Module - 5 PROB

Testing of Hypothesis
Module - 5 1 / 31
Notations
Population parameters
Population mean (µ)
Population standard deviation (σ)
Population size (N)
Population proportion (P)
Sample statistic
Sample mean (x̄)
Sample standard deviation (s)
Sample size (n)
Sample proportion (p)
Module - 5 2 / 31
Hypothesis Testing
Statistical hypothesis
is a statement or a claim about one or more population parameters
An hypothesis to be tested or to answer a question - Alternative

hypothesis (H1 or Ha )
An hypothesis of no difference, opposes H1 - Null hypothesis (H0 )
Example
Suppose we test for population mean. Then
Null Hypothesis H0 : µ = µ0
Alt Hypothesis H1 : µ 6= µ0 or µ > µ0 or µ < µ0
If µ 6= µ0 , then the test is called Two-tailed.
If µ > µ0 , then it is called Right tailed test.
If µ < µ0 , then it is called Left tailed test.
Module - 5 3 / 31
Types of errors
H0 is true H0 is false
Reject H0 Type I error Correct decision
Accept H0 Correct decision Type II error
P(Type I error) = α = level of significance (LOS)

P(Type II error) = β
Module - 5 4 / 31
Steps involved in Hypothesis testing
Formulate Null and Alternate hypothesis

Identify the level of significance
Test statistic
Critical (or) Rejection region
Conclusion
If n ≥ 30 =⇒ Large sample
If n < 30 =⇒ Small sample
Module - 5 5 / 31
Table corresponding to critical values
Test LOS 1% (0.01) 5% (0.05) 10% (0.1)

Two tailed |Z α2 | = 2.58 |Z α2 | = 1.96 |Z α2 | = 1.645
Right tailed Zα = 2.33 Zα = 1.645 Zα = 1.28
Left tailed Zα = −2.33 Zα = −1.645 Zα = −1.28
Module - 5 6 / 31
Test for single proportion
Conditions
nP ≥ 5 and n(1 − P) ≥ 5
Null Hypothesis: H0 : P = P0
Test statistic:
p−P
Z= q
PQ
n
follows Standard normal distribution
q
Standard error of proportion= PQ n
95% confidence limits (that is α = 5%) for P is
r r
pq pq
(p − 1.96 , p + 1.96 )
n n
Module - 5 7 / 31
Problems on single proportion
1. 40 people were attacked by a disease and only 36 survived. At 5% LOS,

test whether the survival rate attacked by this disease is more than 85 %.
Solution:
Let x denotes the number of people survived after getting attacked by a
disease.
Here x = 36, n = 40
sample proportion p = xn = 36
40 = 0.9
1: H0 : P = 0.85 against H1 : P > 0.85 (Right tailed test)
2: Level of significance α = 0.05
3: Test Statistic:
Consider the conditions
nP = 40 × 0.85 = 34 > 5, n(1 − P) = 40 × 0.15 = 6 > 5
Module - 5 8 / 31
Problems
Solution contd:
Hence,
p−P 0.9 − 0.85
Z= q =q
PQ 0.85×0.15
n 40
= 0.8856
Note that Z follows standard normal distribution.
4: Critical region:
Zα = 1.65. The critical region is Z > 1.65 Since Cal Z = 0.8856 lies
in the acceptance region, we accept H0 (or fail to reject H0 )
5: Conclusion:
There is no statistical evidence to prove that more than 85% of the
people are attacked by a disease and survived.
Module - 5 9 / 31
Test of difference of proportions
Conditions
n1 p1 ≥ 5; n1 q1 ≥ 5; n2 p2 ≥ 5; n2 q2 ≥ 5
Null Hyp H0 : P1 − P2 = d where d is the difference in Population

proportions.
Module - 5 10 / 31
Conditions
n1 p1 ≥ 5; n1 q1 ≥ 5; n2 p2 ≥ 5; n2 q2 ≥ 5

proportions.
Test statistic :
(p1 − p2 ) − (P1 − P2 )
Z= q
PQ( n11 + n12 )
Module - 5 10 / 31
Conditions
n1 p1 ≥ 5; n1 q1 ≥ 5; n2 p2 ≥ 5; n2 q2 ≥ 5

proportions.
Test statistic :
(p1 − p2 ) − (P1 − P2 )
Z= q
PQ( n11 + n12 )

If P is unknown, then
n1 p1 + n2 p2
P=
n1 + n2
Module - 5 10 / 31
Problems
1. In a random sample of 100 men taken from Village A, 60 were found to
be consuming alcohol. In another sample of 200 men taken from Village
B, 100 were found to be consuming alcohol. Do the two villages differ
significantly in respect to the proportion of men who consume alcohol?
Module - 5 11 / 31
Problems
Solution:
Let x1 , x2 denotes the number of men consuming alcohol from Village A
and B resptly.
Here x1 = 60, n1 = 100, x2 = 100, n2 = 200
sample proportion p1 = nx11 = 100
60 100
= 0.6, p2 = 200 = 0.5
Module - 5 11 / 31
Problems
Solution:
and B resptly.
Here x1 = 60, n1 = 100, x2 = 100, n2 = 200
60 100
= 0.6, p2 = 200 = 0.5
1: H0 : P1 − P2 = 0 against H1 : P1 − P2 6= 0 (Two tailed test)
Module - 5 11 / 31
Problems
Solution:
and B resptly.
Here x1 = 60, n1 = 100, x2 = 100, n2 = 200
60 100
= 0.6, p2 = 200 = 0.5
Module - 5 11 / 31
Problems
Solution:
and B resptly.
Here x1 = 60, n1 = 100, x2 = 100, n2 = 200
60 100
= 0.6, p2 = 200 = 0.5
3: Test Statistic:
Consider the conditions n1 p1 = 100 × 0.6 = 60 > 5, n1 q1 = 40 >
5, n2 p2 = 100 > 5, n2 q2 = 100 > 5
Module - 5 11 / 31
Problems
Solution contd:
Hence,
n1 p1 + n2 p2 100 × 0.6 + 200 × 0.5
P= = = 0.533
n1 + n2 100 + 200
Module - 5 12 / 31
Problems
Solution contd:
Hence,
n1 p1 + n2 p2 100 × 0.6 + 200 × 0.5
P= = = 0.533
n1 + n2 100 + 200
(p1 − p2 ) − (P1 − P2 )
Z= q
PQ( n11 + n12 )
0.1
=q
1 1
0.533 × 0.467( 100 + 200 )
= 1.6366
Module - 5 12 / 31
Problems
Solution contd:
4: Critical region:
Since it is a two tailed test , The critical value is Z α2 = 1.96.
The critical region is |Z | ≥ 1.96. That is, critical region is
−3 < Z ≤ −1.96 or 1.96 ≤ Z < 3
Since −1.96 ≤ Cal Z = 1.6366 ≤ 1.96, we accept H0 (or fail to reject
H0 )
Module - 5 13 / 31
Problems
Solution contd:
4: Critical region:
Since it is a two tailed test , The critical value is Z α2 = 1.96.
The critical region is |Z | ≥ 1.96. That is, critical region is
−3 < Z ≤ −1.96 or 1.96 ≤ Z < 3
Since −1.96 ≤ Cal Z = 1.6366 ≤ 1.96, we accept H0 (or fail to reject
H0 )
5: Conclusion:
There is no statistical evidence to prove that two villages differ
significantly in respect of proportion.
Module - 5 13 / 31
Test of single mean
Conditions
If Population standard deviation is known, proceed with z-test.
If Population standard deviation is not known, then proceed with
z-test if it is a large sample (n ≥ 30). If small sample, proceed with
t-test (We will study in Module 6).
Module - 5 14 / 31
Test of single mean
Null Hyp H0 : µ = µ0
Module - 5 15 / 31
Test of single mean
Test statistic :
x̄ − µ
Z=
√σ
n
Module - 5 15 / 31
Test of single mean
Test statistic :
x̄ − µ
Z=
√σ
n

Standard error = √σn
95% confidence limits for the mean are
σ σ
(x̄ − 1.96 √ , x̄ + 1.96 √ )
n n
Module - 5 15 / 31
Problems
1. The mean lifetime of a sample of 100 tube lights produced by a

company is found to be 1580 hours with standard deviation of 90 hours.
Test the hypothesis at 1% LOS, that the mean life time of the tubes
produced by the company is 1600 hours.
Module - 5 16 / 31
Problems

Solution:
n = 100, x̄ = 1580, s = 90
Module - 5 16 / 31
Problems

Solution:
n = 100, x̄ = 1580, s = 90
1: H0 : µ = 1600 against H1 : µ 6= 1600 (Two tailed test)
Module - 5 16 / 31
Problems

Solution:
n = 100, x̄ = 1580, s = 90
Module - 5 16 / 31
Problems

Solution:
n = 100, x̄ = 1580, s = 90
3: Test Statistic:
Since n ≥ 30, Z follows standard normal distribution.
x̄ − µ
Z= = −2.22
√σ
n
Module - 5 16 / 31
Problems
Solution contd:
4: Critical region:
Since it is a two tailed test , the critical value is Z α2 = 2.58
The critical region is |Z | ≥ 2.58.
Note that Cal Z = −2.22 ≥ −2.58, we accept H0
5: Conclusion:
We conclude that the mean life time of the tubes produced by the
company is 1600 hours.
Module - 5 17 / 31
Test of difference of means
Conditions
If Population standard deviations are known, proceed with z-test.
If Population standard deviations are not known, then proceed with
z-test if both samples are such that n1 ≥ 30, and n2 ≥ 30. If small
sample, proceed with t-test (We will study in Module 6).
Module - 5 18 / 31
Conditions
Null Hyp H0 : µ1 − µ2 = d
Module - 5 18 / 31
Conditions
Null Hyp H0 : µ1 − µ2 = d
Test statistic :
(x¯1 − x¯2 ) − (µ1 − µ2 )

Z= q 2
σ1 σ22
n1 + n2
Module - 5 18 / 31
Problems
1. Intelligence test given to two groups of boys and girls gave the following
information:
Mean score SD Number

Girls 75 10 50
Boys 70 12 100
Is the difference in the mean scores of boys and girls statistically

significant?
Module - 5 19 / 31
Problems
1. Intelligence test given to two groups of boys and girls gave the following
information:
Mean score SD Number

Girls 75 10 50
Boys 70 12 100
Is the difference in the mean scores of boys and girls statistically

significant?
Solution:
Here x¯1 = 75, x¯2 = 70, s1 = 10, s2 = 12, n1 = 50, n2 = 100.
1: H0 : µ1 − µ2 = 0 against H1 : µ1 − µ2 6= 0 (Two tailed test)
Module - 5 19 / 31
Problems
Solution contd:
3: Test Statistic:
Since n1 , n2 ≥ 30, we proceed with Z -test.
(x¯1 − x¯2 ) − (µ1 − µ2 )

Z= q 2 = 2.6958
σ1 σ22
n1 + n2
4: Critical region:
Since it is a two tailed test , Z α2 = 1.96. The critical region is
|Z | ≥ 1.96
Note that Cal Z = 2.6958 ≥ 1.96, we reject H0
5: Conclusion:
We conclude that the difference in the mean score of boys and girls is
statistically significant.
Module - 5 20 / 31
More problems- Test of single proportion
1. A sample poll of 100 voters chosen at random from all the voters in a
given district indicated that 55% of them were in favour of a particular
candidate. Find (i) 95% and (ii) 99% confidence limits for the proportion
of all the voters in favour of this candidate.
Solution
Given n = 100, p = 0.55 (proportion of voters favouring the candidate.
(i) α = 5% = 0.05 =⇒ Z α2 = 1.96
q q
95% Confidence limits are given by (p − 1.96 pq n , p + 1.96 pq
n )
r r
0.55 × 0.45 0.55 × 0.45
(0.55 − 1.96 , 0.55 + 1.96 )
100 100
(0.4525, 0.6475)
Module - 5 21 / 31
Solution contd.
(ii) α = 1% = 0.01 =⇒ Z α2 = 2.58
q q
99% Confidence limits are given by (p − 2.58 pq
n , p + 2.58 pq
n )
r r
0.55 × 0.45 0.55 × 0.45
(0.55 − 2.58 , 0.55 + 2.58 )
100 100
(0.4216, 0.6784)
Module - 5 22 / 31
2. The population proportion is expected to be around 0.7. Find the
sample size needed to estimate the proportion within 0.02 with confidence
level of 90%.
Solution:
Given P = 0.7, p − P = 0.02
α = 10% = 0.1 =⇒ Z α2 = 1.645
p−P
We know that Z α2 = q PQ
n
q
PQ
p − P = Z α2 ×
qn
0.02 = 1.645 × PQ n =⇒ n = 1420.7
Hence, n = 1421 approx.
Module - 5 23 / 31
3. A die is thrown 9000 times and throw of 3 or 4 is observed 3240 times.
Show that the die cannot be regarded as an unbiased one.
Solution:
Let x be the number of times 3 or 4 occurs.
n = 9000, x = 3240, p = 3240
9000 = 0.36
1 1
1: H0 : P = 3 (unbiased) against H1 : P 6= 3 (biased, Two tailed test)
3: Test Statistic:
Consider the conditions
nP = 9000×0.333 = 2997 > 5, n(1−P) = 9000×0.667 = 6003 > 5
Module - 5 24 / 31
Problems
Solution contd:
Hence,
p−P 0.36 − 0.333
Z= q =q = 5.37
PQ 0.333×0.667
n 9000

4: Critical region:
Z α2 = 1.96. The critical region is |Z | ≥ 1.96 Since Cal Z = 5.37 lies
in the critical region, we reject H0 .
5: Conclusion:
We conclude that the die is biased.
Module - 5 25 / 31
More problems- Test of difference of proportions
1.A cigarette manufacturing firm claims that its brand B cigarette outsells
its brand A by 8%. If it is found that 42 out of a sample of 200 smokers
prefer brand A and 18 out of another random sample of 100 smokers
prefer brand B, test whether the 8% difference is a valid claim.
Solution:
Given n1 = 200, n2 = 100, x1 = 42, x2 = 18
42 18
p1 = 200 = 0.21, p2 = 100 = 0.18
1. H0 : P1 − P2 = 0.08 against H1 : P1 − P2 < 0.08 (Left tailed test)
2. LOS α = 5% = 0.05
3. n1 p1 = 200 × 0.21 = 42 > 5, n1 q1 = 200 × 0.79 = 158 > 5, n2 p2 =
100 × 0.18 = 18 > 5, n2 q2 = 82 > 5
P = n1 np11 +n
+n2 p2
2
= 0.2
Module - 5 26 / 31
Solution contd.
(p1 − p2 ) − (P1 − P2 )
Z= q
PQ( n11 + n12 )
(0.21 − 0.18) − 0.08

=q
1 1
0.2 × 0.8( 200 + 100 )
= −1.02
4: Critical region:
Since it is a left tailed test , The critical value is Zα = −1.645.
The critical region is Z ≤ −1.645.
Since Cal Z = −1.02 ≥ −1.645, we accept H0 (or fail to reject H0 )
Module - 5 27 / 31
Solution contd.
(p1 − p2 ) − (P1 − P2 )
Z= q
PQ( n11 + n12 )
(0.21 − 0.18) − 0.08

=q
1 1
0.2 × 0.8( 200 + 100 )
= −1.02
4: Critical region:
Since it is a left tailed test , The critical value is Zα = −1.645.
The critical region is Z ≤ −1.645.
Since Cal Z = −1.02 ≥ −1.645, we accept H0 (or fail to reject H0 )
5: Conclusion:
The difference of 8% in the sale of two brands is a valid claim.
Module - 5 27 / 31
2. In a year, there are 956 births in a town A of which 52.5% were male,
while in towns A and B combined, this proportion in a total of 1406 births
was 0.496. Is there any significant difference in the proportion of male
births in the two towns?
Hint:
n1 = 956, n1 + n2 = 1406 =⇒ n2 = 450
p1 = 0.525, combined proportion is 0.496 = n1 np11 +n
+n2 p2
2
Hence p2 = 0.434.
H0 : P1 − P2 = 0 against H1 : P1 − P2 6= 0 (two-tailed)
Cal Z = 3.184, Z α2 = 1.96
Cal Z lies in the rejection region. Hence reject H0 (there is significant
difference in the proportion of male births in the two towns)
Module - 5 28 / 31
More problems- Test of single mean
1. A sample of 900 members has a mean 3.4 cm and SD 2.61cm. Is the

sample from a large population of mean 3.25cm and SD 2.61cm. If the
population is normal and the mean is unknown, find the 95% confidence
limits for the mean.
Module - 5 29 / 31
More problems- Test of single mean
1. A sample of 900 members has a mean 3.4 cm and SD 2.61cm. Is the

sample from a large population of mean 3.25cm and SD 2.61cm. If the
population is normal and the mean is unknown, find the 95% confidence
limits for the mean.
Hint:
Given n = 900, x̄ = 3.4, s = σ = 2.61.
H0 : µ = 3.25 against H1 : µ 6= 3.25
Cal Z = 1.724 and Z α2 = 1.96
Critical region is |Z | > 1.96 =⇒ we accept H0 .
95% confidence interval is given by
σ σ
(x̄ − 1.96 √ , x̄ + 1.96 √ )
n n
(3.23, 2.57)
Module - 5 29 / 31
More problems- Test of difference of means
1. The mean production of wheat from a sample of 100 fields comes to

200 kg per acre and another sample of 150 fields gives a mean of 220 kg
per acre. Assuming the SD of the yield at 11 kg for the universe, test if
there is a significant difference between the means of the populations.
Hint:
Given n1 = 100, n2 = 150, x¯1 = 200, x¯2 = 220, σ = 11
H0 : µ1 − µ2 = 0 against H1 : µ1 − µ2 6= 0 (Two tailed)
Cal Z = −14.08, Z α2 = 1.96
Since Cal Z lies in the critical region (−3 < Z < −1.96 or 1.96 < Z < 3),
we reject H0
There is a significant difference between the means of the two populations.
Module - 5 30 / 31
Try this
1. A sample of heights of 6400 Englishmen has a mean of 170cm and a

SD of 6.4cm, while a sample of heights of 1600 Indians has a mean of
172cm and a SD of 6.3cm. Do the data indicate that Indians are on the
average taller than Englishmen?
2. In city A, out of a sample of 50 persons, 26 were regular consumers of
tea. In city B, out of a sample of 60 persons, 43 were regular consumers of
tea. Test whether the tea consuming habit is equally distributed in both
the cities at 5% level of significance.
3. A random sample of 50 observations from normal population gave an
arithmetic mean of 32 units with a standard deviation of 2 units. Test
whether the population mean is 30 at 1% LOS.
4. A sales clerk in a departmental store claims that 60% of the shoppers
entering the store leave without making a purchase. A random sample of
50 shoppers showed that 35 of them left without buying anything. Are
these sample results consistent with the claim of the sales clerk?
Module - 5 31 / 31

Module - 5 PROB

Uploaded by

Copyright:

Available Formats

Module - 5 PROB

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Module - 5 PROB

Uploaded by

Copyright:

Available Formats

Testing of Hypothesis

An hypothesis to be tested or to answer a question - Alternative

P(Type I error) = α = level of significance (LOS)

Formulate Null and Alternate hypothesis

Test LOS 1% (0.01) 5% (0.05) 10% (0.1)

1. 40 people were attacked by a disease and only 36 survived. At 5% LOS,

Null Hyp H0 : P1 − P2 = d where d is the difference in Population

Null Hyp H0 : P1 − P2 = d where d is the difference in Population

follows Standard normal distribution

Null Hyp H0 : P1 − P2 = d where d is the difference in Population

follows Standard normal distribution

follows Standard normal distribution

follows Standard normal distribution

1. The mean lifetime of a sample of 100 tube lights produced by a

1. The mean lifetime of a sample of 100 tube lights produced by a

1. The mean lifetime of a sample of 100 tube lights produced by a

1. The mean lifetime of a sample of 100 tube lights produced by a

1. The mean lifetime of a sample of 100 tube lights produced by a

(x¯1 − x¯2 ) − (µ1 − µ2 )

follows Standard normal distribution

Mean score SD Number

Is the difference in the mean scores of boys and girls statistically

Mean score SD Number

Is the difference in the mean scores of boys and girls statistically

(x¯1 − x¯2 ) − (µ1 − µ2 )

Note that Z follows standard normal distribution.

(0.21 − 0.18) − 0.08

(0.21 − 0.18) − 0.08

1. A sample of 900 members has a mean 3.4 cm and SD 2.61cm. Is the

1. A sample of 900 members has a mean 3.4 cm and SD 2.61cm. Is the

1. The mean production of wheat from a sample of 100 fields comes to

1. A sample of heights of 6400 Englishmen has a mean of 170cm and a

You might also like