CHAPTE k 06

STATISTICAL INFERENCE IL Page No. 123

Data. 30lo5l25

Chi- Squane Test for Goodness of Fit -

’ I t is a non-pavametie te st used to determine whetber cn
cbserved frequency dstibution of single oatsgotcal sooable
Grfers signifcantly from an ekpected theonet a distibutton,.
Null Hypothesis CHo): The observed froquercies mateh the
expeced disteibution.
Y|Atermate typsthesis CHa) : The observed feauencies do net mades
the expected dstibution.
Chi-Squaved Statistics
(Lth degee of treedonn
n-1)
O Observed Freaueng.
E Expected Frequency.
’ The deg nee of treedom is n
where n’ no. of observations
k no af constrajnts. ahen fre- is legs
k=1+nm+ Pooling Frequency umbex
Poolin Eequeney mbext s pocling
is Considered.
m ’No. of Pas ameter
(Binomial Poisson 4 Noml2)

daits o.1,.9 were 23 25, 20, 23, 23, 22, 29,2C, 33, 27. Test
te hypothests thadt they ere randomly drawn.
S9Ho Digits vandonaly draon.
Ha Digt wert not Yandomly drawn.
sSince they were raOdenaly draon, eepeced regucnsy of each.
Aeepdante

eeotion
 Pogo No. 124
Date.

Digats Observed Ertquc niag Ecpecked Ereq uaney o,-E)

(o to 9) (E;)
23 2S b16
4 2 2
20 2 1
23 0.44
4 23 0.46
5 22 2 O.36
29 20 O. 64

33 2-56
9 22 2 616
250 250 x'=5.2
’ =5.2 ’ Calculcted Cht-squar
df =n-1 10- =9
Astune LOS as S ,
. Soos (9) 16.919
oay <rl9) Aept Ho Rand om Drstnbton,
Lea) 0A Cansputer ha 6 I'o chann bls, Out of 15D T/O operectos
observed no- of operattont Loere 22,23, 29 31,26,19. Tet
the hypothests that the Ioad on channels is balanced
S Ho' The load on the channels balanced
# The load on the channels is unbalanced
Expe cked Fréguencg=
Frequentd 190 251
Channel lo,-6)/E; de=6-4 =5
22 25 tr(5) 14.07
23 25 7t4.68
29 25 64
3 31 25 {.44 fo
4 26 25 O.04
19 25 4.44 Load is balknced
150 150 =4.0Z
 Page No. 125
Date.

3 it a binoniad distabutton &test for goodness of At

No of heads 01 2 4 56
26 24 63 64 36 10 1
Freauenay N-254

P() n

Esgected Foeueny = NXx P(2) = 256xIGa (Yasc) =|C

No- of heads o-E)/E,
2 1
6 (Poo le&)
2 24 28 0.54
63 56
4 64 to
5 56 O. 64
6 36 28 2-2R
044
8
CPooled)
2S% 256 7=5.42
51tal 542
df = n-t: n-(1twmtp) = g- (t142) 9-45.
Accept Hol ) Binonsial Distibution.
4Ft a Poisson Distibution a test for goodness of L
2 4 5
29542 30 S 21N=
2 S92
S Mean = fa/N ==189/392 ’ A=0.48. pCK) :=Ae
Expeced Freq. E; =NXPK) = 392(04 Co.64)
E; 239.12 (0.43)/
X 34S
25 2 S 2 1 tast 3 erties Pooled
E 242.1 146.3 28.1 44 0.53 0.05 0.004
|-E)/E:4-42 18|0428 19-21 74o,99
=-(141+3)=2 Sau=5.391 (0t D3or(2)l Not a
PoissonDtbttn
 Pags No. 126
Dete 31 0525

6 fest the kugste'ts Hhat no. o gorts demanded does not

On the dy of the oeek Ho :Not depend on eek
S n:G E= 1692o/s 2236.6 dfS
Dous Monday Tuerdoy Mednery ThusdayFuday Setoday Sum
4424 1125114011120 1426 112 6 114
111 S 16720
L9669 28664 243664 48664 245669 236-64 1620
99203 990.34 1008 31 24920-2 933-64 1002.3 n295041
-Xss(s) 11.0?
Chi-Squaved Test for Independence of Attibutes
t is a statstical test used to deterine hothor there s
Stquificavt associcton between 2 categoical variables
>A Corttngency toble is a type af tabe used in stotistics
to shou te fréquency distibut on oe vaviables. It &
erpecially useful for 2 attyorical vetables to Cxansi'e
elatons hips between tham
’Nll tlyothesis Tuso categonicad vadnble are indagendeat
Alternode tHypthat's' Too Catugo cad vaiable ave dependernt
5 Test Statstics
X2-(O;-E) O: obserued Pirequency
E Expecded Erequeny
ohereE;= (Peco Tot) x (Golumo Totad)
Tota Frequenay
Degrees oP freedom
Af r1) (c-1)ater Poobaq' df= df - Pocled anbt
G Tet ohether the attbutet cpamm b wth imagel ave depe ndet
indeperdert of each otber
Spanm Iwith Iaggs Withet Inaagg Sum
Span 460 240 400
No Span 140 460 600
Sann 300 To00
 Page No. 124
Date.

sY He The 2 vaiables being spam wo'ith imanes rt independet

factors Hai Not ndependent.
’ (-1)(e-s) (2-1) (2-1) =1
Expected Freq uency E= (Roo tofal ) x(Col. Total) /Total Froq
EC460) > 300 X4Ïo 4000 120 At S- Los,
E (240) a toox400 1000 = 280
E(440) Boo x6 00/1 00o 180.
E(460) oo x 6 oo /1000 = 420.
Attabuteg (o,-E)3/E:
Spam Inagcs 160 120 13.33
No Tmag. 240 280 5.4
No Sparm Tmag 140 180

NoSumInagts 460 420 3 81

1000 1000 31.3
Da dacrl1) Reject Ho >They_are defendent.
2) Reseavcher Belauo Aveva_e Aveag Above Aver agg Gentus
X 86 44
40 33 25 2
Test ohetrer the Smeling
smplng teohnques adoptd aue dfkerent
S> Ho Dota obBained at independent of the sanmpllng techuiques
adeped by the 2 eseaychers,, H1: DaBa obtaneis dependeat
’d= (T-1)(C-1) = (2-1) (4-1) = 3
86 60 44 10 200
40 33 2 400
Totel 126 93 69 42 300
’EC36) =(Q00)126) /300 = 34
E (60) =(200)(93)/300 c62 E(40) 42, E(33) 31
E(44) =(200) (69) /B00 46 E(2r) 2?
Af39 we can caludate only 3 valueg of E & the gegt
Can be caleudated by the ditfevence.
able vales
 Paga No. 128
Date.

Attibuty (a,-E)/E:
Reloo A4 86 84 0046

Ava 60 62 0.064
Above Ar 44 46
Genjus 10
Belouw Avg 40 42 0-0952
Avg 31 0.129
25 23
Geius 2 4 Pooled
Total 300 ScO2 32 300
’ =3-(Pooled Cels)= 3-12 (After Poofng)
6a2)= 5.991
nee c zaasl2) Acept Ho Thdeperndest Techniques
) 8l Test cohethey preference of mode 's ndependent of student
ocalrt
Modeocalit! Local HosteleG Tota
Onine 160 140 4o 340
6Pne 40 66 160
totl 200 2o0 100 So0
S df-C2-1) (3-1) = 2
EC160)= 200 XK 340
-t36
EC40) = 64
E(540) 200x 340 E(Go) G4
EC80) 32
Attibates E(O,-¬)/E
Ondine Locel 160 436 4.24
os tel 140 136 0.12
PG 40 68
12.53 sreedk H
offltne Locap
HHorte! 60 64 O.25
They
dependet
PG 60 32 24
Tofo' 500
Nog 43.84
 Page No. 129
Date.

) Oest the hypothests Poisson Drstibutton.

2 4 6
40 13 4 1 N=1oo
3s
SMean =
N 100
EAe -N 11) (o.32) X100 33(1.41)

Tnjuies E
33 0-12
1 40 3663 O.31
13 2o.33 2-64 ’Xas(3) = 4815
t-s2
4 42 2.08) Peoruled
2.62 4-36 Poisson Distibtian
Pooled
Tofe 100 100

9 (10) Agply Chi-Squared Tet to these, data to test Poiscon Dsfibation

No. ce defects 2

LObserved Fequency 32 4 N60

SMean = 4S

ENP CC) = 60 x(0.3)* (o.42) 282 (o- 5)

befecs Co-E)/E,
22 28.2 0.S1
21.1 1.49
2
7-93 O.96
1.98 Pooled
Tote 63.26
d n-= 4-(4 t1f 1) 1 ) 8o.0 (2) = 3.841
Acept Hol PoisSon Digtibutton ft
 Pago No. 12o
Dats.

L6) 65) osed on the t obserrattor s a binomaial distibton o

Pkpgsiate model 9
Value 1 2
Ereg utncy3923 1 2 1 N=S
SMean 0-66
.64/,
’ Expe cked Freauencyr
np.
D-221 ZK
Sudhanv Kulkarni

Valuer o,-E)E
39 25.59 0-32 64
23 30.12 1.63 31
2 12 1.4912
41 0.49 Pooled
Total
dP =n-b = n-(1tmtp =4-(4 t1t1) =1
57,sC (1) 3.941
X'ea Cn'aner (1) Accept Ha ) Binomdal Dstibuten