当代市场调研10E全套配套课件中英文PPTch14.ppt 23页

ChapterFourteen14-1 Census:Populationcanvas-notreallya“sample”AskingtheentirepopulationBudgetAvailable:Avalidfactor–howmuchcanweafford?RuleofThumb:Istheresomeconventionwecanapply?Whatmightmakeanadequatesamplesize?SampleSizeforProbabilitySampling14-2KeyTerms&Definitions NumberofSubgroupsAnalyzed:Inanysamplesizedeterminationproblem,considerationmustbegiventothenumberandanticipatedsizeofvarioussubgroupsofthetotalsamplethatmustbeanalyzed.TraditionalStatisticalMethods:Variance,standarddeviation,andconfidenceintervalplayakeyrole.SampleSizeforProbabilitySampling14-3KeyTerms&Definitions Judgment:Bestguessof“experts”DrawonyourexperiencetodeterminesamplesizeConventional:Whathaveothersdone?SeewhatthesamplesizehasbeenforsimilarstudiesSampleSizeforNonprobabilitySampling14-4KeyTerms&Definitions CentralLimitTheorem:Theideathatadistributionofalargenumberofsamplemeansorsampleproportionswillapproximateanormaldistribution-regardlessofthedistributionofthepopulationfromwhichtheyweredrawn.NormalDistribution:Thecontinuousdistributionthatisbellshapedandsymmetricalaboutthemean.Themean,median,andmodeareequal.Finally,about68%oftheobservationsarewithinonestandarddeviationplus/minusofthemean,96%arewithintwostandarddeviations,and99+%arewithinthreestandarddeviationsofthemeanrespectively.TheNormalDistribution14-5KeyTerms&Definitions ProportionatePropertiesoftheNormalDistribution:Afeaturethatthenumberofobservationsfallingbetweenthemeanandagivennumberofstandarddeviationsfromthemeanisthesameforallnormaldistributions.StandardNormalDistribution:Normaldistributionwithameanofzeroandastandarddeviationofone.TheNormalDistribution14-6KeyTerms&Definitions StandardDeviation:Themeasureofdispersioncalculatedbysubtractingthemeanoftheseriesfromeachvalueinaseries,squaringeachresult,summingtheresults,dividingthesumbythenumberofobservationsminus1andfinallytakingthesquarerootofthisvalue.TheNormalDistribution14-7KeyTerms&Definitions ImportantCharacteristicsofaNormalDistribution14-8KeyTerms&DefinitionsBell-shapedandhasonlyonemode.SymmetricaboutitsmeanUniquelydefinedbyitsmeanandstandarddeviationThetotalareaunderanormalcurveisequaltooneTheareaofaregionunderthenormaldistributioncurvebetweenanytwovaluesofavariableequalstheprobabilityofobservingavalueinthatrangewhenanobservationisrandomlyselectedfromthedistribution.Theareabetweenthemeanandagivennumberofstandarddeviationsfromthemeanisthesameforallnormaldistributions. PopulationDistribution:Thefrequencydistributionofalltheelementsofapopulation.SamplingDistribution:Thefrequencydistributionofalltheelementsofanindividualsample.PopulationandSamplingDistribution14-9KeyTerms&Definitions SamplingDistributionoftheMean:Thetheoreticalfrequencydistributionofthemeansofallpossiblesamplesofagivensizedrawnfromaparticularpopulation;itisnormallydistributed.SamplingDistributionoftheMean14-10KeyTerms&Definitions StandardErroroftheMean:Standarddeviationofadistributionofsamplemeans.StandardErroroftheMean14-11KeyTerms&Definitions RelationshipsoftheThreeBasicTypesofDistribution14-12KeyTerms&Definitions PointEstimate:Theparticularestimateofapopulationvalue.IntervalEstimate:Theintervalorrangeofvalueswithinwhichthetruepopulationvalueisestimatedtofall.PointandIntervalEstimations14-13KeyTerms&Definitions ConfidenceLevel:Theprobabilitythataparticularintervalwillincludethetruepopulationvalue;alsocalledtheconfidencecoefficient.ConfidenceInterval:Theintervalthat,atthespecifiedconfidencelevel,includesthetruepopulationvalue.ConfidenceLevelandInterval14-14KeyTerms&Definitions SamplingDistributionoftheProportion:Therelativefrequencydistributionofthesampleproportionsofmanyrandomsamplesofagivensizedrawnfromaparticularpopulation;itisnormallydistributed.SamplingDistributionoftheProportion14-15KeyTerms&Definitions ProblemsInvolvingMean:DeterminingtheSamplingSize14-16KeyTerms&Definitions DeterminingtheSamplingSizeAllowableSamplingError:Amountofsamplingerrortheresearcheriswillingtoaccept,E.2.TheAcceptableLevelofConfidence:Howconfidentdoestheresearcherwanttobethatanintervalincludesthepopulationmean,Z.3.PopulationStandardDeviation:Standarddeviationofavariablefortheentirepopulation,σ.14-17KeyTerms&Definitions ProblemsInvolvingProportions:DeterminingtheSamplingSize14-18KeyTerms&Definitions PopulationSizeandSamplingSizeIndependenceAssumption:Assumptionthatsampleelementsaredrawnindependently.14-19KeyTerms&Definitions FinitePopulationCorrectionFactor(FPC)StatisticalPower:TheprobabilityofnotmakingatypeIIerror.14-20KeyTerms&Definitions Youneedtocontactmorethann,sinceit’sunlikelythateveryoneyoucontactwillagreetoanswerthequestionnaire.Ifyoudeterminethatyouneed800respondentsandatypicalresponserateforyourtypeofsurveyis30%,thenyouwouldactuallyhavetocontact2,667peopletoget800completedsurveys.HowManyContacts?14-21KeyTerms&DefinitionsNumberofContacts=n/.30Or800/.3NumberofContacts=2,667Youmustcontactapproximately2,667people(with30%responserate),soyoucanexpect800peopletorespond(thefinalsamplesize). TimetoGeneratetheSampleScopeoftheResearchBudgetAvailableExperiencewithSamplingLevelofAccuracyDesiredYourKnowledgeofthePopulationSamplingIssuesToConsider14-22 14-23KeyTerms&DefinitionsCentralLimitTheoremNormalDistributionProportionalPropertyoftheNormalDistributionStandardNormalDistributionSampleSizeforNonprobabilitySamplingStandardDeviationPopulationDistributionSampleDistributionSamplingDistributionoftheMeanLinksandbuttonareactivewhenin“SlideShowMode”KeyTerms&DefinitionsStandardErroroftheMeanPointEstimateIntervalEstimateConfidenceLevelConfidenceIntervalSamplingDistributionoftheProportionAllowableSamplingErrorPopulationStandardDeviationFinitePopulationCorrectionFactor(FPC)StatisticalPower