular diameter. There may be som variations in the
measure
ment. But most of the screws have diameter very near to what is re
quired. Few of them are either much larger or much smaller. Ifwe draw
a histogram of the frequency distribution of measurements, we see that
the histogram has maximum height at the centre and the hight decreases
on either side in almost same rate. So if the class interval of the fre
quency distribution is made smaller and smaller we get a curve of the
shape of a bell. It is the normal curve.
 0.2

Definition follow
A continous random variable Xis said to Normal·
el1 distribu
thon if its probability function is f(%) = where and
standard
constants (4 being the mean obeing the
6 are
The variable X varies between -o and+ o. deviation of x),
Normal distribution as aImiting case of Binominal distribution
Binominal Distribution is an important theoretical distribution e.
dicrete variables. Binominal distribution tends to Normal
under the following conditions.
1. Number of trials (n) is very large
Distribution
2. pand q (ie probability for success in asingle trail and the probahil
ity for its failure) are almost equal.
Then the Binominal distribution can be approximated to normal
Probability density function and parameter of the normal distribution.
IfX is a continous random variable following Normaldistribution
with mean and standard deviation o then its probability density
funcution is given by
f(x)= where z and 'e' are the constants given by
22
7 =3.14159 and e 2.71828. Xmay very between -o and to.
Then mean 4 and standard deviation o are called
the Normal distribution. parameters of
ANormal distribution is described by its
dard deviation. For various values of mean andparameters, mean and stan
get different nomal curves. standard deviation, we
Mrean of the distribution shows the location of the
X-axis and S.D indicates the spread (or shpe) curve along the
of the curve.
Properties (eharacteristics) of the Normal Distribution
1. The normal curve is a'
2. The normal curve is bellcontinous curve.
3: Normal curve is shaped.
4. Mean, Median and symmetric about the mean.
S The hight of Mode are eqùual for a normal
distribution.
Normal curve is at its maximum at the
6. There is only one mean.
7. The ordinate at maximum point, which occurs at the mean.
mean divides the whole area into two equal
 0.3

parts.(ie 0.5 on either side)

8. The curve is asymptotic to the base on either side.ie. the
curve
approaches nearer and nearer to the base,but it never touches.
9. coefficient of skewness is 0 (B, -0)
10. The nomal curve is unimodel, ie, it has only one
|1. The points of inflexion occur atuto. Atthe pointmode.
of inflexion
the curve changes from concavity to convexity.
12. Q, and Q, are equidistant from median.
13. Mcan deviation for Normal distribution is and QD is
14. Allodd moments of the Normal distribution are zero and even
moments are interrelated, by the following fomula.
Hn =(2n -1) o»#n-2.
15. Normal distribution is mesokutic. That is measure of kurtosis
-B, =3
16. Noportion of the curve lies below the x axis,
17. Theoretically the range of thenormal curve is -o to +o. But
practically the range is x 3 oto x+3o.
18. Ifx and yare two independent Normal variates, then their sum
is also a normal variate. This is called the additive property.
19. Area under the normal curve is distributed as follows.
to covers 68.27% area
t2s covers 95.45% area area property
t3o covers 98.73% area.
Importance (or use) of Normal Distribution
The study of the normal distribution is of central importance in sta
tistical analysis because of the following reasons.
1. Most of the discrete probabilty distribútions (eg: Binominal
distribution, Poisson distribution etc) tend to normal distribution as 'n'
becomes large.
2. Almost allsampling distributions such as student's t- distribu
tion, F-distribution, Z-distribution, x2- distribution etc. conform to the
normal distribution for large values ofn.
3. The various tests of significance like t -test, F-test etc are based
on the assumption that the parent population from which the samples
have been drawn follows Normal Distribution.
4. It is extensively used in large sampling theory to find the esti
mates of parameters from statistics, confidance limits etc.
5. Normal distribution has the remarkable property stated in the
central limit theorem. As per the theorem, when the sample size is in
creased, th¹ smple means will tend to be normally distributed. Central
 0.4

limit theorem gives the normal distribution its central place in:the theory
importan1t problems can be.solved by this.
of sampling, since many applied works many
6. In theoritical statistics as well as
can be7.solved only under the assumption of a problems
normal population.
Thee normal distribution has numerous mathermatical properites
which make itt popular and comparatively easy to manupulate. The nor
distributions.
mal curve is reasonably close to many and
in statistical quality control
8. It finds applications industrial
experiments. Many distributions in Social and Economic data are appprO-
ximately normal. Eg: birth, death etc.are normally distributed. In Psy-
chological and Educational data many distributions are of normal type.

Merits and Demerits of Normal distribution

Merits
1 Normal distribution is the mostly used distribution in Inferential
Statistics.
2. Most of error ofmeasurements and a large variety of physical ob.
3
servations have approximiately Normal distributions.
The measurements of linear dimensions of large number of articles
produced may show individual variations. These follow Normal
distribution.
4. The standard normal distribution table shows exhaustively areas
for the different intervals of the values of the variable.
5. The Normal distribution has a number of
mathematical properties.
6. Most of the distributions in nature are either
normal or that can be
approximated to normal.
Demerits
1. The variables which are not
continous cannot be normally distrib
uted. Therefore many distributions in
of number of children per family, Economics like distribution
cannot be studied under Normal
Distribution.
2. The Normal
distribution cannot be applied to situations where the
distribution is highly
is very much skewed.skewed. For example:distribution on income
be appropriate. Therefore here Normal distribution will not
Standard normal variate (Unit normal variate)
IfX is arandom varible following
and standard deviation o then the
Normaldistribution with mean
niard normal variate.
variable Z= is known as sta
This Z' follows normal
disrtibution with mean' 0 and S.D=
 0.5

The distribution of Z is known as Standard Normal Distribution.

The probability function of Zis f(a)-Tfor -on <z< 0

The standard normal distribution table
This is atable showing the probabilty for 'z' taking valuse between
0and a given value. The probability thus obtained is the area of the
standard nomalcurve between the ordinates at z=0 and at the given value.
For example,when z =1.17
table value =0.3790
This table value is the area
between 0 and 1.17 which can be
1.17
written as P(0<z< 1.17) =0.3790
It may be noted that for both negative ).379
and positive values of z, area is the
same. That is, P(0 <z<1.17)
-P(-1.17 <z<0)=.3790 -1.17 0

EL. 1: Find (1) P (-1.78 <z<1.78) (2) P(1.52 <z< 2.01)

(3) P1.52<z<-75) (4) P(2> 1.8) (5) P(z<-1.5) 0.4625s
Ans: 0.4625
(1) P(1.78<z< 1.78)
-P(-1,78<1<)+P(0<|<1.78)
-0.4625 +0,4625 = 0.925 -1.78 0. 1.78.
(Note: Since he values are on either side of the central line, add the
table values) 0.4357 0.0421
(2) P(1.52<z<2.01)
= P(0<z<2.01) -P(0 <z< 1.52)
0.4778-0.4357 = 0.0421 0 1.52 2.01
(Note: Since both the values arè on the same side of the central line,
fake the diference of the table values) 0.1623
0.2734
(3) P-1.52 <z<-0.75)
=P(-1.52<z<0)-P (-0.75 <z <0)
=0,4357 -0,2734 =0.1623
-1.52 -7s 0
(since both the 'Z'values are on the same side, diference of the area is
taken) 0.4641
(4) P(z> 1.8)
=.5-P (0 <z<1.8)
=0.5000 --0.4641=0.0359
0 1.8