Toh

TESTING OF HYPOTHESIS
Population: Any set of objects under study in a statistical

investigation is called population.
Eg:- If we want to study the average life time of the electric bulbs,
here the term population is the number of electric bulbs.
Sample: A finite subset of the population is known as sample and the

number of elements in that sample is known as sample size.
Sampling error: The error involved in approximation about the

population characteristics on the basis of the sample is known as
sampling error and it is inherent and unavoidable in any sample
scheme.
Parameter: Any statistical constant related to the population is

known as parameter.
Eg:- i) Population mean  ii) Population Variance 2
Statistic: Any statistical constant related to the sample is known as

statistic
Eg:- i) Sample mean x ii) Sample variance s2
Tests of significance: A very important aspect of the sampling theory

is the study of the tests of significance which enables us to decide on
the basis of the sample whether
i) the deviation between the observed sample statistic and
the hypothetical parameter value is very less.
ii) the deviation between two sample statistics is significant
or not.
Null hypothesis: For applying the test of significance, we first set up

a hypothesis i.e a definite statement about the population parameter.
Such a hypothesis which is usually a hypothesis of no difference is
called Null hypothesis.
According to R.A.Fisher “Null hypothesis is the hypothesis
which is tested for all possible rejection under the assumption that it
is true”. It is usually denoted by H0.
1
Alternative hypothesis: Any hypothesis which is complimentary to
the Null hypothesis is called an alternative hypothesis. It is usually
denoted by H1.
Degrees of freedom: Degrees of freedom is the number of

independent observations in a set. In a test of hypothesis a sample is
drawn from the population of which the parameter is under a test. The
size of the sample varies since it depends either on the experimenter
or on the resources available. Moreover the test statistic involves the
estimated value of the parameter which depends on the number of
observations. Hence the sample size plays an important role in testing
of hypothesis and is taken care of by degrees of freedom.
Distinguish between type I and Type II errors in testing the

hypothesis.
Decision from Sample
Reject H0 Accept H0
Wrong Correct
H0 True (Type I error)
H0 False Correct Wrong

(Type II error)
Type I error: Probability of rejecting H0 when H0 is true.

Type II error: Probability accepting H0 when H0 is false.
Level of significance() The probability of type I error is known as

level of significance of the test. It is also called the size of the
critical region.
Point estimation: If from the sample information available, we can

find a statistic which is almost equal to the true value of the parameter.
The method of finding such a statistic is known as point estimation.
2
Eg:- sample mean is a point estimate of population mean.
Interval estimation: If from the sample information available, we can

find an interval in which a parameter is expected to lie with a particular
probability, the method of finding out such an interval is called interval
estimation.
Eg:- The 95% confident interval for the population mean  is
   
 x − 1.96 , x + 1.96 
 n n
Procedure for testing the hypothesis.

Step 1: Null hypothesis: Set up the null hypothesis H0.
Step 2: Alternative hypothesis: Set up the alternative hypothesis H1.
Step 3: Level of significance: Choose the appropriate level of
significance  which is fixed in advance.
Step 4: Test statistic: Compute the test statistic under H0.
Step 5: Inference: We compare the computed value with the table
value/ critical value. If the obsolete computed value is less than
or equal to the table value/critical value, we accept H0.
Otherwise we reject
Critical values:
Critical values Z Level of significance ()

1% 5% 10%
Two-tailed test 2.58 1.96 1.645
Right-tailed test 2.33 1.645 1.28
Left-tailed test -2.33 -1.645 -1.28
t-test
Test for Single mean:
H0:  = 0
3
H1:   0 (or)  > 0 (or)  < 0
[Any one of these 3 conditions depends on the given problem]
Level of significance:
Appropriate level of significance is % (given/chosen)
Test Statistic:
To test the above hypotheses the test statistic is given by
x −μ
t= ~ t n−1 at α% level of significance
s n
Inference:
1) If it is a two tailed test (), Reject H0 if tcal  t/2,n-1. Otherwise
accept H0.
2) If it is a right tailed test (>), Reject H0 if tcal  t,n-1. Otherwise
accept H0.
3) If it is a left tailed test (<), Reject H0 if tcal  -t,n-1. Otherwise
accept H0.
Confident Limits: The 100(1-)% confident limits for the

population mean are given by:
1) If it is a two tailed test ():

s s
[ x − t α/2,n −1. , x + t α/2,n −1. ]
n n
2) If it is a right tailed test (>):
s s
[ x − t α,n −1. , x + t α,n −1. ]
n n
3) If it is a left tailed test (<):
s s
[ x − (− t α,n −1 ). , x + (− t α,n −1 ). ]
n n
Problems related to test for single mean.
1) (Dec 2004): Ten individuals are chosen from a normal population
4
and their heights (in inches) are given below. Test whether the
sample comes from a normal population whose mean height is 66
inches or not at 5% level of significance?
63,63,66,67,68,69,70,70,71,71.
Sol:-
H0:  = 66
H1:   66
Appropriate level of significance is 5% (given)
Test Statistic:
x −μ
s n
x=
x i
= 67.8
n
s=
 (x i − x) 2
= 3.0111
n −1
Under H0, the test statistic is given by
67.8 − 66
t= = 1.89 ~ t 9 at 5% level of significance
3.0111 10
Inference:
The tabulated value of t at 5% level of significance for 9
degrees of freedom in a two tailed test is 2.262.
[t/2,n-1=t0.05/2,10-1=t0.025,9=2.262]
Here, tcal < t/2,n-1 . So, we accept H0. Hence we conclude that
the sample comes from the population whose mean value is 66inches
2) A sample of 10 boys had the following I.Qs. 70, 120, 110, 101, 88,
83, 95, 58, 107, 110. Do these data support the assumption of a
population mean I.Q of 100? Test at significant at α=0.05.
5
Sol:-
H0:  =100
H1:   100
Appropriate level of significance is 0.05 (given)
Test Statistic:
x −μ
s n
x=
x i
= 97.2
n
s=
 (x i − x) 2
= 13.541
n −1

92.7 − 100
t= = −0.6203~t 9 at 5% level of significance
13.541 10
Inference:
degrees of freedom in a two tailed test is 2.262
[t/2,n-1=t0.05/2,10-1=t0.025,9=2.262]
population mean. I.Q is 100.
3) A sample of nine items had the following values

45,47,50,52,48,47,49,53,51. Does the mean of the nine items differ
significantly from the assumed population mean of 47.5? Test at
1% level of significance.
Sol:-
H0:  = 47.5
6
H1:   47.5
Test Statistic:
x −μ
s n
x=
x i
= 49.1111
n
s=
 (x i − x) 2
= 2.62
n −1

49.1111 − 47.5
2.62 9
Inference:
[t/2,n-1=t0.01/2,10-1=t0.005,9=3.355]
There is no significant difference between sample mean and the
population mean.
4) 5 measurements of tar content of certain kind of cigarette yielded

14.5, 14.2, 14.4, 14.3, 14.6 milligrams per cigarette. Show that the
difference between the mean of this sample and the average tar
claimed by the manufacturer μ=14.0 mg/cigarette is significant at
α=0.05.
Sol:-
7
H0:  = 14.0
H1:   14.0
Test Statistic:
x −μ
s n
x=
x i
= 14.4
n
s=
 (x i − x) 2
= 0.1581
n −1

14.4 - 14.0
0.1581 5
Inference:
[t/2,n-1=t0.05/2,5-1=t0.025,4=2.776]
Here, tcal > t/2,n-1 . So, we reject H0. Hence we conclude that
  14.0
5) The specifications for a certain kind of ribbon has a mean breaking

strength of 180 pounds. If 5 pieces of the ribbon have a mean
breaking strength of 169.5 pounds with a standard deviation of 5.7
pounds. Test the null hypothesis μ=180 pounds against the
alternative μ<180 pounds at 0.01 level of significance.
Sol:-
H0:  = 180
8
H1:  < 180
Test Statistic:
x −μ
s n
Given x = 169.5, s = 5.7, n = 5

169.5 - 180
t= = −4.12 ~ t 4 at 1% level of significance
5.7 5
Inference:
degrees of freedom in a left tailed test is -3.747
[-t,n-1=-t0.01,5-1=-t0.01,4=-3.747]
Here, tcal < -ttab . So, we reject H0. Hence we conclude that
 <180 pounds.
6) Random sample of 6 steel beams has a mean breaking strength of

58,392 psi (pounds per sq.inch) with a standard deviation of 648
psi. Use this information and level of significance α=0.05 to test
whether the true average compressive strength of the steel from
which this sample has come is 58,000 psi.
Sol:-
H0:  = 58,000 psi
9
H1:   58,000 psi
Test Statistic:
x −μ
s n
Given x = 58,392, s = 648, n = 6

58392 - 58000
648 6
Inference:
[t/2,n-1=t0.05/2,6-1=t0.025,5=2.571]
Here, tcal < t/2,n-1 . So, we accept H0.
7) A random sample of 8 envelopes is taken from a letter box of a post

office and their weights in grams are found to be
12.1,11.9,12.4,12.3,11.9,15.4,13.9,12.8. Find 99% confidence
interval for the population mean weight of the envelopes received
at that post office.
Sol:-
The 100(1-)% confident limits for the population mean are
10
s s
[ x − t α/2,n −1. , x + t α/2,n −1. ]
n n
x=
x i
= 12.8375
n
s=
 (x i − x) 2
= 1.2235
n −1
[t/2,n-1=t0.0 1/2,8-1=t0.005,7=3.499]

[11.3237,14.3513]
8) 10 bearings made by a certain process have a mean diameter of

0.506 cm with a standard deviation of 0.04cm. Construct the 95%
confidence interval for the actual average diameter of the bearings
made by this process.
Sol:-
s s
[ x − t α/2,n −1. , x + t α/2,n −1. ]
n n
Given x = 0.506, s = 0.04, n = 10
[t/2,n-1=t0.05/2,10-1=t0.025,9=2.262]

[0.4773,0.5346]
9) The mean weight loss of n=16 grinding balls after certain length of
time in a mill slurry is 34.2 gms with a standard deviation of 0.68
gms. Construct a 95% confidence interval for the true mean loss of
such grinding balls.
Sol:-
11
s s
[ x − t α/2,n −1. , x + t α/2,n −1. ]
n n
Given x = 34.2, s = 0.68, n = 16
[t/2,n-1=t0.05/2,16-1=t0.025,15=2.947]

[33.7,34.7]
Test for equality of Two means (n1n2) :

H0: x = y
H1: x  y (or) x > y (or) x < y
12
Test Statistic:
(x - y) − (μ x − μ y )
t= ~ t n1 + n 2 −1 at α% level of significance
1 1
S +
n1 n 2
Inference:
1) In a two tailed test (), Reject H0 if t cal  t α/2, n1 + n 2 − 2 .
Otherwise accept H0.
2) In a right tailed test (>), Reject H0 if .
t cal  t α, n1 + n 2 − 2
3) In a left tailed test (<), Reject H0 if t cal  − t α, n 1 + n 2 − 2
1). In a two tailed test ():
1 1 1 1
[( x - y) − t α/2,n −1S + , (x - y) − t α/2,n −1S + ]
n1 n2 n1 n2
2). In a right tailed test (>):
1 1 1 1
[( x - y) − t α,n −1S + , (x - y) − t α,n −1S + ]
n 1 n2 n 1 n2
3). In a left tailed test (<):

1 1 1 1
[( x - y) − ( − t α,n −1 )S + , (x - y) − ( − t α,n −1 )S + ]
n 1 n2 n 1 n2
Problems related to test for equality of two means.

1) The following random samples are measurements of heat
producing capacity in millions of calories per ton of specimens of
coal from two mines.
Mine 1 8260 8130 8350 8070 8340
Mine 2 7950 7890 7900 8140 7920 7840
13
Use 0.01 level of significance to test whether the difference
between the means of these two samples is significant?
Sol:-
H0: x = y
H1: x ≠ y
Test Statistic:
(x - y) − (μ x − μ y )
1 1
S +
n1 n 2
x=
x i
= 8230 y=
y i
= 7940
n1 n2
S=
 (x i − x) 2 +  (y i − y) 2
= 114.31
n1 + n2 − 2
(8230 − 7940) - (0)

1 1
114.31 +
5 6
Inference:
[t/2,n1+n2-2=t0.01/2,9=t0.005,9=3.250]
Here,  tcal  > ttab . So, we reject H0.
2) The heights of 6 randomly chosen sailors are 63,65,68,69,71,72

inches. Those of 10 randomly chosen soldiers are
61,62,65,66,69,69,70,71,72,73 inches. Discuss whether this data
gives a suggestion that the sailors are taller than soldiers.
Sol:-
H0: x = y
H1: x > y
14
Appropriate level of significance is 5% (chosen)
Test Statistic:
(x - y) − (μ x − μ y )
1 1
S +
n1 n 2
x=
x i
= 68, y=
y i
= 67.8
n1 n2
S=
 (x i − x) 2 +  (y i − y) 2
=
60 + 153.6
= 3.906
n1 + n2 − 2 14
(68 − 67.8) - (0)

t= = 0.099 ~ t14 at 5% level of significance
1 1
3.906 +
6 10
Inference:
degrees of freedom in a right tailed test is 1.761.
[t,n1+n2-2=t0.05,14=t0.05,14=1.761]
Here, tcal < t,n-1 . So, we accept H0. Hence we conclude that
there is no significant difference between the mean heights of the
sailors and soldiers.
3) The following are the no. of sales which a sample of 9 sales people
of industrial chemicals in California and a sample of 6 sales people
of industrial chemicals in Oregon made over a certain fixed period
of time ?
California 59 68 44 71 63 46 69 54 48
Oregon 50 36 62 52 70 41
Test the null hypothesis x - y = 0 against the alternative
hypothesis x - y ≠ 0 at 0.01 level of significance.
15
Sol:-
H0: x - y = 0
H1: x - y ≠ 0
Test Statistic:
(x - y) − (μ x − μ y )
1 1
S +
n1 n 2
x=
x i
= 58, y=
y i
= 51.83
n1 n2
S=
 (x i − x) 2 +  (y i − y) 2
= 11.3572
n 1 + n2 − 2
(58 - 51.83) - (0)

1 1
11.3572 +
9 6
Inference:
[t/2,n1+n2-2=t0.01/2,9+6-2=t0.005,13=3.012]
Here, tcal < ttab. So, we accept H0.
4) Samples of two types of electric bulbs were tested for length of life
and the following data is obtained. Is the difference in the means
sufficient to warrant that type I is superior to type II regarding
length of the life.
Type I Type II
Sample No. 8 7
Mean 1234 hours 1036 hours
Sample S.D 36 hours 40 hours
Sol:-
16
H0: x = y
H1: x > y
Test Statistic:
(x - y) − (μ x − μ y )
1 1
S +
n1 n 2
Given, n1 = 8, n 2 = 7, x = 1234, y = 1036, s x = 36, s y = 40
( n1 − 1) s x + ( n2 − 1) s y
2 2
7  36 2 + 6  40 2
S= = = 37.8986
n1 + n2 − 2 13
(1234 - 1036) - (0)

1 1
37.8986 +
8 7
Inference:
degrees of freedom in a right tailed test is 1.771.
[t,n1+n2-2=t0.05,8+7-2=t0.05,13=1.771]
Here, tcal > ttab. So, we reject H0. Hence we conclude that type I
bulbs are superior to type II bulbs regarding length of life.
5) Two random samples gave the following results:
Sample size Sample Mean Sum of squares of deviations from the mean
1 10 15 90
2 12 14 108
Examine the difference of means
Sol:-
H0: x = y
H1: x  y
Appropriate level of significance is 5% (hosen)
Test Statistic:
17
(x - y) − (μ x − μ y )
1 1
S +
n1 n 2
Given, n1 = 10, n 2 = 12, x = 15, y = 14
S=
 (x i − x) 2 +  (y i − y) 2
=
90 + 108
= 3.1464
n1 + n 2 − 2 20
(15 - 14) - (0)

1 1
3.1464 +
10 12
Inference:
[t/2,n1+n2-2=t0.05/2,18=t0.025,18=2.086]
Here, tcal < ttab. So, we accept H0. Hence we conclude that there
is no significant difference between the means.
Test for equality of Two means (n1= n2) (Paired t-test)

H0: d = 0
H1: d  o (or) d > 0 (or) d < 0
Test Statistic:
d − μd
t= ~ t n −1 at α% level of significance
sd n
18
Inference:
1) If it is a two tailed test (), Reject H0 if tcal  t/2,n-1. Otherwise
accept H0.
2) If it is a right tailed test (>), Reject H0 if tcal  t,n-1. Otherwise
accept H0.
3) If it is a left tailed test (<), Reject H0 if tcal  -t,n-1. Otherwise
accept H0.

1) If it is a two tailed test ():

sd s
[ d − t α/2,n −1. , d + t α/2,n −1. d ]
n n
2) If it is a right tailed test (>):
sd s
[ d − t α,n −1. , d + t α,n −1. d ]
n n
3) If it is a left tailed test (<):
sd s
[ d − (− t α,n −1 ). , d + (− t α,n −1 ). d ]
n n
Problems on paired t-test

1) The following are the average weekly loses of working hours due
to accidents in 10 industrial plants before and after a certain safety
program was put into operation. Use the 5% level of significance
to test whether the safety program is effective
Before 45 73 46 124 33 57 83 34 26 17
After 36 60 44 119 35 51 77 29 24 11
Sol:-
H0: d = 0( There is no difference between in the average no.of
accidents before and after the safety program is put into operation)
H1: d > 0 (Training program is effective i.e the average number of
accidents decreased after the safety program is implemented)
19
Test Statistic:
d − μd
sd n
di 9 13 2 5 -2 6 6 5 2 6
d=
d i
= 5.2
n
sd =
 (d i − d) 2
= 4.08
n −1
5.2 - 0
4.08 10
Inference:
degrees of freedom in a right tailed test is 1.833
[t,n-1=t0.05,10-1=t0.05,9=1.833]
Here, tcal < t,n-1 . So, we accept reject H0. Hence we conclude that
The safety program is not useful
2) To compare two kinds of bumper guards 6 of each kind were

mounted on a certain kind of compact car . Then each car was run
into a concrete wall at 5 miles per hour and the following are the
cost of repairs in dollars. Use 0.01 level of significance to test
whether the difference between the two sample means is
significant?
Bumper guard 1 107 148 123 165 102 119
Bumper guard 2 134 115 112 151 133 129
Sol:-
H0: d = 0
H1: d ≠ 0
20
Test Statistic:
d − μd
sd n
di -27 33 11 14 -31 -10
d=
d i
= −1.6666
n
sd =
 (d i − d) 2
= 25.216
n −1
- 1.6666 - 0
t= = −0.1618 ~ t 5 at 1% level of significance
25.216 6
Inference:
[t/2,n-1=t0.005,5=4.032]
Here, tcal < ttab. So, we accept H0.
3) A certain stimulus administered to each of the 12 patients resulted

in the following increase in blood pressure, 5,2,8,-1,3,0,-2,1,5,0,4
and 6. Can it be concluded that the stimulus will in general, be
accompanied by an increase in blood pressure?
Sol:-
H0: d = 0
H1: d < 0
Test Statistic:
21
d − μd
sd n
di 5 2 8 -1 3 0 -2 1 5 0 4 6
d=
d i
= 2.58
n
sd =
 (d i − d) 2
= 3.09
n −1
2.58 - 0
3.09 12
Inference:
degrees of freedom in a left tailed test is -1.796
[-t,n-1=-t0.05,10-1=-t0.05,9=-1.796]
Here, tcal > t,n-1 . So, we accept H0. Hence we conclude that
the stimulus has no effect on blood pressure.
Test for Single Variance/ Single S.D:

H0: σ2=σ02
H1: σ2≠σ02 (or) σ2>σ02 (or) σ2<σ02
Test Statistic:
 2
=
 (x i - x) 2
~  2 n -1 at α% level of significance
 2
22
(n - 1)s 2
 =
2
 2
Inference:
1) If it is a two tailed test (), Reject H0 if χ2cal  χ2α/2,n-1.
2) If it is a right tailed test (>), Reject H0 if χ2cal  χ2α,n-1.
3) If it is a left tailed test (<), Reject H0 if χ2cal ≤ χ21-α,n-1.

population variance are given by:
1.Two- tailed test:
 (n − 1)s 2 (n − 1)s 2 
 2 , 2 
χ
 α/2,n −1 χ 1-α/2,n −1 
2.Right- tailed test and left tailed test:
 (n − 1)s 2 (n − 1)s 2 
 2 , 2 
χ
 α,n −1 χ 1-α,n −1 
Problems related to test for single Variance.

1.A random sample of 6 steel beams has a mean compressive strength of
58,392 psi with a standard deviation of 648 psi. Use this information and
L.O.S α=0.05 to test the null hypothesis σ=600 psi against the alternative
hypothesis σ> 600 psi.
Sol:-
H0: σ=600
H1: σ> 600
Test Statistic:
23
(n - 1)s 2
 =
2
 2

5 * 6482
 =
2
= 5.832
600 2
Inference:
The table value of χ2 at 5% level of significance in a right
tailed test is 11.070
χ2α,n-1= χ20.05,5=11.070
Here, χ2cal < χ2tab . So, we accept
2. The lapping process which is used to grind certain silicon wafers to the
proper thickness is acceptable only if σ, the population standard deviation
is at most 0.5mil. Use 0.05 level of significance to test the null hypothesis
σ=0.5 mil against the alternative hypothesis σ>0.5 mil, if the thickness of
15 wafers have a standard deviation of 0.64 mil.
Sol:-
H0: σ=0.5mil
H1: σ>0.5mil
Test Statistic:
24
(n - 1)s 2
 =
2
 2

14 * 0.64 2
 =
2
= 22.9376
0.52
Inference:
The table value of χ2 at 5% level of significance for 14 d.f in a
right tailed test is 23.685
χ2α,n-1= χ20.05,14=23.685
Here, χ2cal < χ2tab . So, we accept
3. If 12 observations of the specific heat of iron have a standard deviation

of 0.0086, test the null hypothesis that σ=0.01 for such observations. Use
the alternative hypothesis σ≠0.01 and level of significance 0.01.
Sol:-
H0: σ=0.01
H1: σ≠0.01
Test Statistic:
25
(n - 1)s 2
 =
2
 2

12 * 0.0086 2
 =
2
= 8.1356
0.012
Inference:
two tailed test is 26.757
χ2α/2,n-1= χ20.005,11=26.757
Here, χ2cal < χ2tab . So, we accept H0.
4. A random sample of 15 observations from a population are

14,15,13,21,14,12,15,16,18,20,22,24,14,12,10. Test whether the
population variance 7.5 at 5% level of significance.
Sol:-
H0: σ2=7.5
H1: σ2≠7.5
Test Statistic:
26
 2
=
 (x i - x) 2
 2
Given x = 16,  (x i − x ) 2 = 236

236
2 = = 31.4667
7.5
Inference:
two tailed test is 26.119
χ2α/2,n-1= χ20.025,14=26.119
Here, χ2cal > χ2tab . So, we reject H0.
5). While performing a task, the pulse rate of 25 workers increased on the
average by 18.4 beats per minute with a standard deviation of 4.9
beats/min. Find 95% confidence intervals for the corresponding
population standard deviation . Test the null hypothesis σ2=30.0 against
the alternative hypothesis σ2<30.0
Sol:- Part-A
H0: σ2=30.0
H1: σ2<30.0
Test Statistic:
27
(n - 1)s 2
 = 2
 2

(25 - 1)4.9 2
 =2
= 19.208
30.0
Inference:
The table value of χ2 at 5% level of significance for 24 d.f in left
χ21-α,n-1= χ21-0.05,24= χ20.95,24=13.484
Here, χ2cal > χ2tab . So, we accept H0.
Part-B
The 100(1-α)% confident limits for the population variance in the left
tailed test are given by
 (n − 1)s 2 (n − 1)s 2   24  4.9 2 24  4.9 2 
 2 , 2 = 2 , 2 
χ
 α,n −1 χ 1-α, n −1  
  χ 0.05,24 χ 0.95,24 
 576.24 576.24 
 36.415 , 13.484 
[15.8242,42.7351] are the limits of population variance.

[3.978,6.5372] are the limits of population S.D
Test for Two Variances/ Two S.Ds:
H0: σx2=σy2
H1: σx2 ≠ σy2 (or) σx2 > σy2 (or) σx2 < σy2
Test Statistic:
S2x
F = 2 ~ Fn1-1,n2-1 at α% level of significance
Sy
28
Inference:
1) If it is a two tailed test (), Reject H0 if Fcal  Fα,n1-1,n2-1.
2) For both right tailed and left tailed tests (>, <), Reject H0 if
Fcal ≤ F,1-α,n1-1,n2-1. Otherwise accept H0.
Note:-
Fcal≥1 i.e the calculated value of F must be always greater
than or equal to !. If not the test statistic becomes 1/F.
1). Random samples from two normal populations are given below
Sample 1 16 26 27 23 24 22
Sample 2 33 42 35 32 28 31
Do the population variances differ significantly?

Sol:-
H0: σx2=σy2
H1: σx2 ≠ σy2
Test Statistic:
29
S2x
Sy
x = 23, S 2
=
 (x i − x) 2
= 15.2
n1 − 1
x
y = 33.5, S 2
=
 (y i − y) 2
= 22.7
n2 −1
y
15.2
F= = 0.6696  1 ~ F5,5 at 5% level of significance
22.7
22.7
F= = 1.4934
15.2
Inference:
The table value of F at 5% L.O.S for (5,5) d.f is 5.05.
Fα,n2-1,n1-1=F0.05,5,5=5.05
Fcal < Ftab, We accept H0.
2).Two independent random samples of 8 and 7 items respectively have

the following values
Sample 1 9 11 13 11 15 9 12 14
Sample 2 10 12 10 14 9 8 10
Test whether the difference between the variances is

significant at 1% level of significance?
Sol:-
H0: σx2=σy2
H1: σx2 ≠ σy2
30
Test Statistic:
S2x
Sy
x = 11.75, S 2
=
 (x i − x) 2
= 4.786
n1 − 1
x
y = 10.43, S 2
=
 (y i − y) 2
= 3.952
n2 −1
y
4.786
F= = 1.21 > 1 ~F7,6 at 1% level of significance
3.952
Inference:
Fα,n1-1,n2-1=F0.05,7,6=8.26
Fcal < Ftab, We accept H0.
3).It is desired to determine whether there is less variability in the silver

plating done by company 1 than in that done by company2 . If
independent random samples of size 12 of the companies work yield
S1=0.035 mil and S2= 0.062 mil. Test the null hypothesis σ12=σ22
against the alternative hypothesis σ12<σ22 at 5% level of significance
Sol:-
H0: σx2=σy2
H1: σx2 < σy2
Test Statistic:
31
S2x
Sy
0.0352
F= 2
= 0.3186  1 ~ F11,11 at 5% level of significance
0.062
0.062 2
F= = 3.138
0.0352
Inference:
Fα,n1-1,n2-1=F0.05,11,11=2.82
Fcal > Ftab, We reject H0.
4).If the independent random samples of size n1=7 and n2=13 are taken
from a normal population, what is the probability that the variance of
the first sample will be at least three times as large as that of
second sample.
Sol:-
S2x
Sy
3  S2y
F= =3
S2y
From statistical tables F0.05,6,12=3.
Therefore, the probability level is 95%
5). If the independent random samples of size n1=9 and n2=16 are
taken from a normal population, what is the probability that the
variance of the first sample will be at least four times as large as
that of second sample.
Sol:-
S2x
Sy
4  S2y
F= =4
S2y
32
From statistical tables F0.01,8,15=4.
Therefore, the probability level is 99%
6). Find the value of F0.95 for υ1=10 and υ2=20 degrees of freedom.
Sol:-
1
F
The relation is 1−α, ν1, ν2 =
F , ν2, ν1
1 1
F0.95,10,20 = = = 0.361
F0.05, 20,10 2.77
Large Sample Tests(n>30)
Critical values Z Level of significance ()

1% 5% 10%
Two-tailed test 2.58 1.96 1.645
Right-tailed test 2.33 1.645 1.28
Left-tailed test -2.33 -1.645 -1.28
Test for Single mean:

H0:  = 0
33
H1:   0 (or)  > 0 (or)  < 0
Test Statistic:
x −μ
z= ~ N(0,1) at α% level of significance
 n
Inference:
1) If it is a two tailed test (), Reject H0 if Zcal  Zcritical value.
2) If it is a right tailed test (>), Reject H0 if Zcal  Zcritical value.
3) If it is a left tailed test (<), Reject H0 if Zcal  Zcritical value.

 
[ x − Z critical value . , x + Z critical value . ]
n n
1) A trucking firm is suspicious of the claim that the average life time
of certain tyres is at least 28,000 miles. To check this claim the firm
puts 40 of these tyres on its trucks and gets a mean life time of
27,468 miles with a standard deviation of 1,348 miles. What can
we conclude if the probability of type I error is to be
at most 0.01?
Sol:-
H0:  ≥ 28,000 miles
H1:  < 28,000 miles
34
Test Statistic:
x −μ
 n
27468-28000
z= = −2.496
1348 40
Inference:
The critical value of Z at 1% level of significance in a two
tailed test is -2.33
Here, Zcal < Zcritical value. So, we reject H0.
2) According to the norms established for a mechanical aptitude test ,

persons who are 18 years old should average 73.2 with a standard
deviation of 8.6. If 45 random selected persons of that age averaged
76.7 . Test the null hypothesis μ=73.2 against the alternative
hypothesis μ> 73.2 at 0.01 level of significance.
Sol:-
H0:  =73.2
H1:  >73.2
Test Statistic:
x −μ
 n

76.7 - 73.2
z= = 2.73
8.6 45
Inference:
The critical value of Z at 1% level of significance in a right
tailed test is 2.33. Here, Zcal > Zcritical value. So, we reject H0.
35
3) A sample of 400 individuals is found to have a mean height of
67.47 inches. Is it reasonable to regard the sample drawn from the
large population with mean height 67.39 inches and standard
deviation of 1.3 inches. Test at 1% level of significance.
Sol:-
H0:  = 67.39
H1:   67.39
Test Statistic:
x −μ
 n

67.47 - 67.39
z= = 1.23
1.3 400
Inference:
tailed test is 2.58. Here, Zcal < Zcritical value. So, we accept H0.
4) In 64 randomly selected hours of production , the mean and

standard deviation of the number of acceptable pieces produced by
automatic stamping machine are x = 1038, s = 146 . At 0.05 level of
significance does this enable us to reject the null hypothesis
μ=1000 against the alternative hypothesis μ>1000.
Sol:-
H0:  = 1000
H1:  > 1000
Test Statistic:
36
x −μ
 n

1038 - 1000
z= = 2.08
146 64
Inference:
Here, Zcal > Zcritical value. So, we reject H0.
5) A sample of heights of 6400 persons has a mean of 65 inches.

Could it be regarded as having been drawn from the population of
Indians with a mean of 66 inches and a standard deviation of 7
inches?.
Sol:-
H0:  = 66
H1:   66
Test Statistic:
x −μ
 n

65 - 66
z= = −11.4286
7 6400
Inference:
tailed test is 1.96
Here, Zcal > Zcritical value. So, we reject H0. Hence we conclude
that the sample does not belong to Indian population.
37
The mean weekly sales of soap bars in departmental stores was 146.3 bars per store. After
an advertising campaign, the mean weekly sales in 22 stores for a typical week increased
to 153.7 and showed a standard deviation of 17.2. Was the advertisement campaign
successful? Test at 1% level of significance
6) A sample of 100 measurement of breaking strength of cotton

threads gave a mean of 7.5 oz and standard deviation of 1.2 oz.
Find a 95% confidence limits for the mean breaking strength.
Sol:- The 100(1-)% confidence limits for the population mean are:
 
[ x − Z critical value .
, x + Z critical value . ]
n n
1.2 1.2
[7.5 − 1.96 ,7.5 + 1.96 ]
100 100 = [ 7.2648,7.7352]
7) If the average age at death of 64 men engaged in an occupation is

52.4 years with a standard deviation of 10.2 years, obtain 95%
confidence interval for the mean age of all men in the population,
assuming normality.
Sol:- The 100(1-)% confidence limits for the population mean are:
 
[ x − Z critical value .
, x + Z critical value . ]
n n
10.2 10.2
[52.4 − 1.96 ,52.4 + 1.96 ] = [ 49.9,54.9]
64 64
8) A random sample of size 100 is taken from a population having the

mean μ=76 and variance σ2=256. What is the probability that x
will be between 75 and 78?
Sol:-
38
x −μ x −μ
P Z =?
 σ/ n σ/ n 
 75 - 76 78 - 76 
P Z =?
16/ 100 16 100 
P− 0.625  Z  1.25 = 0.6268
9) If a one-gallon can of paint covers on the average 513.3 sq.ft. with

a standard deviation of 31.5 sq.ft. What is the probability that the
sample mean area covered by a sample of 40 of these one-gallon
cans will be anywhere from 510 and 520?
Sol:-
x −μ x −μ 
P Z =?
 σ/ n σ/ n 
 513.3 - 510 513.3 - 520 
P Z =?
 13.5 40 13.5 40 
P− 0.6626  Z  1.3452 = 0.6533
Test for equality of Two means :

H0: x = y
H1: x  y (or) x > y (or) x < y
39
Test Statistic:
(x - y) − (μ x − μ y )
Z= ~ N(0,1) at α% level of significance
2
σx
2
σ
+ y
n1 n2
Inference:
1) If it is a two tailed test (), Reject H0 if Zcal  Zcritical value.
2) If it is a right tailed test (>), Reject H0 if Zcal  Zcritical value.
3) If it is a left tailed test (<), Reject H0 if Zcal  Zcritical value.

 x2  y  x2  y
2 2
[( x - y) − Zcritical value . + , ( x - y) + Zcritical value . +

n1 n 2 n1 n 2
Problems related to test for equality of two means.
1) A company claims that its light bulbs are superior to those of its
main competitor. If a study showed that n1=40 of its bulbs has a
mean life time of 647 hours with a standard deviation of 27 hours,
While a sample of n2=40bulbs made by its main competitor had a
mean lifetime of 638hours with a standard deviation of 31 hours,
does this substantiate the claim at 0.05 level of significance.?
H0: x = y
H1: x > y
40
Test Statistic:
(x - y) − (μ x − μ y )
2
σx σ2
+ y
n1 n2
Under H0, the test statistic becomes,
(647 - 638) − (0)
Z= = 1.3846
2 2
27 31
+
40 40
Inference:
Here, Zcal < Zcritical value. So, we accept H0. Hence we conclude
that there no significant differ between the mean life of two bulbs.
2) A college conducts both day and night classes intended to be

equally effective. A sample of 100 day-students yields examination
results as x 1 = 72.4, S1 = 14.8 . A sample of 200 night-students
yields examination results as x 2 = 73.9, S 2 = 17.9 . Are two means
statistically equal at 10% significance level?
H0: x = y
H1: x  y
41
Test Statistic:
(x - y) − (μ x − μ y )
2
σx
2
σ
+ y
n1 n2
(72.4 - 73.9) − (0)
Z= = −0.7702
14.82 17.9 2
+
100 200
Inference:
Here, Zcal < Zcritical value. So, we accept H0. Hence we conclude
that there no significant differ between the means.
3) To test the claim that the resistance of electric wire can be reduced
by more than 0.050 ohm by alloying. 32 values obtained for a
standard wire yielded x 1 = 0.136 ohm, S1 = 0.004 ohm and 32 values
obtained for alloyed wire yielded x 2 = 0.083 ohm, S 2 = 0.005 ohm .
Does this data support the claim at 0.05 level of significance ?
H0: x - y ≤ 0.050 ohm
H1: x - y > 0.050 ohm
Test Statistic:
42
(x - y) − (μ x − μ y )
2
σx
2
σ
+ y
n1 n2
(0.136 - 0.083) − (0.050)
Z= = 2.65
0.004 2 0.0052
+
32 32
Inference:
Here, Zcal > Zcritical value. So, we reject H0. Hence we conclude
that x - y > 0.050 ohm
.
4) The means of two large sample of sizes 1000 and 2000 are 67.5
and 68 respectively. Test the equality of means of the two
populations each with S.D 2.5 at 5% level of significance.
H0: x = y
H1: x  y
Test Statistic:
43
(x - y) − (μ x − μ y )
2
σx
2
σ
+ y
n1 n2
(67.5 - 68) − (0)
Z= = −5.164
2.52 2.52
+
1000 2000
Inference:
tailed test is 1.96
that the two means are not equal.
5) The mean yield of two sets and their variability are given below.
Test whether the difference in the mean yields of two sets is
significant at 5% level of significance.
Set I Set II
Mean yield 1258 kgs 1243 Kgs
S.D/plot 34 Kgs 28 Kgs
No. of Plots 40 60
H0: x = y
H1: x  y
44
Test Statistic:
(x - y) − (μ x − μ y )
2
σx
2
σ
+ y
n1 n2
(1258 - 1243) − (0)
Z= = 2.3154
34 2 282
+
40 60
Inference:
tailed test is 1.96
that the mean yield of the two sets are not equal.
6) A sample of heights of 6400 soldiers has a mean height of 67.25

inches and with a S.D of 2.56 inches. While a simple sample of
heights of 1600 sailors has a mean of 68.55 inches and a S.D of
2.52 inches. Does the data indicate that the sailors on the taller than
the soldiers?
H0: x = y
H1: x < y
Test Statistic:
45
(x - y) − (μ x − μ y )
2
σx
2
σ
+ y
n1 n2
(67.85 - 68.55) − (0)
Z= = −9.905
2.56 2 2.52 2
+
6400 1600
Inference:
The critical value of Z at 5% level of significance in a left
tailed test is -1.645
Here, Zcal < Zcritical value. So, reject H0. Hence we conclude
that the sailors are on the average taller than the soldiers.
2-test
Test for 2 goodness of fit:

H0: Factors in the contingency table are independent
H1: Factors in the contingency table are dependent.
Test Statistic:
χ =
2 (O − E )2 ~ χ (r2 −1)(c−1) at α% level of significance
E
46
O is the observed value in the (i, j) th cell
where i th row total  j th column total
E=
samplesize
Inference:
If 2 cal  2(r-1)(c-1) , we reject H0. Otherwise we accept H0.
1) In a large manufacturing factory, a survey was conducted regarding

three types of bonus schemes. Total employees were divided into
four categories namely laborers, clerks, technicians and executives.
The results obtained by way of opinion survey are presented in the
form of contingency table as given below. Test the good ness of fit
at 5% level of significance.
EMPLOYEES BONUS SCHEMES

CATEGORY
Type 1 Type 2 Type 3
Labour 190 243 197
Clerks 82 44 44
Technicians 23 78 34
Executives 5 12 8
47
Sol:- H0: Factors in the contingency table are independent.
Appropriate level of significance is 5%(chosen)

BONUS SCHEMES
EMPLOYEES Type 1 Type 2 Type 3
CATEGORY TOTAL
Labour 190 243 197 630

Expected 196.9 247.4 185.7
Count
Clerks 82 44 44 170
Expected 53.1 66.8 50.1
Count
Technicians 23 78 34 135
Expected 42.2 53.0 39.8
Count
Executives 5 12 8 25
7.8 9.8 7.4
Expected
Count
Total 300 377 283 960
The calculated value of χ2=48.101
The table value of χ20.05,6=12.59
χ2cal > χ2tab , H0 is rejected.
2) To determine whether there is really a relationship between

employee’s performance in the company’s training program and his
success in the job, a sample of 400 cases were taken and the
following results were obtained. Test at 1% l.o.s whether the
performance in the training program and success in the job are
independent and the table is as given below.
Performance in training program

Below Avg. Above
Avg. Avg.
Poor 23 60 29
Avg 28 79 60
Very good 9 49 63
48
H1: Factors in the contingency table are dependent
Appropriate level of significance is 1%(given)
Performance in training program Total

Success Below Avg. Above
in job Avg. Avg.
Poor 23 60 29 112
Expected 16.8 52.6 42.6
Count
Avg. 28 79 60 167
Expected 25.0 78.5 63.5
Count
Very good 9 49 63 121
Expected 18.2 56.9 46.0
Count
Total 60 188 152 400
3) 1072 college students were classified according to their intelligence and

economic conditions . Test whether intelligence is independent of
economic condition.
INTELLIGENCE
ECONOMIC
CONDITION
Excellent Good Mediocre Dull
Good 48 199 181 82
Not good 81 185 190 106

49
Appropriate level of significance is 5%(chosen)
INTELLIGENCE TOTAL
ECONOMIC Excellent Good Mediocre Dull
CONDITION
Good 48 199 181 82 510

Expected 61.4 182.7 176.5 89.4
Count
Not good 81 185 190 106 562
Expected 67.6 201.3 194.5 98.6
Count
Total 129 384 371 188 1072

1 Fit a Binomial distribution to the given data and test the goodness of fit.
X 0 1 2 3 4 5 6
F 13 25 52 58 32 16 4
Sol:-
X F fx Px=ncxpxqn-x N.Px
0 13 0 0.02897 5.79 6
1 25 25 0.13984 27.97 28
2 52 104 0.28121 56.24 56
3 58 174 0.30161 60.32 60
4 32 128 0.18196 36.39 36
5 16 80 0.05855 11.71 12
6 4 24 0.00785 1.57 2
200 535
50
Mean = np = Σfx /Σf = 535/200=2.675
P= 2.675/6= 0.4458 , q=1-p=0.5542
The Calculated value of χ2= 15.548

χ2cal > χ2tab , H0 is rejected. Data does not fit B.D
4. Fit a Binomial distribution to the given data and test the goodness of fit.
X 0 1 2 3 4 5 6
F 7 64 140 210 132 75 12
Sol:-
X f fx Px=ncxpxqn-x N.Px
0 7 0 0.014270 9.13 9
1 64 64 0.088230 56.47 56
2 140 280 0.227294 145.47 145
3 210 630 0.312289 199.87 200
4 132 528 0.241350 154.46 154
5 75 375 0.099480 63.67 64
6 12 72 0.017085 10.93 11
640 1949
Mean = np = Σfx /Σf = 1949/640=3.0453
P= 3.0453/6= 0.5075 , q=1-p=0.4925
The Calculated value of χ2=7.607
χ2cal < χ2tab , H0 is accepted. Data fits B.D
Fit a P.D to the following data and test the goodness of fit.
X 0 1 2 3 4 5 6 7 8
F 103 143 98 42 8 4 2 0 0
Sol:-
X f Fx Px=e-λλx /x! N.Px Fe
0 103 0 0.26647 106.59 107
1 143 143 0.35240 140.96 141
2 98 196 0.23303 93.21 93
3 42 126 0.10273 41.09 41
4 8 32 0.03396 13.59 14
5 4 20 0.00898 3.59 4
51
6 2 12 0.00198 0.79 1
7 0 0 0.00037 0.15 0
8 0 0 0.00006 0.02 0
400 529
Mean = λ = Σfx /Σf = 529/400=1.3225
χ cal < χ2tab , H0 is accepted. Data fits the P.D
2
Fit a P. D to the following data and test the goodness of fit.

X 0 1 2 3 4
F 419 352 154 56 19
Sol:-
X F Fx Px=e-λλx /x! N.Px
0 419 0 0.404947 404.95 405
1 352 352 0.366072 366.07 366
2 154 308 0.165464 165.46 165
3 56 168 0.049860 49.86 50
4 19 76 0.011268 11.27 11
1000 904
Mean = λ = Σfx /Σf = 904/1000=0.904

χ2cal < χ2tab , H0 is accepted. Data follows the Poisson law.
Time: 45 Mts. ASSIGNMENT TEST –II Max. Marks: 12

1
1. Evaluate  e − x dx by Simpson’s 1/3 rd rule, by taking 10 subintervals
2
dy
2. Using R-K method of order 4, find y(0.2) from 10 = x 2 + y 2 ,y(0) =1,
dx
3. Using Euler’s method solve for y at x = 0.1 from dy/dx = x + y + xy, y(0) = 1 with h
= 0.025.
4. In a sample of 20 the mean number of defectives is 2.Out of 1000 such samples, how
52
many would be expected to contain at least 3 defectives?(Assume Binomial
distribution).
5. In a normal distribution 31% of the items are under 45 and 8% are over 64.
Find the mean and standard deviation of the distribution.
6. A random sample of 10 boys had the following IQs 70,120,110, 101, 88, 83, 95,
98,107,100.
Do these data support the assumption of a population mean IQ of 100?
53

Toh

Uploaded by

Copyright:

Available Formats

Toh

Uploaded by

Document Information

Original Description:

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

Toh

Uploaded by

Copyright:

Available Formats

TESTING OF HYPOTHESIS

Population: Any set of objects under study in a statistical

Sample: A finite subset of the population is known as sample and the

Sampling error: The error involved in approximation about the

Parameter: Any statistical constant related to the population is

Statistic: Any statistical constant related to the sample is known as

Tests of significance: A very important aspect of the sampling theory

Null hypothesis: For applying the test of significance, we first set up

Degrees of freedom: Degrees of freedom is the number of

Distinguish between type I and Type II errors in testing the

H0 False Correct Wrong

Type I error: Probability of rejecting H0 when H0 is true.

Level of significance() The probability of type I error is known as

Point estimation: If from the sample information available, we can

Interval estimation: If from the sample information available, we can

Procedure for testing the hypothesis.

Critical values Z Level of significance ()

Confident Limits: The 100(1-)% confident limits for the

1) If it is a two tailed test ():

Problems related to test for single mean.

1) (Dec 2004): Ten individuals are chosen from a normal population

Under H0, the test statistic is given by

Under H0, the test statistic is given by

3) A sample of nine items had the following values

Under H0, the test statistic is given by

4) 5 measurements of tar content of certain kind of cigarette yielded

Under H0, the test statistic is given by

5) The specifications for a certain kind of ribbon has a mean breaking

Given x = 169.5, s = 5.7, n = 5

Under H0, the test statistic is given by

6) Random sample of 6 steel beams has a mean breaking strength of

Given x = 58,392, s = 648, n = 6

Under H0, the test statistic is given by

Here, tcal < t/2,n-1 . So, we accept H0.

7) A random sample of 8 envelopes is taken from a letter box of a post

The 100(1-)% confident limits for the population mean are

8) 10 bearings made by a certain process have a mean diameter of

The 100(1-)% confident limits for the population mean are

The 100(1-)% confident limits for the population mean are

Test for equality of Two means (n1n2) :

3). In a left tailed test (<):

Problems related to test for equality of two means.

(8230 − 7940) - (0)

2) The heights of 6 randomly chosen sailors are 63,65,68,69,71,72

Under H0, the test statistic is given by

(68 − 67.8) - (0)

(58 - 51.83) - (0)

Given, n1 = 8, n 2 = 7, x = 1234, y = 1036, s x = 36, s y = 40

Under H0, the test statistic is given by

(1234 - 1036) - (0)

Given, n1 = 10, n 2 = 12, x = 15, y = 14

Under H0, the test statistic is given by

(15 - 14) - (0)

Test for equality of Two means (n1= n2) (Paired t-test)

Confident Limits: The 100(1-)% confident limits for the

1) If it is a two tailed test ():

Problems on paired t-test

2) To compare two kinds of bumper guards 6 of each kind were

di -27 33 11 14 -31 -10