Testing:One-sampleTests
@ Oxford Cereals,Part II
USINGSTATISTICS
ESIS.TESTING
M ETHODOLOGY 9.4 t TESTOF HYPOTHESIS
Null andAlternativeHypotheses FORTHE MEAN (o UNKNOWN)
CriticalValueof the TestStatistic TheCriticalValueApproach
of Rejectionand Nonrejection Thep-ValueApproach
in DecisionMaking UsingHypothesis- CheckingAssumptions
TestingMethodology
9.5 ZTESTOF HYPOTHESIS
TESTOF HYPOTHESIS FORTHE PROPORTION
THEMEAN (o KNOWN) TheCriticalValueApproach
CriticalValueApproachto HypothesisTesting Thep-Value Approach
p-ValueApproachto Hypothesis Testing
9.6 POTENTIALHYPOTHESIS.TESTI
NG
ConnectionBetweenConfidenceInterval
PITFALLSAND ETHICALIssUES
Estimationand HypothesisTesting
9.7 C' (CD.ROMTOPIq THE POWEROF A TEST
.TAILTESTS
CriticalValueApproach EXCELCOMPANIONTO CHAPTER9
p-ValueApproach E9.l UsingtheZTestfor theMean(o Known)
E9.2 UsingtheI Testfor theMean(o Unknown)
E9.3 UsingtheZTestfor theProportion
9.'I HYPOTHESIS.TESTING
METHODOLOGY
Hypothesistestingtypically beginswith sometheory,claim, or assertionabouta particular
meterof a population.For example,your initial hypothesisaboutthe cerealexampleis
processis working properly,so the meanfill is 368 grams,andno correctiveactionis
/10p
: :368
9.1:Hypothesis-GstingMethodology
329
Even though information is availableonly from the sample,the null hypothesisis written in
terms of the population.Remember,your focus is on the populationof all cerealboxes.The
samplestatisticis usedto make inferencesaboutthe entire filling process.One inferencemay
be that the resultsobservedfrom the sampledataindicatethat the null hypothesisis false.If the
null hypothesisis considered false,somethingelsemustbe true.
Whenevera null hypothesisis specified,an alternativehypothesisis also specified,and it
mustbe true if the null hypothesisis false.The alternativehypothesis,Ilr, is the oppositeof
the null hypothesis,/10.This is statedin the cerealexampleas
Hr:1:"*368
Ho: trt:4.5
330 CHAPTERNINE Fundamentals
of Hypothesis
Testing:One-Sample
Tests
The alternative hypothesis is the opposite of the null hypothesis. Because the null
that the populationmeanis 4.5 minutes,the alternativehypothesisis that the population
is not 4.5 minutes.This is statedas
Hr: 1t+4.5
9.1
ionsof rejection
nonrejection in
is testing
\tr
Critical Critical
Value Value
Regionof Regionof Regionof
Rejection Nonrejection Rejection
In the Oxford Cerealsscenario,you makea TypeI error if you concludethat the population
meanfill is not 368 when it ls 368. This error causesyou to adjustthe filling processeven
thoughthe processis working properly.You makea TypeII error if you concludethat the pop-
ulationmeanfill ls 368when itis not 368.Here,you wouldallowtheprocessto continuewith-
out adjustmenteventhoughadjustmentsareneeded.
The confidencecoefficient, (1 - cr),is the probability that you will not rejectthe null
hypothesis,F1o,when it is true and shouldnot be rejected.The confidence level of a
hypothesistestis (l - cr)x 100%.
The power of a statistical test, (l - B), is the probability that you will rejectthe null
hypothesiswhen it is falseand shouldbe rejected.
In the Oxford Cerealsscenario,the powerof the test is the probabilitythat you will
rectlyconcludethat the meanfill amountis not 368 gramswhen it actuallyis not 368
For a detaileddiscussionof thepowerof the test,seeSection9.7 on the StudentCD-ROM.
a. Explainthe risksof committinga TypeI or TypeII error. 9.17 The operations manager at a clothing factory
b. Which type of errorarethe consumergroupstryingto to detenninewhethera new machine is producinga
avoid?Explain. ular tvne of cloth accordins to the manufacturer's
c. Which type of error are the industry lobbyiststryrng to cations, which indicate that the cloth should havea
avoid'?Explain. breaking strength of 70 pounds. State the null andal
d. How would it be possibleto lower the chancesof both tive hvootheses.
Type I and Type Il errors'?
9.18 The managerof a paint supply storewantst0
mine whether the amount of oaint containedin 1
ffi 9 . 1 6 A s a r e s u l t o f c o r n p l a i n t sf r o m b o t h s t u -
dents and faculty about lateness,the registrar at a
T/-sEAlarge university wants to adjust the scheduled
cans purchased from a nationally known manufi
actually averagesI gallon. State the null and al
hypotheses.
ffi class times to allow for adequate travel time
between classesand is ready to undertake a study. Until 9.19 The aualitvcontrolmanaserat a lisht bulb
noq the registrarhas believedthat there shouldbe 20 min- needs to determine whether the mean life of a large
utes betweenscheduledclasses.Statethe null hvpothesis. ment of light bulbs is equal to the specifiedvalueof
Hn, and the alternativehypothesis,H,. hours. State the null and alternativehvootheses. !
H6: P = 363
When the standarddeviation,o, is known, you use the Z test if the populationis normallyd
tributed. If the population is not normally distributed, you can still use the Z test if the sam
size is large enough for the Central Limit Theorem to take effect (see Section 7.4). Equa
(9.1) defines the Z-test statistic for determining the difference between the sample mean,
and the populationmean,p, when the standarddeviation,o, is known.
FORTHEMEAN(o KNOWN)
ZTESTOF HYPOTHESIS
X-p
Z- (e.1)
o
G
ln Equation (9.1)the numerator measureshow far (in an absolutesense)the obser
sample mean, X, is fiom the hypothesizedmean,lu. The denorninatoris the standarderro
the mean, so Z representsthe difference between X and p in standarderror units.
,dS each tail of the normal distribution results in a cumulative area of 0.025 below the lower criti-
ic- cal value and a cumulative areaof 0.975 below the upper critical value. According to the cumu-
fi- lative standardizednormal distribution table (Table E.2), the critical values that divide the
)an rejectionand nonrejectionregionsare -1.96 and +1.96. Figure 9.2 illustratesthat if the mean is
la- actually 368 grams, as Hn claims, the values of the test statistic Zhave a standardizednormal
distributioncenteredat Z :0 (which correspondsto an X value of 368 grams).Valuesof Z
greaterthan +1.96 or less than -1.96 indicate that X is so far from the hypothesizedp:368
.'r-
that it is unlikely that such a value would occur if F1nwere true.
on
'er
ve FIGURE 9.2
Testinga hypothesis
aboutthemean
ry (oknown) at the 0.05
p- level
of significance
75
0
I i
I
R e g i o no f R e g i o no f
Nonrejection Rejection
rg I
ill
Critical
rg Value
le
rl-
,y
Therefore,the decisionrule is
IS
R e j e c tH o l f Z > + 1 . 9 6
orifZ<-1.96;
otherwise, do not reject Hn.
FIGURE 9.3
Testing
a hypothesis
about
the meancereal
weight(o known)at 0.05
level
of significance
i -1.96 0 + ' 1 . 5 0+ 1 . 9 6 I Z
Regionof R e g i o no f R e g i o no f
Rejection Nonrejection Rejection
336 CHAPTER NINE Fundamentalsof HypothesisTesting:One-SampleTests
EXHIBIT9.1 THESIX-STEP
METHODOF HYPOTHESIS
TESTING
l. Statethe null hypothesis, Hr, andthe alternativehypothesis,11,.
2. Choosethe levelof significance,o, andthe samplesize,n. The levelof signifi
basedon the relativeimportanceof the risks of committingTypeI andTypeII
the problem.
J . Determinethe appropriate test statisticand samplingdistribution.
4. Determinethe critical valuesthat divide the rejectionand nonrejectionregions.
5 . Collect the sampledataand computethe valueof the test statistic.
6 . Make the statisticaldecisionand statethe manaserialconclusion.If the test
falls into the nonrejectionregion,you do not rejectthe null hypothesis,110.
Iftk
statisticfalls into the rejectionregion,you rejectthe null hypothesis.The
conclusionis written in the contextof the real-worldoroblem.
E X A M P L E9 . 2 APPLYING
THESIX-STEP
METHODOF HYPOTHESIS
TESTING
AT OXFORDCEREALS
Statethesix-step
methodof hypothesis
testingat OxfordCereals.
SOLUTION
Step I Statethe null and alternativehypotheses. The null hypothesis,
lln, is always
statisticalterms,usingpopulationparameters. In testingwhetherthe meanfill is
grams,the null hypothesisstatesthat p equals368.The alternativehypothesis,ll
alsostatedin statisticalterms,usingpopulationparameters. Therefore,the
hypothesisstatesthat p is not equalto 368 grams.
Step 2 Choose the level of significance and the sample size.You choosethe level of si
cance,o, accordingto the relativeimportanceof the risks of committingTypeI
TypeII errorsin the problem.The smallerthe valueof c, the lessrisk thereis of
ing a Type I error. In this example,a Type I error is to concludethat the
meanis not 368 gramswhenit is 368 grams.Thus,you will takecorrectiveaction
the filling processeventhough the processis working properly.Here,c : 0.05
selected. Thesample,n, is 25.
Step3 Selectthe appropriatetest statistic.Becauseo is known from informationabout
filling process,you usethe normaldistributionandtheZ teststatistic.
Step 4 Determinethe rejectionregion.Critical valuesfor the appropriatetest statistic
selectedso that the rejectionregioncontainsatotal arcaof crwhenI1nis trueand
nonrejection region contains alotal area of I - cr when 11nis true. Becausec =
in the cerealexamDle.the critical valuesof the Z-teststatisticare-1 .96 and+1.
The rejectionregionis thereforeZ < -1.96 or Z > +1.96.The nonrejectionregion
-1.96<Z<+1.96.
Step 5 Collect the sample data and compute the value of the test statistic. In the
e x a m p l e . X : 3 7 2 . 5 . a n d t h e v a l u e o f t h e t e s t s t a t i s t i ci s Z : + 1 . 5 0 .
FIGURE9.4
Findinga p-value
for a two-tailtest
.8684
I
I
I
-1.50 0 +1.50 Z
FIGURE9.5
MicrosoftExcelZ-test
resultsfor the cereal-fill
examote
-s/s0RT(87)
-(88 - 8tyBt1
-lloRXSlllv(85/21
-HORISII{V(I - Blilzl
-2' (1- xoRxsolsr(ABspr4))
-lF(817< 85, 'ReJoctihe null hypotheCr',
"0o not .c.rccllic null hlrpoihcria')
fortheMean(o Known) 339
9.2: ZTestofHypothesis
H6 trt: 4.5
H r : 1 t +4 . 5
_
t-
X - t- -
5.1-4.5 _)50
o 1.2
r
1n ]E
To find the probability of gettinga test statistic Z that is equalto or more extremethan
2.50 standarddeviationunits from the centerof a standardized normal distribution,
you computethe probabilityof a Z valuegreaterthan2.50 alongwith the probability
340 CHAPTER NINE Fundamentalsof HypothesisTesting:One-SampleTests
Z_
X -p,
o
_T
NN
o
x t z -T
\1 n
372.s+(1.e6)+
12s
3 7 2 . 5 +5 . 8 8
so that
366.62(u(378.38
the mean amount dispensedis different from 8 ounces. and earninginterest.Supposethat at a particular
Supposethat the meanamountdispensedin a particular the populationmean amountof money withdrawn
sampleof 50 bottlesis 7.983ounces. AIMs per customertransactionover the weekendis
a. Is thereevidencethat the populationmeanamountis dif- with a populationstandarddeviationof $30.
ferentfrom 8 ounces?(Usea 0.05levelof significance.) a. If a randomsampleof 36 customertransactions
b. Computethep-valueandinterpretits meaning. catesthat the samplemeanwithdrawal amountis
c. What is your answerin (a) if the standarddeviationis is there evidenceto believethat the population
0.05ounce? withdrawalamountis no longer$160?(Usea 0.05
d. What is your answerin (a) if the samplemeanis 7.952 of significance.)
ouncesandthe standarddeviationis 0.15ounce? b. Computethep-valueand interpretits meaning.
9.33 ATMs must be stockedwith enoughcashto satisfy c. What is your answerin (b) if you use a 0.01
customersmakingwithdrawalsoveran entireweekend.But significance?
if too much cashis unnecessarilykept in the ATMs, the d. What is your answerin (b) if the standarddevi
bank is forgoing the opportunityof investingthe money $24(usecr: 0.05)?
9.3 ONE.TAILTESTS
In Section9.2, hypothesis-testing methodologyhasbeenusedto examinethe question
whetherthe populationmeanamountof cerealfilled is 368 grams.The alternative
(Ht: p * 368) containstwo possibilities:Eitherthe meanis lessthan368 grams,or the
morethan 368 grams.For this reason,the rejectionregionis dividedinto the two tailsof
samplingdistributionof the mean.
In many situations,however,the alternativehypothesisfocuseson a particular
one such situationoccursin the following application:A companythat makes
cheeseis interested in determiningwhethersomesuppliersthatprovidemilk for theprocessi
operationareaddingwaterto theirmilk to increasethe amountsuppliedto theprocessing
ation.It is knownthatexcesswaterreducesthe freezingpoint of themilk. The freezingpoint
naturalmilk is normallydistributed,with a meanof -0.545' Celsius(C). The standard
tion of the freezingtemperature of naturalmilk is known to be 0.008"C.Becausethe c
companyis only interestedin determiningwhetherthe freezingpoint of the milk is less
what would be expectedfrom naturalmilk, the entirerejectionregionis locatedin the lower
of the distribution.
The CriticalValueApproach
Supposeyou wish to determinewhetherthe meanfreezingpoint of milk is lessthan-0.5450.
To performthis one-tailhypothesistest,you use the six-stepmethodlistedin Exhibit 9.1on
page336.
Step1 Ho: It>--0.545"C.
Hi trt<-0.545"C.
:- 'i'-1.645 0
rf R e g i o no f R e g i o no f
t-
Rejection Nonrejection
e
n
il Step5 Youselecta sampleof 25 containers of rrrilkand find thatthe sarnplerreanfreezing
p o i n te q u a l-s0 . 5 5 0 " CU. s i n gr r: 2 5 . X : - 0 . 5 5 0 " Co. - 0 . 0 0 8 ' Ca, n dE q u a t i o(n9 . 1)
on page334,
- F -p - ( - 0.s4s)
- 0.ss0 - _1 t?{
o 0.008
T
1tl ,E
Step 6 BecauseZ : -3.125 < - I .645.you rc'jectthe null hypothesis(seeFigure9.6).You con-
clude that the rneanfreezingpoint of the rrilk provided is below -0.545'C. The com-
pany should pursue an irrvestigationof thc milk supplier becausethe mean freezing
point is significantlybelow what is expectedto occur by chance.
Steps l-3 Thcse stepsare the sarneas in the critical value approach.
FIGURE9.7
Determining the
p-valuefor a
one-tailtest
-3.125
FIGURE9.8
MicrosoftExcelZtest
resultsfor the milk
productionexample
.86/5ARTF4
See Section E9.1 to create -(88 _B4yB11
this.
-iloRrslw(B$
-ll0RllS0lST@12)
-lF{816 < Bli, ?eJea tha null hypothrdr',
"llo DotroJ.c{the null hnothcclrJ
E X A M P L E9 . 5 A ONE-TAILTESTFORTHE MEAN
A companythat manufactureschocolatebarsis particularlyconcernedthat the meanweight
a chocolatebar not be greaterthan6.03ounces.Pastexperience allowsyou to assumethat
standarddeviationis 0.02 ounces.A sampleof 50 chocolatebarsis selected,and the
meanis 6.034 ounces.Using the o : 0.01 level of significance,is thereevidencethat the
ulationmeanweightof the chocolatebarsis greaterthan6.03ounces?
SOLUTION Usingthe criticalvalueapproach,
Step I l1o:p ( 6.03
11,:trr> 6.03
The decisionrule is
X -V =w#&=r'4r4
Z_
o
1n So
Step6 BecauseZ:1.414 <2.33, you do not rejectthe null hypothesis. Thereis insufficient
evidenceto concludethat the populationmeanweight is greaterthan 6.03 ounces.
)d
ld FORTHEMEAN(o UNKNOWN)
t TESTOF HYPOTHESIS
re Y -rt
1d
(e.2)
.t
lr"
8). G
ity where the test statistic t follows a I distribution having n - I degreesof freedom.
re-
)ur
To illustratethe use of this 1test, return to the Using Statisticssccnarioconcerningthe
Saxon Home lmprovementCompany on page 284. Over the past five years,the mean amount
per salesinvoice is S120.As an accountantfor the company,you need to inform the finance
iin
departmentif this amount changes.In other words. the hypothesistest is r-rsed to try to prove
that the mean amountper salesinvoice is increasingor decreasing.
/in
;hat
.ore The Critical Value Approach
that
To perform this two-tail hypothesistest, you usethe six-steprnethodlistedin Exhibit9.1 on
the
p a g e3 3 6 .
the
ater Step I H , , :p - $ 1 2 0
H,:p*$120
FIGURE9.9
Testing.ahypothesis
aboutthe mean
(o unknown)at the
0.05levelof significance
with 11 degreesof
freedom
-2.2010 '0
| +2.2010 | t.,
tt
Region
I
I
of Regionof R e g i o no f
Rejection Nonrejection Rejection
I
Critical Critical
Value Value
Step5 The datain the file FilFlllttltl arethe amounts(in dollars)in a random
sampleof
salesinvoices.
108.98 152.22 I I I .45 110.59 t2t .46 107.2o
9 3 . 3 2 9 t . 9 7 I I L 5 6 7 s . 7 | 1 2 8 . 5 8 l 3 s . II
S"
L"I
X=i=t =$112.85
n
FromEquation(9.2) on page347,
x -tt 1 1 2 . 8-5 1 2 0
= -1.1908
.s 20.80
tr
1n E
FIGURE9.10 AB
MicrosoftExcelresults
for the one-samole
t test of salesinvoices
-B8rS0RT{861
-86 -1
See Section E9.2 to create -6n , ar;/Brt
this.
-{nilvps,8.t2)}
-nilv{85,8121
-TDlsT(A8s{Bt3},Bl2,2t
-lF(818< &i, -Rejectrte n{lt hirpothesb-,
'Do
not roJecf $. null hypoth€b"l
fortheMean(oUnknown) 349
9.4:tTestofHypothesis
Step 6 Because-2.2010 < l: -1.1908 < 2.2010,you do not rejectF1o. You haveinsufficient
to
evidence conclude that the mean amount per salesinvoice differs from $120.You
shouldinform the financedepartmentthat the audit suggeststhat the meanamountper
invoicehasnot chansed.
CheckingAssumptions
Youusethe one-sampler testwhenthe populationstandarddeviation,o, is not knownandis esti-
a large sample size matedusingthe samplestandarddeviation,l,S.To usethe one-sample/ test,the dataareassumed
available,S estimates o to representa randomsamplefrom a populationthat is normally distributed.In practice,aslong
enough that there asthe samplesizeis not very small andthe populationis not very skewed,the I distributionpro-
littledifference betwe en videsa good approximationto the samplingdistributionof the meanwheno is unknown.
t andZ distributions. Thereare severalwaysto evaluatethe normality assumptionnecessaryfor using the / test.
You can observehow closely the samplestatisticsmatch the normal distribution'stheoretical
properties.You can also use a histogram,stem-and-leafdisplay,box-and-whiskerplot, or nor-
mal probability plot. For detailson evaluatingnormaliry seeSection6.3 on page234.
Figure9.11 presentsdescriptivestatisticsgeneratedby MicrosoftExcel.Figure9.12 is a
MicrosoftExcelbox-and-whiskerplot. Figure9.13is a MicrosoftExcelnormalprobabilityplot.
9.11
Excel
statisticsfor fl2.qrfi
invoice data 6.txls:f
lrlm
fode #vA
Standard Dcvladon 20.73799
Varlancs f32.t565
Section E3.1 to create
Xurtoglr 0.1727|a
Skormoer 0.1rnn
Range 70-51
tlnlmom 75.71
Xnxlmum $7J,.
Sum 1354.21
Count 12
Largcctfl| tna,
Smallodfl) T5.fI
350 CHAPTER NINE Fundamentalsof Hypothesis Testing: One-SampleTests
Box.and4llhiskerPlot of InvoiceAmount
FIGURE9.12
MicrosoftExcelbox-
and-whiskerplot for the
salesinvoicedata
'1
too
E
See Section E6.2 to create 5ro
this.
:50
Becausethe mean is very close to the median,the points on the normal probabilit
appearto be increasingapproximatelyin a straightline, and the box-and-whiskerplot ap
approximatelysymmetrical.You can assumethat the populationof salesinvoicesis ap
mately normally distributed.The normality assumptionis valid, and thereforethe aud
resultsarevalid.
The / test is a robust test. It doesnot lose powerif the shapeof the populationde
somewhatfrom a normal distribution,particularlywhen the samplesizeis largeenou
enablethe test statisticI to be influencedby the Central Limit Theorem(seeSectio
However,you can makeerroneousconclusionsand can lose statisticalpowerif you ut
/ test incorrectly.If the samplesize,n, is small (that is, lessthan 30) and you cannot(
make the assumptionthat the underlying population is at least approximatelynor
distributed,othernonparametrictesting proceduresare more appropriate(seerefer
1and2).
9.4; tTestof Hypothesisfor theMean(o Unknown) 351
7 a. At the0.05levelof significance, is thereevidence that 9.65 Although many people think they can put a meal on
5 themeandifferenceis not equalto 0.0 inches? the table in a short period of time, a recent article reported
J
b.Determine thep-valuein (a) and interpretits meaning. that they end up spending about 40 minutes doing so
c. Whatassumptionaboutthe differencesbetweenthe (N. Hellmich, 'Americans Go for the euick Fix for
J
actuallengthofthe steelpart andthe specifiedlengthof Dinner," USA Today,February 14,2006). Supposeanother
thesteelpart mustyou makein (a)'i study is conducted to test the validity of this statement.A
d.Evaluatethe assumptionin (c) graphically.Are the sample of 25 people is selected,and the length of time to
)er results of (a) valid?Why/ prepare and cook dinner (in minutes) is recorded,with the
.u-
9.64 In Problem3.63on page138,youwereintroduced following results (stored in the file E[lf[E)
/el to
ue atea-bag-filling operation.An importantqualitycharacter- 44.0 51.949.7 40.055.533.0 43.4 4t.3 45.240.74t.t 4g.t 30.9
isticof interestfor this processis the weightof the tea in
D? 45.255.352.155.138.843.1 39.258.649.843.247.946.6
rst theindividualbags.The data in the file [ll[!![! is an
ordered arrayof the weight,in grams,of a sampleof 50 tea a. Is there evidence that the population mean ume ro pre-
ri- bags produced duringan eight-hourshift. pare and cook dinner is greater than 40 minutes? Use a
a.Is thereevidencethat the meanamountof teaper bag is level of significanceof 0.05.
he differentfrom 5.5grams(usea : 0.0l)? b. What assumptionsare made to perform the test in (a) /
b.Construct a99%oconfidence intervalestimate of thepopu- c. Do you think the assumptionneededin (a) is seriously
lationmeanamountof teaperbag.Interpretthisinterval. violated?Explain.
gs c. Compare theconclusions reached in (a) and(b). d. Determinethe p-value and interpretits meaning.
he
ge
ch
€e 9.5 Z TESTOF HYPOTHESIS
FOR THE PROPORTION
ce
In some situations,you want to test a hypothesisabout the proportion of successesin the popu-
of
lation, n' rather than testing the population mean. To begin, you select a random sample
ny and
compute the sample proportion, p: Xln. You then comparethe value of this statisticto
LCS the
hypothesizedvalue of the parameter,n, in order to decide whether to reject the null hypothesis.
hs
If the number of successes(X) and the number of failures (n - X) are each at least five, the
sampling distribution of a proportion approximately follows a normal distribution. you use the
z
Z test for the proportion given in Equation (9.3) to perform the hypothesistest for the differ-
ence between the sample proportion, p, and the hypothesizedpopulation proportion, n.
7
2
0 ONE SAMPLEZ TESTFORTHEPROPORTION
n-fl
z--: (e.3)
n ( 1- n )
rat
n
t6
where
V
Jo Numberof successesin the sample
n Samplerirr-
3n p : sampleproportionof successes
n: hypothesized
proportionofsuccesses
in thepopulation
rel
andthe teststatisticZ approximately
followsa standardized
normaldistribution.
tn
rd
ne Alternatively,by multiplyingthe numeratoranddenominator
by n, youcanwrite theZ tesl
in statisticin termsof thenumberof successes,
x, asshownin Equation(9.4).
ne
,'n
Z TESTFOR THE PROPORTIONIN TERMSOF THE NUMBEROF SUCCESSES
er
el X -nn
Z _ --- (e.4)
rrt lnn(l - n)
354 CHAPTERNINE Fundamentals
of Hypothesis
Testing:
One-Sample
Tests
The CriticalValueApproach
Becauseyou are interestedin whetherthe proportion of independentgroceryowners
view Wal-Martastheir biggestcompetitivethreatis 0.50,you usea two-tailtest.If you
the a : 0.05 level of significance,the rejection and nonrejectionregionsare setup as
Figure9.14, andthe decisionrule is
t o i fZ < - l . 9 6 o r i f Z > + 1 . 9 6 ;
R e j e cH
otherwise,do not reject110.
FIGURE9.14
Two-tailtest of
hypothesisfor the
proportionat the 0.05
levelof sionificance
Regionof
Rejection
78 =
'p = - 0.5166
lsl
Using Equation(9.3),
0 . 5 1 6-6 0 . s 0 o'0166=
Z_ 0.4069
ru(l- r) 0.50(l- 0.s0) 0.0407
n
or, usingEquation(9.4).
X-nn 7 8 - ( 1 5 1 X 0 . 5=0 ) 2 . 5 = 0 . 4 0 6 9
Z_
,,lnn(l- n) 6.144r
./tst1o.soy1o.so;
9.5: ZTcstof Hypothesis
for the Proportion 355
ro- Because-1.96 < Z:0.4069 < 1.96,you do not rejectHo. There is insufficientevidencethat the
reir proportion of independentgrocery owners who view Wal-Mart as their biggest competitive
ner threat is not 0.50. Figure 9.15 presentsMicrosoft Excel resultsfor thesedata. In other words,
rey the currentbelief that half of al1independentgrocersview Wal-Mart as their bigget competitive
lnd threatis not contradictedbv the Z test.
For
FIGURE
9.15
1
MicrosoftExcel
results
for the survey 3.
4
of independentgrocery 5
ownersand Wal-Mart
7-
asbiggestcompetitive
threat g
10. -S187
1 1. -soRT(8411- B{)r87}
l1 -{810- 84}1811
't3
'14'
vho 15: -t{oRf,stilv(852}
SeeSectionE9.3 to create 16 -llORf,SlNV(1 ,852)
lect -2' (1 - ilORilSDrsT(ABS(Br?)))
this. 18' -lF{817< 85, -Rejad rhe null hypothe€is",
sin To not rejecrlho ilull hypoth6ls'J
The p-ValueApproach
As an alternativeto the critical value approach,you can computethe p-value. For this two-tail
test in which the rejectionregion is locatedin the lower tail and the uppertail, you needto find
the area below a Z value of -0.4069 and above a Z value of +0.4069. Figure 9.15 reports a
p-value of 0.684L Becausethis value is greaterthan the selectedlevel of significance(o : 0.05),
vou do not reiect the null hvnothesis.
(9.3)on page353,
UsingEquation
Yq4
p=1)_ =0.94
n 100
oog
Z - i <1
-J/r (|r - 0.8s(Lq.8l) 0 . 0 3 5 7
n 100
and then cancelledtheir transactionwas less Plotting Mix of Work and Family Won't Find perfectplan,"
0,50with the new system? The WallStreetJournal, June11, 2002,p. Bl). Assumethat
the resultsof(a) and (b) and discussthe effect the group of 187 women was a randomsamplefrom the
samplesize has on the outcome,and,in general,in populationof all successfulwomenexecutives.
a. What was the sampleproportion of successfulwomen
executiveswho had children?
Moreprofessionalwomenthan everbefore are forgo- b. At the 0.05 level of significance,can you statethat more
becauseof the time constraintsof their thanhalf of all successfulwomenexecutiveshad children?
However,manywomenstill manageto find time to c. At the 0.05 level of significance,canyou statethat more
the corporateladderand settime asideto havechil- than two-thirds of all successfulwomen executiveshad
A surveyof I 87 attendeesat FortuneMagazine'sMost children?
Womenin Businesssummit in March 2002 found d. Do you think the random sampleassumptionis valid?
133had at least one child (C. Hymowitz, "Women Explain.
9.6 POTENTIALHYPOTHESIS-TESTING
PITFALLS
AND ETHICALISSUES
To this point, you havestudiedthe fundamentalconceptsof hypothesistesting.You haveused
hypothesistestingto analyzedifferencesbetweensampleestimates(that is, statistics)and
hypothesizedpopulation characteristics(that is, parameters)in order to make businessdeci-
sionsconcerningthe underlyingpopulationcharacteristics. You havealso learnedhow to eval-
uatethe risks involvedin making thesedecisions.
When planning to carry out a test ofthe hypothesisbasedon a survey,researchstudy,or
designedexperiment,you must ask severalquestionsto ensurethat you useproper methodol-
ogy.You needto raiseand answerquestionssuchas the following in the planningstage:
1 . What is the goal of the survey,study,or experiment?How can you translatethe goal into a
null hypothesisand an alternativehypothesis?
2. Is the hypothesistest a two-tail test or one-tailtest?
3. can you selecta randomsamplefrom the underlyingpopulationof interest?
4. What kinds of datawill you collect from the sample?Are the variablesnumericalor cate-
gorical?
5. At what significancelevel, or risk of commiuing a Type I error, shouldyou conductthe
hypothesistest?
6. Is the intendedsamplesize large enoughto achievethe desiredpower of the test for the
level of significancechosen?
7. What statisticaltestprocedureshouldyou useand why?
8. What conclusionsand interpretationscanyou reachfrom the resultsof the hypothesistest?
Failing to considerthesequestionsearly in the planningprocesscan lead to biasedor
incompleteresults.Properplanningcanhelp ensurethat the statisticalstudywill provideobjec-
tive informationneededto makegood businessdecisions.
9.16
for selecting
L -l@
One-Sample
test
.
I
tY9"
Cateqorical Numerical
l- D;1"-
lr 6
Tf-.i:".-
g;zr"*t to, tt e No -- -
LP.*"il; f .*n?*n
lir
lJ
t fTest ::
$* *i.t
360 CHAPTERNINE Fundamentals
of Hypothesis
Testing: Tests
One-Sample
In decidingwhich test to use,you should ask the fol- . If the test involves a numerical variable. do you
lowing questions: thepopulationstandard deviation?Ifyou knowthe
ulation standarddeviation.use the Z test for the
. Doesthe test involve a numericalvariableor a categori- you do not know the populationstandarddeviati
cal variable?If the test involvesa categoricalvariable, the I test for the mean.
usethe Ztestfor the proportion. Table9.4 providesa list of topicscoveredin this
ZTestof Hypothesis for the Mean (o Known) One Sample ZTest for the Proportion
X -p, p-n
Z_ (e.1)
o
T=
"tln
/Test of Hypothesis for the Mean (o Unknown) ZTest for the Proportion in Terms
x -1t of the Numberof Successes
(e.2)
,s Z _ ---
X -nn
T, lnn(r - n)
How cana confidenceinterval estimatefor the pop- d. How do changesin the rejection criterion affect the
meanprovide conclusionsto the corresponding probabilitiesof committingTypeI andTypeII errors?
sistestfor the populationmean? 9.85 In 2006,Visa wantedto move awayfrom its long-
running televisionadvertisingthemeof "Visa, it's every-
What is the six-step critical value approachto
where you want to be." During the Winter Olympics,Visa
is testing?
featuredOlympiansin commercialswith a broadermessage,
Whatare someof the ethical issuesinvolvedwith including securiry check cards,and paymenttechnologies
ing a hypothesistest? suchascontactless processing.One of the first commercials
featuredsnowboarderLindsey Jacobellisbeing coachedto
the Concepts calm down before a big race by imaginingthat her Visa
An article in Marketing News (T. T. Semon, CheckCardgot stolen.A key metric for the successof tele-
ider a StatisticalInsignificanceTest," Marketing vision advertisements is the proportionof viewerswho "like
Februaryl,1999) arguedthat the level of signifi- the adsa lot." HarrisAd ResearchServiceconducteda study
usedwhen comparingtwo products is often too of 903 adultswho viewedthe new Visa advertisementand
is, sometimesyou shouldbe using an ct,value reported that 54 indicated that they "like the ad a lot."
than0.05.Specifically,the article recountedtesting According to Harris, the proportion of a typical television
ion of potentialcustomerswith a preferencefor advertisement receivingthe "like the ad a lot" scoreis 0.21
I overproduct2.The null hypothesiswas that the (Extractedfrom T. Howard "Visa to ChangeStrategiesin
ion proportionof potential customerspreferring UpcomingAds," usatoday.com,January23, 2006.).
I was0.50,and the alternativehypothesiswasthat a. Use the six-stepcritical value approachto hypothesis
notequalto 0.50.Thep-valuefor the testwas 0.22. testingand a 0.05 level of significanceto try to prove
articlesuggestedthat in somecases,this shouldbe that the newVisa ad is lesssuccessfulthan a typical tele-
evidenceto rejectthe null hypothesis. vision advertisement.
the null and alternativehypothesesfor this exam- b. Use the five-step p-valueapproachto hypothesistesting
plein statisticalterms. and a 0.05 level of significanceto try to prove that the
Explainthe risks associatedwith Type I and Type II new Visa ad is less successfulthan a typical television
in thiscase. advertisement.
wouldbe the consequences ifyou rejectedthenull c. Comparethe resultsof (a) and(b).
hesisfor a p-valueof 0.22? 9.86 The ownerof a gasolinestationwantsto studygaso-
do you think the article suggestedraising the line purchasinghabitsby motoristsat his station.He selects
of s? a randomsampleof 60 motoristsduring a certainweek,
wouldyou do in this situation? with the following results:
is your answerin (e) if thep-value equals0.12? . The amount purchasedwas .F : ll.3 gallons,
if it equals0.06? ^t: 3.1gallons.
. 1l motoristspurchasedpremium-gradegasoline.
La QuintaMotor Innsdevelopeda computermodelto
predictthe profitability of sitesthat are being consid- a. At the 0.05 level of significance,is there evidencethat
aslocationsfor new hotels.If the computermodelpre- the meanpurchasewas different from l0 gallons?
largeprofits, La Quinta buys the proposedsite and b. Determinethep-valuein (a).
a newhotel.If the computermodelpredictssmall or c. At the 0.05 level of significance,is there evidencethat
profits,La Quintachoosesnot to proceedwith that fewer than 20%oof all the motorists at the station pur-
(Extracted from S. E. Kimes and J. A. Fitzsimmons, chasedpremium-gradegasoline?
ing Profitable Hotel Sites at La Quinta Motor d. What is your answerto (a) if the samplemeanequals
Vol. 20, March-April 1990,pp.12-20). 10.3gallons?
l' Interfaces,
decision-making procedurecan be expressedin the e. What is your answerto (c) if 7 motoristspurchased pre-
is-testingframework.Thenull hypothesisis thatthe mium-grade gasoline?
is nota profitablelocation.The alternativehypothesisis 9.87 An auditor for a governmentagencyis assignedthe
thesiteis a profitablelocation. task of evaluatingreimbursementfor office visits to physi-
ain the risks associated with committinga Type I cianspaid by Medicare.The audit wasconductedon a sam-
in this case. ple of 75 of the reimbursements,
with the following results:
in the risks associatedwith committing a Type II . ln 12 of the office visits, an incorrectamountof
errorin this case. reimbursementwas provided.
Whichtype of error do you think the executivesat La . The amountof reimbursementwas X : S93.70.
intaMotor Inns aretrying hard to avoid?Explain. ^s:s34.55.
362 of Hypothesis
CHAPTERNINE Fundamentals Testing:
One-sample
Tests
a. At the 0.05 level of sigaificance, is there evidencethat 9.9O An important quality characteistic used by the
themeanreimbursement waslessthan$I00? ufacturerof Boston and Vermontasphaltshinglesis
b. At the 0.05 level of significance,is thereevidencethat amount of moisturethe shinelescontain when thev
the proportionof incorrectreimbursements in the popu- packaged.Customersmay feel that they have
lationwasgreaterthan0.10? product lacking in quality if they find moistureand
c. Discusstheunderlyingassumptions of thetestusedin (a). shinglesinsidethe packaging.In somecases,
d. Whatis your (a)
answerto if the samplemeanequals$90? moisturecancausethe sranulesattachedto the shinsle
e. What is your answerto (b) if 15 office visits had incor- textureandcoloringpurposesto fall offthe shingle,
rect reimbursements? ing in appearanceproblems.To monitor the amount
moisturepresent,the companyconductsmoisture
9.88 A bankbranchlocatedin a commercialdistrictof a
s h i n g l e i s w e i g h e d a n d t h e n d r i e d . T h e s h i n g l ei s
city has developedan improvedprocessfor servingcus-
reweighed,and,basedon the amountof moisturetaken
tomersduring the noon-to-1:00p.m. lunch period.The
of the product,the poundsof moistureper 100square
waiting time (definedas the time the customerentersthe
are calculated.The companywould like to showthat
line until he or shereachesthe teller window) of all cus-
meanmoisturecontentis lessthan 0.35 poundsper
tomersduringthis hour is recordedovera periodof I week.
A random sample of 15 customers(see the data file
squarefeet.Thedatafile @@fs includes36
ments(in poundsper 100squarefeet) for Bostonshi
!l@!) is selected,and the resultsare as follows:
and31 for Vermontshinsles.
4.21 5.55 3.02 5.13 4.77 2.34 3.54 3.20 a. For the Boston shingles,is there evidenceat the
4.50 6.10 0.38 5.12 6.46 6.r9 3.79 level of significancethat the mean moisture
a. At the 0.05 level of significance,is thereevidencethat lessthan0.35poundsper 100squarefeet?
the meanwaitingtime is lessthan5 minutes? b. Interpretthe meaningof thep-valuein (a).
b. What assumptionmust hold in order to perform the test c. For the Vermontshingles,is thereevidenceat the
in (a)? level of significancethat the mean moisturecontent
c. Evaluatethe assumptionin (b) through a graphical lessthan0.35poundsper 100squarefeet?
approach. Discuss. d. Interpretthe meaningof thep-valuein (c).
d. As a customerwalks into the branchoffice during the e. What assumptionmusthold in orderto performthe
lunch hour, she asksthe branchmanagerhow long she in (a) and(c)?
canexpectto wait.Thebranchmanagerreplies,'Almost f. Evaluatethe assumptionin (e) for Boston shingles
certainlynot longerthan5 minutes."On the basisof the Vermontshinglesby usinga graphicalapproach.Di
resultsof (a), evaluatethis statement. 9.91 Studiesconductedby the manufacturerof
9.89 A manufacturingcompanyproduceselectricalinsula- and Vermont asphaltshingleshave shownproductwei
tors. If the insulatorsbreakwhen in use,a short circuit is to be a major factor in the customer'sperceptionof
likely to occur.To testthe shengthof the insulators,destruc- Moreover,the weight representsthe amountof raw
tive testing is carried out to determinehow muchforce is als being usedand is thereforevery importantto the
requiredto break the insulators.Force is measuredby pany from a cost standpoint.The last stageofthe
observingthe numberof poundsof force appliedto the line packagesthe shinglesbefore the packagesare
insulatorbeforeit breaks.The following data(storedin the on woodenpallets.Oncea palletis full (a palletfor
[!@! file) arefrom 30 insulatorssubjectto thistesting: brandsholdsl6 squares of shingles),it is weighedand
measurementis recorded.The data file [!fts@
Force(in theNumberof PoundsRequired
the weight (in pounds)from a sampleof 368 pallets
to BreaktheInsulator)
Bostonshinglesand330 palletsof Vermontshingles.
t,870 1,728 1,6561,610 1,6341,784 1,5221,6961,5921,662 a. For the Bostonshingles,is thereevidencethatthe
weightis differentfrom 3,150pounds?
1,8661,764 1,7341,6621,7341,774 1,5501,7561,762 1,866
b. Interpretthe meaningof thep-valuein (a).
1,8201,744 1,7881,6881,8101,752 1,6801,810 1,6521,736 c. For the Vermont shingles,is there evidencethat
meanweight is different from 3,700pounds?
a. At the 0.05 level of significance,is thereevidencethat
d. Interpretthe meaningof thep-valuein (c).
themeanforceis greaterthan 1,500pounds?
e. What assumption is neededfor (a) and(c)?
b. What assumptionmust hold in order to perform the test
f. Evaluatethis assumptionthrougha graphical
in (a)?
Discuss.
c. Evaluatethe assumptionin (b) through a graphical
approach. Discuss. 9.92 The manufacturerof Boston and Vermont
d. Basedon (a), what canyou concludeaboutthe strength shinglesprovidesits customerswith a 20-yearwarranty
of the insulators? most of its products.To determinewhethera shingle
References363
its monitoringof the blacknessof the news- is evidence that the mean blackness is less than 0.97. Write
print, first describedin the Chapter6 Managing a memo to management that summarizes your conclusions.
'Springtille
Herald case,the productiondepartmentof
Blackness of 50 Newspapers
newspaper wantsto ensurethat the meanblacknessof the
for all newspapers is at least0.97 on a standardscale 0 . 8 5 41 . 0 2 31 . 0 0 51 . 0 3 01 . 2 t 90 . 9 7 7r . 0 4 40 . 7 7 8t . t 2 2 1 . t 1 4
ich the targetvalue is 1.0.A randomsampleof 50 1.0911.0861.1410.9310.7230.9341.0601.0470.8000.889
hasbeenselected,and the blacknessof each 1 . 0 1 20 . 6 9 50 . 8 6 90 . 7 3 41 . 1 3 10 . 9 9 30 . 7 6 20 . 8 1 41 . 1 0 80 . 8 0 5
hasbeenmeasured(and storedin the !fi! file). r . 2 2 3 r . 0 2 40 . 8 8 40 . 7 9 90 . 8 7 00 . 8 9 80 . 6 2 10 . 8 1 8l . l l 3 1 . 2 8 6
thesamplestatisticsand determinewhetherthere 1 . 0 5 20 . 6 7 81 . 1 6 20 . 8 0 81 . 0 1 20 . 8 5 90 . 9 5 1L l t 2 1 . 0 0 30 . 9 7 2
I
r--*-; About the lF Function
Both worksheetsof theEEEEEIE*orkbook and all other
hypothesis-testingworksheets presented in the restofthe
TaSWkns
book use the IF (comparison,what to do if comparison
rd T$n-TC Test holds,what to do if comparisonfails) functionto suggest
|* lSpcr-TailTart whether you should reject the null hypothesis.(You
f Uxner-td te* shouldalwaysverify the resultsof worksheetsandnot
blindly acceptresultsreportedby formulascontainingthe
0utfut Options IF function.)
Ttle; T--- You need to supply threethings when using an IF
functionin a formula.The comparisonis an algebraicor
logicalcomparisonbetweentwo things.In FigureE9'1,
thep-valuesarecompared to the levelof significance'
E9.2: Using the I Testfor the Mean (o Unknown) 365
E9.1
FIGURE
1 ZTest oflhe
worksheet
ZMean_All I
3
t
5
b
7
I
I
10 lntermedialeCalculalions
11 -86isQRT{87}
*(88 - B{lrB11
IJ
K 14
f 15 Lowel =l,lORMSll{V{85/21
tb =l{ORlrlSll{V{t- 8512}
f l7 o-Valrte =2 - {1 - }|ORh|SO|ST{ABS{812D}
1g *lF{Bt7 < 85, "Re.lectthe null hypothesis",
'19 'llo rot reJecllhe null hypothesis'J
2n
*HORHSINV{B5l
n -NORIiSDTST{812}
ZJ *lF(822 < 85, "Rej6crtho nrll hypotfteris",
)1 "Do ool rejgct ths null hypolheels']
23
x -xoRtislNv{l - 85)
27 -1 " ilOR['tSDtST{812}
2S -lF{827 < 85, "Rojeclthe n$ll hypolhesl$",
"0o ttot rejectlhc rull hyporhesic'J
E9.2USINGTHE t TESTFOR
SsfiSleStondardDevi&ionr t*
THE MEAN (o unknown)
{*' 5ampleSsistics Ur*q"rary*n
*** --_:
Youconduct a / test for the mean (o unknown) by either
t-*
the PHStat2 t Test for the Mean, sigma unknown
selecting
r-
procedureorby making entries in the@l$frworkbook.
Test0plions
r Two-TdTes*
UsingPHStat2 t Test r-' 1$Pr-TeilTest
for the Mean, Sigma Unknown ,l".'Lower-TnilTe*
SelectPHStat ) One-Sample Tests ) t Test for the
Mean,sigma unknown. In the t Test for the Mean, sigma Odp.rt Options
unknowndialog box (shown at right), enter values for the Titler :
Null Hypothesis and the Level of Significance. lf you
ry
knowthe sample size, sample mean, and sample standard
deviation,click Sample Statistics Known and enter those
- -t*"*-i _.. jl
li_..__a-K csrcd j
---
tochapter
366 ExcELcoMPANIoN 9
FIGUREE9.2 A B
Td for thr linort-b of tlu Xrat
TMeanAllworksheet 2
3 Oata
rnol|r.*
E
lwrl of Slonficarca 0.
5 r Slzr
r larll llit.fi
I r Silndanl D.Yldon m,
9
lntarmsdiste Celculatione
l 1 ilandard Enor ofthg Mean 6.trlt( -B0xsORT{86}
armes of Fmsdom -86-1
Trd Staddc l.tllr -{87 -il}Alr
14
Two.Tdl Tet
1 E ow.r Ctl{crl Vilu. 2mil --(nl{vps,814)
t7 )oa1 ua 2.Z$11 -nilvGt, Br2l
r.Vr||r. 0. -TI,IST(ABS(Bt3), Br2,2)
19 Do nol ral6cl $. null lwooth.dr -lF{818< Bli, 'Rsl.ct tln null lrypolbedr',
n 1)o not r.r..t ilr. null hypo{hrlbJ
21 Loser-Tall Ted
2. Low.r Crl0c.l V!lu. -t -{n|wc.85, Bt2,,
n o-V.lu. 0.t29( -lF(Bl3<0,EA,ffil
Ilo not rel.cl ft. null lmolh* -lFlBil3 < Hi, 'ReJcc{th. nufl lrSpothe*',
x 'Do mt rcJ.c{$. null hypothrdr'}
UDO'I.T'TI Id
zl rDal crldcrl yrlua t.lgt! -milvg.s,8r4)
B o-V!lr|. 0r70a -lF(Bl3< O,E/l.,r7'l
Do nol rcle cl th. null hwolhedr -lF(ElE < 8li, 'R.r.cl lfte null hpolhcrb',
'0o
not rdea rhc mll lrypo&crb'|
llel-6eu
CollEl2: -TIIIST0ABS{B|3},
812,t}
C-cllg}3: -1 -sn
r $pr-f*rr*
sheet(seethe Excel Companionto Chapterl). For a lower-
tail-test-onlyworksheet,selectand deleterows 26 through f uru-rdta*
29 andthen selectand deletecell rangeAl5:B20. For an
upper-tail-test-onlyworksheet,selectand deletecell range
Al5:B25. Theseinstructionsask you to selectand delete , *rf
certaincell rangesin order to preservethe D20:E23calcu-
ot( Crt6l
lationsareacell range(which is not shownin FigureE9.2).
E9.3:UsinctheZTestfbr theProoortion 367
Levelof Significance,the Number of Successes, and the includes both the two-tail test and the upper- and lower-tail
SampleSize.Click one of the test options, enter a title as testson one worksheet.To adapt theseworksheetto other
theTitle,and click OK. problems,change(if necessary)the null hypothesis,level
o f s i g n i f i c a n c e , n u m b e r o f s u c c e s s e sa, n d s a m p l e s i z e
values in the tinted cells 84 through 88. To better under-
Z Proportion.xls
Using stand how messagesget displayed in these worksheets,
'About
Youopenand use either the ZProp-TT or ZProp-All read the the IF Function" part of Section E9.l on
worksheets of theEEEEEEEEIEworkbook Io apply the Z p a g e3 6 4 .
testfor the proportion. These worksheetsuse the lf you want the ZProp-All worksheetto show only one
NORMSINV(P<X) function to determinethe lower and of the single-tail tests, first make a copy of that worksheet
upper criticalvaluesandtheNORMSDIST(Z value)func- (seethe Excel Companionto Chapter 1). For a lower-tail-
tionto compute thep-values. test-only worksheet, select and delete rows 25 through 28
TheZProp_TT worksheet(seeFigure9.15 on page and then select and delete rows 14 through 19. For an
355)applies thetwo-tailZtestrotheSection 9.5competrtive upper-tail-test-onlyworksheet,select and delete rows 14
threat
example. TheZProp-All worksheet (seeFigureE9.3) throush 24.
E9.3
FIGURE
worksheet
ZProp_All
-46187
-soRT(8r11 - 8CFB4
-{810 - B4yB1I
-t{0RMSlt{V{85,2i
ng =NORfttSlNV{l .85,?}
*2' (1 . r'loRilSDlsT{ABs{812)}}
k- *lF{817 < 85, 'Tsjecl lhe null hypotlresi*",
'?o not feject the '|ull hypotbesis'J
-ltoRMsrilv{85}
-NoRllilSDlsT{812}
*lF(822< 85,'?ejedthe null hyporhasls",
"Do nol rajecl lhe null hypcthesis'J
he -l{oRils|l{v{l - 85}
ox *1 - t{oRfrlsDrsT(B12}
*lF{827 < 85, "Rejesttbe null hypothesis".
he "0o not rejecl the null hypothesis"]