4英文翻译模板参考样本.docx

4英文翻译模板参考样本.docx

《4英文翻译模板参考样本.docx》由会员分享，可在线阅读，更多相关《4英文翻译模板参考样本.docx（20页珍藏版）》请在冰点文库上搜索。

4英文翻译模板参考样本

毕业设计（英文翻译）

译文内容

UsingMCAtoSegmentNewCarMarkets（用多元类别分析进行新车市场细分）

译文出处

JournalofMarketingResearch

Aug70,Vol.7Issue3,p360-363,4p,2charts

系别：

经济管理学院

专业：

汽车营销

班级：

T553-1

学生姓名：

晏子淇

学号：

20050530124

指导教师：

魏仁干

UsingMCAtoSegmentNewCarMarkets

By:

Peters,WilliamH..JournalofMarketingResearch（JMR）,Aug70,Vol.7Issue3,p360-363,4p,2charts;（AN5002921）

PInthisstudythemid-196O'snewcarmarketisanalyzedbyusinganew,dummyvariable,multivariatetechniqueknownasMultipleClassificationAnalysis（MCA）.Thisapproachhelpsprovideabetterunderstandingofthiscomplexmarketandgivessomeinterestinghintsastoitspossiblefuturedirection.

INTRODUCTION

IntheyearssinceSmith'sarticle[3]arousedinterestinmarketsegmentation,therehasbeengrowingconcernaboutthepowerofdemographicvariablestosegmentmarketsformanyproducts.Infact,whetherdemographicanalysisisofmuchhelpatallindefiningmarketsorunderstandingbuyerbehaviorhasbeenquestioned.In1964Yankelovich[6,p.84]wrote:

"Inneitherautomobiles,soap,norcigarettesdodemographicanalysesrevealtothemanufacturerswhatproductstomakeorwhatproductstoselltowhatsegmentsofthemarket."

MorerecentlythepessimismabouttheusefulnessofdemographicdatainmarketsegmentationwascounteredbyBass,Tigert,andLonsdale[2].Theysaythatdemographicvariableshavehadsuchapoorperformancerecordinpaststudiesbecausethesestudiesattemptedtopredictindividualfamilybuyingbehaviorwithdemographics.Onthisbasis,becauseofwidevariationsofbuyerbehaviorinanymarketgroup,themultipleregressionR^valuesarealmostalwaysverylow.ButBassetal.writethat:

"...implementationofthestrategyofmarketsegmentationinvolvespostulatesaboutthecharacteristicsandbehaviorofgroups,notpersons"[2,p.265].Theyshowhowimportantbuyergroupscanbeidentifiedwithdemographicsandtherebyrefutetheideathatitisnotfeasibletodefinemarketsegmentsbysocioeconomicmeasurements.Theessenceoftheirargumentisthat"Formarketsegmentation,theessentialquestioniswhetheritispossibletoidentifygroupsofconsumerswithdifferentmeanpurchaseratesdependentoncertainvariables,suchasincome,age,andoccupation"[2,p.265].

Thisauthorfeelsthatanew,easytouse,dummyvariablemultipleregressionprogramcalledMultipleClassificationAnalysis（MCA）isideallysuitedtoperformthistaskefficiently.TopresentandexplaintheMCApro-gramisthefirstobjectiveofthisarticle.ThesecondistoputMCAtoworkanalyzingthemid-1960'snewcarmarketandtoshowtheresults.Theseresultsallowforamuchbetterunderstandingofthismajor,complexconsumermarket.Also,theyprovidesomeusefulinsightsintothepossiblefuturedirectionofthenewcarmarket—inparticularahintatapossiblereasonforthecurrentpopularityofsmallercarsandwhatthismaymeanforthefutureofsmallandmedium-sizedcars.

THESTUDY

DataUsed

Thedatainthisstudywerecollectedbypersonalinterviewforthe1966and1967SurveyofConsumer

Finances（SCF）bytheSurveyResearchCenteratthe

UniversityofMichigan.Theyarecombinedhere.Thesubsampleinthisstudyusedtheautomobileownedastheunitofanalysis,withappropriateeconomicanddemographicdataaboutthefamilyunitassociatedwitheachcar.Thissubsamplewasallcarspurchasednewwhosemodelyearwas1963-67（N=2,051）.

Variables

Thedependentvariableusedthroughoutthestudywastheautomobileclasstowhichthecarbelongs,withfivenewcarclassesestablishedgenerallyonthebasisofsize:

compacts,intermediate-sized,medium-sized,large,andforeigneconomycars.Theexplanatoryvariableincluded:

totalfamilyunitincome（allearnedandunearnedincomeexceptcapitalgains）,ageofthefamilyhead,head'soccupation,family'shousingstatus（rent/own）,numberofmajorincomereceivers（over$600ayear）,numberofchildrenunder18,numberofcarsowned,numberofadults（over18）,educationandraceofthehead,numberofdrivers,whoJournalofMarketingResearch,Vol.VII（August1970）,360-3USINGMCATOSEGMENTNEWCARMARKETSnormallydrives,distancefromcenterofcentralcity,andhousetype.Totalfamilyunitincomewasexpressedintermsofdeciles—thelowestnumberedoneandthehighestten.

Methodology

ThecurrentversionofMCAwasdevelopedinthe

early196O'sattheInstituteforSocialResearchatthe

UniversityofMichigan.Itisacomplexcomputerprogramthatmakesdummyvariablemultipleregressioneasy.WhileMCAhasbeenusedasananalyticaltoolinotherbranchesofsocialscienceresearch（especiallyattheInstituteforSocialResearch）,ithasyettobecomewellknowninmarketing.Ithaspotentialadvantagesinexploringmultivariatemarketingproblems,andespeciallyforgroup-typemarketingsegmentation,assuggestedbyBass　et　al.

ThemainadvantagesofMCAare:

1.Itmakesdeckpreparationwhendoingdummyvariableregressionsverysimple.。

takesallthedatacodedintomutuallyexclusivecategories（bothnominalandintervalscalepredictors）andautomaticallyconvertseachsubcategoryofeachpredictorvariableintoa1-0dummyvariable.（Itonlyusesonecomputercardcolumnforeachpredictorvariable,regardlessofthenumberofsubcategories.Withconventionaldummyvariableregressions,theresearcherhastowriteaspecialcomputerprogramtohelpdothistime-consumingjob,andeventhenonecomputercardcolumnisusedforeachsubcategoryofeachpredictorvariable.

2.MCAhandlesnonlineardata.

3.Theoutputisveryeasytounderstand（abrief

examplefollows）.MCAgivestheeffectofeach

explanatoryvariableuponthedependentvariable

intwoways—withandwithouttakingintoac-

counttheotherexplanatoryvariables.

ThechiefdisadvantageofMCAisthatitcannot

efficientlyhandleinteractioneffects.Itassumestheabsenceofinteraction.。

WhiletheMCAprogramwasdesignedtooperateonanintervalscaledependentvariable,itcanhandleaone-zeronominalscaledependentvariable,which,asinthefollowingexample,isusefulforstudyingproductownershipamongmarketsegments.

Table1showspartofatypicaltableconstructedfromMCAoutputduringthisstudy.Inthisexample,thedependentvariablewasaone-zerodichotomy:

oneifthecarownedwasacompactandzeroifitwasnot.（InthestudyaseparateMCAregressionwasrunforeachcarclass.）

ThemeaningoftheimportantstatisticsinTable1isasfollows:

1.Grandmean—Theoutputfortheregressionwillshowagrandmeanforalltheinputdata.Inthisexamplethegrandmeanis.1424,the14.24percentoftheentiresamplewhoownacompactcar.

2.Differencefromgrandmean,gross总体平均数离差值。

—

Thisshowstheunadjustedeffectofbeinginaspecific

categoryofanexplanatoryvariable.Thisgrossdifferenceiscomputedwithoutaccountingfortheeffectsoftheotherexplanatoryvariables.

3.Differencefromgrandmean,net—Thisdifferencegivestheeffectofbeinginagivensubgroupofthepopulationaftertheeffectsoftheotherexplanatoryvariableshavebeentakenintoaccount.。

Thenetdifferenceisthedeviationofthesubgroupmeanfromthegrandmeanthatwouldoccurwhenonlytheeffectofmembershipinthesubgroupaffectstheresult.Thisgetsatthetrueeffectofbeinginthatsubgroupofthepopulation.Inthereportoftheresultsthatfollows,onlyadjustedfrequenciesbasedonnetdifferencesthatwereatleastsignificantatthe.05levelarereported（significantadjustedfrequencies）.

4.β＾２Statistic

Thismeasurestheexplanatoryvariable's（theentirevariablewithallitssubgroups）abilitytoexplainvariationinthedependentvariableafteradjustingfortheeffectsofallotherexplanatoryvariables（notintermsofpercentofvarianceexplained）.Thosevariableswithhigher/8'levelsarebetterpredictors.

RESULTSANDDISCUSSION

BeforetheMCAresultsarepresenteditshouldbe

notedthatmarketsegmentationviathetraditional　JOURNALOFMARKETINGRESEARCH,AUGUST

1970methodofcross-tabulatingtheexplanatoryvariablesoneatatimeagainstthedependentvariable（carclass）wasattemptedwithdata.Withoutshowingthedetailsofthisprocedure,thisanalysisindicatedthatseparatemarketsegmentsforthefiveclassesofnewcarscannotbeclearlydifferentiatedusingsuchexplanatoryvariablesasincome,ageofhead,occupation,numberofchildren,etc.Theonlynewcarmarketsegmentthatdidlenditselftofairlycleardifferentiationfromthewholewasforlargecars.Themassmarket,however,couldnotbeseparatedsothatclearmarketsegments,intermsofthesedemographicvariables,weredistinguishableforcompactsvs.intermediatesvs.medium-sizedcarsvs.theforeigneconomyclass.。

Fromthepointofviewofthistypeofanalysis,thenewcarmarketseemstobeaveryuniformone.。

ThenextstepintheanalysiswastodotheMCA

regressions—aseparateregressionforeachcarclasswiththevalueofthedependentvariablebeingoneifthecarownedwasintheclassbeingstudiedandzeroifitwerenot.Thefirstthingonedoeswithsuchresults,ofcourse,istolookattheR-valuestoseethejointeffectofthevariables.Asinthepreviouslypublishedworkwheredemographicswereusedaspredictorsofbuyerbehavior,theR"^valueswereallverylow.Noneofthemwentabove.05forthenewcardata

Butnowtothemoreinterestingresultsfromthe

MCAanalyses—thetrueeffectofeachexplanatoryvariableuponnewcarownershipasexpressedbythenetdifferencesofthesubgroupmeansfromthetotalownershipmeanoftheentirepopulation.Theeffectonthegrandmeanprobabilityasexpresse