应用spss进行对数转换教学课件_第1页
应用spss进行对数转换教学课件_第2页
应用spss进行对数转换教学课件_第3页
应用spss进行对数转换教学课件_第4页
应用spss进行对数转换教学课件_第5页
已阅读5页,还剩31页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

ComputingTransformationsTransformingvariablesTransformationsfornormalityTransformationsforlinearity1Transformations:

TransformingvariablestosatisfyassumptionsWhenametricvariablefailstosatisfytheassumptionofnormality,homogeneityofvariance,orlinearity,wemaybeabletocorrectthedeficiencybyusingatransformation.Wewillconsiderthreetransformationsfornormality,homogeneityofvariance,andlinearity:thelogarithmictransformationthesquareroottransformation,andtheinversetransformationplusafourththatmaybeusefulforproblemsoflinearity:thesquaretransformation2TransformationschangethemeasurementscaleInthediagramtotheright,thevaluesof5through20areplottedonthedifferentscalesusedinthetransformations.Thesescaleswouldbeusedinplottingthehorizontalaxisofthehistogramdepictingthedistribution.Whencomparingvaluesmeasuredonthedecimalscaletowhichweareaccustomed,weseethateachtransformationchangesthedistancebetweenthebenchmarkmeasurements.Allofthetransformationsincreasethedistancebetweensmallvaluesanddecreasethedistancebetweenlargevalues.Thishastheeffectofmovingthepositivelyskewedvaluestotheleft,reducingtheeffectoftheskewingandproducingadistributionthatmorecloselyresemblesanormaldistribution.3Transformations:

ComputingtransformationsinSPSSInSPSS,transformationsareobtainedbycomputinganewvariable.SPSSfunctionsareavailableforthelogarithmic(LG10)andsquareroot(SQRT)transformations.Theinversetransformationusesaformulawhichdividesonebytheoriginalvalueforeachcase.Foreachofthesecalculations,theremaybedatavalueswhicharenotmathematicallypermissible.Forexample,thelogofzeroisnotdefinedmathematically,divisionbyzeroisnotpermitted,andthesquarerootofanegativenumberresultsinan“imaginary”value.Wewillusuallyadjustthevaluespassedtothefunctiontomakecertainthattheseillegaloperationsdonotoccur.4Transformations:

TwoformsforcomputingtransformationsTherearetwoformsforeachofthetransformationstoinducenormality,dependingonwhetherthedistributionisskewednegativelytotheleftorskewedpositivelytotheright.BothformsusethesameSPSSfunctionsandformulatocalculatethetransformations.Thetwoformsdifferinthevalueorargumentpassedtothefunctionsandformula.Theargumenttothefunctionsisanadjustmenttotheoriginalvalueofthevariabletomakecertainthatallofthecalculationsaremathematicallycorrect.5Transformations:

FunctionsandformulasfortransformationsSymbolically,ifweletxstandfortheargumentpassestothefunctionorformula,thecalculationsforthetransformationsare:Logarithmictransformation:computelog=LG10(x)Squareroottransformation:computesqrt=SQRT(x)Inversetransformation:computeinv=-1/(x)Squaretransformation:computes2=x*xForalltransformations,theargumentmustbegreaterthanzerotoguaranteethatthecalculationsaremathematicallylegitimate.6Transformations:

TransformationofpositivelyskewedvariablesForpositivelyskewedvariables,theargumentisanadjustmenttotheoriginalvaluebasedontheminimumvalueforthevariable.Iftheminimumvalueforavariableiszero,theadjustmentrequiresthatweaddonetoeachvalue,e.g.x+1.Iftheminimumvalueforavariableisanegativenumber(e.g.,–6),theadjustmentrequiresthatweaddtheabsolutevalueoftheminimumvalue(e.g.6)plusone(e.g.x+6+1,whichequalsx+7).7Transformations:

ExampleofpositivelyskewedvariableSupposeourdatasetcontainsthenumberofbooksread(books)for5subjects:1,3,0,5,and2,andthedistributionispositivelyskewed.Theminimumvalueforthevariablebooksis0.Theadjustmentforeachcaseisbooks+1.Thetransformationswouldbecalculatedasfollows:ComputelogBooks=LG10(books+1)ComputesqrBooks=SQRT(books+1)ComputeinvBooks=-1/(books+1)8Transformations:

TransformationofnegativelyskewedvariablesIfthedistributionofavariableisnegativelyskewed,theadjustmentofthevaluesreverses,orreflects,thedistributionsothatitbecomespositivelyskewed.Thetransformationsarethencomputedonthevaluesinthepositivelyskeweddistribution.Reflectioniscomputedbysubtractingallofthevaluesforavariablefromoneplustheabsolutevalueofmaximumvalueforthevariable.Thisresultsinapositivelyskeweddistributionwithallvalueslargerthanzero.Whenananalysisusesatransformationinvolvingreflection,wemustrememberthatthiswillreversethedirectionofalloftherelationshipsinwhichthevariableisinvolved.Ourinterpretationofrelationshipsmustbeadjustedaccordingly.9Transformations:

ExampleofnegativelyskewedvariableSupposeourdatasetcontainsthenumberofbooksread(books)for5subjects:1,3,0,5,and2,andthedistributionisnegativelyskewed.Themaximumvalueforthevariablebooksis5.Theadjustmentforeachcaseis6-books.Thetransformationswouldbecalculatedasfollows:ComputelogBooks=LG10(6-books)ComputesqrBooks=SQRT(6-books)ComputeinvBooks=-1/(6-books)10Transformations:

TheSquareTransformationforLinearityThesquaretransformationiscomputedbymultiplyingthevalueforthevariablebyitself.Itdoesnotmatterwhetherthedistributionispositivelyornegativelyskewed.Itdoesmatterifthevariablehasnegativevalues,sincewewouldnotbeabletodistinguishtheirsquaresfromthesquareofacomparablepositivevalue(e.g.thesquareof-4isequaltothesquareof+4).Ifthevariablehasnegativevalues,weaddtheabsolutevalueoftheminimumvaluetoeachscorebeforesquaringit.11Transformations:

ExampleofthesquaretransformationSupposeourdatasetcontainschangescores(chg)for5subjectsthatindicatethedifferencebetweentestscoresattheendofasemesterandtestscoresatmid-term:-10,0,10,20,and30.Theminimumscoreis-10.Theabsolutevalueoftheminimumscoreis10.Thetransformationwouldbecalculatedasfollows:ComputesquarChg=(chg+10)*(chg+10)12Transformations:

TransformationsfornormalityBoththehistogramandthenormalityplotforTotalTimeSpentontheInternet(netime)indicatethatthevariableisnotnormallydistributed.13Transformations:

DeterminewhetherreflectionisrequiredSkewness,inthetableofDescriptiveStatistics,indicateswhetherornotreflection(reversingthevalues)isrequiredinthetransformation.IfSkewnessispositive,asitisinthisproblem,reflectionisnotrequired.IfSkewnessisnegative,reflectionisrequired.14Transformations:

ComputetheadjustmenttotheargumentInthisproblem,theminimumvalueis0,so1willbeaddedtoeachvalueintheformula,i.e.theargumenttotheSPSSfunctionsandformulafortheinversewillbe:netime+1.15Transformations:

ComputingthelogarithmictransformationTocomputethetransformation,selecttheCompute…commandfromtheTransformmenu.16Transformations:

SpecifyingthetransformvariablenameandfunctionFirst,intheTargetVariabletextbox,typeanameforthelogtransformationvariable,e.g.“lgnetime“.Second,scrolldownthelistoffunctionstofindLG10,whichcalculateslogarithmicvaluesuseabaseof10.(Thelogarithmicvaluesarethepowertowhich10israisedtoproducetheoriginalnumber.)Third,clickontheuparrowbuttontomovethehighlightedfunctiontotheNumericExpressiontextbox.17Transformations:

AddingthevariablenametothefunctionFirst,scrolldownthelistofvariablestolocatethevariablewewanttotransform.Clickonitsnamesothatitishighlighted.Second,clickontherightarrowbutton.SPSSwillreplacethehighlightedtextinthefunction(?)withthenameofthevariable.18Transformations:

AddingtheconstanttothefunctionFollowingtherulesstatedfordeterminingtheconstantthatneedstobeincludedinthefunctioneithertopreventmathematicalerrors,ortodoreflection,weincludetheconstantinthefunctionargument.Inthiscase,weadd1tothenetimevariable.ClickontheOKbuttontocompletethecomputerequest.19Transformations:

ThetransformedvariableThetransformedvariablewhichwerequestedSPSScomputeisshowninthedataeditorinacolumntotherightoftheothervariablesinthedataset.20Transformations:

ComputingthesquareroottransformationTocomputethetransformation,selecttheCompute…commandfromtheTransformmenu.21Transformations:

SpecifyingthetransformvariablenameandfunctionFirst,intheTargetVariabletextbox,typeanameforthesquareroottransformationvariable,e.g.“sqnetime“.Second,scrolldownthelistoffunctionstofindSQRT,whichcalculatesthesquarerootofavariable.Third,clickontheuparrowbuttontomovethehighlightedfunctiontotheNumericExpressiontextbox.22Transformations:

AddingthevariablenametothefunctionSecond,clickontherightarrowbutton.SPSSwillreplacethehighlightedtextinthefunction(?)withthenameofthevariable.First,scrolldownthelistofvariablestolocatethevariablewewanttotransform.Clickonitsnamesothatitishighlighted.23Transformations:

AddingtheconstanttothefunctionFollowingtherulesstatedfordeterminingtheconstantthatneedstobeincludedinthefunctioneithertopreventmathematicalerrors,ortodoreflection,weincludetheconstantinthefunctionargument.Inthiscase,weadd1tothenetimevariable.ClickontheOKbuttontocompletethecomputerequest.24Transformations:

ThetransformedvariableThetransformedvariablewhichwerequestedSPSScomputeisshowninthedataeditorinacolumntotherightoftheothervariablesinthedataset.25Transformations:

ComputingtheinversetransformationTocomputethetransformation,selecttheCompute…commandfromtheTransformmenu.26Transformations:

SpecifyingthetransformvariablenameandformulaFirst,intheTargetVariabletextbox,typeanamefortheinversetransformationvariable,e.g.“innetime“.Second,thereisnotafunctionforcomputingtheinverse,sowetypetheformuladirectlyintotheNumericExpressiontextbox.Third,clickontheOKbuttontocompletethecomputerequest.27Transformations:

ThetransformedvariableThetransformedvariablewhichwerequestedSPSScomputeisshowninthedataeditorinacolumntotherightoftheothervariablesinthedataset.28Transformations:

AdjustmenttotheargumentforthesquaretransformationInthisproblem,theminimumvalueis0,noadjustmentisneededforcomputingthesquare.Iftheminimumwasanumberlessthanzero,wewouldaddtheabsolutevalueoftheminimum(droppingthesign)asanadjustmenttothevariable.Itismathematicallycorrecttosquareavalueofzero,sotheadjustmenttotheargumentforthesquaretransformationisdifferent.Whatweneedtoavoidarenegativenumbers,sincethesquareofanegativenumberproducesthesamevalueasthesquareofapositivenumber.29Transformations:

ComputingthesquaretransformationTocomputethetransformation,selecttheCompute…commandfromtheTransformmenu.30Transformations:

SpecifyingthetransformvariablenameandformulaFirst,intheTargetVariabletextbox,typeanamefortheinversetransformationvariable,e.g.“s2net

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

最新文档

评论

0/150

提交评论