MINIAB軟件應(yīng)用手冊英文

INTRODUCTIONTOMINITABVERSION13WorksheetConventionsandMenuStructuresMinitabInteroperabilityGraphicCapabilitiesPareto HistogramBoxPlotScatterPlotStatisticalCapabilitiesCapabilityAnalysisHypothesisTestContingencyTablesANOVADesignofExperiments(DOE)

MinitabTrainingAgenda

WorksheetFormatandStructureSessionWindowWorksheetDataWindowMenuBarToolBarTextColumnC1-T(Designatedby-T)NumericColumnC3(NoAdditionalDesignation)

DataWindowColumnConventionsDateColumnC2-D(Designatedby-D)ColumnNames(Type,Date,Count&AmountEnteredDataforDataRows1through4DataEntryArrowDataRows

OtherDataWindowConventions

MenuBar-MenuConventionsHotKeyAvailable(Ctrl-S)SubmenuAvailable(…attheendofselection)

MenuBar-FileMenuKeyFunctionsWorksheetFileManagementSavePrintDataImport

MenuBar-EditMenuKeyFunctionsWorksheetFileEditsSelectDeleteCopyPasteDynamicLinks

MenuBar-ManipMenuKeyFunctionsDataManipulationSubset/SplitSortRankRowDataManipulationColumnDataManipulation

MenuBar-CalcMenuKeyFunctionsCalculationCapabilitiesColumnCalculationsColumn/RowStatisticsDataStandardizationDataExtractionDataGenerationMenuBar-StatMenuKeyFunctionsAdvancedStatisticalToolsandGraphsHypothesisTestsRegressionDesignofExperimentsControlChartsReliabilityTestingMenuBar-GraphMenuKeyFunctionsDataPlottingCapabilitiesScatterPlotTrendPlotBoxPlotContour/3DplottingDotPlotsProbabilityPlotsStem&LeafPlotsMenuBar-DataWindowEditorMenuKeyFunctionsAdvancedEditandDisplayOptionsDataBrushingColumnSettingsColumnInsertion/MovesCellInsertionWorksheetSettingsNote:TheEditorSelectionisContextSensitive.Menuselectionswillvaryfor:DataWindowGraphSessionWindowDependingonwhichisselected.MenuBar-SessionWindowEditorMenuKeyFunctionsAdvancedEditandDisplayOptionsFontConnectivitySettingsMenuBar-GraphWindowEditorMenuKeyFunctionsAdvancedEditandDisplayOptionsBrushingGraphManipulationColorsOrientationFontMenuBar-WindowMenuKeyFunctionsAdvancedWindowDisplayOptionsWindowManagement/DisplayToolbarManipulation/DisplayMenuBar-HelpMenuKeyFunctionsHelpandTutorialsSubjectSearchesStatguideMultipleTutorialsMinitabontheWebMINITABINTEROPERABILITYMinitabInteroperabilityExcelMinitabPowerPointStartingwithExcel...Loadfile““Sample1”inExcel….StartingwithExcel...ThedataisnowloadedintoExcel….StartingwithExcel...HighlightandCopytheData…….MovetoMinitab...OpenMinitabandselectthecolumnyouwanttopastethedatainto…….MovetoMinitab...SelectPastefromthemenuandthedatawillbeinsertedintotheMinitabWorksheet….UseMinitabtodotheAnalysis...LetssaythatwewouldliketotestcorrelationbetweenthePredictedWorkloadandtheactualworkload…….SelectStat…Regression….FittedLinePlot……..UseMinitabtodotheAnalysis...Minitabisnowaskingforustoidentifythecolumnswiththeappropriatedate…….Clickintheboxfor““Response(Y):Notethatouroptionsnowappearinthisbox.Select““ActualWorkload”andhittheselectbutton…..Thiswillenterthe“ActualWorkload””dataintheResponse(Y)datafield...UseMinitabtodotheAnalysis...NowclickinthePredictor(X):box….Thenclickon“PredictedWorkload””andhittheselectbutton……Thiswillfillinthe““Predictor(X):”datafield...Bothdatafieldsshouldnowbefilled….SelectOK...UseMinitabtodotheAnalysis...Minitabnowdoestheanalysisandpresentstheresults...NotethatinthiscasethereisagraphandananalysissummaryintheSessionWindow…Let’’ssaywewanttousebothinourPowerPointpresentation…….TransferringtheAnalysis...Let’’stakecareofthegraphfirst….GotoEdit….CopyGraph...TransferringtheAnalysis...OpenPowerPointandselectablankslide….GotoEdit….PasteSpecial...TransferringtheAnalysis...Select““Picture(EnhancedMetafile)…Thiswillgiveyouthebestgraphicswiththeleastamountoftrouble.TransferringtheAnalysis...OurMinitabgraphisnowpastedintothepowerpointpresentation….Wecannowsizeandpositionitaccordingly…….TransferringtheAnalysis...NowwecancopytheanalysisfromtheSessionwindow…..Highlightthetextyouwanttocopy…….SelectEdit…..Copy…..TransferringtheAnalysis...Nowgobacktoyourpowerpointpresentation…..SelectEdit…..Paste…..TransferringtheAnalysis...Wellwegotourdata,butitisabitlarge…..Reducethefontto12andweshouldbeok…..Presentingtheresults....Nowallweneedtodoistunethepresentation……..Herewepositionthegraphandsummaryandputintheappropriatetakeaway...Thenwearereadytopresent…….GraphicCapabilitiesParetoChart....Let’’sgenerateaParetoChartfromasetofdata…….GotoFile……OpenProject…….LoadthefilePareto.mpj…….Nowlet’’sgeneratetheParetoChart...ParetoChart....Goto:Stat……QualityTools……ParetoChart…….ParetoChart....Filloutthescreenasfollows:OurdataisalreadysummarizedsowewillusetheChartDefectstable...Labelsin““Category”…Frequenciesin““Quantity””….AddtitleandhitOK..ParetoChart....MinitabnowcompletesourparetoforusreadytobecopiedandpastedintoyourPowerPointpresentation….Histogram....Let’’sgenerateaHistogramfromasetofdata…….GotoFile…OpenProject….Loadthefile2_Correlation.mpj…….Nowlet’sgeneratetheHistogramoftheGPAresults...Histogram....Goto:Graph……Histogram…Histogram....Filloutthescreenasfollows:SelectGPAforourXvalueGraphVariableHitOK…..Histogram....MinitabnowcompletesourhistogramforusreadytobecopiedandpastedintoyourPowerPointpresentation….Thisdatadoesnotlooklikeitisverynormal….Let’suseMinitabtotestthisdistributionfornormality…...Histogram....Goto:Stat…BasicStatistics…DisplayDescriptiveStatistics…….Histogram....Filloutthescreenasfollows:SelectGPAforourVariable….SelectGraphs…..Histogram....SelectGraphicalSummary….SelectOK…..SelectOKagainonthenextscreen...Histogram....NotethatnowwenotonlyhaveourHistogrambutanumberofotherdescriptivestatisticsaswell….Thisisagreatsummaryslide...Asforthenormalityquestion,notethatourPvalueof.038rejectsthenullhypothesis(P<.05).So,weconcludewith95%confidencethatthedataisnotnormal…..Histogram....Let’slookatanother“Histogram”toolwecanusetoevaluateandpresentdata….GotoFile……OpenProject…….Loadthefileoverfill.mpj…….Histogram....Goto:Graph……MarginalPlot…Histogram....Filloutthescreenasfollows:Selectfiller1fortheYVariable….SelectheadfortheXVariableSelectOK…..Histogram....NotethatnowwenotonlyhaveourHistogrambutadotplotofeachheaddataaswell...Notethatheadnumber6seemstobethesourceofthehighreadings…..ThistypeofHistogramiscalleda““MarginalPlot”..Boxplot....Let’slookatthesamedatausingaBoxplot….Boxplot....Goto:Stat…BasicStatistics…DisplayDescriptiveStatistics...Boxplot....Filloutthescreenasfollows:Select““filler1””forourVariable…….SelectGraphs……..Boxplot....SelectBoxplotofdata…….SelectOK……..SelectOKagainonthenextscreen...Boxplot....WenowhaveourBoxplotofthedata...Boxplot....ThereisanotherwaywecanuseBoxplotstoviewthedata...Goto:Graph……Boxplot...Boxplot....Filloutthescreenasfollows:Select“filler1”forourYVariable….Select“head”forourXVariable….SelectOK…..Boxplot....Notethatnowwenowhaveaboxplotbrokenoutbyeachofthevariousheads..Notethatheadnumber6againseemstobethesourceofthehighreadings…..Scatterplot....Let’slookatdatausingaScatterplot….GotoFile……OpenProject…….Loadthefile2_Correlation.mpj…….Nowlet’sgeneratetheScatterplotoftheGPAresultsagainstourMathandVerbalscores...Scatterplot....Goto:Graph……Plot...ScatterPlot....Filloutthescreenasfollows:SelectGPAforourYVariable….SelectMathandVerbalforourXVariables…..SelectOKwhendone...Scatterplot....WenowhavetwoScatterplotsofthedatastackedontopofeachother……Wecandisplaythisbetterbytilingthegraphs….Scatterplot....Todothis:GotoWindow…Tile...Scatterplot....NowwecanseebothScatterplotsofthedata……Scatterplot....Thereisanotherwaywecangeneratethesescatterplots….Goto:Graph……MatrixPlot...ScatterPlot....Filloutthescreenasfollows:Clickinthe“Graphvariables”blockHighlightallthreeavailabledatasets…Clickonthe“Select””button...SelectOKwhendone...Scatterplot....WenowhaveaseriesofScatterplots,eachonecorrespondingtoacombinationofthedatasetsavailable…NotethatthereappearstobeastrongcorrelationbetweenVerbalandbothMathandGPAdata….MinitabStatisticalToolsPROCESSCAPABILITYANALYSISLet’sdoaprocesscapabilitystudy…….OpenMinitabandloadthefileCapability.mpj…….SETTINGUPTHETEST….GotoStat……QualityTools…….CapabilityAnalysis(Weibull)….Select“Torque”foroursingledatacolumn...Enteralowerspecof10andanupperspecof30.Thenselect“OK”….SETTINGUPTHETEST….Notethatthedatadoesnotfitthenormalcurveverywell...NotethattheLongTermcapability(Ppk)is0.43.ThisequatestoaZvalueof3*0.43=1.29standarddeviationsorsigmavalues.ThisequatestoanexpecteddefectratePPMof147,055.INTERPRETINGTHEDATA….HYPOTHESISTESTINGLoadthefilenormality.mpj……..SettingupthetestinMinitabCheckingtheDataforNormality….It’simportantthatwecheckfornormalityofdatasamples.Let’sseehowthisworks…….GotoSTAT….BasicStatistics...NormalityTest….SetuptheTestWewilltestthe““Before””columnofdata…….CheckAnderson-DarlingClickOKAnalyzingtheResultsSincethePvalueisgreaterthan.05wecanassumethe““Before”dataisnormalNowrepeatthetestforthe““After”Data(thisislefttothestudentasalearningexercise..)Checkingforequalvariance..Wenowwanttoseeifwehaveequalvariancesinoursamples.Toperformthistest,ourdatamustbe“stacked”.ToaccomplishthisgotoManip…Stack…StackColumns….Selectbothoftheavailablecolumns(BeforeandAfter)tostack....Typeinthelocationwhereyouwantthestackeddata…….InthisexamplewewilluseC4….Typeinthelocationwhereyouwantthesubscriptsstored……InthisexamplewewilluseC3….SelectOK…….Checkingforequalvariance..Nowthatwehaveourdatastacked,wearereadytotestforequalvariances.…GotoStat…ANOVA…….TestforequalVariances...Checkingforequalvariance..Settingupthetest….Ourresponsewillbetheactualreceiptperformanceforthetwoweekswearecomparing.InthiscasewehadputthestackeddataincolumnC4….Ourfactorsisthelabelcolumnwecreatedwhenwestackedthedata(C3)..WesetourConfidenceLevelforthetest(95%).Thenselect““OK””.Here,weseethe95%confidenceintervalsforthetwopopulations.Sincetheyoverlap,weknowthatwewillfailtorejectthenullhypothesis.TheFtestresultsareshownhere.WecanseefromtheP-Valueof.263thatagainwewouldfailtorejectthenullhypothesis.NotethattheFtestassumesnormalityNotethatwegetagraphicalsummaryofbothsetsofdataaswellastherelevantstatistics…….Analyzingthedata…….Levene’’stestalsocomparesthevarianceofthetwosamplesandisrobusttononnormaldata.Again,theP-Valueof.229indicatesthatwewouldfailtorejectthenullhypothesis.Herewehaveboxplotrepresentationsofbothpopulations.Letstestthedatawitha2SampletTest--UnderStat…BasicStatistics….Weseeseveralofthehypothesistestswhichwediscussedinclass.Inthisexamplewewillbeusinga2SampletTest…….GotoStat….BasicStatistics..2Samplet…..Sincewealreadyhaveourdatastacked,wewillloadC4foroursamplesandC3foroursubscripts.Settingupthetest….Sincewehavealreadytestedforequalvariances,wecancheckoffthisbox……NowselectGraphs….Settingupthetest….Weseethatwehavetwooptionsforourgraphicaloutput.Forthissmallasample,Boxplotswillnotbeofmuchvaluesoweselect““Dotplotsofdata”andhit““OK””.HitOKagainonthenextscreen…….Inthesessionwindowwehaveeachpopulation’sstatisticscalculatedforus..NotethatherewehaveaPvalueof.922.Wethereforefindthatthedatadoesnotsupporttheconclusionthatthereisasignificantdifferencebetweenthemeansofthetwopopulations...Interpretingtheresults…….Thedotplotshowshowclosethedatapointsinthetwopopulationsfalltoeachother.Theclosevaluesofthetwopopulationmeans(indicatedbytheredbar)alsoshowslittlechancethatthishypothesiscouldberejectedbyalargersampleInterpretingtheresults….PairedComparisonsInpairedcomparisonswearetryingto“pair”observationsortreatments.Anexamplewouldbetotestautomaticbloodpressurecuffsandanursemeasuringthebloodpressureonthesamepatientusingamanualinstrument.Itcanalsobeusedinmeasurementsystemstudiestodetermineifoperatorsaregettingthesamemeanvalueacrossthesamesetofsamples.Let’’slookatanexample:2_Hypothesis_Testing_Shoe_wear.mpj2_Hypothesis_Testing_Shoe_wear.mpjInthisexamplewearetryingtodetermineifshoematerial““A””wearrateisdifferentfromshoematerial““B””.Ourdatahasbeencollectedusingtenboys,whomwereaskedtowearoneshoemadefromeachmaterial.Ho:Material““A””wearrate=Material““B””wearrateHa:Material““A””wearrateMaterial““B””wearratePairedComparisonGotoStat….BasicStatistics……Pairedt……..PairedComparisonSelectthesamples…GotoGraphs…….PairedComparisonSelecttheBoxplotforourgraphicaloutput..ThenselectOK..PairedComparisonWeseehowthe95%confidenceintervalofthemeanrelatestothevaluewearetesting.Inthiscase,thevaluefallsoutsidethe95%confidenceintervalofthedatamean.Thisgivesusconfirmationthattheshoematerialsaresignificantlydifferent.CONTINGENCYTABLES(CHISQUARE)Enteringthedata….Enterthedatainatableformat.Forthisexample,loadthefileContingencyTable.mpj...Let’’ssetupacontingencytable….ContingencytablesarefoundunderStat…….Tables…ChiSquareTest…….Selectthecolumnswhichcontainthetable.Thenselect““OK””Settingupthetest….Notethatyouwillhavethecriticalpopulationandteststatisticsdisplayedinthesessionwindow.Minitabbuildsthetableforyou.Notethatouroriginaldataispresentedanddirectlybelow,Minitabcalculatestheexpectedvalues.Here,MinitabcalculatestheChiSquarestatisticforeachdatapointandtotalstheresult.ThecalculatedChiSquarestatisticforthisproblemis30.846.PerformingtheAnalysis….ANalysisOfVArianceANOVALet’ssetuptheanalysisLoadthefileAnovaexample.mpj…StackthedatainC4andplacethesubscriptsinC5Setuptheanalysis….SelectStat…ANOVA……Oneway…SelectC4ResponsesC5FactorsThenselectGraphs….Setuptheanalysis….Chooseboxplotsofdata...ThenOKSetuptheanalysis….NotethatthePvalueislessthan.05thatmeansthatwerejectthenullhypothesisAnalyzingtheresults…….Let’’sLookAtMainEffects…….ChooseStatANOVAMainEffectsPlot…….MainEffectsSelectC4ResponseC5FactorsOKAnalyzingMainEffects..Formulation1HasLowestFuelConsumptionDESIGNOFEXPERIMENTS(DOE)FUNDAMENTALSFirstCreateanExperimentalDesign...GotoStat……DOE……Factorial...CreateFactorialDesign...FirstCreateanExperimentalDesign...Select2LevelFactorialdesignwith3factorsThengotoDisplayAvailableDesigns…….BowlingExample(continued)Wecannowseetheavailableexperimentaldesigns…….WewillbeusingtheFull(Factorial)for3factorsandwecanseethatitwillrequire8runs……Now,selectOKandgobacktothemainscreen.OnceatthemainscreenselectDesigns...BowlingExample(continued)Selectyourdesign…….WewillbeusingtheFull(Factorial)andagainwecanseethatitwillrequire8runs……Now,selectOKandgobacktothemainscreen.OnceatthemainscreenselectFactors...BowlingExample(continued)Fillinthenamesforyourfactors….Thenfillintheactualconditionsforlow(-)orhigh(+)Now,selectOKandgobacktothemainscreen.OnceatthemainscreenselectOptions...BowlingExample(continued)RemovetheoptiontoRandomizeRuns…….Now,selectOKandgobacktothemainscreen.OnceatthemainscreenselectOK...BowlingExample(continued)Minitabhasnowdesignedourexperimentforus…….Now,typeyourDatafromeachofyourexperimentaltreatmentsintoC8.Wearenowreadytoanalyzetheresults……BowlingExample(continued)GotoStat…….DOE……Factorial...AnalyzeFactorialDesign...BowlingExample(continued)HighlightyourDatacolumnanduseSelecttoplaceitintheResponsesbox.Then,selecttheTermsOption.BowlingExample(continued)NotethatSelectedTermshasalloftheavailablechoicesalreadyselected.Weneeddonothingfurther.SelectOK.Then,atthemainscreenselectGraphsBowlingExample(continued)SelectyourEffectsPlotsandresetyourAlphato.05.SelectOKtoreturntothemainscreenandthenselectOKagain.BowlingExample(continued)Notethatonlyoneeffecthasa