6-1 - QSM 754 Minitab PowerPoint Slides v8

单击此处编辑母版标题样式,单击此处编辑母版文本样式,第二级,第三级,第四级,第五级,#,INTRODUCTION TO MINITAB VERSION 13,1,WorksheetConventionsand MenuStructures,MinitabInteroperability,Graphic Capabilities,Pareto,Histogram,BoxPlot,Scatter Plot,StatisticalCapabilities,Capability Analysis,Hypothesis Test,ContingencyTables,ANOVA,DesignofExperiments,(,(DOE),MinitabTrainingAgenda,WorksheetFormat andStructure,Session Window,WorksheetDataWindow,Menu Bar,Tool Bar,Text ColumnC1-T,(Designatedby,-,-T),Numeric ColumnC3,(NoAdditionalDesignation),Data WindowColumnConventions,Date ColumnC2-D,(Designatedby,-,-D),ColumnNames,(Type,Date,Count,&,&Amount,Entered Data for Data Rows 1 through 4,Data Entry Arrow,Data Rows,OtherData WindowConventions,Menu Bar,-,- MenuConventions,HotKey Available,(,(Ctrl-S),Submenu Available,(,(attheend of selection),Menu Bar,-,- FileMenu,KeyFunctions,WorksheetFileManagement,Save,Print,Data Import,Menu Bar,-,- EditMenu,KeyFunctions,WorksheetFileEdits,Select,Delete,Copy,Paste,Dynamic Links,Menu Bar,-,-ManipMenu,KeyFunctions,Data Manipulation,Subset,/,/Split,Sort,Rank,RowDataManipulation,ColumnDataManipulation,Menu Bar,-,-CalcMenu,KeyFunctions,CalculationCapabilities,ColumnCalculations,Column,/,/Row Statistics,Data Standardization,Data Extraction,Data Generation,Menu Bar,-,-StatMenu,KeyFunctions,AdvancedStatisticalToolsandGraphs,Hypothesis Tests,Regression,DesignofExperiments,Control Charts,ReliabilityTesting,Menu Bar,-,- Graph Menu,KeyFunctions,Data Plotting Capabilities,Scatter Plot,TrendPlot,BoxPlot,Contour/3Dplotting,DotPlots,ProbabilityPlots,Stem &LeafPlots,Menu Bar,-,-Data WindowEditorMenu,KeyFunctions,AdvancedEdit andDisplay Options,Data Brushing,ColumnSettings,ColumnInsertion/Moves,Cell Insertion,WorksheetSettings,Note:TheEditor SelectionisContext Sensitive.Menuselectionswill varyfor:,Data Window,Graph,Session Window,Dependingonwhichisselected,.,.,Menu Bar,-,-Session WindowEditorMenu,KeyFunctions,AdvancedEdit andDisplay Options,Font,ConnectivitySettings,Menu Bar,-,-GraphWindowEditorMenu,KeyFunctions,AdvancedEdit andDisplay Options,Brushing,GraphManipulation,Colors,Orientation,Font,Menu Bar,-,- WindowMenu,KeyFunctions,AdvancedWindowDisplayOptions,WindowManagement/Display,Toolbar Manipulation/Display,Menu Bar,-,- HelpMenu,KeyFunctions,Help andTutorials,Subject Searches,Statguide,MultipleTutorials,MinitabontheWeb,MINITAB INTEROPERABILITY,18,MinitabInteroperability,Excel,Minitab,PowerPoint,Startingwith Excel.,.,.,Load file,“,“Sample1”inExcel,.,.,Startingwith Excel.,.,.,Thedataisnowloadedinto Excel.,Startingwith Excel.,.,.,Highlightand Copythe Data,.,Move toMinitab.,OpenMinitabandselect thecolumnyou wanttopastethe datainto,.,.,Move toMinitab.,SelectPastefromthemenuandthe datawillbeinsertedinto theMinitabWorksheet,.,UseMinitabtodotheAnalysis,.,.,Lets saythat we would liketotestcorrelationbetween thePredictedWorkloadand theactualworkload,.,SelectStat Regression,.FittedLinePlot.,.,.,UseMinitabtodotheAnalysis,.,.,Minitabisnowasking forustoidentifythecolumnswith theappropriatedate.,Clickinthebox for,“,“Response,(,(Y,),):Notethat ouroptions nowappearinthisbox,.,.,Select,“,“ActualWorkload”and hittheselect button,.,.,This willenterthe “ActualWorkload,”,” datainthe Response (Y)data field.,.,.,UseMinitabtodotheAnalysis,.,.,Nowclickinthe Predictor,(,(X),:,: box,.,.Then click on “PredictedWorkload,”,” andhitthe selectbutton,This willfillinthe,“,“Predictor,(,(X),:,:”data field.,.,.,Both datafields shouldnowbefilled.,SelectOK,.,.,UseMinitabtodotheAnalysis,.,.,Minitabnowdoestheanalysisand presents theresults.,.,.,Note thatinthiscase there is agraphand an analysis summaryinthe SessionWindow,Let,ssaywewanttousebothinourPowerPointpresentation,.,Transferringthe Analysis.,.,.,Let,stake careofthe graph first.,GotoEdit.CopyGraph,.,.,Transferringthe Analysis.,.,.,Open PowerPointand selecta blank slide.,GotoEdit.Paste Special,.,.,Transferringthe Analysis.,.,.,Select,“,“Picture,(,(Enhanced Metafile)Thiswillgive youthebestgraphicswith theleastamountoftrouble.,Transferringthe Analysis.,.,.,OurMinitabgraphisnowpasted intothepowerpointpresentation,.Wecan nowsize andpositionitaccordingly,.,.,Transferringthe Analysis.,.,.,Nowwecan copythe analysis fromthe Sessionwindow.,.,.,Highlightthetext youwant to copy,.,SelectEdit,.,.Copy,.,.,Transferringthe Analysis.,.,.,Nowgobacktoyourpowerpointpresentation,.,SelectEdit,.,.Paste,.,Transferringthe Analysis.,.,.,Well we gotourdata,but it is abit large.,.,.,Reducethe fontto12and we shouldbeok,.,.,Presenting theresults.,.,.,Nowall we needtodoistunethepresentation.,.,.,Here we position thegraphandsummaryandput in theappropriatetakeaway.,.,.,Then we arereadytopresent.,Graphic Capabilities,37,ParetoChart,.,.,.,.,Let,sgeneratea ParetoChartfrom aset of data,.,GotoFileOpenProject.Loadthe filePareto.mpj.,Nowletsgeneratethe ParetoChart.,.,.,ParetoChart,.,.,.,.,Goto:,Stat,Quality Tools,ParetoChart,.,ParetoChart,.,.,.,.,Fill outthescreen as follows,:,:,Ourdataisalready summarizedsowewillusethe Chart Defectstable,.,.,Labelsin,“,“Category”,Frequenciesin,“,“Quantity,”,”.,Addtitleand hitOK.,.,.,ParetoChart,.,.,.,.,Minitabnowcompletes ourparetofor us ready to be copiedandpasted intoyourPowerPoint presentation,.,.,Histogram,.,.,.,.,Let,sgeneratea Histogramfrom aset of data,.,GotoFileOpenProject.Loadthe file2_Correlation.mpj.,Nowletsgeneratethe HistogramoftheGPA results,.,.,Histogram,.,.,.,.,Goto:,Graph,Histogram,Histogram,.,.,.,.,Fill outthescreen as follows,:,:,SelectGPA forourXvalueGraphVariable,HitOK,.,Histogram,.,.,.,.,Minitabnowcompletes ourhistogramfor us ready to be copiedandpasted intoyourPowerPoint presentation,.,.,This datadoesnotlooklike it is verynormal.,Let,suseMinitabtotest thisdistribution fornormality,.,.,.,Histogram,.,.,.,.,Goto:,Stat,BasicStatistics,Display Descriptive Statistics,.,Histogram,.,.,.,.,Fill outthescreen as follows,:,:,SelectGPA forourVariable,.,SelectGraphs.,.,.,Histogram,.,.,.,.,SelectGraphical Summary,.,SelectOK,.,SelectOKagainonthe nextscreen.,.,.,Histogram,.,.,.,.,Note thatnow we notonly haveour Histogrambutanumberofotherdescriptivestatisticsaswell.,This is agreatsummaryslide.,.,.,Asforthe normalityquestion,notethat ourP value of .038rejectsthenullhypothesis (P.05),.,.So,weconcludewith95%confidencethat thedata is notnormal,.,Histogram,.,.,.,.,Let,slook at another,“,“Histogram”toolwecanuse to evaluate andpresent data,.,GotoFileOpenProject.Loadthe fileoverfill,.,.mpj.,Histogram,.,.,.,.,Goto:,Graph,MarginalPlot,Histogram,.,.,.,.,Fill outthescreen as follows,:,:,Selectfiller 1for theY Variable.,Selectheadforthe XVariable,SelectOK,.,Histogram,.,.,.,.,Note thatnow we notonly haveour Histogrambutadotplotofeach headdataaswell.,.,.,Note thatheadnumber6seemstobethesource of thehigh readings.,.,.,This typeofHistogram is calleda “MarginalPlot”.,.,.,Boxplot.,.,.,Let,slook at thesame datausingaBoxplot.,Boxplot.,.,.,Goto:,Stat,BasicStatistics,Display Descriptive Statistics,.,.,Boxplot.,.,.,Fill outthescreen as follows,:,:,Select,“,“filler1”forour Variable.,SelectGraphs.,.,.,Boxplot.,.,.,SelectBoxplotofdata.,SelectOK,.,SelectOKagainonthe nextscreen.,.,.,Boxplot.,.,.,WenowhaveourBoxplotofthedata.,.,.,Boxplot.,.,.,Thereisanother waywecanuseBoxplotstoview thedata.,.,.,Goto:,Graph,Boxplot.,Boxplot.,.,.,Fill outthescreen as follows,:,:,Select,“,“filler1”forour YVariable,.,Select,“,“head,”,” forourXVariable,.,.,SelectOK,.,Boxplot.,.,.,Note thatnow we nowhave abox plotbroken outbyeach of thevarious heads.,Note thatheadnumber6againseemstobethesource of thehigh readings.,.,.,Scatter plot,.,.,.,.,Let,slook at datausingaScatterplot.,GotoFileOpenProject.Loadthe file2_Correlation.mpj.,NowletsgeneratetheScatterplotoftheGPA resultsagainstourMathandVerbal scores.,.,.,Scatter plot,.,.,.,.,Goto:,Graph,Plot.,.,.,Scatter Plot,.,.,.,.,Fill outthescreen as follows,:,:,SelectGPA forourYVariable,.,.,SelectMathandVerbal forourXVariables,.,SelectOKwhendone.,.,.,Scatter plot,.,.,.,.,WenowhavetwoScatterplotsofthedatastacked on topofeach other,Wecandisplaythis betterbytilingthe graphs,.,.,Scatter plot,.,.,.,.,Todothis:,GotoWindow,Tile.,.,.,Scatter plot,.,.,.,.,Nowwecan seeboth Scatterplotsofthe data,Scatter plot,.,.,.,.,Thereisanother waywecangeneratethesescatterplots,.,.,Goto:,Graph,MatrixPlot.,.,.,Scatter Plot,.,.,.,.,Fill outthescreen as follows,:,:,Clickinthe,“,“Graph variables”block,Highlightall three availabledata sets,Clickonthe,“,“Select”button.,.,.,SelectOKwhendone.,.,.,Scatter plot,.,.,.,.,Wenowhavea seriesofScatter plots,each onecorresponding to acombinationofthedata setsavailable,Note thatthereappearstobea strongcorrelationbetween VerbalandbothMath andGPAdata,.,.,Minitab Statistical Tools,71,PROCESS CAPABILITYANALYSIS,72,Let,sdoa processcapabilitystudy,.,.,OpenMinitabandloadthefileCapability.mpj.,SETTING UP THETEST.,GotoStat QualityTools,.Capability Analysis (Weibull).,Select,“,“Torque”for oursingledatacolumn,.,.,Entera lower specof10and an upper specof30,.,.Then select,“,“OK,”,”.,SETTING UP THETEST.,Note thatthe datadoesnotfit thenormalcurveverywell.,.,.,Note thatthe LongTermcapability (Ppk) is 0,.,.43,.,.This equatestoaZ value of 3,*,*0.43=1.29 standard deviationsorsigmavalues.,This equatestoanexpecteddefect ratePPM of 147,055,.,.,INTERPRETINGTHE DATA,.,HYPOTHESIS TESTING,77,Load thefilenormality,.,.mpj.,Setting up thetest inMinitab,CheckingtheDataforNormality.,Its importantthat we check fornormalityofdatasamples.,Let,sseehow thisworks,.,GotoSTAT.BasicStatistics.,.,.NormalityTest,.,.,Setupthe Test,Wewill testthe “Before,”,” columnofdata.,CheckAnderson-Darling,ClickOK,Analyzingthe Results,SincethePvalueisgreater than,.,.05 we canassumethe “Before,”,” dataisnormal,Nowrepeat thetest forthe,“,“After”Data (this is lefttothe studentasalearningexercise.,.,.),Checkingforequalvariance,.,.,Wenowwanttoseeifwehaveequalvariancesinour samples,.,.,Toperform thistest,our datamustbe,“,“stacked”,.,.,Toaccomplish thisgotoManipStackStackColumns,.,.,Selectbothoftheavailable columns,(,(BeforeandAfter,),) to stack.,.,.,Type in thelocationwhereyouwantthestackeddata.In thisexamplewewill useC4,.,.,Type in thelocationwhereyouwantthesubscriptsstored, In thisexamplewewill useC3,.,.,SelectOK,.,Checkingforequalvariance,.,.,Nowthatwehave ourdata stacked, we arereadytotest forequalvariances,.,.,GotoStatANOVA,.,.Test forequalVariances,.,.,Checkingforequalvariance,.,.,Setting up thetest.,Ourresponsewillbetheactual receiptperformancefor thetwoweeksweare comparing.In thiscasewehadput thestacked dataincolumn C4.,Ourfactorsisthelabelcolumn we createdwhenwestacked thedata (C3),.,.,Wesetour ConfidenceLevelfor thetest (95%,),).,Then select,“,“OK,”,”.,Here,weseethe 95%confidence intervalsforthe twopopulations.Since theyoverlap,weknowthat we willfailtorejectthe nullhypothesis.,TheFtest resultsare shown here,.,.Wecansee fromthe P,-,-Valueof,.,.263that again we would failtoreject thenull hypothesis,.,. NotethattheFtest assumesnormality,Note thatweget agraphical summaryofbothsets of dataaswellastherelevantstatistics,.,.,Analyzingthe data,.,Levene,stest alsocomparesthe variance of thetwosamplesandisrobust tononnormaldata.Again, theP-Value of .229indicates thatwewouldfailtorejectthe nullhypothesis.,Here we havebox plotrepresentations of bothpopulations,.,.,Lets testthe datawitha 2Sample tTest,-,UnderStat Basic Statistics,.Weseeseveralofthehypothesistestswhichwediscussedinclass,.,.Inthis examplewewillbeusinga 2Sample tTest,.,.,GotoStat.BasicStatistics.2Sample t,.,Sincewealready haveour datastacked,wewillload C4 foroursamplesandC3for oursubscripts.,Setting up thetest.,Sincewehave alreadytested forequalvariances, we cancheckoffthisbox,Nowselect Graphs,.,.,Setting up thetest.,Weseethatwehave twooptions forourgraphical output.Forthis small asample,Boxplotswill notbeofmuch value so we select,“,“Dotplotsofdata”andhit “OK”,.,.Hit OK again on thenext screen,.,.,Inthesessionwindowwehaveeach population,sstatistics calculatedfor us.,Note thatherewehave aPvalueof,.,.922.Wetherefore findthatthedatadoes notsupport theconclusion thatthereisasignificantdifference betweenthe means of thetwopopulations,.,.,Interpretingthe results,.,Thedotplotshowshowclosethedatapointsinthetwo populations falltoeachother.Theclosevaluesofthe twopopulation means (indicatedbythe redbar,),) alsoshowslittle chancethat thishypothesiscouldberejectedbya largersample,Interpretingthe results,.,PairedComparisons,Inpairedcomparisonsweare tryingto,“,“pair”observations or treatments,.,. An examplewouldbetotestautomaticbloodpressurecuffsand anursemeasuring thebloodpressureonthesamepatient using amanual instrument,.,.,Itcanalsobeused in measurement systemstudies to determineifoperatorsare gettingthe samemeanvalueacrossthe sameset of samples,.,.,Let,slook atanexample:2_Hypothesis,_,_Testing_Shoe_wear,.,.mpj,2_Hypothesis,_,_Testing_Shoe_wear,.,.mpj,Inthisexample we aretryingtodetermine if shoematerial,“,“A,”,” wearrateisdifferentfromshoe material “B”.,Ourdatahasbeencollectedusingten boys, whomwereaskedtowear oneshoe madefromeach material.,Ho: Material “A”wear rate,=,=Material,“,“B”wearrate,Ha: Material “A”wear rate Material “B”wear rate,PairedComparison,GotoStat.,BasicStatistics,Pairedt,.,.,PairedComparison,Selectthe samples,GotoGraphs,.,PairedComparison,SelecttheBoxplotforour graphicaloutput,.,.,Then selectOK.,.,.,PairedComparison,Weseehow the95%confidenceintervalofthemeanrelates to thevaluewearetesting.In thiscase,the value falls outsidethe 95%confidence interval of thedata mean,.,.This gives us confirmationthat theshoe materialsaresignificantlydifferent,.,.,CONTINGENCYTABLES,(,(CHISQUARE,),),98,Enteringthedata,.,.,Enterthedataina table format.Forthis example, loadthe fileContingencyTable,.,.mpj.,Let,ssetupacontingencytable,.,.,Contingencytablesare found underStat.TablesChiSquare Test,.,Selectthe columnswhichcontainthetable,.,.Then select,“,“OK,”,”,Setting up thetest.,Note thatyou willhavethecriticalpopulationandteststatistics displayedinthesessionwindow,.,.,Minitabbuildsthe table foryou,.,.Note thatour original dataispresented anddirectlybelow,Minitabcalculates theexpectedvalues,.,.,Here,Minitabcalculates theChiSquare statisticforeachdata point andtotalsthe result.Thecalculated ChiSquarestatistic forthis problemis30,.,.846.,Performing theAnalysis,.,.,ANalysisOfVAriance,ANOVA,Let,ssetupthe analysis,Load thefileAnovaexample.mpj,StackthedatainC4andplacethe subscriptsinC5,Setupthe analysis.,SelectStat,ANOVA,Oneway,Select,C4Responses,C5Factors,Then selectGraphs,.,Setupthe analysis.,Chooseboxplotsofdata.,.,.,Then OK,Setupthe analysis.,Note thatthe Pvalueislessthan .05,that means thatwereject thenull hypothesis,Analyzingthe results,.,Let,sLook At MainEffects,.,.,ChooseStat,ANOVA,Main EffectsPlot,.,.,Main Effects,Select,C4Response,C5Factors,OK,AnalyzingMainEffects.,Formulation1 HasLowestFuelConsumption,DESIGNOFEXPERIMENTS,(,(DOE)FUNDAMENTALS,112,FirstCreateanExperimental Design.,.,.,Goto,Stat,DOE,Factorial,.,.,CreateFactorial Design.,.,.,113,FirstCreateanExperimental Design.,.,.,Select2LevelFactorialdesign with3factors,Then go to DisplayAvailable Designs,.,114,Bowling Example,(,(continued),Wecannow seetheavailable experimentaldesigns.WewillbeusingtheFull(Factorial)for3factors andwecansee thatitwillrequire 8runs,Now, selectOKandgobacktothemainscreen,.,.,Once at themain screenselectDesigns.,115,Bowling Example,(,(continued),Selectyourdesign,.,Wewill be using theFull(Factorial)andagainwecan seethat it willrequire8 runs,Now, selectOKandgobacktothemainscreen,.,.,Once at themain screenselectFactors.,116,Bowling Example,(,(continued),Fill inthenamesfor yourfactors,.,.,Then fillinthe actualconditions forlow,(,(-,),) or high,(,(+,),),Now, selectOKandgobacktothemainscreen,.,.,Once at themain screenselectOptions.,117,Bowling Example,(,(continued),RemovetheoptiontoRandomize Runs,.,Now, selectOKandgobacktothemainscreen,.,.,Once at themain screenselectOK.,118,Bowling Example,(,(continued),Minitabhasnow designed ourexperiment forus,.,.,Now, typeyourData fromeachofyour experimentaltreatments intoC8,.,.,Wearenow ready to analyzethe results,119,Bowling Example,(,(continued),Goto,Stat.,DOE,Factorial,.,.,Analyze FactorialDesign,.,.,120,Bowling Example,(,(continued),HighlightyourData columnanduseSelecttoplaceitintheResponsesbox,.,.,Then,selecttheTermsOption,.,.,121,Bowling Example,(,(continued),Note thatSelectedTermshas alloftheavailable choicesalreadyselected.We needdonothingfurther.,SelectOK.,Then,atthemainscreenselectGraphs,122,Bowling Example,(,(continued),SelectyourEffects Plots andresetyour Alpha to .05.,SelectOKtoreturntothe mainscreen andthen selectOKagain.,123,Bowling Example,(,(continued),Note thatonlyoneeffect hasa significancegreater than95,%,%.,Allthe remainingfactors andinteractionsare notstatistically significant.,124,Bowling Example,(,(continued),Another waywecanlookatthedataistolook at theFactorialPlotsofthe resultingdata.,Goto,DOE,.,Factorial,FactorialPlots,.,125,Bowling Example,(,(continued),SelectMainEffects Plotand thenSetup,126,Bowling Example,(,(continued),SelectC8asyourresponse,Select,“,“Wristband”, “Ball”and,“,“Lane,”,”asyour factors,.,.,Then select,“,“OK,”,” andOKagainonthemainscreen,.,.,127,Bowling Example,(,(continued),Themagnitude of theverticaldisplacementindicates thestrengthofthemaineffectfor thatfactor.Hereweseethatthewristband hasdramaticallymoreeffectthananyotherfactor.Weknowfrom ourearlier plots thatthe wristbandistheonlystatistically significant effect, 95%confidence.,This plotalsoshowsyouthe directionofthemaineffects.Weclearlyseethatthe,“,“with,”,” conditionisrelated to thehigherlevelofperformance,.,.,128,Bowling Example,(,(continued),Nowletslook at theinteractions,.,.,.,.,Goto,DOE,.,Factorial,FactorialPlots,129,Bowling Example,(,(continued),SelectInteractionPlotandthenSetup,.,.,130,Bowling Example,(,(continued),SelectC8asyourresponsevariable.,Select,“,“Wristband”, “Ball”and,“,“Lane,”,”asyour factors,.,.,Then select,“,“OK,”,” andOKagainonthenextscreen,.,131,Bowling Example,(,(continued),Themorethelinesdivergefrom being parallel,themoretheinteraction,.,.,Weseethatthestrongest interaction (stillnotsignificant,),) is betweenthe laneand theball.,Weknow fromour earlieranalysisthatnone of these interactionswere statisticallysignificantfor thisexperiment,.,.,132,Bowling Example,(,(SessionWindow),Youcan alsosee thatthereiszeroerror,This is becauseonly1 runwasperformed withnoreplications,This is whereMinitabshowsustheMainEffects andInteractionEffects.,Note thatWristband hasthestrongest effectfollowedbytheinteractionbetweentheWristband andtheLane.,.,.,133,