motortokyohotn0744

M.
A.
Wimmeretal.
(Eds.
):EGOV2010,LNCS6228,pp.
275–288,2010.
IFIPInternationalFederationforInformationProcessing2010WhatIstheIssuewithInternetAcceptanceamongElderlyCitizensTheoryDevelopmentandPolicyRecommendationsforInclusiveE-Government*BjoernNiehavesandRalfPlattfautWestflischeWilhelms-UniversittMünster,EuropeanResearchCenterforInformationSystems,Leonardo-Campus3,48149Münster{bjoern.
niehaves,ralf.
plattfaut}@ercis.
uni-muenster.
deAbstract.
Digitaldivideisstillabigtopicininformationsystemsande-governmentresearch.
Inthepast,severaltracksandworkshopsonthistopicexisted.
Asinformationtechnologyandespeciallytheinternetbecomemoreandmoreimportantgovernmentscannotignorethefactthatelderlycitizensareexcludedfromthebenefitsrelatedtointernetusage.
Althoughe-Inclusionprogrammesandinitiativeschangedovertheyearsand,moreover,althoughtheamountofe-Inclusionliteratureisconstantlygrowing,thereisstillnothoroughunderstandingofpotentialfactorsinfluencingprivateinternetusage.
Hence,inthisstudyweidentifyimportantinfluencingfactorsbasedontheliteratureontechnologyacceptanceanddigitaldivide.
Wedevelopamodelbasedonthesefactorsandtestitagainstcomprehensivesurveydata(n=192).
Ourtheoreticalmodelisabletoexplainmorethan70%ofthevariationinprivateinternetusage.
Wederivepolicyrecommendationsbasedontheresultsanddiscussimplicationsforfutureresearch.
Keywords:DigitalDivide,e-Inclusion,UTAUT,QuantitativeStudy.
1IntroductionToday'swesternsocietiesfacetwocommontrends:First,today'ssocietiesaroundtheworldtendto"age"or"grey"[26].
Theshareofpopulationolderthan65yearsis15.
9%andwillriseupto25.
9%by2050.
Second,theimportanceofinformation,informationprocessing,andcommunicationisconstantlygrowing.
Thisphenomenonhasbeencondensedtotheterminformationsociety[34,15].
Societalagingbearsseveralrisksforaninformationsociety.
Ontheonehand,anincreasingshareofelderlycitizensresultsinproblemsforlocalgovernmentssuchasfiscalstressandincreasingexpenditureonhealthcareorpensions[19].
Ontheotherhand,largepartsofthepopulationareexcludedfromtheinformationsociety.
Theyneitherhaveaccessnorskillstousemodernmedialiketheinternet.
Adigitaldivideamongon-linersandnon-linersexists[22].
Especiallyseniorcitizensareoftenexcludedfrommoderntechnology[6,4].
*TheauthorsaregratefultothefinancialsupportofthisresearchbytheVolkswagen-Foundation.
276B.
NiehavesandR.
PlattfautHowever,governmentswanttomakeuseofthegrowingimportanceofICT.
EspeciallylocalauthoritiescanenhancetheeffectivenessandefficiencyoftheirprocessesandorganisationalstructureusingICTand,bythis,levertheirproductivitytoanewlevel(electronicortransformationgovernment[31]).
Moreover,governmentagenciescanprovidetheirservices"online"andsupportthembymeansofICT.
However,inadigitallydividedworldthenon-linersareexcludedfromthebenefitsofICTsupportedgovernmentalservices.
TheEuropeanUnionrecognisedboththeimportanceofICTandtheexistenceofadigitaldivide.
Therefore,theministersofthememberstatesoftheEUcalledforaninclusiveinformationsocietyanddeclaredtofocusonmultiplegoalstoreachthisaim[21].
ThiswasalsocapturedbythecabinetofficeoftheUnitedKingdomwhichcalledfortackling"overallissuesofdigitalinclusion"[9]andworks"towardsachievingequitableaccesstonewtechnologyandremovethebarrierstotake-up"[10].
Bothdefineelectronicinclusion(e-inclusion)asanintegralpartof(especiallylocal)governmentalpolicies.
Projectstobridgethedigitaldividehavealonghistory.
Firstgenerationprojectsincludedgrantstoprovidemoreseniorcitizenswithcomputers[16],freeinternetaccessatlocallibrariesorcomparablecentres,aswellasinternetcoursesspeciallydesignedforelderlypeople[32].
However,technologyacceptanceresearchsuggestsseveralotherbarriersthatcouldbetackledbygovernmentale-inclusionprojects.
TheUnifiedTheoryofAcceptanceandUseofTechnology(UTAUT)suggeststhatnexttoEffortExpectancy,whichistackledbyinternetcourses,andFacilitatingConditions,whichare(amongothers)establishedthroughtheprovisionofaccess,PerformanceExpectancyandthesocialmilieuplayanimportantroleinexplainingusagebehaviour.
Hence,itisdoubtablewhetherthemereprovisionofcomputercoursesorfreeinternetaccessaresufficienttoreachaninclusiveinformationsociety.
Moreover,thereisthepossibilitythatthegroupofnon-linersisfragmentedandthatdifferentmeasuresshouldbeestablishedfordifferentgroups.
Hence,thisstudyaimsatclarifyingthefollowingresearchquestions:RQ1Howcanweexplaintheprivateinternetusageandnon-usageofseniorcitizensRQ2Whatareimportantfactorsforseniorcitizens'usageandnon-usageoftheinternetRQ3DoesanextensionofUTAUTusingmoremoderatingvariablesfromthedigitaldivideliteratureprovideabenefitinexplainingprivateinternetusageamongtheelderlyRQ4Whatcanpractitionerslearnfromamorecomprehensiveviewonseniorcitizens'internetusageToanswerthisquestion,wequantitativelystudythecitizensofage50orhigherinamedium-sizedcityinWesternEurope.
WecreatedaquestionnairebasedonthetheoreticalbackgroundoftheUTAUT[47]andtheDigitalDivideliterature[48,45,2,5].
Thisquestionnairewashandedouttomorethan3,000randomlychoseninhabitants.
Insum,wereceived192questionnairesfromrespondentsaged50orhigher.
Fordataanalysis,weusethepartialleastsquares(PLS)method[35].
Thepaperisstructuredasfollows.
Inthenextsection,wewillpresentsometheoreticalbackground.
Afterwards,wewilldevelopourresearchmodelbasedontheUTAUTandDigitalDivideliterature.
Insectionfour,wewillpresentourresearchWhatIstheIssuewithInternetAcceptanceamongElderlyCitizens277methodologyindetail.
Theresultsarepresentedinsectionfive.
Wewilldiscussthemintermsofrelevancefortheoryandpracticeinsectionsix.
Thelastsectionisconcernedwithlimitations,conclusions,andfutureresearch.
2TheoreticalBackgroundE-andT-governmenthavebeenestablishedasamainconceptingovernmentchangeprocessesandintegratestechnical,social,andorganisationalthemes[31,42].
Beingreadytochangeandimprovehasbecomeanecessityforpublicadministrationsinordertocopewithincreaseddemandsinacomplexchangeenvironment.
Exploitationofbenefitsrealisedbyelectronicgovernment(e-government)istheessentialpartofthisstrategy.
Beingpartofthisagenda,initstransformationgovernmentimplementationplan,theCabinetOffice[10,page4]acknowledgesthattheexploitationofthefullpotentialofelectronicservicedeliveryincludesmakingwideruseofonlineprovisioninordertomakeservicesmoreaccessibletothepublic(seeforinstance,onlinecentres[9,36]).
However,researchdiscussesage-relatedfactorsanddemographictrendsthatmightcounteracttheseefforts.
Societalagingisamajordemographictrendinindustrialisedsocieties.
Hauser&Duncan[28,p2]definedemographyas"thestudyofthesize,territorialdistribution,andcompositionofpopulation,changestherein,andthecomponentsofsuchchanges,whichmaybeidentifiedasnatality,mortality,territorialmovement(migration),andsocialmobility(changeofstatus).
"Threemajorfactorsconstitutethedevelopmentofdemography:a)fertility,b)mortality,andc)migration.
Inthiscontext,especiallyfertilityandmortalityhaveundergonesignificantchangesinmostindustrialisedcountriesoverthelastdecades.
Ontheonehand,fertilityhasbeendecliningdueto,forinstance,changedlifemodelsorfamilyplanning[38].
Ontheotherhand,regardingmortality,lifeexpectancyhasincreasedsubstantiallybecauseof,e.
g.
,improvedmedicalcare.
Forinstance,between1995and2003,lifeexpectancyatbirthinEuropeancountries,nowbeing78yearsonaverageformenand83forwomen,wentupbyanaverageof3monthseachyearformenand2monthsforwomen[17].
Asaconsequence,societalaging(synonym:populationaging)hasestablisheditselfasalong-termtrendthatwillcontinueforgenerationstocome.
Demographicprojectionsindicatethatthegroupof65yearsandolderwillcontinuetoconstituteagrowingshareofpopulation.
Forinstance,atpresent,14oftheworld's15"oldest"countriesintermsofpercentageofpeopleaged65orolder,areinEurope,whileJapanheadsthisranking[40].
In2050,fortheEuropeanUnion(EU)thepopulationshareofthoseaged65andmoreisprojectedtoincreaseto29.
9%andforJapanto39.
6%.
Similarly,intheUnitedStates(USA)andCanada,thepopulationshareofthoseaged65andmore,isestimatedtoincreaseto21%and23.
7%respectively.
Whilethedemographictrendofsocietalagingisparticularlydistinctinmoredevelopednations,lessandleastdevelopednationsalsosharethisgeneraltendency.
Societalagingposeschallengestothedevelopmentoft-governmentande-inclusionstrategies.
Oneofthesechallengesisthe(here:age-related)digitaldivide[45,2,5,3],inthiscontextunderstoodasanemergingpolarisationphenomenoninsociety,creatingagapbetweenthosewhodohaveaccesstoandusethepotentialities278B.
NiehavesandR.
PlattfautofICTs,andthosewhodonot[18].
Thedemographicgaprefers,amongstothers,tothefactthatseniorpeopleoftendonotuseICTonaregularbasis[6,39,5].
Thereasonsforthisgapresultsfromamultitudeofchallengeswhichseniorpeopleoftenface.
Theseincludeforinstanceisolation,physicaldisabilities,orlowretirementpension[33].
Disabilitiescandebarpeoplefromactivelyusinginformationtechnology.
Fortheusageofonlineservicesthemostimportantdisabilitiestoconsiderarevisualhandicaps,cognitivedefectsandlimitationsofmotorskills.
GeographicaldifferencesrefertogapsinICTusagebetweendifferentregions.
Socio-economicgapsincludedifferencesinoccupation,incomeandeducationwhereasethnicalandculturalgapsidentifybarriersintheICTusageofmigrantsandethnicalminorities.
Here,e-inclusionfocusesontheeliminationofthesebarriersfortheuseofICT.
ThedeclarationofRigagivesthefollowingdefinitionofE-inclusion:"'eInclusion'meansbothinclusiveICTandtheuseofICTtoachievewiderinclusionobjectives.
Itfocusesonparticipationofallindividualsandcommunitiesinallaspectsoftheinformationsociety.
E-inclusionpolicy,therefore,aimsatreducinggapsinICTusageandpromotingtheuseofICTtoovercomeexclusion,andimproveeconomicperformance,employmentopportunities,qualityoflife,socialparticipationandcohesion.
"[21,p.
1]Themainfocusofe-inclusionisoncreatingaccessibleservicesoverICT.
Thiseffortcanbedividedintoaccessibilityandusabilityaspects[33].
Accessibilitymeansthepossibilityforhandicappedpeopletoaccesstherelevantservice(e.
g.
Braillesupport).
Usabilityfocusesontheuser-friendlinessofaweb-service(e.
g.
easydiscoveryandfastnavigationwithinawebsite[20]).
3ResearchModelAgainstthebackgroundofourresearchobjective,ourresearchmodelisinformedbytwostreamsofresearch:acceptanceanduseoftechnologyaswellasdigitaldivideresearch.
Asforresearchonacceptanceanduseoftechnology,Venkateshetal.
[47]undertakeacomprehensivecomparisonoftheoriesinthisfieldinordertodeveloptheirUTAUT.
Theauthorsprovideevidencethat,forthecaseofinformationtechnologyacceptance,theirmodelshowsbestexplanatorypower,comparingwith,forinstance,thetheoryofreasonedaction[24,23],thetechnologyacceptancemodel[13],orthetheoryofplannedbehaviour[43].
Therefore,wewillapplyUTAUTforexplainingbehaviouralintentiontowardspersonaluseoftheinternet(BI)aswellasforexplainingusebehaviourregardingpersonalinternalusage(USE).
Here,Venkateshetal.
[47]provideevidencefortheinfluenceofthefollowingindependentvariables:PerformanceExpectancy(PE),EffortExpectancy(EE),SocialInfluence(SI),andFacilitatingConditions(FC).
Asfortherepresentationofthedigitaldivideperspective,fouradditionalvariableswereincludedinourmodel:education[45,2,5],gender[27,7,2,5],income[48,7,2,5],andmigrationbackground[2,5].
Here,weargue–inlinewithotherstudies–thatthesefactorsmoderatetherelationshipsdescribedintheoriginalUTAUTmodel.
11Pleasecontacttheauthorforinformationontheconstructs,questions,measures,andtheirroots.
WhatIstheIssuewithInternetAcceptanceamongElderlyCitizens279Accordingtostudiesoftechnologyacceptance,specificallyUTAUT,andtakingintoaccountdigitaldivideresearch,weformulatethefollowinghypothesesinordertoexplainbehaviouralintentiontowardspersonaluseoftheinternet:1)OntheinfluenceofPerformanceExpectancy:H1a:PerformanceExpectancywillpositivelyinfluenceBehaviouralIntention.
H1b:TheinfluenceofPerformanceExpectancyonBehaviouralIntentionwillbemoderatedbyeducation,gender,income,andmigrationbackground(digitaldividevariables).
2)OntheinfluenceofEffortExpectancy:H2a:EffortExpectancywillpositivelyinfluenceBehaviouralIntention.
H2b:TheinfluenceofEffortExpectancyonBehaviouralIntentionwillbemoderatedbyeducation,gender,income,andmigrationbackground(digitaldividevariables).
3)OntheinfluenceofSocialInfluence:H3a:SocialInfluencewillpositivelyinfluenceBehaviouralIntention.
H3b:TheinfluenceofSocialInfluenceonBehaviouralIntentionwillbemoderatedbyeducation,gender,income,andmigrationbackground(digitaldividevariables).
AsfortheexplanationofinternetpersonalusebehaviourweformulatethefollowinghypothesesbasedonVenkateshetal.
[47]aswellasdigitaldivideresearch:4)OntheinfluenceofBehaviouralIntention:H4:BehaviouralIntentionwillpositivelyinfluenceUseBehaviour.
5)OntheinfluenceofFacilitatingConditions:H5a:FacilitatingConditionswillpositivelyinfluenceUseBehaviour.
H5b:TheinfluenceofFacilitatingConditionsonUseBehaviouralwillbemoderatedbyeducation,gender,income,andmigrationbackground(digitaldividevariables).
WeassumethattheoriginalUTAUThassignificantpowertoexplainvariationsinbehaviouralintentiontowardspersonalinternetuseandinusebehaviour.
Moreover,weassumethattakingintoaccountinsightsfromdigitaldivideresearch,specificallyvariablessuchaseducation,gender,income,andmigrationbackground,willfurtherincreasetheexplanatorypowerofthemodel.
WethusseektoapplyUTAUTforstudyingpersonalinternetusageandtoextendthemodelbyintegratinginsightsfromdigitaldivideresearch.
4ResearchMethodologyDatacollectionphase.
Beforethedatacollectionphase,weconstructedaquestionnaireaccordingtotheresearchmodelpresentedabove.
Here,weappliedwellestablishedconstructsanditemsformeasurement.
Also,weconductedapilotstudywith7respondentsforthepurposeofquestionnairevalidation.
Itledtopositivefeedbackanddidnotresultinanychangesinthesetofquestions,items,orconstructs.
Thequestionnairewasusedtogatherdatawithinamedium-sizedcitylocatedinEuropebetweenSeptemberandOctober2009.
Weemployedamulti-channelstrategy280B.
NiehavesandR.
Plattfauttoreachtherespondents:Wecontacted100peopleviaphoneand1500viamail(bothrandomlychosen).
Moreover,weplacedadditional1,500questionnairesatthecities'town-hallandlocallibraries.
Potentialrespondentswereassuredoftheconfidentialityoftheirresponses.
Furthermore,weraffledthreematerialprizesamongallrespondents.
Thankstoanactiveinvolvementofthemayorourstudyfoundgoodcoverageinthelocalmedia.
Thus,wereceived518questionnaires(192fromrespondentsofage50orhigher).
Anadditionalnon-responseanalysisdidnotrevealanybiases.
Dataanalysisphase.
ThestructureddatawasfirstanalysedusingSPSS17.
0.
0.
Here,weselectedonlydatarecordsfromrespondentsofage50orhigher(seniorcitizens)whichledto192cases.
Tofurtheranalyseourdataset,weemployedthepartialleastsquares(PLS)pathmodellingalgorithmasitissuitablefordatasetswithlowerthan200cases[35,38].
ThesoftwarepackagetosupportthiswasSmartPLS[41].
Exceptinternetusage(formativemeasurement),allconstructsweremodelledusingreflectiveindicators(cf.
[47];foradetaileddiscussiononformativeversusreflectiveindicators,cf.
[14]).
Thedatausedincorporatessomemissingvalues(Averageof2percase).
Thesemissingvaluesweretreatedusingthemeanreplacementalgorithm[1].
Intheanalysisphasewecomparedtwodifferentmodels,onewithoutmoderatingeffectsandonewithmoderationthroughvariablesfromthedigitaldivideknowledgebase.
Thisdataanalysisprocedureallowsustoevaluatetheabovestatedhypotheses.
SampleDemographics.
Oursampleconsistsofdataof192seniorcitizens.
Themeanageoftherespondentswasslightlyabove62.
Theyspentonaverage11.
6yearsinschooloruniversitywhichprovesadecenteducation.
Concerninggender,oursampleisalmostequallydistributed(51.
56%werefemale).
Theincomevariableshowsthemostmissingvalues(52).
However,wecanobservequitehighincomesforthesamplepopulation(Table1).
Moreover,sampledemographicsshowthatthenumberofpeoplewithmigrationbackgroundisratherlow.
98%oftherespondentshavethecitizenshipofthecountrystudiedand97%arenativespeakersofthecorrespondinglanguage.
Hence,itisquitedifficulttoanalyseanyresultsrelatedtomigrationbackground.
Table1.
DemographicsoftheanalysedsampleQuestionNMinMaxMeanStd.
Dev.
AGE(inyears)19250,0083,0062,33858,41371EDU(inyearsofeducation)18002011.
633.
853INC(0=lessthan1000;1=between1000and2000;2=between2000and3000;3=morethan3000)140031,83,9525ResultsWewillpresentourresultsderivedusingtheabovementionedmethodologyinathree-steppedapproach.
First,wewillstudythevalidityofourconstructs(outerWhatIstheIssuewithInternetAcceptanceamongElderlyCitizens281model)usingstandardisedmeasures[7,46,47].
Second,wewillpresenttheinnermodel:thepathsandtheircoefficientsinbothmodels(withandwithoutmoderatingdigitaldividevariables).
Third,wewillpresentandcomparethecoefficientofdeterminationofbothmodels.
OuterModel.
Wemeasuredtheinternalconsistencyreliability(ICR)ofalllatentvariablesusingCronbach'sAlpha.
Generally,anICRabove.
9isconsideredasexcellent,onebetween.
7and.
9ashigh,onebetween.
5and.
7asmoderatelyhigh,andonebetween.
5aslow[30].
Thereliabilitiesinthepresentedstudyarecomparablyhigh,onlysocialinfluenceisinthehighmoderatearea.
ThehighICRsshowthattheitemsmeasurethecorrespondingconstruct.
Allcorrelationsbetweentheconstructswerelowerthanthesquarerootsofthesharedvariancebetweentheconstructsandtheirmeasuresineverycase.
AccordingtoFornellandLarker[25]thissupportsconvergentanddiscriminantvalidity.
2Weemployedabootstrappingmethod(500iterations)usingrandomlyselectedsub-samplestothesignificanceofourPLSmodel.
Analysingtheitemloadings,wecouldgenerallyobservethatourlatentvariablesaremeasuredbythecorrespondingitems.
AllitemsexceptPE4andFC4havecomparablyhighitemloadings(Table2).
However,analysingtheaveragevarianceextractedinallcasesshowsthatourconstructscanbeconsideredvalid[30].
Table2.
ItemLoadings(withmoderatoreffect–significanceofitemsisstable)LVItemLoadingLVItemLoadingPE1.
8910***BI1.
9301***PE2.
8190***BI2.
8323***PE3.
7681***BIBI3.
9235***PEPE4.
3629***USE01INFO.
5894EE1.
8473***USE02COMM.
2515EE2.
8244***USE03BUSI.
1113EE3.
8142***USE04BANK.
1475EEEE4.
7042***USE05HEAL.
0582SI1.
6820***USE06TOUR.
0829SI2.
5839***USE07GOVE.
0556SI3.
5977***USE08EDUC.
0217SISI4.
7666***USE09SOCI.
0147FC1.
8779***USE10GAME-.
0678FC2.
8835***USEUSE_PRI_MINPERW.
0744FC3.
8887***LANGUAGE.
9507***FCFC4.
2518*MIGNATIONALITY.
9530***a)USEwasmeasuredinaformativeway,thereforewepresentthecorrespondingweights.
b)Education,Income,andGenderweremeasuredwithonevariable.
InnerModel.
Inthefirstmodelwithoutmoderatoreffects(UTAUT),allpathshavetobeprovensignificantusingthebootstrappingmethod(Table3).
WeobservedahighinfluenceofPerformanceExpectancyonBehaviouralIntentionandof2DataforthemeasurementmodelestimationcanbefoundintheAppendixforreviewpurposesonly.
282B.
NiehavesandR.
PlattfautBehaviouralIntentiononUSE.
Theotherpathcoefficientsarecomparablylow.
However,astheanalysissuggeststhateveryconsideredpathiscorrect,wedidnotdropanyforthesecondmodelwithmoderatoreffects.
Inthesecondmodel(UTAUTanddigitaldividevariables),severalrelationshipsweremoderatedbyeducation,gender,income,andmigrationbackground.
Bythis,16interactiontermswereaddedtotheanalysis.
Themoderatorvariablemigrationbackgroundwasadded;however,asthesamplepopulationshowsalmostnomigrationbackgroundtherelatedresultsarenotinterpretable.
Bootstrappingsuggeststhatonlyaminorityofallpathsusedissignificant.
Thisisduetothehighamountofmoderatingconstructsinthemodelandcanbeignored[47].
However,somepathcoefficientsarehighandwillbefurtheranalysedinthediscussionsection.
Table3.
PathCoefficientsDependentVariable:BIDependentVariable:USEwithoutmoderatoreffectswithmoderatoreffectwithoutmoderatoreffectswithmoderatoreffectR.
5181.
6378R.
7120.
7440PE.
4651***.
0867BI.
7065***.
6469***EE.
2106**.
3892FC.
1770**.
1274SI.
1947***.
2223EDU-.
0243EDU-.
2678*GEN-.
2206GEN.
1682INC.
0320INC-.
0519MIG-.
0679MIG-.
0741FC*EDU.
1265PE*EDU.
6236*FC*GEN.
3307*PE*GEN-.
0502FC*INC.
0471PE*INC.
0394FC*MIG-.
1191PE*MIG.
0989EE*EDU-.
2068EE*GEN.
1472EE*INC.
0536EE*MIG-.
1460SI*EDU.
1354SI*GEN-.
1956SI*INC-.
0600SI*MIG-.
0736CoefficientofDetermination.
Thecoefficientofdetermination(R)isdefinedastheproportionofvariabilityinthedataexplainedbythestatisticalmodel(andnotbyrandomerrortermsornotincludedconstructs).
TheoriginalUTAUTachievedanRforBIbetween.
51and.
77andforUSEbetween.
41and.
52[47].
Ouranalysisalreadyshowsahighcoefficientofdeterminationof.
5181forBIand.
7120forUSEinthefirstmodelwithoutmoderatingeffects.
InthesecondcasewithmoderatingWhatIstheIssuewithInternetAcceptanceamongElderlyCitizens283effectswecanevenobservehigherR-ValuesforbothBI(.
6378)andUSE(.
7440).
Thus,themodelcombiningUTAUTandDigitalDivideisabletoexplainmoreofthevarianceinusagebehaviourofseniorcitizens(Table3).
6DiscussionOuterModel.
Asshownabove,allconstructsarevalidwhichisinlinewiththetheoreticalfoundation.
However,theUTAUT-originatingconstructSocialInfluencehasanICRof.
59.
ThisisonlyconsideredmoderatelyhighbyHintonetal.
[30].
Furthertheorydevelopmentcouldtrytofindbetterfittingitems,forinstancebyincludingitemsfromtheModelofAdoptionofTechnologyinHouseholds[8].
InnerModelandHypotheses.
Theresultsforthepaths'coefficientsoftheinnermodelcanbemappedwiththehypothesesmentionedinsection3.
Especiallythepathcoefficientofthemoderatingdigitaldividevariablesareofhighinterest.
Theexpectedperformanceofinternetusageisthemaindriverforelderlycitizens.
Withthehighestpathcoefficientofall,performanceexpectancyhashighinfluenceontheinternetusage.
Therefore,governmentsaimingataninclusiveinformationsocietyshouldevaluatetheire-inclusiont-governmentalstrategieswithspecialregardstoraisingthepositiveexpectationsofseniorcitizens.
Thus,ouranalysisconfirmshypothesisH1a.
TheinfluenceofPerformanceExpectancyonBehaviouralIntentionishighlypositivemoderatedbyeducation.
Especiallyforhighereducatedseniorstheexpectedperformanceisagoodpredictorfortheintentiontousetheinternet.
Othermoderatorvariablesprovideonlymarginalpowersofexplanation.
Hence,ouranalysispartiallyconfirmshypothesisH1b.
TheinfluenceofEffortExpectancyisoverestimated.
AlthoughEffortExpectancydoessignificantlyinfluenceBehaviouralIntentioninahighpositiveway,itisnotamongthemaindriversforinternetusage.
Apparently,EffortExpectancyisoverestimatedasitsinfluenceisnotashighasexpected.
However,theanalysispartiallyapprovedourhypothesisH2a.
TherelationshipbetweenEffortExpectancyandBehaviouralIntentionismoderatedbyeducationandgender.
Ontheonehand,especiallyforlesseducatedpeople,theexpectedeffortisofhighimportancefortheirBehaviouralIntention.
Ontheotherhand,thesamefactholdstrueformen.
Theinfluenceofothermoderatorvariablesislow.
Therefore,ouranalysispartiallyvalidatesthehypothesesH2b.
SocialfactorsinfluenceBehaviouralIntention.
TheimpactofSocialInfluencesonBehaviouralIntentioniscomparabletotheoneofEffortExpectancy.
Thus,hypothesisH3acanberegardedaspartiallyconfirmed.
Moreover,ouranalysisshowsthatespeciallywomenareinfluencedbytheirsocialmilieuwiththepathcoefficientforthecorrespondingmoderatorvariableat-.
1956.
Thesecondmoderatorvariableinfluencingtheimportanceofsocialfactorsiseducation.
Highlyeducatedseniorcitizensaremoreinfluencedbytheirsocialsettingthanlesseducatedones.
Thus,hypothesisH3bcanberegardedaspartiallyconfirmed.
TheinfluenceofBehaviouralIntentiononactualinternetusageishigh.
Inbothmodelstested,theinfluenceoftheintentiontouseontheactualuseisbothhighandsignificant.
Thus,wecanregardthehypothesisH4asproven.
284B.
NiehavesandR.
PlattfautFacilitatingConditionsisnotthemaindriverforinternetusage.
OuranalysisprovidesevidencethattheimpactofFacilitatingConditionsonactualusageisnotashighasexpected.
Materialaccessaspartoffacilitatingconditionsisneithertheonlynorthemaindriverforinternetusageasthecorrespondingpathcoefficientisthelowestofallconstructrelatedpathcoefficientsinthewholemodel(adH5a).
However,theimpactofFacilitatingConditionsishighlymoderatedbyeducationandgender.
Apparently,especiallyforwelleducatedmen,facilitatingconditionsarecrucialforinternetusage.
ModelComparison.
Bothpresentedmodelsexplainthevarianceofprivateinternetusesignificantly.
OurquantitativeanalysisshowsthatthefusionofUTAUTandDigitalDivideconstructsprovidesgreatvalueinpredictingboththeintentiontouseandtheuseoftheinternetinaprivatemanner.
WecanshowthatamodelthatintegratesbothapproachesisbetterthanamodelbuildingontheoriginalUTAUT-constructsonly.
However,theUTAUThastobeprovenasvaluableforpredictingprivateinternetusage.
Ourresultsbearseveralimplicationsforpractice.
Today'slocalgovernmentuseICTtolevertheirorganisationandprocessestoamoreeffectiveandefficientlevelintermsofe-governmentort-government.
However,tomaketheirICTsupportedgovernmentalservicesaccessedbyeveryonetheyneedtobridgethedigitaldivide.
AsPerformanceExpectancyisthemaindriverforbehaviouralintentiontousetheinternetlocalauthoritiesshouldthinkaboutthecommunicationandmarketingofbenefitsofinternetusageingeneralandtheusageofICTsupportedgovernmentalservices(t-government)inspecialtoelderlycitizens.
Here,especiallymoreeducatedcitizenscanbereached.
Sofar,alotofcoursestoprovidetherightskillsettoelderlycitizenshavebeeninitiatedorsupportedbylocalgovernments.
However,thestudyshowsthattheinfluenceofEffortExpectancyiscomparablylow.
Authoritiesshouldevaluatetheirundertakingsintermsofcomputercoursesandespeciallyfocusonlesseducatedpersons.
Decisionmakersshouldalsothinkaboutworkingonthesocialenvironmentoftheirinhabitantsand,e.
g.
addressstrongdisseminatorsenrootedinthecorrespondingmilieu.
Oneideawouldbetotrainlocalopinionleaderstousetheinternetandgivethemtheopportunitytotalkabouttheirpathtobecoming"experts"onthelocalradio.
Thesilverbulletoflocalgovernmentstobridgethedigitaldividehasbeentoprovideinternetaccesstoexcludedgroups.
However,ourstudysuggeststhatthisapproachisoutdated:Materialaccessaspartoffacilitatingconditionsisneithertheonlynorthemaindriverforinternetusage.
Thecorrespondingpathcoefficientisthelowestofallconstructrelatedpathcoefficientsinthewholemodel.
Apparently,purematerialaccessisnotthecrucialfactoranymore.
LocalauthoritiesshouldthereforerethinktheirengagementsinthisdirectioninordertomaketheirICTsupportedservicesusedbyeveryone.
WhatIstheIssuewithInternetAcceptanceamongElderlyCitizens2857ConclusionThispaperexaminesinfluencingfactorsforseniorcitizens'useoftheinternetforprivatepurposes.
Wepresentaresearchmodelanddevelopacorrespondingquestionnairebasedontechnologyacceptanceanddigitaldivideresearch.
Our2009surveyyields192responsesfromseniorcitizens(age50yrsandabove).
TheresultingdatasetwasanalysedusingPLSpathmodelling[41].
OurresultssuggestthatUTAUTisparticularlyusefulforanalysingprivateinternetusageachievinganR2ashighas.
7120.
Wealsofoundthatthemaindriverforseniorcitizensinternetusageisperformanceexpectancy:Thehighertheexpectedperformanceorutility,thehighertheintentiontousetheinternet.
Drawingfromdigitaldivideresearch,weextendedtheUTAUT-modelbyfouradditionalvariablesthatarehypothesisedtomediateoriginalUTAUT-relationships.
Includinginteractionterms,weobservedthate.
g.
especiallyforwomenthesocialinfluencethroughtheircorrespondingmilieuisextremelyimportantandthatmenaremoreinfluencedbythefacilitatingconditions.
Allinall,ourextendedmodelisabletoexplainasmuchas74%ofthevariationininternetusageand,therefore,isbetterthantheoriginalUTAUTmodelforthisspecificpurpose.
WethusprovideevidencethattheinclusionofdigitaldivideconstructsyieldsgreaterexplanatorypowerthanUTAUTconstructsonly.
However,ourstudyisbesetwithcertainlimitations.
First,thetotalpopulationstudieddidnotincludemanypeoplewithmigrationbackground(only3%oftherespondents).
Therefore,wecouldnotwellinterprettheresultsontheinfluenceofthisspecificvariable.
Moreover,ourstudywascarriedoutinaspecificregioninWesternEurope.
Webelievethatourresultswill,toagreatextent,holdtrueinothersettingsaswell.
Futureresearchcouldaimattestingthisassumptionbycarryingoutacomparablestudyinothernational/social/culturalsettings.
Inaddition,longitudinalstudiescouldshowthedevelopmentofprivateinternetusageanditsinfluencingfactorsamongseniorcitizensovertimeandcouldthusberegardedanotherpotentiallyfruitfulavenueforfutureresearch.
Otherfutureresearchcouldcoverthematchingofexistinglocalgovernmente-inclusionprojectswiththegivenexplainingvariables:Whichprojectscontributetoperformanceoreffortexpectancy,howissocialinfluencestimulatedandhowcanfacilitatingconditionsbeimprovedWhichprojectsaddresstheneedsofspecificgroups(seedigitaldividevariables)bestSuchoverview,webelieve,couldbeveryvaluablebutdoesnotyetexisttoourknowledge.
Asforfuturetheorydevelopment,wewereabletoexplainthelargestshareofvarianceinprivateinternetusageamongseniorcitizensbyemployingninevariables,takenfromtechnologyacceptanceanddigitaldivideresearch.
Here,webelieve,furthertestingofinfluencingfactors,forinstancepsychologicalvariables(e.
g.
,theBigFive,cf.
[12])couldstillincreaseexplanatorypower.
References1.
Afifi,A.
A.
,Elashoff,R.
M.
:ObservationsinMultivariateStatistics:I.
ReviewoftheLiterature.
JournaloftheAmericanStatisticalAssociation61(315),595–604(1966)2.
Agerwal,R.
,Animesh,A.
,Prasad,K.
:SocialInteractionsandthe"DigitalDivide":ExplainingVariationsinInternetUse.
InformationSystemsResearch20(2),277–294(2009)286B.
NiehavesandR.
Plattfaut3.
Al-Shafi,S.
,Weerakkody,V.
:UnderstandingCitizens'BehaviouralIntentionintheAdoptionofE-GovernmentServicesintheStateofQatar.
In:EuropeanConferenceonInformationSystems,Verona,Italy(2009)4.
Becker,J.
,Niehaves,B.
,Bergener,P.
,Fielenbach,K.
,Rckers,M.
,Wei,B.
:TheDigitalDivideinE-Government-AQuantitativeAnalysis.
In:ProceedingsoftheIADISInternationalConferencee-Society,Algarve,Portugal,pp.
337–344(2008)5.
Bélanger,F.
,Carter,L.
:TheImpactOfTheDigitalDivideOnE-GovernmentUse.
CommunicationsoftheACM52(4),132–135(2009)6.
Brandtweiner,R.
,Donat,E.
:DigitalDivideinAustria-AnyReasonsforEnthusiasmTheCaseofAustria.
In:Proceedingsofthe20thBledeConference–eMergence:MergingandEmergingTechnologies,ProcessesandInstitutions,BledSlovenia,June3-6(2007)7.
Brown,S.
A.
,Venkatesh,V.
:ModelofAdoptionofTechnologyinHouseholds:ABaselineModelTestandExtensionIncorporatingHouseholdLifeCycle.
MISQuarterly29(3),399–426(2005)8.
Brown,S.
A.
,Venkatesh,V.
,Bala,H.
:HouseholdtechnologyUse:IntegratingHouseholdLifeCycleandtheModelofAdoptionofTechnologyinHouseholds.
TheInformationSociety22,205–218(2006)9.
CabinetOffice:TransformationalGovernment–EnabledbyTechnology.
StrategyDocument,London,UK(2005)10.
CabinetOffice:TransformationalGovernment–ImplementationPlan,London,UK(2006)11.
Compeau,D.
R.
,Higgins,C.
A.
,Huff,S.
:SocialCognitiveTheoryandIndividualReactionstoComputingTechnology:ALongitudinalStudy.
MISQuarterly23(2),145–158(1999)12.
Costa,P.
T.
,McCrae,R.
R.
:RevisedNEOPersonalityInventory(NEO-PI-R)andNEOFive-FactorInventory(NEO-FFI)professionalmanual.
PsychologicalAssessmentResources,Odessa(1992)13.
Davis,F.
D.
:Perceivedusefulness,perceivedeaseofuse,anduseracceptanceofinformationtechnology.
MISQuarterly13(3),319–339(1989)14.
Diamantopoulos,A.
,Siguaw,J.
A.
A.
:FormativeversusReflectiveIndicatorsinOrganizationalMeasureDevelopment:AComparisonandEmpiricalIllustration.
BritishJournalofManagement17(4),263–282(2006)15.
Duff,A.
S.
,Craig,D.
,McNeill,D.
A.
:Anoteontheoriginsof'informationsociety'.
JournalofInformationScience22(5),122–171(1996)16.
Eastman,J.
K.
,Iyer,R.
:Theelderly'susesandattitudestowardstheinternet.
JournalofConsumerMarketing21(3),208–220(2004)17.
EHEMU:Arewelivinglonger,healthierlivesintheEUEuropeanHealthExpectancyMonitoringUnitTechnicalreport2,Montpellier,France(2005),http://www.
hs.
le.
ac.
uk/reves/ehemutest/pdf/techrep20507.
pdf18.
EuropeanCommission:eInclusion@EU:StrengtheningeInclusion&eAccessibilityacrossEurope.
Analyticframework-eInclusionandeAccessibilitypriorityissues(2004a)19.
EuropeanCommission:Theimpactofageingonpublicexpenditure:projectionsfortheEU25MemberStatesonpensions,healthcare,long-termcare,educationandunemploymenttransfers(2004-2005),EuropeanCommission,Brussels(2006)20.
EuropeanCommission:Topoftheweb.
UserSatisfactionandUsageSurveyofE-Governmentservices,Kopenhagen(2004b)21.
EuropeanUnion:RigaMinisterialDeclaration(2006),http://ec.
europa.
eu/information_society/events/ict_riga_2006/doc/declaration_riga.
pdfWhatIstheIssuewithInternetAcceptanceamongElderlyCitizens28722.
Ferro,E.
,Gil-Garcia,J.
R.
,Helbig,N.
:TheDigitalDivideMetaphor:UnderstandingPathstoITLiteracy.
In:Wimmer,M.
A.
,Scholl,J.
,Grnlund,.
(eds.
)EGOV2007.
LNCS,vol.
4656,pp.
265–280.
Springer,Heidelberg(2007)23.
Fishbein,M.
,Ajzen,I.
::Belief,attitude,intention,andbehaviour:Anintroductiontotheoryandresearch.
Addison-WesleyPub.
Co.
,Reading(1975)24.
Fishbein,M.
:Attitudeandthepredictionofbehaviour.
In:Fishbein,M.
(ed.
)Readingsinattitudetheoryandmeasurement,pp.
477–492.
Wiley,NewYork(1967)25.
Fornell,C.
,Larker,D.
F.
:EvaluatingStructuralEquationModelswithUnobservableVariablesandMeasurementError:AlgebraandStatistics.
JournalofMarketingResearch18(3),382–388(1981)26.
Fougère,M.
,Mérette,M.
:PopulationageingandeconomicgrowthinsevenOECDcountries.
EconomicModelling16,411–427(1999)27.
Gilly,M.
C.
,Enis,B.
M.
:RecyclingtheFamilyLifeCycle:AProposalforRedefinition.
AdvancesinConsumerResearch9,271–276(1982)28.
Hauser,P.
M.
,Duncan,O.
D.
:TheStudyofPopulation.
AnInventoryandAppraisal,Chicago(1959)29.
Henseler,J.
,Fassott,G.
:TestingModeratingEffectsinPLSPathModels:AnIllustrationofAvailableProcedures.
In:HandbookofPartialLeastSquares:Concepts,MethodsandApplicationsinMarketingandRelatedFields.
Springer,Heidelberg(2009)30.
Hinton,P.
R.
,Brownlow,C.
,McMurvay,I.
,Cozens,B.
:SPSSexplained.
RoutledgeInc.
,EastSussex(2004)31.
Irani,Z.
,Elliman,T.
,Jackson,P.
:ElectronictransformationofgovernmentintheU.
K.
:aresearchagenda.
EuropeanJournalofInformationSystems16,327–335(2007)32.
Kiel,J.
:Thedigitaldivide:Internetande-mailusebytheelderly.
MedicalInformatics&theInternetinMedicine30(1),19–23(2005)33.
Kraner,S.
:BridgingtheDigitalDivideinE-Government,Zürich(2004)34.
Machlup,F.
:TheProductionandDistributionofKnowledgeintheUnitedStates.
PrincetonUniversityPress,Princeton(1962)35.
Marcoulides,G.
A.
,Chin,W.
W.
,Saunders,C.
:ACriticalLookatPartialLeastSquaresModeling'.
MISQuarterly33(1),171–175(2009)36.
Milner,H.
:UKonlinecentres:TransformationalGovernmentfortheCitizen.
Researchreport,Sheffield,UK(2009)37.
Moore,G.
C.
,Benbasat,I.
:DevelopmentofanInstrumenttoMeasurethePerceptionsofAdoptinganInformationTechnologyInnovation.
InformationSystemsResearch2(3),192–222(1991)38.
Morgan,S.
P.
,Hagewen,K.
J.
:HandbookofPopulation(2006)39.
Niehaves,B.
,Becker,J.
:TheAge-DivideinE-Government-Data,Interpretations,TheoryFragments.
In:Proceedingsofthe8thIFIPConferenceone-Business,e-Services,ande-Society(I3E2008),Tokyo,Japan,pp.
279–287(2008)40.
PopulationReferenceBureau:WorldPopulationDataSheet2006,Washington,DC(2006),http://www.
prb.
org/pdf06/06WorldDataSheet.
pdf41.
Ringle,C.
M.
,Wende,S.
,Will,S.
:SmartPLS2.
0(M3)Beta,Hamburg(2005),http://www.
smartpls.
de42.
Smith,P.
:ABlueprintforTransformation.
E-GovernmentBulletin(2006),http://www.
cio.
gov.
uk/documents/news/pdf/ABlueprintForTransformation.
pdf43.
Taylor,S.
,Todd,P.
A.
:UnderstandingInformationTechnologyUsage:ATestofCompetingModels.
InformationSystemsResearch6(4),570–591(1995)288B.
NiehavesandR.
Plattfaut44.
Thompson,R.
L.
,Higgins,C.
A.
,Howell,J.
M.
:PersonalComputing:TowardaConceptualModelofUtilization.
MISQuarterly15(1),124–143(1991)45.
vanDijk,J.
A.
G.
M.
:Digitaldivideresearch,achievementsandshortcomings.
Poetics34(4-5),221–235(2006)46.
Venkatesh,V.
,Brown,S.
A.
,Maruping,L.
M.
,Bala,H.
:PredictingDifferentConceptualizationsofSystemUse:TheCompetingRolesofBehaviouralIntention,FacilitatingConditions,andBehaviouralExpectation.
MISQuarterly32(3),483–502(2008)47.
Venkatesh,V.
,Morris,M.
G.
,Davis,G.
,Davis,F.
:UserAcceptanceofInformationTechnology:TowardaUnifiedView.
MISQuarterly27(3),425–478(2003)48.
Wagner,J.
,Hanna,S.
:TheEffectivenessofFamilyLifeCycleVariablesinConsumerExpenditureResearch.
JournalofConsumerResearch10,281–291(1983)