PROBABILITY AND STATISTICS

UNIT -1
Trial : Conducting an experiment only one time is called trial.
Ex: Tossing of coin, Rolling of dice, we draw one card from a pack of cards, etc.,

OUTCOME: The result of a trial in a experiment is called an outcome.

EVENTS : The outcomes of an experiment are known as events.

RANDOM EXPERIMENT : An experiment whose result is not determined by us but determined by

randomly and naturally is called random experiment.

SAMPLE SPACE: A set of all possible outcomes of an experiment is called sample space. Or The
total no. of favourable and un favourable events is also called sample space. It is denoted by S or Ω

EXHAUSTIVE EVENTS: The total no of possible outcomes in any trial is know as exhaustive events
or exhaustive cases. It is also called sample space.
Ex: i) In tossing of coin there are two exhaustive events i.e., =2 viz., head (H) and tail (T)
ii) In throwing of a die, there are six exhaustive events i.e., =6viz., 1,2,3,4,5,6 since any one of the
six faces may come upper most
iii) In drawing two cards from a pack of cards the exhaustive number of cases is ways.
In general in tossing of n coins the exhaustive events is i.e., , in throwing of n dice the exhaustive
events is i.e.,

FAVOURABLE EVENTS :The no. of events are favourable for happening of an event is said to be
favourable events.
Ex: In tossing of two coins at a time the no. of events favourable to getting the exactly one head
is 2 i.e., { HT,TH } or at least one head is 3 i.e., { HH,HT,TH }

EQUALLY LIKELY EVENTS : Events are said to be equally likely events, if happening of one event
has equal chance to happening of other event.
Ex: In tossing of a coin Head or Tail are equally likely events ii) In throwing of a die all the six
faces are equally likely events.

MUTUALLY EXCLUSIVE EVENTS: Events are said to be mutually exclusive events, if happening of
one event exclude the happening of other event.
Ex: In tossing of a coin Head and Tail are mutually exclusive events ii) In throwing of a die all
the six faces numbered 1 to 6 are mutually exclusive events.

INDEPENDENT EVENTS: Events are said to be independent if the happening of an event is not
affected by the happening of another event.
Ex: To getting a head on the coin is independent of obtaining a five on the die.
Let A and B are any two independent events then P(A B) = P(A). P(B)

DEFINITIONS OF PROBABILITY: There are three definitions of probability

I) MATHEMATICAL OR CLASSICAL OR A PRIORI DEFINITION OF PROBABILITY : If a trial result
in ‘n’ exhaustive ,equally likely and mutually exclusive events and ‘m’ events are favourable to the
happening of an event E, then the probability of happening of an event E is denoted by P(E) .
It is defined as P(E) = i.e., P(E) =
LIMITATIONS : i) If the various outcomes of the trial are not equally likely or equally probable
ii) If the exhaustive no. of cases in a trial is infinite.
II) STATISTICAL OR EMPIRICAL OR VON MISS DEFINITION OF PROBABILITY: If a trial is
repeated a no. of times under essentially homogeneous and identical conditions then the limiting
value of the ratio of the no. of times the event happens as the no. of trials, become indefinitely large,
is called statistical definition of probability. Symbolically, if in ‘n’ trials an event E happens ‘m’ times
lim
then the probability of the happening of E is given by P(E) =
LIMITATIONS : i) If an experiment is repeated a large number of times, the experiment conditions
may not remain identical and homogeneous.

III) THE AXIOMS OF PROBABILITY: The axioms of probability are

i) The probability value always must be positive i.e., P(E) ≥ 0
ii) The probability value always lies in between 0 to 1 i.e., 0 ≤ P(E) ≤ 1
iii) The total probability equal to one i.e., P(S) = 1 or P(Ω) = 1 or P E = 1
iv) If E 1, E2, … En are n mutually exclusive events the P(E 1 U E2,U … U E n) = P(E 1) + P(E 2) + … + P(E n)

CONDTIONAL PROBABILITY: If A, B are any two events then the conditional probability of
occurrence of A when the event B has already happened is called conditional probability A. It is
P
denoted by P(A/B) and this is defined as P(A/B) = , P( ) ≠0 Similarly, The conditional
P
probability of occurrence of B when the event A has already happened is called conditional
P
probability B. It is denoted by P(B/A) and this is defined as P(B/A) = , when P ≠ 0.
P

ADDITION THEOREM ON PROBABILITY : If A, B and C are any three events then

I. (i) P(A U B) = P(A)+ P(B) - P(A B)
(ii) P(A U B UC) = P(A)+ P(B)+P(C) - P(A B)- P(B C)- P(A C)+ P(A B C)
II. If A, B and C are any three mutually exclusive events then
(i) P(A U B) = P(A)+ P(B)
(ii) P(A U B UC) = P(A)+ P(B)+P(C)

MULTIPLICATION THEOREM ON PROBABILITY : If A, B and C are any three events then

I i) P = P . P(A/B) or
P = P . P(B/A)
ii) P = P . P(B/A) . P(C/A B)
II If A, B and C are any three independent events then
i) P = P(A) . P(B)
ii) P = P(A) . P(B) . P(C)

BAYES THEOREM : Suppose E1, E2, …., En are mutually exclusive events of a sample space S such that
, i = 1,2,.., and ‘A’ is any arbitrary event of S such that P(A) > 0, and A U E i = S then the
conditional probability of Ei given A is denoted by P(Ei/A) and it can be defined as
P(Ei/A) =
Where P(E1), P(E2), …., P(En) are termed as the a priori probabilities
, i = 1,2,..,n are called posterior probabilities
 UNIT - 2
LARGE SAMPLE TESTS : If the sample values (n ≥ 30) then it is called large samples. The tests
belongs to small samples is called large sample tests.

Test (1): Test for significance difference between sample mean and
population mean
(i)If are n sample values are drawn from a normal population with mean and variance
then we calculate sample mean and sample S.D (s)in the following way.

Sample Mean , Sample Standard Deviation(s)= , : No. of sample values

(ii) For testing the significant difference between sample mean and population mean we formulate the
null hypothesis Ho as
Ho : There is no significant difference between sample mean and population mean
i.e Ho : Here = sample mean, = Population mean (iii)
To test the above hypothesis the suitable test statistics ‘ Z ’ as

= → = ∼ as by C LT

(iv) We substitute different values in the above test we get calculated value of
we take table value of at α % level of significance (los) and compare this value calculated value of
(a) If calculated value of table value then we reject the null hypothesis.
i.e ., There is significant difference between sample mean and population mean.

(b) If calculated value of < table value then we do not reject(accept) the null hypothesis.
i.e., There is no significance difference between sample mean and population mean
Conclusion can be drawn according to rejection or non-rejection of the null hypothesis.

(v) The 100(1-α)% Confidence limits for un known parameter is

Statistic ± ( Standard Error) = ±

Test (2): Test for significance difference between two sample means
Case I: (i)Let be the sample values are drawn from normal population with mean
and variance and be the sample values are drawn from another normal
population with mean and variance . By using sample values we calculate sample means
, sample S.Ds , in the following way .

, ,

(ii) For testing the significant difference between two sample means we formulate the null hypothesis as
Ho: There is no significant difference between two sample means

i.e Ho : , Here = First sample mean, = Second sample mean

(iii) To test the above hypothesis the suitable test statistics ‘ z ’ as

= ∼ as by CLT , if known population S.D( )

Or

= ∼ as by CLT , if un known population S.D ( )

Rest of the procedure as usual

(v) The 100(1-α)% Confidence limits for un known parameter - is

Statistic ± ( Standard Error) = ±

Case II : If two sample values are drawn from Single population then for testing the significant
difference between two sample means we use the following test statistic as

= ∼ as by CLT , if known population S.D ( )

Or
= ∼ as by CLT , if un known population S.D( )

Where

Test (3): Test for significance difference between sample proportion and
population proportion. (i)
If be n sample values are drawn from a binomial population with mean and
variance . Let we denote the special no of characters in the sample is denoted with . with the
help of sample values we find sample proportion by using the following formula
Sample Proportion ( ) , i.e.,
(ii) For testing significant difference between sample proportion and population proportion we
formulate the null hypothesis Ho as
Ho: There is no significant difference between sample proportion and population proportion
i.e Ho : = Here = sample proportion, = Population Proportion
(iii) To test the above hypothesis the suitable test statistics ‘ z ’ as

= → = ∼ as by C LT

(v) The 100(1-α)% Confidence limits for un known parameter is ±

Test(4): Test for significance difference between two sample proportions
Case I: (i) Let be the sample values are drawn from binomial population with mean
and variance and be the sample values are drawn from another
binomial population with mean and variance .
By using sample values we calculate sample proportions in the following way .

(ii)For testing the significant difference between two sample proportions. We formulate the null
hypothesis as Ho: There is no significant difference between two sample proportions
i.e Ho : , Here = First sample proportion , = Second sample proportion

(iii) To test the above hypothesis the suitable test statistics ‘ z ’ as

= ∼ as by CLT , if known population proportion (P)

Or

= ∼ as by CLT , if un known population proportion( )

Rest of the procedure as a usual

(v) The 100(1-α)% Confidence limits for un known parameter - is

Statistic ± ( Standard Error) = ±

Case II : If two sample values are drawn from Single population then for testing the significant
difference between two sample proportions we use the following test statistic as
= ∼ as by CLT , if known population proportion(P)

Or
= as by CLT , if un known population proportion(P)

Where
 UNIT - 3

Small Sample Tests : If the sample values (n < 30) then it is called small samples. The tests
belongs to small samples is called small sample tests.

Small sample tests for Means

Test (1): Test for significance difference between sample mean and
population mean
If are n sample values drawn from a normal population with mean and variance then
we calculate sample mean and sample S.D (s)in the following way Sample

Mean , Sample Standard Deviation(s)= , : No. of sample values

For testing significant difference between sample mean and population mean we formulate the null
hypothesis Ho as
Ho : There is no significant difference between sample mean and population mean
i.e Ho : Here = sample mean, = Population mean
To test the above hypothesis the suitable test statistics ‘ t ’ as

= → = ∼

We substitute different values in the above test we get calculated value of and compare this value
with table value of at α % level of significance (los) with

Rest of the procedure as a usual

Test (2): Test for significant difference between two sample means for
small samples
If two independent sample values and of sizes n1 and n2 have been
drawn from two normal populations with means and . By using sample values we calculate
sample means , sample variances , in the following way .

, , ,

For testing the significant difference between two sample means we formulate the null hypothesis as
Ho: There is no significant difference between two sample means
i.e Ho : , Here = First sample mean, = Second sample mean

To test the above hypothesis the suitable test statistics ‘ t ’ as

= ∼

Where

We substitute different values in the above test we get calculated value of and compare this value
with table value of at α % level of significance (los) with

Rest of the procedure as a usual

Paired - test for difference of means (or) Correlated variables
Let be the paired sample values are drawn from a normal population. we
consider when the sample size are equal i.e . The problem is to test the sample means
differ significantly or not.

For example we want to test the performance of the students before giving the coaching and after giving
the coaching , the performance of the patients before giving the drug and after giving the drug. Let
be the readings in hours of sleep, on the i th individual, before and after the drug
is given respectively. For testing the significant difference between the means of paired sample values or
correlated samples , we formulate the null hypothesis as follows.
Ho : The drug is not significant.
To test the above hypothesis we use the following test statistic as

= ∼

Where = , ,

We substitute different values in the above test we get calculated value of and compare this value
with table value of at α % level of significance (los) with

Rest of the procedure as a usual

F-TEST Or Testing the significant difference between the two sample variances.
If two independent sample values and of sizes n1 and n2have been drawn
from two normal populations with same variances .

With the help of sample values we calculate sample variances by using the following formulae.

First sample variance , Second sample variance ,

For testing the significant difference between two sample variances, we formulate the null hypothesis as
Ho : There is no significant difference between two sample variances or population variances
i.e Ho : =

To test the above hypothesis we use the following test statistic F as

∼ if >

Or

∼ if >

We substitute different values in the above test we get calculated value of F and compare this
value with table value of F at α % level of significance (los) with

Rest of the procedure as usual

 TEST (Or) Testing the significant difference between the sample variance and
population variance :
If are sample values drawn from a normal population with a specified variance . With
the help of sample values we calculate sample variance by using the following formula.

Sample Variance

For testing significant difference between sample variance and population variance we formulate
the null hypothesis Ho as
Ho : There is no significant difference between sample variance( ) and population variance( )

To test the hypothesis the test statistic is = ∼

We substitute different values in the above test we get calculated value of and compare this value
with table value of at α % level of significance (los) with

Rest of the procedure as usual

Test for goodness of fit

Let (i=1,2,….,n) be the observed(given) frequencies and (i=1,2,….,n) be the corresponding expected
frequencies then to test the significant difference between observed and expected frequencies, we
formulate the null hypothesis Ho as follows

Ho: There no significant difference between observed and expected frequencies

To test the above hypothesis the test statistic as , = ∼

We substitute different values in the above test we get calculated value of

we take table value of at α % level of significance (los) with

Rest of the procedure as usual

Test for independence of attributes.

In the given table cell frequency are known as observed frequency i.e . From this observed
frequencies we calculate expected frequencies i.e ( Eij ) by using the following formula

Expected Frequency Eij = i.e Eij =

For testing independence of the attributes or Testing the significant difference between observed and
expected frequencies , we formulate the null hypothesis as follows.

Ho: The two attributes are independent or

Ho: There is no significant difference between observed and expected frequencies.
To test the above hypothesis to be true the suitable test statistics as

= ∼ , r = No. of rows, c = No. of columns

If we substitute different values in the above formula then we get calculated value of and compare
this value with table value of at α % level of significance (los) with