JournalofEnvironmentalPsychology25(2005)273 - PDF document

JournalofEnvironmentalPsychology25(2005)273…291AreviewofinterventionstudiesaimedathouseholdenergyconservationWokjeAbrahamse,LindaSteg,CharlesVlek,TalibRothengatterDepartmentofPsychology,UniversityofGroningen,GroteKruisstraat2/1,9712TSGroningen,TheNetherlandsThisarticlereviewsandevaluatestheeffectivenessofinterventionsaimingtoencouragehouseholdstoreduceenergyconsumption. Correspondingauthor.Tel.:+31503636452.E-mailaddress:W.Abrahamse@rug.nl(W.Abrahamse). energy-intensiveappliances),economicgrowth(e.g.in-creaseofhouseholdincomes),demographicfactors(e.g.populationgrowth),institutionalfactors(e.g.governmen-talpolicies)andculturaldevelopments(e.g.emancipation,increasingmobilityofwomen)(seeGatersleben&Vlek,).Inturn,theseTEDICfactorsshapeindividual(viz.,micro-level)factorssuchasmotivationalfactors(e.g.preferences,attitudes),abilitiesandopportunities(theMOA-model,seelander&Thøgerson,1995).Iftheaimofinterventionsistoreducenegativeenvironmentalimpactbychanginghouseholdsconsumptionpatterns,itisnecessarytoconsidermacro-levelaswellasmicro-levelvariables(seealsorlingetal.,2002).Behavioralinterventionsmaybeaimedatvoluntarybehaviorchange,bytargetinganindividualsperceptions,preferencesandabilities(i.e.MOAvariables).Alternatively,interventionsmaybeaimedatchangingthecontextinwhichdecisionsarebeingmade,forinstance,through“nancialrewards,laws,ortheprovisionofenergy-ef“cientequipment(i.e.TEDICfactors).Thelatterstrategyisaimedatchangingthepay-offstructure,soastomakeenergy-savingactivitiesrelativelymoreattractive.Asthisreviewwillshow,interventionswithintherealmofsocialandenvironmentalpsychologypredominantlyfocusonvoluntarybehaviorchange,ratherthanchangingcontextualfactorswhichmaydeterminehouseholdsbehavioraldecisions.Behaviorsrelatedtohouseholdenergyconservationcanbedividedintotwocategories:ef“ciencyandcurtailmentbehaviors(Gardner&Stern,2002).Ef“ciencybehaviorsareone-shotbehaviorsandentailthepurchaseofenergy-ef“cientequipment,suchasinsulation.Curtailmentbeha-viorsinvolverepetitiveeffortstoreduceenergyuse,suchasloweringthermostatsettings.Studiesreviewedinthispaperwereaimedatbothef“ciencyand/orcurtailmentbehaviors,withthelatterseemingsomewhatoverrepresented.Thisisstriking,becausetheenergy-savingpotentialofef“ciencybehaviorsisconsideredgreaterthanthatofcurtailmentbehaviors(e.g.Gardner&Stern,2002).Forinstance,householdsmaysavemoreenergybyproperlyinsulatingtheirhomesthanbyloweringthermostatsettings.Itshouldbenotedhowever,thatenergy-ef“cientappliancesdonotnecessarilyresultinareductionofoverallenergycon-sumptionwhenpeopleusetheseappliancesmoreoften(theso-calledreboundeffect,seeBerkhout,Muskens,&Veldhuijsen,2000).Here,theimportanceoftheinterplaybetweenmacro-level(e.g.technologicalinnovations)andmicro-levelfactors(e.g.knowledgeofef“cientuseoftechnologicalinnovations)becomesapparent.Varioussocialandenvironmentalpsychologicalstudieshaveembarkedonissuesrelatedtohouseholdenergyuse.Onelineofresearchfocusesontestingtheeffectivenessofinterventionstrategiesaimingtochangeenergy-relatedbehaviors.Anotherlineofresearchistheorydrivenandaimstoidentifyunderlyingdeterminantsofenergyuse,suchasattitudes(e.g.Becker,Seligman,Fazio,&Darley,)andsocio-demographics(e.g.Black,Stern,&Elworth,1985).Insomestudies,boththeeffectivenessofaninterventionaswellas(changesin)underlyingdeterminantsofenergyusearemonitoredsimultaneouslyGeller,1981Staats,Wit,&Midden,1996).Thelattergiveadditionalinsightintoreasonswhyinterventionsweresuccessfulornot,andassuch,theyareastartingpointforthefurtherenhancementofaninterventionseffectiveness.Thepurposeofthispaperistwofold.First,empiricalstudiesontheeffectivenessofinterventionstopromotehouseholdenergyconservationarereviewed.Theaimistocometoconsistent“ndingswithrespecttotheeffectivenessoftheseinterventions.Itisexaminedwhichfactorsdetermineaninterventionssuccessorfailure.Interven-tionsaremoreeffectivetotheextentthattheytargetdeterminantsofenergyuseandenergysavings(e.g.attitudes,knowledge).Second,basedonthestrengthsandshortcomingsoftheresearchreviewedhere,sugges-tionsaregivenonhowtoimproveourunderstandingandknowledgeofeffectiveinterventionplanning.Indoingso,thisreviewaimstocomplementandupdatepreviousreviewsonenergyconservationandotherpro-environ-mentalbehaviors(e.g.Cook&Berrenberg,1981DeYoung,1993Dwyer,Leeming,Cobern,Porter,&Jackson,1993Geller,2002Schultz,Oskamp,&Mainieri,Stern,1992Winett&Kagel,19842.Method2.1.SelectionprocedureVarioussocialandenvironmentalpsychologicaljournalsanddatabases(e.g.PSYCHLit,WebSPIRS)werecon-sulted.Further,referencelistsofarticleswereusedtolocateadditionalpublishedmaterial.Thissearchresultedinatotalof38peer-reviewed(i.e.qualityguarantee)studies,datingfrom1977to2004.Thesestudiesweremostly“eldexperiments,usingquasi-experimentaldesigns.Onesinglestudywasconductedinalaboratorysetting.Inordertobeselectedforreview,thestudyhadtoincludeadesignallowingforeffectstobemeasuredeithercomparedtoabaseline(pretest/post-testdesign)ortoacontrolgroup.Anotherimportantselectioncriterionwasthatthetargetgroupunderstudybehouseholds.Themainreasonforthisisthedifferentialeffectaninterventionmayhavedependingonthetargetgroup.Forinstance,comparativefeedback(i.e.feedbackabouttheperformanceofothers)hasbeenshowntohavepositiveeffectsonreducingenergyuseintheworkplace(seeSiero,Bakker,Dekker,&VandenBurg,1996).Asthisreviewwillindicate,theresultsarenotasclear-cutforhouseholds(e.g.VanHouwelingen&VanRaaij,1989Theselectedstudiesareclassi“edaccordingtothetaxonomyforbehaviorchangeinterventionsasproposedGelleretal.(1990)(seealsoDwyeretal.,1993Schultzetal.,1995),inwhichadistinctionismadebetweenantecedentandconsequencestrategies.Antecedentinterven-tionsareassumedtoin”uenceoneormoredeterminantsARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 priortotheperformanceofenvironmentallysigni“cantbehaviors.Forinstance,providinghouseholdswithin-formationaboutenergy-savingoptionsmayresultinenergysavings,becausepeoplehaveacquired(more)knowledge.Aconsequencestrategyisassumedtoin”uencedeterminantsaftertheoccurrenceofapro-environmentalbehavior,bymeansofprovidingaconsequencewhichiscontingentontheoutcomeofthebehavior.Forinstance,givinghouseholdsfeedbackabouttheirenergysavingsmayencouragethemto(further)reduceenergyuse,becausetheirlevelofself-ef“cacy(i.e.perceivedpossibilitiestoconserveenergy)hasincreased.2.2.EvaluationcriteriaToassesstheeffectivenessofinterventionsaimedatreducingenergyuse,thefollowingcriteriawereconsidered.First,theextenttowhichinterventionsresultedinbehavioralchangesand/orreductionsofenergyuseisreported.Itisimportanttomonitorboth,becausehouse-holdsmayhaveadoptedenergy-savingbehaviorswithoutdecreasingoverallenergyuse.Second,anindicationisgivenoftheextenttowhichthesechangescanbeattributedtotheintervention(s),bycomparingexperimentalgroupswithacontrolgroupnotexposedtotheintervention(s).Wheresuf“cientquantitativeinformationwasreported,effectsizeswerecalculated.TheeffectsizeindexusedforthispurposewasCohensd,whichwasestimatedbydividingthebetween-groupsdifferenceinmeanscoresbythepooledwithin-groupstandarddeviation(Hunter&Schmidt,1990).Theeffectsizewasgivenapositivevaluewhentheexperimentalgroupssavedmoreenergythanthecontrolgroup,andanegativevaluewhentheyusedmoreenergy.Effectsizesthusrepresentthenumberofstandarddeviationunitsbywhichtheinterventiongroupout-performedthecontrolgrouponacertainoutcomevariable(e.g.gassavings,electricitysavings).Whenmeansandstandarddeviationswerenotprovided,effectsizeswerecalculatedfromotheravailablestatisticalinformation(e.g.-ratios,-tests)accordingtotablesandformulasproposedGlass,McGraw,andSmith(1981),andSeifert(1991)Notinfrequentlyhowever,thenecessarystatisticalinfor-mationwasnotreported,andconsequently,effectsizescouldnotbecalculated.Therefore,conductingathoroughmeta-analysiswasnotdeemedfeasible.Third,itwasexaminedwhyinterventionswere(in)effective,bymeansofreportingchangesinunderlyingbehavioraldetermi-nants.Ourassumptionisthatinterventionsaremoreeffectivetotheextentthattheytargetandchangeimportantdeterminantsofenergyuse.Forexample,aninformationcampaignmaynothavebeeneffectivebecausenoincreaseinknowledgeoccurred.Suchconclusionscanbedrawnonlywhenchangesinbehavioraldeterminantsandinactualbehavioraremonitoredsimultaneously.Finally,itisreportedwhethertheeffectsoftheinterven-tionsweremonitoredoveralongerperiodoftime,inordertoassesswhetherhouseholdspersistedinbehavinginanenergy-ef“cientway,wellaftertheinterventionwasdiscontinued.Inthispaper,studiesusingantecedentinterventions(commitment,goalsetting,information,andmodeling)willbediscussed“rst,followedbystudiesusingconsequenceinterventions(feedbackandrewards).Variousstudiesexaminedtheeffectofacombinationofantecedentand/orconsequencestrategiesandthesestudiesaregroupedaccordingtotheinterventionbeingvariedacrossexperi-mentalconditions.Inthetext,mainissuesconcerningthestudieswillbeaddressed.AdditionalinformationcanbefoundinTableA1(seeAppendix),whichlistsallstudiesreviewedhere.Thistablegivesanoverviewoftypeofintervention(s),designofthestudy,totalnumberofparticipatinghouseholds,targetbehavior(andwhetherthisinvolvedcurtailmentbehaviors,ef“ciencybehaviors,orboth),measurementofdeterminants,durationoftheintervention,effectivenessoftheintervention,effectsizes,andlong-termeffects.3.AntecedentInterventionsInthissection,studiesarediscussedusingantecedentinterventionstopromotehouseholdenergyconservation.Asmentionedearlier,antecedentinterventionsin”uenceoneormoredeterminantspriortotheperformanceofbehavior.Thatis,interventions(e.g.information)areaimedatin”uencingunderlyingbehavioraldeterminants(e.g.knowledge),whichinturnarebelievedtoin”uencebehavior.Thefollowinginterventionsareconsideredantecedentinterventions:commitment,goalsetting,infor-mation,andmodeling.3.1.CommitmentAcommitmentisanoralorwrittenpledgeorpromisetochangebehavior(e.g.toconserveenergy).Moreoftenthannot,thispromiseislinkedtoaspeci“cgoal,forinstance,toreduceenergyuseby5%.Thispromisecanbeapledgetooneself,inwhichcaseitmayactivateapersonalnorm(viz.,amoralobligation)toconserveenergy.Thepromisecanalsobemadepublic,forinstance,bymeansofanannouncementinthelocalnewspaper.Then,socialnorms(viz.,expectationsofothers)mayplayaroleasdetermi-nantsofconservationbehavior.KatzevandJohnson(1983)measuredtheeffectofcommitmentonelectricityconsumption,bymeansoftheso-calledfoot-in-the-doortechnique.Theassumptionbehindthistechniqueisthatcompliancetoa“rst(smaller)requestwillresultincompliancetoasubsequent(bigger)request.Inthe“rststudy,householdseitherreceiveda(small)requestto“lloutaquestionnaire,a(bigger)requesttosignacommitmenttoconserveenergyby10%,orbothrequests.Thecommitmentwasaccompaniedbyinforma-tionaboutenergyconservation.Householdswhohadreceivedeitherrequest,orboth,savedmoreenergycomparedtoacontrolgroup.ThiseffectdidnotemergeARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 duringtheintervention,butduringafollow-upperiod.Inasubsequentstudy(Katzev&Johnson,1984),twoexperi-mentalgroupswereadded:householdswhoreceivedareward(dependingontheamountofelectricitysaved),andhouseholdswhoreceivedallinterventions(questionnair-e+commitment+information+reward).Incontrasttothepreviousstudy,theeffectoccurredonlyduring(the“rstweekof)theinterventionperiod:thecommitmentgroupandthecombinedtreatmentgroupshowedthelargestdeclineinelectricityuse.Arelativelylownumberofrespondentsperconditionmayhavereducedthestatisticalpowerofbothdesigns.PallakandCummings(1976)usedcommitmenttopromotegasandelectricityconservationamonghouse-holds.Thosewhohadsignedapubliccommitment(i.e.publicationinalea”et)showedalowerrateofincreaseinbothgasandelectricityconsumptionthanthoseineithertheprivatecommitmentorthecontrolgroup.Thiseffectwasmaintainedoveraperiodof6monthsfollowingdiscontinuationoftheintervention.3.2.GoalsettingGoalsettingentailsgivinghouseholdsareferencepoint,forinstancetosave5%or15%energy.Agoalcanbesetbytheexperimenters,orbythehouseholdsthemselves.Itisoftenusedincombinationwithotherinterventions,suchasfeedback(toindicatehowhouseholdsareperformingrelativetothegoal),oraspartofacommitmenttoconserveacertainamountofenergy.Inthissection,wediscussstudiesinwhichtheuniquecontributionofgoalsettingcouldbeestablished.Forgoalsettingaspartofotherinterventions,thereaderisreferredtothesectionsoncommitmentandfeedback.Becker(1978)gavehouseholdseitherarelativelydif“cultgoal(20%)orarelativelyeasygoal(2%)toreduceelectricityuse.Thegoalwaseithercombinedwithfeedback(threetimesaweek),ornot.Allhouseholds(includingthecontrolgroup)receivedinformationonwhichappliancesusedmostelectricity.Householdswhoreceivedadif“cultgoalandfeedbackconservedmost(15.1%)andweretheonlygrouptosigni“cantlydifferfromthecontrolgroup.Thisindicatesthatinorderfora(dif“cult)goaltowork,householdsneedfeedbackonhowtheyareperforminginrelationtothegoal.Aneasygoalappearednottobeeffectiveatall;2%mayhavebeenperceivedasnotbeingworththeeffort.Oneofthefewstudiestotakethe“eldintothelabwasconductedbyMcCalleyandMidden(2002).Theyappliedgoalsettingincombinationwithfeedbacktoonespeci“cenergy-relatedbehavior:doingthelaundry.Inalaboratorysetting,agoalsettingprocedurewasused,andimmediatefeedbackwasgivenaboutaverageamountofenergy(kWh)usedperwashingtrial,displayedinasimulatedcontrolpanelofawashingmachine.Participantswhohadbeengivenagoalaswellasfeedbacksavedmoreenergyperwashingtrialthanparticipantswhohadonlyreceivedfeedback(withoutagoal).Nosigni“cantdifferenceemergedbetweenparticipantswhohadbeenabletosetagoalthemselvesandthosewithanassignedgoal.Socialvalueorientation(i.e.theextenttowhichonevaluesoutcomesforoneselforforothers)wasmeasuredaswell,andinterestingly,therewasasigni“cantinteractionbetweensocialvalueorientationandtypeofgoal.Forpro-selfrespondents,anassignedgoalresultedinlowerenergysavingsthanaself-setgoal,whileforpro-socialrespondentsthereversewastrue.3.3.InformationInformationisacommonlyusedstrategytopromoteenergyconservationbehaviors.Thismaybegeneralinformationaboutenergy-relatedproblems,orspeci“cinformationaboutpossiblesolutions,suchasinformationaboutvariousenergy-savingmeasureshouseholdscanadopt.Providinginformationservestoincreasehouse-holdsawarenessofenergyproblemsandtheirknowledgeaboutpossibilitiestoreducetheseproblems.Informationaboutenergyconservationcanbeconveyedtohouseholdsinseveralways.Inthissection,wediscussworkshops,massmediacampaignsandtailoredinformation.3.3.1.WorkshopsGeller(1981)measuredtheeffectivenessofaworkshop,inwhichinformationaboutenergy-savingmeasureswasgiven.Inaddition,eachparticipantreceivedashower-”owrestrictorandabookletwithinformationaboutenergyconservation.Theworkshopledtohigherlevelsofconcernabouttheenergycrisis,toanincreaseinknowledgeaboutenergyconservation,andstrongerintentionstoadoptenergy-savingmeasures.Home-visitsrevealednodiffer-encesbetweenattendeesandnonattendeesinthenumberofadoptedenergy-savingmeasures.So,althoughinformationdidin”uenceunderlyingdeterminantsofenergyuse,itdidnotresultinbehavioralchanges.3.3.2.MassmediacampaignsLuyben(1982)evaluatedtheeffectivenessofPresidentCarterstelevisedpleatolowerthermostatsettingsinviewofapotentialgasshortage.Threedaysaftertheappeal,randomlyselectedresidentsweresurveyedeitherbytelephoneoradoor-to-doorinterview.Thereappearedtobenodifferenceinthermostatsettingsbetweenthosewhohadheardthepleaandthosewhohadnot.Also,nodifferenceinknowledgeofthefactthatloweringthermo-statsettingswouldhelpreduceenergyusewasfoundbetweenthosewhohadandthosewhohadnotheardtheplea.Interestingly,self-reportedthermostatsettingsap-pearedtobesigni“cantlylowerthanthoseobservedbyinterviewers,pointingtoapossiblein”uenceofsocialdesirability.HuttonandMcNeill(1981)evaluatedtheLowCost/NoCostenergyconservationprogramoftheUSDepartmentofEnergy.Abookletofenergy-savingtipsandashowerARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ”owcontroldevicewassentto4.5millionhouseholds.Inaddition,amassmediacampaignwaslaunched.Toevaluateitssuccess,atelephonesurveywasconducted.Householdswhohadreceivedthebookletandtheshowerdevicereportedimplementingtheenergy-savingtipsmoreoftenthanhouseholdswhohadnot.Overall,thosewhohadinstalledtheshower”owdevicereportedapplyingsigni“cantlymoretipsthanthosewhohadnot.Itisnotreportedwhethertheinterventionresultedinactualenergysavings.AstudybyStaatsetal.(1996)evaluatedamassmediacampaignoftheDutchgovernment,aimedatcommu-nicatingthenatureandcausesofglobalwarming,andpossiblewaysofdealingwithit.Apretest/post-testsurveyrevealedaslightincreaseinknowledge,butlevelsofawarenessoftheproblemremainedunchanged.Will-ingnesstobehavepro-environmentallyincreased,butonlyamongthosewhohadalreadybeenbehavingpro-environmentallybeforethecampaign.Knowledgeandproblemawarenesswerenotrelatedtoself-reportedpro-environmentalbehaviors.3.3.3.Tailoring:homeauditsTailoredinformationishighlypersonalizedandspeci“cinformation.Anadvantageofthisapproachisthatparticipantsreceiverelevantinformationonly,ratherthangettinganoverloadofgeneralinformation,whichmaynotalwaysapplytotheirhouseholdsituation.Tailoringhasalreadygaineditsmeritsinotherdomains,suchashealthcare(seeKreuter,Farrell,Olevitch,&Brennan,1999).Examplesoftailoringintherealmofenergyconservationareenergyaudits,i.e.ahomevisitbyanauditorwhogiveshouseholdsarangeofenergy-savingoptions(ef“ciencyandcurtailmentbehaviors)basedontheircurrentsituation.Forinstance,theymayadviseahouseholdtoapplyinsulationandlowerthermostatsettings.Severalstudiesinvestigatedtheeffectofhomeenergyaudits.AstudybyWinett,Love,andKidd(1982…1983)showedthathouseholdswhohadreceivedanenergyaudit(providinginformationonheatingandairconditioning)used21%lesselectricity,comparedtoacontrolgroup.HirstandGrady(1982…1983)comparedgasconsump-tionofhouseholdswhohadreceivedhomeauditstothosewhohadnot.Oneyearfollowingtheaudit,householdshadsaved1…2%ongasuse,comparedtothecontrolgroup;2yearsfollowingtheaudit,thisamountedto4%.Inaddition,householdsintheauditedgroupreportedapplyingmoreenergy-savingmeasuresthanthecontrolgroup.Contrarytowhatwasexpected,amorepositiveattitudetowardsgasconserva-tionwasassociatedwithhighergasuse.Theauthorsdrawcautiousconclusions,expressingdoubtsaboutdataquality.Anotherstudyexaminedwhetherenergyauditorstrainedtousepersuasionprinciples(e.g.useofvivid,personalizedinformation)wouldbemoresuccessfulinencouraginghouseholdstoadoptenergy-savingmeasuresthanauditorswithoutsuchtraining(Gonzales,Aronson,&Costanzo,1988).Householdsinthetrained-auditorgroupreportedasigni“cantlygreaterlikelihoodofmakingtherecommendedchangesthanhouseholdsintheothergroup.Aftermakingthesechanges,homeownerscouldapplyforarebate.Thenumberofapplicationsservedasabehavioralmeasureforevaluatingtheauditssuccess.Signi“cantlymorehomeownersfromthetrainedauditorconditionappliedfora“nancialrebate.However,nodifferenceinactualenergyusewasfound.McDougall,Claxton,andRitchie(1982…1983)evaluatedtheCanadianENERSAVEprogram.Householdswhoparticipatedintheprogramcompletedaquestionnaireaboutcurrentbehaviorsrelatedtoenergyuse(e.g.amountofinsulation,thermostatsettings).Onegroupofpartici-pantsreceivedtailoredinformationaftersendinginthequestionnaire;theothergroupofparticipantsdidnot.Twoyearslater,participantswerecontactedagain,andnodifferenceswerefoundinreportedenergy-savingactions,orinactualenergyusebetweenthosewhohadreceivedtailoredinformationandthosewhohadnot.Apossibleexplanationistherelativelylongtimethatelapsedbetweenimplementationoftheinterventionanditseffectmeasure-Recently,McMakin,Malone,andLundgren(2002)appliedtailoringtoenergyconservationamonghouseholdslivingattwoUSmilitaryinstallations.Thetailoredinformationwasbasedonfocusgroupinterviewscon-ductedpriortotheintervention.Informationaboutenergyconservationinthe“rstinstallation(inthestateofWashington)targetedheating-relatedenergyuseandthesecond(locatedinArizona)targetedcooling-relatedenergyuse.Resultsweremixed:householdsinWashingtonsaved10%ontheirgasandelectricityuse,andhouseholdsinArizonaused2%moreelectricity,comparedtobaseline3.4.ModelingModeling,basedonBanduraslearningtheory,entailsprovidingexamplesofrecommendedbeha-viors.Itisassumedthattheseexampleswillbefollowedwhentheyareunderstandable,relevant,meaningfulandrewarding(intermsofpositiveresults)topeople.Winett,Leckliter,Chinn,Stahl,andLove(1985)usedmodelingbymeansofcableTV.Theprogramwastailoredinthesensethatitwastargetedatmiddle-classhomeownersanditshowedvariousenergy-savingmeasures.Viewersalsoreceivedaninformationbookletcontainingcartoonsdepictingenergy-savingmeasures.TheTVmodelinggroupsigni“cantlyreducedenergyuseby10%,comparedtoacontrolgroup.Beforeandaftermeasuresrevealedasigni“cantincreaseinknowledgefortheexperimentalgroup,butnotforthecontrolgroup.Afollow-upstudyoneyearlatershowedthattheenergysavingswerenotARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 4.Antecedentinterventions:conclusionsCommitmentmaybeasuccessfulstrategyforreducinghouseholdenergyuse,especiallyinviewofthelong-termeffectsfoundinseveralstudies(Katzev&Johnson,1983Pallak&Cummings,1976).However,KatzevandJohn-sonssecondstudy(1984)onlyfoundshort-termeffectsofcommitment.Studiesongoalsetting(Becker,1978McCalley&Midden,2002)showedthatcombininggoalsettingwithfeedbackwasmoreeffectivethangoalsettingalone.Informationhasalsoproventobemoreeffectivewhenusedincombinationwithotherinterventions(e.g.VanHouwelingen&VanRaaij,1989).Theeffectsofinformationseemtodependlargelyonitsspeci“city.Massmediacampaignstendtoresultinanincreaseinattitudesorknowledge(e.g.Staatsetal.,1996),butthereisnoclearevidencethatthisresultsinreductionsofenergyuse.Itmaywellbethatamorepersonalizedapproachsuchastailoringismoreeffective.Homeenergyaudits,usingtailoredenergyadvice,hadpositiveeffectsonhouseholdenergyuseWinettetal.,1982…1983)andontheextenttowhichef“ciencyactionsweretaken(Gonzalesetal.,1988).Also,tailoredinformationwassuccessfulamonghouseholdsinamilitaryinstallation(McMakinetal.,2002),butMcDou-galletal.(1982…1983)failedto“ndanyreductionsinenergyuseasaresultoftailoring.Finally,modelingWinettetal.,1985)resultedinaknowledgeincrease,andwasalsoeffectiveinreducingenergyuse.Strikingly,anumberofstudiesincludedrelativelylownumbersofhouseholdsperexperimentalcondition,whichmayhavebeenacausefornot“ndinganystatisticallysigni“canteffects.Also,ithasbeenshownthatcombinationsofinterventionsareespeciallyeffectiveinreducingenergyuse.However,studiesoftenemploydesignswithoutincludingexperimentalconditionsinwhichthesingleinterventionswereused,inwhichcaseitisdif“culttoestablishanyadditionaleffectsofthecombinationoverandabovetheinterventionsseparately.5.ConsequenceinterventionsConsequencestrategiesarebasedontheassumptionthatthepresenceofpositiveornegativeconsequenceswillin”uencebehavior.Pro-environmentalbehaviorwillbe-comeamoreattractivealternativewhenpositiveconse-quencesareattachedtoit(e.g.byprovidingamonetaryincentive),andenvironmentallyunsoundbehaviorwillbecomelessattractivewhennegativeconsequencesareattachedtoit.Feedbackandrewardswillbediscussedinthissection.5.1.FeedbackFeedbackisoftenappliedtopromoteenergyconserva-tion.Feedbackconsistsofgivinghouseholdsinformationabouttheirenergyconsumption,orenergysavings.Itcanin”uencebehavior,becausehouseholdscanassociatecertainoutcomes(e.g.energysavings)withtheirbehavior.Ideally,feedbackisgivenimmediatelyafterthebehavioroccurs(Geller,2002).First,wediscussstudiesfocusingonthedifferentialeffectoffeedbackfrequency,followedbystudiessystematicallyvaryingfeedbackcontent.5.1.1.ContinuousfeedbackMcClellandandCook(1979…1980)gavehouseholdscontinuousfeedbackoveraperiodof11monthsaboutmonetarycostsofelectricityusebymeansofamonitordisplayingelectricityuseincentsperhour.Onaverage,householdswhohadamonitorinstalledintheirhomesused12%lesselectricitythanacontrolgroup.Inasimilarvein,Hutton,Mauser,Filiatrault,andAhtola(1986)testedwhethercontinuouscost-relatedfeedbackbymeansoftheso-calledEnergyCostIndicator(ECI)wouldbeeffectiveinreducinggasandelectricityuse.Theyalsoprovidedparticipantswithinformationaboutenergyconservation.ThestudywasconductedintwoCanadianandoneAmericancity.IntheCanadiancities,behavioralchangeswereobserved:householdswhohadeitherreceivedinformationonly,orinformationcombinedwithfeedbackused4…5%lessenergythanacontrolgroup.However,nochangesinknowledgewereobserved.IntheAmericancityhowever,anincreaseinknowledgeoccurred,butnobehavioraleffectwasfound.TheauthorsattributethistothepossibilitythatknowledgeofenergyissueswasalreadyhigherinCanadathanintheUS(i.e.ceilingeffect).AnotherstudyusingasimilarIndicatorwasdonebyVanHouwelingenandVanRaaij(1989).Theyinvestigatedthedifferentialeffectofcontinuousversusmonthlyfeedbackongasconsumptionbymeansofafeedbackmonitordisplayingdailygasuseaswellasdailytargetconsumption(basedonannualgasuse),thelatterservingasconservationgoal.Allhouseholdsreceivedinformationaboutenergyconservation.Itappearedthathouseholdswhohadreceivedcontinuousfeedbacksavedmoregas(12.3%)thanthosewhohadreceivedmonthlyfeedback(7.7%),thosewhohadbeentaughttoreadtheirgasmeter(5.1%)andthosewhohadonlyreceivedinformation(4.3%).Nosigni“cantchangesingasusewereobservedinacontrolgroup.Thisstudyfoundthatlowusersofgasactuallyincreasedgasuseduringtheintervention.Oneyearafterterminationoftheinterven-tion,gasusehadincreasedforallgroups,comparedtobaselinelevels.Sexton,Brown,Johnson,andKonakayama(1987)continuousfeedbackaboutthedifferencebetweenmone-tarycostsofelectricityusedinon-andoff-peakperiods(thelatterhavingacheaperrate).Feedbackdidresultinashiftinconsumptiontooff-peakperiods,whichwaslargestforthosehouseholdswhohadbeengivenahigherpricedifference.However,totalelectricityconsumptiondidnotdecrease.Inthiscase,thefeedbackmayhavealsoservedasa“nancialincentive(viz.,reward).ARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 5.1.2.DailyfeedbackBittle,Valesano,andThaler(1979)assignedhouseholdsintoeitheradailyfeedbackgrouporacontrolgroup.Thefeedbackgroupsavedanaverageof4%ontheirelectricityuse(comparedtobaselineconsumption),andalsosavedmorethanthecontrolgroup.Then,thetreatmentwasreversed.Householdsinitiallypartofthefeedbackgroup(nownolongerreceivingfeedback)continuedsavingmoreelectricitythanthecontrolgroup(nowreceivingfeedback).Thisisprobablyduetoacarry-overeffect:newhabitsmayhaveformed;persistingevenafterfeedbackwasdiscon-Bittle,Valesano,andThaler(1979…1980)exploredthedifferentialeffectoffeedbackcontent.Allhouseholdsreceiveddailyfeedback,butadistinctionwasmadebetweenfeedbackaboutpreviousdayselectricityuseandcumulativefeedback(electricityusesince“rstofthemonth).AnotherdistinctionwasmadebetweenfeedbackintermsofkWhandintermsofcost.Highconsumersofelectricityshowedalowerrateofincreaseinelectricityuse,withcumulativefeedbackbeingsomewhatmoreeffectivethanfeedbackaboutdailyuse.Formediumandlowconsumershowever,feedbackappearedtohavetheoppositeeffectandresultedinanincreaseofelectricityuse.Katzev,Cooper,andFisher(1980…1981)eithergavehouseholdsdailyfeedbackaboutelectricityuse(kWh,costandcomparedtootherhouseholds),feedbackeverythirdday(kWh,cost,andcomparedtoothers),ornoncontin-gent(viz.,regardlessofwhetherhouseholdshadactuallysavedelectricityornot)feedback(kWhandcost).Nosigni“cantgroupdifferencesinelectricityusewerefound,possiblyduetoalownumberofrespondentsineachexperimentalgroup.Feedbackandself-monitoringwerecomparedinastudyonelectricityuse(Winett,Neale,&Grier,1979).House-holdsweregiveninformationabouthowtoconserveandtheywereaskedtochooseanenergyconservationgoal.Resultsshowthathouseholdswhohadreceiveddailyfeedbackused13%lesselectricity,andhouseholdswhoweretaughttoreadtheiroutdoormeters(self-monitoring)used7%lesselectricitythandidacontrolgroup.Thiseffectwasstillpresentduringafollow-upmeasurement.SeligmanandDarley(1977)foundfeedbacktohaveapositiveeffectonelectricityconservation.Allparticipatinghouseholdsweretoldthatairconditionerswerethelargestusersofelectricityinhomes.Halfofthemreceivedfeedbackaboutelectricitysavings(fourtimesaweekduringonemonth),whiletheotherhalfdidnotreceiveanyfeedback.Householdsinthefeedbackgroupused10.5%lesselectricitythanthecontrolgroupdid.Therewasnofollow-upmeasurementtodeterminewhethertheeffectwasmaintained.5.1.3.WeeklyandmonthlyfeedbackArecentstudybyllinkandMeertens(1999)usedacombinationofweeklyfeedback,goalsetting(householdshadachoicebetween5%,10%,or15%energysavings),andinformation(energy-savingtips)throughtextTV.Householdswhoweresubjecttothecombinationofinterventionssavedmoreenergythanthecontrolgroupdid.However,sinceparticipantswerelivinginenergy-ef“cienthomes,theresultscannoteasilybegeneralizedtothegeneralpopulation.HayesandCone(1981)examinedtheeffectofmonthlyfeedbackonelectricityuse,bothintermsofkWhaswellasintermsofmoney.Householdswhohadreceivedfeedbackreducedelectricityuseby4.7%,whilehouseholdsinthecontrolgroupincreasedelectricityuseby2.3%.Afterthefeedbackwaswithdrawn,electricityusewasmonitoredoveraperiodof2months,andduringthisperiod,thepatternwasreversed:householdsintheexperimentalgroupused11.3%more,whilehouseholdsinthecontrolgroupsaved0.3%comparedtobaselineAstudybyHeberleinandWarriner(1983)focusedonthepricedifferencebetweenon-andoff-peakperiods(thelatterhavingacheaperrate).Householdsreceivedmonthlyfeedback(throughtheirelectricitybill)abouttheamountofkWhtheyhadusedinon-andoff-peakperiods.Knowledgeofpriceratioandbehavioralcommitmenttoshiftconsumptionfromon-peaktooff-peakperiodsweremeasured.Largerpricedifferencesresultedinlargershiftstooff-peakperiods.Regressionanalysisrevealedknowl-edgeandbehavioralcommitmenttohavestrongereffectsonthisshiftthanpricehad.ItwasnotreportedwhetherfeedbackaboutpriceratioledtoreductionsinelectricityKantola,Syme,andCampbell(1984)usedacombina-tionoffeedbackandinformationamongaboveaverageconsumersofelectricity.Theyusedfeedbacktoevokecognitivedissonancebyinforminghouseholdsthateventhoughtheyhadpreviouslyindicatedfeelingadutytoconserveenergy,theywerehighconsumersofelectricity.Allparticipantsreceivedenergy-savingtips.The“rstgroupreceivedcognitivedissonancefeedback,thesecondgroupreceivedfeedbackthattheywerehighconsumersofelectricity(withoutinducingcognitivedissonance),thethirdgrouponlyreceivedenergy-savingtips.Duringthe“rst2weeksoftheinterventionperiod,thecognitivedissonancegroupsavedsigni“cantlymoreelectricitythantheothergroupsdid.Forthesecond2weeks,thedissonancegroupdifferedfromthecontrolgrouponly.5.1.4.ComparativefeedbackFeedbackaboutindividualperformancerelativetoperformanceofothersmaybehelpfulinreducinghouse-holdenergyuseaswell.Bygivingcomparativefeedback,afeelingofcompetition,socialcomparison,orsocialpressuremaybeevoked,whichmaybeespeciallyeffectivewhenimportantorrelevantothersareusedasareferenceARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 Midden,Meter,Weenig,andZieverink(1983)theeffectivenessofcomparativefeedback,individualfeedback,monetaryrewardsandinformation.Thecomparativefeedbackconsistedofacomparisonwithconsumptionlevelsofhouseholdsinsimilarsettings.Onlymarginallysigni“cantdifferencesemergedbetweenthegroups.Forelectricityuse,householdswhohadeitherreceivedcomparativefeedback,individualfeedbackorrewardstendedtosavemorethanthecontrolgroupdid.Forgasuse,householdswhohadreceivedeitherindividualfeedbackorrewardstendedtosavemost.Overall,comparativefeedbackwasnotmoreeffectivethanindividualfeedback,andprovidinghouseholdswithinformationalonewasnoteffectiveatall.Inarecentstudy,adistinctionwasmadebetweencomparativefeedback(i.e.ownsavingscomparedtootherparticipants),individualfeedback,feedbackabout“nancialcosts,andfeedbackaboutenvironmentalcosts(Brandon&Lewis,1999).Also,onegroupofhouseholdsreceivedfeedbackonalea”et,whileanotherreceivedcomputerizedfeedback.Thedifferenceinenergysavingsbetweenallfeedbackgroupscombinedandthecontrolgroupwasonlymarginallysigni“cant.Computerizedfeedbackappearedtoberelativelysuccessful,thenumberofconserversinthisgroupbeingsigni“cantlyhigherthanthenumberofnonconservers.Highandmediumconsumerssavedenergy(3.7%and2.5%,respectively)whereaslowconsumersincreasedenergyuse(by10.7%),whichcorroboratesasimilar“ndingbyBittleetal.(1979…1980).Environmentalattitudesandbeliefsweremarginallysigni“cantpredictorsofenergysavings.Theauthorsindicatethatalownumberofhouseholdsperconditionandlargewithin-groupvariancesmayhavereducedthestatisticalpoweroftheComparativefeedbackwasalsopartoftheso-calledEcoTeamProgram(ETP).EcoTeamsaresmallgroups(e.g.neighbors,friends,family)whocometogetheronceeverymonthtoexchangeinformationaboutenergy-savingoptions.Theyalsoreceivefeedbackaboutownenergysavings,andsavingsofotherEcoTeams.Staats,Harland,andWilke(2004)evaluatedtheETPintheNetherlands,targetingvariousbehaviorsrelatednotonlytogasandelectricityuse,butalsotowateruse,transportation,foodconsumptionandwastemanagement.Acomparisongroupwasusedforasubsetofeightenergy-relatedbehaviors.RepeatedmeasuresanalysisrevealedthatETPparticipantsincreasedthefrequencyofpro-environmentalbehaviorsovertime,whereasthecomparisongroupdidnot.Aftertheprogram,ETPhouseholdshadsaved20.5%ongasuse,4.6%onelectricityuse,2.8%onwateruse,andhadreducedtheirwasteby28.5%.Twoyearslater,thesesavingswere16.9%forgasuse,7.6%forelectricityuse,6.7%forwateruse,and32.1%forwastereduction.ETPseemstobeapromisinginterventioninthatitprovedtobesuccessfulinreducingenergyuseinseveraldomains,bothshortlyaftertheprogramandduringafollow-uptwoyearslater.Sinceacombinationofinterventionswasused,itisdif“culttoattributeitssuccesstocomparativefeedback.Also,respondentsinthisstudypresumablywerehighlymotivatedparticipants,makingitdif“culttogeneralizeits5.2.RewardsMonetaryrewardsmayserveasanextrinsicmotivatortoconserveenergy.Rewardscaneitherbecontingentontheamountofenergysaved,ora“xedamount(e.g.whenacertainpercentageisattained).HayesandCone(1977)testedtheeffectofrewards,feedbackandinformationonelectricityuse.Theimple-mentationoftheinterventionswasdonesequentially,inamultiplebaselinedesign.Allparticipatinghouseholdsreducedelectricityconsumption.However,basedonasamplesizeoffourandwithoutreportinganystatisticaltests,theseresultscannotbegeneralized.Winett,Kagel,Battalio,andWinkler(1978)studiedtheeffectofhighversuslowmonetaryrewardsincombinationwithfeedbackandinformation.Duringthe“rstfourweeksoftheintervention,householdsinbothhighandlowrewardgroupssavedmoreenergythantheothergroups.Duringthesecondpartoftheinterventionperiod,householdswhoinitiallyhadonlyreceivedin-formationwerenowgivenahighreward,resultinginsavingsof7.6%.Overaperiodofeightweeks,householdswhohadreceivedahighreward,feedbackandinformationreducedelectricityusebyabout12%.Itisnotclearwhetherthiseffectwasmaintainedafterdiscontinuationoftherewards.Inastudyfocusingonhouseholdsinmaster-meteredapartments(i.e.whodonothaveowngasmeters),contestswereheldwitharewardfortheapartmentblockabletosavemost(McClelland&Cook,1980).Contendinggroupsreceivedweeklyfeedbackongassavingsoftheownaswellastheothergroups,andinformationonhowtosaveenergy.Thecontestgroupsused6.6%lesselectricitythanacontrolgroupofmaster-meteredapartments.However,thesavingsdecreasedasthetreatmentperiodprogressed,suggestingashort-termeffectofrewards.Rewardswerealsousedintwootherstudiesconductedwithmaster-meteredapartments(Slavin,Wodanski,&Blackburn,1981).The“rststudyinvestigatedthecom-binedeffectofinformation,prompts(reminders),bi-weeklyfeedback(abouttheperformanceoftheentiregroup)andrewards(100%ofthevalueofelectricitysavings).Allparticipantsreceivedthesamecombinationofinterventions.Theinterventionlasted14weeksforgroup1(savingsof11.2%),12weeksforgroup2(1.7%)and8weeksforgroup3(4%),andresultedinaveragesavingsof6.2%relativetobaseline.Theeffectsappearedtobestrongestimmediatelyfollowingimplementationoftheintervention.Theirsecondstudywassetupalongthesamelines(Slavinetal.,1981,Study2).Insteadofreceivingthefullamount,participantsnowreceived50%oftheARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 monetaryvalueofelectricitysavings,andabonusamountwasgiveniftotalgroupsavingsexceeded10%.Thecombinationofinterventionsresultedinelectricitysavingsof9.5%(group1),4.7%(group2),and8.3%(group3),withanaverageof6.9%.Incontrasttostudy1,theeffectsdidnotdeclineduringthetreatmentperiod,whichmaybeattributedtotheextrabonus.Duetothedesignofbothstudies,itisnotpossibletodifferentiatebetweentheeffectsofthevariouscomponentsoftheintervention.PittsandWittenbach(1981)evaluatedtheeffectoftaxcreditsonconsumersdecisionstoinsulatetheirhomes.Thecreditconsistedofadeductionfromtotalincometaxes,whichwasgivenafterhouseholdshadinstalledinsulation.Atelephonesurvey,conducted2yearsafterthetaxcredithadcomeintoeffect,revealedthatthecredithadhadnoeffectonthedecisionwhetherornottoinstallinsulation.6.Consequenceinterventions:conclusionsFeedbackappearstobeaneffectivestrategyforreducinghouseholdenergyuseinmoststudiesreviewedhere(e.g.Seligman&Darley,1977),althoughsomeexceptionsexistKatzevetal.,1980…1981).Resultsofstudiesusingfeedbackseemtosuggestthatthemorefrequentthefeedbackisgiven,themoreeffectiveitis.Positiveeffectshaveforinstancebeenfoundforcontinuousfeedback(e.g.McClelland&Cook,1979…1980).Threestudiesfounddifferentialeffectsforhighandlowconsumersofenergy,thelattergroupincreasingtheirenergyuseasaresultoffeedback(Brandon&Lewis,1999Bittleetal.,1979…1980VanHouwelingen&VanRaaij,1989Kantolaetal.showedthathighfrequencyisnotnecessarilythekeytosuccess:bygivingfeedbackevokingcognitivedissonanceonesingletime,householdssigni“cantlyre-ducedenergyuse.Itisnotclearwhetheritmakesadifferencetogivefeedbackintermsofmonetaryratherthanenvironmentalcosts,sincestudiesinvestigatingthisdifferencedidnot“ndany(e.g.Brandon&Lewis,1999Bittleetal.,1979…1980).Studiesusingcomparativefeed-back(e.g.Brandon&Lewis,1999;Middenetal.,1989)didnot“ndittobemoreeffectivethanindividualfeedback.Combiningcomparativefeedbackwithrewardsinacontestsettingprovedtobesuccessful(McClelland&Cook,1979…1980EcoTeams,whoreceivebothindividualandcomparativefeedback,weresuccessfulinreducingenergyuse,alsointhelongrun.Combiningfeedbackwithgoalsettingresultedinreductionsinenergyconsumption(McCalley&Midden,),especiallywhencombinedwithadif“cultgoalBecker,1978).Studieswhoexaminedtheeffectofgivingfeedbackaboutthepricedifferencebetweenon-andoff-peakhoursfoundthistoresultinshiftsinconsumptiontooff-peakhours,butnodifferenceinoverallconsumptionwasfoundorreported(Heberlein&Warriner,1983Sextonetal.,1987Overall,rewardsseemtohaveapositiveeffectonenergysavings:allstudiesreviewedherereportsigni“cantdifferencesbetweenhouseholdswhohadreceivedarewardandthosewhohadnot(e.g.Winettetal.,1978).Resultsofseveralstudies(McClelland&Cook,1980Slavinetal.,)dohoweversuggestthattheeffectofrewardsisrathershort-lived.PittsandWittenbach(1981)foundthatgovernmentalin”uenceintheformoftaxcreditswasnotadecisivefactorinconsumersdecisionstobuyandinstallin-homeinsulation.Severalstudiesmayhavefailedto“ndanystatisticallysigni“canteffectsduetoarelativelylownumberofhouseholdsperexperimentalconditionand/orlargewithin-groupvarianceinenergyuse(e.g.Brandon&Lewis,1999).Also,insomecases,itisdif“culttomakegeneralizationsbasedontheresults,forsamplesconsistedofhighlymotivatedparticipants(e.g.Staatsetal.,2004Anotherissueconcernsconfoundingofeffects:duetotheuseofcombinationsofinterventions,itisdif“culttoestablishthecontributionofeachinterventionseparately.7.DiscussionInterventionstopromoteenergyconservationamonghouseholdshavebeenemployedwithvaryingdegreesofsuccess.Theantecedentinterventionscommitmentandgoalsettingappearedsuccessfulinbringingaboutchangesinenergyuse,especiallywhenusedincombinationwithotherinterventions(e.g.Becker,1978).Generally,informa-tionaloneisnotaveryeffectivestrategy(e.g.Houwelingen&VanRaaij,1989).Informationaboutenergyproblemsasconveyedbymassmediacampaignstendstoresultinincreasesofknowledgeandof(self-reported)conservationbehaviors,butlittleisknownabouttheeffectsonactualenergyuse(e.g.Staatsetal.,1996However,energysavingswereachievedbygivinghouse-holdstailoredinformationthroughhomeenergyauditsWinettetal.,1982…1983Asforconsequenceinterventions,rewardsareeffective,butthereissomeindicationofthispositiveeffectdisappearingassoonastheinterventionisdiscontinued(seealsoDwyeretal.,1993Geller,2002).Providinghouseholdswithfeedback,andespeciallyfrequentfeed-back,hasproventobeasuccessfulinterventionforreducingenergyconsumption(e.g.Seligman&Darley,1981).However,exceptionsexist(e.g.Katzevetal.,1980…1981).Somestudiesfoundadifferentialeffectforhighandlowconsumersofenergy,theformerreducingenergyuseandthelatterincreasingenergyuseasaresultoffeedback(e.g.Bittleetal.,1979…1980).Thisisanimportant“ndingfromapolicyperspective,inthesensethatpoliciesaimingtoreduceenergyusemayespeciallywanttotargethighusersofenergy,becauseofahigherenergy-savingThestudiesdiscussedhererevealthatunderlyingdeterminantsofenergyuseandenergy-relatedbehaviorshavehardlybeenexamined.Insomecases,determinantsofenergyuseorenergysavingsweremeasured,anditARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 appearedthatattitudeandknowledgearegenerallypositivelyrelatedtoenergysavings(e.g.Brandon&Lewis,1999Heberlein&Warriner,1983).Studiesusingbeforeandaftermeasurementshavefoundanincreaseinknowledgelevelsaftermassmediacampaigns(e.g.Staatsetal.,1989),andworkshopsaboutenergyconservationGeller,1981),butthisdidnotnecessarilyresultinbehavioralchangesorreductionsinenergyuse.Winettetal.(1985)foundmodelingtoresultinsigni“cantenergysavingsandhigherknowledgelevels.AnotherstudyHuttonetal.,1986)foundabehavioraleffectbutnolearningeffectinonesampleofparticipants,whereasintheothersamplealearningeffectbutnobehavioraleffectwasAnumberofcriticalremarkscanbemadewithrespecttotheinterventionstudiesdiscussedhere.The“rstissueconcernsthefactthatinsomestudies,theexactcontentoftheinterventionwasnotclearlyspeci“ed.Forinstance,insomecasesitisunclearwhetherinformationwasprovidedaboutenergyproblems,aboutenergy-savingmeasures,orboth.Also,authorsdidnotalwaysspecifywhichbehaviorsweretargetedbytheintervention(e.g.ef“ciencyorcurtailmentbehaviors).Thesespeci“cationscanbeadecisivefactorinevaluatinganinterventions(in)effective-ness.Inaddition,interventionsarenotalwaysexplicitlymentioned.Mainconclusionsappeartobefocusedontheeffectivenessofoneintervention(e.g.feedback),butwhenreadingthestudydesign,otherinterventions(e.g.informa-tion)appeartohavebeenusedaswell.Manystudieshaveshownthatacombinationofstrategiesisgenerallymoreeffectivethanapplyingonesinglestrategy.However,confoundingofeffectsmakesitmoredif“culttodeterminewhichstrategiesactuallycontributedtotheoveralleffect.Moresystematicresearchontheeffectivenessofinterven-tionsundervariouscircumstanceswouldbeadvisableinthisrespect.Thismaywellinpartbedoneinexperimental(laboratory)studies.Thesecondtypeofproblemsemergingfromthisreviewconcernsmethodologicalissues.Verysmallsamplesizes,especiallyinconjunctionwithlargewithin-groupvariances(inenergyuse)mayhavereducedthestatisticalpowerofdesignsandconsequently,thesestudiesmayhavefailedto“ndanystatisticallysigni“canteffects.Moreover,house-holdswhoparticipateinthistypeofstudiestendtobehighlymotivated,tendtohavehigherthanaverageincomes,andhigherthanaverageeducationlevels,makinggeneralizationsbasedonthesestudiesratherdif“cult.Toillustrate,astudyusingasampleoffourhouseholdsclaimstohaveestablishedunequivocallytheindependenteffec-tivenessofpaymentsandtohavedemonstratedtherelativesuperiorityofsuchaprocedureoverfeedbackandinformation(Hayes&Cone,1977,p.433),withoutperforminganykindofstatisticaltesttosubstantiatethisassertion.Itishardlywarrantedtogeneralizeresultsbasedonsamplessizesthissmall.Interestingly,thisstudyhasbeencitedinotherarticlesassupportfortheeffectivenessofrewards.Thethirdtypeofproblemsisrelatedtothesizeoftheeffectsfoundinthestudies.First,effectsizeswerenotveryhigh:Cohensdwasfoundtorangefrom0.07to1.41.Moststudiesdidnotreportsuf“cientstatisticalinforma-tionneededtocalculateeffectsizes(viz.,meansandstandarddeviationsforexperimentalandcontrolgroups).Itwouldbeadvisableforauthorstoprovidethesedatasoastoenableathoroughmeta-analysis.Moreinsightintoeffectsizesmayalsoserveasvaluableinputforpoliciesaimingtoreducehouseholdenergyuse.Second,differentindicatorsareusedtotesttheeffectivenessofinterventions.Anumberofstudiesreportaneffectoftheinterventionbasedonchangesinself-reportedenergy-relatedbehaviors.Itisimportanttoexamineactualenergyuseaswell,asbehavioralchangesdonotnecessarilyresultinenergysavings.Besides,itmaywellbethatself-reportedbehaviorswerein”uencedbysocialdesirability.Forinstance,Luybenfoundself-reportedthermostatsettingstobesigni“cantlylowerthanthoseobservedbyinterviewers.Warriner,McDougallandClaxton(1984)comparedobservedenergy-relatedbehaviorswithself-reportedmeasuresandfoundnosigni“cantdifferencesbetweenthem.Assumingthatself-reportedbehaviorsdore”ectreality,itstilldoesnotbecomeclearwhethertheinterventionhadanyimpactonactualenergyuse(e.g.aso-calledreboundeffectmayhaveoccurred,seeBerkhoutetal.,2000).Householdsmayhavespentthemoneytheyinitiallysavedbyreducingenergyuseonenergy-intensiveproducts,therebyincreasingoverallenergyuse.Or,reductionsinenergy-relatedbehaviorsmayhaveoccurred,whichwerenotmonitoredbecausetheywerenottargetedbytheintervention(spill-overeffect).Inaddition,fewstudiesreportontheactualimpactofenergysavings.Toillustrate,areductionof10%basedonatotalenergyuseof1000MJisnotthesameasareductionof10%basedonatotalenergyuseof10,000MJ.Thisway,itdoesnotbecomecleartowhatextentchangesinbehaviorresultedinenergyreductions.Lastly,asmentionedelsewhere(seeDeYoung,1993Dwyeretal.,1993),relativelylittleisknownaboutthelong-termeffectsofinterventions.Amajorityofthestudiesdidnotmonitortheeffectsoftheinterventionsoverlongerperiodsoftime.Consequently,itisnotclearwhetherbehavioralchangesweremaintainedandwhethernew(energy-saving)habitswereformed,orwhetherenergyusereturnedtobaselinelevels.Whenafollow-upwasincluded,oftenitappearedthatthepositiveeffectsoftheinterven-tionwerenotmaintained.Thereareexceptionstothis,withsomestudies(Pallak&Cummings,1976Staatsetal.,)reportingpromisinglong-termeffects.8.RecommendationsManyenvironmentalproblems,suchasenergyuse,arerelatedtohumanbehaviorand,consequently,maybereducedthroughbehavioralchanges.Comparingpreviousreviewsoninterventionsaimedatchangingenergy-relatedARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 behaviors(e.g.Dwyeretal.,1993Schultzetal.,1995)tothecurrentonerevealssimilaritiesanddifferences.Thesepreviousreviewshadalreadypointedoutissuesinordertoimproveinterventionstudies,suchastheinclusionoflong-termmeasurementsandtheuseofstudydesignsprecludingaconfoundingofeffectswhenusingmultipleinterventions.Studiesthathavebeenpublishedsincethenhaveprovidedadditionalinsightintoeffectiveinterventionplanningbyaddressingtheseissuesofconcern.AstudybyStaatsetal.(2004)examinedandfoundlong-termeffectsofinterventionsonbothenergyuseandtheadoptionofenergy-savingbehaviors.Laboratoryresearchhasextendedourknowledgeoftheseparateandcombinedeffectofinterventions(McCalley&Midden,).Inaddition,newmethodsofapproachinghouse-holdshavebeenused,suchastextTV(llink&Meertens,)andcomputerizedfeedback(Brandon&Lewis,),bothshowingpromisingeffectsintermsofenergyreductions.Takentogether,thesenewstudieshaveaddedtoourunderstandingofhowtoencouragehouseholdenergyconservation.Thepresentreviewalsocomestodifferentconclusions.AsopposedtoDwyeretal.s(1993)recommendationtouseantecedentinterventionsonly,wefoundthatsingleantecedentinterventionsarenotveryeffective.Rather,wefoundanantecedentinterventionseffectiveness(e.g.goalsetting)toincreasewhencombinedwithconsequencestrategies(e.g.feedback,seeBecker,1978McCalley&Midden,2002VanHouwelingen&VanRaaij,1989).Inthisconcludingsection,severaladditionalguidelinesareproposedtohelpresearchersandpolicymakerseffectivelydesign,implementandevaluateinterventionprogramstoreducehouseholdenergyuseinthefuture.Animportant“rststepindesigningandimplementinginterventionsaimedatreducingenergyuseamonghouse-holdsisathoroughproblemdiagnosis(Geller,2002).First,byidentifyingbehaviorsthatsigni“cantlycontributetoenvironmentalproblems,andsecond,byexaminingfactorsthatmakesustainablebehaviorpatterns(un)attractive,suchasmotivationalfactors(e.g.attitudes),opportunities,andperceivedabilities.Itisimportantthatinterventionsaddressandchangepossiblebarrierstobehavioralchange(seealsoGardner&Stern,2002).Therefore,aproblemdiagnosisisnecessaryinexaminingwhichbehaviorsandwhichbehavioraldeterminantsshouldbetargetedbytheintervention.Forexample,“nancialincentiveswillbeeffectiveonlywhenpeopleinfacttakepricesintoconsiderationwhenmakingchoicesandeducationalcampaignsmayespeciallybeadvisablewhenpeopleareunawareofenergyuseandenvironmentalproblemsassociatedwiththeirbehavior.Intermsofreducingenvironmentalimpact,itisimportanttoidentifytargetbehaviorsthathavearelativelylargeenergy-savingpotential.Bykeepingenvironmentalgoalsinmind,researchersandinterventionplannerscanfocusonbehaviorsandtargetgroupsthatsigni“cantlyin”uenceenvironmentalqualities.Thefollowingrecommendationisrelatedtotheobserva-tionthatinterventionsstudiestypicallyhaveamono-disciplinaryfocus.Interventionstudiesfromapsychologi-calperspectivetendtofocuspredominantlyonchanging(individual-level)MOA-variables(e.g.attitudes,abilities).Itisequallyimportanttotargetmacro-levelfactorscontributingtohouseholdenergyuse,suchasdemographicorsocietaldevelopments(e.g.TEDICfactors),whichshapethephysicalinfrastructureandtechnicalapparatusthatconditionbehavioralchoicesandenergyuseassociatedwiththesechoices.Itisthereforeimportanttoconsiderhouseholdenergyconservationfromamultidisciplinaryperspective.Forinstance,sociologistscanprovidemoreinsightintomacro-levelfactorsthatshapehouseholdenergyuse.Also,inputfromenvironmentalscientistscanbeofvaluableimportancetofurtherimproveinterventionstudies.Theenvironmentalsciencescanhelptranslateenergy-relatedbehaviorsofhouseholdsintotheirenviron-mentalimpact,e.g.intermsofCOemissions,andhelpselecthigh-impactbehaviors.Finally,evaluationsofaninterventionseffectivenessshouldbefocusedon(changesin)behavioraldeterminantsaswellas(changesin)energy-relatedbehaviors.Moststudiesrevealonlytowhatextentinterventionshavebeensuccessful,withoutprovidinginsightintothereasonswhy.Forinstance,failureofamassmediacampaigntochangebehaviormaywellbeattributabletothefactthattargetgroupswerealreadyfamiliarwiththeinformationpro-vided.Inotherwords,theeffectivenessofinterventionsandpossibledeterminantsofbehaviorshouldbeexaminedsimultaneously.Athoroughmonitoringofdeterminantsofenergyuseandenergysavingsmayincreasesourunder-standingofthesuccessorfailureofinterventionprograms.TheguidelinesproposedheremayhelpresearchersandpolicymakersdesignandimplementeffectiveinterventionprogramstoencourageamoresustainablebehaviorARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESS TableA1Overviewofinterventionstudiesonhouseholdenergyconservation(includingauthor(s),typeofintervention,designandnumberofgroups,totalsamplesize,targetbehavior,behavioraldeterminants,duration,effectsduringtheintervention,effectsizesandlong-termeffect)Author(s)Intervention(s)DesignTargetbehaviorDurationEffectduringinterventionEffectsizeLong-termeffectBecker(1978)Goalsetting20%goal,feedbackperweek2%goal,feedbackperweek20%goal2%goal100ElectricityuseNotmeasured1month(1)20%-feedback:15.1%(2)2%-feedback:5.7%(3)(4)NotmeasuredBittleetal.(1979)(1)FeedbackDailyfeedback(costs)30ElectricityuseNotmeasured42daysFeedbackgroupreducedelectricityuseby4%,comparedtobaseline,andconservedmorethanthecontrolgroup.24-dayreversalExperimentalgroupnolongerreceivedfeedback;stillusedlesselectricitythancontrolgroup,nowreceivingfeedback.Bittleetal.(1)FeedbackCumulativefeedbackCumulativefeedbackDailyusefeedbackDailyusefeedback353ElectricityuseNotmeasured35daysForhighconsumersofelectricity,allfourtypesoffeedbackresultedinalowerrateofincrease,butformediumandlowconsumersofelectricityitresultedinanincreaseinNotmeasuredBrandonandLewis(1)FeedbackIndividualfeedbackCostfeedbackLea”etfeedbackComputerfeedback120GasandelectricityuseOfsavingsAttitudes(+)OfenergyuseAge(+)Size(+)Income(+)2months(1)Comparative:4.6%(2)(3)Cost:4.8%(4)(5)Lea”et:0.4%(6)Computerized:4.3%(7)NotmeasuredMarginallysigni“cantdifferencebetweenfeedbackgroupscombinedandcontrol.Feedbackvs.noGeller(1981)(1)Information117Electricity,gasandwateruse(C,E)Attitudes(+)Knowledge(+)Intention(+)3hoursTheworkshopresultedinanincreaseinlevelsof6–12weeksafterworkshopNobehavioraleffectwasW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESSGonzalesetal.(1)(2)Information(trainedauditors),rebate(nontrainedauditors),408Gasandelectricityuse(C,E)Notmeasured1–2weeksafterauditHouseholdsintrained-auditgroupreportedagreaterlikelihoodoffollowingthroughon4monthsafterauditHouseholdsintrained-auditgrouphadfollowedrecommendationsmoreoften,butnodifferenceinenergyconsumption.HayesandCone(1)(2)(3)Multiplebaselinedesign:interventionssequentially4ElectricityuseNotmeasured91daysAllhouseholdsreducedelectricityconsumption,comparedtobaseline.NotmeasuredHayesandCone(1)FeedbackMonthlyfeedback40ElectricityuseNotmeasured4monthsFeedbackgroup:4.7%Controlgroup:2-monthfollow-upControl:0.3%HeberleinandWarriner(1983)(1)FeedbackMonthlyfeedback(priceratio)600ElectricityuseOfon-peakelectricityuseKnowledge(commitment(3yearsLargerpricedifferencesbetweenon-peakandoff-peakperiodsresultedinlargerreductionsofon-peakelectricityuse.NotmeasuredHirstandGrady(1)Information850Gasuse(C,E)OfgasuseIncome(+)Attitudes(+)Oneyearafterhomevisits:gassavingsof2%,comparedtocontrol2yearsafterauditGassavingsof4%,relativetocontrol.HuttonandMcNeill(1981)(1)InformationInformation(mediacampaign),shower”owdevice1811Gas,electricityandwateruse(C,E)Notmeasured1monthExperimentalgroupadoptedmoreenergysavingtipsthanthecontrolgroup.NodatareportedonactualenergyNotmeasuredHuttonetal.(1986)(1)(2)300GasandelectricityuseKnowledge(+)(onlyinUScity)groupandinformationonlygroupconservedmoreenergythancontrols(butonlyinCanadianNotreportedKantolaetal.(1)(2)Dissonancefeedback,118ElectricityuseElectricityusePersonaldutytosave(n.s.)Importanceofconservation(n.s.)4weeksThecognitivedissonancegroupsavedsigni“cantlymoreelectricitythantheothergroups.Forthesecondtwoweeks,thisgrouponlydifferedfromNotmeasuredKatzevetal.(1)FeedbackDailyfeedbackFeedbackevery3rd44ElectricityuseNotmeasured2weeksNosigni“cantdifferencesbetweenexperimentalgroupsandcontrolgroup.2-weekfollow-upNosigni“cantdifferences.W.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESS TableA1(Author(s)Intervention(s)DesignTargetbehaviorDurationEffectduringinterventionEffectsizeLong-termeffectKatzevandJohnson(1983)(1)(2)Bothrequests66ElectricityuseNotmeasured4weeksNosigni“cantdifferencesbetweengroups.12-weekfollow-upExperimentalgroupsconservedmoreelectricitythancontrolgroup.KatzevandJohnson(1984)(1)(2)(3)BothrequestsBothrequests,90ElectricityuseNotmeasured2weeksThecommitmentonlyandthegroupreceivingallinterventionsconservedmoreelectricitythantheothergroups(butonlyin“rstweek).2monthsfollow-upNosigni“cantdifferencesbetweenthegroups.Luyben(1982)(1)Information(televisedplea)Noplea210Lowerthermostatsettingsto65Post-testonlyNodifferenceinbetweengroups.ThreedaysfollowingpleaNodifferenceinthermostatsettingsbetweenthosewhohadandhadnotheardtheNotmeasuredMcCalleyandMidden(2002)(1)(2)GoalsettingFeedback,self-setFeedback,assigned100Doinglaundry(loadandtemp.respondentssavedmoreenergywithself-setgoalandlesswithassigned20washingFeedbackcombinedwithgoalsettingwasmoreeffectivethanfeedbackalone.Participantswithaself-setgoalsaved21.9%,thosewithanassignedgoalsaved19.5%.NotmeasuredMcClellandandCook(1979…1980)(1)FeedbackContinuousfeedback101ElectricityuseNotmeasured11monthsContinuousfeedbackresultedinaveragesavingsof12%,comparedtoNotmeasuredMcClellandandCook(1980)(1)(2)(3)Reward,feedback,500GasuseNotmeasured12weeksOnaverage,6.6%gaswassavedbythecontestNotmeasuredMcDougalletal.(1)InformationNoinformation1451Variousbehaviorsrelatedtoheating(C&E)NotmeasuredNotreportedAfter2yearsNodifferenceinnumberofenergysavingactivities.W.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESSMcMakinetal.(1)Information(1)Information1231Gasandelectricityuse(relatedtoNotmeasured1yearHouseholdssaved10%energycomparedtoNotmeasuredMcMakinetal.(1)Information(1)Information175Electricityuse(relatedtoNotmeasured4monthsHouseholdsused2%moreelectricity,comparedtobaseline.NotmeasuredMiddenetal.(1)(2)(3)Individualfeedback,feedback,informationfeedback,information91GasandelectricityuseElectricityuseAttitude(+)GasuseAttitude(+)Attitude(n.s.)12weeksElectricity18.8%Gas18.4%Electricity18.4%Gas5.8%Electricity19.4%Gas17.5%Electricity7.6%Gas0%Electricity5.6%Gas11.6%NotmeasuredPallakandCummings(1976)(1)CommitmentPubliccommitmentPrivatecommitmentGasandelectricityuseNotmeasured1monthPubliccommitmentconditionshowedalowerrateofincreaseingasandelectricityusethanprivatecommitmentorcontrol.6monthsfollow-upThiseffectwasPittsandWittenbach(1981)(1)Financial(taxcredit)(1)Incentive146BuyinghomeNotmeasuredNotmeasured2yearsaftertaxTaxcredithadnoeffectoninsulationpurchaseSeligmanandDarley(1977)(1)Feedback40ElectricityuseNotmeasured1monthFeedbackgroupused10.5%lesselectricitythan0.85NotmeasuredSextonetal.(1987)(1)Feedback600ElectricityuseNotmeasured22monthsElectricityconsumptionshiftedtooffpeakhours,buttotalconsumptiondidnotdecrease.NotmeasuredSlavinetal.(1981)Study1(1)(2)(3)(4)(14weeks)(12weeks)(8weeks)166ElectricityuseNotmeasured8…14weeksCombinedinterventionsresultedinsavingsof11.2%(group1),1.7%(group2),and4%(group3),andanaverageofNotmeasuredSlavinetal.(1981)Study2(1)(2)(3)(4)(5)Goalsetting(14weeks)(11weeks)(8weeks)255ElectricityuseNotmeasured8…14weeksCombinedinterventionsresultedinsavingsof9.5%(group1),4.7%(group2),and8.3%(group3),anaverageofNotmeasuredW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESS TableA1(Author(s)Intervention(s)DesignTargetbehaviorDurationEffectduringinterventionEffectsizeLong-termeffectStaatsetal.(1996)(1)Information(massmediaN/A704Willingnesstoshowpro-Knowledge(+)OfbehaviorKnowledge(n.s.)awareness(n.s.)AftermediacampaignSlightincreaseinwillingnesstoshowpro-environmentalbehaviors,butonlyforthosewhoalreadyactedpro-NotmeasuredStaatsetal.(2004)(1)(2)(3)individual&comparativefeedback150Gas,water,electricityuse,waste,food,Reducecaruse5kmIntention(+)Habit(8monthsGasuse:20.5%Electricityuse:4.6%Wateruse:2.8%Waste:32.1%After2yearsGasuse:16.9%Electricityuse7.6%Wateruse:6.7%Waste:32.1%VanHouwelingenandVanRaaij(1)(2)Goalsetting(3)(4)Continuousfeedback,goalsetting,Monthlyfeedback,goalsetting,Monitoring,goalsetting,informationGoalsetting,285GasuseNotmeasured1yearContinuousfeedback:Monthlyfeedback:Self-monitoring:5.1%Information:4.3%Control:0.3%Allgroupssigni“cantlyreducedgasusecomparedtobaselinelevelsandcomparedtocontrol(1)vs.(2):(1)vs.(3):After1yearGasuseincreasedforallgroups,comparedtobaseline;differencebetweengroupsllinkandMeertens(1999)(1)(2)Goalsetting(3)Feedback,goalsetting,information48Gas,electricityandwateruseNotmeasured5monthsExperimentalgroupused18%lesswater,23%lessgasand15%lesselectricitythancontrol.NotmeasuredWinettetal.(1978)(1)(2)(3)feedback,highrewardfeedback,lowreward129ElectricityuseNotmeasured8weeksFirst4weeks:Highreward:3.5%Lowreward:4.5%Control:0.9%After8weeks:highrewardgroupsaved12%,andinformationonly(nowonhighrewards)saved7.6%.NotmeasuredW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ARTICLEINPRESS Winettetal.(1979)(1)(2)(3)(4)Goalsettinginformation,goalinformation,goal71Electricityuse(C&E)Notmeasured1monthFeedbackgroupreducedelectricityuseby13%andtheself-monitoringgroupby7%.10-weekfollow-upEffectwasmaintained.Winetttetal.(1)Information51Electricityuse(waterheating,(C&E)Notmeasured1monthAftertheaudit,householdsreducedelectricityuseby21%,relativetothecontrolNotmeasuredWinettetal.(1985)(1)(2)150GasandelectricityuseKnowledge(+)OfenergysavingsKnowledge(n.s.)5weeksExposuretoTVprogramresultedinelectricitysavingsof10%.After1yearEffectwasnotmaintainedACreferstocurtailmentbehaviors,anEtoef“ciencybehaviors.Somestudiesdidnotreportwhichtypeofbehaviorwastargeted.Withrespecttothemeasuredchangesindeterminantsofenergyuseenergysavings,therearetwopossibilities:(1)Measurementofrelationshipsbetweendeterminantsandenergyuse/energysavingsbymeansofaregressionanalysis:(+)apositiverelationshipbetweendeterminantandenergyuse/savings;)anegativerelationshipbetweendeterminantandenergyuse/savings;(n.s.)norelationshipbetweendeterminantandenergyuse/savings.(2)Measurementofchangesinlevelsofdeterminantsasaresultsoftheinterventionbymeansofpretest/post-testmeasurements:(+)anincreaseofthelevelofadeterminantwasobserved;)adecreaseofthelevelofadeterminantwasobserved;(n.s.)nochangewasobservedwithrespecttothelevelofadeterminant.W.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 ReferencesBandura,A.(1977).Sociallearningtheory.NewYork:Prentice-Hall.Becker,L.J.(1978).Jointeffectoffeedbackandgoalsettingonperformance:A“eldstudyofresidentialenergyconservation.ofAppliedPsychology(4),428…433.Becker,L.J.,Seligman,C.,Fazio,R.H.,&Darley,J.M.(1981).Relatingattitudestoresidentialenergyuse.EnvironmentandBehavior590…609.Berkhout,P.H.G.,Muskens,J.C.,&Velthuijsen,J.W.(2000).De“ningthereboundeffect.EnergyPolicy(6/7),425…432.Biesiot,W.,&Noorman,K.J.(1999).Energyrequirementsofhouseholdconsumption:AcasestudyofTheNetherlands.EcologicalEconomics,367…383.Bittle,R.G.,Valesano,R.,&Thaler,G.(1979).Theeffectsofdailycostfeedbackonresidentialelectricityconsumption.BehaviorModication(2),187…202.Bittle,R.G.,Valesano,R.M.,&Thaler,G.M.(1979…1980).Theeffectsofdailyfeedbackonresidentialelectricityusageasafunctionofusagelevelandtypeoffeedbackinformation.JournalofEnvironmental,275…287.Black,J.S.,Stern,P.C.,&Elworth,J.T.(1985).Personalandcontextualin”uencesonhouseholdenergyadaptations.JournalofApplied(1),3…21.Brandon,G.,&Lewis,A.(1999).Reducinghouseholdenergyconsump-tion:Aqualitativeandquantitative“eldstudy.JournalofEnviron-mentalPsychology,75…85.Cook,S.W.,&Berrenberg,J.L.(1981).Approachestoencouragingconservationbehavior:Areviewandconceptualframework.JournalofSocialIssues(2),73…107.DeYoung,R.(1993).Changingbehaviorandmakingitstick.Theconceptualizationandmanagementofconservationbehavior.EnvironmentandBehavior(4),485…505.Dwyer,W.O.,Leeming,F.C.,Cobern,M.K.,Porter,B.E.,&Jackson,J.M.(1993).Criticalreviewofbehavioralinterventionstopreservetheenvironment.Researchsince1980.EnvironmentandBehavior275…321.Gardner,G.T.,&Stern,P.C.(2002).EnvironmentalProblemsandHuman.Boston:Pearson.rling,T.,Eek,D.,Loukopoulos,P.,Fujii,S.,Johansson-Stenman,O.,Kitamura,R.,Pendyala,R.,&Vilhelmson,B.(2002).AconceptualanalysisoftheimpactoftraveldemandmanagementonprivatecarTransportPolicy,59…70.Gatersleben,B.,&Vlek,Ch.(1998).Householdconsumption,quality-of-lifeandenvironmentalimpacts:Apsychologicalperspectiveandempiricalstudy.InK.J.Noorman,&A.J.M.SchootUiterkampGreenhouseholds?Domesticconsumers,environment,and(pp.141…179).London:Earthscan.Geller,E.S.(1981).Evaluatingenergyconservationprograms:Isverbalreportenough?JournalofConsumerResearch,331…335.Geller,E.S.(2002).Thechallengeofincreasingproenvironmentbehavior.InR.G.Bechtel,&A.Churchman(Eds.),HandbookofEnvironmental(pp.525…540).NewYork:Wiley.Geller,E.S.,Berry,T.D.,Ludwig,T.D.,Evans,R.E.,Gilmore,M.R.,&Clark,S.W.(1990).AconceptualframeworkfordevelopingandevaluatingbehaviorchangeinterventionsforinjuryHealthEducationResearch:TheoryandPractice125…137.Glass,G.V.,McGraw,B.,&Smith,M.L.(1981).Meta-analysisinsocial.BeverleyHills,CA:SagePublications.Gonzales,M.H.,Aronson,E.,&Costanzo,M.A.(1988).Usingsocialcognitionandpersuasiontopromoteenergyconservation:Aquasi-experiment.JournalofAppliedSocialPsychology1049…1066.Hayes,S.C.,&Cone,J.D.(1977).Reducingresidentialelectricalenergyuse:Payments,information,andfeedback.JournalofAppliedBehavior,425…435.Hayes,S.C.,&Cone,J.D.(1981).Reductionofresidentialconsumptionofelectricitythroughsimplemonthlyfeedback.JournalofAppliedBehaviorAnalysis,81…88.Heberlein,T.A.,&Warriner,G.K.(1983).Thein”uenceofpriceandattitudeonshiftingresidentialelectricityconsumptionfromon-tooff-peakperiods.JournalofEconomicPsychology,107…130.Hirst,E.,&Grady,S.(1982…1983).EvaluationofaWisconsinutilityhomeenergyauditprogram.JournalofEnvironmentalSystemsHunter,J.E.,&Schmidt,F.L.(1990).Methodsofmeta-analysis.Correctingerrorandbiasinresearchndings.BeverlyHills,CA:SageHutton,R.B.,Mauser,G.A.,Filiatrault,P.,&Ahtola,O.T.(1986).Effectsofcost-relatedfeedbackonconsumerknowledgeandconsumptionbehavior:A“eldexperimentalapproach.JournalofConsumerResearch,327…336.Hutton,R.B.,&McNeill,D.L.(1981).Thevalueofincentivesinstimulatingenergyconservation.JournalofConsumerResearchKantola,S.J.,Syme,G.J.,&Campbell,N.A.(1984).Cognitivedissonanceandenergyconservation.JournalofAppliedPsychology(3),416…421.Katzev,R.,Cooper,L.,&Fisher,P.(1980…1981).Theeffectoffeedbackandsocialreinforcementonresidentialelectricityconsumption.JournalofEnvironmentalSystems(3),215…227.Katzev,R.D.,&Johnson,T.R.(1983).Asocial…psychologicalanalysisofresidentialelectricityconsumption:Theimpactofminimaljusti“cationJournalofEconomicPsychology,267…284.Katzev,R.D.,&Johnson,T.R.(1984).Comparingtheeffectsofmonetaryincentivesandfoot-in-the-doorstrategiesinpromotingresidentialelectricityconservation.JournalofAppliedSocialPsychol-(1),12…27.Kreuter,M.W.,Farrell,D.,Olevitch,L.,&Brennan,L.(1999).healthmessages:CustomizingcommunicationwithcomputertechnologyMahwah,NJ:LawrenceErlbaum.Luyben,P.D.(1982).Promptingthermostatsettingbehavior:Publicresponsetoapresidentialappealforconservation.Environmentand(1),113…128.McCalley,L.T.,&Midden,C.J.H.(2002).Energyconservationthroughproduct-integratedfeedback:Therolesofgoal-settingandsocialJournalofEconomicPsychology,589…603.McClelland,L.,&Cook,S.W.(1979…1980).Energyconservationeffectsofcontinuousin-homefeedbackinall-electrichomes.JournalofEnvironmentalSystems,169…173.McClelland,L.,&Cook,S.W.(1980).Promotingenergyconservationinmaster-meteredapartmentsthroughgroup“nancialincentives.ofAppliedPsychology(1),20…31.McDougall,G.H.G.,Claxton,J.D.,&Ritchie,J.R.B.(1982…1983).Residentialhomeaudits:AnempiricalanalysisoftheENEVERSAVEJournalofEnvironmentalSystems(3),265…278.McMakin,A.H.,Malone,E.L.,&Lundgren,R.E.(2002).Motivatingresidentstoconserveenergywithout“nancialincentives.EnvironmentandBehavior(6),848…863.MilieuCentraal(2005).http://www.milieucentraal.nl.ConsultedJanuary21,2005.Midden,C.J.H.,Meter,J.E.,Weenig,M.H.,&Zieverink,H.J.A.(1983).Usingfeedback,reinforcementandinformationtoreduceenergyconsumptioninhouseholds:A“eld-experiment.JournalofEconomicPsychology(1),65…86.lander,F.,&Thøgerson,J.(1995).Understandingconsumerbehaviourasaprerequisiteforenvironmentalprotection.JournalofConsumer,345…385.Pallak,M.S.,&Cummings,N.(1976).CommitmentandvoluntaryenergyPersonalityandSocialPsychologyBulletin(1),27…31.Pitts,R.E.,&Wittenbach,J.L.(1981).Taxcreditsasameansofin”uencingconsumerbehavior.JournalofConsumerResearchARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291 Schultz,P.W.,Oskamp,S.,&Mainieri,T.(1995).Whorecyclesandwhen?Areviewofpersonalandsituationalfactors.JournalofEnvironmentalPsychology,105…121.Seifert,T.L.(1991).DeterminingeffectsizesinvariousexperimentalEducationalandPsychologicalMeasurement,341…347.Seligman,C.,&Darley,J.M.(1977).Feedbackasameansofdecreasingresidentialenergyconsumption.JournalofAppliedPsychologySexton,R.J.,BrownJohnson,N.,&Konakayama,A.(1987).Consumerresponsetocontinuous-displayelectricity-usemonitorsinatime-of-usepricingexperiment.JournalofConsumerResearch,55…62.Siero,F.W.,Bakker,A.B.,Dekker,G.B.,&VandenBurg,M.T.C.(1996).Changingorganizationalenergyconsumptionbehaviorthroughcomparativefeedback.JournalofEnvironmentalPsychology,235…246.Slavin,R.E.,Wodanski,J.S.,&Blackburn,B.L.(1981).Agroupcontingencyforelectricityconservationinmaster-meteredapartments.JournalofAppliedBehaviorAnalysis(3),357…363.Staats,H.,Harland,P.,&Wilke,H.A.M.(2004).Effectingdurablechange.AteamapproachtoimproveenvironmentalbehaviorintheEnvironmentandBehavior(3),341…367.Staats,H.J.,Wit,A.P.,&Midden,C.Y.H.(1996).Communicatingthegreenhouseeffecttothepublic:Evaluationofamassmediacampaignfromasocialdilemmaperspective.JournalofEnvironmentalManage-,189…203.Stern,P.C.(1992).Whatpsychologyknowsaboutenergyconservation.AmericanPsychologist(10),1224…1232.USDepartmentofEnergy(2005).Energyinformationadministration:Carbondioxideemissions.RetrievedJanuary21,2005fromwww.eia.doe.gov/oiaf/1605/ggrpt/carbon.htmlVanHouwelingen,J.H.,&VanRaaij,F.W.(1989).Theeffectofgoal-settinganddailyelectronicfeedbackonin-homeenergyuse.JournalofConsumerResearch,98…105.llink,T.,&Meertens,R.M.(1999).Deeffectiviteitvanelektronischefeedbackoverhetenergie-enwaterverbruikdoormiddelvanteletekstbijhuishoudens.(TheeffectivenessofelectronicfeedbackonhouseholdenergyuseandwaterusebymeansoftextTV).InR.M.Meertens,R.Vermunt,J.B.F.DeWit,&J.F.Ybema(Eds.),Socialepsychologieenhaartoepassingen(pp.79…91).ASPO.Delft:Warriner,G.K.,McDougall,G.H.,&Claxton,J.D.(1984).Anydataornoneatall?Livingwithinaccuraciesinself-reportsofresidentialenergyconsumption.EnvironmentandBehavior,503…526.Winett,R.A.,&Kagel,J.H.(1984).Effectsofinformationpresentationformatonresourceusein“eldstudies.JournalofConsumerResearch,655…667.Winett,R.A.,Kagel,J.H.,Battalio,R.C.,&Winkler,R.C.(1978).Effectsofmonetaryrebates,feedback,andinformationonresidentialelectricityconservation.JournalofAppliedPsychologyWinett,R.A.,Leckliter,I.N.,Chinn,D.E.,Stahl,B.,&Love,S.Q.(1985).Effectsoftelevisionmodelingonresidentialenergyconserva-JournalofAppliedBehaviorAnalysis,33…44.Winett,R.A.,Love,S.Q.,&Kidd,C.(1982…1983).Theeffectivenessofanenergyspecialistandextensionagentsinpromotingsummerenergyconservationbyhomevisits.JournalofEnvironmentalSystemsWinett,R.A.,Neale,M.S.,&Grier,H.C.(1979).Effectsofself-monitoringandfeedbackonresidentialelectricityconsumption.JournalofAppliedBehaviorAnalysis(2),173…184.ARTICLEINPRESSW.Abrahamseetal./JournalofEnvironmentalPsychology25(2005)273–291