Chapter 4:statustucal Inferen ce I

iampling: An aqqreg ate oF obj ects under stud y Ls

atled.population or unverse

subset of a uLniverse is called a sample.

sampleis a sma portion of the universe. The no
tindiviau0n aa sample
sample is called sample size

rocess ot SeleTung a sample from a unvers e

aledsampling
ghe

,T1ypes oF Sampling:
Simple Random sample: Jt is a subset of a
y in which each member of
statisti cal population in
ahstt has
Subset an equal probability of beino chosen
meant to be an unbias ed representation
ASRS Ls
of a gz20p.
2) Stratified sampling dn statistics , stratified
sarnpling s a method of sampling from a populatn.
dividinq the membeks
QtratifLCotisn s the pro Cess of
before
cf a population into homoqeneD us subg roups
cungling GrpsAS S.R.S

Sample
Grp3 s R.S S.R.S

3) Systematic Sampling: t is a stotistical method of:

involyung theselectisn of elem ents from an
Dndered sampin frame
4) Cluster sam pling: gt is a sam plung plan usod
mutuay homogeneous yet ntern o
when are evident in a Statlstico,
heteroqeneous qro upung
poplation gt Ls Dften used
in
marketlnq esearch
Sn this sam plunq plan, total pop laticn IS davis
into subqro ups Ga simple random m group of
Sample Of 9roups is selected.

Sample

statistics :
o Parameters of
ghe statistical constants of the populati on such as
mean; variance
Variance 4 proportion are known as
Parameters.
Stan dard Error Cs.E).
he stan d ar d deviation of the Sampling distr. of
a statistic.
oF Larg e samples
3t plays an ump role in th cory of
Lt forms a basis for testng of hypothes is
n- Sample siRe P- population
variance of population proportuon
- mean of the population 9=1-P
g'- variance of the sanm ple psam ple
- mean of the sample proportion
size of the popul ation
Np 9=1-p
 statistic. S.E

n.
P
oufferenceof
2 means:

pifferenceof
)4)2 propOvtions:

PPa
dLata.Cx.%n)
Varuance for the raw
(x -a)
V:
Nean Variance =1
Mean = XË
data:
Variance for grouped
Mean= Zix Variance =If;(%-)
=
N
Ef, =N
set of sample of certain sre 'n.
Suppose we draw each.
popuulartor we tind the mean x of
Cor this thLs mean s
lhe frea. distr. Of
nf th ese SampleS Suppose. the
sam plung digtr. of m
caLled infinite. with
with sL2e Np
without
poput ation (s related to u & o
n & 5 are
replacement , then
the fouOwing for mula:
through.
S
= x Np-n
Vn. Np-1
 the popul ation S unfinite or
Sample ls flnite.
Tcplacement then,
with
Vn.
Innp A population contauns 4 units: 37, |,15.
the saMpina distr. of sZe 2 when
Samples ae dro
with.neplacement & without replacement.
Sbtaun
Np = 4
n = 2

5ample without replace ment i Mean = = ?

(3,7) c3, |1) (3, 15) S.D = S =?
C11|) (7, 15)
Cl, 15)
Find u & (ie for Po

Mean for the poputation

x, == I(3+1+ 11+15) = q

i(3-9)+ (7-4)+ (l|-9)+ (|5-4)3

g= 20
i is the mid value of ea ch sample.
i : 5, 1, 9 , 9, 11, 13
Mean= i A= 1(5+1+9 + 9 +11+ 13)
n 6

C1l-q)2 + C13 -4)1

S= 2.58

Verficationi u = 9

.47 xX 4-2 = 2.58

= 4.47
-1. V 4-1
Np
amplewith

(3,77)
replacem ent
3)
7) (3,11) (3, 15)
3)
(7,
(1,11) (1,15)
3)(11,7)
(11,11) C1,15)
(U5,7) (15,11) 45,15)
Nidvalue.

3
fi Mean- = Sfi
=1
1
3
3x1+ 1x3+ 11x3
11 +15x1+5x2t qx44
15 1

S= H.H1 3.16
2 =
13
44-H w.

W:
fopln 5, 4,7 n=.

Mean S.a Ot the pop ulation

u=2Xi =L5 +4+7] ’=5.33.
H

e1Z(-)=(5-5.33)+(-5.33)+(7-5331
n
R
T
=1.556

.u5.33 G =1.241
Sinple ran dom sanple
without epiacerm ent
(5, 4) (5, 1) (4, 1)
G 5.5
[4.5 +6+ 5. 5] = 5.33
3
2
Verity
with
S'=4 (H.5-g.33)*+ (6-5.33)+ (5.5-3-333 f01mula
=0.3g89 ’S = 0.6236
 |3-2
Np -n .211 V3-1

(55) (5, 4) (5,1) 5 1

(4,5) (H,4) CH1)
(1y5) (1, 4) (7,1)
4.5 2

5.5

X 5.5X27
x1 + 4 X1 4 7X1 + H,5X 2 +6
|5

}5.33

s= 1N f; (x;-)" = 1f(5-5.33)XI + (4-533) x1

9 (7-5.33)x 1+ (45-5.33)x2+
(6-5.33)x2 + (5.5- 5. 33)x24
S =0. 8818 (Fiz ca) c to
244 1 paces
2 - chetk)
(L.0.S)
o evel of sianificance

2..57
2.5 O.0025
(reje ctiorn raqion

ran dom vae of the statistic. T,

The
the region. 1S known as. Level Of
belong to
Sugnuficance
 ntdcnce nterVal (CI)
estimn aleen whch
whuch qwes the
LO hich the
parameter of theendpoints
populatcn ues
of the
I
for mean
populaticn () is gven by
(5.E)) ’ (7+ Zay, ()
when s
S unknown
For the proPorubn OF the population
p*
(P Zae (5.E)) ’ (p± ) (p*zp(Psn

P= p when P LS unknown

dLfference ot means:
I for
+92 )

Tfor difference of proportions (p, -p, )

((Pi-Pe) Pi+P292

Mazimum eror to estmate is qiven by :

2
E= ’ n=(,
NOTE: For 457. CI, =1.9
q9/. CI
-|.96 1.96
=1.6H5

1P 96) n=?
E = |0000 at q5/ .confidence
30000
 at q57 con fidence =1.4G
(I96 3000Oy = 3 4. 574 35
1000o

E = n- (Z4s at confdence 952

Z, -|.4G
2
n (.96x_2 62
0.5

KP CS) n= 1o0 )
P= 0.91 (given)
’ = 0.09

C.I for proportion 9 5). 49/. ?

0.91x0.09
=1.4G (0.9|± 1.9G
(o.91t(0.56 1))
(O.91-0.05 61 0.9l+ 0.056) )
Co.539 ) O.9661)
At 957. proportion of the popuLation lies between
(o-8539, o.96GI)

At q9. confud ence ) R, =2.58 -

0. 91 t 2.5
(° o.4l x0. 09

(o.91 t0.o73 8)
(0.363 0.9838)

At q97 proportion of the poPulation çes between

(0.83G3, O.9838)
 500
500

preportion of the population:

er

= 2.05 2.35

O. 12 X SS
o1 2 2.05
5O0
0.0666 )
(o. 12+o.0299)
(o. o 402, 0. I498)
= 2.5
At99 )
2.58(02x08%
(o.12 V:
0375)
12 t0.
(0.
(o.0S2, 0157)

S=2.42 n=100.

mean of the popul atuon'

( Ifor
(E)
At 957. =1.96

(67. H541.96 (i (67.45 t 0. 5723)

=
(G1.H5t1.96 (o.292))
= (G6.811,68.022) 1:5

= 2.1 2.11

(G1.45 t 2.)7 C0.292))

(61.45t 0.6336)
(G6SIGH, G6 -0S 36)
above tincS
(11- tdishh. - s all sanple
 n= 275 = n, =n
0.4 |1
P=5S9
92 = -p = 0. 3s1
differen ce of portLon :

fo the
Cs.E))
(C P, -P2)t
))
=|.96 (-o.I O.o4)
At q 5/
cal cul alions

PSI4) n,=R50 X = I20 S |2

L) S. E for the dufference O+ means s

S. E = s,2 + S =1.o325

n2
means. t 9 2.
y SE CI for the difFerence of
confid.ence:
= 2.5
(S. E) = (CI 20- |2 4)t 2.58 (1-0325)
=(-6 G638, - . 336 2)

Suoport
newspaper is condu cting a surVey. 50
e AAuv ersity Ln fees
yeQY Ln crease
of the students, a Rs. 200 per
stadium. Hou bua
to pay for a new fo otbal
Sampe is requred to ensure the n arqin 0f error
 than O. O4
lcss Using 45/. Confid en ce ?
E<0.oH
E is for
Maximum or for the proportuon
propor tion
E=

95% =1.96 ) P= 0.5 ) q = o. 5

At
(V) < 0.04

L.46(JD. 5X013 <o.04

0.5
1.96

1.46. (5x0.5) < n.

No.5 xo-5) <n
GO0 25 < n nn G01

Null hypothesis:
definite statement about the population para meter.
A
Denoted by H

ALternate hypoth esis (H,)

is compim entary to the
Ary hypothesis which ls
l l hypoth esis.

We reyect a hy pothesis when Lt shoçd

.Type- I error:
is accepted)
be accepted. (True one s
a type I error has occurred.
Then we ay that
Type - TL erYOY: We accept a hypothesis when it
Should berejected. ( False. -I K 2ccepted)
hen we Say type Il erroy has occurY ed,
Critucal region
Aregion in the sample Space S whicn amon
rejection of Ho s termed oLS CYLtical
rejection. rcaon of regor oF
Yeect3n.

Two tarl
9ne-toul.