Development and Validation of Technology Acceptance Modelling. for Evaluating User Acceptance of an E learning framework YALDA DANESH SEDIGH

Transcription

1 DevelopmentandValidationofTechnologyAcceptanceModelling forevaluatinguseracceptanceofanelearningframework By YALDADANESHSEDIGH AthesissubmittedtoTheUniversityofBirminghamforthedegreeof MASTEROFPHILOSOPHY TheSchoolofElectrical,ElectronicandComputerEngineering TheUniversityofBirmingham 11 th June2013

2 University of Birmingham Research Archive e-theses repository This unpublished thesis/dissertation is copyright of the author and/or third parties. The intellectual property rights of the author or third parties in respect of this work are as defined by The Copyright Designs and Patents Act 1988 or as modified by any successor legislation. Any use made of information contained in this thesis/dissertation must be in accordance with that legislation and must be properly acknowledged. Further distribution or reproduction in any format is prohibited without the permission of the copyright holder.

3 ACKNOWLEDGEMENTS ThisthesishasbeenagreatachievementformeandIwouldliketoacknowledgethesupport ofmysupervisor,drtheodorosn.arvanitis,throughoutmystudyandsupportgivenbymy cosupervisordralexgibbintheinitialyearofmystudy.iamgratefulforthefinancialand emotionalsupportfrommyfamily.iwouldalsoliketothankmycolleaguesinthehuman ComputerInteractionResearchCentre.AquestionnairesurveyedbyOudeRengerink s(2011) aidedthisthesis.thethesisresearchworkwascarriedoutattheuniversityofbirmingham;i enjoyedthefacilitiesprovidedbytheuniversity,suchasthemainlibraryandaccesstoonline databasesofjournalpapers. i

4 ABSTRACT ThethesisaimedtodevelopandvalidateanewtheoreticalmodeltoassessEvidenceBased Medicine(EBM)trainers technologyacceptanceofanelearningapplicationassociatedtothe EUEBMTeachtheTrainersecourse.Modellinguserinteractionswithelearningapplications allowsthepredictionofhowebmtrainersaremotivatedtousetheecourseintheclinical setting.aspartofthisresearch,asurveywasconstructedandanalysedusinganempirical modelcalledthetechnologyacceptancemodel(tam).thetamwasdevelopedintothee TAM,anewmodel,whichcanassessauser sadoptionofonlinelearningapplications.this thesisusedasurveytodeveloptheetamasanextensionfortechnologyacceptanceofon linepublicationssuchastheblendedlearningapproachforebmstudy.thethesisvalidated the TAM and etam, which followed an assessment of EBM trainers acceptance of the application.statisticalanalysis,includingreliabilitywithcronbach salpha,factoranalysisand multipleregression analysis, were carried out on TAM, etam and data from the questionnairesthatshowedthemodelswerevalidforthisfieldofstudy.thisassessment foundtheebmtrainers experience,perceivedusefulnessandattitudetowarduseasstrong predictorsofuserbehaviouralintentionandacceptanceoftheapplication.overall,themost influentialfactorintheetammodelwasexperience,whichexplainedoverafifthofthetotal varianceoftheuseracceptanceoftheelearningapplication. ii

5 iii TABLEOFCONTENTS ACKNOWLEDGEMENTSI ABSTRACTII TABLEOFCONTENTSIII LISTOFFIGURESV LISTOFTABLESVII 1 INTRODUCTION1 1.1 EVIDENCEBASEDMEDICINE2 1.2 MOTIVATION5 1.3 THEORETICALMODELLINGOFUSERACCEPTANCE6 1.4 ANALYSISOFUSERACCEPTANCEINFORMATION8 1.5 AIMANDOBJECTIVES9 1.6 CONTRIBUTIONS THESISOUTLINE11 2 BACKGROUNDANDLITERATUREREVIEW INTRODUCTION THEUSEOFELEARNINGFOREBMTRAINERSANDITSACCEPTANCE THEORETICALMODELLINGOFUSERACCEPTANCE:FROMTHEORYANDMODELLINGTOPRACTICEANDANALYSIS EXTERNALFACTORSFORTHETAMMODEL SUMMARY69 3 METHOD UNDERSTANDINGTHECORECONCEPTSOFTHEPROJECT TECHNOLOGYACCEPTANCEMODELLING HYPOTHESES PROPERTIESOFTHEQUESTIONNAIRE TESTINGRELIABILITYOFFACTORSWITHINTHEQUESTIONNAIRE TESTINGTHERELIABILITYOFTHEQUESTIONS FACTORANALYSIS REGRESSIONANALYSIS REGRESSIONANDHYPOTHESESTESTING SUMMARY102

6 iv 4 TAMRESULTS INTRODUCTION RELIABILITY FACTORANALYSIS REGRESSION DISCUSSION SUMMARY114 5 EXTENDEDTAMRESULTS INTRODUCTION RELIABILITY FACTORANALYSIS REGRESSION DISCUSSION SUMMARY132 6 CONCLUSIONANDFURTHERWORK PROPOSEDDEVELOPMENTSFORTTTEBMECOURSE OPPORTUNITIESPRESENTEDFORNEWRESEARCH CONSIDEREDFURTHERWORKFORTHISSTUDY135 7 REFERENCES137 8 APPENDIX DAVIS EXPLANATIONOFTHEFISHBEINMODEL THETECHNOLOGYACCEPTANCEQUESTIONNAIRE IMPLEMENTATIONBARRIERSTOTEACHINGEVIDENCEBASEDMEDICINEINCLINICALPRACTICEQUESTIONNAIRE ETAMQUESTIONNAIREFACTORISATION GLOSSARY TABLEOFACRONYMS161

7 LISTOFFIGURES FIGURE1:TRADITIONALCLINICALPRACTICEANDIMPROVEMENTCYCLE,REPRODUCEDFROMTHANGARATINAMETAL.(2009)...19 FIGURE2:IMPROVEDLEARNINGANDUSEOFEBMINCLINICALENVIRONMENTWITHADDITIONALLEARNINGRESOURCES,REPRODUCED FROMTHANGARATINAM(2009)...20 FIGURE3:TTTEBMELEARNINGCOURSEACCESSIBLEVIATHEINTERNET:MODULE1 WARDROUND...23 FIGURE4:THESIXCLINICALSETTINGSTHATARETAUGHTINTHETTTEBMELEARNINGCURRICULUM,REPRODUCEDFROM(ZANREI, 2009)...24 FIGURE5:THEINNOVATIONDECISIONPROCESSCOMMUNICATIONCHANNELS,REPRODUCEDFROMROGERS(1995)...28 FIGURE6:ORIGINALVERSIONOFTHEORYOFREASONEDACTION(TRA),REPRODUCEDFROMAJZEN&MADDEN(1986)...31 FIGURE7:THEORYOFREASONEDACTION(TRA),REPRODUCEDFROMAJZEN&FISHBEIN(1980)...32 FIGURE8:SOCIALCOGNITIVETHEORY,REPRODUCEDFROMBANDURA(1986)...34 FIGURE9:THEORYOFPLANNEDBEHAVIOUR,REPRODUCEDFROMAJZENANDCOTE(2008)...36 FIGURE10:THETAMMODELBASEDON(DAVIS,1989)...38 FIGURE11:TAMWITHNUMBEREDLINKSBETWEENTHEFACTORSANDEXTERNALFACTORS,REPRODUCEDFROMDAVIS(1986)...39 FIGURE12:TAM2EXTENSIONOFTHETAM,REPRODUCEDFROMVENKATESHANDDAVIS(2000B)...42 FIGURE13:UNIFIEDTHEORYOFACCEPTANCEANDUSEOFTECHNOLOGY,REPRODUCEDFROMVENKATESHETAL.(2003)...44 FIGURE14:TECHNOLOGYACCEPTANCEMODEL3(TAM3),REPRODUCEDFROMVENKATESHANDBALA(2008)...48 FIGURE15:RESEARCHMODELANDRESULTSOFTHEHYPOTHESISTEST,REPRODUCEDFROMONGETAL.(2004)...50 FIGURE16:THETECHNOLOGYACCEPTANCEMODEL,CONSIDEREDASABASEFORTHENEWMODELBYPITUCHANDLEE,REPRODUCED FROMPITUCHANDLEE(2006)...52 FIGURE17:MODELASHOWS FULLYMEDIATEDMODEL FORELEARNINGUSEDBYPITUCHANDLEE(2006),BASEDONTHETAMWITH EXTERNALVARIABLES...52 FIGURE18:MODELBSHOWSTHE PARTIALLYMEDIATEDMODEL FORELEARNINGUSE,REPRODUCEDFROMPITUCHANDLEE(2006)...54 FIGURE19:THEMODELFORTHEACCEPTANCEOFWEBBASEDLEARNINGSYSTEMS,REPRODUCEDFROMNGAIETAL.(2007)...55 FIGURE20:THERESULTOFREGRESSIONANALYSIS,REPRODUCEDFROMMASROM(2007)...56 FIGURE21:FLOWCHARTOFTHEMETHODOLOGYSHOWINGTHEPROCESS...73 v

8 FIGURE23:SHOWSTHEEXTENDEDTAM...77 FIGURE24:SHOWSGROUPINGFOREXTERNALFACTORS...78 FIGURE25:SHOWINGEXTERNALFACTORSGROUPINGANDTHEMETHODOFCONNECTIONTOTHETAM...79 FIGURE26:NEWILLUSTRATIONOFTHEGENERALISATIONOFEXTERNALFACTORS,WHEREAGEANDGENDERHAVEBEENOMITTED,AND QUALITYISCONSIDEREDASPARTOFADESIGNDIMENSION...81 FIGURE27:THETECHNOLOGYACCEPTANCEMODEL(TAM),SHOWINGTHEHYPOTHESESASSUMEDFORTHISRESEARCH...83 FIGURE 28: EXTENDED TECHNOLOGY ACCEPTANCE MODEL (ETAM), SHOWING THE HYPOTHESES ASSUMED FOR THIS PART OF RESEARCH...86 FIGURE29:EXAMPLEOFTHELIKERTSCALEFORMATUSEDINTHEQUESTIONNAIRES...90 FIGURE30:THEEXAMPLESOFTHEFITLINE(SAPSFORD,ETAL.,2006) FIGURE31:ANEXAMPLEOFTHECORRELATIONCOEFFICIENT(RVALUE)(SAPSFORD,ETAL.,2006) FIGURE32:CRONBACH'SALPHAIFITEMDELETED(CAIID)WITHREDHORIZONTALLINESHOWINGTHESTANDARDISEDCRONBACH S ALPHA FIGURE33:CORRECTEDITEMTOTALCORRELATION(CITC)WITHYAXISSETTOMINIMUMRECOMMENDEDVALUE FIGURE34:FACTORLOADINGFOREACHFACTOR1TO4FROMTOPLEFTTOBOTTOMRIGHT FIGURE35:THETECHNOLOGYACCEPTANCEMODELSHOWSTHEWEIGHTOFEACHFACTORONEACHOTHERWITHACOEFFICIENT FIGURE36:CORRECTEDITEMTOTALCORRELATION WITH RED HORIZONTAL LINE SHOWING THE MINIMUM RECOMMENDED VALUES FIGURE37:CRONBACH SALPHAIFITEMDELETEDWITHREDHORIZONTALLINESHOWINGTHESTANDARDISEDCRONBACH SALPHA120 FIGURE38:PERCENTAGEOFVARIANCEEXPLAINED FIGURE39:THEETAMSHOWSTHEWEIGHTOFEACHFACTORONTHEOTHERONEWITHCOEFFICIENT FIGURE40:SHOWINGTHELIMITATIONSWITHFACTORSANDHYPOTHESESINREDDASHEDLINEBOXESANDWITHDASHEDLINES FIGURE42:THEDEMONSTRATIONOFTHERELATIONSANDINFLUENCEOFEXTERNALFACTORSONPERCEIVEDUSEFULNESS FIGURE43:THEDEMONSTRATIONOFTHERELATIONSANDINFLUENCEOFEXTERNALFACTORSONPERCEIVEDEASEOFUSE vi

9 LISTOFTABLES TABLE1:ALISTOFTHEHYPOTHESESTHATDISTINGUISHBETWEENDEPENDENTANDINDEPENDENTVARIABLES...83 TABLE2:LISTOFTHEHYPOTHESESTHATDISTINGUISHBETWEENDEPENDENTANDINDEPENDENTVARIABLES...87 TABLE3:SHOWINGEACHFACTORANDRELATEDCOMPONENTFOREACHFACTORSEPARATELY...89 TABLE4:SHOWINGEACHFACTORANDRELATEDCOMPONENTFOREACHFACTORSEPARATELY...92 TABLE5:AVERAGERELIABILITYANALYSIS TABLE6:RELIABILITYANALYSISINDETAILFOREACHQUESTION TABLE7:ROTATEDFACTORMATRIXOFTAMROTATEDWITHVARIMAXWITHKAISERNORMALIZATIONANDLESSTHAN0.5CUTOFFTO DRAWOUTFACTORS(DARKGREYSHADE),0.3CUTOFF(LIGHTGREYSHADE) TABLE8:MULTIPLEREGRESSIONRESULTOFTECHNOLOGYACCEPTANCEMODEL TABLE9:AVERAGERELIABILITYANALYSIS TABLE10:RELIABILITYANALYSISINDETAILFOREACHQUESTION TABLE11:FACTORLOADINGWITHDRAWOUTVARIABLES(DARKGREYSHADE) TABLE12:REGRESSIONRESULTOFETAM TABLE13:THELISTOFTHEQUESTIONSFROMTTTEBMELEARNINGCURRICULUM SQUESTIONNAIREFROMOUDERENGERINK(2011) THATHAVEBEENDROWNEDOUTFORTHEEXPERIENCEPREDICTORINTHEEXTERNALFACTORS vii

10 1 INTRODUCTION Medicalpractitionershavetoadaptquicklytolearn,understandandadoptnewtechnology asgeneraltrendsshowacontinuousintegrationoftechnologyandnewmedicalstrategies intotheworkplace.thereisaneedforthemtolearnhowtouseelearningtechnologiesand associatedinnovativesystemsonaroutinebasis,becausetheintegrationanduseofthese systemsintheworkplacewillimprovepatients healthcare(ouderengerink,etal.,2011).this meanstrainersofnewmedicalknowledgeandevidenceneedtolearnhowtoteacheffectively theuseofavailableinformationthroughelearningsystems,suchastheelearningframework for teachingthetrainersevidencebasedmedicine (Thangaratinam,etal.,2009).Resulting from better training, medical practitioners can manage to use innovative approaches to handleinformationandknowledge,suchasevidencebasedmedicine,whileprovidingbetter healthcarepracticeonadailybasis.overall,thisdevelopmentintheclinicalenvironmentwill benefitpatientswithbetterdiagnoses,treatment,andfollowupprocesses. Ehealth is a new approach to healthcare practice and is a framework facilitated by information systems such as libraries, the Internet, online clinical databases, research databases,decisionsupportinformationsystems,telephonesupportcentres,etc.(broderick &Smaltz,2003).Thesefacilitiesincreasetheaccessibilityandavailabilityofinformationfor healthcareservices.ehealthcanbeintegratedintootherframeworks,suchaselearning.e learning utilises Information and Communication Technology (ICT) systems to give users access to knowledge worldwide. Elearning lets users study with the aid of interactive tutorials,prerecordedlecturesandotherdownloadableorstreamingmedia(arbaugh,2004; 1

11 Singh,etal.,2005;Ngai,etal.,2007).Theinnovationofelearningwithinthecontextofe health in healthcare institutions can train healthcare practitioners how to use associated technologywithoutrestrictionsonthepractitioner stimeorplace. 1.1 EvidencebasedMedicine Evidencebased medicine (EBM) is a medical information framework and considered as a subsetofehealthsystems(eysenbach,2001). EBMcanbesupported byicttechnologies (Schaper&Pervan,2007)andupgradesthetraditionalmedicaltheoryandpracticebyfocusing moreonthespecificneedsofpatients(coppus,etal.,2007).withtheaidoftraining,medical professionals can access EBM information through online softwarebased technologies, as well as use and update medical evidence as an integral part of the routine (Kulier, et al., 2008b),asnewtheoriesandpracticesbecomeavailableinrealtime(Coppus,etal.,2007; Hatala,etal.,2006). Onlinetrainingsoftwareapplicationscansupportprofessionalstowardstheuseofevidence basedmedicalknowledge(kulier,etal.,2008a)andtherebysupporttheintegrationanduse ofebmintheclinicalenvironment;furtherdiscussiononthistopicinsection2.2.inaddition, access to EBM information from an online medical database and support with elearning trainingsessionshasthepotentialtoimprovetheaccessibility,availabilityandawarenessof medicalinformationformedicalprofessionalsandtrainersofebmintheclinicalenvironment (Coppus,etal.,2007;Hatala,etal.,2006).Itisbeneficialtoutilisetheadvantagesofelearning forebmstudy,specificallyintermsofitsabilityinprovidingathandinformationonaglobal scale. 2

12 Medicalandhealthcareevidence,availablefromonlineelearningenvironments,cansave teachingtimewhenmanypeoplewanttolearntheebmapproach,invariouslocationsand organisations.moreover,publishingonlinecantakemediaformsofebooks,andstreaming audioorvideo,whichallowsamorecosteffectivewayoflearningebmcomparedtohard multimediaformslikecds,dvdsorbooks(kulier,etal.,2008a).whenprofessionalssubmit and author EBM information from reliable sources, it provides official EBM support for practitionersandtrainers(hatala,etal.,2006). EBM gives practitioners a solid reference for the duration of the of patienthealthcare practitionerinteractionincludingdiagnosis,treatmentandfollowup(kulier,etal.,2008b). ThatreferenceallowsEBMtoprovideabasisforconvincingbothpatientandpractitionerto maketherightdecisionthrougheachstepinthecontinuityofhealthcare.however,ebm baseddecisionshavealimitedeffectasaconfidencebuilderifthepractitionersareunsure about the use of EBM due to trainers lacking confidence in demonstrating its potential (Thangaratinam,etal.,2009).TrainerscanteachEBMinvariousways,whichmightexplaina confusionoverwhichcurriculumtofollow(coppus,etal.,2007).ifthisisthecase,thenthere shouldbearecognisedandofficialcourseforteacherstouse(thangaratinam,etal.,2009). OnlinetrainingformedicalprofessionalshasgrownindemandinEurope,whiletheapproach of elearning needs approval as an official qualification (Oude Rengerink, et al., 2011). In addition, organisations have attempted to analyse how users accept such softwarebased systems(schaper&pervan,2007)andthereforedevelopelearningsystemstoincreaseits adoptioninhealthcare. 3

13 TherearemanybarriersassociatedwiththeuseracceptanceofanonlineEBMtrainingsystem (OudeRengerink,etal.,2011),butthesedonotoutweighthebenefitsofitswidespreaduse. Becauseofthis,thethesisconsidersfindinginfluentialbarriers/factorsimportantasitcanaid thedevelopmentofatechnologyacceptancemodelthatcanassesstheteachingthetrainers EBM(TTTEBM)elearningapproachandtherebyitsadoption. Thangaratinam (2009) pointed out that trainers of EBM should use the available clinical environmental facilities, teach EBM practice and demonstrate how to apply technology associatedwithebmineverydayclinicalactivitiesforthebenefitofthetrainees.inherstudy, thisteachingmethodissetintoacurriculumcalled TeachingtheTrainersEBM(TTTEBM), whichprovidesablendedlearningapproach,includinganelearningframework.thetttebm projectisalsoknownastheeuebmtttprojectinthangaratinam(2009),butthisthesisrefers toitastheformerthroughoutthethesis.thetttebmelearningcurriculumanditsblended learning approach is part of a Europeanwide project, which plays a role in providing standardised continuing professional development for EBM trainers across the European Union(Thangaratinam,etal.,2009).Thiscurriculumisaccessedasanecourseandisthefirst example of elearning training system, designed for building trainer confidence and integrating learning within the clinical environment. It includes learning and assessment facilitatedbyelearningtechnology;thereexistemodulesinvolvedwiththeuseofjournal clubs,morbidity/mortalitymeetingsandauditpresentations,whichtogetherisconsideredan ecourse. ResearcherscanusetheresultspresentedinthisthesistoimprovetheTTTEBMproject.As thisthesis outcomeshowed,thefactorsofexperience,educationlevel,technologysupport 4

14 andtimeconstraintaccountedforoverhalfofthevarianceintheuseracceptanceoftheebm elearning application. Meaning those factors are influential to the TTTEBM project s adoption.therefore,thatprojectcanimprovewithdevelopmentsfocusedontheelimination ofnonessentialbarriersandtheinclusionorreinforcementoftheinfluentialfactorsshown inthisthesis. 1.2 Motivation EBMtrainersneedtoteachpractitionerswithconfidence,whichaproposedEuropeanwide ecoursecangive,whereacceptanceoftheecourse sapplicationneedstobeassessedto ensureitswidespreaduse.trainersneedtobemotivatedandconvincedtobeinvolvedwith thatapplicationthatteachesthemhowtoteachmedicalpractitionerstheuseandbenefitsof EBM in the workplace (Oude Rengerink, et al., 2011). However, there is a shortage of appropriatemodelstopredicttheuseracceptanceofthatelearningapplication.therefore, thisthesisproposesanewmodelforuseracceptanceassessmentofthetttebmelearning curriculum,deliveredviatheecourse. TheneedforEBMtoachieveimprovedqualityhealthcarehasbeenacceptedandpublished inmanycountries,especiallyineuropeanjournalssuchasfromdawesetal.(2005),daviset al.(2007),andcoppusetal.(2007).ideally,theoutcomesofthetttebmprojectwillhelp EBMtrainersimproveclinicalstaff sknowledgeandskillsofebmpracticeineverydayclinical activities. Thus, moving healthcare and medical professionals to upgrade EBM knowledge independently,askulier(2008b)hadpointedout.hence,thecentralobjectiveofthisthesisis toevaluatetheuseracceptanceandassociatedhumanfactorconsiderationsforintegrating thetttebmblendedelearningapproachfortheworkplace. 5

15 Theworkofthisthesiscontributedtouseracceptanceoftechnologyresearchbyeliminating irrelevant barriers and finding influential factors of a user s acceptance of the TTTEBM project.thetttebmprojectimprovestheeuropeanhealthcaresectorthroughthedesign, development,promotionandpilotingofaeuropeantrainingprogramme(thangaratinam,et al.,2009;ouderengerink,etal.,2011).thattttebmprojectfocusesonteachinghealthcare trainers of EBM, through an intensive integrated elearning curriculum within a clinical practice.thedesignofthecurriculumandblendedelearningcourseshouldbebeneficialfor theflexibilityofthetrainers timeandplace.thisthesisfocusesonhowtoimprovethequality of that project by evaluating factors that influence acceptance, future attitudes and behaviouralintentionsofusersenrolledonthetttebmelearningcurriculum. Modellingandpredictinghowusersaccepttheelearningapplicationisafocusofthisthesis, whichitattemptstoassesswiththedevelopmentofatechnologyacceptancemodelsand associatedfactors,section1.3furtherdiscussesthispoint.themodelsdevelopedinthethesis evaluatesurveysofprofessionalcliniciansinthescopeofthetttebmproject,namelyfrom Europeanclinicians.Inthisthesis,oursurvey,althoughsmallinscale,takesapanEuropean perspectiveofevaluatingthetttebmelearningframeworkanddevelopsawidelyaccepted empiricalmodeltodetecttheinfluenceofbarriersassociatedtoauser sadoptionofthee course. 1.3 TheoreticalModellingofUserAcceptance TheTechnologyAcceptanceModel(TAM)isawellrecognisedempiricalmodelforitsability inpredictinghowandwhyitandcomputersystemsusersapproach,startandcontinueusing technology(davis,1986).itcandeterminetheprobabilityofacceptanceauserhastoanew 6

16 typeoftechnology(davis,1986).thetamhasevolvedintothetam2(venkatesh&davis, 2000b)andTAM3(Venkatesh&Bala,2008),withtheimplementationofadditionalfactors (formoredetail,seesection2.4).theseadditionalfactorswereusedtoproduceresultsthat arebettersuitedforspecificcontextssuchasauser sexperience(venkatesh&davis,2000b; Venkatesh&Bala,2008),jobrelevance(Venkatesh&Davis,2000b;Venkatesh&Bala,2008) andcomputerselfefficacy(venkatesh&davis,1996;venkatesh&davis,2000b)aswellasto furtherexplainthevarianceofuserbehaviouralintentionofthetam soriginalfactors. NumerousotherstudieshavedevelopedtheTAMwithadditionalfactorsofuseracceptance oftechnologyinhealthcareittoprovideaspecificresearchmodeltofitastudyarea(holden &Karsh,2010).ThefactorsthatthisthesisconsidersfordevelopingtheTAM,alsocalledsocial organisational barriers (Oude Rengerink, et al., 2011), include apparent challenges and limitationsoftheclinicalenvironmentformedicalstaffandtrainers.holden&karsh(2010) reviewedthelimitationsofthetamandfoundareasofhealthcareitrelatingtoelearning andinternetbasedsystemssuchaswebbasedelectronicmedicalrecords,mobilemedical informationsystems,onlinedisabilityevaluationsystems,telemedicinetechnology,ictand Internetbasedhealthapplications.HethenevaluatedthosefactorswiththeTAM. DevelopingtheTAM scontextualassessmenttosuitablendedlearningapproach,byadding factorsrelatedtoelearninginthehealthcarecontext,isnecessarybecauseitwouldmakeit avalidmodelforusersassociatedwiththisthesis,whereasdavisdevelopedthetamtoassess ITandcomputersystemusers(Holden&Karsh,2010).Eventhoughelearningapplications are a subset of IT and computer systems, EBM and the concept of TTTEBM can function independentofitandcomputersystems,whichmeansusingthetamtoassessusersofthe 7

17 ecourse would not necessarily fit within the original scope of the TAM. This clarifies the reasoningtodevelopthetamwithadditionalfactorstoachieveastrongerpredictionofthe useracceptanceoftechnologyfortheapplicationbecausenoothermodelshavetheunique abilityforassessingusersinthisspecificcontext.anoutputofthisthesisisthedevelopment ofanextendedtechnologyacceptancemodel(etam),whichincludesthechosenadditional externalfactors.section3.2.2hasmoreinformationonthedevelopmentanddefinitionofthe etam. 1.4 AnalysisofUserAcceptanceInformation Theoreticalmodelsrepresentatheoryandresearchersassumethatthemodelcanpredictan outcomeofthetheoryoridentifyinfluentialfactors,suchaspredictingusersacceptinge learningtechnologiesinthehealthcaresetting,byanalysingindividualdatafrommultiple choicequestionnaires.thetheoreticalmodel sanalysisshowstheinfluenceoffactorsrelated bypredeterminedhypotheses,whichfitaroundthemaintheoryofthesubject,whichinthis thesisistheacceptanceoftechnologyandapproachtodeliverebmtraining.hypothesesand questionnaires, similar to research models, need a design that is specific for the research purposestogetreliableresultsforthemodeltoanalyse(seesection3.4).thisthesishasused two questionnaires designed in mind of understanding the user acceptance of elearning technologiestostudyebminaclinicalsetting.ouderengerink(2011),whoconstructedone ofthequestionnairesfromaliteraturereview,distributed,surveyedandtestedit usingthe nonparametrickruskal WallistestortheWilcoxonRankSumtest.Thattestfoundbarriers andfacilitatorsofteachingebminclinicalenvironments. 8

18 Pretestingquestionnaires,withareliabilityanalysis,isnormallyaprerequisiteinthedesign process of questionnaires. It validates the results and assesses whether each question achievesthesamescorewhenputunderdifferentconditionsfromthesamecandidates,or from different candidates under the same conditions (Bailey & Pearson, 1983; Kripanont, 2007;Field,2009).Moreinformationontheuseofreliabilityanalysis,inthisthesis,iscovered insection3.5.thecronbach salphaapproach,asusedinthisthesis,seesection3.5,iswidely usedforreliabilityassessmentbecauseithasbeenrecognisedasoneofthemostsignificant tests of reliability in various fields of social science (Cortina, 1993). The reliability of the questionnaire directly relates to the quality of the results from the research models evaluation. This thesis has validated both questionnaires to support the TAM and etam models resultsandfindings. Usingtechnologyacceptancemodels,suchastheTAM,facilitatesthecollectionofdatafor examiningwhichfactorhasaninfluenceontheuser sadoptionofthetttebmelearning applicationanddrawsoutbarrierstowarditsacceptance.thisthesispredictsuserbehaviour, intention,attitude,useracceptanceandadoptionofanonlineapplicationfortrainerstoteach EBMinaclinicalenvironment.Thisincludestheselectionoffactorsassociatedtoelearning, aswellasanalysisofresultsfromtworesearchmodels(tamandetam).thisevaluationalso quantifies the effect of important barriers that improve the adoption of the TTTEBM e learningapplication. 1.5 AimandObjectives Theaimofthisthesisinvolvesthetheoreticaltestingoffactorsforthedevelopmentofthe TAMintoamodelthatassessesablendedlearningapproachtowardteachingEBMtrainersin 9

19 aclinicalsetting.thefactorscomefromanevaluationofpreviousresearchmodelsanda literature review based on elearning, EBM and TTTEBM. This study has established the followingobjectivestoreachthataim: todrawouttheinfluentialfactorsofthetamandetamthathavearoleontheuser acceptance, future attitudes and behavioural intentions of users of the blended learningapproachadoptedinthetttebmproject tofindandevaluatethemostinfluentialfactorsofthetamandetamforpredicting theuseracceptanceoftheblendedlearningapproach tocriticallyselectsuitableexternalfactorsforuseinthedevelopmentofthetamthat havebeenweighedasbarriersinelearning toevaluatethetam sabilityinassessingindividualdataontheuseracceptanceofa blendedlearningapproachtoebmstudy to evaluate the etam in assessing the user acceptability of the blended learning approach touseapretestedquestionnairetovalidatetamandassesstheuseracceptanceof thetttebmprojectanduseanotherquestionnairetovalidatetheetamtofindout themostimportantbarriertotttebm toprovideabasisforotherresearchersinterestedincontinuingthedevelopmentof researchmodelsandimprovementoftheblendedlearningapproach toprovideinsightintohowtheelearningapplicationcanbeimprovedforthebenefit ofmedicalpractitionersandthetrainer suseoftheapplication 10

20 1.6 Contributions This study designed a questionnaire for the TTTEBM elearning curriculum and blended learning approach, and used it to study and evaluate the TAM. A different questionnaire designedbyouderengerink(2011)wasusedtoevaluatetheeffectivenessoftheetamin assessingtheuseracceptanceofthetttebmelearningapplication.itsresultscontributeto research involved with the evaluation of barriers and factors for the user acceptance e learningtechnologiesassociatedtoehealthormorespecificallyebm,wheredesignershave blended the study material into the workplace. The thesis results are beneficial to developmentofthetttebmelearningapplicationbecauseithaspredictedtheecourse s useracceptance. Thereiscontributiontotheoreticalmodellingdesignanddevelopmentresearchbydeveloping thetamwithexternalfactorsandanestablishmentofanewmodel,theetam.thisthesis criticallyreviewedawiderangeofresearchthathadevaluatedtheeffectofaddingexternal factorstothetamanditsuseinassessingelearningapplications. 1.7 ThesisOutline SecondChapter:BackgroundandLiteratureReview Sufficient literature and background knowledge is reviewed to understand the merits of similarresearchtothisthesisandtofindpotentialgapsortheoriesthatareunexploredinthe context of developing technology acceptance models, elearning, EBM and TTTEBM. The literature review in Section 2.4 covers potential external factors that other research has evaluated the user acceptance of technology in different models, where some areas of researchrelatedtowebbasedlearningorelearning. 11

21 1.7.2 ThirdChapter:Methods TheinformationfromChapter2exploresthecoreconceptsofdevelopingmodelsofuser acceptanceoftechnologyforelearningapplications.thisfollowsthedesignanddevelopment oftworesearchmodels,thetamandetam,twoquestionnairesandthreeanalysismethods, reliability,factoranalysisandregression;usedtoevaluatetheuseracceptanceoftechnology FourthChapter:ResultsTAM This chapter provides and explains the results of the TAM s evaluation of the ecourse. It shows how this study identified the most influential factor using reliability, factor and regressionanalyses.finally,anyanomaliesintheresultsareexplored,whichareexplained withreferencetopastworksofotherresearchers FifthChapter:eTAMResults ThischaptershowsthedevelopedTAMmodel,theeTAM,anditsassessmentinbeingable to evaluate the user acceptance of technology for elearning applications. A threestage analysisprocessofreliability,factorandregressionisusedtoevaluatethepredictivestrength ofthenewmodelandbringoutanyweaknessforfurtherdevelopment.themostinfluential factoroftheetamisidentified,whichconstitutesasthegreatestbarrierforuseradoption ofthetttebmelearningapplication SixthChapter:ConclusionandFurtherWork ThefurtherworkchapterprovidesusefulinsightintohowtheTTTEBMprojectmayimprove withtheknowledgegeneratedinthisthesis.newopportunitiesarepresentedforresearchers of technology acceptance and technology acceptance modelling. Finally, pointing out the 12

22 potentialflawsoftheetamandquestionnaires,aswellasexploringhowresearcherscould resolvethoseflaws. 13

23 2 BACKGROUNDANDLITERATUREREVIEW 2.1 Introduction Evidencebasedmedicine(EBM),whenusedeffectivelybymedicalpractitioners,isconsidered beneficial for the quality of healthcare (Oude Rengerink, et al., 2011), job performance competency (Hatala, et al., 2006) and ethics of medical practitioners approaching the decisionmaking process of patientpractitioner interactions (Thangaratinam, et al., 2009). ThereisalargenumberoflearningresourcestoteachEBMuse,whichmayconfusethetrainer ofwhatcurriculumtofollow(coppus,etal.,2007).moreover,thereisaneedformodelling and prediction of how EBM trainers teach medical practitioners to use EBM in clinical environments. An approach to this challenge is modelling the trainer s acceptance of a blendedlearningapproachtoebmstudy. EBM trainers, as their profession, are responsible for providing medical practitioners an approachtopracticeebmandrelatedtechnologiesinworkplaces.thiscreatesalinkfromthe traininggivenbyebmtrainerstothebettermedicalpracticereceivedbypatients.becauseof thislink,wewanttoknowandquantifythetrainers barriersassociatedtoebmteaching. The identification of barriers affecting EBM teaching involves an assessment of tried and testedmethodsinsimilarareas.factorscanrepresentbarriersforuseinaresearchmodel; Section2.4discussesthecategorizationoffactors.Establishingaresearchmodeloftrainers acceptanceofinnovativeebmtrainingmethodscanprovideasolidbasistoimproveebm trainers teaching methods. Ultimately it is assumed the outcomes of betterebm training 14

24 methods reflect in medical practitioners providing better EBM utilisation and thereby the patientsreceivingbetterhealthcare. Publishedresearchfromthe1950 suptothepresenttimehaveshownthedevelopmentof models that have become increasing able to predict the user acceptance of innovative technologies,wherepresentdaymodelscanassessuseradoptionofelearningtechnology andsystemsassociatedtoehealthorhealthcare.outcomesofpreviousdevelopmentonuser acceptance modelling has shown a general trend of fitting models to specified users or applicationssuch as modelling trainers to further develop the innovation process of e learningapplications,asdiscussedinsection2.4.byfurtherunderstandingthereasonsbehind theoutcomesofrelevantresearchmodels,whendevelopingones ownmodel,itispossible to have a better approach to modelling by avoiding the pitfalls and setbacks of previous researchers. Moreover, the selection of external factors from previous research gives a backgroundtocompareresultswiththisthesis. Theliteraturereviewaddressesthreeresearchareasassociatedtodevelopingmodelsthat predict user behavioural intention and assessment of EBM trainers that teach medical practitionersinaclinicalenvironment.thefirstsection,whichexplainstheneedsofpatients, medicalpractitionersandebmtrainers,addressestheuseofelearningforebmandteach thetrainersebm(tttebm),wheresection2.2.4explainsmoreaboutthetttebmproject. Thesecondsectiondescribesandexplorestheoreticalmodellingincludingmodelsthatcan assess a user s acceptance of technology. This section also discusses similarities of the differentpropertiesofthemodels,suchasfactorsandhypotheses,aswellaswhatfeatures ofthemodelsarebetteratassessingtheuseracceptanceofinnovativetechnologiesincluding 15

25 elearning.thefinalsectiondiscussesandanalysestheimportantexternalfactorsconsidered asbarriersorfactorsinresearchjournalssuchasajzenandfishbein(1980),davis(1986), Masrom (2007), Thangaratinam (2009), Oude Rengerink (2011) and many other studies associatedtotechnologyacceptancemodelling.thesechosenexternalfactorsareusedinthe developmentoftheetammodel,whichfitstheusersofthetttebmelearningprojectand isillustratedinsection TheuseofeLearningforEBMtrainersanditsacceptance Knowledgeofelearningandhowusersfeelaboutusingitasateachingtechnologyapplication leadstounderstandinghowclinicianscanintegratethelearningintoaclinicalsetting(sun,et al.,2006;saadé&kira,2009;pituch&lee,2006).hereareafewwellknownbenefitsofe learninginthecontextofthisthesis: Elearningfacilitatesuserstolearncoursesthatareaccessibleonline,whichprovides flexibilityoftimefortheebmtrainerstolearn(sun,etal.,2006;ong,etal.,2004). EBM can be sourced on an elearning framework, which gives EBMtrainers the freedom of learning in front of any Internetenabled computer (Sun, et al., 2006; Masrom,2007). An elearning qualification, on how to teach EBM to medical practitioners, gives conformity to EBMtrainers on how to teach (Thangaratinam, et al., 2009; Oude Rengerink,etal.,2011;Coppus,etal.,2007). TheEBMtrainerbeforeadoptionoftheelearningEBMcourse,needstoconsiderwhetherit fits into their routine, encourages overall work productivity and delivers a learning achievement. 16

26 Generally,elearningcanbeusedforInternetenabledlearningofvarioussubjectsandhas increasingimportanceinteachingsystems(pituch&lee,2006;sun,etal.,2006;ngai,etal., 2007).Theadoptionofelearningintoinstitutesworldwidehasseenanincreaseinrecent years(sun,etal.,2006;saadé&kira,2009).accordingtoareportfromtheinternationaldata Corporation,theelearningmarketintheUnitedStateshadanestimatedincreasefrom$2.2 billionto$23billionfrom2000to2004(pituch&lee,2006).elearningisconsideredasan opportunityforhighereducationstudentstodevelop.eventhoughithasadvantagesover traditional education (Sun, et al., 2006), it is not considered as a replacement for the traditionalclassroomcourses(masrom,2007) ELearning Elearning systems use ICT (Information Communication Technology) to facilitatelearning, whichisbeneficialfordistancelearners,learnerswithlimitationsoftimeandplaceaswellas learnerswholiketolearncollectivelyorasagroup.similartotheabovestatement,singhet al. (2005) mentioned that there are various ways to define elearning, such as distance learning, online learning and networked learning. The IEEE Learning Technology Standard Committeedescribeselearningasasystemsimilartowebbasedlearningsystems,where userscanaccessandlearntheircoursematerialonline.peoplemayaccessthissystemthrough webbrowsers,whichisaninterfaceforuserstolearnandpracticewithapplications(ieee, 2009).ElearningusesICTsopeoplecanaccessinformation,suchasmedicalrecords,that administrators could store on a single server, which is also updatable from a distance. ThereforeusingICTforteachingandlearningpurposesisalsoconsideredelearning(Masrom, 2007).Thesefeaturesfacilitateandsupportlearningforamasspopulation. 17

27 There have been many studies on the benefits of elearning, such as shown in Hofmann (2002),KekkonenMonetaetal.(2002),Pituch&Lee(2006)andNgaietal.(2007).Kekkonen Moneta(2002)foundstudentsfromelearningclasseshaveasimilarprogressasagroupin facetofacelectures.infavourofelearning,hofmann(2002)stateddistancelearningwas betterduetousersneedingatobecome activelearners,ortodevelopahighselfesteem to learn individually without force from a teacher or supervisor. Pituch & Lee (2006) also statedelearningasbettersuitedtodistancelearnersbygivingtheuserfreedomoftimewith advantagesofprovidingsynchronousorasynchronouscommunicationbetweenteacherand student.moreover,ngaietal.(2007)foundthatstudentshaveapositiveapproachtoonline learning and better learning outcomes using elearning and WebCT (The University of Birmingham,2012)comparedtolecturebasedcourses.Overall,theyshowedthatelearning enablesflexibilityofstudy,studyimprovementaswellasreduction ofcostsforacademic institutions(ngai,etal.,2007) LearningEBMwithelearning The fundamentals and reasoning behind the use of EvidenceBased Medicine (EBM) were discussedinsection1.1.ebmisamedicaldecisionmakingapproach,enablingpractitioners to improve and evaluate patient care by finding the best available evidence through the considerationofuptodateavailableinformation(kulier,etal.,2008b;bmjpublishing,2009). TeachingandlearningEBMinanelearningtrainingenvironmenthassimilaradvantagesto other lecturebased lessons. Davis et al. (2007) found equal knowledge gains between e learning and traditional courses, whereas Ngai et al. (2007), in a literature review, found studentstakingtheonlinecourseoutperformedthosetakingthetraditionalclassroombased course. There are many different ways to get involved with EBM by attending courses, 18

28 conferences, workshops, journal clubs, educational meetings, including having access to medicalliteratureandguidelines(khan&coomarasamy,2006;thangaratinam,etal.,2009). Figure1:Traditionalclinicalpracticeandimprovementcycle,reproducedfrom Thangaratinametal.(2009) Figure 1 illustrates the traditional approach to improving medical practice in the clinical environment (Thangaratinam, et al., 2009). The process involves medical practitioners assessing patients problems by using their experience and understanding of medical practices.fromthat,theymakeadecisiononadiagnosis,andtheprocesscontinueswitha selfimprovement cycle from treatment to follow up. A significant limitation of this EBM frameworkisthelackofuptodateknowledge,thelackofevidencetorecogniseapatients problemandthelackofsolutionsforit(hatala,etal.,2006).thestudyfromhatalaetal. (2006)highlightstheneedforEBMtobeintegratedintotheworkplaceandtaughtforusein clinician s daily activities. Hatala et al. (2006) suggest EBM and its training facilities be improvedtoincreasethequalityofthedecisionmakingprocessinclinicalpractice. 19

29 Figure2:ImprovedlearninganduseofEBMinclinicalenvironmentwithadditionallearning resources,reproducedfromthangaratinam(2009) AnimprovementmodeltoFigure1isFigure2,whereseveralEBMlearningprocessesare proposed to be blended into the workplace (Thangaratinam, et al., 2009). This includes a blendedlearningapproachwithelearningcoursesthatdirectlyrelatetoebmpractice.such ecourses include audit meetings, journal clubs, mortality meetings and ward rounds. The blendedlearningapproachmeetsthedemandforclinicianstolearnebmintheworkplace,as otherebmcoursesinterruptedtheflowofroutineclinicalactivities(khan&coomarasamy, 20

30 2006).CliniciansshouldlearntoworkwithEBMsotheycanutilisethebenefitsofimproved job performance and approach medical decisions with confidence, but more importantly, trainersneedtotraincliniciansto useebmwithoutitdisturbingclinicians workflow. The impactofpredictingthetechnologyacceptanceoftheblendedlearningapproachwithane course would allow elearning designers to develop an application that fits the needs of cliniciansandtheirroutine ImprovingEBMusewithTTTEBM Trainers have traditionally taught medical practitioners EBM outside the context of the trainees workplace;thishascausedagapbetweenlearningandpractice(korenstein,etal., 2002).Korensteinetal.(2002)undertookastudythatwouldchangethecurriculaofEBM trainers,byplacingebmtrainingintotheclinicalsetting.dawesetal.(2005)alsostudiedthe skillsandtrainingneededtopracticeebmapproachesandcurricula.intheirdiscussion,they pointedouttheneedforcleartrainingthatconnectstoskills:peoplewhoworkinhealthcare organisationsneedtohavetheabilitytoprocessnewknowledgebyobtaining,quantifying, usingandinvolvingitintheirjobs(dawes,etal.,2005).meaningtheircareerswillinvolve takingonprogressivedevelopmentsofevidence,trainingandpractices. Davisetal.(2007)studiedcomputerbasedlecturingforpostgraduatesandundergraduates. InsupportofKekkonenMoneta s(2002)results,davisetal.showedthatcomputerbased teaching is asgoodas facetoface lecturing for EBM. The two studies from Davis have introducedcomputerbasedteachingasaneffectiveandinnovativemeansofteachingebm. Computerbased learning and teaching has an advantage over facetoface teaching and learningasfollows: 21

31 Flexibilityintimeandworkplace Interactiveforlearnerswithcontrollableprogresssuchaspauseorrevise Standardisedteachingqualityforamassaudience LearnershaveaprovenacceptanceofEBMwhileusingcomputerbasedteachingsystems (Davis, et al., 2007). Their results highlights that computerbased teaching or elearning systemscanbeusedtoreachthesamelearningoutcomesasthetraditionalapproach. ThetrainersofEBMcanbothgivetraditionalandelearningcoursestomedicalpractitioners, buttrainersmustunderstandthedifferentcontextsoftheirtrainingsessions.beforethettt EBM project initiated, the need to move the teachers from standalone courses to more integratedworkbasedteachingwasprovenbyotherstudiessuchas,delmaretal.(2004), KhanandCoomarasamyin(2004;2006),Coppusetal.(2007)andThangaratinametal.(2009). Theseauthorssummarisedthereasonsforthiskindoftrainingbecausetrainershadalackof confidence: IndemonstratingEBMintheworkplace InteachinghowtoapplyEBMatthesametimeasotherclinicalactivities InunderstandingabouttheavailableopportunitiesandfacilitiestoteachEBMina clinical workplace to distinguish, incorporate and apply uptodate evidence to improvehealthcare TTTEBMelearningcurriculum EBMtrainers lackofconfidencewasachallengetoaddressforthangaratinametal.(2009), who designed an elearning curriculum for teaching the trainers, aiming in the successful 22

32 deliveryofebmtrainingsessionsineverydayclinicalpractice.thiscurriculumaimstohelp cliniciansbecomingfamiliarwiththefacilitiesavailabletotrainebm,suchasonlinelearning inamultimediaformatseefigure3,andtherebygivethemconfidenceinprovidingtheebm trainingcourse.amainobjectivetothetttebmprojectistostandardizeaeuropeanwide curriculumofteachingebminclinics,whichisintegraltoclinicians workschedule.thecourse useselearningtoincorporatetrainingandonthejobpractice,whilecarryingoutroutine tasks,whichgivesthedoctorsflexibilityoftimeandplace.theoverallgoalisthiselearning applicationincreasesebmtrainers confidenceleadingtothedeliveryofbetterhealthcare provisionandorganisationsthatbenefitthepatient(zanrei,2009). Figure3:TTTEBMelearningcourseaccessibleviatheInternet:Module1 WardRound The TTTEBM elearning curriculum contains six steps that are accessible in five different languagesandcoveredinthefollowingsequentialorder:wardround,evidencebasedjournal club,formalclinicalassessment,outpatientclinic,formalclinicalmeetingsandclinicalaudit meetings(thangaratinam,etal.,2009)assowninfigure4.eachcoursecomprisesofdifferent 23

33 learningmaterialincludingvideoclipsofrealworldebmpracticeandelearningcontentto informhowtoteachebm,whichalsoprovidesresourcesonthemodules(zanrei,2009).an exampleoftheelearningpartofonemoduleincludesa10 15minuteseminarfollowedby anonlinemultiplechoicequestiontest. Figure4:ThesixclinicalsettingsthataretaughtintheTTTEBMelearningcurriculum, reproducedfrom(zanrei,2009) TheTTTEBMprojecthasasitsmainpurposetohelpEuropeanhealthcareoverall.Asstated in the introduction of this thesis, this study supports the aims of the TTTEBM project by developingthetamintoamodelthathasfactorstoassessebmtrainers acceptanceofan elearning application, where the TTTEBM designed the application. In Chapter 4 and Chapter5,thisstudyusedtechnologyacceptancemodelstoanalysequestionnairesrelatedto thetttebmproject,wherethequestionnaires studygroupwereebmtrainers.asignificant outcomeofthosechapterswasnotjustevidencetosupportthemodels,butalsoprovided resultsthatthetttebmprojectmayusetoimprovetheuseracceptanceoftheirelearning course.thenextsectiongoesintothedetailoftheoreticalmodellingandcollatesfactorsfor developmentofthisthesis etammodel;thedesignofthemodeliscoveredinsection

34 2.3 TheoreticalModellingofUserAcceptance:FromTheoryandModellingto PracticeandAnalysis EBMtrainerscanuseelearning sadvantagesoftimeandplaceflexibilitywhenstudyingebm; however,thedisadvantageistheymustacceptnewtechnologyintotheirworkinglives.design anddevelopmentoftheoreticalmodelsfacilitatesthemeasurementofauser sacceptanceof technology. A user, being either medical practitioner or EBMtrainer, will have individual reactionstotechnology.thesereactionsareclassifiedintobehaviouralconstructs,alsoknown asfactors,whichebmcoursedesignersormanagerscantrytoinfluence,orbarrierstheycan reducetoimprovetheirsystem. Researchers often develop their own technology acceptance model based on empirical evidence or knowledge of established models, such as the Technology Acceptance Model (TAM)seeSection Theymaydevelopamodelintoamodelthathasfactorsspecificto theapplication,whichrepresentsademographicusergroup sacceptanceoftechnology,such asngaietal. s(2007)modelfortheacceptanceofwebbasedlearningsystems.section2.3.2 showssomeotherapplicationspecificmodels. Varioussourcesofresearchtrytoexplainpeople sbehaviour,intention,beliefandattitude towardsusingtechnology.researchinthisfieldhasdevelopedtheoreticalframeworksofuser technologyacceptancetoimprovetheusers adoptionofnewtechnology,suchaselearning systems.thissectionofthechaptercontinueswithexamplesofsuchstudiesinthefieldofe learning, and the use of the associated technology acceptance theories. These build up a stronginformationbasisforthedevelopmentofaspecificmodelforusewithinthecontextof thetttebmblendedelearningapproach,asdiscussedinsection

35 2.3.1 TheoryandModellingofUserAcceptanceofTechnology Thissectiondiscussesthefollowingfrequentlycitedeightresearchmodelsrelatedtouser behaviourassessment: InnovationDiffusionTheory TheoryofReasonedAction(TRA) SocialCognitiveTheory TheoryofPlannedBehaviour(TPB) TechnologyAcceptanceModel(TAM) TechnologyAcceptanceModel2(TAM2) TheUnifiedTheoryofAcceptanceandUseofTechnology TechnologyAcceptanceModel3(TAM3) ThereviewconsiderstheTAM sdevelopment,aswellasstudytheuseandpredictivepower of other wellknown models for user acceptance of technology. This study uses this informationtodevelopanewmodel InnovationDiffusionTheory InnovationDiffusionTheoryexplainstheinnovationdecisionprocess(Kripanont,2007),being theeventsleadinguptotheacceptanceorrejectionofusinganinnovation.rogers(1995) developedandimplementedthistheoreticalmodel.itistheprocessofadaptinginnovation bypredictinghowusersadoptit,haveanattitudetowardit,decidetoacceptorrefuseit,put ittoeffectiveuseandvalidatetheirreasontouseit(rogers,1995;kripanont,2007).the theoreticalmodelismainlypopularamongresearcherstomodelandpredicttheinnovation decisionprocessofusers. 26

36 Innovationdiffusioncansucceedusingaprocessoffivestepsovertime,resultinginadecision makingprocessasshowninfigure5.rogers(1995)cameupwiththefollowingelementsin theprocess: 1. Knowledgeauser sunderstandingoracquaintancewithaninnovation 2. Persuasionanindividual spositiveornegativeattitudetowardsinnovation 3. Decisionanindividualweighsadvantagesanddisadvantagesofacceptingorrejecting aninnovation 4. Implementation auserstartsusingtheinnovationithelpstheindividualrealisethe usefulnessoftheinnovation 5. Confirmationwhen the relation between an innovation and a user's decision is verified 27

37 Figure5:TheInnovationDecisionProcessCommunicationChannels,reproducedfrom Rogers(1995) Venkatesh(2003)presentedtheapplicationofInnovationDiffusionTheorywithsevencore constructs,basedonmooreandbenbasat s(1991)studyofindividualtechnologyacceptance asfollows: 28

38 1. RelativeAdvantage howmuchitisperceivedasbeingbetterthanpresentorpast technology 2. EaseofUsehowmuchitisperceivedasbeingdifficulttouse 3. Imagehowmuchtheuseofitisperceivedtoenhanceone sreputationorsocial status 4. Visibilityhowmuchawarenessthereisofothersusingit 5. Compatibilityhowmuchitisperceivedasbeingconsistentwiththeexistingvalues, needs,andpastexperiencesofpotentialusers 6. ResultsDemonstrabilitythetangibilityoftheresultstouseit,includingtheirability toobserveandcommunicate 7. VoluntarinessofUsehowmuchtheuseofitisperceivedasbeingvoluntary,orof freewill Innovation Diffusion Theory has been used to understand the linear and time dependent connection between an innovation and the decision processes associated to that new technology or idea. Rogers (1995) initially put the process into a linear format. Then the properties of the innovationdecision process were categorised by Venkatesh (2003) into sevencoreconstructsforstudyingindividualtechnologyacceptance.itisusefultocategorise behaviouralpropertiesofusers,becauseitisaninitialstepindevelopingatheoreticalmodel. To sum up Innovation Diffusion Theory has been used to aid the development of other theoriesandmodels,shownbelow,whichhavetheabilitytounderstandthefundamentaland coreconstructsofhownewtechnologiesorideascanbeadoptedbyusers. 29

39 TheoryofReasonedAction TheTheoryofReasonedAction(TRA)isamodelthathasbeenusedtopredictthebehaviour ofindividuals,whentheywanttoaccomplishavoluntarypredeterminedtaskorobjective (Sheppard,etal.,1988).FishbeinandAjzen(1975)havedevelopedit;ithasalsocontributed totheinitialdevelopmentofthetam,(asdiscussedinsection ).trahasbeenusedin studies to identify the relationships between behavioural intention (BI), attitude (A) and subjectivenorm(sn),where BI=A+SN,wherethisequationpredictsauser sbehaviourin doingavoluntaryaction(davis,1986). Fishbein and Ajzen (1975) and Ajzen and Fishbein (1980) defined behavioural intention, attitude, belief and subjective norm as follows: Behaviour can be commonly described as beingunpredictableandaseitherrebellingorconformingtosocialacceptance.forexample, culture,attitudes,emotions,values,ethics,authority,rapport,persuasionandcoercioncan influence human behaviour. Intention is the action of a person that drives them to do something specific, it is the relation between a person and their action. An attitude is an individual sstateofmindorperceptionthateitherfavoursordisfavourssomething.itcanbe a type of bias for evaluative response, which is positive or negative. Belief is a form of connectionthatpeopleassumebetweentheattitudesofeachotherortoinanimateobjects. Subjectivenormistheinfluencefromotherpeopleon ones behaviour.italsoshowsthe impactotherpeoplehaveonone sbeliefs,whichhasaconsequentialeffectonbehavioural intention. More information regarding behavioural intention and attitude can be found in Section

40 AsoutlinedbyVenkatesh(2003),thecoreconstructsofTRAareattitudetowardbehaviour andsubjectivenorm,whichhaveadirectinfluenceontheuser sbehaviouralintentiontodo atask.thiscomplementssheppard s(1988)explanationofhowauser sbehaviouralintention to accomplish a certain task is very dependent on their attitude or behaviour toward the necessaryprocessesindoingthattaskandhowmuchtheyregarditasasubjectivenorm.as illustratedinfigure6,whenauserhasastrongattitudeorbehaviourtowarddoingataskand a high regard of their task being a subjective norm, they will ultimately have a strong behaviouralintentiontocarryouttheirtask.trahasbeenasuccessfulmodelinpredicting theuserintentionsandbehaviourwhenwantingtocarryoutavoluntaryaction(sheppard,et al.,1988). Figure6:OriginalversionofTheoryofReasonedAction(TRA),reproducedfromAjzen& Madden(1986) AjzenandFishbein(1980)includednewtheoriestothemodelshowninFigure6,andfurther expandedandillustrateditasshowninfigure7.asthefigureillustrates,auserhasabelief andevaluationaboutcarryingoutabehaviour,whichhasadirectinfluenceonattitude.the userwouldalsohaveanormativebelieffromtheirsociety,i.e.beliefsthatfollowthoseof people close to them, such as family or friends, and a degree of motivation to behave in accordancewiththeirsociety sbeliefs.thesetwofactorscombinedhaveaneffectonthe intentions of the user, which depends on how much the user feels their social group is 31

41 pressurizingthempeerpressure.theattitudeorbehaviourtowarduseoftechnologyand subjective norm are directly related to a user s behavioural intention. Meaning the user s belief,subjectivenormandevaluationonhowtoactaswellasmotivationtocomplyhavean overalleffectonthebehaviouralintentionofausertodotheirvoluntaryaction. Thismodelpresentsthepositiveornegativeoutcomeofasystem.Theoutcomeispositiveif thebehaviourofapersondevelopsintoapositiveattitudeaboutthebehaviourandifitisa negativeoutcome,itisviceversa(kripanont,2007). Relativeimportance ofattitudinal& normative considerations Beliefs and Evaluations Attitude Toward Behaviour Normative Beliefs& Motivation tocomply Subjective Norm (SN) Behaviour Intention (BI) Behaviour (Actual Usage) Figure7:TheoryofReasonedAction(TRA),reproducedfromAjzen&Fishbein(1980) TRAiswellknownasbeingthebackboneforthemajorityofstudiesassociatedtotherelation ofattitudeandbehaviour.however,bagozzietal.(1992)identifiedaweaknessoftrainthe determinantsofattitudesandpredictedintentions,statingthatdeterminantscloselylinkto actualbehaviour.sunandzhang(2006)alsoidentifiedaweaknessoftrainitsabilitytoassess fully the technology acceptance of users. They reached this conclusion by comparing the 32

42 explainedvarianceofexplanatorypowersofthetechnologyacceptancemodelwithtra(sun &Zhang,2006). Since1975,whenTRAwasfirstintroducedbyFishbeinandAjzen(1975),thismodelhasbeen analysedanddevelopedtoimproveitsabilityofpredictingauser sbehaviouralintentionto carryoutavoluntaryaction.davis(1986)addressedsomechallengesofthetrainassessing auser sbehaviouralintentionandtherebydevelopeditintoanewtheoreticalmodel,which helatervalidated,asshowninsection8.1. Modellingbehaviouralintention,attitudeandsubjectivenorm,aswellasotherbeliefsinthe TRA model, has provided insight into a user s intention to perform a voluntary action. Section discussedhowstudiesusedinnovationdiffusiontheory,whichsimilartoTRA explainshowauser sperceptionoftechnologyhasaneffectonitsadoption.asshownin Section ,Davis(1986)consideredthecoreconstructsofTRAtodevelophistechnology acceptance model. Background information on innovation diffusion theory and TRA representsthebasisofthecoreconstructsofdavis model,whichwasneededforthisthesis tounderstandanddevelopdavis modelintotheetam.thefollowingparagraphsreport furthertheoriesandassociatedmodelstodistinguishadvantagesfromdisadvantagesanduse themforimplementingthisthesis model SocialCognitiveTheory Theuser sbehaviouralintentiontocarryoutavoluntaryactionmaychangedependingonthe influencefromtheirsituation,societyandgovernmentlegislation.thiscanhappenifaperson hasachangeintheirlives,whichhasadirectimpactontheirvoluntaryaction,suchaswhen 33

43 wantingtouseanelearningsystem,buttheirinternetserviceproviderdecreasestheirusage allowanceandthereforecreatingabarriertousingelearning. Social Cognitive Theory provides a model to predict the human behavioural changes, as introducedbybandura(1986).heassumedthattheperson,behaviour,andenvironmentare allcombinedtocreatelearninginanindividual.themodelshowninfigure8representsthe linkbetweenapersonandtheirbehaviour,thisrelationincludestheperson sthoughtsand actions. Behaviour Environmental Factors Figure8:SocialCognitiveTheory,reproducedfromBandura(1986) People learn by observing others, where the environment, behaviour, and cognition are factorsthatinfluencetheirdevelopment.venkateshetal.(2003)studiedthetheoriesrelated toanindividual sacceptanceoftechnology.hestatedthatsocialcognitivetheoryisoneof the most powerful theories that can describe the behaviour of people. In addition, he commented on how researchers used Social Cognitive Theory in the context of computer utilisation.helistedthebehaviouralconstructsofthismodelaspersonalandperformance expectations,aswellasselfefficacy,affectandanxiety. Personal Factors: Cognitive, Affective, and Biological Events 34

44 Personalorperformanceexpectationsarewhattheuserassumestogetasanoutcomeof achievingapersonalorjobrelatedgoal.auser sselfefficacyistheirabilityandskillinusing technology.venkatesh(2003)describedtheuser semotionstousingtechnologyas liking it, astheaffect,andhavingdoubtstowarditsuse,astheanxiety TheoryofPlannedBehaviour Auser sbehaviourtocarryingoutavoluntaryactionisrelatedtotheinfluentialeffectoftheir surroundings and their ability and motivation for adopting that behaviour. The Theory of PlannedBehaviour(TPB)hastheabilitytoassesstheuser sbehaviouralintention,basedon theirattitude,socialcognitionandintentionalmotivation,forspecificactivitiesorvoluntary actions (Ajzen, 1985). This being useful because, as stated by Ajzen (1985), the user s behavioural intention differs depending on the voluntary action, they also recommend avoidingassessingtheuser sbehaviouralintentioningeneralvoluntaryactivities. TPBwasintroducedbyAjzen(1985)anditcompletestheTRAmodelbyaddingperceived behaviourcontrolasafactor.thelatterisdefinedashowpeopleseetheirbehaviourtotheir intendedactionaseasyordifficult(ajzen,1991).inaddition,ajzenandcote(2008)identified itasthepotentialthatapersonhastocompleteaspecificbehaviour.theperceivedbehaviour controlsectionwasaddedtotpbtosolvetheapparentchallengeswiththetra,suchaswhen dealingwithbehavioursoverwhichpeoplehaveincompletefreedomofcontrol(kripanont, 2007). TPB explains human behaviour as an action that is influenced by behavioural beliefs, normativebeliefsandcontrolbeliefs,whichasfigure9illustratesarepredictorsforattitude towardthebehaviour,subjectivenormandperceivedbehaviourcontrolrespectively.there 35

45 was an assumption that supporting those user constructs will increase the person s acceptanceofthevoluntaryactionandwillconvincethepersontoperformthatvoluntary action(kripanont,2007). Figure9:TheoryofPlannedBehaviour,reproducedfromAjzenandCote(2008) Behaviouralbeliefsarethosethatareproducedfrompositiveornegativeattitudestowards thebehaviour(kripanont,2007).normativebeliefsarethoseofindividualsorgroupsthat comefromperceivedbehaviouralexpectations(kripanont,2007).controlbeliefsarethose thatarisefromtheexistenceoffactorsthatcaninfluencethebehaviourandspecifypeople s feelingsofbeingincontrol.thiscouldincreasethepowerofperceivedbehaviourcontrol (Kripanont, 2007). Attitude toward the behaviour is identified as an analysis of the user behaviourthatisbiasedeithertoencourageortodiscourageanaction(ajzen&cote,2008). The subjective norm is identified as the influence that society has on the behaviour of a person. The influence of society can change the user s original plan or behaviour to that voluntaryactionnegativelyorpositively(ajzen&cote,2008). 36

46 TPBisknownasthemostpopulatedsocialpsychologicalmodelforbehaviouralstudies(Ajzen & Cote, 2008). However, Mathieson (1991), Chau and Hu (2001; 2002) evaluated the TPB modelandfoundthatperceivedbehaviourcontrol,oneofitsfactors,isgenerallyweakerthan thefactorsofthetamintheirabilitytopredicttheuseracceptanceoftechnology.moreon thetamandthereasonsbehinditschoiceasthebasetodevelopthethesis modelfollows TechnologyAcceptanceModel The Technology Acceptance Model (TAM), established as an empirical model, is also a theoreticalframeworkthatdesignersusetopredicttheuser sbehaviouralintentiontoward newtechnology.tamcanbeappliedtoevaluatedatacollectedfromaquestionnaireorsurvey of a target group of users of a system; the model outputs an assessment of the users acceptanceoftechnologyandidentifiesthemostinfluentialtechnologyacceptancefactor, representedasthecoreconstructsoftheuser.thetamhasbeenwidelyusedandstudied, thefollowingreviewshowitdevelopedincludingitsstrengthsandweaknessesasabasisfor itsdevelopmentintothethesis smodel,theetam. TheTAMwasderivedfromTRAtomodeltheuseracceptanceofITsystems(Davis,1986).This model has enabled designers to understand the advantages and disadvantages of design elements of technology, while the technology is in early implementation stages, and thereaftertheycanimprovethesystems design. Davis (1986) initially tested the characteristics of user acceptance of computerbased informationsystems.thetraasstudiedinsection andthetamshowninfigure10 havesimilarities,becausedavisusedthetraasabasisformodeldevelopment.thecore constructsofthetamareperceivedusefulnessandperceivedeaseofuse,whiletheother 37

47 coreconceptsofbehaviouralintentionandattitudetowardusehavebeenadaptedfromthe TRAmodel.Theirgenericmeaningsareasfollows. PerceivedEaseofUse(PEOU),auser spointofviewofhowhardoreasydoingan actionis(davis,1989) PerceivedUsefulness(PU),auser sconfidencethatusingasystemwillimprovetheir jobperformance(davis,1989) Behavioural Intention (BI), a user s motivation to start, continue and complete an action(fishbein&ajzen,1975) AttitudeTowardUse(ATU),auser spositiveornegativeinteresttousingthesystem (Fishbein&Ajzen,1975) Figure10showstheconnectionbetweenthefourfactorsoftheTAM;theyhavearrowsto indicateacorrelationbetweeneachofthem.eacharrowrepresentsatheoreticalrelationship, orhypothesis,thatpredictsonefactorwillhaveaneffectontheother.thetamoutputsa magnitudeanddirectionforeachhypothesis,moreinformationonhypothesesinsection3.3. Figure10:TheTAMmodelbasedon(Davis,1989) Davis(1989)reviewedawiderangeofarticlespurposelytoestimatethefactorsheneededin the TAM. He found hypotheses between the main factors (from the original TAM), 38

48 representedwithsolidlines1to6infigure11,andpredictedtheinfluenceofexternalfactors, designfeaturesandperformanceimpact,onthemainfactors,whicharehypothesesshown bydashedlines7to9infigure11. Figure11:TAMwithnumberedlinksbetweenthefactorsandexternalfactors,reproduced fromdavis(1986) Initially,itwaspopulartousetheTAMintheareasofInformationSystemsandIT(Information Technology)however,uptonowithasbeenusedinseveralotherareas.TheTAMispopular amongresearchersformodellingauser sintentiontouseinnovativesystems,wheretheir researchoutcomeshaveindicatedhowtomotivateindividualstoadoptandusetechnology. Theseincludeusingelearning(Ngai,etal.,2007;Moon&Kim,2001;Zhang,etal.,2008)(see Section 2.2.1), healthcare event reporting systems (Wu, et al., 2008; Chau & Hu, 2002; Goetzinger, et al., 2007), personal computers (Igbaria, et al., 1997), word processors and spreadsheets(chau,1996),knowledgemanagementinagriculture(folorunso&ogunseye, 2008)andpredictingintranetandportalusage(Chang,2004). 39

49 SincetheTAM sintroductionin1986,davisandotherresearchershavebeenevaluatingthe predictiveabilityofthetamtoexpanditsabilityinpredictinguserintentionandthereby proposingimprovementsforassessinguseracceptanceoftechnology.davis(1993)stated thatimprovementstothetamshouldbeabletoexplainauser smotivationtouseasystem becauseoftheinfluencefrommanagementorhierarchicalstructuresintheirsociety. ResearchershaveimprovedtheTAM scapabilitytopredictuserbehaviouralintentiontoward usingtechnology.venkateshanddavis(2000b)didthisbyintroducingadevelopmentofthe TAM,calledtheTAM2.TheTAM2isasignificantadvancementinpredictinguserbehavioural intention and acceptance of technology (Venkatesh & Davis, 2000b); more on TAM2 in Section MoonandKim(2001)havealsomadedevelopmentsfromtheoriginalTAM. They extended the TAM model through the addition of a predictor called Perceived Playfulness,whichsignifieshowmuchauserbelievesthesystemasplayful;associatedtofun activitieslikegames.withhypotheses,moonandkim smodelconnectspeoutoperceived PlayfulnessandPerceivedPlayfulnesstoATUandBI.Theirstudyincludedananalysisof152 graduatestudentsusingtheworldwideweb(moon&kim,2001).aftercomparinganalysis resultsoftheirextendedtamwiththeoriginaltam,theyfoundthattheirextendedtamhad a5%greaterexplanationofthevarianceinattitudeand4%greaterexplanationofthevariance ofbehaviouralintentionfromthegroup. InChapter4ofthisstudytheabilityoftheTAMisassessedinmodellingEBMtrainersasusers ofthetttebmelearningsystem,providingasbasisforthetam sdevelopmentintothee TAMbyaddingfactorsthatexplainthevarianceoftheTAM scoreconstructs. 40

50 TechnologyAcceptanceModel2 Theoreticalmodellingofuserbehaviouralintentionandacceptanceoftechnologyshouldhave a good balance between generalisation and being specific in terms of choosing factors. Generalisationoffactorsintogroupsallowsresearcherstoapplytheoutcomesofastudyto variousapplications,whereasbeingspecificdrownsoutsomedeterminantsofmainfactors thatareirrelevanttoausers acceptanceoftechnology. VenkateshandDavis(2000b)havedevelopedtheTAMtobecomelessgeneralisedandmore specifictoproduceresultsthatallowthemtodesignorganizationalinterventions,byincluding external factors of experience, image, job relevance, output quality, subjective norm and voluntariness.experienceandvoluntarinessaremoderatorstothehypothesesofsubjective norm, where both moderate the influence of subjective norm on intention to use and experience also moderates subjective norm s influence on perceived usefulness. These additionalorexternalfactorswereaddedtoseetheeffectthatdeterminantsoftheoriginal factorsofthetamhaveontheuser sbehaviouralintention,asshowninfigure12. 41

51 Experience Voluntariness SubjectNorm Image Perceived Usefulness JobRelevance Intention ToUse Usage Behaviour OutputQuality Result Demonstrability Perceived EaseofUse Figure12:TAM2ExtensionoftheTAM,reproducedfromVenkateshandDavis(2000b) Figure12showsinwhitefonttheexternalfactors,Experience,JobRelevance,Image,Output Quality,ResultDemonstrability,SubjectNormandVoluntarinessseetheGlossary,Section8.4 forthedefinitionofeachterm.theseadditionsalongwiththetambecamethetechnology AcceptanceModel2(TAM2).Eachadditionalfactor,apartfromexperienceandvoluntariness, representsthedirectdeterminantsofpu,whichcanexplainandpredictauser spuandhence theuseracceptanceoftechnology(venkatesh&davis,2000b). TechnologyAcceptanceModel(TAM) VenkateshandDavis(2000b)conductedastudyfortheTAM2toevaluate156participants from four different organisations. They chose two organisations where participants were given voluntary usage of the system in their study and another two organisations where participants were given mandatory usage of the system. This division of voluntary and mandatoryusagewasdesignedinordertoquantifytheeffectsofauserhavingachoiceto 42

52 carryouttheactionand toexplicitlyexaminethetheorizedmoderatingroleofvoluntariness (Venkatesh&Davis,2000b).Venkatesh&Davis(2000b)collectedthelongitudinaldatafrom foursystemsinfourorganisationsandtestedthetam2atpreimplementation,atonemonth after postimplementation and at three months after postimplementation. The results showedthetam2accountedfor 40% 60%ofthevarianceinusefulnessperceptionsand 34% 52%ofthevarianceinusageintentions (Venkatesh&Davis,2000b).Subjectivenormin thetam2exerted asignificantdirecteffect ontheuser sbehaviouralintention,whichwas greater than the TAM s perceived usefulness and perceived ease of use, but only when organizationsmakeuseofthesystems mandatory (Venkatesh&Davis,2000b).Theyalso found the role of social influence on perceived usefulness decreased as the user gained experienceovertime(venkatesh&davis,2000b).therefore,thereisasignificantpositive effectofaddingnewpredictorsofsubjectivenorm,experienceandvoluntarinesstothetam. Based on evidence from this literature review, there is an argument that adding direct determinantsofthetam sfactorsasexternalmeasurablevariablescangivegreateranalysis oftheoriginalfactorsofthetamorpredictabilityoftheusers intentions.itfollowstheidea thatthisthesishasoffindingbarrierstothetttebmelearningcourseandintegratethemas additionalfactorstothetam(seesection2.4.).thisjustasvenkateshanddavis(2000b)had done,isspecificenoughtofitthecontextofthisstudywhilebeinggeneralenoughtoapplyit tootheronlinelearningcourses.thisthesis workhasselectedfactorsspecifictousersofan EBMonlinecourse,whicharegeneralenoughforapplicationtootherfieldsofelearning. Section3.2.2showstheevolutionoftheTAMintotheeTAM,whichmodelsusersenrolledon anebmelearningcourse. 43

53 UnifiedTheoryofAcceptanceandUseofTechnology Themodelspreviouslymentionedinthischapter;InnovationDiffusionTheory,TRA,Social Cognitive Theory, TPB, TAM and TAM2 have different abilities in predicting the user acceptanceoftechnology.venkateshetal.(2003)decidedthatitwouldbebeneficialifthere wereonemodeltoexplaintheuser stechnologyacceptancethatiscomparativelybetterthan the majority of technology acceptance models. He presented the Unified Theory of Acceptance and Use of Technology (UTAUT), which takes into account conceptual and empiricalsimilaritiesacrosstheeightmodels.themodelisshowninfigure13. Performance Expectancy Effort Expectancy Social Influence Behavioural Intention User Behaviour Facilitating Conditions Gender Age Experience Voluntariness of Use Figure13:UnifiedTheoryofAcceptanceandUseofTechnology,reproducedfromVenkatesh etal.(2003) 44

54 These similarities were summarised after reviewing a wide range of literature to cross examinetheeightresearchmodels.theseareasfollows:thetra,thetam,themotivational model,thetpb,thecombinedmodelofthetamandthetpb,themodelofpersonalcomputer utilization,theinnovationdiffusiontheoryandthesocialcognitivetheory(venkatesh,etal., 2003).Theycomparedthemodelswithdatacollectedfromquestionnairesasadaptedfrom Davis(1989)andDavisetal.(1989).Venkatesh(2003)appliedhisstudythreetimesoverasix month period to four organisations. The results from the study led them to identify four predictors as direct determinants for user acceptance and usage behaviour namely performance expectancy, effort expectancy, social influence, and facilitating conditions. Venkatesh(2003)explainedtheseasfollows: Performanceexpectancyishowmuchauserbelievesusingasystemwillimprovetheir jobperformance Effortexpectancyistheamountofworktheuserthinkstheyneedtoputintousea system Social influence is the effect a user gets from the weight that is put on them by associatedpeoples ideas Facilitating conditions is the confidence level a user has of the support from an organisationortechnicalcentre Venkatesh(2003)analysedtheUTAUT,theresultsindicateditasastrongtheoreticalmodel. InVenkatesh s(2003)studytheutautaccountedfor70%ofthevarianceinusageintention comparedtotheeightmodels,whichexplained17to53%ofthevarianceinuserintentions touseinformationtechnology. 45

55 Theyshoweditsmainbenefactorstopredictinguserbehaviourandtheiracceptance.Firstly, it provides the direct influence of performance expectancy, effort expectancy, and social influenceonintention.secondly,thedirectinfluenceofintentionandfacilitatingconditions on usage behaviour. Finally, the most important founding was that the factors of age, experience,genderandvoluntarinesswerefoundashavingastronginfluenceandasessential touseinutaut(venkatesh,etal.,2003). TheUTAUTisinagreementwithTAM2;thedevelopmentofthosetwomodelsissimilarto howthisthesischoosesthebestfactorsfromothermodelstoexplainuserintention.ithas shownthatcombiningthefactorsofothermodelstoformonehadgivengoodresultsfortheir study,whichiswhatthisthesisexplainsfortheetaminsection3.2.2.thefactorsweretaken fromothermodels,whichisasimilarmethodusedinthisthesis,wherefactorsarecritically reviewedinsection TechnologyAcceptanceModel3 VenkateshandBala(2008)introducedtheTechnologyAcceptanceModel3(TAM3)forfurther explanationofhowandwhyusersacceptanduseitintheworkplace.theyfocusedtheirwork onclarifyingtheuseracceptanceandthepredictorsoftheperceivedusefulness(pu)and PerceivedEaseofUse(PEOU),whichoriginatefromtheTAM.Theymentionedhowimportant useracceptanceisfromanorganizationalpointofviewfordevelopmentofprojects,when theirusershaveagreateracceptanceandbetteruseofit.addingitwouldgiveorganizations anopportunitytoreducetheriskofprojectfailureandahugefinancialloss(venkatesh&bala, 2008).Theydrewoutafewexamplesofprojectsthatdidnotsucceedduetohavingalackof informationonuseradoptionoftechnology.forinstancein2004,hewlettpackardhada$160 46

56 millionfailurewhiletryingtoimplementit.inaddition,rolandwolframwhowasthegeneral managerofnike'sasiapacificdivisionin2000hada20%stocklosscostinghim$100million (Koch,December.1,2004;Koch,Jun,15,2004;Venkatesh&Bala,2008).Theyemphasisedthe valueofuseracceptancebecauseofits5.1%growthinresearchfrom2000to2004and7.7% from 2004 to The growth occurred because the academic and industrial disciplines agreedontheneedformanagerstodevelopandincorporatebetterittrainingmethodsfor theirstaff(jasperson,etal.,2005;venkatesh&bala,2008). VenkateshandBala(2008)usedtheTAMmodelwithfocusontheBehaviouralIntention(BI) as the base for their model. They also used the TAM2 s predictors for PU, which was introducedbyvenkateshanddavis(2000b),seesection andpeouaspredictors,which was introduced by Venkatesh (2000a) as anchor, adjustments and experience. The TAM3 itself,wasfinallyintroducedwithtwosetsofpredictors,oneforpu:subjectivenorm,image, jobrelevance,outputquality,resultdemonstrabilityandtheotherforpeou:computerself efficacy,perceptionsofexternalcontrol,computeranxiety,computerplayfulness,perceived enjoymentandobjectiveusability,seefigure14.thesechangesmadethetam3different fromthetam2withthreenewrelations: 1. betweenexperience,peouandpu 2. betweencomputeranxietyandpeou 3. betweenpeouandbi 47

57 Experience Voluntariness SubjectiveNorm Image JobRelevance OutputQuality Perceived Usefulness Result Demonstrability Anchor ComputerSelf efficacy Perceptionsof ExternalControl Behavioural Intention Use Behaviour Computer Anxiety Computer Playfulness Perceived EaseofUse Adjustment TechnologyAcceptanceModel(TAM) Perceived Enjoyment Objective Usability Figure14:TechnologyAcceptanceModel3(TAM3),reproducedfromVenkateshandBala (2008) 48

58 VenkateshandBala(2008)studiedtheTAMandTAM2fordevelopingamodeltopredict technologyacceptanceintheworkplace.theirmodelwassimilartothedevelopmentofthe TAM2 in finding the determinants of PU, but different because they also found the determinantsofpeou.asshowninthissection,thedevelopmentofthetam3wasdesigned inawaythatsupportstheideasofdevelopingresearchmodelforthisthesis,seesection2.4. ThesignificanceoftheTAM3isitsabilitytoassesstechnologyusageintheworkplace,which issimilartotheaimsofthisthesis.asshowninsection3.2.2,thisthesisdevelopsamodel specifictotheworkingenvironmentofitsusers theclinicalenvironmentofebmtrainers TechnologyAcceptanceofeLearning ElearninghasbeenintroducedandexploredforitsuseinSection2.2.Thisshowedthatthere isaneedforebmtrainerstolearnhowtoteachmedicalpractitionersintheirworkplaceand elearningcanfacilitatewidespreadtimeindependentlearning.togainthefullbenefitsof learningwithelearning,itisneededtoknowmoreabouthowusersorebmtrainersinteract and accept elearning technology. This section reviews studies that developed research modelsandhaveassessedtheuser stechnologyacceptanceofelearning. Ong et al. (2004) studied barriers that influence an engineer s acceptance of elearning systems. Using the TAM as a base, they developed new constructs, such as perceived credibility and computer selfefficacy and excluded attitude towards use. Both of these becameadditionalfactorsintheirmodeltoenhancethepredictionofuseracceptance. Perceived credibility was defined as the degree to which a person believed that using a particularsystem,suchasanelearningapplication,wouldbefreeofprivacyandsecurity threats(ong,etal.,2004).computerselfefficacywasdefinedas anindividual sperceptions 49

59 ofhisorherabilitytousecomputersintheaccomplishmentofataskratherthanreflecting simplecomponentskills (Compeau&Higgins,1995). InOngetal. s(2004)study,theyexaminedhowapplicablethetamwasconcerningtheir project.statingthattherewasasignificantincreaseinuseracceptancebecauseofashiftfrom productbasedtoknowledgebasedeconomy (Ong,etal.,2004),thisbeinganincreasein demandofintelligentusersforcomplexsystems.onemethodthathelpstheshiftisusinge learning,whichprovideseducationandtrainingandtherebyincreasesauser sknowledgeof a system. Their literature review presented the benefits of elearning when involved with limitationsoftimeandlocation,coupledwithotherbenefitssuchas: reducedcost,regulatory compliance, meeting business needs, retraining of employees, low recurring cost, and customersupport (Ong,etal.,2004). Figure15:Researchmodelandresultsofthehypothesistest,reproducedfromOngetal. (2004) 50

60 Figure15showstheresultsofthehypothesistest,wheretheyusedtheCALIS 1 procedureof SAS8.1,which providesestimatesofparametersandtestsoffitforlinearstructuralequation model (Ong,etal.,2004).Asshowninthediagramcomputerselfefficacyhasanegative effect on perceived credibility. However, the most significant overall effect was that of PerceivedEaseofUse(PEOU)onBehaviouralIntention(BI)(Ong,etal.,2004).Themodel supportedallofthehypotheseswhereitexplained44%ofthevarianceofauser sbehavioural intention.theadditionalfactorsofcomputerselfefficacyandperceivedcredibilitymanaged to explain the variance of the TAM s original factors. Out of the two additional factors, computerselfefficacyhadthegreatestinfluence,whichwasonpeou,indicatingthegreater proficiencyauserhaswithcomputerstheeasierthesystemseemstouse.thisresultinitially provedthatthetamcouldbeappliedforelearningandsuggestedthatifanengineerhada highercomputerselfefficacythentheywouldalsohavegreaterdoubtstowardsthesystem. Pituch and Lee (2006) studied what effect elearning has on learners and they developed theoreticalmodelstotestparticipantsandtherebyprovideinformationonhowtoimprovee learning.usingquestionnaires,theycollecteddatafrom259collegestudents.forthiswork, theyusedtheextendedtam,showninfigure16asthebasefortheirmodelsandadded systemandparticipantcharacteristics.theyusedaninstitutionallearningenvironmentthatis similartowebct,calledcyberuniversity,whichisanadvancedelearningsystem.thebenefit ofthesystemwasflexibilityinthetimeandlocationofstudy. 1 ThestatisticssoftwarepackageSAShasaprocedure,CALIS,whichisageneralstructuralequationmodelling (SEM)toolthatestimatesandfindsalinearfittotheSEM. 51

61 Figure16:Thetechnologyacceptancemodel,consideredasabaseforthenewmodelby PituchandLee,reproducedfromPituchandLee(2006) PituchandLee s(2006)firstresearchmodelintroducedimportantpredictorsofthebeliefsof users in other technology adoption studies, which can improve the development of an e learningsystem,seefigure17. Figure17:ModelAshows fullymediatedmodel forelearningusedbypituchandlee (2006),basedontheTAMwithexternalvariables Model A in Figure 17 has external variables of system functionality, system interactivity, systemresponse,selfefficacy,internetexperience,useforsupplementarylearninganduse fordistanceeducation.thesewerefoundtobeimportantpredictorsoftheusers beliefsfor 52

62 adopting technology. System functionality, system interactivity, and system response are categorisedassystemcharacteristics.systemcharacteristicsarewellestablishedpredictors becauseoftheirinfluenceonusers beliefsandtheiracceptanceoftechnology.pituchandlee (2006)dividedsystemcharacteristicsintothreeparts: Systemfunctionality,definedastheaccessibilityoffacilitiesthatanelearningsystem hastopresenttrainingandassessmentmediatousers System interactivity, defined as being the interaction of learning procedures to facilitatedevelopmentbetweenindividualstudentsandstudentswiththefaculty Systemresponse,definedaswhentheuserexperiencespromptreactiontimesfrom theelearningsystem,whicharereliableandpracticaltotheirareaofworkorstudy Pituchand Lee(2006) categorisedselfefficacyand Internetexperienceasuserattributes. Theydefinedselfefficacyasauser sproficiencyinlearningspecifictasksfromtheelearning system.theyalsodefinedinternetexperienceashowmuchunderstandingapersonhasof usingtheinternetandtheeffectthatthishasontheuseofelearning.pituchandlee(2006) statedthattheincreaseduseofsupplementarylearninganddistanceeducationareoutcomes oftheirstudy.theseoutcomesaretwospecificpurposesofthecyberuniversityelearning system. ThesecondmodelfromPituchandLee,ModelBshowninFigure18,hasthesamevariables andhypothesesbetweenfactorsasmodela,apartfromthreeadditionalrelationsexplained asfollows.inmodelb,itpredictsimprovementsinsystemfunctionalitywillincreasetheuse oftheelearningsystemasacomplementtotheclassroomlearningexperience,denoteduse forsupplementarylearning(usl)onthediagram.inaddition,thediagramillustratesthat 53

63 directeffectsonbothsystemfunctionality(sf)andsysteminteractivity(si)willboostauser s intentiontousetheelearningsystemtoaidlearningwheredistancepresentsabarrier,also calledusefordistanceeducation(ude). Figure18:ModelBshowsthe partiallymediatedmodel forelearninguse,reproduced frompituchandlee(2006) ThemostimportantfindingsfromPituchandLee(2006)aresummarisedasfollows: Systemfunctionalityhasthestrongesteffectonbothoutcomes SysteminteractivityhasastrongeffectonPerceivedUsefulness(PU) Overall,thesystemcharacteristicsarethemostimportantfactorsforuseracceptanceofthe elearningsystem.similartopituchandlee,anobjectiveofthisstudy saimistofindsuitable systemcharacteristics(externalfactors)fromarangeofresearchtodevelopatheoretical model(seesection3.2)thatcaneffectivelyassesstheuseracceptanceoftechnologyfora TTTEBMelearningapplication. Ngaietal.(2007)alsostudiedtheuseracceptanceofanInstitutionalLearningEnvironment, calledwebct,whenstudentsandlearnersinhighereducationwereusingitasanelearning facility.similartopituchandlee(2006),ngaietal.(2007)usedanextensiontothetamand 54

64 aimedtoidentifythefactorsthatinfluenceuseracceptanceandtoassessthecapabilitytheir modelhasatidentifyingthosefactors.theysummarisedliteraturebasedonelearning,user satisfactionandadoptionoftechnologytodevelopthetammodelintoanewtheoretical model,seefigure19. Figure19:ThemodelfortheacceptanceofWebbasedlearningsystems,reproducedfrom Ngaietal.(2007) This model included technical support as an extension to the TAM with the purpose of improvingtheuseracceptanceofwebbasedlearningsystems.technicalsupportincludesthe assistancefromtrainedandexperiencedemployeesfortheuseofcomputerhardwareand softwareproducts,andcaninvolvehelpdesks,hotlines,onlinesupportservices,machine readable support knowledge bases, faxes, automated telephone voice response systems, remotecontrolsoftwareandotherfacilities(wilson,1991;ngai,etal.,2007).ngaietal. s (2007)studyusedaquestionnairetocollateindividualdatafrom1263surveys.Theanalysis resultsprovedthatthetamisastrongempiricalmodelindemonstratingtheuseracceptance forthewebbasedcourseapplication,webct,asanelearningcourseforhighereducation. 55

65 Inaddition,theyunderlinethesignificantdirecteffectoftechnicalsupportonthePEOUand PU,andthestrongindirecteffectonattitude. Masrom(2007)usedtheTAMwithtwoimportantpredictors,PEOUandPU,toexaminee learninginuniversitiesasbeingeffectiveforlearningworkrelatedtasks.he,similartopituch andlee(2006),andngaietal.(2007),usedthetamtotestthehypothesespresentedby Davis,suchasATUpredictsBI,inthecontextofauser sadoptionofelearning.heintroduced elearningasanapplicationthatgivesnewopportunitiesforteachingandlearningbetween teachersandstudents.theoutcomesofhisworkhelpeddrawoutfactorsthathaveaneffect onauser sbehaviouralintentiontowardselearning. Figure20:Theresultofregressionanalysis,reproducedfromMasrom(2007) Figure20showstheresultsoftheregressionanalysisandfactoranalysisinMasrom sstudy, andillustratesthesupportfortherelationbetweenpuandbi,peouandatu,puandatu andpeouandatu.thecoefficientshowsthestrengthofthetheoreticalrelationbetween each factor, where a value closer to ±1 is strong and a value closer to zero is weak; the magnitudeshowsthedirectionofthecorrelationbetweenfactors(seefurtherdiscussionin 56

66 Section3.9.).ThestrongestinfluentialrelationshipisbetweenPEOUandPUwith= However, the relation between ATU and BI is not supported because of the result from Masrom sfactoranalysis.theresultsfromhisstudyshowhowthetamcanbeappliedin general,wheretherewasasmalllimitationontherelationbetweenatuandbi(masrom, 2007). Thereviewinthissectionhassummarisedthestrengthsandweaknessesofthetheoretical modelsthisthesiswantstoapplyinitsmethod.mostresearchstudiesinthissectionhave establishedresearchmodelsthattheyhaddevelopedfromabaseofthetam,explainedin Section to evaluate user acceptance of elearning technology. They used external factorsthatarelinkedtopuandpeou.ongetal.(2004)aswellaspituchandlee(2006)made furtherdevelopmentsbylinkingtoandmodifyingbehaviouralintentionandsystemusageof thetam.thisthesishassimilarobjectivesoffindingtheuseracceptanceofanelearning application, associated to the TTTEBM curriculum. Section 2.4 determines barriers to the systemandsection3.2.2showsthedevelopmentofthetamintoanetamthatisspecificto anebmtrainerusergroup. 57

67 2.4 ExternalFactorsfortheTAMModel Thissectionreviewspapersthatdevelopedtechnologyacceptancemodelsandcollects backgroundinformationaboutfactorsthatinfluenceauser sadoptionofaninnovation.the reviewisusefulfortheselectionofadditionalfactorsfordevelopingthetam(more informationontaminsection )intothisthesis model,theetam.selectionofthe technologyacceptancefactorsshouldberelatedtotamaswellasclinicalenvironments, EBM,ehealthandelearning(seeSection2.2).Externalfactorschosenaspartofdeveloping theetamincludeage,gender,timeconstraint,experience,educationlevel, OrganisationalSupportandQuality.Afteranalysingtheadditionalfactors,Section3.2.2 and3.3.2choosefactorsandhypotheticallinkstodeveloptheetam.thechoiceofthe factors,alsocalledpositivemoderatorsormoderatingfactors,allowsinvestigationoftheir moderatingeffectonanebmtrainer stechnologyacceptance Age Thereisawidevariationofagebetweenusersofasysteminthismoderntechnologically mindedsocietysuchas,computer,internetormobilephoneusers,whocanbefromasyoung asachildtoasenior.however,intermsofebmusersacceptingelearningapplications,the agerangefallsintoamoredistinctadultagedgroup.manyotherstudieshaveconsideredage asafactorthatinfluencesauser sadoption oftechnology,suchasfromvenkateshetal. (2003),Morrisetal.(2005),Luetal.(2006)andMorrisandVenkatesh(2000). MorrisandVenkatesh(2000),aswellasSunandZhang(2006),foundthatAgeisanimportant factorfortheadoptionoftechnologyandthatitmoderatestheuser sjudgment.theystate thatitisstrongerthanthemajorityofothermoderatingfactors,withtheexceptionofgender. 58

68 MorrisandVenkatesh s(2000)foundthatyoungerworkerswereaffectedbytheirattitudeto usingtechnologywhereasolderworkers,incontrastwereinfluencedbysubjectivenormand perceivedbehaviourcontroltousetechnology.justasmorrisandvenkatesh(2000)hadfound itimportantfortechnologyacceptancemodelling;venkateshetal.(2003)alsofounditas beingthemaininfluentialfactoroftheirmodel.theystatedthatyoungerusershavegreater behaviouralintentionstousetechnologywhenthereisaperformancebasedrewardandthis motivates their attitude towards using that system. However, older users felt it hard to process complicated interfaces and it would pose a barrier to their attention on the job, leadingtoincreasedeffortexpectancy.olderusershavebehaviouralintentionstowardusing technologydependingontheirworkingconditionsandavailablefacilities(venkatesh,etal., 2003);moreover,Luetal.(2006)foundthatolderwomenhavealargetendencytochoose whethertousetechnologydependingontheirjobenvironment. ThereisalinkbetweenAgeandothermoderatingfactors,suchasExperienceorGender;a more detailed discussion on Experience follows in Section Venkatesh et al. (2003), similartomorrisandvenkatesh(2000),statedthatolderusershaveagreaterdependenceon socialinfluenceandthisdecreaseswithgreaterexperience.thelinkofageandgenderisthat youngermenhavegreaterperformanceexpectancyandolderwomanhavegreatereffort expectancy(venkatesh,etal.,2003).morrisetal.(2005)andluetal.(2006)alsofoundthat Age and Gender are interrelated in their effect on adoption and use of technology; they suggestedthatfactorsofageandgendershouldberegardedinunison.thedifferenceof technologybasedusagedecisionsbetweengendersbecomesmoreprominentwithincreasing age;thecontraryhappenswithyoungerusersandleanstowardaunisexpattern(morris,et al.,2005). 59

69 IntermsofInternetusage,agehasanimpactonthedeterminantsofbehaviouralintention anddirectlyontheuser sbehaviouralintention(kripanont,2007).kripanont s(2007)review foundthatyoungusershadahigherprobabilitytousetheinternet.hesupportedtheworkof Venkatesh et al. (2003), stating age has an influence on attitude, subjective norm and perceivedbehaviouralcontrolofauseroftheinternet.kripanont sfindingsindicatedthat older users needed more attention to improve their Internet use by focusing on social influence and facilitating conditions, which supports Lu et al. s (2006) findings. Kripanont concluded by adding that older users could have an increase in behavioural intention by building selfefficacyviatraining forinternetuse Gender Thereareassumptionsthatmenandwomenmayhavedifferentperceptionsandmayuse informationandcommunicationsystems(ict)differently.manystudiesconsidergenderasa factor that has influenced the use of technology; these include Gefan and Straub (1997), SpeierandVenkatesh(2002),Venkateshetal.(2003),Venkateshetal.(2004),Schuler(1975) andvenkateshetal.(2000d).someofthesestudieshadresearchedthejointeffectofage andgender,asdiscussedinsection2.4.1.thissectionexplorestheeffectthatgenderhason users behaviouralintentionsandthedeterminantsofbehaviourintention. Venkatesh et al. (2000d) found that there are large gaps between male and female co workers technology acceptance in an organisation. Their findings supported Gefan and Straub s(1997)viewsbyhighlightingtheimportanceformanagerstoassesstheiremployees, especiallywithconsiderationtogender,sotheycouldimprovetheiremployees productivity in the workplace (Venkatesh, et al., 2000d). Venkatesh et al. (2000d) stated that the 60

70 employees appraisals, from a good productivity viewpoint, should be in relation to their gender,sincemenandwomenvalueperformancebenefitsdifferently.bothgefanandstraub (1997) and Venkatesh et al. (2000d) concluded that the satisfaction of women about a computerbasedsystem,suchaselearning,isaffectedbythepeouandthesatisfactionof menisaffectedbypu(venkatesh,etal.,2000d). Schuler (1975) found that women, compared to men, had a greater desire to work with colleagueswhowerewellmannered.incontrast,theyfoundmendesiretoinfluenceother peopleintheworkplaceandtobethedominantdecisionmaker.inarbaugh s(2002)review, hearguedabouttheperceptionmenhavewheninteractingwithcyberspace,statingtheyuse it for communication and educational purposes, to get lower costs and greater access worldwide.inagreementwithschuler(1975),arbaughstatedthatmenoftenusecyberspace inacompetitivebehaviourthatwouldeitherboosttheir ego ordegradeothers.thisalso backsupmenhavingagreaterintenttotechnologyusebasedonitspuasstatedbyvenkatesh etal.(2000d).arbaugharguedwomen sinteractionwithcyberspacediffersfrommen susage. He said women perceive cyberspace as a communication gateway, where each additional memberofapublicnetworkcancontributetothelearninglevelofagroupwhohavejoined anonlinenetwork,suchasablogorchatroom.healsomentionedthatwomenhaveagreater involvement in webbased learning class discussions compared to traditional classrooms (Arbaugh&Duray,2002) TimeConstraint Workplacesarenaturallytimeconstrainedenvironmentssuchasinaclinicalsetting,where theorganizationsetsascheduleforclinicians.however,despitetheschedule,trainersand 61

71 cliniciansmaybeabletofindsomefreetimewithintheirroutines.thisfreetime,however frequentorinfrequentitmaybe,createsaflexibilityoftimeinthetrainer sschedule.clinical EBMtrainerscoulduseapartoftheirfreetime,whenevertheyfindit,tostudyEBM.Hence, EBMstudymethodsneedtofitintoatrainer sschedule.inotherwords,becauseauser s routineistimedependent,tobecompatiblewithatrainer sschedule,ebmlearningneedsto betimeindependent. Elearningsystemsfeaturecontenttheusercanrepeatedlyaccessandviewatanytime.This timeindependencygivesusersanopportunitytostudyinawaynotfeaturedintraditional classroomlessons.theblendedlearningapproachforebmusershasthistimeindependency attributebyproviding24houraccess,however,howmuchusersperceiveitasbeneficialin theworkplaceneedsquantification. StudiesthatconsideredtheeffectofTimeConstraintsandtheuseofwebbasedlearning applications,suchasarbaughandduray(2002)andarbaugh(2004),modelledtheeffectofa user sconstraintoftimeontheirbehaviouralintentiontouseawebbasedcourse. Arbaugh and Duray (2002) combined the effect of a user s Time Constraint together with restraintsintheirlocationorplace.theythenmodelledhowthesetwobarriers,whichthey defined as Perceived Flexibility, change the user s perception of the benefit of an online course. They found that when participants of the online course thought it had fewer constrictionsontimeandplace,theywouldstrengthentheirperceptionsaboutthecourse contributingtotheirlearning.moreover,theyfoundusersperceivingthecoursewithfewer constraintsontimeandplacewouldbemoresatisfiedwiththewebbasedcourse. 62

72 Arbaugh(2004)studiedhowstudentsparticipatedwithmultipleonlinelearningcourses.He identifiedthedifferencebetweentraditionalclassroomandwebbasedcourses.insupportof ArbaughandDuray(2002),oneofthefactorsArbaughdiscoveredbeneficialinonlinelearning wasgreaterflexibilityoftimeandplace.thefactoroftimeconstraintwasoneimportant aspectinlearningonlineforusers(arbaugh,2004).arbaughandduray(2002)alsopointed outthestrongconnectionbetweenperceivedflexibilityandexperience,moreonexperience below Experience(KnowledgeandSkills) Theuseofelearningapplicationsonaregularbasisovertimebuildsexperiencewiththat system. Speeds at which users can learn something new through experience with the application,variesbetweenthelevelsofaptitudeoftheusers.ebmtrainersmaynaturally encounter challenges when interacting with the application, to which they would make a decisiontotackleormoveonfromthechallenge.theniftheuserdiscoversthedecisionmade wasright,theywouldpresumablyreusethesuccessfulmethod. Users should also be able to adapt their experience from one technology application to another.experiencegainedfromsimilarapplicationstotheblendedelearningebmcourse shouldbeabasisforuserstoreferbacktowhenusingtheecourse.asanexternalfactorto thetam,experiencecouldbeastrongpredictorofauser susefulnessperceptions,attitude andbehaviouralintentionstousetheelearningtechnology. ManyresearchershaveconsideredthefactorofExperienceasamoderatoroftechnology acceptance, such as from Taylor and Todd (1995), Szajna (1996), Venkatesh (2000c), Venkateshetal.(2000d),VenkateshandDavis(2000b),ArbaughandDuray(2002),Venkatesh 63

73 andspeier(2002),venkateshetal.(2003),venkateshetal.(2006),sunetal.(2006),jones (2008) and Saadé and Kira (2009). Section included a partial study of Experience combined with Age. This section reviews a few important studies on Experience as a moderator. SunandZhang(2006)consideredExperienceasamoderatingfactorthatinfluencesthePEOU. TheystatedthatExperienceisachievableandtheleveloftheuser sexperiencehasaneffect onthepuandbi.therefore,becauseofitsoverallinfluenceonbehaviouralintentionand behaviouralintention sdeterminants,theystatedthatexperiencehasadirecteffectonthe decisionsusersmakeabouttheiractualuseofthesystem(sun&zhang,2006). Venkatesh(2000c)discussedwhetherExperiencehasadirecteffectonPEOU.Inaddition,he studiedhowthelevelofexperiencehadaneffectontheperceptionofthesystem,forexample thecomparisonofearlystageexperiencewithawellgroundedexperience.heanalysedother researchstudiesonexperience,particularlyemphasisingthatthejudgmentofusersisbased onthreeaspectspastexperience,contextorbackgroundandstimulus EducationLevel Thefactorofauser seducationlevel,similartoexperience,contributestoauser sabilityto handletechnology,suchaselearningsystems.usersoftheblendedlearningapproachfor EBMstudywouldhavedifferenteducationandintelligencelevels.EducationlevelsofEBM users,particularlyebmtrainersintheclinicalenvironment,aremostlikelytohaveamedical sciencedegreefromundergraduatetopostdoctoratelevel.intelligencelevelsarealsolikely tovarybetweenusers,whichcouldhaveanimplicitrelationtotheuser seducationlevel,and couldhaveaneffectonhowquicklytheuseradoptsthetechnology.reasonswhythefactor 64

74 ofeducationlevelneedsassessmentinthisstudy scontextincludesaneedtoquantifythe effectofeducationontheuser sbehaviouralintentionstousetheelearningcourseandto seeifthelevelofeducationmediatesthateffect. Technology acceptance studies have rarely studied education for its use as a moderating factor(kripanont,2007).venkateshetal.(2000d)investigatedgenderandotherfactors,such astheeffecteducationhasontheadoptionofusingtechnologyintheworkplace.fromtheir review,theyfoundthateducationlevelisconsideredaslinkedtogenderasaninfluential moderator.however,educationonitsownhadaninsignificantinfluenceonthebehavioural intentionofuserstodotheaction(venkatesh,etal.,2000d).theysuggestedthatabetter indicatorofauser sintelligencewasneededratherthanlevelsoftheireducationtogivea clearerindicationofthismoderator sinfluence. ContrarytoVenkateshetal. s(2000d)review,recentstudies(abushanab,2011;mardikyan, etal.,2012)haveshownthesignificantinfluenceofeducationlevelasamoderatingfactor. AbuShanab(2011)foundthateducationinfluencedusers acceptanceofinternetbanking technologyasamoderator.mardikyan(2012)supportedthisbystatingitsuse,adoptionand integrationincreaseswithagainintheuser seducation. Clearly,thereisagreaterneedtorealizetheimpactsofauser seducationontheirbehavioural intentiontocarryoutavoluntaryaction.venkateshetal.(2000d)suggestedimprovingthe categorisationofeducationtoincludeintelligenceratings.othermoderatingfactorsshowa connection,i.e.ageandgenderwitheducationlevel,oninfluencingbehaviouralintention and behavioural intention s determinants. Perhaps the improvement of education as a moderatorcouldbebylookingintoitslinkswithotherfactors,suchasthoseofthetam. 65

75 2.4.6 OrganisationalStructure ThisthesisfocusesonmodellingtheuseracceptanceofanelearningapplicationthatEBM userswouldstudyintheclinicalsetting.aclinicalorganizationalstructureisunique,butalso hassimilaritiestomostbusinesses,suchasneedingjobroles,coordinationandmanagement. In the scope of EBM trainers, this includes trainertrainee, trainertrainer and trainer supervisorrelations.generally,theclinicalenvironmentisconsideredasaflatorhierarchical structure, which directly influence the trainerwith organisational aims set by supervisors, collaboration or competition between trainers and dependence or expectations from trainees. These features of the workplace could either reinforce or weaken a trainer s behaviouralintentionstousetheecourse(singh,etal.,2005).asanexternalfactortothe TAM,thisthesiscanweightheeffectsanorganisation sstructurehasonuserperceptions, attitudetouseandbehaviouralintention. Venkatesh et al. (2000d) reviewed other researchers who had studied organisational structuresthatindicateditspositiveinfluenceonbehaviouralintention.however,aftertheir study, they concluded that the factor of Organisational Structure has a weak effect on behaviouralintention. Singhetal.(2005)reviewedotherresearchpapersonorganisationalstructuresandrealised thatthefactorisabarriertoelearning.hesuggestedthatorganisationsshouldmovetowards astructurewithgreaterflexibility:flexibilitybeinganorganisation sabilitytoincorporatenew technologiesintothestructureofcompanies,institutionsorworkplaces.inaddition,aflexible structurecouldemphasiseastowhetheraninstitutioncanadoptnewtechnologiessuchase learning(singhetal.,2005). 66

76 2.4.7 Quality AnelearningsystemforEBMtrainers,justlikeotherapplications,shouldconsiderqualityas adesignfactorthatisflexiblearoundtheuser sneeds.somedeterminantsofqualityinthe contextofthisstudyincludehowtheuserinteractswiththeapplication,learnsorunderstands howtouseitaswellasensuringthereisminimalsystemdowntime,i.e.beingreliableor providingtroubleshootingguidelines.however,theextentofthesefeaturesandmeasuring theireffectontheuser sacceptanceoftheelearningsystemisneeded.thisthesisproposes thefactorofsystemqualityasadeterminantofthetam sbehaviouralconstructs. Systemqualityhasfrequentlybeenoneofthefactorsthatraisethelevelofusersatisfaction (Sun&Xiao,2006;Sun,etal.,2006;Venkatesh&Bala,2008).Itisalsoatopicofdiscussionin theareaofwebbasedresearch,likeelearning,whereahypotheticalconnectionhasbeen established between factors of the Knowledge Support and satisfaction of Technology Support KnowledgeSupport(EvidenceAvailability) Knowledge Support is one of the factors that represent the quality of the elearning applicationforebmtrainers.theuserneedstounderstandhowtousethesystemwithhelp fromtheapplicationitself,amanualorothersourceofinformation.presentingtheknowledge supportinafree,easilyaccessibleanduserfriendlywayshouldimproveauser sperceptions of the system s usefulness. Other research has recognised this factor as important in predicting a user s satisfaction. The significance of this factor has been proven by other research,suchasthatbysun(2006),whoconsideredknowledgesupportasintegraltothe 67

77 qualityofasystem.inaddition,sun(2006)showedthatknowledgesupporthadasignificant effectonusersatisfactioninwebbasedlearning. Venkatesh (2000c) and Venkatesh et al. (2003) studied the use of a factor called the PerceptionsofExternalControl,whichexplainsauser sassuranceinknowingthatthereare enoughresourcesfromanorganisationortechnicalcentretoprovidesupportforthesystem, such as availability of knowledge, resources, and opportunities required to perform the specificbehaviour.theyfoundtheuser sperceptionofavailableknowledgeresourceshada significanteffectonthepeou.thisfactorcouldbeinfluentialtoauser sadoptionofthee learningapplication,whichthisthesishasmodelledinsection TechnologySupport Technology Support can increase the trust EBM users have with the use of elearning applications,astherearefewersystemfailures.alsointheresultofafailure,thedowntime wouldbelimitedtohowquickthesupportcanrestorethesystem sfunctionality.thefactor oftechnologysupportshouldinfluenceusers perceptionsofusefulnessandeaseofuseof theelearningcoursebecausewhenthesystemisreliabletheuserbelievesthattheirworkis safeintheresultofasystemcrash.thisthesishasmodelledthisfactortoquantifyitseffect ontheuseracceptanceoftheelearningapplication. Inmanystudies,theinclusionoftechnologysupporthasshownthatithasasignificanteffect onthesatisfactionofusers.roca(2006)andsun(2006)showedthequalityoftechnology supportprovideddirectlyinfluencestheusersatisfactionandtheiradoptionoftechnology. Sun also mentioned that regarding user satisfaction, quality (technology and knowledge support)hadastrongereffectthanflexibility. 68

78 In another paper, Piccoli (2001) reported an improvement in user acceptance with the inclusionoftechnologysupportasafactorinassessinguseracceptance.thisfactor,asshown intheresultsfrompiccoli,showsthedirecteffectthatthequalityoftechnologysupporthas in a learning environment. Arbaugh (2004) also noted a strong connection between user satisfactionandsystemqualityininternetbasedlearningsystems.heconcludedthattheuser becomesmoresatisfiedfromhigherqualitysystemcharacteristicscomparedtosuperioruser characteristics.thestudiesfrompituch(2006)alsofoundoutthatexternalfactorshavean influence on the adoption of elearning systems. He emphasised that the quality of the contentwasanimportantfactor,whichalsohadaneffectontheleveloftheusersatisfaction. 2.5 Summary ThischapteridentifiedexternalfactorsaspreliminaryworkfordevelopingtheTAM,including arelevantreviewofresearchmodelsthatassessuseracceptanceoftechnologytoseekout thepotentialsandflawsofthetamforsupportingdevelopmentofthisthesis model,thee TAM.Thecriticalinformationcollectedinthisliteraturereviewformsthebasisofdeveloping thisthesis extendedtechnologyacceptancemodel,questionnairesandinformationthatcan beusedtoanalyseandvalidatetheresults. EBM was explained and studies were reviewed to identify the benefits and drawbacks associatedwithtrainingebminworkplaces.elearningwasintroducedandexplainedasan aidforteachingoftrainersofebm.theintroductionofthetttebmprojectshowedthatthere isaneedtoimprovetheebmtrainers acceptanceofusingtheelearningsystemtodevelop theirexperience,knowledgeandskillofusingebmintheworkplace. 69

79 Theliteraturereviewincludedpaperspublishedontheoriesofvariousmethodstopredict userperceptionsofusefulness,easeofuse,attitudeandbehaviouralintentions.someofthe papersinthecriticalreviewdevelopedthetamwithexternalfactors,whichdemonstrated animprovedexplanationofthevarianceofthetam sfactors. Thenextchapteristhemethodchapter,whichexplainseachstepofthisstudyandallthe specialconsiderationsforachievingtheaimandobjectivesoftttebmprojects. 70

80 3 METHOD 3.1 Understandingthecoreconceptsoftheproject Astheliteraturereviewstates,whenmedicalpractitionersuseEBMonaregularbasistheir performance on their duties are improved and they are able to improve their abilities to correctlydiagnose,maketherapeuticdecisionsandfollowupthehealthandwellbeingoftheir patients. Due to the active nature of EBM, clinicians need to learn and practice it in the workplace.ebmlearningneedsintegrationintothefulltimejobsofpractitioners,whichcan befacilitatedbyanonlinelearningsystem.moreover,thelackingconfidenceofebmtrainers in teaching EBM for a clinical environment needs improvement. These challenges led the TeachtheTrainersEBM(TTTEBM)projecttoemergeasasolutiontoimproveEBMpractice withanofficialeuropeanwidecurriculumforteacherstofollow.theprojectisusingane learningframework,whichthisthesisaimstoassessbydevelopingatheoreticalmodelthat fits,ormodels,theebmtrainersasusersandthereforefindsfactorsthatcanpredicttheir behaviouralintentiontousetheelearningebmapplication. ThisthesispredictshowEBMtrainersusetheTTTEBMelearningapplication.Themethod wasbydesigningaquestionnaireandevaluatingthefactorsofthetechnologyacceptance Model (TAM). This followed by developing a theoretical model based on the TAM, an establishedmodel,byintegratingadditionalfactorsthatcouldhaveadecisiveeffectonthe user sbehaviouralintention.thenewmodelcanevaluatetheebmtrainer sacceptanceofthe TTTEBMprojectfromaquestionnaire,ascarriedoutinChapter5,andtherebystateinsight inhowtoimprovetheelearningapplication. 71

81 The methodology of this thesis uses background information on modelling and literature review of factors, as discussed in Chapter 2, to develop the TAM into the extended TAM (etam),whichcanpredictebmtrainer sacceptanceofthetttebmelearningapplication. TheTTTEBMecourseusesablendedlearningapproachtoEBMtraining,whichcombinese learningintotheclinicalworkplaceforaeuropeanwideebmelearningcurriculumthettt EBMprojectwascoveredinSection2.2.4.ThefactorsreviewedinSection2.4arechosenfor the etam based on their significance of being a determinant of the TAM s factors and relevance of predicting the trainer s acceptance of EBM in a clinical environment. The additionalfactorsareconnectedwithhypotheses,withrespecttothetttebmecourse,to theconstructsofthetammodel. Two research models, the TAM and etam, are developed with hypotheses, where each hypothesis acts as a link between factors. Hypotheses allow evaluation of the influential relationships between factors, which is a quasiquantitative evaluation. In the etam, the chosen additional factors also have hypotheses that relate to the TAM s factors including PerceivedUsefulness(PU)andPerceivedEaseofUse(PEOU),AttitudeTowardsUse(ATU)and BehaviouralIntention(BI).Overall,theeTAMasanewmodelisdevelopedfromtheTAM, wherebothmodelsareassessedtoseeiftheycanpredictebmtrainers acceptanceofthe TTTEBMapplication. Thischaptercoversthedesignoftwoquestionnairesthatwereusedtocollectparticipant data,whichwasusedtoevaluatethetamandetam.thedesignofaquestionnaireshould collect,filterandcategoriseinformationfromusersorebmtrainers.followingareviewof questionnairedesign,section3.4discussesscalesizingmatters.thisstudyhasdesigneda 72

82 questionnairerelatedtoebmtrainers acceptanceoftheelearningcourse,whichwasused toevaluatethetam.theprocessrevealedthemostinfluentialfactorofthetamforthee course.adifferentquestionnaireisusedtovalidateandfindthemostimportantfactorofthe etam.thisstudyteststhereliabilityofthefactorsoftamandetamandthequestionsof thequestionnaires. Start ETAMModelDesign * HypothesisSelection DesignTAMquestionnaire ** Quantitativedata fromsuveys Testreliabilityofquestionnaires Factoranalysis Regression End *TAMmodelwasdesignedbyDavis (1986) **QuestionnairedesignedbyOude Rengerink(2011)wasusedtotesteTAM Figure21:Flowchartofthemethodologyshowingtheprocess Figure21showsaflowchartoftherelevantstepsneededtoproduceresultsfromtheTAM andtheetamforassessingtheuseracceptanceoftheapplication.inaddition,itshowsthe 73

83 validationprocessusedintheresultsfromthetamandtheetam.asshown,thisthesisuses astepbystepprocesstoitsmethodologyfromstarttoend.itisimportanttomentionthat theprocessshowninthediagramstartswiththetamandleadsintotheetamdevelopment asfollows: 1. AssessmentandvalidationofTAMforitscapabilityinpredictingtheuseracceptance oftheecourse. 2. TheTAMmodel,nowvalidfortheecourse,canbedevelopedintotheeTAMwith additionalfactors. 3. TheadditionalfactorsofeTAMarevalidatedandthemostinfluentialfactorisfound. 3.2 TechnologyAcceptanceModelling Technologyacceptancemodelsormodelsthatpredictauser sbehaviouralintentiontoward asystemtrytofindfactorsthatcaninfluenceausertoadopttechnology.thisthesisdevelops apredeterminedmodel,thetam,toassesshowuserscanadoptelearningtechnologythat teachesebmintheworkplace.somemodelsandfactorsrelevanttotheaimsofthisthesis werereviewedinsection2.3,wherethetamwasusedasabasisfordevelopingnewmodels. Other literature shows the TAM as a strong predictor of user intention, hence this thesis assumeittohavepredictivepowerinassessingtheebmtrainers acceptanceofthetttebm elearning course. In the following section, the TAM s factors are reclassified into terms suitableforthisproject.section3.2.2includestheadditionofexternalfactorstothetamthat arepotentialbarrierstothecourse.thismakesitthefirstresearchmodelfortheassessment oftheuseracceptanceofthetttebm sblendedlearningapproach. 74

84 3.2.1 TAM Section2.3discussedtheoreticalmodeldevelopmentsincludingtheTAMbyDavis(1986),its predecessorsandsuccessors.section2.3.1explainedthepurposeofthetam,whichisto model and predict relationships between system characteristics and user acceptance. The TAM model contains four main behavioural constructs, which are stated below in terms relativetothetttebmelearningcurriculumastheecourse: Perceivedusefulness(PU);howanEBMtrainervaluestheecourseashavingapositive effectonachievingorworkingtowardsusingandimplementingit Perceivedeaseofuse(PEOU);theassumedorexperiencedlevelofeaseorsimplicity anebmtrainerhaswhentakingpartintheecourse Attitude towards use (ATU); being the motivational feelings and interest an EBM trainerhastocarryoutandcontinuelearningfromtheecourse Behaviouralintention(BI);howanEBMtrainerhasdesiresandinspirationstomotivate themselvesinachievingtheirgoalsintheecourse OtherdefinitionsoftheTAM sconstructsaredetailedinsection andtheglossary, Section

85 Figure22:TheoriginalTAMconsideredforthisresearch Figure22showstheconnectionsbetweenthefourpartsoftheTAM.Section3.3introduces theseconnectionsashypotheses.chapter4coverstheanalysisoftheresultsfromthetam andrelativequestionnaires,whereitshowsthequalityofthemodelintermsofassessingebm trainers technologyacceptance ETAM TheeTAM,asintroducedinSection1.3,isapurposebuilttheoreticalmodelforelearning technologiesandthetttebmproject.theetam srootscomefromthetamandthisstudy has chosen factors from a wide researchbased survey on elearning and technology acceptance,seesection2.4.thisfoundasuitablesetofmoderatorsthathadanimpacton users behaviouralintentionandthefactorsrelatedtothebehaviouralintention.thisthesis, fortheetam sdevelopment,considersmoderatorsasfactorsbecausetheirbasisisformed fromtheoreticalresultsofotherresearchers.aswellasstandingforanextensiontothetam, the e intheetamstemsfromthecategoryofallsimilardesignations,suchaselearning,e health and EBM, which is similar to how Eysenbach (2001) described the e in ehealth. 76

86 Therefore,theeTAMmodelisdesignedforassessingusersofthosesubjectsandespecially forebmtrainersofthetttebmelearningcurriculum. In this section, we show the grouping of external factors and their relation to the TAM s factors,whichformsthebaseoftheetam.thetermexternalfactorsinthisstudyrepresent the new group of predictors or barriers for the etam. The chosen external factors are organisational barriers to the TTTEBM elearning project, which include Age, Gender, Experience,EducationLevel,OrganisationalStructure,TimeConstraint,KnowledgeSupport andtechnologysupport. Figure23showsthebasisofthe TAMwithablueboxrepresentingtheplacementofthe externalfactorsandhencethetam sextension. Figure23:ShowstheextendedTAM TheexternalfactorstotheeTAMhavethreedivisions.Thesearehumandimension,design dimensionandquality,seefigure24.eachuserhastheirownpersonalattributes,suchasage, Gender, Experience, and Education Level; therefore, these factors are part of the human dimension field, which is similar to how Piccoli (2001) and Sun (2006) defined them as 77

87 discussed in Section 2.4. The other classification of these factors could be individual characteristics(venkatesh&agarwal,2006),learnerdimensionsorhumaneffectiveness(sun, etal.,2006). Figure24:Showsgroupingforexternalfactors TheOrganisationalStructureconsistsoftheeffectsofmanagement,levelofhelpfromother trainers and demand of trainees to learn EBM coinciding with a system. Time Constraint includestheroutineofebmtrainers,wheretasksintheirworkplaceeachhaveanallotted amountoftime.theflexibilityofatrainer sroutineintheworkplace,asstatedinsection2.4.3 and 2.4.6, depends on their work s demand or restriction of time and place. This thesis assessedthetrainers acceptanceoftheecourse sapplicationintheworkplace,wherethe applicationneedsadesignthatblendsintotheirroutine.hence,thedesigndimensionfield includesorganisationalstructureandtimeconstraint. Similar to other elearning systems, TTTEBM elearning application needs technology, includingacomputerwithaninternetconnection,tofacilitatethedistancelearningprocess. Therefore,thequalityoftheapplicationincludesbothknowledgeandTechnologySupport; seefigure24,becausethesefactorstelltheuserhowtousethesystemandkeepthesystem 78

88 from crashing or having faults. The quality of the system is part of the design dimension becausedesignersoftheapplicationneedtoconsiderthesystemsquality. Figure25showstheeTAM,anewmodelthatincludestheexternalfactorsandtheTAM s factors of Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Attitude Toward Use (ATU) and Behavioural Intention (BI). Section3.3.2 describes the methodology behind the connection of these parts, where each arrow represents a hypothesis to describe their correlation. Figure25:ShowingexternalfactorsgroupingandthemethodofconnectiontotheTAM 79

89 The critical literature review in Section and showed Age and Gender to be interrelatedfactorsthathadastronginfluenceontheuser sacceptanceoftechnologyine learningandotherictapplications.however,thisthesishasomittedthesetwofactorsfrom themodelshowninfigure25.thiswasdecidedafterpreliminarytestingbecauseoftheir instability with low sample sizes from the questionnaire. In addition, the literature survey showedthattheyhadbeenthoroughlyanalysedinotherresearchandtheyhaveastrong moderatingeffecttogether.moreover,theireliminationallowsotherfactorstobestudiedin greater depth. In this study, there is no further analysis of the effect of Age and Gender. However,thisstudyconsidersthesefactorsasfurtherwork,needingmoreparticipantsfor assessmentofageandgender.section4.3and5.3discusstheextentoftheeffectsofsample sizeonthisstudy. Figure26showsthefinalconfigurationofthemodel.Thissectionhasshownhowtheincluded factorsrelatetothetam sfactors.asshowninchapter5,thisthesisusedtheetamtoassess EBMtrainers acceptanceofthetttebmecourseandapplicationbyusingparticipantdata fromaquestionnaire.theanalysisofthefactorswithdatafromthequestionnairediscovers themostinfluentialfactorofetamtowardadoptionoftheecourse sapplication,italso showstheexplainedvarianceofalladdedfactorsonthetam sconstructs.thenextsection showstheselectionofhypothesesthatpredictrelationshipsbetweenfactorsandtheebm trainers behaviouralintentiontowarduseoftttebm selearningapplication. 80

90 Figure26:Newillustrationofthegeneralisationofexternalfactors,whereAgeandGender havebeenomitted,andqualityisconsideredaspartofadesigndimension. 3.3 Hypotheses AsstatedinSection3.1thefactorsintheTAMandeTAMhaverelationsdependingontheir hypotheses.thissectionshowshowhypothesesweregenerated,testedandmeasuredtogive asignificancelevelwithpositiveornegativerelations.theliteraturereviewontheoretical modellingandexternalfactors,insection2.3and2.4,aswellasbackgroundknowledgeon theconceptsofelearning,ebmandtttebm,insection2.2,providesthebasisofgenerating thehypotheses. ThesehypothesesenabletestingofrelationshipsbetweenanEBMtrainer sconstructand theirbehaviouralintentiontothetttebmelearningapplication.theresultsandanalysis following in Chapters 4 and 5 identifies the strongest hypothetical connection and hence drawout the most significant factors, of TAM and etam respectively, in assessing EBM trainersofthetttebmproject.theprooforrejectionofhypothesesclarifieswhichconstruct thetttebmprojectshouldfocusontoimprovetheirecourse sapplication. 81

91 3.3.1 TAM TheTAMasshowninSection hasfiveconnectionsbetweenitsfactors,whichDavis (1986)hadestablished.ThisstudyhasgenerateditshypothesesfortheTAMmodelbasedon thefiveconnectionsofdavis model.testingandanalysisofthehypothesesbuildsonresearch ofthetamaswellasdrawsoutconstructsthatmotivateebmtrainerstousethesystem.for example,whenanebmtrainerconsiderstheecourse sapplicationaseasytouse,theyare moreenthusiasticaboutusingitandthereforewillhaveagreaterperformancewhileusing thesystem. Figure27showsthediagramofthehypothesesgeneratedbetweenPEOU,PU,ATU,andBI. Thearrowindicatesthedirectionofthehypothesis,e.g.forH1,PEOUwillhaveasignificant effectonpuandsoon.incontext,whereecoursemeanstheblendedelearningapproachto EBMstudy,thesemean: H1.EBMtrainerswhoperceivetheecourseaseasytousewillbelieveitisusefulforthem H2.EBMtrainerswhoperceivetheecourseasusefulwillhaveadrivenattitudetouseit H3.EBMtrainerswhoperceivetheecourseaseasytousewillhaveadrivenattitudeto useit H4.EBMtrainerswhohaveanattitudethatmotivatesthemtousetheecoursewillhave abehaviourintentiontothesystem H5.EBMtrainerswhoperceivetheecourseasusefulwillhaveabehaviourintentionto thesystem 82

92 Figure27:TheTechnologyAcceptanceModel(TAM),showingthehypothesesassumedfor thisresearch Thedirectionofahypothesisnotonlysignifiesarelationship,butalsoshowswhatvariablesa factordependson.theperformanceofadependentvariable(dv)isshowntobemoderated byanindependentvariable(iv).inthecaseofh2andh3,wherethedvisatu,whichhasivs ofpuandpeourespectively,thismeans anebmtrainer spointofvieworattitudetoward usingtheecoursedependsonwhethertheybelievethecourseisbeneficialortheythinkitis easy to use, or both. Table 1 shows the categorisation of IVs and DVs related to the hypothesesoffigure27. Hypothesis Dependentvariables(DV) Independentvariables(IV) H1 PerceivedUsefulness(PU) PerceivedEaseOfUse(PEOU) H2 AttitudeTowardsUse(ATU) PerceivedUsefulness(PU) H3 AttitudeTowardsUse(ATU) PerceivedEaseOfUse(PEOU) H4 BehaviourIntention(BI) AttitudeTowardsUse(ATU) H5 BehaviourIntention(BI) PerceivedUsefulness(PU) Table1:Alistofthehypothesesthatdistinguishbetweendependentandindependent variables ETAM TheeTAMmodel,asdescribedinSection3.2.2,hasexternalfactorsconnectingtothePUand PEOUfactorsoftheTAM.Eachconnectionrepresentsarelationshipbetweentwofactors,as studiedinsection2.4.thisstudygenerateshypothesestotestandanalyserelationshipsof 83

93 factorsintheetamandtoidentifythemostsignificantconstructofebmtrainers,asshown infigure28.thisstudyhasgenerateditshypothesesfortheetambasedonthefactorsthis thesisassumedwerepredictorsofauser sbehaviouralintention. ThefactorsdiscussedinSection2.4werefoundtoexplainthevarianceofPUorPEOUfor otherstudies;thisstudyhasalsohypothesizedtherelationsofthechosenexternalfactorsfor theetam.organisationalstructureandtimeconstraintarepartoftheclinicalenvironment andareimportanttoassessbecauseebmtrainingneedstobeanintegralintoebmtrainers routines,seesection1.2.theclinicalenvironmentcanalsoincludefacilitiesforknowledge andtechnologysupport,whichwhenmoderatedcouldchangetheebmtrainer sperception ofusingtheecourse application,seesection2.4.7.moreover,researchersrarelystudied EducationLevelasamoderatingfactor;Experienceisalmostaconversetoeducationandthey arebothmembersofhumandimension.hence,followingfigure25wehavegeneratedtwo additionalhypothesesconnectingeducationlevelandexperiencetoatuandbi.therefore, intotal,thisresearchprojectassumessixteenhypothesislinksfortheexternalfactorsthat connecttopeou,pu,atuandbiwhicharerelevanttothetttebmelearningapplication. Provingthehypothesesmayproduceresultsinsupportofotherresearch;ontheotherhand whendisproved,itwouldpresenttheopportunityforanindepthstudyofthecontradictory result.theresultsandanalysesoftheinfluenceoffactorsandproofofthehypothesesare showninchapter5. Figure28fromH1toH16showsthegeneratedhypotheseswithrespecttotheTTTEBMe learningcurriculumandblendedlearningapproach,alsodenotedastheecoursehere.inthe contextofthisthesis,theirgroupingisasfollows: 84

94 Set1. (H1,H3,H13,H14): An improvement in an EBM trainer s ability and knowhowofelearning,ebmtrainingsystemsorsimilarecourseswillincrease theiracceptanceoftheecourse Set2. (H2,H4,H15,H16):WhenEBMtrainersaregreaterinformedandtrained intheuseoftheonlinetrainingsystemsandebmteachingpractice,theywill haveanoverallgreateracceptanceoftheecourse Set3. (H5,H7):AnEBMtrainerwhoseclinicalenvironmentisflexibleinplace andisbetterequippedwithfacilitiesforelearningandebmwillfinditeasier tousetheecourseandwillhaveastrongerbeliefthatitsuseisadvantageous Set4. (H6,H8):AmoreflexibleworkingscheduleforEBMtrainerstousethe ecoursewillenhancetheirabilitytouseitandtheywillhavestrongerfeelings toitsvalue Set5. (H9,H10,H11,H12):WhenthequalityofsupportforICTandinformation resourcesoftheecourseincrease,ebmtrainerswillthinkitiseasiertouse andwillbelievethatitisbeneficialforthem The hypotheses in Figure 28 have directions as indicated by the arrows that shows the influenceofonefactoronanother. 85

95 Figure28:ExtendedTechnologyAcceptanceModel(eTAM),showingthehypotheses assumedforthispartofresearch Table2showshowIVsandDVsoftheeTAMmodelarecategorisedbasedontheexplanation ofivsanddvsfromsection

96 Hypothesis Dependentvariables (DV) Independentvariables(IV) H1 PerceivedUsefulness(PU) Experience(Ex) H2 PerceivedUsefulness(PU) EducationLevel(EL) H3 PerceivedEaseOfUse(PEOU) Experience(Ex) H4 PerceivedEaseOfUse(PEOU) EducationLevel(EL) H5 PerceivedUsefulness(PU) OrganisationalStructure(OS) H6 PerceivedUsefulness(PU) TimeConstraint(TC) H7 PerceivedEaseOfUse(PEOU) OrganisationalStructure(OS) H8 PerceivedEaseOfUse(PEOU) TimeConstraint(TC) H9 PerceivedUsefulness(PU) KnowledgeSupport(KS) H10 PerceivedUsefulness(PU) TechnologySupport(TS) H11 PerceivedEaseOfUse(PEOU) KnowledgeSupport(KS) H12 PerceivedEaseOfUse(PEOU) TechnologySupport(TS) H13 BehaviourIntention(BI) Experience(Ex) H14 AttitudeTowardsUse(ATU) Experience(Ex) H15 AttitudeTowardsUse(ATU) EducationLevel(EL) H16 BehaviourIntention(BI) EducationLevel(EL) Table2:Listofthehypothesesthatdistinguishbetweendependentandindependent variables NowtheinitialdevelopmentoftheeTAMiscomplete.Thissectionhasincludedhypotheses toconnecttheexternalfactorstothefactorsofthetam.ithasalsogeneratedhypothesesfor thetam.bothsetsofhypothesesarewithrespecttoadoptionofthetttebmelearning applicationandhaveapredicteddirection.chapter4includestheassumeddirectionsofthese hypotheses,whichthisthesistestsforthetamandfortheetaminchapter5. AmainobjectiveofthisthesisistodevelopandtesttheeTAMbecauseithasapotentialto improve the TTTEBM curriculum and blended learning approach. The etam assessed a questionnaire designed by Oude Rengerink (2011), which is a questionnaire that uses the Likert7scalerating,asexplainedinSection3.4.Themodelcanassesstheuseracceptanceby identifyingfactorsthathaveaninfluenceontheuser sadoptionoftheecourseanddrown 87

97 out irrelevant barriers. This thesis has analysed the results from that questionnaire, as discussedinchapter PropertiesoftheQuestionnaire Aquestionnaireinitsfinalformisasetofquestionsonamaintopicthatrequirespeopleto answer each question. The purpose of the questionnaire is to provide information to researchersasevidencetosupporttheirhypothesesandresearchclaims.whendesigninga questionnairethereshouldbeaprocedure(diem,2004).thereshouldbeameasurablescale fortheanswers,sotheparticipantscanaccuratelyquantifytheiranswer.thesedesignsteps areaimedatgivingreliableinformationsothehypothesescanbeansweredandmeasured. Toanswertheresearchaims,questionsaredesignedarounduserconstructs,hypothesesand aresearchtopic(diem,2004). EachfactoroftheTAMandeTAMhasagroupofcomponentsthatdescribeapartofthe factor, also called components. The components of factors have a similar meaning to the questions, which serves as their relationship. Table 3 lists components of the TAM s questionnaire and relations to TAM s factors. These components become the basis of questions in the questionnaire, which then frames the questions. Kim and Mueller(1978) recommended extracting at least three components per factor when designing a questionnaire.hence,theamountoffactorsandquestionsaredirectlyrelated,wheremore factorsmeansmorequestions. 88

98 Technology Acceptance Model Factors Perceived ease of use (PEOU) Perceived usefulness (PU) Behavioural intention (BI) Attitude towards use (ATT) Components Table3:Showingeachfactorandrelatedcomponentforeachfactorseparately Easy to use. Easy to learn. Easily understandable interaction. Finding information easily. Making user productive quickly. Enhance effectiveness in learning. Improve performance. Increase productivity at work. Useful. Plan for future use. Used on all occasions. Intended for use in case of need. Interacting with other knowledge holders. Favourable during use. Good idea. Positive tool. Pleasant experience. Scalemeasurementsofquestionshavedifferentformatsandresearchershavetoselectthe right type to gain better reliability in their results. They usually choose multiscale measurementsoversinglescalemeasurementbecauseoftheirgreaterabilitytoindicatethe user ssatisfactionlevel(davis,1986).inadditiontothat,researchersgenerallychooselikert scaleformatsthatlimittherangeofanswerstoafewverbalstatementsornumbersforeach question(dawes,2008).therangesinthelikertscaleareusuallyin5,7or10pointrating scales,however5and7pointratingscaleshaveagreaterlikelyhoodofgiving highermean scoresrelativetothehighestpossibleattainablescore,comparedtothatproducedfroma10 pointscale (Dawes,2008).Theanswersavailablefortheparticipantstoselectusuallyrange fromstronglydisagreetostronglyagree,butcanhavesimilarphrasestodistinguishpointsin 89

99 thescale.figure29showsthelikertscaleformatthatisatypeofthemultiscaleformat,as usedinthisthesis squestionnaireandthequestionnaireusedfromouderengerink(2011). Q1.IfindtheEBMTTTinterac tiveelearningmaterialeasytoaccess VeryUnlikely Unlikely Likely VeryLikely Figure29:ExampleoftheLikertscaleformatusedinthequestionnaires Likert scales help to control a large number of answers from questions, enabling their measurementasasingleconstruct;thismakesthequestionnairemorereliable.toelaborate, thismeansinsteadofusingwordsorotherformatsofanswers,theanswerisanumberwith aspecificmeaning. ShownintheAppendix,Section8.2and8.3,arethequestionnairesusedinthisstudythat gatheredinformationformodelsthathaveassessedebmtrainer sacceptanceofthetttebm ecourse, namely the TAM and the etam. The design of the TAM questionnaire was developedfromdavis (1986)questionnaire,whereastheeTAMquestionnairewasdesigned byouderengerink(2011).thesequestionnaireshaveamultiscalemeasurementwitha7 pointratingscaleformat,whichdavis(1989)anddecoster(2000)definedasbeingthemost reliableratingscale.thesetwoindividualquestionnaireshavebeenusedseparatelyandonly bytherelevantmodelinthisthesis.thisisbecausetheamountandtypeofquestionsineach questionnairearerelatedtothefactorsinthemodels.hence,eachmodel sanalysisofthe participant data of the relevant questionnaires will be different, but when the results are combinediteffectivelymeanstwoquestionnaireshavebeenanalysedbymodelsforassessing theebmtrainers technologyacceptanceoftheecourse. 90

100 This thesis initially structured the TAM questionnaire s questions on Davis (1986) questionnaire,whichconstructsalreadyfittedthetam sfactors.basedonthatthisstudythen developed the questions to suit the hypotheses of this study. A statistical analysis of the questionswithapretestindicatedwhetherthequestionsweresuitableforprovidingtheright type of information, from the questionnaire s participants, for the TAM. The pretest was carriedoutinareliabilitytestusingthecronbach salpha,discussedfurtherinsection3.5. AfterverifyingtheTAMquestionnaireasreliable,DrT.ArvanitiscarriedoutthesurveyonEBM trainers,whoansweredthetamquestionnaireinmarchapril2009.theparticipantswerea totalof56including34men:twointheirlatetwenties,nineinthirties,19inforties,andfour above50yearsold;and22women:fiveintheirlatetwenties,sixinthirties,seveninforties, andfourabove50yearsold.thenextsectioncoverstheetam squestionnaireindepth ETAM TheeTAMquestionnairehasasimilarstructuretotheTAMquestionnaire.Itusesamulti scale measurement, the Likert scale, as described in Section 3.4. Oude Rengerink (2011), designed the structure, developed the questions and conducted the survey. There are 23 questions in the questionnaire and this thesis has analysed them using the etam. One hundredandtwenty clinical EBM teachers from 12 countries including Belgium, Canada, Finland,Germany,Greece,Hungary,Italy,Netherlands,Poland,Switzerland,UnitedKingdom, andtheunitedstatescompletedthequestionnaireonline,oronpaper.thisstudyhasused thesameparticipantsfromthetamquestionnaire,being56clinicians. ThisthesishasgroupedthequestionsintofactorsthatrelatetotheeTAMmodel.Factorising thequestionsisastepindesigningthemodel(field,2009).wefoundthelistoftopicsinthe 91

101 questionnaireareattitude,availabletime,hospitalhierarchy,levelofunderstandingenglish, available resources, knowledge & skills of trainers and requirements for EBM teaching in curriculaoratworkplace.similaritiesbetweenthosetopicsandhypothesesofthefactors werefoundandtheirrelevancewasconsultedwithdrt.arvanitis. ExternalFactorswhichaddedtoTAM SubFactors Components < Age Male Gender Female PreviouslyusedofPC Workedwithdatabases Understandsthesystem Experience Determiningthesystem Definingthesystem Appraisingthesystem Applicabilityofusingthesystemtohelppatients Undergraduate EducationLevel Postgraduate Medicaleducationdegree Flat OrganisationStructure Pyramid Teacherstime TimeConstraint Learnerstime Resourceavailability Knowledge Lackofevidence Support Toomuchknowledge Databaseaccess Technology Internet Support Lackofassistance Table4:Showingeachfactorandrelatedcomponentforeachfactorseparately HumanDimension DesignDimension Quality Table4showsthefactorsandcomponentsoftheeTAM.Thisstudyhasdrownedouteach predictor (component) in the list of external factors from those questions. Shown in the Appendix,Section8.4,arethelistsofquestionsfromtheTTTEBMelearningcurriculum s 92

102 questionnaire that have been grouped into factors of Experience, Education Level, OrganisationalStructure,TimeConstraint,KnowledgeSupportandTechnologySupport. Thissectionhasdiscussedthatdesigningquestionnairesneedsalotofconsiderationtofituser constructs,hypothesesandfactors.thesestepsincreasethelikelihoodofobtainingreliable results.thenextsectiondiscussessomesignificantmethodologiesavailabletoexaminethe reliabilityofthequestionnaire. 3.5 TestingReliabilityofFactorswithintheQuestionnaire Whenmultiplecomponentsaremeasured,theymightproducedifferentresultsdependingon thetestingconditions.pretestingthequestionnaireorestimatingthereliabilityofthefactors inthequestionnaireisimportantforresearchers.theycanuseittodeterminetheabilityofa measure. For example, it is a good measure when it produces the same result from componentsunderdifferentsituations(field,2009).thisstudyhasalsousedpretestingof thetamandetamquestionnairesasasteptowardvalidatingthem,whichisdiscussedin Section4.2and5.2.Thisisaninitialstepinfindingthereasonbehindtheinfluentialfactorof thetamandetammodelsbecausetheresultsshouldbetrustable. Researchers often use pretesting as a prerequisite to the validation of research questionnaires.pretestingquestionnairesisasurveyonasmallselectionofparticipants.this processinvolvestheidentificationofquestionsthathavealowreliability,asshowninfield (2009),Kinnear&Gray(2008)andSchumacker&Lomax(2004).Thisthesisusespretesting toidentifyanyanomaliesorrestraintswithinthequestionnaires,whichisawaytogetbetter results. 93

103 Reliabilityofquestionsisastepthisthesistakestoverifytheaccuracyoftheresearchmodels assessmentsbecausethedatatheyareanalysingcomesfromthequestionnaires.thebest waytodefinereliabilityistoputitincontextofanexample,asfollows.firstly,agroupis testedusingdifferentmeasurementprocesses,i.e.givenaquestionnairetofillout,wherethe questionnaire has questions relevant to the factors of the model and the answer to each questionhasaweight(1to7,orstronglydisagreetostronglyagree).afterthat,itispossible to compare the variation between the results of each application of the measurement process.iftherewereasmallvariationbetweenmeasurements,thentheparticipantswould have produced similar results, which is a way of declaring the questionnaire s reliability (Rudner,etal.,2001). Nunnally(1978)definedreliabilityastheleveloftrustorprovenaccuracyoftheresultsfrom experiments.thetrustworthinessoftheresultsisproportionaltotheamountofrepetitions oftheexperimentwhereithasproducedthesamesetofresults.repeatingtheexperiment severaltimes,replicatestheresults.thisrepetitionwillrevealthenumberoferrorsthathave occurredinthemeasurementprocessandallowseacherrorintheprocesstobequantified (Nunnally,1978). TherearedifferentprocessestomeasurereliabilitysuchassplithalfreliabilityandCronbach s Alpha.Splithalfreliabilityisamethodthatdividesadatasetintotworandomparts.Ifthe source of the data is reliable, the results from the reliability will be the same or similar. However,thismethodofmeasuringreliabilityhasdifficultieswhenwantingtodividethedata intotwoseparatepartseachtimeindifferentways.cronbach salphasolvesthisproblemand isamorepopularmethod(hinton,etal.,2004;field,2009).cronbach salphawasinitially 94

104 introducedin1951,anditshowsthecorrelationbetweenthecomponents.manyresearchers haveuseditasareliabilityscale,suchasnunnally(1978),davis(1989)andhinton(2004). Afterareviewofmanyrelatedpapers(Davis,1989;Hinton,etal.,2004;Field,2009),itis conclusivethatthecronbach salphamethodisabetterwaytotestthequestionsofthetam andetamquestionnaires.theresultsofthecronbach salphacalculationcanbecategorised intolevelsofreliability: Cronbach salphawith0.6orhigherisanacceptablereliability Cronbach salphawith0.8orhigherisagoodreliability Cronbach salphawith0.95orhigherisaveryhighreliability ThereliabilityandCronbach salphaisalsodiscussedinthecontextofameasureofreliability in Section 3.6, whereas in Section 4.2 and Section 5.2, this thesis analyses the results of Cronbach salphaforthetamandetamrespectively. 3.6 Testingthereliabilityofthequestions Field(Field,2009)definedreliabilityas theabilityofameasuretoproduceconsistentresults when the same entities are measured under different conditions. So, in terms of a questionnaire s scale that measures a group of users, the amount of random error in the results shows the questionnaire s reliability. Reliability can be assessed using different methodologies as discussed in Section 3.5. This thesis assesses the TAM and etam questionnairesreliabilitytoshowthecredibilityoftheresultsfromthequestions.reliability alsohasadirecteffectonthecorrelationandregressionanalysis,whichthisthesisshowsin 95

105 Section4.4and5.4.Thereforetestingthereliabilityofthequestionsisimportantbecauseit explainsdeviationsinmeasurementsinthequestionnaire. ThisresearchusesCronbach salphaasamethodtomeasurethereliabilityofthequestions inthequestionnaires.cronbach salphaisareliabilityassessmentmethodandhasbeenused bymostresearcherssuchas,nunnally(1978),davis(1989)andhintonetal.(2004)because ofits accuracyasindicatedbyfield(davis,1986).nunnallysuggestedthatareliabilitylevel of 0.8 is enough for testing the reliability in simple examples. Davis, based on Nunnally s research,considered0.8asanacceptablelevelofreliabilityinhisanalysis.hedidthisbecause increasingthereliabilitylevelabove0.8isnotnecessaryasithasverylittleeffectonthe correlationlevel. ThevalueofCronbach salphaexplainstheinternalvalidity,beingthedegreeofcertaintythat ahypothesisbetweenfactorshastherightdirectionornot,andconsistencyofthequestions thatwereusedinthequestionnaire.inthisstudy,cronbach salpha,iscalculatedusingthe Scaleifitemdeleted method,whichisshowninsection4.2and5.2.thismethodassesses thequestionnaire sreliabilityunderconditionswhereaquestioncouldhavebeeneliminated. Accordingtothismethod,aquestionnaireandallitsquestionshaveagoodreliabilityafter meetingthreeconditions: 1. IfthevalueofthestandardisedCronbach salphaisgreaterthan Ifallthevaluesofthe Cronbach salphaifitemdeleted (CAIID)columnarearound thesamevalueasthestandardisedcronbach salpha 3. Ifthevaluesinthe CorrectedItemTotalCorrelation (CITC)columnarefrom0.2to 0.3orhigher 96

106 ConditionNo.1aboverelatestothereliabilitylevelsshowninSection3.5.Intheeventof conditionno.1beinglessthan0.8,buthigherthan0.6,thequestionnairehasanacceptable reliability, but not a good one. In that case, the standardised Cronbach s Alpha may be increasedbylookingatconditionno.2.deletingthecaiidvaluesthatarehigherthanthe standardised Cronbach s Alpha increases the standardised Cronbach s Alpha to the higher valueofthedeletedcaiid.whenconditionno.3isnotmet,theitem/questionbelow0.2or 0.3maynotbemeasuringwhatotheritemsaremeasuring(Field,2009),whichmeansits resultisnotconsistentwiththeotheritemsanditshouldbedeletedtoincreasethereliability. Whenallthreeconditionshavebeensatisfied,thereliabilityofeachquestionandthewhole questionnairehasbeenshowntobegood. Thenextsectioncoversfactoranalysis.Factoranalysisisthefirststatisticalanalysismethod thisstudyappliestothedataofthequestionnaires. 3.7 FactorAnalysis Assessing data from questionnaires with the TAM and etam should be able to prove or disprovethehypothesesanddrawouttheinfluentialfactortoauser sbehaviouralintention. ThisthesisusedSPSStocalculatethefactoranalysis,whichshowstherelationshipsbetween componentsandfactorstochoosethebestcomponentforeachfactor.factoranalysisisa methodthatidentifiesfactors,orcorrelatescomponentswithfactors.factoranalysisisthe statisticalanalysistechniqueusedtoreducethedatasetintoafeweramountofcomponents (Dancey&Reidy,2002).Mulaik(1986)definedfactoranalysisasamethodthatisusedto condenseinformation,givingtheresearcheragreaterabilityindefiningfactors.thisstudyhas usedtheprincipalaxisfactoringextractionmethodinspsstoreducethedata. 97

107 TherearethreemainareasreportedbyPedhazur(1991)asbeneficialforusingfactoranalysis, asfollows: Whenthereisnotenoughinformationabouttherelationshipofacomponent When researchers are not sure about the results from other authors questionnaires Whenthereisaproblemwiththescaleofthesampling,sizeofthesample,the numberoffactorstakingthemethodofextractionortherotationoffactors. Themethodoffactoranalysisfindsoutwhethertherightcomponenthasbeenchosenfor eachfactorandwhetherthefactorshouldbeloadedornot(spearman,1904;thurstone, 1947;Guilford,etal.,1954).ItalsovalidateseachfactorthatisusedintheTAMandeTAM models. Tocarryoutfactoranalysis,thisstudyhasusedcorrelationtofindouttheconnectionbetween thecomponentsandfactorsinthetamandetammodels.thecalculationofthecorrelation forcomponentsisshownbelow: whereisthenumberoftheitems.thefactoranalysisofafactorcanberepresentedas (Field,2009): Whereb n representthefactorloadingand i representerror. 98

108 After,thisthesishasidentifiedwhichfactorscanbeindicatedasvalidforthemodel,thenext logicalstepinthemethodology,beingsimilartothatofbydancey&reidy(2002),isregression analysisandhypothesestesting. 3.8 Regressionanalysis ThisthesisusesaregressionanalysisonthequestionnairestoshowhowfactorsoftheTAM and etam complement each other in influencing an EBM trainer s behavioural intention towardtheecourse.forinstance,saytheregressionanalysisfindsarelationbetweenthe user sexperienceandtheirtimeconstraints,whichdirectlyrelatestotheirhighertestscore at the end of the ecourse s module. For this instance, we can assess the outcome by comparinganindividual stimeconstraintswiththeirexperience. Regression is explained as a statistical analysis that first discovers whether there is a relationshipbetweenthevariablesandsecondlyfindsthebestcorrelationbetweendvs,or outcomes,andivs,orpredictors,asshowninhintonetal.(2004).hintonetal.(2004)stated correlationwastheanalysisofhowdatafromtwovariablesinfluenceeachother.thechange ofdatafromonevariablecouldaffecttheothervariable.itisinterestingtoseethevariation thatonevariablehasontheotherbecauseitshowsthedominantfactor. There are two main types of regression analysislinear regression and multiregression. Regressionisamethodthatalsofindsthebestfitforadataset.Thebestfitofthedatacan berepresentedasaline.alineisplottedbetweenthedatapointsbyconsideringthesmallest amountofvariancebetweeneachdataelement.threeexamplesofthefittedlineareshown belowinfigure30(pedhazur,etal.,1991;sapsford,etal.,2006). 99

109 Figure30:Theexamplesofthefitline(Sapsford,etal.,2006) Agoodfithasaregressionwithin±1,howeveravalueclosertozeromeansthereisaweaker relation or poor fit. A good fit is defined by the correlation coefficient. The correlation coefficientcanbeestimatedbyconsideringthedistancebetweenthedatapointclustersand thefittedline.figure31showsanexampleofthecorrelationcoefficient(rvalue)foradata set(pedhazur,etal.,1991;sapsford,etal.,2006). Figure31:Anexampleofthecorrelationcoefficient(rvalue)(Sapsford,etal.,2006) Theregressionanalysisshowstheassociationbetweenthehypothesesofaparticularmodel. Moreover,italsoidentifiesthevalidityoftherelations,orhypotheses,betweentheDVsand 100

110 IVs.ThisthesiscarriesoutregressionanalysisinSection4.4andSection5.4toidentifythe causalrelationshipinthetamandetammodels. 3.9 Regressionandhypothesestesting TheTAMandeTAMmodelsneedvalidationtoensurethataconsistentlevelismaintainedso itproducesthesameresultingtheoreticalassessmentfortopicsrelatedtoelearning.validity is defined as the degree to which a measurement (i.e. TAM or etam models and questionnaires) can achieve the standard levels in which it was intended to measure, as presentedinthestatisticalbookbyfield(2009).regressionisthevalidationmethodchosen becauseithasbeenshowntoprovethesignificancelevelofthehypothesis,whichthisthesis considersasagoodwaytoprovethatthemodelhasbeendesignedcorrectly(field,2009). This research used simple and multiregression methods to calculate the standardised regressioncoefficient(),whichshowstheimpactoffactorsoneachother. indicatesthe strengthofrelationshipbetweenagivenpredictorandoutputinstandardisedform (Field, 2009).Thiscoefficientshouldbenonzeroandvarybetween±1.Thecloseritgetsto+1or1, thestrongerrelationoreffectthereisbetweenthefactors.ifislessthanzeroitshowsthe connectionassumedisinthewrongdirection(miller,etal.,2002;muijs,2004;field,2009). For example, while referring to adoption of the TTTEBM elearning application, if the hypothesis says PEOU will have a significant effect on PU, which results in value of equalling0.87.thatresultmeansthepuhadasignificanteffectonthepeou(miller,etal., 2002;Muijs,2004;Field,2009). Multipleregressionisastatisticaltechniquethatenablesastudyoftherelationbetweenthe DVwithtwoormoreIVs(Davis,1986;Sapsford,etal.,2006).Byusingtheregressionanalysis, 101

111 itispossibletofindoutthesignificancelevelofthehypothesis.theconfidenceorsignificance levelofthehypothesisiscalledthepvalue. ThePvalueindicatesconfidenceinthevalidityofahypothesis.Thestandardcutoffvalues forthepvalueare: Pvalue<0.05meanslessthan5in100chanceoftheerrorinthesignificancelevel ofthehypothesis Pvalue<0.01meanslessthan1in100chanceoftheerrorinthesignificancelevel ofthehypothesis Pvalue<0.001meanslessthan1in1000chanceoftheerrorinthesignificance levelofthehypothesis InSections4.4and5.4,thisthesisexplainstheresultsoftheregressionanalysisforTAMand etamrespectively Summary ThischapterhasexplainedhowtheeTAM,anextensiontotheTAM,wasdevelopedwiththe additionofresearchbasedexternalfactors.hypothesesweregeneratedtoformtheetam withlinksfromtheexternalfactorstothebehaviouralconstructsofthetam.thetamwas alsoimplementedwithhypotheticallinksbetweenitsfactors.thischapteralsoexploredhow todevelopquestionnairesbasedonthetamandetammodels factorsanditspreliminary testing of reliability. The design of two questionnaires was discusses; stating that the questionsneedtofittheusergroupunderassessment,categorisetheanswersusingalikert scaleandprovidereliableresultstoanswerhypothesesofthetamandetam.reliabilityof 102

112 thechosenfactorsandthequestionsinthequestionnaireswereexploredwithastudyof methodsincludingfactor,regressionandreliabilityanalyses.altogether,thismethodologyhas establishedanewtheoreticalmodel,theetam,aswellasdrawnuphypothesesforboth models assessmentofauser sacceptanceofthetttebm secourseandblendedlearning approach.chapter4and5analysetheresultsofthetamandetamrespectivelytofindthe influentialfactors,drownoutbarriersanddiscusstheimpactoftheresultsontheecourse s application. 103

113 4 TAMRESULTS 4.1 Introduction Thethesishassofardiscussedthebenefitofintegratingonlinelearningintotheworkplace and improving an Evidence Based Medicine (EBM) trainer s efficacy to teach medical practitionershowtouseebminthepatient sdiagnosis,treatmentandfollowupstages.the TeachtheTrainersEBM(TTTEBM)projecthasdesignedanelearningapplicationtotrainEBM study in a workplace and this thesis has proposed a set of hypotheses and developed a questionnaire for the application s assessment using the Technology Acceptance Model (TAM),asintroducedinSection3.2.1and3.3.1.Theresultsandanalysisofthischapterhas foundthetam smostinfluentialfactoratpredictingtheebmtrainer suseracceptanceofthe TTTEBMelearningapplication.Thischapterhasputhypothesesbetweenfactorsintothe contextofadoptionoftheonlinelearningsystemtodiscoverthemostinfluentialfactorand thenvisualisewhereimprovementsarepossibleonthetttebmproject.however,theactual improvementofthetttebmprojectisoutofthescopeofthisthesis. Thissectioninvolvesanindepthanalysisanddiscussionoftheresultsfromthequestionnaire thathasbeenanalysedusingthetammodel.theresultsoftheanalysisarecategorisedinto threeparts,reliability,factoranalysisandregression,wheresection3.5to3.9coveredtheir methodologies. The reliability section shows the calculation of Cronbach s Alpha for each questionfromthequestionnaire,andasanaveragevalueforthewholequestionnaire.the factoranalysissectionincludesastudyofthechosenfactorsthatwereusedforvalidationand theirapparentconnectiontotherelatedgroupofquestions.theregressionsectionshowsthe 104

114 impactofeachfactoronthewholeofthesystem,includingtheconnectionsbetweeneach partofthetammodel. 4.2 Reliability Table5showsthetotalandstandardisedaveragesofCronbach salphaandtable6shows Cronbach salphaforeachquestionandthecorrelationvalues.thestandardisedcronbach s Alphahasbeencalculatedgivingaresultof0.961,asexplainedinSection3.5,thismeansthat thetamquestionnairehasaveryhighreliability.therefore,thequestionnairehasbeenwell designedforitsuseinthisstudyandthetamhastheappropriatefactorsforassessingthe questions. Tolookattheresultsmoreclosely,Table6showsthevaluesofCronbach salphaforeach questionfromthequestionnaire,whichthisstudycalculatedwiththe Scaleifitemdeleted method.asillustratedinfigure32,allbutoneofthevaluesfromthecaiidcolumnarelower thanthestandardisedcronbach salpha(redhorizontalline).theoneaboveispeou1with 0.963,whichis0.002abovethealpha.DeletingPEOUwouldincreasethealpha,butthiswould have been insignificant in terms of the classification of the questionnaire s reliability. The lowestisbi1with0.956,whichmakesitthemostreliableitem.thevaluesfromthecitc column,illustratedinfigure33,areallgreaterthan0.3,wherethelowestispeou1with0.412 andhighestisbi1with0.86.peou1isindicatedinthecitccolumnasmeasuringthesameas theotherfactors;however,ithasaslightdifferentmeasurethantherest.overall,theresults of Table 6 meet the three stated conditions in Section 3.5 for the results to be reliable; therefore,thisstudyregardsthisquestionnaireandallitsquestionsincludingthefactorsof thetamashavingaveryhighreliability. 105

115 Questions ScaleMeanif ScaleVariance ItemDeleted ifitemdeleted Numberofquestions 20 TotalCronbach salpha StandardisedCronbach salpha Table5:Averagereliabilityanalysis CorrectedItem TotalCorrelation (CITC) Squared Multiple Correlation Cronbach'sAlpha ifitemdeleted (CAIID) PEOU PEOU PEOU PEOU PEOU PU PU PU PU PU ATU ATU ATU ATU ATU BI BI BI BI BI Table6:Reliabilityanalysisindetailforeachquestion 106

116 Cronbach'sAlphaifItemDeleted(CAIID) Figure32:Cronbach'sAlphaifItemDeleted(CAIID)withredhorizontallineshowingthe standardisedcronbach salpha CorrectedItemTotalCorrelation(CITC) Figure33:CorrectedItemTotalCorrelation(CITC)withyaxissettominimumrecommended value Thenextstepistousefactoranalysis,whichhastheabilitytoshowthenumberoffactorsthat canfitthedatasetwiththeircomponents. 4.3 Factoranalysis Factor analysis is the assessment of a factor with what it depends on. This analysis can determine the influence of known or unknown components of a factor and examine the 107

117 hypothesesofthefactors;seesection3.7formoreonthemethodologyoffactoranalysis.the link between each question/component and the factors were defined as hypotheses in Section3.3.1.Hypothesesaresetupsothefactoranalysiscandeterminetheinfluentialfactor ofthetammodelforeachcomponentinthequestionnaire. This section validates the number of components from the 20 questions in the TAM questionnairearefourfactorsofthemodel,wheretheresultsareshownintable7.factor analysiswascarriedoutusingstatisticssoftwarecalledspss,whichrecommendsacutoff varianceaboveorequalto0.5todrawoutcomponentswithsignificantvariance.thisstudy haschosentoassesscomponentswiththeprincipalaxisfactoringextractionmethod,which isintegraltothestatisticssoftware. Illustrated in Figure 34, Factors 1 to 4 account for a large amount of variance, and their componentsaregreaterthan0.5intheirgroupsofpu{2,3,4,5},bi{1,2,3,4,5},peou{1,2,3,4} andatu{2,3}respectively.thetableshowsthepercentagevarianceexplainedafterrotation forfactors1to4is26%,19%,18%and13%respectively,whichexplainsthreequartersofthe totalexplainedvarianceinuseracceptanceoftheapplication.factor1,puhasahighvariance andfactors2to4haveamediumvarianceintheircomponents,whichmeansthattheyare valid.factor4,withapercentagevarianceof13%isonlypartlyvalidatedbyatu{2,3},because ithastwocomponentsabove0.5,whichmakesiteligiblefordeletion,howeverforresearch purposesandbecauseofitsvariancethisthesisconsidersitasbeingsupported. 108

118 Question Factor1 Factor2 Factor3 Factor4 PEOU PEOU PEOU PEOU PEOU PU PU PU PU PU ATU ATU ATU ATU ATU BI BI BI BI BI PercentageVariance 26.0% 18.6% 17.6% 13.1% Table7:RotatedFactorMatrixofTAMrotatedwithVarimaxwithKaiserNormalizationand lessthan0.5cutofftodrawoutfactors(darkgreyshade),0.3cutoff(lightgreyshade) 109

119 Factor Factor2 PEOU1 PEOU2 PEOU3 PEOU4 PEOU5 PU1 PU2 PU3 PU4 PU5 ATU1 ATU2 ATU3 ATU4 ATU5 BI1 BI2 BI3 BI4 BI5 PEOU1 PEOU2 PEOU3 PEOU4 PEOU5 PU1 PU2 PU3 PU4 PU5 ATU1 ATU2 ATU3 ATU4 ATU5 BI1 BI2 BI3 BI4 BI Factor3 PEOU1 PEOU2 PEOU3 PEOU4 PEOU5 PU1 PU2 PU3 PU4 PU5 ATU1 ATU2 ATU3 ATU4 ATU5 BI1 BI2 BI3 BI4 BI5 Figure34:FactorloadingforeachFactor1to4fromtoplefttobottomright Factor4 PEOU1 PEOU2 PEOU3 PEOU4 PEOU5 PU1 PU2 PU3 PU4 PU5 ATU1 ATU2 ATU3 ATU4 ATU5 BI1 BI2 BI3 BI4 BI5 AsstatedintheSPSSguidebook,factorsvalidatewhengreaterthanthreecomponentsofone grouphaveavariancegreaterthan0.5.accordingtospssguidelines,factor4hasaweak validity,whichmightbeduetothequestionnaire ssamplesize.however,0.5asacutoff variance is not obligatory to choose. Other literature has defined the cutoff variance as insignificantbelow0.3(pahnila,2006;child,2006),whereafactorloadingabove0.3ishigh andabove0.6isveryhigh(burgess,2006).the0.3cutoffvariancewouldvalidatefactor4 withatu{1,2,3,4}andtheotherthreefactorswouldnotbeaffected,shownintable7with lightgreyshade.forthepurposesofresearch,atuisloadedasafactorbasedonitsmedium varianceandthisthesiscontinuedwithregressionanalysisofatualongwiththeotherfactors. 4.4 Regression AspreviouslydescribedinSection3.9,multiregressionanalysisinvolvesthecomparisonof connections between a dependent variable and two or more independent variables. For example,anebmtrainer sbehaviouralintentiontowardusingtheecoursedependsonhow muchtheybelieveittohelpthemandtheirattitudetousingthesystem.table8showsthe detailofmultipleregressionanalysisdoneon theresultsfromthetamquestionnaire.as 110

120 shownintheresults,allhypothesesaretruebecausetheestimatedpvalueislessthan0.05 and the regression coefficient () is between 0 and 1. The hypotheses were defined in Section Hypotheses H1 H2 H3 H4 H5 Dependent variables(dv) Independent variables(iv) Standardised regression coefficient() Significantlevel (pvalue) Perceived PerceivedEaseOf Usefulness(PU) Use(PEOU) AttitudeTowards PerceivedUsefulness Use(ATU) (PU) AttitudeTowards PerceivedEaseOf Use(ATU) Use(PEOU) Behaviour AttitudeTowards Intention(BI) Use(ATU) Behaviour PerceivedUsefulness Intention(BI) (PU) Table8:MultipleregressionresultofTechnologyAcceptanceModel Figure35showstheclearconnectionbetweenthefactorsthathadbeenconsideredinthis researchproject.asshown,thestrongestisatutobiwith0.82andtheweakestispeouto PUwith0.65.ThismeansthattheEBMtrainer sapproachorattitudetousingthetttebme learning application has the greatest influence on their intention to use the system. In addition, increasing their perception of the system s usefulness for improving their job performanceencouragesthemtouseitmorecomparedtohoweasytheyperceiveitsuse. 111

121 Figure35:TheTechnologyAcceptanceModelshowstheweightofeachfactoroneachother withacoefficient ThetestoftheregressionhasresultedinallthehypothesesofSection3.3.1beingprovenas true for EBM trainers using the TTTEBM elearning application. The results support the findingsofdavis(1986),whooriginallyfoundedthetam,byprovingtheuser speouandpu haveaninfluenceonthedeterminantsofbehaviouralintentionandpuhasadirecteffecton theuser sintentiontousetechnology.theresultsofthefactoranalysisloadedallfactors, includingatuasafactordespiteneedingalowercutoffvarianceof0.3. The regression results have shown there is a significant link between the EBM trainer s attitudeandtheirbehaviouralintentiontotheecourseandtheirbeliefsofitsbenefitsfor theirworkisamoderatoroftheirattitude.therefore,theresultsareindicatingthatthettt EBMprojectshouldfocusontheusefulnessoftheecourseandplaceemphasisonmotivating andinfluencingtheebmtrainer sattitudetousingtheapplication. 4.5 Discussion ThischapterhasevaluatedtheTAM sabilitytoexplaintheacceptanceofthetttebme learningcurriculumandblendedlearningapproachbasedondatacollectedfrom56clinical 112