管理統(tǒng)計學課件

第1章概率論

BasicProbabilityProbabilityDistributions本章概要

BasicProbabilityConcepts:

SampleSpacesandEvents,Simple Probability,andJointProbability(聯(lián)合概率)。

ConditionalProbability(條件概率)

BayesianTheorem（貝葉斯定理）

TheProbabilityofaDiscreteRandomVariable

Binomial,Poisson,andHypergeometricDistributions

Covariance(協(xié)方差)anditsApplicationsinFinanceSampleSpaces（樣本空間）

CollectionofallPossibleOutcomes

e.g.All6facesofadie:

e.g.All52cardsofabridgedeck:

Events（事件）SimpleEvent:OutcomefromaSampleSpace with1Characteristic

e.g. A

RedCard

fromadeckofcards.

JointorCompoundEvent:Involves2Outcomes

Simultaneously（同時發(fā)生稱為“Joint”）

e.g. AnAcewhichisalsoa

RedCard

fromadeckofcards. AnAcegiventhatitisa

RedCard.

VisualizingEvents（形象化事件）ContingencyTables（列聯(lián)表）TreeDiagrams

（樹狀圖） AceNotAceTotal

Red22426Black22426Total44852SimpleEvents（簡單事件）

TheEventofaHappyFaceThereare5happyfacesinthiscollectionof18objectsJointEvents（聯(lián)合事件---交）TheEventofaHappyFaceAND

LightColored3HappyFaceswhicharelightincolor12Items,5happyfacesand7otherlightobjectsCompoundEvents（複合事件---並）TheEventofHappyFaceORLightColored

SpecialEvents（特殊事件）

NullEvent（無效事件）

Club&Diamondon

1CardDraw

Complement（補集）

of

EventForEventA, AllEventsNotInA:Null

Event除非它是次品補集的文氏圖Theevent“notA”happenswheneverAdoesnotVenndiagram:A(incircle),“not

A”(shaded)Prob(notA)=1–Prob(A)IfProb(Succeed)=0.7,thenProb(Fail)=1–0.7=0.3AnotA3Items,3HappyFacesGiventheyareLightColoredDependentorIndependentEvents相依或獨立事件給定的TheEventofaHappyFaceGIVENitisLightColoredE=HappyFace

LightColorContingencyTable（列聯(lián)表）ADeckof52Cards（去掉大、小王）

Ace(么點)NotanAceTotalRedBlackTotal224224262644852SampleSpace(樣本空間)RedAceTreeDiagram（樹圖）EventPossibilitiesRedCardsBlackCardsAceNotanAceAceNotanAceFullDeckofCardsProbability（概率）Probability

isthenumericalmeasureofthelikelihoodthattheeventwilloccur.Valueisbetween0

and1.Sum

oftheprobabilitiesofallmutuallyexclusive（互斥的）eventsis

1.Certain必然的Impossible不可能的.510計算事件的概率----關鍵確定有效事件與全部事件TheProbabilityofanEvent,E:EachoftheOutcomeintheSampleSpace equallylikelytooccur(等概率發(fā)生).e.g.P(

)=2/36(Thereare2waystogetone6andtheother4)有效事件數(shù)樣本空間事件總數(shù)

JointProbability（聯(lián)合概率）TheProbabilityofa

JointEvent,A

and

B:

e.g.P(RedCard

and

Ace)=P(AandB)=NumberofEventOutcomesfrombothAandBTotalOutcomesinSampleSpaceA與B同時發(fā)生交集的文氏圖IntersectionhappenswheneverbotheventshappenVenndiagram:Intersection“Aand

B”shaded)e.g.,A

=“signcontract”,B=“getfinancing”Didtheintersectionhappen?Great!Projecthasbeenlaunched!e.g.,DidIhaveeggsandcerealforbreakfast?“No”ABP(A2

and

B2)P(A1

and

B2)P(A2andB1)P(A1

and

B1)EventEventTotal

Total1交集事件的聯(lián)合概率JointProbabilityMarginal(Simple)Probability邊際概率P(A1)A1A2B1B2P(A2)P(B1)P(B2)CompoundProbability（並集事件概率）“或”的關係TheProbabilityofa

CompoundEvent,AorB:

e.g.P(RedCard

orAce)

並集的文氏圖Unionhappenswhenevereither(orboth)happenVenndiagram:Union“Aor

B”shaded)e.g.,A

=“getInteljoboffer”,B=“getGMjoboffer”Didtheunionhappen?Congratulations!Youhaveajobe.g.,DidIhaveeggsorcerealforbreakfast?“Yes”ABA+B重複部分理應減去P(A1

andB1)P(B2)P(B1)P(A2andB2)P(A2and

B1)EventEventTotal

Total1CompoundProbability

AdditionRule（可加性準則）P(A1

andB2)P(A1)A1A2B1B2P(A2)P(A1

orB1)=P(A1)+P(B1)-P(A1and

B1)ForMutuallyExclusiveEvents:P(AorB)=P(A)+P(B)

互斥的或不相容或不相交事件ConditionalProbability（條件概率）TheProbabilityoftheEvent:Event

A

given

thatEvent

B

hasoccurred

P(A

B)=e.g.P(RedCardgiventhatitisan

Ace)=若將A、B對換，有效與全部事件數(shù)會怎樣變化？條件概率（續(xù)）GiventheextrainformationthatBhappensforsure,howmustyouchangetheprobabilityforAtocorrectlyreflectthisnewknowledge?Thisisa(conditional)probabilityaboutATheeventBgivesinformationUnconditionalTheprobabilityofAConditionalAnewuniverse,sinceBmusthappenProb(A

given

B)=Prob(A

and

B)Prob(B)ABAandBB現(xiàn)實中的條件概率Prob(WingivenAheadathalftime)HigherthanProb(Win)evaluatedbeforethegamebeganProb(SucceedgivenGoodresultsintestmarket)HigherthanProb(Succeed)evaluatedbeforemarketingstudyProb(GetjobgivenPoorinterview)LowerthanProb(GetjobgivenGoodinterview)Prob(HaveAIDSgivenTestpositive)HigherthanProb(HaveAIDS)forthepopulation-at-largeBlackColorTypeRedTotalAce224Non-Ace242448Total262652條件概率的列聯(lián)表分析ConditionalEvent:Draw1Card.NoteKind&ColorRevisedSampleSpace

條件概率及事件的獨立性

ConditionalProbability:P(A

B)=P(AandB)=P(A

B)

P(B)EventsareIndependent:P(A

B)=P(A)Or,P(AandB)=P(A)P(B)EventsAandBareIndependentwhentheprobabilityofoneevent,Aisnotaffectedbyanotherevent,B.MultiplicationRule:貝葉斯公式

后驗概率P(Bi

A)=AddingupthepartsofAinalltheBi事件Bj之間獨立且構成必然事件的完全劃分；分母利用全概率公式。Whatarethechancesofrepayingaloan,givenacollegeeducation?

貝葉斯公式的列聯(lián)表分析LoanStatusEducationRepayDefaultProb.College.2.05.25NoCollege

Prob.1

P(Repay

College)=

????

?

拖欠DiscreteRandomVariable

(離散型隨機變數(shù))RandomVariable:representsoutcomesofanexperiment.

e.g.Throwadietwice:

Countthenumberoftimes

4comesup(0,1,or2times)DiscreteRandomVariable:

ObtainedbyCounting(0,1,2,3,etc.)Usuallyfinitebynumberofvaluese.g.Tossacoin5times.Countthenumberoftails.

(0,1,2,3,4,or5times)

離散事件實例ProbabilityDistributionValues

Probability

0

1/4=.25

1

2/4=.50

2

1/4=.25Event:Toss2Coins.

CountofTails.TTTT離散事件的概率分佈Listof

AllPossible

[Xi,P(Xi)]

Pairs

Xi=ValueofRandomVariable(Outcome) P(Xi)=ProbabilityAssociatedwithValueMutuallyExclusive

(NoOverlap)CollectivelyExhaustive

(NothingLeftOut)

0

P(Xi)

1

P(Xi)=1離散隨機變數(shù)的數(shù)字特徵ExpectedValue（數(shù)學期望值）

TheMeanoftheProbabilityDistributionWeightedAverage

=E(X)

=

XiP(Xi)

e.g.Toss2coins,Counttails,ComputeExpectedValue:

=0

.25+1

.50+2

.25=1Variance（方差）

WeightedAverageSquaredDeviationaboutMean

=E[(Xi-

)2]=

(Xi-

)2P(Xi)

e.g.Toss2coins,Counttails,ComputeVariance:

=(0-1)2(.25)+(1-1)2(.50)+(2-1)2(.25)=.50NumberofTails常見的離散(變數(shù))概率分佈DiscreteProbabilityDistributionsBinomial二項分佈Hypergeometric超幾何分佈Poisson泊松分佈二項分佈的性質（1）2IdenticalTrials,e.g.15tossesofacoin, 10lightbulbstakenfromawarehouse2MutuallyExclusiveOutcomes,

e.g.headsortailsineachtossofacoin, defectiveornotdefectivelightbulbsConstantProbability

foreachTrial,

e.g.probabilityofgettingatailisthesame eachtimewetossthecoinandeach lightbulbhasthesameprobabilityof being defective

二項分佈的性質（2）

2SamplingMethods:

InfinitePopulationWithoutReplacement FinitePopulationWithReplacement

TrialsareIndependent:

TheOutcomeofOneTrialDoesNotAffectthe OutcomeofAnother二項分佈函數(shù)(Function)P(X)=probabilitythatX

successesgivenaknowledgeofn andp

X

=numberofsuccesses?in

sample,(X=0,1,2,...,

n)

p = probabilityofsuccess

n = samplesizeP(X)nX!nXppXnX!()!()

1

Tailsin2TossofCoin

X

P(X)

0 1/4=.251 2/4=.502 1/4=.25

二項分佈的數(shù)字特徵

標準差n=5

p=0.1n=5

p=0.5Mean（均值）StandardDeviation

EXnpnpp

()()10.2.4.6012345XP(X).2.4.6012345XP(X)e.g.

=

5(.1)=

.5e.g.

=

5(.5)(1-.5)

=1.118

0泊松分佈

區(qū)間PoissonProcess:

DiscreteEvents

inaninterval

TheProbabilityof

OneSuccess

in IntervalisStable

TheProbabilityofMorethanOne SuccessinthisIntervalis0ProbabilityofSuccessis

Independent

fromIntervaltoInterval

e.g.#CustomersArrivingin15min.#DefectsPerCaseofLight Bulbs.PXxxx(|!

e-泊松分佈函數(shù)P(X)=probabilityof

X

successesgiven

= expected(mean)numberofsuccessese = 2.71828(baseofnaturallogs)X = numberofsuccesses

perunitPXXX()!

ee.g.Findtheprobabilityof

4customersarrivingin3minuteswhenthemeanis

3.6.P(X)

=e-3.63.64!4=

.1912泊松分佈的數(shù)字特徵

=0.5

=6MeanStandardDeviation

iiNiEXXPX

=1/

()()10.2.4.6012345XP(X)0.2.4.60246810XP(X)

超幾何分佈

試驗

n

Trials

inaSampleTakenFroma

FinitePopulation

ofsize

N

Sampletaken

Without

Replacement

Trialsare

Dependent總體

ConcernedWithFindingthe

Probability

ofX?Successesinthe

Sample

wherethereareA?Successesin the

Population

超幾何分佈函數(shù)P(X)=probabilitythat

X

successesgiven

n,

N,and

An

= samplesizeN

= populationsize

A

=numberofsuccesses

inpopulationX

= numberofsuccesses

insample(X=0,1,2,...,

n)PX)()()AX(

()N

-

An

-

XNn3

Lightbulbswereselected

from10.Ofthe

10

therewere

4defective.Whatistheprobabilitythat

2ofthe

3selectedaredefective?P(2)=()()426

110

3(

)=.30(組合數(shù)超幾何分佈的數(shù)字特徵MeanStandardDeviation

EXnnA

N

A

()N

nN

1FinitePopulationCorrectionANN2隨機變數(shù)的協(xié)方差(Covariance)

X=discreterandomvariableXXi=

ithoutcomeofXP(XiYi)=probabilityofoccurrenceofthe ithoutcomeofYY=discreterandomvariableYYi=ithoutcomeofYi=1,2,…,N數(shù)字特徵應用：平均投資回報

MeanforInvestmentReturnsReturnper$1,000fortwotypesofinvestmentsP(XiYi) Economiccondition DowJonesfundXGrowthStockY

.2 Recession -$100

-$200

.5 StableEconomy +100

+50.3 ExpandingEconomy +250

+350

InvestmentE(X)=

X=(-100)(.2)+(100)(.5)+(250)(.3)=$105E(Y)=

Y=(-200)(.2)+(50)(.5)+(350)(.3)=$90投資回報的方差P(XiYi) Economiccondition DowJonesfundXGrowthStockY

.2 Recession -$100

-$200

.5 StableEconomy +100

+50.3 ExpandingEconomy +250

+350

InvestmentVar(X)==(.2)(-100-105)2+(.5)(100-105)2+(.3)(250-105)2

=14,725,

X=121.35Var(Y)

=

=(.2)(-200-90)2+(.5)(50-90)2+(.3)(350-90)2

=37,900,

Y=194.68投資回報的協(xié)方差P(XiYi) Economiccondition DowJonesfundXGrowthStockY

.2 Recession -$100

-$200

.5 StableEconomy +100

+50.3 ExpandingEconomy +250

+350

Investment

XY

=(.2)(-100

-

105)(-200-90)

+(.5)(100-

105)(50

-

90)

+(.3)(250

-105)(350

-

90)=

23,300TheCovarianceof23,000indicatesthatthetwoinvestmentsarestronglyrelatedandwillvarytogetherinthesamedirection.本章小結DiscussedBasicProbabilityConcepts:

SampleSpacesandEvents,SimpleProbability, andJointProbabilityDefined

ConditionalProbabilityDiscussed

BayesianTheoremAddressedtheProbabilityofa

DiscreteRandom

Variable

Discussed

Binomial,Poisson,andHypergeometric DistributionsAddressedCovarianceanditsApplicationsinFinance

第3章統(tǒng)計概述SummaryofStatistics本章概要StatisticalThinkingandManagementDescriptive

versusInferentialStatisticsTypesofDataandtheirSourcesTypesofSamplingMethodsTypesofSurveyErrorsStatisticalThinkingandManagement統(tǒng)計思想與管理ThreeAspectsofQualityImprovementManagementPhilosophy管理哲學BehavioralTools行為工具StatisticalMethods統(tǒng)計方法StatisticalMethodsDescriptiveStatisticsInferentialStatisticsCollectinganddescribingdata.Makingdecisionsbasedonsampledata.

統(tǒng)計描述統(tǒng)計推斷統(tǒng)計描述收集資料、整理資料、描述資料CollectData

e.g.SurveyPresentData

e.g.TablesandGraphsCharacterizeData

e.g.Mean

ACharacteristicofa: (1)PopulationisaParameter(2)SampleisaStatistic.統(tǒng)計推斷參數(shù)估計、假設檢驗EstimationHypothesisTestingMakingdecisionsconcerningapopulationbasedonsampleresults.TypesofData分類資料、數(shù)字資料（離散、連續(xù)）DataSources（觀測、調研、實驗）（數(shù)據(jù)編輯、出版及電子介質）PrimaryDataCollectionSecondaryDataCompilationObservationExperimentationSurveyPrintorElectronicQuotaTypesofSamplingMethods抽樣方法類型SamplesNon-ProbabilitySamples非概率抽樣JudgementChunkProbabilitySamples概率抽樣SimpleRandomSystematicStratifiedCluster概率抽樣ProbabilitySamplesSimpleRandomSystematicStratifiedClusterSubjectsofthesamplearechosenbasedonknownprobabilities.SimpleRandomSamples簡單隨機抽樣Everyindividualoritemfromthetargetframehasan

equalchance

ofbeingselected.Selectionmaybe

withreplacement

or

withoutreplacement.

Onemayusetableofrandomnumbersforobtainingsamples.SystematicSamples系統(tǒng)抽樣Decideonsamplesize:nDividepopulationofNindividualsintogroupsof

k

individuals:

k

=N/n

Randomlyselectoneindividualfromthe1stgroup.Selecteveryk-thindividualthereafter.

N=64n=8k=8FirstGroupStratifiedSamples分層抽樣Populationdividedinto2ormoregroupsaccording tosomecommoncharacteristic.

Simplerandomsampleselectedfromeach.Thetwoormoresamplesarecombinedintoone.ClusterSamples整群抽樣Populationdividedintoseveral

clusters?

eachrepresentativeofthepopulation.

Simplerandomsampleselectedfromeach.Thesamplesarecombinedintoone.Populationdividedinto4clusters.TypesofSurveyErrors調研誤差類型CoverageErrorNonResponseErrorSamplingErrorMeasurementErrorExcludedfromselection.Followupon nonresponses.

Chancedifferencesfrom

sample

tosample.

BadQuestion!本章小結Describedtheuseof

StatisticalThinking

toimprovequality.Addressedthenotionof

Descriptiveversus

InferentialStatistics.Definedanddescribeddifferent

TypesofData

and

SourcesListedTypesof

SamplingMethods.DescribeddifferentTypesof

SurveyErrors.第4章問卷設計

QuestionnaireDesign本章概要（1）Tolearntheobjectivesofquestionnairedesign.Tounderstandtheroleofthequestionnaireinthedatacollectionprocess.Tobecomefamiliarwiththecriteriaforagoodquestionnaire.Tolearntheprocessforquestionnairedesign.本章概要（2）Tobecomeknowledgeableofthethreebasicformsofquestions.Tounderstandthekeyroleofthequestionnaireindatacollectioncosts.Tolearnthenecessaryproceduresforsuccessfulimplementationofasurvey.QuestionnaireDefined

問卷的定義Aquestionnaireisasetofquestionsdesignedtogeneratethedatanecessaryforaccomplishingtheobjectivesoftheresearchproject.1、問卷是數(shù)據(jù)收集方法的一般思路；2、問卷提供了標準化和統(tǒng)一化的數(shù)據(jù)收集程式；3、問卷在數(shù)據(jù)收集過程中起著重要作用。TheQuestionnaire’sPositionintheResearchProcess

問卷在調研過程中的地位QuestionnaireSurveyObjectivesRespondent’sInformationDataAnalysisFindingsRecommendationsManagerialActionCriteriaforaGoodQuestionnaire

一份優(yōu)秀問卷的標準Todesignagoodquestionnaire,thefollowingissuesshouldbeconsidered:DoesitProvidetheNecessaryDecision-MakingInformation?DoesitConsidertheRespondent?

Doesitsatisfyediting,codingandprocessingdata?EditingandDataProcessingRequirements

編輯與數(shù)據(jù)處理的要求EditingReferstogoingthroughthequestionnairetomakecertaintheskippatternsarefollowedandrequiredquestionsarefilledout.Askippatternisthesequenceinwhichquestionsareasked.Open-EndedQuestions（開放式問題）Anopen-endedquestionisonethatdoesnotcontainprerecordedpossibleresponses.TheQuestionnaireDevelopmentProcess

問卷的設計過程DetermineSurveyObjectivesandConstraints（1）DetermineDataCollectionMethodsEvaluatetheQuestionnaireandLayoutDescribeQuestionnaireFlowandLayoutDecideQuestionWordingDetermineQuestionResponseFormatObtainApprovalfromallRelevantPartiesPretestandRevisePrepareFinalCopyImplementation（10）StepOne:DetermineSurveyObjectives,Resources,andConstraints

確定調研目的、來源和限制因素Survey(information)objectivesshouldbespelledoutasclearlyandpreciselyaspossible.Ifthisstepiscompletedinathoroughfashion,therestoftheprocesswillfollowmoresmoothlyandefficiently.StepTwo:DeterminetheDataCollectionMethod

確定數(shù)據(jù)收集方法Chapter6discussedthevarietyofwaysthatsurveydatacanbegathered,suchasin-person,telephone,mail,orself-administration.Eachmethodwillhaveanimpactonsurveydesign.StepThree:DeterminetheQuestionResponseFormat

確定問題的回答形式（1）Open-EndedQuestions（開放式問題）Open-endedquestionsarethoseinwhichtherespondentcanreplyinhisorherownwords.Exampleofanopen-endedquestion:1.Whatadvantages,ifany,doyouthinkorderingfromamailordercatalogcompanyofferscomparedwithlocalretailoutlets?StepThree:DeterminetheQuestionResponseFormat

確定問題回答形式（2）Closed-EndQuestions（封閉式問題）Aclosed-endedquestionisonethatrequirestherespondenttomakeaselectionfromalistofoptions.Theprimaryadvantageofclosed-endedquestionsissimplytheavoidanceofmanyoftheproblems(suchassubjectivity)ofopen-endedquestions.StepThree:DeterminetheQuestionResponseFormat

確定問題回答形式(3)DichotomousQuestions(二項式問題)Thissimplestformofaclosed-endedquestionisthedichotomouschoice.Exampleofadichotomousquestion:1.Wouldyouratherliveinasmalltownoracity?A.SmallTown1B.City2StepThree:DeterminetheQuestionResponseFormat

確定問題回答形式（4）Multiple-ChoiceQuestions（多選題）Multiple-ChoiceQuestionspermittherespondenttoselectfromseveralalternatives.Exampleofamultiple-choicequestion:1.WhodoyouthinkwillwintheWorldSeries?A.NewYorkYankeesB.St.LouisCardinalsC.LosAngelesDodgersD.AtlantaBravesStepThree:DeterminetheQuestionResponseFormat

確定問題回答形式（5）Scaled-ResponseQuestions(量表應答式問題)Questionformatthatpermitsthemeasurementofthe搃ntensity?ofarespondents?answers.Exampleofascaled-responsequestion:1.NowthatyouhavereadasampleoftheWallStreet

Journal,wouldyousaythatyouwould?Checkone)DefinitelybuyitProbablybuyitMightormightnotbuyitProbablywillnotbuyitDefinitelywillnotbuyitStepFour:DecidetheQuestionWording

決定問題的用詞FourGuidelinesinSelectingWordingThewordingmustbeclear.Selectwordssoastoavoidbiasingtherespondent.Considertheabilityoftherespondenttoanswerthequestion.Considerthewillingnessoftherespondenttoanswerthequestion.StepFive:EstablishQuestionnaireFlowandLayout

確定問卷流程和編排Afterthequestionshavebeenproperlyformulated,thenextstepistosequencethemanddevelopalayoutforthequestionnaire.Thegeneralguidelinesconcerningquestionnaireflowareincludedinthenextslide.HowaQuestionnaireShouldbeOrganized

問卷如何組織UsethescreenerquestionstoidentifyqualifiedrespondentsAfterobtainingaqualifiedrespondent,beginwithaquestionthatobtainsarespondent’sinterestAskgeneralquestionsfirstPositionsensitive,threatening,anddemographicquestionsastheendInsertpromptersatstrategicpointsAskquestionsthatrequireworksinthemiddleofthequestionnaireStepSix:EvaluatetheQuestionnaire

問卷評價（1）IssuestoConsiderinReviewingtheDesignoftheQuestionnaireTheissuestoconsiderinreviewingthedesignofthequestionnaireareshownonthenextslide:StepSix:EvaluatetheQuestionnaire

問卷評價（2）IstheQuestionNecessary?IstheQuestionnaireTooLong?WilltheQuestionsProvidetheDesiredInformationtoAccomplishtheResearchObjectives?AppearancesofMailandSelf-AdministeredQuestionnairesAvoidaClutteredLookAllowPlentyofSpaceforOpen-EndedResponsesConsiderColor-CodingtheQuestionnairesInstructionsPrintedWithinTheQuestionnaireShouldbeinCapitalLettersStepSeven:ObtainApprovalofAllRelevantParties

獲得各方面的認同Atthispointinthequestionnairedesignprocess,thefirstdraftofthequestionnairehasbeencompleted.Approvalforgoingforwardwiththequestionnaireistheobjectiveofthisstep.

StepEight:PretestandRevise

預先測試和修訂Whenfinalmanagerialapprovalhasbeenobtained,thequestionnairemustbepretested.Theprimarypurposeofapretestistomakecertainthatthequestionnairegivestherespondentclear,understandablequestionsthatwillevokeclear,understandableresponses.StepNine:PrepareFinalCopy

準備最後印刷Eventhefinalcopyphasedoesnotallowtheresearchertorelax.Precisetypinginstructions,spacing,numbering,andprecodingmustbesetup,monitored,andproofread.StepTen:ImplementingtheSurvey

實施調研（1）Supervisor’sInstructions（管理者說明）Informthemofthenatureofthestudy,startandcompletiondates,quotas,reportingtimes,equipmentandfacilityrequirements,samplinginstructions,numberofinterviewersrequired,andvalidationprocedures.Interviewer’sInstructionsCovermanyofthesamepointsassupervisor’sinstructionsbutaregearedtotheactualinterview.StepTen:ImplementingtheSurvey

實施調研（2）CallRecordSheets（通話記錄單）Interviewers?logslistingthenumberandresultsofacontact.Callrecordsheetsareusedtomeasuretheefficiencyoftheinterviewers.VisualAidsandOtherSupplements

（可視化輔助材料和其他附錄）Manystudiesusevisualaidstofacilitatetheinterviewingprocess.StepTen:ImplementingtheSurvey

實施調研（3）FieldManagementCompanies(實地調研)FieldmanagementcompaniessuchasQfact,On-LineCommunications,andDirectResourcegenerallyprovidequestionnaireformatting,screenerwriting,developmentofinstructionalandperipheralmaterials,shippingdepartments,fieldauditing,andallcoordinationofdatacollection,coding,andtabservicesrequiredfortheproject.QuestionnaireChallenges

問卷在成本和獲利性方面的作用（1）TheDifficultyofFindingaQualifiedRespondentinaCentralLocationTelephoneInterview1.FailedAttempts -Busy -NoAnswer -AnsweringMachine -BusinessNumber -Phone/LanguageProblem -DiscontinuedLine2.CooperationProblems -RespondentNotatHome -RefusedtobeInterviewed QuestionnaireChallenges

（2）TheDifficultyofFindingaQualifiedRespondentinaCentralLocationTelephoneInterview3.ScreenerDeterminesRespondentNotEligible -Failedsecuritytest(worksformarketingresearchfirm,advertisingagency,ortheclient) -Doesn抰usetheproduct -Demographicdisqualification(wronggender,age,etc.) -Quotafilled(surveyhasquotaof500usersofTideand500usersofotherclothingwashingpowder.Intervieweralreadyhas500Tideusers;thecurrentrespondentusesTide.4.RespondentTerminatesDuringInterview5.Completedinterview

本章小結（1）Thischapterexaminestheobjectivesofthequestionnaire,aswellasitsconstruction,evaluation,andimpactonprojectcostsandprofitability.Thecriteriaforagoodquestionnairefallintothefollowingtopicarea:(1)achievingthegoalsofthestudy;(2)fittingthequestionnairetotherespondent;and(3)editing,coding,anddataprocessing.本章小結（2）Theprocessofdevelopingaquestionnaireisdiscussed.Theprocesscontains10sequentialsteps.Thechapterconcludeswithadiscussionoftheroleofthequestionnaireinsurveyresearchcosts.第5章抽樣技術

SamplingTechnology一、基本抽樣問題基本概念制定抽樣計畫的步驟抽樣誤差與非抽樣誤差概率抽樣方法非概率抽樣方法本節(jié)概要（1）Tounderstandtheconceptofsampling.Tolearnthestepsindevelopingasamplingplan.Tounderstandthedifferencesbetweenprobabilitysamplesandnonprobabilitysamples.Tounderstandtheconceptsofsamplingerrorandnonsamplingerror.本節(jié)概要（2）Toreviewthetypesofprobabilitysamplingmethods.Togaininsightintononprobabilitysamplingmethods.基本概念Population（總體）Thetotalgroupofpeoplefromwhominformationisneeded.Census（普查）Dataobtainedfromorabouteverymemberofthepopulationofinterest.Sample（樣本）Asubsetofthepopulationofinterest.StepsinDevelopingaSamplePlan

設計一個抽樣方案的步驟Step1:DefinethePopulationofInterestStep2:ChooseDataCollectionMethodStep3:ChooseSamplingFramesStep4:SelectaSamplingMethodStep5:DetermineSampleSizeStep6:DevelopandSpecifyOperationalPlanStep7:ExecuteOperationalSamplingPlanStepsinDevelopingaSamplingPlan

（1）Step1:DefiningthePopulationofInterestBasesfordefiningthepopulationofinterestinclude:GeographyDemographicsUseAwarenessStepsinDevelopingaSamplingPlan

（2）Step2:ChoosingaDataCollectionMethodTheselectionofadatacollectionmethodhasimplicationsforthesamplingprocess.Step3:ChoosingaSamplingFrameSam