BES Tutorial Sample Solutions, S1/13: WEEK 13 TUTORIAL EXERCISES (To Be Discussed in The Week Starting June 3)
BES Tutorial Sample Solutions, S1/13: WEEK 13 TUTORIAL EXERCISES (To Be Discussed in The Week Starting June 3)
E
0
: o
2
= 1Su
2
; E
1
: o
2
< 1Su
2
UnderE
0
(n -1)s
2
o
0
2
=
4u s
2
1Su
2
~_
41-1
2
Decisionrule:
RejectE
0
ifcolculotcJ _
2
< _
0.95,40
2
= 26.Su9S
Since calculated
_
2
=
4u 14u
2
1Su
2
= S4.84 > _
0.95,40
2
= 26.Su9S
we do not reject E
0
and conclude that the manufacturers claim of lower bulb
life variability is not supported by the data.
2. uiven the infoimation of Question 1 above, calculate a 9S% confiuence
inteival foi the vaiiance of bulb life.
Sincen=41and=0.05wehave:
2
_u
2
,n-1
2
= _
0.025,40
2
= S9.S417
_
1-
u
2
,n-1
2
= _
0.975,40
2
= 24.4SSu
Theconfidenceintervalisthengivenby
_
(n - 1)s
2
_
u
2
,n-1
2
,
(n -1)s
2
_
1-
u
2
,n-1
2
_ or _
4u 14u
2
S9.S417
,
4u 14u
2
24.4SSu
_ or (1S,211.6, S2,u87.6)
S. In a busy maiketing consulting fiim, Su% of the employees aie ageu
2u29 yeais, 4u% aie ageu SuS9 yeais, 2u% aie ageu 4u49 yeais, anu
1u% aie ageu Su yeais anu ovei. The peisonnel managei of the fiim is
inteiesteu in knowing whethei age is a factoi in the taking of sick uays.
She takes a ianuom sample of 1uu sick uays anu finus the following
uistiibution of sick uays by age gioup:
Age No. of Sick Bays
2u29 yeais
SuS9 yeais
4u49 yeais
Su yeais anu ovei
26
S7
24
1S
0sing these uata, test at the = u.uS level of significance the hypothesis
that age is not a factoi in taking sick uays.
E
0
: Agc is not o octor in tbc toking o sick Joys
E
A
: Agc is o octor in tbc toking o sick Joys
If age is not a factor in taking sick days, then the expected proportion of the sick
days taken by a particular age group should be the same as the proportion of the
firms employees in that age group. We would have:
3
Age oi ei
2u29 yeais
SuS9 yeais
4u49 yeais
Su yeais anu ovei
26
S7
24
1S
Su
4u
2u
1u
1uu 1uu
_
2
=
(o
-c
)
2
c
c
=1
=
(26 -Su)
2
Su
+
(S7 - 4u)
2
4u
+
(24 -2u)
2
2u
+
(1S -1u)
2
1u
= 2.46
c1=41=S anu hence the ciitical value is _
0.05,3
2
= 7.8147
Since_
2
< _
0.05,3
2
wedonotrejectthenullhypothesisthatageisnotafactor
inthetakingofsickdays.
4
4. A public health ieseaichei is inteiesteu in testing whethei the level of
beei consumption of inuiviuuals is inuepenuent of theii level of
cigaiette consumption. A ianuom sample of 1,uuu inuiviuuals aie
inteivieweu about theii weekly consumption of beei anu cigaiettes,
yieluing the following bivaiiate fiequency uistiibution:
uiven the uata, peifoim a test of inuepenuence of beei anu cigaiette
consumption using a S% significance level.
E
0
: Bccr onJ cigorcttc consumption orc inJcpcnJcnt
E
A
: Bccr onJ cigorcttc consumption orc not inJcpcnJcnt
The expected cell frequencies are given in brackets in the following table.
Beei consumption
nil 1S
glasses
41u
glasses
0vei 1u
glasses
Row
total
Cig.
cons.
nil 2S6
(182.962)
68
(1u4.78)
S4
(7S.749)
2S
(41.Su9)
4uS
12u 1uS
(1S6.176)
11u
(89.44)
87
(62.9S2)
42
(SS.4S2)
S44
216u 71
(8u.812)
S6
(46.28)
2S
(S2.S74)
26
(18.SS4)
178
>6u 22
(S4.uS)
26
(19.S)
17
(1S.72S)
1u
(7.72S)
7S
Col.
Total
4S4 26u 18S 1uS 1uuu
Beei consumption
nil 1S
glasses
41u
glasses
0vei 1u
glasses
Cigaiette
cons.
nil 2S6 68 S4 2S
12u 1uS 11u 87 42
216u 71 S6 2S 26
>6u 22 26 17 1u
5
e.g.
4uS 4S4
1uuu
= 182.962
As v = (r -1)(c -1) = S S = 9
_
0.05,v
2
= _
0.05,9
2
= 16.919u
Now the test statistic is given by
_
2
=
(o
]
-c
]
)
2
c
]
c
]=1
=1
=
(2S6 -182.962)
2
182.962
++
(1u -7.72S)
2
17.72S
= 1u1.9u
And since _
2
> _
0.05,9
2
we reject E
0
and conclude that beer and cigarette
consumptionarenotindependent.