You are on page 1of 3

Nonparametric- 340

. Tutorial No 8 . ,
Problem I: A random sample of 300 customers was selected in a marketing study. Each
person tasted 5 different brands of soft drinks and asked to specify his preferred brand. The
results are given below. Test the null hypothesis that Brand D and E equally share 52% of the
market while Brand A,B and C equa 11ly Share 48%00 f t he rna rket Find p-value ?
i
Brand Number

A (I) 50

B (2) 65 i

C (3) 45
i

D (4) 70

L
I

E(5) 70

Solution: Ho: PI=0.16, P2=0.16, P3=0.16, p4=0.26, Ps=0.26 vs. HI: Not Ho

Brand A B C D E

OJ 50 65 45 70 70
L
Ej 48 48 48 78 78 i

I
(OJ-Ej)LIE\ 0.083 6.02 0.188 0.821 0.821
I
Since observed X 2 =7.933 < X(~05 4 = 9.488, we accept Ho.

P-value P( X2 > 7.933) = .094

Problem II: The table below shows the number of males in the first seven children born to
l334 Swedish ministers of religion. The data were reported by Edwards and Fraccaro (El3).
Test the hypothesis that the number of males is binomially distributed with p = 0.51.

Number of males in first seven births of children born to a sample of 1334 Swedish ministers
of religion
-_.
! # males 0 I 2 3 :4 5 6 7
1
1

#01' families in : 6 57 206 362 365 . 256 1


69 13
! which that i I
i i
I number occur~_1
I I I

Solution: Ho: data are 8(7, 0.5') versus HI: data are not 8(7, 0.51)
-_ ..

I # males 10 1 2 3 14 5 16 7
I
1 i

i OJ 16 57 206 362 365 256 69 13


i

Pi . 0.0068 0.0494 0.1543 0.2676 0.2786 0.1740 I 0.0604 0.0090


i

i Ei 1334*P i i 9.07 65.9 205.84 356.98 371.65 232.12 80.57 112.01

(0 ­ Et/E I

I
1.0391 11.202
.
.0001 1.0706 1. 119 2.4567
1.6615 .L::J
Since observed X =6.631 < X~057
2
11.071, we accept Ho.

Problem III: A random sample of size 10 resulted in

~_14061461 .1474 .1338 1457 1469 1 498


Test the null hypothesis that the data follows a Normal population.
Solution:

X=450.4 S2 = 6823.59 then S = 82.61

x Z=(X­ Fo(x) Sex) lFo(x)-S(x)1


450.4)/82.61
1 303 -1.78 .0375 0.1 .0625

I 2 338 -1.36 .087 0.2 .113


I

3 406 - .54 .2946 0.3 .0059


I
4 457 .08 .5319 0.4 .1319
I
5 461 .13 .5510 0.5 .0517
6 469 I .225 .5909 0.6 .0091
I
I 7 474 .286 .6122 0.7 .0878

1 8 498 .576 .7174 0.8 .0826


9 515 .782 .782 0.9 .1177
10 583 1.605 .9458 1.0 .0542 i

1 - - . -..

Kolmogorov Distance 0=.132 and cut-of point =d(0.05, 10)=0.409 . Since observed 0=.132
< 0.409, we accept Ho: data follow normal distribution.

Problem IV: Two methods of performing an assembly-line task were compared by having a
sample of 10 workers perform the task by method A and an independent sample of workers
perform the task by method B. The times in seconds required to complete the task are shown
in the following table. Do these data provide sufficient evidence to indicate a difference in
population distributions? Use a 0.05 level of significance.
Times required to perform a task by two methods

lMethod A I 25 22 17 . 16 19 . 24 i 26 21 19 27
Method B ! 21 22 14 13 22 22 15 21 23 21

Method A Met~SI(X) S2(X)

~
I IS1-S21
i
I 13 0 1110 1110
i

i
14 0 2/10 2110
I
15 0 3/10 3/10

I 16 1110 3110 2110

~ 17 2/10 3110 1110

19(2) 4110 3110 1110

I 21 i 21(3) 5110 6110 1110


i II
I 22 22(3) 6110 9110 3/10
i
I
i
23 6110 10110 4110(max)
i

24 I 7110 10110 3/10


i

25 8/10 10110 2110

26 9110 10110 1110


I
27 10110 10110 0
I

D 4/10
d(.05,10,1O) =.7
Since D < d(.05,10,10)
We accept Ho
This means that the two pop's have same distribution.

You might also like