Escolar Documentos
Profissional Documentos
Cultura Documentos
BUSINESS
STATISTICS
by
AMIR D. ACZEL
&
JAYAVEL SOUNDERPANDIAN
7th edition.
Prepared by Lloyd Jaisingh, Morehead State
University
Chapter 7
Hypothesis Testing
McGraw-Hill/Irwin
7-2
7 Hypothesis Testing
Using Statistics
The Concept of Hypothesis Testing
Computing the p-value
The Hypothesis Test
Pre-Test Decisions
7-3
LEARNING OBJECTIVES
7-4
A hypothesis is a statement or assertion about the state of nature (about the true
value of an unknown population parameter):
A hypothesis is either true or false, and you may fail to reject it or you may
reject it on the basis of information:
7-5
Decision-Making
Correct
Incorrect
7-6
nullhypothesis,
hypothesis,denoted
denotedby
byHH0,0,isisan
anassertion
assertionabout
aboutone
oneorormore
morepopulation
population
AAnull
parameters. This
Thisisisthe
theassertion
assertionwe
wehold
holdtotobe
betrue
trueuntil
untilwe
wehave
havesufficient
sufficient
parameters.
statisticalevidence
evidencetotoconclude
concludeotherwise.
otherwise.
statistical
H0: = 100
H
0: = 100
Thealternative
alternativehypothesis,
hypothesis,denoted
denotedby
byHH1,1,isisthe
theassertion
assertionofofall
allsituations
situationsnot
not
The
coveredby
bythe
thenull
nullhypothesis.
hypothesis.
covered
H1: 100
H
1: 100
andHH1are:
are:
HH00and
1
Mutually
Mutuallyexclusive
exclusive
Onlyone
onecan
canbe
betrue.
true.
Only
Exhaustive
Exhaustive
Togetherthey
theycover
coverall
allpossibilities,
possibilities,so
soone
oneor
orthe
theother
othermust
mustbe
be
Together
true.
true.
7-7
H0: p 40%
H1: p < 40%
H0:
H1:
7-8
Thenull
nullhypothesis:
hypothesis:
The
Oftenrepresents
representsthe
thestatus
statusquo
quosituation
situationor
oran
anexisting
existingbelief.
belief.
Often
Is maintained, or held to be true, until a test leads to its rejection
Is maintained, or held to be true, until a test leads to its rejection
favorof
ofthe
thealternative
alternativehypothesis.
hypothesis.
ininfavor
Isaccepted
acceptedas
astrue
trueor
orrejected
rejectedas
asfalse
falseon
onthe
thebasis
basisof
ofaa
Is
considerationof
ofaatest
teststatistic.
statistic.
statistic
consideration
statistic
7-9
teststatistic
statisticisisaasample
samplestatistic
statisticcomputed
computedfrom
fromsample
sampledata.
data. The
Thevalue
valueof
of
AAtest
thetest
teststatistic
statisticisisused
usedinindetermining
determiningwhether
whetherorornot
notwe
wemay
mayreject
rejectthe
thenull
null
the
hypothesis.
hypothesis.
Thedecision
decisionrule
ruleofofaastatistical
statisticalhypothesis
hypothesistest
testisisaarule
rulethat
thatspecifies
specifiesthe
the
The
conditionsunder
underwhich
whichthe
thenull
nullhypothesis
hypothesismay
maybe
berejected.
rejected.
conditions
ConsiderHH00::=
=100
100.. We
Wemay
mayhave
haveaadecision
decisionrule
rulethat
thatsays:
says:Reject
Reject
Consider
thesample
samplemean
meanisisless
lessthan
than95
95or
ormore
morethan
than105.
105.
HH00ifif the
Inaacourtroom
courtroomwe
wemay
maysay:
say:The
Theaccused
accusedisisinnocent
innocentuntil
untilproven
proven
In
guiltybeyond
beyondaareasonable
reasonabledoubt.
doubt.
guilty
7-10
Decision Making
There are
are two
two possible
possible states
states of
of nature:
nature:
There
true
HH00isistrue
false
HH00isisfalse
There are
are two
two possible
possible decisions:
decisions:
There
Fail to
to reject
reject H
H00 as
as true
true
Fail
Reject H
H00 as
as false
false
Reject
7-11
Decision Making
decision may
may be
be correct
correct in
in two
two ways:
ways:
AAdecision
Fail to
to reject
reject aa true
true H
H00
Fail
Reject aa false
false H
H00
Reject
decision may
may be
be incorrect
incorrect in
in two
two ways:
ways:
AAdecision
Type II Error:
Error: Reject
Reject aa true
true H
H00
Type
The Probability
Probability of
of aaType
Type II error
error isis denoted
denoted
The
by .
.
by
Type IIII Error:
Error: Fail
Fail to
to reject
reject aa false
false H
H00
Type
The Probability
Probability of
of aaType
Type IIII error
error isis denoted
denoted
The
by .
.
by
7-12
decisionmay
maybe
beincorrect
incorrectin
intwo
twoways:
ways:
AAdecision
TypeIIError:
Error:Reject
Rejectaatrue
trueHH00
Type
TheProbability
ProbabilityofofaaType
TypeI Ierror
errorisisdenoted
denotedby
by.
.
The
calledthe
thelevel
levelofofsignificance
significanceofofthe
thetest
test
isiscalled
TypeIIIIError:
Error:Accept
Acceptaafalse
falseHH00
Type
TheProbability
ProbabilityofofaaType
TypeIIIIerror
errorisisdenoted
denotedby
by..
The
calledthe
thepower
powerofofthe
thetest.
test.
11- -isiscalled
andare
areconditional
conditionalprobabilities:
probabilities:
and
= P(Reject H 0 H 0 is true)
= P(Accept H 0 H 0 is false)
7-13
7-14
The p-Value
Thep-value
p-valueisisthe
theprobability
probabilityof
ofobtaining
obtainingaavalue
valueof
ofthe
thetest
teststatistic
statisticas
as
The
extremeas,
as,or
ormore
moreextreme
extremethan,
than,the
theactual
actualvalue
valueobtained,
obtained,when
whenthe
thenull
null
extreme
hypothesisisistrue.
true.
hypothesis
Thep-value
p-valueisisthe
thesmallest
smallestlevel
levelof
ofsignificance,
significance,,
,atatwhich
whichthe
thenull
null
The
hypothesismay
maybe
berejected
rejectedusing
usingthe
theobtained
obtainedvalue
valueof
ofthe
thetest
teststatistic.
statistic.
hypothesis
RULE:When
Whenthe
thep-value
p-valueisisless
lessthan
than, ,reject
rejectHH0. .
RULE:
0
7-15
The power
power of
of aa statistical
statistical hypothesis
hypothesis test
test isis the
the
The
probability of
of rejecting
rejecting the
the null
null hypothesis
hypothesis when
when the
the
probability
null hypothesis
hypothesis isis false.
false.
null
Power== (1
(1 -- )
)
Power
7-16
0.1261
0.1261
0.2695
0.2695
0.4423
0.4423
0.6387
0.6387
0.8037
0.8037
0.9123
0.9123
0.9682
0.9682
0.9908
0.9908
0.9972
0.9972
1.0
0.9
Power
61
61
62
62
63
63
64
64
65
65
66
66
67
67
68
68
69
69
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
60
61
62
63
64
65
66
67
68
69
70
7-17
The power depends on the distance between the value of the parameter under
the null hypothesis and the true value of the parameter in question: the greater
this distance, the greater the power.
The power depends on the population standard deviation: the smaller the
population standard deviation, the greater the power.
The power depends on the sample size used: the larger the sample, the greater
the power.
The power depends on the level of significance of the test: the smaller the level
of significance,, the smaller the power.
7-18
Example
companythat
thatdelivers
deliverspackages
packageswithin
withinaalarge
largemetropolitan
metropolitan
AAcompany
areaclaims
claimsthat
thatitittakes
takesan
anaverage
averageof
of28
28minutes
minutesfor
foraapackage
packageto
to
area
bedelivered
deliveredfrom
fromyour
yourdoor
doorto
tothe
thedestination.
destination. Suppose
Supposethat
thatyou
you
be
wantto
tocarry
carryout
outaahypothesis
hypothesistest
testof
ofthis
thisclaim.
claim.
want
Set the null and alternative hypotheses:
H0: = 28
H1: 28
Collect sample data:
n = 100
x = 31.5
s=5
Construct a 95% confidence interval for
the average delivery times of all packages:
x z
. 025
s
5
315
. 196
.
n
100
315
. .98 30.52, 32.48
7-19
7-20
Example
Anautomatic
automaticbottling
bottlingmachine
machinefills
fillscola
colainto
intotwo
twoliter
liter(2000
(2000cc)
cc)bottles.
bottles. AAconsumer
consumeradvocate
advocatewants
wants
An
testthe
thenull
nullhypothesis
hypothesisthat
thatthe
theaverage
averageamount
amountfilled
filledby
bythe
themachine
machineinto
intoaabottle
bottleisisatatleast
least2000
2000
tototest
cc. AArandom
randomsample
sampleofof40
40bottles
bottlescoming
comingout
outofofthe
themachine
machinewas
wasselected
selectedand
andthe
theexact
exactcontent
contentofof
cc.
theselected
selectedbottles
bottlesare
arerecorded.
recorded. The
Thesample
samplemean
meanwas
was1999.6
1999.6cc.
cc. The
Thepopulation
populationstandard
standard
the
deviationisisknown
knownfrom
frompast
pastexperience
experiencetotobe
be1.30
1.30cc.
cc.
deviation
Computethe
thep-value
p-valuefor
forthis
thistest.
test.
Compute
2000
HH0:0:2000
2000
HH1:1:2000
40,0==2000,
2000,x-bar
x-bar==1999.6,
1999.6,
nn==40,
0
1.3
==1.3
Thetest
teststatistic
statisticis:
is:
The
x 0
x0 =1999.6
1999.6--2000
2000
x
z
0
z =
1.3
1.3
40
nn
40
1.95
==1.95
valueP(Z
P(Z--1.95)
1.95)
pp--value
0.5000--0.4744
0.4744
0.5000
0.0256
0.0256
7-21
7-22
7-23
actionisistotobe
betaken
takenififaaparameter
parameterisisless
lessthan
thansome
somevalue
valuea,a,then
thenthe
the
IfIfaction
alternativehypothesis
hypothesisisisthat
thatthe
theparameter
parameterisisless
lessthan
thana,a,and
andthe
thetest
testisisaa
alternative
left-tailedtest.
test.
50
left-tailed
HH0:0:50
50
HH1:1:50
actionisistotobe
betaken
takenififaaparameter
parameterisiseither
eithergreater
greaterthan
thanor
orless
lessthan
thansome
some
IfIfaction
valuea,a,then
thenthe
thealternative
alternativehypothesis
hypothesisisisthat
thatthe
theparameter
parameterisisnot
notequal
equaltotoa,a,
value
andthe
thetest
testisisaatwo-tailed
two-tailedtest.
test. HH:0:50
50
and
0
50
HH1:1:50
7-24
Let==5,5,==5%,
5%,and
andnn==100.
100. We
Wewish
wishtotocompute
computewhen
when==1==998.
998.
Let
1
Referto
tonext
nextslide
slide
Refer
Thefigure
figureshows
showsthe
thedistribution
distributionof
ofx-bar
x-barwhen
when==0==1000,
1000,and
andwhen
when
The
0
998.
==11==998.
Notethat
thatHH0will
willbe
berejected
rejectedwhen
whenx-bar
x-barisisless
lessthan
thanthe
thecritical
criticalvalue
valuegiven
given
Note
0
by(x-bar)
(x-bar)crit==0-z-z/n
/n==1000
10001.6455/
1.6455/100
100==999.18.
999.18.
by
crit
0
Conversely,HH0will
willnot
notbe
berejected
rejectedwhenever
wheneverx-bar
x-barisisgreater
greaterthan
than(x-bar)
(x-bar) crit. .
Conversely,
0
crit
7-25
Computing (continued)
7-26
Computing (continued)
When==1==998,
998,will
willbe
bethe
theprobability
probabilityof
ofnot
notrejecting
rejectingHH0which
which
When
1
0
impliesthat
thatP{(x-bar
P{(x-bar>>(x-bar)
(x-bar)crit}.}.
implies
crit
When==,1,x-bar
x-barwill
willfollow
followaanormal
normaldistribution
distributionwith
withmean
mean1and
and
When
1
1
standarddeviation
deviation==/n.
/n. Thus,
Thus,
standard
X crit 1
P Z
P( Z 1.18 / 0.5) P( Z 2.360)
/ n
0.0091
Thepower
powerof
ofthe
thetest
test==110.0091
0.0091==0.9909.
0.9909.
The
7-27
7-28
Tests
Testsof
ofhypotheses
hypothesesabout
aboutpopulation
populationproportions.
proportions.
Tests
Testsof
ofhypotheses
hypothesesabout
aboutpopulation
populationvariances.
variances.
7-29
isisknown
knownand
andthe
thesample
samplesize
sizeisisatatleast
least30.
30. (The
(Thepopulation
population
neednot
notbe
benormal)
normal)
need
The formula
formula for
forcalculatin
calculatingg ZZ isis::
The
x
x
zz
n
n
7-30
thepopulation
populationisisnormal.
normal.
the
The formula
formula for
forcalculatin
calculatinggttisis::
The
x
x
tt s
s
n
n
7-31
Rejection Region
7-32
Nonrejection Region
7-33
= 28
95% confidence
interval around
observed sample mean
30.52
x = 31.5
32.48
It seems reasonable to reject the null hypothesis, H 0: = 28, since the hypothesized
value lies outside the 95% confidence interval. If we are 95% sure that the
population mean is between 30.52 and 32.58 minutes, it is very unlikely that the
population mean will actually be 28 minutes.
Note that the population mean may be 28 (the null hypothesis might be true), but
then the observed sample mean, 31.5, would be a very unlikely occurrence. There
is still the small chance (= 0.05) that we might reject the true null hypothesis.
represents the level of significance of the test.
7-34
Nonrejection Region
If the observed sample mean falls within the nonrejection region, then you fail
to reject the null hypothesis as true. Construct a 95% nonrejection region
around the hypothesized population mean, and compare it with the 95%
confidence interval around the observed sample mean:
0 z.025
s
5
28 1.96
n
100
28.98 27,02 ,28.98
27.02
0=28
28.98
95% Confidence
Interval
around the
Sample Mean
30.52
x z .025
s
5
315
. 1.96
n
100
315
. .98 30.52 ,32.48
32.48
The nonrejection region and the confidence interval are the same width, but
centered on different points. In this instance, the nonrejection region does not
include the observed sample mean, and the confidence interval does not include
the hypothesized population mean.
7-35
.95
0.6
0.5
0.4
0.3
0.2
.025
.025
0.1
27.02
0=28
28.98
7-36
.95
0.6
0.5
0.4
0.3
0.2
.025
.025
0.1
0.0
27.02
0=28
28.98
x
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Do not reject H0 if the sample mean falls within the nonrejection region (between the
critical points).
Reject H0 if the sample mean falls outside the nonrejection region.
7-37
Example 7-5
An automatic bottling machine fills cola into two liter (2000 cc) bottles. A consumer advocate wants to test the
null hypothesis that the average amount filled by the machine into a bottle is at least 2000 cc. A random sample of
40 bottles coming out of the machine was selected and the exact content of the selected bottles are recorded. The
sample mean was 1999.6 cc. The population standard deviation is known from past experience to be 1.30 cc.
Test the null hypothesis at the 5% significance level.
2000
HH0:0:2000
2000
HH1:1:2000
40
nn==40
For==0.05,
0.05,the
thecritical
criticalvalue
value
For
ofzzisis-1.645
-1.645
of
Thetest
teststatistic
statisticis:
is: z
The
x 0
40
nn==40
1999.6
xx==1999.6
1.3
==1.3
1999.6--2000
2000
xx 0 1999.6
0
z
=
z
=
1.3
nn
1.3
40
40
Donot
notreject
rejectHH0if:
if:[z[z-1.645]
-1.645]
Do
0
RejectHH0if:
if:zz]
]
Reject
0
1.95
Reject
RejectHH
==1.95
00
7-38
2000
HH0:0:2000
2000
HH1:1:2000
40
nn==40
For==0.05,
0.05,the
thecritical
criticalvalue
value
For
ofzzisis-1.645
-1.645
of
Thetest
teststatistic
statisticis:
is:
The
x 0
Donot
notreject
rejectHH0if:
if:[p-value
[p-value]
]
Do
0
RejectHH0if:
if:p-value
p-value]
]
Reject
0
1999.6--2000
2000
xx 0 1999.6
0
z
=
z
=
1.3
nn
1.3
40
40
1.95
==1.95
valueP(Z
P(Z- -1.95)
1.95)
pp- -value
0.5000- -0.4744
0.4744
0.5000
0.0256
Reject
RejectHH since
since0.0256
0.02560.05
0.05
0.0256
0
0
7-39
n = 40
n = 20
7-40
7-41
Results when
is unknown
7-42
7-43
7-44
7-45
7-46
7-47
Because
Fromthe
thebinomial
binomialtables,
tables,with
withnn==25,
25,pp==0.5,
0.5,this
thisvalue
value
From
2*0.054==0.108.
0.108.
2*0.054
Since0.108
0.108>>==0.05,
0.05,then
then
Since
donot
notreject
rejectHH0
do
0
7-48
7-49
7-50
7-51
2
0
where isisthe
theclaimed
claimedvalue
valueof
ofthe
thepopulation
populationvariance
variancein
inthe
the
where
nullhypothesis.
hypothesis. The
Thedegrees
degreesof
offreedom
freedomfor
forthis
thischi-square
chi-square
null
randomvariable
variableisis(n
(n1).
1).
random
2
Note:Since
Sincethe
thechi-square
chi-squaretable
tableonly
onlyprovides
providesthe
thecritical
criticalvalues,
values,itit
Note:
cannotbe
beused
usedto
tocalculate
calculateexact
exactp-values.
p-values. As
Asin
inthe
thecase
caseof
ofthe
thet-tables,
t-tables,
cannot
onlyaarange
rangeof
ofpossible
possiblevalues
valuescan
canbe
beinferred.
inferred.
only
7-52
Example 7-8
manufacturerofofgolf
golfballs
ballsclaims
claimsthat
thatthey
theycontrol
controlthe
theweights
weightsof
ofthe
thegolf
golfballs
balls
AAmanufacturer
accuratelyso
sothat
thatthe
thevariance
varianceofofthe
theweights
weightsisisnot
notmore
morethan
than11mg
mg2.2. AArandom
randomsample
sample
accurately
31golf
golfballs
ballsyields
yieldsaasample
samplevariance
varianceof
of1.62
1.62mg
mg2.2. IsIsthat
thatsufficient
sufficientevidence
evidencetoto
ofof31
rejectthe
theclaim
claimatatan
anofof5%?
5%?
reject
Let22denote
denotethe
thepopulation
populationvariance.
variance. Then
Then
Let
2
1
HH0:0:2 1
2
HH1:1:2
Inthe
thetemplate
template(see
(seenext
nextslide),
slide),enter
enter31
31for
forthe
thesample
samplesize
size
In
and1.62
1.62for
forthe
thesample
samplevariance.
variance.Enter
Enterthe
thehypothesized
hypothesizedvalue
value
and
of11inincell
cellD11.
D11. The
Thep-value
p-valueof
of0.0173
0.0173appears
appearsinincell
cellE13.
E13. Since
Since
of
thisvalue
valueisisless
lessthan
thanthe
theof
of5%,
5%,we
wereject
rejectthe
thenull
nullhypothesis.
hypothesis.
this
7-53
7-54
P- value
7-55
not12
12pounds.
pounds. The
Theanalyst
analystwants
wantstototest
testthe
thenull
nullhypothesis
hypothesisatat=
=0.05.
0.05.
isisnot
=12
12
HH0:0:=
12
HH1:1:12
.95
0.6
For==0.05,
0.05,critical
criticalvalues
valuesof
ofzzare
are1.96
1.96
For
x 0
z
Thetest
teststatistic
statisticis:
is:
s
The
n
Donot
notreject
rejectHH0if:
if:[-1.96
[-1.96z
z1.96]
1.96]
Do
0
RejectHH0if:
if:[z[z<-1.96]
<-1.96]or
or zz1.96]
1.96]
Reject
0
0.5
0.4
0.3
0.2
.025
.025
0.1
0.0
-1.96
Lower Rejection
Region
1.96
Nonrejection
Region
Upper Rejection
Region
7-56
.95
0.6
0.5
0.4
0.3
0.2
.025
.025
0.1
0.0
-1.96
1.96
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsininthe
theupper
upperrejection
rejectionregion,
region,HH 0isisrejected,
rejected,and
andwe
wemay
may
Since
0
concludethat
thatthe
theaverage
averageamount
amountof
ofcarry-on
carry-onbaggage
baggageisismore
morethan
than12
12pounds.
pounds.
conclude
7-57
=2000
2000
HH0:0:=
2000
HH:1:2000
1
For==0.01,
0.01,critical
criticalvalues
valuesof
ofzzare
are2.576
2.576
For
Thetest
teststatistic
statisticis:
is:
The
x 0
z
s
n
Donot
notreject
rejectHH0if:
if:[-2.576
[-2.576z
z 2.576]
2.576]
Do
0
RejectHH0if:
if:[z[z<-2.576]
<-2.576]or
or zz2.576]
2.576]
Reject
0
z z
xx00
ss
nn
==
700
700
2700- -2000
2000
2700
=
=
947
947
100
100
.39
Reject
RejectHH
77.39
94.7
00
94.7
7-58
.99
0.6
0.5
0.4
0.3
0.2
.005
.005
0.1
0.0
-2.576
2.576
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
theupper
upperrejection
rejectionregion,
region,HH00
the
rejected,and
andwe
wemay
may
isisrejected,
concludethat
thatthe
theaverage
average
conclude
insuranceliability
liabilityper
perboard
board
insurance
seatin
insmall
smallcompanies
companiesisis
seat
morethan
than$2000.
$2000.
more
7-59
=3.24
3.24
HH0:0:=
3.24
HH1:1:3.24
For==0.05,
0.05,critical
criticalvalues
valuesof
ofzzare
are1.96
1.96
For
Thetest
teststatistic
statisticis:
is:
The
x 0
s
n
Donot
notreject
rejectHH0if:
if:[-1.96
[-1.96z
z1.96]
1.96]
Do
0
RejectHH0if:
if:[z[z<<-1.96]
-1.96]or
or zz1.96]
1.96]
Reject
0
200
nn == 200
3.48
xx == 3.48
2.8
s s == 2.8
z z
xx00
ss
nn
==
3.48--3.24
3.24
3.48
==
2.8
2.8
200
200
0.24
0.24
.21
Do
Donot
notreject
rejectHH
11.21
0.20
00
0.20
7-60
.95
0.6
0.5
0.4
0.3
0.2
.025
.025
0.1
0.0
-1.96
1.96
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
thenonrejection
nonrejectionregion,
region,HH00isis
the
notrejected,
rejected,and
andwe
wemay
may
not
concludethat
thatthe
theaverage
average
conclude
performancetime
timehas
hasnot
not
performance
changedfrom
from3.24
3.24seconds.
seconds.
changed
7-61
=49
49
HH0:0:=
49
HH1:1:49
18
nn==18
For==0.01
0.01and
and(18-1)
(18-1)==17
17df
df, ,
For
criticalvalues
valuesof
ofttare
are2.898
2.898
critical
Thetest
teststatistic
statisticis:
is:
The
x 0
s
n
Donot
notreject
rejectHH0if:
if:[-2.898
[-2.898t
t 2.898]
2.898]
Do
0
RejectHH0if:
if:[t[t<<-2.898]
-2.898]or
or tt2.898]
2.898]
Reject
0
n = 18
n = 18
x = 38
x = 38
s = 14
s = 14
t
t
x
38 - 49
x 0
0 == 38 - 49
s
14
s
14
n
18
n
18
- 11
=- 11 3.33 Reject H
= 3.3 3.33 Reject H 0
0
3.3
7-62
.99
0.6
0.5
0.4
0.3
0.2
.005
.005
0.1
0.0
-2.898
2.898
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
therejection
rejectionregion,
region,HH0isis
the
0
rejected,and
andwe
wemay
mayconclude
conclude
rejected,
thatthe
theaverage
averageprice
pricehas
hasnot
not
that
risenby
by49%.
49%. Since
Sincethe
thetest
test
risen
statisticisisin
inthe
thelower
lower
statistic
rejectionregion,
region,we
wemay
may
rejection
concludethat
thatthe
theaverage
average
conclude
pricehas
hasrisen
risenby
byless
lessthan
than
price
49%.
49%.
7-63
=27
27
HH0:0:=
27
HH1:1:27
24
nn==24
For==0.05
0.05and
and(24-1)
(24-1)==23
23df
df, ,
For
criticalvalues
valuesof
ofttare
are2.069
2.069
critical
Thetest
teststatistic
statisticis:
is:
The
x 0
s
n
Donot
notreject
rejectHH0if:
if:[-2.069
[-2.069t
t 2.069]
2.069]
Do
0
RejectHH0if:
if:[t[t<<-2.069]
-2.069]or
or tt2.069]
2.069]
Reject
0
24
nn == 24
24.6
xx == 24.6
7.4
s s == 7.4
t t
xx00
ss
nn
24.6--27
27
24.6
==
7.4
7.4
24
24
-2.4
-2.4
.59
==
1.159
1.51
1.51
Do
Donot
notreject
rejectHH
00
7-64
.95
0.6
0.5
0.4
0.3
0.2
.025
.025
0.1
0.0
-2.069
2.069
Lower Rejection
Region
Nonrejection
Region
Upper Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
thenonrejection
nonrejectionregion,
region,HH00isis
the
notrejected,
rejected,and
andwe
wemay
maynot
not
not
concludethat
thatthe
theaverage
average
conclude
speedisisdifferent
differentfrom
from27
27
speed
copiesper
perminute.
minute.
copies
7-65
Statistical Significance
While the null hypothesis is maintained to be true throughout a hypothesis
test, until sample data lead to a rejection, the aim of a hypothesis test is
often to disprove the null hypothesis in favor of the alternative hypothesis.
This is because we can determine and regulate , the probability of a Type I
error, making it as small as we desire, such as 0.01 or 0.05. Thus, when we
reject a null hypothesis, we have a high level of confidence in our decision,
since we know there is a small probability that we have made an error.
A given sample mean will not lead to a rejection of a null hypothesis unless
it lies in outside the nonrejection region of the test. That is, the nonrejection
region includes all sample means that are not significantly different, in a
statistical sense, from the hypothesized mean. The rejection regions, in turn,
define the values of sample means that are significantly different, in a
statistical sense, from the hypothesized mean.
7-66
0.70
HH0:0:pp==0.70
0.70
HH1:1:pp0.70
210
nn==210
For==0.05
0.05critical
criticalvalues
valuesof
ofzzare
are1.96
1.96
For
Thetest
teststatistic
statisticis:
is: z p p0
The
p0 q 0
n
Donot
notreject
rejectHH0if:
if:[-1.96
[-1.96z
z 1.96]
1.96]
Do
0
RejectHH0if:
if:[z[z<<-1.96]
-1.96]or
or zz1.96]
1.96]
Reject
0
210
nn == 210
130
130
p
=
.619
p = 21000.619
210
p p- -pp 0
0
z z==
==
p
q
p 0q 0
0 0
nn
==
0.619--0.70
0.70
0.619
(0.70)(0.30)
(0.70)(0.30)
210
210
-0.081
-0.081
.5614
Reject
RejectHH
22.5614
0.0316
0.0316
00
7-67
55
HH0:0:55
55
HH1:1:55
100
nn==100
For==0.01,
0.01,the
thecritical
criticalvalue
value
For
ofzzisis2.326
2.326
of
Thetest
teststatistic
statisticis:
is:
The
x 0
s
n
Donot
notreject
rejectHH0if:
if:[z[z2.326]
2.326]
Do
0
RejectHH0if:
if:zz2.326]
2.326]
Reject
0
100
nn == 100
60
xx == 60
20
s s == 20
z z
xx00
ss
nn
==
60--55
55
60
20
20
100
100
==
55 2.5
2.5
22
Reject
RejectHH
00
7-68
0.99
f(z)
0 .3
0 .2
0 .1
0 .0
-5
2.326
2.5
Nonrejection
Region
Rejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
therejection
rejectionregion,
region,HH00isis
the
rejected,and
andwe
wemay
mayconclude
conclude
rejected,
thatthe
theaverage
averageconcentration
concentration
that
ofvinyl
vinylchloride
chlorideisismore
morethan
than
of
55ppm.
ppm.
55
7-69
12
HH0:0:12
12
HH1:1:12
144
nn==144
For==0.05,
0.05,the
thecritical
criticalvalue
value
For
ofzzisis-1.645
-1.645
of
Thetest
teststatistic
statisticis:
is:
The
x 0
s
n
Donot
notreject
rejectHH0if:
if:[z[z-1.645]
-1.645]
Do
0
RejectHH0if:
if:zz]
]
Reject
0
144
nn == 144
11.8
xx == 11.8
ss == 66
11.8-12
-12
xx 0 11.8
0
z
=
z
=
ss
66
144
nn
144
-.2
-.2
=
Do
Donot
notreject
rejectHH
= .5 00.4.4
00
.5
7-70
0.95
f(z)
0.3
0.2
0.1
0.0
-5
-1.645
-0.4
Rejection
Region
Nonrejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
thenonrejection
nonrejectionregion,
region,HH00isis
the
notrejected,
rejected,and
andwe
wemay
maynot
not
not
concludethat
thatthe
the
conclude
manufacturerisisunderfilling
underfilling
manufacturer
packageson
onaverage.
average.
packages
7-71
65
HH0:0:65
65
HH1:1:65
21
nn==21
For==0.01
0.01an
an(21-1)
(21-1)==20
20df,
df,the
the
For
criticalvalue
value-2.528
-2.528
critical
Thetest
teststatistic
statisticis:
is:
The
Donot
notreject
rejectHH0if:
if:[t[t-2.528]
-2.528]
Do
0
RejectHH0if:
if:zz]
]
Reject
0
7-72
0.95
f(t)
0 .3
0 .2
0 .1
0 .0
-5
-2.528
-3.82
Rejection
Region
Nonrejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
therejection
rejectionregion,
region,HH00isis
the
rejected,and
andwe
wemay
mayconclude
conclude
rejected,
thatthe
themanufacturers
manufacturersclaim
claim
that
false,that
thatthe
theaverage
average
isisfalse,
floodlightlife
lifeisisless
lessthan
than65
65
floodlight
hours.
hours.
7-73
0.0096
HH0:0:pp0.0096
0.0096
HH1:1:pp0.0096
600
nn==600
For==0.10
0.10 the
thecritical
criticalvalue
value1.282
1.282
For
Thetest
teststatistic
statisticis:
is:
The
Donot
notreject
rejectHH0if:
if:[z[z1.282]
1.282]
Do
0
RejectHH0if:
if:zz]
]
Reject
0
7-74
0.90
f(z)
0 .3
0 .2
0 .1
0 .0
-5
1.282
0.519
Nonrejection
Region
Sincethe
thetest
teststatistic
statisticfalls
fallsin
in
Since
thenonrejection
nonrejectionregion,
region,HH00isis
the
notrejected,
rejected,and
andwe
wemay
maynot
not
not
concludethat
thatproportion
proportionof
ofall
all
conclude
hotelsin
inthe
thecountry
countrythat
thatmeet
meet
hotels
theassociations
associationsstandards
standardsisis
the
greaterthan
than0.0096.
0.0096.
greater
Rejection
Region
7-75
0.4
0.4
f(z)
0.2
0.2
0.1
0.1
0.0
p-value=area to
right of the test statistic
=0.0062
0.3
f(z)
p-value=area to
right of the test statistic
=0.3018
0.3
0.0
-5
0.519
Additional Example k
-5
2.5
Additional Example g
Thep-value
p-valueisisthe
theprobability
probabilityof
ofobtaining
obtainingaavalue
valueof
ofthe
thetest
teststatistic
statisticas
asextreme
extremeas,
as,
The
ormore
moreextreme
extremethan,
than,the
theactual
actualvalue
valueobtained,
obtained,when
whenthe
thenull
nullhypothesis
hypothesisisistrue.
true.
or
Thep-value
p-valueisisthe
thesmallest
smallestlevel
levelof
ofsignificance,
significance,,
,atatwhich
whichthe
thenull
nullhypothesis
hypothesis
The
maybe
berejected
rejectedusing
usingthe
theobtained
obtainedvalue
valueof
ofthe
thetest
teststatistic.
statistic.
may
7-76
7-77
f(z)
0.3
0.2
0.1
0.0
-5
-0.4
0.4
Inaatwo-tailed
two-tailedtest,
test,we
wefind
findthe
thep-value
p-valueby
bydoubling
doublingthe
thearea
areain
in
In
thetail
tailof
ofthe
thedistribution
distributionbeyond
beyondthe
thevalue
valueof
ofthe
thetest
teststatistic.
statistic.
the
7-78
7-79
Onecan
canconsider
considerthe
thefollowing:
following:
One
Sample
SampleSizes
Sizes
versusfor
versusforvarious
varioussample
samplesizes
sizes
The
ThePower
PowerCurve
Curve
The
TheOperating
OperatingCharacteristic
CharacteristicCurve
Curve
7-80
Computing and
Plotting Required
Sample size.
7-81
7-82
Note: Similar
analysis can
be done when
testing for a
population
proportion.
7-83
7-84