samplingMethods

theobservedsample.If this probability is less than acertain assigned value, the hypothesis is1645
to be
ejected.

6211 NULL HYPOTHESIS (H,)

Nullhypothesisis based for analysing the problem. Null hypothesis isthe hypothesis of no
iference.Thus. we shall persume that there is no significant difference between the observed
alueandexpected|value. Then, we shall test whether this hypothesis is satisfied by the data or
Ifthe hypothesiS 1S not approved the difference is considered to be significant. If hypothesis
isapprovedthen the difference would be described as due to sampling fluctuation. Null hypothesis
H,.
sdenotedby
62,12ERRORS
K nsampling theory to draw valid inferences about the population parameter on the basis of the
sarmpleresults.
We decide to accept or to reject the lot after examining a sample from it. As such, we are liable
(oCommit the following two types of errors.
Type 1Error. If H, is rejected while it should have been accepted.
Type II Error. If H,is accepted while it should have been rejected.
62:13 LEVEL OF SIGNIFICANCE
There are two critical regions which cover 5% and 1% areas of the normal curve. The shaded
nortions are the criticalregions.
Area of Area of
Acceptance Acceptance
2.5% 2.5%
0.5% 0.5%

Z=-0.674 Z= 0.674
Z=1.966 Z=1.966
Thus, the probability of the value of the variate falling in the critical region is the level of
significance. If the variate falls in the critical area, the hypothesis is to be rejected.
62.14 TEST OF SIGNIFICANCE
The tests which enables us to decide whether to accept or to reject the null hypothesis is
called
population values are
the tests of significance. If the difference between the sample values and the
so large (lies in critical area), it is to be rejected
62.15 CONFIDENCE LIMITS
l- 1.96 o,l+ 1.96g are 95% confidence limits as the area
between l-1.96 o and + 1.96 o is 95%. If asample statistics lies in -1.960 +1.96o

the interval u-1.96 o, l+ 1.96 o, we call 95% confidence interval.

Similarly. ! -2.58 o, + 2.58 o is99% çonfidence limits as the
area between l- 2.58 G and L+ 2.58 o is 99%. The numbers 1.96. -2.586 +2.586
2.58 are called confidence çoefficients.
62.16 TEST OF SIGNIFICANCE OF LARGE SAMPLES (N > 30)
Normal distribution is the limiting case of Binomial distribution when n is large enough. For
of the items lie outside
normal distribution 5% of the items lie outside l 1.96 G while only 1%
H2.586 o.
 sarmplingMethods
1649
THEt-DISTRIBUTION
20 (FOR SMALL SAMPLE)
students distribution is used to
) The mean of a small sample. test
The the
significance of
() The Aiference between the mneans oftwo small samples or to compare two small samples.
a The correlation coefficient,
the members of random
T be the mean'ofthe sample drawn fromn anormal population with
sample then

where
n-1

Example 6. A machine which produces mica

insulating washers for use in electric device to
turn out washers having a thickness of 10 mm. A sample of 10 washers has an
average thickness 9.52 mm with astandard deviation of0.6 mm. Find out t.
Salution. X =9.52, M= 10, S' = 0.6, n = 10
-M 9.52-10 0.48/10 4
0.6 0.6
Vn
=0.8x 3.16 =2.528 Ans.
2.21 WORKING RULE
To calculate significance of sample mean at 5% level.

Calculate t = n and compare it to the value of t with (n- 1) degrees of freedom at 5%

level, obtained from the table. Let this tabulated value of t be t,.
Ifr<4,then we accept thehypothesis ie, we say that the sample is drawnfromthe population.
Ift> 4, we compare it with the tabulated value of t at 1% level of significance for (n - 1)
degrees of freedom.Denote it by t, If, <t<t, then we say that the value of tis significant.
Ift> t, we reject the hypothesis and the sample is not drawn from the population.
Exampl7. A maufacturer intends that his electric bulbs have a life of 1000 hours. He
tests a sample of 20 bulbs, drawnat random from a batch and discovers that
the mean life of the sample bulbs is 990 hours with a S.D of 22 hours. Does
this signify that the batch is not up to the standard?
[Given:The table value oft at 19% level is significance with 19 degrees offreedom is 2.539]
Solution. T=990, o = 22; x=1000
990-1000 10V20 22.36
= -2.033
22 22 11
N20 o 2.033 2.539
Since the calculated value of t(2.032) is less than the value of t (2.539) firom the table. Hence.
S not correct to say that this batch is not upto this standard. Ans.
Bxample 8. Ten individuals are chosen at random from a population and their heights are
found to bein inches 63, 63, 64, 65, 66, 69, 69, 70, 70, 71. Discuss the suggestion
that the Mean height ofuniverse is 65.
2.262.
or 9 degree of freedomt at 5% level of significance =
 One-Tailed and Two-Tailed Tests
Two-Tailed Test
Testing to see if your calculated value is either above or below where
it's expected to be.

One-Tailed Test
Testing onlyto see if your calculated value isabove where it's
expected to be.

OR

Testing only to see if your calculated value is below where it's

expected to be.
 Rajshree
PAGE NO.
DATE

to cheek
Cangigsiample
iSeple belonys to popid SE.
Variane not -kmoesn Z-4
S.E

Taso Sampe t- test

j Usgure t-test (Indepedest Sample t-tet
Variene korsun
butwesn tuo empla talen

df Mitn-2

+ +S

Piud t- fes (Dependeut samale t-tet)

Vaiane knoun anol Same
Cenparison bkueen teuo san
Sanples deaon rem
he Same fter)

t )
 Eram

1650 Higher Engineering Mathematics

Solution.
*-67 (x-67) 670
=67,
63 10
-4 16 Solu

63 -4 16
88
64 -3 9 =3.13
S =
n-1 9
65 -2 4
1 67-65 14
66 -1 2V10
=

4 S 3.13 -2.02 13
69 +2 3.13
69 +2 4 J10 12
9
13
70 +3 2.02 < 2.262
12
70 +3 9 13
Region of
71 +4 16 acceptance 14
X(*-)' =88 2.02 2.262
Zx=670 the table value of 1(2.262). The hypothesis ie
12
Calculated value of t (2.02) is less than
65 inches. Ans.
accepted the mean height of universe is bulbs produced hy a
Example 9. The mean life time of sample ofhours l00 fuores cent light hoNe Exa
with a standard deviation of 20
1
company is computed to be 1570 bulbs producedby it is 1600
The company claims that the average life of the claim acceptable?
hours. Using the level of significance of0. 05, is the
Solution. I=1570, s = 120, n = 100, u = 1600
L=_
1570]600 1570-1600
= 2.5
S 120 12 So

100
At 0.05 the level of significance, t = 1.96
Calculated value of t > Table value of t.
2.5 > 1.96
Ans.
Hence the claim is to be rejected.
Example 10.A sample of6 persons in an office revealed an average daily smoking of 10, 12,8,
9, 16, 5 cigarettes. The average level of smoking in the whole office has to be
estimated at 90% level of confidence.
I= 2.015for 5 degree offreedom
Solution.
Efd
X10 (x-10) Mean =at =10+10
6
10 22
|70
12 2 4 S= =3.74
n-1
8 -2 4
At 90% level of confidence, t = +2.015.
9 -1 1
16 +6 36 +2.015 = 10
- 3.74
-5 25

Total 0 X«-10) - 70 3.74 +10=6.92.13.08 Ans.

=2.015x:
 Methods

Eramplee11. Afertiliser mixing machine is set to give 12

Ten 100 kg bags are kg of nitrate for 1651
examined The
percentages fnitrate quintal bag
1. 14, 13, 12, 13, 12, 13, 14,
11. 12 ofi per bag are as offertiliser:
Is there any reason to
believe that the follows:
degrees of freedom is 2.262. nachine defective?
is
Value of t for 9
Solution. The calculation of x and s
is given in the following table:
d=x-12 d2
-1 1 = 12 kg, n =10.F=L* 125
14 2 4 10 -=12.5
1
13 13 2

12 0
10
13 1 21 105
1 10
13 1 10 4 20
100
12
13 1 1
Value oftfor 9 degrees S=1.024
of freedom
14 2 4
2.262
-1 Also t= 12.5-12
12
Since 1.024 -J10 =1.54
the
Sr=125 d=s a'=13 to believevalue oft is less than
that machine is 2.262,
defective. there in no reason
Example 12.A random sample of size 16 Ans.
of 53 and asun of values from a normal population
this sample be squares of deviation from the mean eguals showed a mean
regarded
Obtain 95% and 99% as taken from the population to 150. Can
confidence limits of the mean of the having 56 as mean?
Y=15, a=0.05, t = 2.131 population.
a =0.01, t=2.947
Solution.
=56, n= 16, x =
53,>(*-} = 150

X(-'
n-1
150 =10 Region of
15 rejection
S=10
t-u 53-56 -3x4 3.79 -2.131 O
=-3.79
V10

When a = 0.5 then 3.79 t=3.79

When a, = 0.01 then 3.79>>2.131
2.947
Ihus, the sample cannot be regarded as taken
from the population. Ans.
62.22 TESTING FOR DIFFERENCE BETWEEN MEANS OF TWO SMALL SAMPLES
ne mean and variance of the first population be u, and o.' and #,. o, be the mean and
Variance of the second population.
L be the mean of small sample of size n, from first population and , the meal o
dnple of size n, from second
We know that population.
2

)=4and Var ()=L

n
 Higher Engineering Malthematics
1652
Var (5,)= 9
and
EZ,)=} 4 also independent.
and (x, ) are
are independent, then (x,)
Ifthe samples 4.
2

(a,-,)
andVar
E(x, -*,) =4 -4y +
then n

(%-x)-(H - )

n,

the same then

If the population is
t=
1+ 1 (H, -; =-H =0)

of 8and 7 items respectively had the followino wet

13. Two independent samplesounces):
Example variable (weight in
of the 12 14
Sample l: 9 1113 11 15 9
10
Sample 2: 1012 1014 98
means of thesample significant?
Is the difference between the
[Given for V= 13, ,o=2.16]
Solution.
ofy = 10
Assumed mean of x= 12, Assumed mean
(*-12)² (-10) (-10
(r-12)
-3 10
9 4
1 12 2
13 1 10
1 14 4 16
-1
15 9 -1
-2 4
-3 9 8
12 10 0
14 2 4
94 -2 3 25
34 73
-

x= Z:94
n8
= 11.75

2_2(x-12)²
o,=
n
2(-13n
34
= 4.1875

V=73=10.43
n 7

2 X(y-10) E(y-10)
n

S=, (x-+oy
n +n, -2
34+ 25 59
-=4.54 = 2.13
V8+7-2 V13
 Methods

11.75-10.43 1.32 1653

1 1 1.32
1
2.13Jo.268 1.32

Thus,5%
less
value

than 2. 1l6, the

n,
of t for 13 degree of2.13|*7 2.13x0.518 =
freedom is given to be 2.16. Since
difference between the means of
1.103 =1.2

calculated value of tis

2is
samples is not significant.
random sample of six steel beams
EXERCISE 62.1 Ans.
has mean
A
square inch) with a standard deviation compressive
of s strength of 58.392 psi (pounds per
= 648 psi.
H,-u=58,000 psi against the alternative Test the null
(valuefor tat 5 degree of freedom and 5%
hypothesis H,; W> 58,000 psi at 5% level of hypothesis
mean.
(A.M.I.E.,
significance
Summer 2000)
level is 2.0157). Here
denotes the
significance
A
certain cubical the was thrown 96
times and population
1 shows 2 upwards 184 times. Is
sample of 100
the the biased?
i In a residents of a colony 60 are Ans.
articles are equally popular?found to be wheat eaters and 40 rice
assume that both food die is biased.

4. Out of 400 children, 150 are found to be

eaters. Can we
estimate the percentage of under
children who are weight. Assuming the conditions of
per-centage probably lies. underweight in, and assign limits simple sampling,
within which the
5 so0eggs are taken at Ans. 37.5% approx.
random from
and 50 are found to beLimits 37.5+3(2.4)
a large
orcentage of bad eggs in the consignment,
consignment and assign limits within which the bad. Estimate the
6 A machine puts out 16 imprefect Ans.
percentage
10%, 10 ±3.9
probably lies.
3 imprefect articles in a
articles in a batch of 100. Has the sample of 500. After the machine is
machine been improved? repaired, puts out
1. In a city A, 20% of Ans.
another city B, 18.5% of
a random
sample of 900 school boys had The machine has not been improved.
a certain slight
a random physical defect. In
between the proportions sample of 1600 school boys had the
significant? same defect. Is the difference
B. In two large Ans. z =0.37, Difference between
populations there are 30% and 25% respectivelyof fair
likely to be hidden
proportions is significant.
in samples of 1200and haired people. s this difference
900 respectively from the
9. two populations?
One thousand articles Ans, z = 2.5, not hidden at S% level of
significance.
from a factory are
hundred similar articles from a second examined and found to be three percent defective. Fifteen
concluded that the product of the firstfactory are found to be only 2 percent defective. Can it reasonably be
factory is inferior to the second?
Ans. It cannot be reasonable concluded that the product of
the first factory is inferior to that of the second.
10. A
manufacturing
show a tolerance ofcompany
about 26% of his better
claims 90% assurance that
than 5%. The capacitors are
the capacitors manufactured by them will
packaged and sold in lots of 10. Show that
52.23 CHI customers ought to complain that capacitors do not reach the specified standard.

The SQUAREdistribution
Chi-square TEST is one of the most extensively used distribution tunction in statistics.
It was
first
discovered
by Helmert in 1875 and later on Karl Pearson's in 1900.
 Higher Engineering Mathematics san

1654
62.24Ihe
CHI-SQUARE VARIATES
square of a standard normal yariateis known as Chi-square variate %) with one degree

function:
Z= is a nomal
variate.

variate.
Hence is a Chi-square
be arandom sample of
Ifx bea normaly
distributedvariate andx,, X,
has X dist1bution with n-degree of fre
in-values
from
w=x +x5 +...+x
this populationthen 62
TEST
CONDITIONS FOR CHI-SQUARE
62.25 necessary for ChÉ square test.
conditions which are
There are some total of cell frequency should not be
sample under study must be large and may be less
1. The
than 50.
2. The member of the
cells should be independent.
cellfrequency of each cell should be greater than 3. If any cell has frequency lees at
3 The
5then it should be combined withthe next or preceding cell until the total frequency exceeds
5.
frequencies they should be linear
4. If there are any constraint on the cell +
i.e.; a, x,t a, x,ta, X, t..... t a,x,+ trigonometric relation
Note: Cell frequency should not involve any logarithmic, exponentialor
62.26 CHI-SQUARE (y2) IS USED AS:
(1) Test of independence (2) Test of goodness of fit
(3) To test if the hypothetical value of the population variate is o
(4) To test the homogeneity of independent estimate of the population variance.
We shall mainly use the first two test
62.27 CHI-SQUARE TEST OF GOODNESS OF FIT
This test is used to test significance of the discrepancy between theory and experiment.
helps us to find if the deviation of the experiment from the theory is just by chance or it is due to
the inadequacy of the theory to fit the observed data.
The theoretical frequencies for various classes are calculated from the assumption 0T o
population. The significant deviation between the observed and theoretical frequencies is tested by
means of this test.
x is calculated by means of the following formula
(0,-E,)?
i=] E, and Z0, ZE, =N
where O, is the observed frequency E, is the expected (Theoretical) frequency of the cell.
62.28 WORKING RULE TO
CALCULATE y2:
Step 1. Calculate the expected
Step 2. Calculate the difference frequencies.
between each observed frequency 0, and the corresponding
expected frequency E, for each class i.e.; to find 0,- E,
 Methods

step3.
Squarethe difference obtained
in step 2 for each
value i.e., Calculate (O, -1655
Step4. Divide (O,- E)' by the expected frequency Eto get (o- E, E}.
E

Step5.
these quotients obtained in step 4.
Add all Z (0,- E,}
Then y2i=1
noted
Itis to
be E,
(1) The value of y is always positive. (2)) y' will be zero if each pair is zero.
(3) The value of x* lies between 0 and
DEGREE OF FREEDOM
CaseI. If the data is given in the form of a
degree of freedom = (No. of items inseries of
the series)variables
- 1
in a row or
column then the
2When the number of
frequencies are put
The degree of freedom = (R 1) (C- 1)in cells in a contingency table.
swhere R is number of rows
andCis the
Erample 14. A survey of 320 families with 5 number of columns.
No. of boys 5
children is given below :
4 2
0 Total
No. of girls 1 2 3 4
No. of families 14 S6 110 88 40
Is this result 12 320
consistent
equally possible.
with hypothesis i.e.; the male
and female birth are
Solution.
Null Hypothesis H
(1) Male and Female birth are
Alternate Hypothesis H,: Male and female equally probable.
birth are not equally probable.
Calculation of expected frequencies (g + p)
Probability of female birth =

Probability of male birth q= 1

2
(9tp)" = q + "C, p + "C, p² -2 +"C, p t ... tp'

B-@9GG-GG-GG-¬
No. of girls = 320
|32
1 5
32
10 -+10 +
32 32 32 32
5 10 10 5
=320 x +320 x +320 x +320 x
+320 x +320 x 32
32 32 32 32
= 10 + 50 + 100 + 100 + 50 + 10
These are the expected frequencies of the female births.
 Higher Engineering Matthematics
1656
(0- E) (0- E)?
O- E E
E
16 1.60
4
10 36 0.72
14 6
50 100 1.00
56 10
100 144 1.44
110 - 12
100 100 2.00
88 10
40 50 4 0.40
2
12 10 7.16
Total

a =0.05
Level of significance Let
Critical value. The table value of y at a =0.05 for (6-- 1) (2- 1) =5 degree of freedom
is 11.07
of y (7.16) < Table value of y at level of
Decision Since the calculated value
significance 0.05 for 5df= 11.07
female birth is equallyprobak,
Hence, the null hypothesis is accepted i.e., the male and
Example 15. The table below give the number of air craft accidents that occured during the
various days of the week. Test whether the accidents are uniformly distribt
Over the week.
Mon. Tue. Wed. Thu. Fri Sat Sun Total no. accidents
Days
18 12 15 14 14 98
No. of accidents 14
Solution. H, : Null Hypothesis : The accidents are uniformly distributed over the week
98
=14
The expected frequencies of the accidents on each day =

0- E (0- E)'
(0- E)'
E
14 14 0 0
18 14 4 16 1.14
12 14 2 4 0.29
11 14 0.64
15 14 0.07
14 14 0
14 14
98
Total 2.14
Levelof significance :
Let a = 0.05
Critical value: The table vaue of y at a = 0.05 is for (1-
y= 12.592 1) (2 - 1) i.e.; o dv
Since the calculated value of x (7.16) <Table value of x at level of significance 0.05for
six degree = 12.592
Hence, the null Hypothesis is accepted i.e.; distributed
over the week. the air craft accident are uniformly Ans.