CONSPIRACYTHEORYANDPARTISANPOLITICSByAdamM.EndersADISSERTATIONSubmittedtoMichiganStateUniversityinpartialentoftherequirementsforthedegreeofPoliticalScience{DoctorofPhilosophy2016ABSTRACTCONSPIRACYTHEORYANDPARTISANPOLITICSByAdamM.EndersWhileresearchonconspiracytheoriesandthosewhobelievethemhasrecentlyundergoneanempiricalrenaissance,therestillexistsagreatdealofuncertaintyaboutthemeasure-mentofconspiratorialbeliefsandpredispositions,andtheconsequencesofaconspiratorialmindsetwhenitcomestoexpresslypoliticalattitudesandbehaviors.Unfortunately,currentstrategiesformeasuringthelatenttraitthatinformsconspiratorialbeliefsarebyerrorattributabletopartisanmotivatedreasoning,socialdesirabilityconcerns,andthepecu-liarcontentandcontextofspconspiracytheories.Inthisproject,Idemonstrate,usingdatafromthe2012AmericanNationalElectionStudy,thatbeliefsinavarietyofspconspiracytheoriesaresimultaneouslytheproductofbothageneralpredispositiontowardconspiratorialthinkingandpartisanship.IfurtherusingsurveydatagatheredviaAmazon'sMechanicalTurkplatform,thatthelevelofpartisancontentingrainedwithinconspiracytheoriesvariestremendouslyacrossconspiracytheories,includingeveninherently\political"conspiracies.Next,Iuseuniquedatagatheredviaamoduleonthe2014CooperativeCongressionalElectionStudy(CCES)andreigningtheoryonthecharacteristicsofconspiratorialthinkingtoconstructandvalidateacumulative,unidimensionalscaleofconspiratorialthinking.Ithatasubstantialproportionofthepopulationishighlysuspiciousofgovernmentalauthority.Inadditiontobeingrelatedtootherconstructsandattitudessuchasauthoritarianism,trust,andperceptionsofgovernmentalcorruption,theconspiratorialthinkingscalestronglypredictsconspiratorialattitudesaboutgenetically-mofoods,thelinkbetweenchildhoodvaccinesandautism,thebirthplaceofBarackObama,governmentalknowledgeaboutthe9/11terroristattacks,andtheassassinationofJohnF.Kennedy.Finally,Idemonstratetheofconspiratorialthinkingonavarietyofpoliticalphe-nomenaofnormativeinteresttopoliticalscientists.First,Iconsiderthemoderatingroleofaconspiratorialmindsetwhenitcomestotraditionalrelationshipsliketheonebetweenpartisanshipandattitudesabouttheroleofgovernment.Moresp,Idemonstratethatasone'slevelofconspiratorialthinkingincreases,thetiesbetweenpartisanshipandfeelingsaboutthefederalgovernment,attitudesaboutthescopeandreachofgovernment,preferencesaboutthelevelofgovernmentalspendingandproductionofservices,andsup-portfortheTeaPartysimultaneouslydissolve.Second,Ithatconspiratorialthinkingisnegativelyrelatedtoapprovalofmajorgovernmentalinstitutions,includingthepresident,Congress,andtheSupremeCourt.Lastly,Icorroboraterecentworkshowingthenega-tiverelationshipbetweenconspiratorialthinkingandparticipationincampaignactivities,andextendthisworkbydemonstratingthenegativeofconspiratorialthinkingontheprobabilityofidentifyingasapartisanandthepositiverelationshipbetweenconspiratorialthinkingandTeaPartysupport.ACKNOWLEDGEMENTSFirst,IwouldliketothankanumberofgraduateschoolcolleagueswhohavemadethisprojectbetterthanitwouldotherwisehavebeenhadIbeenleftsolelytomyowndevices.Ingeneral,allthemembersoftheMSUAmericanPoliticsResearchGroupprovidedconstructiveandusefulfeedbackonnearlyeveryaspectofthisproject{thewordsandnumbersonsubsequentpagesareundoubtedlyimprovedbythem.Inparticular,IthankBobLuptonandMilesArmalyforexceptionallythoughtfuladviceonthisprojectandmanyothers.IwouldberemissifIfailedtoprovidespecialrecognitiontoStevenSmallpage.Withouthisassistanceandencouragement,thisprojectwouldnotonlybefarworse{itwouldbenonexistent.Anumberofthanksregardingthedataemployedinthisprojectarealsoinorder.First,IwouldliketothankProfessorsThomasHammondandCharlesOstromforprovidingfundingforthedatathatwasgatheredviaAmazon'sMechanicalTurkplatform.ThisdataisusedinChapters2and3ofthemanuscript,andevenmoredatahasyettobeanalyzed.IalsothankChrisHareforincludingtheuniquesuspicionquestionsemployedinChapters3and4onthe2014CooperativeCongressionalElectionSurvey.Ofcourse,Iowemanythankstoallmembersofmycommittee:WilliamJacoby,SaundraSchneider,CorwinSmidt,andDanielBergan.Eachindividualprovidedmanythoughtfulsuggestionsandadvicethatundoubtedlyimprovedmywork.Inparticular,CorySmidtpushedmetoconsidermoreseriouslytherelationshipbetweenconspiratorialthinkingand(dis)trust,aswellasthenormativeofconspiratorialthinking.SandySchneider'sencouragementthroughoutallofmygraduatestudentexperience{especiallyviaemployingmeattheICPSRSummerprograminmanytcapacities{hashelpedshapemeintothescholarIamtoday.Finally,andmostimportantly,IexpressmydeepestgratitudetoBillJacoby.InmysemesterofgraduateschoolIattendedProfessorJacoby'sdepartmentalpresentationivofhis\vectormodel"ofcorevaluepreferences.ItisthistalkwhichalertedmetoProfessorJacoby'swhollyunique{and,atthetime,enchanting{orientationtowardthemeasurementandinterpretationofsocio-politicalphenomena.ThispresentationalsospurredaconversationbetweenfellowgraduateStevenSmallpageandI{aconversationthathasleadtoastrongfriendshipandprosperousworkingrelationship.TheimpactofthistalkaloneisworthyofgratitudebeyondwhatIcancogentlyexpressinthislimitedspace.ProfessorJacobyengagedwithmeasayeargraduatestudentbyvolunteeringhistimetomeandotherswhennocoursesonpublicopinionorpoliticalbehaviorwerebeingprovided(andcontinuestoprovide)mewithmanyuniqueemploymentexperiences{thebofwhichhaveandwillcontinuetoextendfarbeyondamerepaycheck{atMSU,theICPSRSummerProgram,andtheAmericanJournalofPoliticalScience,andhasguidedmyeducationandresearchincountlessways.Ithankhimforallofthesedeeds,theimpactofwhicharetrulyimmeasurableandstillforthcoming.Lastly,IthankProfessorJacobyforhisperspectiveonandapproachtoscieninquiry.Thoughitmayseemoddtothanksomeonefortheiroutlook,itismyunderstandingthatIamonlythelatestinalonglineofindividualswhoseliveshavebeentrulychangedbytheapproachtosciencethatProfessorJacobyhaspasseddowntome.ThisperspectivepositivelyanduniquelyallaspectsofmyworkandIexpectitwillcontinuetodosofortherestofmyprofessionalcareer.IcanonlyhopethatIwillonedaybefortunateenoughtocontinuethetraditionbystudentsofmyowninthesameway.vTABLEOFCONTENTSLISTOFTABLESviiLISTOFFIGURESixChapter1:Introduction11.1Organization3Chapter2:HowareConspiratorialThinkingandPartisanshipRelated?52.1Background62.2Study1:SourcesofSpeConspiracyBeliefs92.2.1EmpiricalResults122.2.2ValidatingtheConspiratorialThinkingFactor152.3Study2:WhatConspiraciesArePartisan?202.3.1EmpiricalResults232.4Conclusion29APPENDIX31Chapter3:ConstructingaBetterMeasureofConspiratorialThinking403.1Background413.2EstimatingSuspicionintheMassPublic443.3DataandAnalyticalStrategy473.4EmpiricalResults503.4.1ValidatingtheSuspicionScale533.4.2SuspicionandSpConspiracyBeliefs573.4.3RobustnessCheck613.5Conclusion66APPENDIX68Chapter4:TheAttitudinalandBehavioralofConspiratorialThinking824.1TheConditioningofConspiratorialThinking834.2onApprovalofActorsandInstitutions934.3onPoliticalEngagement954.4Conclusion101APPENDIX103Chapter5:Conclusion1105.1NormativeImplications1125.2DirectionsforFutureResearch114BIBLIOGRAPHY116viLISTOFTABLESTable2.1:Exploratoryfactoranalysisoffourspconspiracyitems.13Table2.2:Structuralequationmodelestimatesandstatistics.16Table2.3:Correlationsbetweenconspiratorialthinkingandtheoreticallyrelatedvariables.19Table2.4:DemographiccompositionofMTurksample.22Table2.5:PercentageofrespondentsattributingagivenconspiracytheorytoDemocrats,Republicans,orneitherparty.25Table2.6:Percentageofself-idenDemocraticandRepublicanrespondentsattributingagivenconspiracytheorytoDemocratsorRepublicans.27Table2.7:Correlationsbetweenbeliefsinspconspiracytheories,2012ANES.38Table3.1:OLSregressionvalidatingthesuspicionscale.56Table3.2:Regressionsofspconspiratorialbeliefsonsuspicionscale.58Table3.3:Regressionsofspconspiratorialbeliefsonalternatesuspicionscale(MTurk).65Table3.4:Correlationsbetweensuspicionitems,2014CCES.73Table3.5:ExploratoryfactoranalysisofCCESsuspicionitems.74Table3.6:ExploratoryfactoranalysisofMTurksuspicionitems.74Table3.7:factoranalysisofCCESsuspicionitems.75Table3.8:factoranalysisofMTurksuspicionitems.76Table3.9:OLSregressionsdemonstratingthecovariatesacrossspconspiracytheories.80Table3.10:Correlationsbetweenbeliefsinspconspiracytheories,2014CCES.81Table4.1:OLSregressionsdemonstratingtheindependentadditiveofpartisanshipandconspiratorialthinking(ANES).88Table4.2:OLSregressionsdemonstratingtheofpartisanshipconditionalonconspiratorialthinking(ANES).90Table4.3:Independentofsuspicion(CCES)onapprovalofgovernmentalactorsandinstitutions.96Table4.4:Independentofsuspicion(CCES)onpoliticalengagement.99viiTable4.5:OLSregressionsdemonstratingtheindependentadditiveofpartisanshipandconspiratorialthinking.109viiiLISTOFFIGURESFigure2.1:DistributionsofresponsestoANESspconspiracyitems,bypartisanidentity.11Figure2.2:Two-factormodelofthecompetingsourcesofvariationinspconspiracybeliefs:conspiratorialthinkingandpartisanmotivatedreasoning.14Figure2.3:DistributionsofresponsestoMTurkspconspiracyitems,bypartisanidentity.23Figure3.1:Distributionsofindividualconspiratorialthinkingquestions.48Figure3.2:Distributionofsuspicionscalescores.52Figure3.3:Predictedprobabilityofregisteringaspconspiratorialattitudeovertherangeofthesuspicionscale.60Figure3.4:Distributionofresponsestoalternatesuspicionquestions.63Figure3.5:Distributionofsuspicionscalescores.64Figure3.6:DistributionofANESconspiracyitems.78Figure4.1:Marginalofpartisanshipconditionalonconspiratorialthinking(ANES).92ixChapter1:IntroductionIsthegeneticmooffoodasafeandemethodforproducinglargequantitiesofnutritiousfood,orisitsimplyaploybylargecorporationstoregulatefoodproduction,protecttheirandcontrolaverageconsumers?Dovaccinesreallyprotectchildrenfrompotentiallydevastatingillnesses,ordotheymerelyreplaceostensiblyextinctdiseaseswithlifelongconditionslikeautism?Recentresearchrevealsthatapotenttypeofmisinformationnon-triviallytheformationofmassattitudestowardrelevantpo-liticaltopicsrangingfromhealthcarepracticesandpoliciestoopinionsaboutthebirthplaceofhighlyvisiblepolitical(e.g.,Nyhan&2010,Nyhan,&Ubel2013).Yet,agreatdealoftheworkattemptingtoexaminethecharacteristicsofpeoplewhosub-scribetoaconspiratorialstyleofthinkingsuggeststhatthesebeliefsareheldbyonlyadiscontentedandhighly-suspiciousfew.Theseindividualsareinthegripofthe\paranoidstyle"(Hofstadter2008).AccordingtoHofstadter,the\paranoidstyle"isa\wayofseeingtheworldandofexpressingoneself"by\moreorlessnormalpeople"thathasthe\qualitiesofheatedex-aggeration,suspiciousness,andconspiratorialfantasy"(2008,3-4).Whilebeliefin,andpropagationof,conspiracytheoriesisoftenconsideredanactivityofthosewhooccupythefringesofsociety,ageneralsuspiciontowardgovernmentandauthoritieshasbeenahallmarkofAmericandemocracysinceitsveryinception{itisthebedrockoftheAmericanCreed(Bailyn1992,Hofstadter2008,Wood1993).Indeed,suspiciousorconspiratorialthinkingsurroundsomeofthemostimportanthistoricaleventsofthe20thand21stcenturies,fromthehumanstepsonthemoontotheassassinationofJohnF.Kennedy,andfromtheterror-1istattacksofSeptember11,2001tothedenialofthemassacresatSandyHookElementarySchoolandUmpquaCommunityCollege.Furthermore,theparanoidstyleunderwritingmassbeliefincertainconspiracytheoriescanbeofseriouspracticalconsequencetoboththeconspiracytheoristandthosearoundher.Consider,forinstance,thebeliefthatchildhoodvaccinationsarelinkedtoautism(Goertzel2010).Suchabeliefcan,anddoes,infact,suppressindividualwillingnesstovaccinateone'schildren(Jolley&Douglas2014a,Nyhan,Richey&Freed2014).Widespreadrefusaltovaccinatechildrenfromdiseasessuchasmeasles,mumps,andrubella(MMR)couldleadtoaresurgenceoftheharmfuldiseasesthatthevaccineswerecreatedtocontrol.Conspiracytheoriesaboutbirthcontrol(Bird&Bogart2003),HIV/AIDS(Bogart&Bird2006,Bird&Bogart2005),andevenhealthcarereform(Nyhan,&Ubel2013)cansimilarlyresultinseriouspublichealthconsequences.Eventhoughseriouspracticalreasonswhyweshouldunderstandconspiracytheoristsabound,empiricalresearchonthesetypesofindividualsandthemindsetthatcolorstheirviewoftheworldaroundthemisinearlystages.Aswillbediscussedatgreatlengthbelow,theverystartingpointfromwhichsoundempiricalinquirymustbegin{themeasurementofkeyconcepts{isexceedinglydisjointedacrossindividualstudiesand,morebroadly,socialsciendisciplines.Assuch,wehavenowayofdeterminingthesourcesofdisparateAretheydotomeasurementstrategy?Or,perhaps,culturalsetting?Qualitiesofthesurveysample?Evenmoredisconcerting,wecannotevenbetthatwearecapturingconspiracytheoristsorthepredispositiontothinkinaconspiratorialwaywithcurrentmeasurementstrategiesbecause,withveryfewexceptions(Brotherton,French&Pickering2013,Oliver&Wood2014),suchstrategieshaveneitherbeenconstructedwiththeoryinmind,norempiricallyvalidated.Inthisproject,mybroadgoalistwofold:toinvestigatetheproblematicpropertiesofthemodalmeasurementstrategyemployedintheburgeoningliteratureonconspiracytheorywithaparticularfocusontherolepartisanshipinconspiracybelief,and,second,to2propose,execute,andvalidateasuperiorstrategyformeasuringconspiratorialthinkingthatcircumventstheproblemsplaguingexistingts.Asaby-productofthesecentralgoalsIwillalsoextendcurrentknowledgeontheofconspiratorialthinking,withafocusonthepotentialattitudinalandbehavioralconsequencesofconspiratorialthoughtwhenitcomestoengagementwithestablishmentpartisanpolitics.1.1OrganizationTheprojectiscomprisedofthreecentral,substantivechapters.Thesecondchapterfocusesontherelationshipbetweenbeliefsinspconspiracytheories,thelatentpredispositiontoviewtheworldthroughalensofconspiracism,andpartisanattachments.InthischapterIemploylatentvariablemodelingtechniquesanddatafromthe2012AmericanNationalElec-tionStudyonbeliefsinfourspcconspiracytheoriestodemonstratethatthesuchbeliefsaresimultaneouslytheproductofpartisanmotivatedreasoningandageneralpredispositiontowardconspiratorialthought.Byconstrainingoneoftwoestimatedlatentfactorstocap-turepartisanattachments,Iamableidentifythesecondfactorasconspiratorialthinkinganditassuchusingtheretainedindividualfactorscoresandtraditionalmethodsofvalidation.Next,IemployuniquedatagatheredviaAmazon'MechanicalTurkplatformtoinvestigatethepotentialpartisannatureofalargerbatteryofspconspiracytheo-ries.Ithatwhilesomeconspiracytheorieshaveasubstantial{ifnotoverwhelming{partisancomponent,manyarenotwidelyperceivedasbeingattachedtoacertainpoliticalparty.Furthermore,Ithatpartisansarelargelywillingtoacceptownershipofspconspiracytheoriesthatmembersoftheirpartyperpetuate.InthethirdchapterIfurtherdocumenttheproblemsassociatedwithusingquestionsaboutbeliefsinspconspiracytheoriestoestimateconspiratorialthinking,proposeanalternativemeasurementstrategy,andvalidatetheproposedmeasurementstrategy.Al-thoughIhavedocumented,bythispoint,theproblemofpartisanshipwhenitcomestoestimatingconspiratorialthinkingusingsurveyquestionsaboutbeliefsinspconspiracy3theories,thereareotherproblemswithsuchastrategy.Majoradditionalproblemsincludesocialdesirabilitybias,andinthesocio-politicalcontentandcultural/temporalcontextofspconspiracytheories.Inlightofallthesesourcesoferrortakentogether,Iusethetheoreticalliteratureonconspiratorialthinkingtoguidetheconstructionofseveralquestionsdesignedtoserveasindicatorsofthegeneralconspiratorialthinkingtrait.Uponestimationofthistrait,Iestablishthereliabilityandvalidityofthemeasureusingscalingtechniquesandbeliefsinspconspiracytheories,respectively.Inthesubstantivechapter,Iusethetmeasuresofconspiratorialthinkingdevelopedinearlierchapterstodemonstratetheofconspiracismonavarietyofpoliticalattitudesandbehaviors.Moresp,Ishowthatconspiratorialthinkingandpartisanshipare,forthemostpart,atoddswitheachother.Indeed,Ithattherobustofpartisanshiponattitudesabouthighlypartisanandideologicaltopicssuchastheroleofgovernmentandtheappropriatebalancebetweengovernmentalspendingandtheprovisionofpublicservicesaregreatlyattenuatedaslevelsofconspiratorialthinkingrise.Relatedly,Ithatasconspiratorialthinkingincreases,thepredispositiontoidentifywithoneofthethetwomajorpartiesdecreases,asdoesparticipationincampaignactivities,whilesupportforananti-establishmentparty{theTeaParty{increases.Finally,Idiscussthenormativeconsequences{bothpositiveandnegative{ofconspiratorialthinking,andoutlinefuturedirectionsforresearchonthistopic.4Chapter2:HowareConspiratorialThinkingandPartisanshipRelated?Today,asubstantialproportionofthemasspublicendorsesatleastonespconspir-acytheory(Brotherton,French&Pickering2013,Hargrove2006),andtheseconspiratorialattitudesappeartobepoliticallymotivated.Forexample,RepublicansaremorelikelytobelieveboththatBarackObamawasbornoutsideoftheU.S.(Pasek,Stark,Krosnick&Tompson2015)andthatglobalwarmingisahoax(Blank&Shaw2015,McCright&Dunlap2011)despitethereleaseofabirthandwidespreadscienconsensus,respectively.Theintimateconnectionbetweenpartisanshipandconspiratorialthinkinghigh-lightsproblemswiththeconceptualandempiricaldistinctionsbetweenpartisanmotivatedreasoningandthepredispositiontosubscribetoconspiracytheories.Are,forexample,allconspiracytheoristspartisanmotivatedreasoners?Orisitpossibletobea\birther"ora9/11\truther"withoutbeingpartisan?Icontendthatmuchofthisconceptualandempiricalopacityisduetothemeasurementstrategymostfrequentlyemployedtoinvestigatethesetopics:surveyquestionsaboutbe-liefsinspconspiracytheories1.Inferencesaboutthepsychologicalcharacteristicsofconspiracytheoristsfromstatedbeliefsinspconspiracytheoriesareseverelycompli-catedbytheofpartisanmotivatedreasoning.Inlightofthisproblem,Iarguethattobetterunderstandthenature,frequency,andofconspiratorialthinking1Thisisnottosay,however,thatquestionsaboutbeliefsinspconspiracytheoriesaretheonlymeasurementdevicesemployedinthisliterature.Indeed,somescholarshavebeguntoproposeandteststrategiesformeasuringconspiratorialthinkingusingmoregeneralsentimentsthatmightbeindicativeofaconspiratorialmindset(e.g.,Brotherton,French&Pickering2013,Bruder,e,Neave,Nouripanah&2013,Uscinski&Parent2014).Ultimately,however,thesefairlyfewexamplesarestillmuchmoretheexceptiontotherule,andIknowofnosuchalternatestrategythathasbeenagreeduponasofthewritingofthismanuscript.5whenitcomestotheAmericanmasspublic,wemustseparatebothconceptuallyandempiricallyconspiratorialthinkingfromotherknownpsychologicalprocesses{inparticular,partisanship.BelowIdemonstratethatresponsestoquestionsaboutbeliefsinspconspiracytheoriesfoundonthe2012AmericanNationalElectionStudy(ANES)arethesimultaneousproductofbothageneralconspiratorialthinkingtraitandpartisanship.Bothexploratoryandlatentvariablemodelsareusedtoexploreand,ultimately,estimatetheofthesetwoconceptsonstatedbeliefsinspconspiracytheories.Althoughbeliefsinsuchovertlypartisanconspiraciesas\birtherism"arefoundtobeheavilybypartisanship,amoregeneralconspiracytheorytraitalsosubstantivelyresponsestoeachofthespconspiracytheoryquestionsemployed.Finally,IemployuniquesurveydatagatheredviaAmazon'sMechanicalTurkplatformtoexplorejustwhatspconspiracytheoriesarethoughtofaspartisanbyaverageindividu-als.Ontheonehand,Ithatsomeconspiracies{suchasthosesurroundingtheIraqWarandclimatechange{arewidelyperceivedasbeingsystematicallyperpetuatedbyonepartyoranother.Ontheotherhand,Ithatotherconspiracies{suchasthosesurroundingtheassassinationofJFKoraboutthevaportrailsleftbyaircraft{arenotwidelyperceivedashavinganypartisancontent.Furthermore,Iobservethatpartisans\own"theirrespectiveconspiracytheories.Inotherwords,Republicans,forexample,overwhelmingrecognizetheideathatclimatechangeisahoaxasbeingperpetuatedbyfellowRepublicans.2.1BackgroundPartisansareincreasinglyconspiratorial:recentresearchintoconspiratorialthinkinginthemasspublicrevealsthatpartisansincreasinglyassenttoconspiracybelief.Miller,Saunders,andFarhart(forthcoming),forinstance,thatconspiratorialbeliefsabouttopicsandeventssuchasthebirthplaceofBarackObama,theinclusionofa\deathpanels"provisioninthe2010CareAct,andliesonthepartoftheBushadministrationwith6respecttotheexistenceofweaponsofmassdestructioninIraqareallsubstantiallydrivenbypartisanmotivatedreasoning.Inasimilarvein,Pasek,Stark,Krosnick,andTompson(forthcoming)that\birther"beliefsthatBarackObamawasbornoutsideoftheU.S.areinformedbypartisanmotivatedreasoning.Indeed,theydemonstratethatpartisanship,ideologicalself-idenations,racialresentmenttowardblacksandnegativefeelingstowardBarackObamaarealltlyandsubstantivelyrelatedto\birther"beliefs.Finally,HartmanandNewmark(2012){viaImplicitAssociationTesting{thatconservatives,Republicans,andthosewhosimplydidnotlikeBarackObamaaremorelikelythanothers(particularlyliberalsandDemocrats)toregistererroneousbeliefsthatheisaMuslim.Onereasonpartisanshipmaybedrivingconspiracybeliefsisthatpartisanshipislargelyaprocessof\motivatedreasoning."Motivatedreasoningisadecision-makingmechanismbywhichindividualstiallyprocessandintegrateinformationbasedontheirpriorbeliefs,attitudes,andemotions,anditpervadesthepsychologyofindividuals'reasoningaboutpoliticalphenomenaandstimuli(Kunda1990,Lodge&Taber2013).Indeed,theofpartisanmotivationsinattitudeformationandexpressionhavebeenextensivelydocumentedwithrespecttomanytphenomena(e.g.,Bartels2002,Bullock,Gerber,Hill&Huberforthcoming,Taber&Lodge2006).Kraft,Lodge,andTaber(2015)andBlankandShaw(2015)evendemonstratethatpartisanmotivationsextendtothe(mis)interpretationoffactsproducedbynon-partisannamelyscientistsandothertypesofexperts.Ofcourse,thereisanalternativeexplanationfortheriseandcentralityofseeminglypar-tisanconspiracies:ageneralofwhatRichardHofstadtercalledthe\paranoidstyle"(Hofstadter2008).FollowingHofstadter's\paranoidstyle,"conspiratorialthinkingisbestunderstoodasastyleofreasoningaboutthepoliticalworldandourplaceinit(Hof-stadter2008).Numerousempiricalstudiesonconspiratorialthinkinglargelyagreethataconspiracytheoryisaninterpretationofaneventorpublicactioncenteringonasecretplanofasmallgroupofindividualsorgroups,whosegoalsandintentionsarepartiallyhidden,thoughusuallydirectedatassumingpower(Brotherton,French&Pickering2013,Bruder7etal.2013,Keeley1999,Moscovici1987).Likepartisanship,conspiratorialthinkingisa\per-ceptualscreen"throughwhichinformationissiftedandtheworldisinterpreted(Campbell,Converse,Miller&Stokes1960).Indeed,accordingtoGoertzel(1994),conspiratorialthink-ingisa\monological"beliefsystem,whereanindividualditiallyassesses(orignores)evidencethatcounterspriorbeliefs.Amonologicalbeliefsystemisonecontainingconsis-tentandcoherentbeliefs,butattheexpenseofmaintainingthemagainstcountervailingevidence.Empiricalstudiesonconspiracismortheparanoidstylepresentageneralpsy-chologicalofindividualswhopracticesuchconspiratorialthinking:theseindividualsaremoreauthoritarian,lesstrustingingovernment,lesseducatedandpoliticallysophisti-cated,morereligious,andoftentimesbelongtosocio-economicand-demographicminoritygroups(e.g.,Brotherton&French2015,Douglas&Sutton2010,Fenster2008,Jolley&Douglas2014b,Oliver&Wood2014,Swami,Chamorro-Premuzic&Furnham2010,Swami,Pietschnig,Tran,Nader,Stieger&Voracek2013,Wood,Douglas&Sutton2012).Thistypeofthinkingislargelyasignofdis-idenwithmainstreampolitics,whichmakesitaprimecandidateforanalternativeexplanationtothepartisanmotivatedreasoningview.Itcould,inotherwords,bethecasethatthexofconspiraciestodayisduetotheriseintheparanoidstyle,orconspiracism,andnotsimplypartisanpolarization.Perhapsthemostimportantaspectoftheparanoidstyleisthatconspiratorialthinkingdirectlyrelatestoafeelingofbeingbrandedapoliticaloutsider(Hofstadter2008).Indeed,whilethespconspiracytheoriesemployedinpreviousstudieshavevariedwidelyintheircontentandsocio-politicalcontext(Brothertonetal.2013),noneofthemhaveuncoveredarelationshipbetweenpartisanattachmentsandapredispositiontowardconspiratorialthought.Thisisnottosaythatpartisanscannotsubscribetoaspeconspiracy(asMiller,Saunders,andFarhart[forthcoming],forinstance,show),butitdoesraisetheimportantquestionastotheprimarypsychologicalmotivationbehindthatbelief:partisanorconspiratorialsuspicion?Ofcourse,neitherviewnecessarilyentailsdenyingtheother:anindividualmayhold8aconspiracybeliefbecauseofpartisanshipandconspiracism.And,whilethisperspectiveseemsmostlikely(especiallyinthecaseofconspiracytheoriesinvolvinghighlypartisancurrentmodelsofconspiratorialthinkingexcludeoneortheother.Millerandcolleagues(forthcoming),forexample,explorethecorrelatesofatwo-factormodelofcon-spiracieswherethetwofactorsrepresentliberalandconservativeconspiracies.Underthisconceptualizationandmeasurementstrategy,then,itisnosurprisethatideologicalmoti-vatedreasoningisfoundtoheavilyebeliefsinspconspiracytheories2.ThesameistrueforrecentworkdonebyPasekatal.(2015)whichshowsthatpartisanshipandsymbolicracismarethemostimportantonbeliefsthatBarackObamawasbornoutsideoftheU.S.Putsimply,thecarefulstepofempiricallytiating,andsubsequentlymodeling,thesourcesofbeliefsinspconspiracytheorieshasnotbeenmade.Withoutthistiation,welosepotentiallyvaluableinsightintothenatureofbothconspiratorialthinkingandpartisanship,promulgateconfusionregardingthesourcesofconspiratorialbeliefs,anddrawpotentiallymisleadinginferencesaboutvariouspoliticalbehaviorsfrommisspstatisticalmodels.2.2Study1:SourcesofSpConspiracyBeliefsInordertorevealthedynamicrelationshipbetweenpartisanshipandconspiratorialthinking,Iusethefourconspiracyitemsincludedonthe2012ANESpost-electionsurvey.Theseques-tionsaskabouttheextenttowhichrespondentsbelieveinconspiraciesaboutthebirthplaceofBarackObama(henceforthreferredtoasthe\birther"theory),theinclusionofaprovi-sionauthorizingthecreationofpanelstomakeend-of-lifedecisionsforpeopleonMedicareinthe2010CareAct(henceforthreferredtoasthe\deathpanel"theory),theamountofrelatedknowledgethefederalgovernmentpossessedpriortotheterroristattacksofSeptember11,2001(henceforthreferredtoasthe\truther"theory),andtheroleofthe2Ialsonotethattheideaofideologicalmotivatedreasoninginthecontextofspeconspiracybe-liefsaboutovertlypartisanandobjectsseemshighlysuspect.Althoughgeneralideologicalself-idenarehighlycorrelatedwithpartisanones,Iseelittlereasonwhyideologicalorientationsshouldbemoreofanthanpartisangroupidenwhichareomittedfromtheirmodels.9federalgovernmentinbreachingtheodleveesinNewOrleansduringHurricaneKatrina(henceforthreferredtoasthe\leveebreach"theory)3.Theresponsestothesequestionsarecodedsuchthathighervaluesdenotemoreconspiratorialattitudes.Thedistributionsofresponsestothesequestions{byDemocraticandRepubli-canpartisanship{arepresentedinFigure14.Thoughtheexactshapesofthesedistributionsvaries,eachofthedistributions,onaverage(acrosspartisanorientations,thatis),exhibitsomedegreeofpositiveskew,denotingthatlessthanhalfofindividualsbelieve{probabilis-ticallyorcertainly{anyofthesespconspiracytheories5.Evenso,however,thedegreeofsupportfortheseconspiracytheoriesvariestremendously,withnon-trivialproportionsofindividualseitherfullysupportingorexhibitinguncertaintyaboutagivenconspiracy.Theshapeofthedistributionsalsovary{heterogeneously{bypartisanship.AlthoughDemocratsmaybeslightlymorelikelytoascribetobeliefsinthe\truther"and\leveebreach"conspiracytheories,thereareclearlyvisiblebetweenpartisanswhenitcomestothe\birther"and\deathpanel"conspiracytheories.Moresp,RepublicansaretlymorelikelythanDemocratstoascribetoconspiracybeliefsregardingthebirthplaceofObamaanddeathpanels.Thisisarstpieceofevidenceforthepartisannatureofspconspiracytheories(atleast,thoseemployedhere),butwiththeaddednuancethatthestrengthofthepartisancontentcanvarybetweenconspiracytheories.Theremainderofmyempiricalanalysisproceedsasfollows.First,Iproposeatheo-reticalmodelofresponsestothespconspiracyquestionsfoundonthe2012ANES.3Iadmitthattheseconspiracytheoriesarefairly\political"(i.e.,aboutgovernmentandpolitics)innature,and,inthatsense,arenotfullyrepresentativeofallpossibletypesofconspiracytheories.Regardless,politicalconspiracytheoriesaretheonesthathavethemostapparentrealworldconsequenceandwhichreceivethemostattentionintheconspiracytheoryliterature,withothertypesofconspiracies(i.e.,thoseaboutextraterrestrials)consideredtobequitetintheirgeneralnature(Brothertonetal.2013).4Forinterpretationofthereferencestocolorinthisandallothergures,thereaderisreferredtotheelectronicversionofthisthesis(ordissertation).5IndividualswhoparticipatedintheinternetversionoftheANESsurveywereslightlymorelikelytoexpressmoreconspiratorialattitudesthanthoseindividualsintheface-to-facesample,likelybecauseofsocialdesirabilitybias.Inotherwords,individualsmaybelesslikelytodiscloseconspiratorialbeliefswhensittingacrossfromanotherindividualwhomayjudgethem.Inspiteoftheseslightdiscrepancies,weseenoreasonwhytherelationshipbetweenspconspiracybeliefs,orbetweenspconspiracybeliefsandotherconstructswouldbysurveymode.Regardlessofourhunch,wefurtheraddressandtestforsurveymodebelow.10Figure2.1:DistributionsofresponsestoANESspconspiracyitems,bypartisanidentity.Atthecoreofmyperspectiveisthatbothpartisanshipandageneralpredispositionto-wardconspiratorialthinkingencebeliefsaboutspconspiracytheories.Itestthistheoreticalmodelusinglatentvariablemodelingfamiliarlyknownasstructuralequationmodeling(SEM).Next,Ivalidatethemeasureofconspiratorialthinkingproducedfromthismeasurementapproachbyexaminecorrelationsbetweenconspiratorialthinkingandahostofcriterionvariablesknowntoberelatedtoconspiracism.Thesevariablesinclude:trustingovernment,liberal-conservativeideologicalself-idenreligiosity,knowledge,in-terestinpolitics,campaignactivity,authoritarianism,,attitudesaboutthelevelof11corruptioningovernment,education,andage.Finally,Iuseregressiontechniquestoexaminetheinteractionbetweenpartisanshipandconspiratorialthinkingwhenitcomestofeelingsaboutthegovernmentverygenerally,andattitudesabouttheroleofgovernment,moresp.Indeed,IestimateaseriesofinteractiveOLSmodelswithmultiplicativetermsdesignedtocapturethemoderatingimpactofconspiratorialthinkingonpartisanship'sonattitudesaboutthesizeandscopeofthefederalgovernment,thetensionbetweengovernmentspendingandtheproductionofgovernmentalservices,towardthefederalgovernment,andevensupportfortheanti-establishmentTeaParty.ThequestionwordingandcodingprocedureassociatedwitheachofthevariablesemployedinthefollowinganalysescanbefoundintheAppendix.2.2.1EmpiricalResultsIbeginmyanalysisbyexaminingtheextenttowhichtheobservedresponsestothefourANESconspiracyitemsaresimultaneouslytheproductoftwolatenttraits:conspiratorialthinkingandpartisanship.Inparticular,Iconductanexploratoryfactoranalysisofthefouritemsinordertobetterunderstandthelatentsourcesofvariancesharedbyresponsestothesesurveyitems6.Suchananalysiscouldproduceseveralerentresults.Ontheonehand,iftheregisteredbeliefsinthesespconspiracytheoriesreallyaretheproductsofonlya\monological"beliefsystemasmostconspiracyresearchtheorizes(Goertzel1994,Wood,Douglas&Sutton2012,Oliver&Wood2014,Uscinski&Parent2014),thenwewouldexpectthatasinglefactorcapturesmostofthevarianceintheresponsestotheitems.Ontheotherhand,ifresponsestotheconspiracyitemsarestrictlypartisaninnature,wemightexpectasinglefactorwithbothpositiveandnegativefactorloadings,assuming{asothershave(e.g.,Miller,Saunders&Farhartforthcoming){thatbothDemocraticandRepublicanconspiraciesarerepresentedinthisset7.Finally,theexploratoryfactoranalysis6AmatrixofPearsonpairwisecorrelationsbetweenresponsestotheconspiracyitemscanbefoundintheAppendixtothischapter.7Relatedly,wecouldalsondamulti-factorstructurewhereeachfactorcorrespondstoatparty(e.g.,onefactorforDemocraticconspiracytheories,andoneforRepublicanones.12Table2.1:Exploratoryfactoranalysisoffourspconspiracyitems.Factor1Factor2FactorLoadings:Birther0.480-0.422DeathPanels0.548-0.328Truther0.4980.369LeveeBreach0.4460.445Eigenvalue:0.9770.618VarianceAccountedFor:59.7%37.8%Note:Methodofestimationisiteratedprincipalfactors.n=4814.couldproduceasolutionwheremorethanonefactorisrequiredtoaccuratelydescribetheunderlyingstructureinthespconspiracydata.TheresultsoftheexploratoryanalysisarepresentedinTable1.Theestimatesaresuggestiveofacombinationofthetheandsecondscenariosdescribedabove.Indeed,whileallconspiracyitemshavestrongpositiveloadingsonthefactor,theyalsohavestrongloadingsonthesecondfactor8.Moreover,thesecondfactorloadingsforwhatcouldbeconsideredtheRepublicanconspiracies{the\birther"and\deathpanel"theories{arenegative,whereastheloadingsforwhatcouldbeconsideredtheDemocraticconspiracies{the\truther"and\leveebreach"theories{arepositive.Atwodimensionalsolutionisalsomoreappropriatethanaunidimensionalonegiventhehighproportionofvariance{approximately38%{accountedforbythesecondfactor9.InordertoprovideaformaltestofthestructureinthespconspiracybeliefdataIturntoalatentvariablemodelingframework.MytheoreticalmodelofthesourcesofvariationintheresponsestotheconspiracyitemsispresentedinFigure2.Accordingto8Thoughitisimpossibletosaywhatexactlyconstitutesa\strong"loading,themostcommonruleofthumbisthatanyloadingof0.30orgreatershouldbeconsiderednon-trivialatleast,and\strong"atbest.9Ialsoexaminedascreeplotoftheeigenvaluesversusthenumberoffactorsandconductedaparallelanalysis,bothofwhichsuggestedthatatwo-factorsolutionbestaccountedfortheunderlyingstructureintheresponsestothefourconspiracyitems.13Figure2.2:Two-factormodelofthecompetingsourcesofvariationinspconspiracybeliefs:conspiratorialthinkingandpartisanmotivatedreasoning.BirtherDeathPanelTrutherLeveeBreachPartyIDObamaFTBushFTConspiratorialThinkingPartisanship11bothmytheoreticalexpectationsaswellastheresultsoftheexploratoryfactoranalysis,twolatentvariablesproducebeliefsinallfourofthespconspiracytheories:conspiratorialthinkingandpartisanship.Allestimatesassociatedwiththeconspiratorialthinkingfactorappearinblue,whilethoseassociatedwithpartisanshipappearinred.Inadditiontothefourspconspiracyitems,Iincludethreeadditionalitemsrelatedtopartisanshipinthepartisanshipfactoronly:partisaniden(measuredviathestandardseven-pointscale),feelingstowardBarackObama,andfeelingstowardGeorgeW.Bush(bothofwhicharemeasuredviathefamiliarfeelingthermometer[\FT"]items).TheestimatesofthemodelinFigure2arepresentedinTable210.Everyfactorloadingforboththeconspiratorialthinkingandpartisanshipfactorsisstatisticallyntatthep<0:001level(assumingatwo-tailedtest).Perhapsmoreimportantly,traditionalmeasuresofsuggestthatthismodelaccountsfortheunderlyingstructureinthedataverywell.Indeed,theRootMeanSquaredErrorofApproximation(RMSEA)is0.052andthelargep-valueassociatedwiththisestimate(0.345)suggeststhatIcannotrejectthe10Thoughthepathsfromconspiratorialthinkingfactortothe\birther"itemandfromthepartisanshipfactortothepartisanidenitemareusedtoscaletherespectivefactors,non-1valuesappearinTable2.ThisisbecauseIpresentstandardizedfactorloadings,whichallowacomparisonofofthelatentfactorsontheobservedvariables.14nullhypothesisthatthetrueRMSEAisequaltoorlessthan0.05.SinceRMSEAsof0.05orlessaremostdesirable(e.g.,Browne&Cudeck1992,Kline2011),Ihaveapieceofpromisingevidenceforanappropriatemodel.TheStandardizedRootMeanResidual(SRMR)alsodemonstratessuperiormodelasitisquiteabitlowerthanthesuggested0.08forgood(Hu&Bentler1999).BoththeComparativeFitIndex(CFI)andTucker-LewisIndex(TLI)provideevidenceofasuperiormodelaswell.Theruleofthumbregardingtheseindicesisthatvaluesgreaterthan0.95suggestverygoodmodel(Hu&Bentler1999).ThevaluesforbothindicessurpassthisAlthoughthechi-squaredtestforequivalenceoftheobservedandmodel-impliedcovariancematricesisstatisticallysigt,thechi-squaredtestisknowntobeparticularlysensitivetosamplesize(amongotherpeculiaritiesofthedata),andIemployanunusuallylargesamplehere(4790individuals).Thus,itdoesnotworrymethatthismeasureofdoesnotfullycomportwiththeotherindicators.2.2.2ValidatingtheConspiratorialThinkingFactorThoughIcanbequitecertainoftheinterpretationofthepartisanshipfactorduetotheidenstrategyIemployed,Imusttakeaslightlytapproachwiththecon-spiratorialthinkingfactor.Inordertovalidatethatthisfactorreallycanbeinterpretedasageneralizedconspiratorialthinkingtrait,Icorrelatetheindividualfactorscoreswithahostofvariablesknowntoberelatedtoconspiracybeliefs,aswellassomepreviouslyunexploredvariablessuchasattitudesabouttheproportionofgovernmentalwhoarecorruptand.Ofcourse,aquestionremainsofwhethersymbolicpoliticalpredispositions,suchaspartyidenandliberal-conservativeself-idencation,arerelatedtothegeneralpredispo-sitiontowardconspiratorialthinkingcapturedinthefactor.Accordingtothetheoreticaldistinctionbetweenconspiratorialthinkingandpartisanshipmadeabove,aswellasthefactthatthebettermodelofspconspiratorialbeliefswasonewheretheconspiratorialthink-15Table2.2:Structuralequationmodelestimatesandstatistics.ConspiratorialThinkingPartisanshipFactorLoadingsBirther0.2480.683(0.016)(0.013)DeathPanels0.3090.464(0.017)(0.014)Truther0.572-0.074(0.022)(0.017)LeveeBreach0.627-0.195(0.024)(0.017)PartyIden0.678(0.015)ObamaThermometer-0.816(0.013)BushThermometer0.629(0.016)FitStatistics˜2(7df),p-value97.412,0.000RMSEA0.052Prob(RMSEA0.05)0.345SRMR0.019CFI0.992TLI0.975n4790StandardizedMLEcotswithstandarderrorsinparentheses.Allestimatesstatisticallytatthep<0:001levelwithrespecttoatwo-tailedtest.16ingandpartisanshipfactorswereconstrainedtonotcovary,Iknowthatpartisanshipisnotacorrelateofthegeneralpredispositiontowardconspiratorialthinking.Whenitcomestoideology,however,itcouldbethecasethatconservativesaremorepronetoconspiratorialthinking.Itis,afterall,partoftheconservativemantratodistrustgovernment.Ialsoconsiderwhethertrustingovernmentalinstitutionsisrelatedtoconspiratorialthinking.Morespeci,Ihypothesizethatgeneralizedtrustingovernmentandconspir-atorialthinkingarenegativelyrelatedsuchthatastrustdecreases,thepropensitytowardconspiratorialthinkingincreases(e.g.,Goertzel1994,Miller,Saunders&Farhartforthcom-ing).Ialsosuspectthatconspiratorialthinkingwillberelatedtoattitudesabouttheper-ceivedlevelofgovernmentalcorruption.The2012ANESaquestionabouthowmanyofthepeoplerunningthegovernmenttherespondentthinksarecorrupt,withresponseoptionsrangingfromalltonone.Ihypothesizethattherewillbeapositiverelationshipbetweenconspiratorialthinkingandattitudesaboutgovernmentalcorruptionsuchthatasconspiratorialthinkingincreases,theperceivedproportionofcorruptgovernmentalincreases.Next,Ihypothesizethateducationalattainmentandknowledgemayberelatedtocon-spiratorialthinking.Iexpectthateducation/knowledgeandconspiratorialthinkingareneg-ativelyrelatedsuchthataseducationalattainmentandknowledgeaboutpoliticsincreases,thepropensitytothinkconspiratoriallydecreases.Thisislargelyduetomyexpectationthathigherlevelsofeducationandknowledgeaboutthepoliticalprocesstendtoreducetheer-rorsinreasoningthattheoreticallyunderwritethepropensitytowardconspiratorialthinking(Brothertonetal.2013).Recentworkalsosuggeststhatconspiratorialthinkingisnegativelyrelatedtoparticipationinpolitics(e.g.,Einstein&Glick2015,Jolley&Douglas2014b),asentimentIendorsehere.Participationisoperationalizedbyanadditiveindexofcampaignactivitiesanindividualparticipatedinduringthe2012electionseason.Ialsoconsidertherelationshipbetweenconspiratorialthinkingandexternal.Morespec,Ihy-pothesizethatthemoreconspiratorialoneis,thelowerthelikelihoodtheywillbelievethat17publiccarewhattheythink.Thisrelationshipwouldbeconsistentwiththeideathatconspiracytheoristsbelievethatpoliticsisreallyrunbysecretgroups,andthatpoliticiansareliarsthatareconstantlyschemingontheirownbehalves(Brotherton2015).Forsimilarreasons,Iexpectthatinterestinpoliticsisnegativelyassociatedwithconspiracism.Finally,Iexaminewhetherreligiosity,authoritarianism,andagearerelatedtothecon-spiratorialthinkingfactor.Somehavefoundthat\supernatural"or\paranormal"thinkingisrelatedtobeliefsinparticularconspiracies(Newheiser,Farias&Tausch2011,Oliver&Wood2014).Thus,religiosity{operationalizedasthefrequencyofattendanceataplaceofworship{maybepositivelyrelatedtoconspiratorialthinking.Thereisalsosomeevi-dencethatolderindividualsaremorepronetoconspiratorialthinking(Goertzel1994,Us-cinski&Parent2014){anexpectationIshare.Finally,oneofthelongeststandingcor-relatesofconspiratorialthinkingisauthoritarianism(Abalakina-Paap,Stephan,Craig&Gregory1999,Goertzel1994,&Bruder2014,McHoskey1995).Thisworknotesthatthosewithauthoritarianpredispositionsaremorelikelytoblamesocio-political\others"fortheirdisadvantagedpositionsinlifeandpossessheightenedlevelsofaggressionandhostil-ity.Beliefinconspiraciesaboutcertainindividualsandgroups,itishypothesized,providesanoutletforthesetendencies,resultinginapositiverelationshipbetweenconspiratorialthinkingandauthoritarianism.1819Table2.3:Correlationsbetweenconspiratorialthinkingandtheoreticallyrelatedvariables.(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(1)ConspiratorialThinking1.000(2)TrustinGovernment-0.0901.000(3)Lib-ConIdeology0.008-0.1701.000(4)Authoritarianism0.2010.0030.2101.000(5)Religiosity0.0360.0650.2430.2141.000(6)Knowledge-0.269-0.0950.040-0.306-0.0331.000(7)Interest-0.1300.0380.037-0.1340.0560.3401.000(8)CampaignActivity-0.0440.029-0.023-0.0820.0890.2460.3511.000(9)Education-0.222-0.012-0.035-0.3340.0250.4210.2160.1581.000(10)Age-0.1270.0030.1170.0060.1060.2100.2290.1750.0091.000(11)Corruption0.260-0.3540.0950.078-0.025-0.078-0.072-0.070-0.121-0.0931.000(12)Externalcacy-0.1530.352-0.070-0.0680.0430.0220.1110.1220.1180.007-0.3091.000CellentriesarePearsoncorrelationcots.=p<0.05levelwithrespecttoatwo-tailedtestAmatrixofpairwisePearsoncorrelationsbetweenalloftheaforementionedvariablescomprisesTable3.Mostimportanttoconsideristhecolumnwhichdisplaysthecor-relationsbetweenconspiratorialthinkingandallothervariables.Allcorrelationcotsarecorrectlysignedexceptfortheoneassociatedwithliberal-conservativeideology,whichalsohappenstobetheonlystatisticallytrelationshipintheset.Thus,allofthehypothesizedrelationshipshold,providingstrongevidencethatmylabelingofthisfactoras\conspiratorialthinking"isappropriate.Thereareparticularlystrongrelationshipsbetweenconspiratorialthinkingandknowledge,education,andauthoritarianism.Additionally,thetwonewcriterionvariables{attitudesaboutgovernmentalcorruptionandexternal{exhibitfairlystronglinearrelationshipswithconspiratorialthinking,especiallygiventhenebulousnessoftheconceptsandtheiroperationalizations11.ThisoperationalizationofconspiratorialthinkingwillbeemployedinChapter4whereIdemonstratetheofconspiratorialthinkingonvariouspoliticalbehaviorsandatti-tudes.Inkeepingwiththethemeoftherelationshipbetweenconspiratorialthinkingandpartisanship,IturnnowtoasecondstudywhereIinvestigatethepartisannatureofsetofcommonconspiracytheoriesasdeterminedbyindividualsingeneral,andself-describedpartisansinparticular.2.3Study2:WhatConspiraciesArePartisan?InthepreviousstudyIdemonstratedthatbeliefsinsometypesofspconspiracytheoriesarethesimultaneousproductofbothpartisanorientationsandconspiratorialthinking.Ofcourse,however,Iwaslimitedbythenumberandtypeofconspiracytheoriesincludedonthe2012ANES.Unsurprisinglygiventhedataset,thespconspiracytheoriesincludedontheANEScouldverywellbemorepartisaninnaturethanotherconspiracytheories,even11Idonotintendtosuggestthatconceptssuchas{ortrustorinterestinpolitics,forthatmatter{areunimportant,buttonotethatmeasuringthesemulti-facetedconceptsismorethanmeasuringsomethinglikegeneralfeelingsaboutthepresident.Assuch,modelsofthesevariableshavehistoricallyaccountedforfairlylittlevarianceintheconcepts,suggestingthattherelationshipsIobservehere,whileindeedsomewhatweak,areparfortheproverbialcourse.20otheronesaboutthegovernmentorpolitics.Thus,therelationshipbetweenpartisanshipandconspiratorialthinkingobservedabovemaycharacterizesometypesofconspiracytheoriesmuchbetterthanothers.Inthissecondstudy,Ibuildontheformerbyexploringjustwhichofseveralwell-knownconspiraciesarepopularlyconceivedofaspartisaninsomeway.ToexaminethepotentialpartisancontentofconspiracytheoriesIaskedsurveyrespondentstoindicatewhichofthemajorpoliticalparties,ifeither,wasmorelikelytopromoteagivenconspiracytheory.Thefullpromptisasfollows:\BelowisalistofcontroversialideasrelatedtoAmericanpolitics.Pleaseindi-catewhetheryouthinkeachideaismorelikelytobepromotedbyDemocrats,Republicans,orneitherDemocratsnorRepublicans."Followingthepromptwasalistof8statementswhichsummarizedthecentralthrustofconspiracytheoriesthathavebeenwidelyemployedinthescholarlyliteraturetodate.ThefulllistofconspiraciesandthestatementsdescribingthemarepresentedinTables5and6below.Thedataemployedinthisstudycomesfromasampleof1543U.S.adultsgatheredviaAmazon'sMechanicalTurk(heretofore\MTurk")platform.Individualswhere$0.80tocompletea6-8minutesurveyinwhichtheywouldbeaskedabouttheir\opinionsaboutpoliticsandthenews."AsiscustomarywithMTurksamples,individualsinthisstudytendedtobeabitmoreliberal/Democratic,havehigherlevelsofeducationalattainment,andbeyoungerthantheaverageadultU.S.population.Table4belowliststhecodings/ranges,means,andstandarddeviationsofacoupleofkeydemographiccharacteristicsoftheindi-vidualsinthissample.Regardlessoftheslightunrepresentativenessofthesample,Ihavelittlereasontobelievethattheseindividualsthinkaboutconspiracytheoriesinfundamentallytwaysthanamorerepresentativesamplemight.Furthermore,Ihavenoreasontobelievethatthepartisansinthissamplewouldbehaveintlytwaysthanmore\average"21Table2.4:DemographiccompositionofMTurksample.Coding/Mean/StandardDemographicVariableRangeProportionDeviationPartyIden(3=StrongRepublican)-3-3-0.7591.898Ideology(3=Extremelyconservative)-3-3-0.8842.017Education(5=Advanceddegree)1-53.5650.864Age(inyears)18-7834.58710.831Female(1=female)0-10.4390.496Black(1=black)0-10.0690.254Hispanic(1=hispanic)0-10.0520.222Totaln=1,543.partisans.Asanadditionalcheckonthisclaim,IdisplayinFigure3thedistributionsofresponsestothetwospconspiracybeliefquestionsinthissurveythatoverlappedwiththoseonthe2012ANES,bypartisanship.Thequestionwordingandavailableresponsecategoriesfortheseitemsareexactlythesameacrossthetwosurveys.ThedistributionsofquestionresponsesintheMTurksamplearenearlyidenticaltothoseinthe2012ANESsample.Onaverage,theindividuals{acrosspartisaniden{intheMTurksamplearelesslikelytobelieveeitherthe\birther"or\truther"conspriacytheo-riesthanthoseintheANESsample,thoughtheisfairlyminor.Moreimportantly,theshapeofthedistributions,forthewholesampleandbypartisanship,areextremelysim-ilar.Indeed,thedistributionofresponsestothequestionabout\birther"beliefsisstronglypositivelyskewed,withDemocratsbeingmuchlesslikelytoexpressbeliefinthetheorythanRepublicans.AswiththeANESsample,thedistributionofbeliefsinthe\truther"theoryareslightlypositivelyskewedforallindividuals,withveryminorbetweenmembersofpartisangroups.ItakethesesimilaritiesasastrongindicatorthatneithertheresultspresentedbelownortheconclusionsImakefromthemwouldnotbesubstantiallyalteredhadIemployedamorerepresentativesample.22Figure2.3:DistributionsofresponsestoMTurkspconspiracyitems,bypartisanidentity.2.3.1EmpiricalResultsThestepofmyanalysisentailsaninvestigationofthepotentialpartisannatureofconspiracytheoriesasstatedbyallindividuals.Thus,Ipresent,inTable5,atabulationofresponsestothequestionoutlinedaboveintofourcategories.Sinceweknowsolittleabouthowindividualsthinkaboutconspiracytheories,whethertheysubscribetothemornot,Ialsoincludetheproportionofindividualswhostatedthattheyeitherdidn'tknowwhichpoliticalpartywasmorelikelytopropagateagivenconspiracytheory,orthatneitherpoliticalpartywasmoreorlesslikelythantheothertopromotethetheory.Eachoftherowsofthetableshouldsumtoapproximately100,withminordeviationsbeingattributabletoroundingerror.AcursoryglanceatTable5revealsawidevarianceinthedirectionandstrengthof23partisancontentascribedtocertainconspiracytheories.ConspiraciesaboutthebirthplaceofBarackObama(1),globalwarming(8),anddeathpanels(6)areoverwhelmingviewedbyindividualsasconspiracytheoriesmostoftenpromotedbyRepublicans.Approximately86%oftheindividualsperceivethe\birther"conspiracyasbeinganinherentlyRepublicanone,withnearly74%and60%ofindividualssayingthesamewithrespecttotheglobalwarminganddeathpanelconspiracies,respectively.ThoughIexpectedthatthesethreeconspiracieswouldlikelybeperceivedasRepublican-\owned,"theproportionofindividualswhoconcur{inthefaceofbothsocialdesirabilityandendorsementsbyoneofmodernpolitics'mostunlikeableRepublicanpresidentialcandidates{isquiteremarkable.WhilesomeoftheconspiracytheoriesconfrontingrespondentswereperceivedasmoreDemocraticthanRepublican,theproportionofsuchbeliefsismuchlowerthanthatwithrespecttoRepublicanconspiracies.Conspiracytheoriesregardingthe9/11terrorattacks(3),invasionofIraq(5),andHurricaneKatrina(7)areallperceivedasbeingmorelikelytobepromotedbyDemocratsthanRepublicans.The\most"Democraticconspiracytheory(regardingtheIraqwar),however,wasconsideredprimarilyDemocraticbyonly45%ofrespondents,ashyofamajorityandsubstantiallylowerthantheRepublicanconspir-acytheories.EachoftheDemocraticconspiracytheoriesalsohadmuchhigher\neither"and\don'tknow'ratesthandidthemajorityRepublicanconspiracytheories.Indeed,thecombined\neither"and\don'tknow"rates(43%)forthemostovertlyDemocraticconspir-acytheory{abouttheIraqwar{nearlyequalledtherateofindividualsperceivingtheconspiracytheoryasprimarilyDemocratic(45%).ThoughitmaybetemptingtoconcludefromthesenumbersthatRepublicansaremoreconspiratorial{or,atleast,morewillingto\use"conspiracytheories{thanareDemocrats,suchaconclusioncannotbemadefromtheseresults.ThereareatleasttworeasonswhytheDemocraticconspiracytheoriesmaybe\less"DemocraticthantheRepublicanonesareRepublican.Thereasonregardsthesurveysample.SincetheMTurksampleincludesmoreDemocratsthanRepublicans,theaggregateresultsmayanunwillingnessof24Table2.5:PercentageofrespondentsattributingagivenconspiracytheorytoDemocrats,Republicans,orneitherparty.ConspiracyTheoryDemocratsRepublicansNeitherDon'tKnow1)BarackObamawasbornoutsideof2.7286.397.193.69theU.S.2)LeeHarveyOswalddidnotactalone12.7011.5453.7322.03inassassinatingJohnF.Kennedy.3)Seniorhadknowledgeofthe28.1313.0345.4313.429/11terrorattackbeforetheyoccurred.4)Vaportrailsleftbyaircraftareactually14.7114.6553.8616.79chemicalagentsdeliberatelysprayedinasecretprogramdirectedbygovernment5)TheU.S.invasionofIraqwasnotpart45.2411.2831.8911.60ofacampaigntotterrorism,butwasdrivenbyoilcompaniesandJewsintheU.S.andIsrael.6)Thehealthcarelawpassedin20109.6659.7817.1114.45authorizedtheuseofdeathpanelstomakeendoflifedecisionsforpeopleonMedicare.7)Thefederalgovernmentintentionally34.809.5940.9614.65breachedodleveesinNewOrleansduringHurricaneKatrinasothatpoorneighborhoodswouldbeodedandmiddleclassneighborhoodswouldbespared.8)Globalwarmingisahoax,perpetuated3.9573.5615.946.55byenvironmentalscientistswhohavetheirownpoliticalagenda.25self-describedDemocratsto(implicitly)claimownershipofconspiracytheories.Alsowithrespecttosampling(albeitofaenttype),ImayhaveinadvertentlychosenforinquiryconspiracytheoriesthataremoreoverwhelmingRepublicanthanDemocratic.ThesecondreasonwhyIamcautiousinconcludingthatRepublicansaremorecon-spiratorialthanDemocratsisduetothe\conspiraciesareforlosers"hypothesis(Uscinski&Parent2014).UscinskiandParenthypothesize{though,donotempiricallyc{thatconspiracytheoriesareusedbypartisanstopunish,insomeway,membersofthepartisanoutgroup.Take,forexample,thepast8yearsunderthepresidencyofBarackObama.AccordingtoUscinskiandParent'shypothesis,wewouldexpecttoseeagrowthinRepublican-leadconspiracytheoriesaboutthe(Democratic)presidentandhispolicies,drivenprimarilybyadesiretodiscreditandregisterdisapprovalofanoutgroupmemberwithtotalpoliticalpower.Thus,theoverwhelmingextenttowhichtheRepublicancon-spiracytheoriesareregardedasRepublicancouldbedotoarealgrowthinandincreasingprominenceofsuchconspiracytheories.Ofcourse,thisphenomenoncouldalsobesensitivetopartisanpolarization,astateofpoliticswhichhasonlyincreasedinintensityoverthepastdecadeorso.Inalsoappearsthatsomeconspiracytheoriessimplydonotcontainmuchpartisancon-tent,and,therefore,arenotperceivedasparticularlypartisaninnature.Suchconspiraciesincludetheoriesaboutbothaircraftvaportrails(4)andtheassassinationofJFK(2).Fewerindividualsfelttinassigningaparticularpoliticalpartytoeitheroftheseconspir-acytheoriesthandidindividualswhosimplyhadnoopiniononthem.Perhapsthisistobeexpectedwhenitcomestothevaportrailconspiracy,asthepromptisvoidofanyexplicitorwidelyunderstoodpartisancontent.Thisisnotthecase,however,withtheJFKconspiracy.JohnF.Kennedywas,ofcourse,aDemocraticpresident;so,ifindividualswereusingonlypartisancontent(asopposedto\natureofthetimes"information)toassignconspiracythe-ories,itshouldhavebeenperceivedasprimarilyRepublicaninnature.Thisprovidessomeadditionalnuancetotheexplanationsregardingthecausesofpartisanattachmentsto26Table2.6:Percentageofself-idenDemocraticandRepublicanrespondentsattributingagivenconspiracytheorytoDemocratsorRepublicans.DemocratsRepublicansConspiracyTheoryDemocratsRepublicansDemocratsRepublicans1)BarackObamawasbornoutsideof2.9897.023.5796.43theU.S.2)LeeHarveyOswalddidnotactalone51.1148.8957.4242.58inassassinatingJohnF.Kennedy.3)Seniorhadknowledgeofthe64.5035.5074.2425.769/11terrorattackbeforetheyoccurred.4)Vaportrailsleftbyaircraftareactually44.1055.9059.3240.68chemicalagentsdeliberatelysprayedinasecretprogramdirectedbygovernment5)TheU.S.invasionofIraqwasnotpart76.9623.0482.8317.17ofacampaigntotterrorism,butwasdrivenbyoilcompaniesandJewsintheU.S.andIsrael.6)Thehealthcarelawpassedin201011.7688.2417.9382.07authorizedtheuseofdeathpanelstomakeendoflifedecisionsforpeopleonMedicare.7)Thefederalgovernmentintentionally77.0722.9379.1220.88breachedodleveesinNewOrleansduringHurricaneKatrinasothatpoorneighborhoodswouldbeodedandmiddleclassneighborhoodswouldbespared.8)Globalwarmingisahoax,perpetuated4.3295.686.6893.32byenvironmentalscientistswhohavetheirownpoliticalagenda.conspiracytheories;indeed,itseemsthatrelevance(culturalorsimplytemporal)maybeatplayaswell.27Lastly,Iconsiderthedistributionofresponsestothesequestionsbypartisan-ship.Table6includestheproportionofDemocratsandRepublicanswhoattributethespconspiracytheoriestofellowDemocratsandRepublicans.Thisanalysiscanreveal,insomesense,thestrengthofpartisanshiprelativetotheconspiracytheorylabel.If,forinstance,oneoftheprovidedconspiracystatementswasperceivedmoresoasacrazycon-spiracybelief,membersofthepartytowhichtheconspiracyisattachedmaybelesswillingtoclaimownershipoftheconspiracy.If,ontheotherhand,partisanconspiracytheoriesareusedbypartisansaspoliticaltoolsfordiscreditingmembersofpartisanoutgroups,wemightobservethatpartisansarewillingtoclaimownershipoftheconspiraciesattachedtotheirparty.TheproportionsinTable6aremoresupportiveofthelatterpossibilitythantheformer.Eachofthepartisanconspiracytheoriesinthetotalsampleretain,directionally,thesamepartisanattachmentwhenthesampleisbyDemocratsandRepublicans.Thestrengthofthepartisanvalenceisalsopreservedinthisanalysis.Inotherwords,the\birther"andglobalwarmingconspiracytheoriesarestillthemostpartisan{thebiggestdiscrepancybetweenpartisanattachmentsisstillassociatedwiththeseconspiracytheories,withtheprimarilyDemocraticconspiracytheoriesbeinglessstronglydividedalongpartisanlines.Finally,theaircraftvaportrailandJFKconspiracytheoriesarestillattributedtobothpartiesequally.Thatmembersofpartisangroupsaresowillingtoadmittheirownparty'sculpabilityinpromotingcertainconspiratorialideasistellingofthetruenatureofthe\conspiratorial"elementofthetheories.Ifthe\birther"orglobalwarmingconspiracytheorieswererecog-nizedasbeingtrulyconspiratorial,ratherthansimplypartisan,membersoftheattachedpoliticalgroups(Republicansinthiscase)wouldlikelyattempttodistancethemselvesandtheirpartyfromsuchacrazyandradicalidea.However,weobservejusttheopposite.Theoverwhelmingpartisancontentofmanyoftheconspiracytheoriesinquiredabouthereservesasapieceofevidenceagainstthetrulyconspiratorialelementoftheideas,andfor28theprimarilypartisancontentoftheideas.2.4ConclusionLabelingsomeonewhoascribestoaparticularconspiracytheoryeithera\conspiracytheo-rist,"fromaheatedparanoidpsychopathology,orsimplya\partisan,"drivenbygroup-basedpartisanistomissthenuanceoftherelationshipbetweenthegeneralpsychologicalpredispositionsthatareconspiratorialthinkingandpartisanmotivatedrea-soning.IthatbothpartisanmotivatedreasoningandconspiratorialthinkingplaysystematicandsubstantialrolesintheformationofbeliefsinspconspiracytheoriesaboutthebirthplaceofBarackObama,deathpanels,thetruthbehindthe9/11terroristattacks,andtheroleofthegovernmentindisastersrelatedtoHurricaneKatrina.Ialsofoundthatnotallconspiracytheoriesarecreatedequally.Thatistosay,notonlyarethereanumberofheterogeneitiesacrossspconspiracytheorieswhichwecanregardaserrortobeaveragedoutofourmeasurementsviamultipleitemmeasures,buttherearesystematicsourcesofvariance{namely,partisanship{whichmustbeconsideredmoredeliberately.Althoughsomeconspiracytheoriessuchasthoseaboutvaportrails,theassassinationofJFK,andtheroleofnegligenceinthe9/11terroristattacksarenotperceivedaspartisaninnature,otherconspiracytheoriessuchasthoseaboutthebirthplaceofBarackObamaorthedubiousnessofclimatechangeverymuchsoare.Whatismore,partisansarequitewillingtoclaimownershipoftheseovertlypartisanconspiracytheories,insomesensethatsuchbeliefshavemuchlesstodowithapredispositiontowardconspiracismthantheydowithpartisanallegiances.Theseareimportantforanumberofreasons.First,thatbeliefsinspcon-spiracytheoriesaresosignitlytheproductoflatentpredispositionsbeyondtheassumedconspiracismshouldcauseconspiracyresearcherstostronglyreconsiderthemeasurementstrategiestheyemployandhowtheyemploythem.Indeed,previousworkhaseitheras-sumedthedomineeringofpartisanshipoverconspiratorialthinking(e.g.,Milleret29al.forthcoming,Paseketal.2015),orjusttheopposite{notevenincludingitinthevariousmodelsofconspiracybelief(e.g.,nearlyallworkonthetopicconductedbysocialpsychologists).Assuch,theempiricalrelationshipsreportedinworkusingconspiratorialthinkingasoperationalizedby\raw"questionsaboutbeliefsinspconspiracytheoriesaresuspect.Morethananystatisticalproblemsassociatedwiththismeasurementstrategy,Iwishtoemphasizethattheresultanttheoreticalopacityisofmuchgreaterconsequence.Futureresearchshould,minimally,boththeoreticallyandempiricallyprovidefortheofnon-conspiratorialattitudes,traits,andpredispositionsonbeliefsinspconspiracytheories,and,maximally,consideralternatemeasurementstrategies.Itisthislatergoalthatconspiracyresearchershavejustbeguntoadvance(e.g.,Brothertonetal.2013,UscinskiandParent2014),andthetasktowhichIturninthenextchapter.30APPENDIX31Variablecoding/wordingforANESvariablesWasBarackObamabornintheUnitedStates,probablybornintheUnitedStates,probablyborninanothercountry,orborninanothercountry?(nonmain_born)0.itelybornintheU.S.1.ProbablybornintheU.S.2.Probablyborninanothercountry3.itelyborninanothercountryDoesthehealthcarelawpassedin2010authorizegovernmentpanelstomakeend-of-lifedecisionsforpeopleonMedicare,probablyauthorizegovernmentpan-elstomakeend-of-lifedecisionsforpeopleonMedicare,probablynotauthorizegov-ernmentpanelstomakeend-of-lifedecisionsforpeopleonMedicare,ornotauthorizegovernmentpanelstomakeend-of-lifedecisionsforpeopleonMedicare?(nonmain_endlife)0.itelydoesnotauthorize1.Probablydoesnotauthorize2.Probablyauthorizes3.itelyauthorizesDidseniorfederalgovernmentknowabouttheterroristattacksonSeptember11,2001beforetheyhappened,probablyknewabouttheterroristattacksonSeptember11,2001beforetheyhappened,probablydidnotknowabouttheterroristattacksonSeptember11,2001beforetheyhappened,ordidnotknowabouttheterroristattacksonSeptember11,2001beforetheyhappened?(nonmain_govt911)320.itelydidnotknow1.Probablydidnotknow2.Probablyknew3.itelyknewSomepeoplesaythatwhenHurricaneKatrinahittheGulfCoastinthesummerof2005,thefederalgovernmentintentionallybreachedodleveesinNewOrleanssothatpoorneighborhoodswouldbeodedandmiddle-classneighborhoodswouldbespared.Doyouthinkthefederalgovernmentdidthis,probablydidthis,probablydidnotdothis,ordidnotdothis?(nonmain_hurric)0.itelydidnotdothis1.Probablydidnotdothis2.Probablydidthis3.itelydidthisHowmanyofthepeoplerunningthegovernmentarecorrupt?(trustgov_corrpt)0.None1.Afew2.Abouthalf3.Most4.All7-pointPartyID(pid_x)-3-3,where-3strongDemocrat,3strongRepublican7-pointIdeology(libcpre_self)-3-3,where-3extremelyliberal,3extremelyconservative33Importanceofreligion(relig_import)0-3,where0notatallimportant,3veryimportantChurchattendance(relig_churchoft)0-5,where0never,5morethanonceaweekPoliticalInterest(interest_attention)0-3,where0Hardlyatall,3MostofthetimeDuringthecampaign,didyoutalktoanypeopleandtrytoshowthemwhytheyshouldvotefororagainstoneofthepartiesorcandidates?(mobilpo_rmob)0.No1.YesDidyougotoanypoliticalmeetings,rallies,speeches,dinners,orthingslikethatinsupportofaparticularcandidate?(mobilpo_rally)0.No1.YesDidyouwearacampaignbutton,putacampaignstickeronyourcar,orplaceasigninyourwindoworinfrontofyourhouse?(mobilpo_sign)0.No1.YesDidyoudoany(other)workforoneofthepartiesorcandidates?(mobilpo_otherwork)0.No1.Yes34Duringanelectionyearpeopleareoftenaskedtomakeacontributiontosupportcampaigns.DidyougivemoneytoanINDIVIDUALCANDIDATErunningforpublic(mobilpo_ctbcand)0.No1.YesDuringanelectionyearpeopleareoftenaskedtomakeacontributiontosupportcampaigns.DidyougivemoneytoaPOLITICALPARTYrunningforpublic(mobilpo_ctbpty)0.No1.YesDuringanelectionyearpeopleareoftenaskedtomakeacontributiontosupportcampaigns.DidyougivemoneytoaANYOTHERGROUPrunningforpublic(mobilpo_ctboth)0.No1.YesKnowledgequestions(preknow_prestimes,preknow_sizedef,ofcrec_speaker_correct,ofcrec_vp_correct,ofcrec_pmuk_correct,ofcrec_cj_correct)0.Incorrect1.CorrectHowoftencanyoutrustthefederalgovernmentinWashingtontodowhatisright?(trustgov_trustgrev)0.Never351.Someofthetime2.Abouthalfthetime3.Mostofthetime4.AlwaysHowmuchdopublicscarewhatpeoplelikeyouthink?(efficpo_carerev)0.Notatall1.Alittle2.Amoderateamount3.Alot4.AgreatdealPleasetellmewhichoneyouthinkismoreimportantforachildtohave:INDEPEN-DENCEorRESPECTFORELDERS(auth_ind)0.Independence1.RespectforeldersPleasetellmewhichoneyouthinkismoreimportantforachildtohave:CURIOSITYorGOODMANNERS(auth_cur)0.Curiosity1.GoodmannersPleasetellmewhichoneyouthinkismoreimportantforachildtohave:OBEDIENCEorSELF-RELIANCE(auth_obed)0.Self-reliance361.ObediencePleasetellmewhichoneyouthinkismoreimportantforachildtohave:BEINGCONSIDERATEorWELLBEHAVED(auth_consid)0.Considerate1.Well-behavedIrish,Italians,Jewishandmanyotherminoritiesovercameprejudiceandworkedtheirwayup.Blacksshoulddothesamewithoutanyspecialfavors.0.Disagreestrongly1.Disagreesomewhat2.Neitheragreenordisagree3.Agreesomewhat4.AgreeStronglyGenerationsofslaveryanddiscriminationhavecreatedconditionsthatmakeitforblackstoworktheirwayoutofthelowerclass.0.Agreestrongly1.Agreesomewhat2.Neitheragreenordisagree3.Disagreesomewhat4.DisagreeStronglyOverthepastfewyears,blackshavegottenlessthantheydeserve.0.Agreestrongly371.Agreesomewhat2.Neitheragreenordisagree3.Disagreesomewhat4.DisagreeStronglyIt'sreallyamatterofsomepeoplenottryinghardenough;ifblackswouldonlytryhardertheycouldbejustaswellaswhites.0.Disagreestrongly1.Disagreesomewhat2.Neitheragreenordisagree3.Agreesomewhat4.AgreeStronglyBarackObamafeelingthermometer(ft_dpc)0-100,where0correspondsto\cold"feelings,100correspondsto\hot"(veryfavorable)GeorgeW.Bushfeelingthermometer(ft_gwb)0-100,where0correspondsto\cold"feelings,100correspondsto\hot"(veryfavorable)Education(dem_edugroup_x)0-5,where0nohighschool,5postgradAge(dem_age_r_x)Ageinyears38CorrelationsBetweenSpConspiracyBeliefsTable2.7:Correlationsbetweenbeliefsinspconspiracytheories,2012ANES.(1)(2)(3)(4)(1)Birther1.000(2)Truther0.0911.000(3)DeathPanels0.3890.1341.000(4)LeveeBreach0.0200.3700.1031.000CellentriesarePearsoncorrelationcots.p<0:05levelwithrespecttoatwo-tailedtest.39Chapter3:ConstructingaBetterMeasureofConspiratorialThinkingInthischapter,Iadvancethestudyofsuspiciousthinkingthroughtheconstructionandsubsequentvalidationofascalecapturingthelatentsuspicionthatdrivesbeliefinparticularconspiracies.Inthesectionofthechapter,Iarguethatthemodalmeasurementdeviceusedtoassessconspiratorialthinking{surveyquestionsaskingrespondentsiftheyendorseaspconspiracytheory{biasesourunderstandingofthenumber,nature,andgeneralcharacteristicsofconspiracytheorists.Moreprecisely,Icontendthatquestionsregardingtheendorsementofspconspiracytheoriesareerror-ladenmeasuresofconspiratorialthinkingforthreereasons:heterogeneousbiasesduetothevaryingcontentandcontextofindividualconspiracies,contaminationviapartisanmotivatedreasoning,andsocialdesir-abilitybias.Inthesecondpartofthechapter,insteadofrelyingonthisproblematicmeasurementstrategy,Ioperationalizethecomponentsoffrequentlyemployedandlargelyagreeduponofconspiratorialthinkingtocreateareliable,valid,andtheoreticallysoundmulti-itemmeasureoftheunderlyingtraitthatIlabel\suspicion."Indeed,Iestablishthevalidityofthescalethroughbothitsstrongrelationshipswiththeoreticalvariablesofinterestanditsabilitytopredictspconspiratorialattitudes.Asexpected,Ithatsuspiciousthinkingispositivelyassociatedwiththebeliefinthelinkbetweenchildhoodvaccinationsandautism,thatgeneticallymofoods(GMOs)aredangerousbecausecorporateinterestscontroltheprocess,andthatpresidentsaregreatlybypoliticalmotiveswhenallocatingdisasterrelief.403.1BackgroundMuchoftheliteratureinvestigatingconspiratorialthinkingfocusesonindividuals'beliefsinparticularconspiraciesinordertorecordthefrequencyofsuchheterodoxbeliefsandtomea-suretheircorrelationswithotherpsychologicalpredispositionsandpersonalitytraits.Earlyliteraturealsosuggestedthatconspiratorialthinkingwasatraitofthesocioeconomicallydis-advantaged,ingmuchofRichardHofstadter's\pathologymodel"(Hofstadter2008,Fenster2008).Mostempiricalstudiesofconspiratorialthinkingfoundthatthesefringeormarginalizedindividualswhobelieveincertainconspiraciesarelargelydistrustfulandpronetofeelingsofpowerlessness,andlikelytosubscribetoconservativesocialideologieslikeright-wingauthoritarianism(Abalakina-Paapetal.1999,Goertzel1994,Darwin,Neave&Homes2011,Swami,Coles,Stieger,Pietschnig,Furnham,Rehim&Voracek2011,Swamietal.2013).Takentogether,thesesuggestthatconspiratorialthinkingisaphe-nomenonfoundonlyamongpeculiar,isolatedindividuals.However,theempiricalexaminationofconspiracytheoriesismoreimportantthanmerelytallyingthefrequencyofwhobelievesinwhatconspiracy.Amajornormativethrustofthestudyofconspiratorialbeliefsistogainabettertheoreticalandempiricalunderstandingofthenatureofconspiratorialthinkingandthosewhoengageinit(Uscinski&Parent2014,Goertzel1994).Yet,recentworkhashasbeenhamstrungbyafaultymeasurementstrategythatonmeasuringbeliefsinparticularconspiracies,creatingtinferentialproblemsaboutthenatureofconspiratorialthinking.InthefollowingsectionIdiscussthreesourcesofbiasinthemodalmeasurementstrategy:partisanmotivatedreasoning(andresponses),conspiracy-spheterogeneities,andsocialdesirabilitybias.Recentresearchsuggeststhatself-reportedbeliefsinparticularconspiraciesareheav-ilybypartisanmotivatedreasoning(e.g.,Hartman&Newmark2012,Miller,Saunders&Farhartforthcoming).Motivatedreasoningistheunconscioustendencytoseekout,process,andintegrateinformationinsuchawaythatpriorbeliefsarepreserved41(Kunda1990,Lodge&Taber2013,Taber&Lodge2006).Partisanmotivatedresponsesthusthreatenthevalidityofrecentndingssinceindividualsmayrecordanresponsetoaparticularconspiracyifthatconspiracytapsintonegativedirectedtowardapoliticalout-group.Thatis,beliefinaspconspiracytheorymaybedueexclusivelytopartisanallegiances,ratherthananygeneralpredispositiontowardconspiratorialthinking.Millerandcolleagues(forthcoming)demonstrate,usingquestionsaboutthe\birther"and\deathpanel"conspiracytheories,respectively,thatideologicalmotivatedreasoningdoes,infact,beliefsaboutthosespconspiracytheories.Similarly,PasekandcolleaguesthatRepublicansandthosewhopossessnegativeattitudestowardblacksaremorelikelythanotherstoendorsethe\birther"conspiracy(Paseketal.2015).Finally,OliverandWood(2014)unearthapartisancomponenttoquestionsaskingaboutthe\birther"conspiracy,aswellastheallegedlysecretcabalbeingplannedbyliberalbillionaireGeorgeSoros.Thisinferentialsourceoferror1wasexploredatgreatlengthinthepreviouschapter.inthecontentandcontextofspconspiracytheoriesalsomakethisgeneralmeasurementstrategysuspect.Theseinthegeneraltopicsandfociofspconspiracytheoriescanleadsomepeopletobemorelikelytobelieveagivenconspiracytheorythanothers,regardlessoftheindividuals'generalpredispositiontowardconspiratorialthought.Forexample,believingintheexistenceofaliensisatquestionthanbelievinginArea51(Brotherton,French&Pickering2013).Spconspiracytheoriesareboundbytime,place,socio-politicalcontext,andculture,whichmakesarrivingatgeneralizableconclusionsexceedinglyespeciallygiventheseeminglyarbitrarychoicebytheinvestigatorastowhatspconspiracytheoriesshouldbeincludedinthestudyintheplace(Brotherton,French&Pickering2013,Bruderetal.2013).MaterialintheAppendixdemonstratesthecharacteristicsofthoseregisteringbeliefsinthevarious1Iamreferringtotheofpartisanshipas\error"inthiscomponentofthemanuscriptbecausemyfocushasshiftedtounderstandingthenatureofconspiratorialthinking,exclusively.Thus,whatwaspreviouslyconceivedofaatheoreticallyinterestingsourceofvarianceinthepreviouschapterisnowbeingreferredtoasmeasurementerror.Inotherwords,theonlyrealinmyreferencetopartisanshipbetweenchaptersisashiftinanalyticalgoals.42conspiracytheoriesincludedonthe2012AmericanNationalElectionStudy(ANES).Indeed,theofkeypredictorvariablessuchaspartyidenreligiosity,thecognitiveneedtoevaluate,politicalsophistication,age,andracethatallconspiracytheoriesarenotcreatedalike.Finally,questionsaboutbeliefsinspconspiraciescanbeplaguedbysocialdesirabil-itybias.Socialdesirabilitybiasisthebiasintherecordedsurveyresponseduetorespon-dents'proclivitytoaltertheirstatedansweraccordingtotheirbeliefsaboutwhatissocially(un)acceptable(Edwards1957).Sinceconspiracytheoriesandthosewhobelieveinthemhavelongbeenconsideredparanoid,delusional,orgenerallycrazy,it'squiteeasytoimaginewhyquestionsaskingrespondentswhetherornottheyendorsespconspiracytheoriesareparticularlypronetothissortofbias(Fenster2008,Hofstadter2008,Robins&Post1997).Indeed,thecombined\don'tknow"andrefusalratefortheconspiracybeliefquestionsonthe2012ANESwereashighas14%,substantiallyhigherthanevenotherquestionsthatarenotoriouslysensitiveinnaturesuchasthoseaboutraceandracism(Berinsky1999).Forthesethreereasons,Iarguethatpreviousdrawnfrombeliefsinspconspiracytheorieslikelybiasandobscureourunderstandingoftheextenttowhichagen-eraltendencytowardsuspiciousorconspiratorialthinkingexistsinthemasspublic,andpotentiallymischaracterizethespnatureofthoseofaconspiratorialpredisposition.Whilemanyhaverecognizedtheneedforameasureofthegeneralconspiratorialthink-ingtraitthatismorecongruentwithacceptedtheoryaboutconspiracytheorists(e.g.,Goertzel1994,Oliver&Wood2014),onlyahandfulofsuchstudieshaveattemptedtoestimatethisgeneral,\context-free"trait(Brotherton,French&Pickering2013,Im&Bruder2014,Oliver&Wood2014,Uscinski&Parent2014).And,whileIappreciatethedirectionofthisresearch,thesefewpreviousattemptshaveseriouslimitations.Brothertonandcolleagues,forinstance,employ75separatequestionsinantoexploratorilyunearthaunidimensionalconspiracytrait(2013).Unfortunately,astheau-thorsthemselvesnote,theyestimateamuchlessparsimonious,andtheory-incongruent,43e-dimensionalconstruct(13).Furthermore,thedauntingquestionbatterywasonasmallconveniencesampleofBritishcollegestudents,promptingconcernsaboutgeneraliz-abilitytothebroaderpopulationandacrosscultures.OliverandWoodsuggestthattheirofconspiratorialthinkingmaybeindicativeofanelusive\centralorganizingprin-ciplebehindpublicbelief"(2014,13),thoughtheythatconspiratorialthinkingisakintoaparanoidbeliefinthesupernaturalandparanormal{bothofwhichsuggestthattheyarecapturingsomethingotherthanmerelyconspiratorialthinkng(indeed,supernaturalandparanormalbeliefsareseparatefactorsinBrothertonetal.[2013]'se-factorsolution).Furthermore,thefactoranalysistheypresentrevealsthatpartisanmotivatedreasoningper-vadesmanyoftheresponsestoquestionsaboutbeliefsinthespconspiraciesaboutwhichtheyinquire,namelythoseinvolvingthebirthplaceofBarackObamaandthesecretdesignsofbillionaireliberalGeorgeSoros.Finally,byUscinskiandParent(2014),whileturningupaseeminglyunidimensionalmeasureofconspiratorialthinking,areneitherempiricallyvalidatednorusedtoexaminepotentialcausesorconsequencesofconspiratorialthinking.Giventheinadequaciesofthesefewpreviousattempts,Iturnnexttomyownproposalforcarryingoutthisloftytask.3.2EstimatingSuspicionintheMassPublicConspiratorialthinkingisastyleofreasoningaboutthepoliticalworldandourplaceinit(Hofstadter2008).Aconspiracytheoryisoftenunderstoodtobeanexplanationofagivenevent(orasetofevents)byreferencingthesecretplanofasmallcollectionofunknownindividuals(orgroups)thathavetheintention(oftendisguised)toassumemorepower(Bruderetal.2013,Keeley1999).Ratherthandebatingthedetailsofanyoneconspiracy,themorefruitfulscholarlydirectionistostudytheprevalenceofthis\narrativeform"or\style"ofthinking(Barkun2003,Fenster2008,Hofstadter2008).Whilethetheoreticalstructureofhowconspiratorialtendenciesguideourthinkingaboutpoliticsiswelldocumented,ithasonlybeenpartiallyexaminedempirically,andnopreviousstudieshavesoughttovalidatea44scalethattracesthegeneraltraitintheAmericanmasspublic.Theproperstartingpointofanyconspiracytheoryissuspiciontowardepistemicauthority{theaccountofwhathappenedandwhy(Brotherton,French&Pickering2013,Bruderetal.2013,Keeley1999,Uscinski&Parent2014).Therefore,thequestionIaskrespondentsistheirlevelof(dis)agreementwithastatementthatpoliticianslie:1.Politiciansoftenlie,dblame,andotherwaystolookinnocent.(PoliticiansLie)Thisis,perhaps,alowbar;wehavelongknownthatthevastmajorityoftheAmer-icanpopulationdoesnotthinkhighlyofpoliticiansortheirrepresentatives,andaspre-viousresearchshows,trustisaconsistentlynegativepredictorofconspiracybeliefs(e.g.,Abalakina-Paapetal.1999,Goertzel1994,Swamietal.2011,Swamietal.2013).However,Iarguethatthiscynicalassumptionmasksagenuineandpervasivesuspicionofauthority.Indeed,thoughwemightexcusethisassumptionthatpoliticianslie,thebeliefthatotherepistemicauthoritiesroutinelylieisoftenconsideredtobequitetroubling.Forexample,recentingsthatsizeableportionsofindividualsareincreasinglyskepticalordistrustfuloftheauthoritativeclaimsofthesciencommunityandofscience'sroleinpublicpolicyhasmademanyquestionthebeliefinarational,deliberativepublicspherealtogether(Blank&Shaw2015,Bolsen,Druckman&Cook2015,Kraft,Lodge&Taber2015).Whilequestion-ingepistemicauthorities(betheypoliticiansorscientists)iscertainlyasignofasuspiciousbeliefsystem,itdoesnotnecessarilymeanthatsomeoneiswhollya\conspiracytheorist."Instead,weshouldconcludethereverse:Thebeliefthatpoliticianslieisalogicallynecessarybeliefforconspiracytheorists,butnotalogicallycientone.ThesecondquestionIaskfollowspreviousresearchinsuggestingthatanotheraspectofconspiratorialthinkingisthebeliefthatthedemocraticmachineryhasbrokendownandbeenovertakenbyeliteinterests:2.Governmentinstitutionsarelargelycontrolledbyeliteoutsideinterests.(Out-45sideInterests)Americanpoliticalpopulismcannotsurvivewithoutthisbeliefinapowerfulelite(Hofstadter2008,Fenster2008),anditisatthecoreofUcisnskiandParent's(2014)ofwidespreadconspiratorialthinking.Again,Idonotarguethatsomeonewhobelievesthatasmallandpowerfulelitecontrolsourgovernmentalinstitutionsisnecessarilyaconspiracytheorist,butratherthatthe\paranoidstyle"requiresthisbelief(Hofstadter2008,Fenster2008,Keeley1999,Pigden2007,Uscinski&Parent2014).Recentresearchshowsthatindividualswhobelieveinconspiracytheoriesaremorelikelytocommitthe\conjunctionfallacy"inreasoningabouttheprobabilityoftwoindependenteventsoccurring.Forthisreason,conspiracytheoristsoftenbelievethateverythingiscon-nected,orthattherearenoaccidents(Barkun2003,Brotherton&French2014,Keeley1999).Inordertotapthispeculiarpsychologicalprocess,Iaskindividualsabouttheextenttowhichthey(dis)agreethatthereare\accidents"innationalpolitics:3.Innationalpolitics,nothinghappensbyaccident.(NoAccidents)Again,thoughindividualswhodonotbelieveinconspiracytheoriesarealsopronetomakethesameerroneousprobabilitycalculations,itismoreprevalentamongconspiracybelieverslargelybecausetheynecessarilypositaworldwheretherearenoaccidents,wheretheworldispotentiallywhollydiscernible.Finally,theparanoidstyleculminatesinthebeliefthatnotonlyaretherenoaccidents,butalsothattheindividualcan\see"oruncoverthe(otherwise)secretplotsorplansofothers.Theideathateveryacthasacauseblossomsintoastrongerbeliefthateveryacthasanintentionalcause.Thisphenomenoniscalled\intentionalitybias,"andindividualswhobelieveinconspiraciesaremorepronetobelievethatanactisintentionalthannot(Brotherton&French2015).ThisintentionalitybiasisakintoOliverandWood's(2014)\unseenforces"aspectoftheparanoidstyle.Mylastquestionaskstheextenttowhichtherespondent(dis)agreesthatshecanseethesecretpatternsordesignsaroundher:464.Youcanseepatterns,designs,andsecretactivitieseverywhereonceyouknowwheretolook.(SecretDesigns)Thislastaspectofsuspicionisthepeakofthe\paranoidstyle,"ahabitofthinkinginwhicheverythinghasa(known)causeand,moreover,iscausedbyalyingelite.Thislastquestiontakestheconceptofsuspiciontoitsnaturalconclusionthatthesurfaceofamatterisacleverlydesigneddeception.Mymainhypothesesregardingtheproposedmeasureareasfollows:First,theseitemsallrelatetoasingleunderlyingdimension,whatIcallsuspicion;second,Ihypothesizethattheseitemsdiscriminatesystematicallybetweenthoseindividualswithmoresuspicionandthosewithless.Thiscumulativeorderingwouldpreciselymyintuitionthattheseitemsnotonlycanbeaggregatedintoareliablemeasureofsuspicion,butalsothattheyreveal,step-by-step,thelevelsofthe\paranoidstyle."Ultimately,anoperationalofafullysuspiciousorconspiratorialpersoncapturessomeonewhobelievesthatthingshappenduetoasecretplan,thatnothinghappensbyaccident,thateliteshaveusurpedtraditionalavenuesoflegitimatepower,andthatepistemicauthoritiesareunreliable.3.3DataandAnalyticalStrategyIeachofthequestionsoutlinedabove,inadditiontoonescapturingbeliefsinseveraltypesofspconspiracytheories,viaagroupcontentmoduleofthe2014CooperativeCongressionalElectionStudy(CCES)2.TheCCESisalarge-scale,nationalsamplesurveyonlinebyYougov/Polimetrix.ThedataIemploywerecollectedduringthepre-electionwaveofthesurveyinOctober2014.ThedistributionsofresponsestothefoursuspicionquestionsappearinFigure1.ThoughIammoreinterestedinasummarymeasureoftheseindividualitemswhichwillprovideamuchmorereliablemeasureofconspiratorialthinking(Ansolabehere,Rodden2SpecialthanksgotoChristopherHareforprocuringthisdataforme.47Figure3.1:Distributionsofindividualconspiratorialthinkingquestions.&Snyder2008),thedistributionsuniformlypossessanotablecharacteristic:afairlypro-nouncednegativeskew.Whileitmaynot(andprobablyshouldnot)comeasasurprisethatavastmajorityofindividualsbelievetosomeextentthatpoliticianslie(94%),itisremark-ablethatapproximately78%ofindividualsagreetosomeextentthatonecanseepatterns,designs,andsecretactivitiesonceoneknowswheretolook.Additionally,85%ofindividualsagreethatgovernmentinstitutionsarelargelycontrolledbyeliteoutsideinterests,and83%agreethatnothinghappensbyaccidentinnationalpolitics.TheoverwhelminglypositivesupportforthesepropositionslendssomedescriptiveevidenceforseeminglyhighlevelsofsuspicionintheAmericanmasspublic.48Therearethreecomponentstotheempiricalanalysispresentedbelow.First,Iconductanon-parametricitemresponsetheory(IRT)analysisoftheresponsestothefourquestionsdesignedtoserveasindicatorsofageneralpredispositiontowardsuspiciousthinking.Myintentionistodecipherwhethertheitemsserveasindicatorsofasinglecumulativetrait,and,thus,meetthestandardsofconstructvalidity.Furthermore,Iconsiderthedistributionofindividualscalescoresinordertoobtainapictureoftheelectorate'soveralllevelofconspiratorialthinking.Second,Iempiricallyvalidatethenewlyconstructedsuspicionscale.Forthiscomponentoftheinvestigation,Iuseinformationabouttherespondent'ssymbolicpoliticalpredisposi-tions(i.e.,partyidenandliberal-conservativeideology),leveloftrustingovernment,levelofpoliticalsophistication,educationalattainment,religiosity,andotherdemographiccharacteristicstoassesscriterionvalidity.Ialsousethesuspicionscale,alongwithappropri-atecontrolvariables,topredictbeliefsinthreemorespconspiracytheoriesregardingthelinkbetweenchildhoodvaccinesandautism,thedangerouscorporatecontrolofgenetically-mofoods,andtheextenttowhichpresidentialdisasterreliefaregovernedbypoliticalconcerns,respectively.Finally,afterthesuspiciousthinkingscalehasbeenvalidated,Iconsiderpotentialconse-quencesofsuspiciousthinkingforvariousformsofpoliticalengagement.Moresp,Iinvestigatewhethersuspicionisrelatedtothestrengthoftraditionalpartisanidentitiesasmeasuredbythefoldedpartyidenscale,strengthofanon-traditionalpoliticalidentityasmeasuredbysupportfortheTeaParty,andparticipationincommoncampaignactivitiesasmeasuredbyanadditiveindexofcampaignactivitiesundertakenduringthe2012presidentialelection.ThecodingschemeandexactquestionwordingassociatedwitheachofthesevariablescanbefoundintheAppendix.493.4EmpiricalResultsThefoursurveyitemsIemploytocapturesuspiciousthinkinginthemasspublicwerede-signedwithtwosptheory-derivedempiricalhypothesesinmind:thattheyshouldtapasingletrait,andthatthetraitshouldbecumulative.A\singletrait"canbeoperational-izedsimplyasaofunidimensionalitywhenitcomestothelatentvariablethatthefoursuspicionitemsaremeasuring.Unidimensionalitycouldbeassessedviaoneofseveraltypesofdimensionalexploration/reductiontechniquessuchasexploratoryfactoranalysis,factoranalysis,oroneofmanymorescalingmethods(e.g.,unfoldinganalysis).Theoperationalizationofthecumulativepropertyisslightlymoreinvolved.By\cumulative,"ImeanthatbothrespondentsanditemsprobabilisticallyfollowaGuttman-stylescalogramsuchthatIcanarrangebothsubjectsanditemsontoasingledimensionrepresentingtherelationshipsbetweentherespondentsanditemsinatheoreticallymeaningfulway.Taketheexampleofgeneralspellingability.Say,forinstance,thatstudentsinagivengradeschoolgrammarclassaregiventhreetypesofwordstospell:three-letterwords,e-letterwords,andseven-letterwords.Wewouldprobablyassumethatstudentswhocouldspelle-letterwordscouldalsospellmostthree-letterwords,andthatstudentswhocouldspelltheseven-letterwordscouldspellmoste-andthree-letterwords.Ifthisisempiricallytrue,thenIcanmeaningfullyarrangeboththestudentsandthewordsontoasinglecontinuumsuchthattherelationshipbetweeneachpairofobjectstellsusabouttherelativeabilityofthestudentsortherelativecultyoftheword.ThisexampleofspellingabilityparallelsthesuspiciousthinkingconstructIamattempt-ingtocapture.Althoughthesuspicionitemscontainordinalresponsecategoriesratherthanthedichotomouscorrect/incorrectresponsesinthespellingexample,thebasicprincipleisthesame.Iamaskingwhethersomeindicatorsofsuspiciousthinkingtaphigherlevelsofgeneralsuspicionthanothers,justasIsuggestedwhenoutliningthereasoningbehindtheparticularsuspicionquestions.Inordertotestthiscumulativemodel,Iturntothe50non-parametricIRTmodeldevelopedbyMokken(1971).Non-parametricIRTallowsmetoinvestigatesystematicallywhetherthecumulativepatterndescribedabove(probabilistically)existstotheextentthatIcancomfortablyassertunidimensionalityandthejointscalabilityofrespondentsanditems.ThecriterionofvalidatingthesuspicionscaleisthemagnitudeoftheLoevingersHcot,whichsummarizesinasinglevaluetheextenttowhichthedataacumulativepattern.AccordingtotherulesofthumbdevelopedbyMokken(1971),anHvaluebetween0.40and0.50indicatesa\good"ofthedatatoaGuttman-stylecumulativepattern.Theempirically-derivedLoevinger'sHcotforascaleofthefoursuspicionitemsIemployhereisastrong0.425,suggestingthatwithminorerror(tobeexpectedinalargedataset),thedataacumulativeresponsemodel.Furthermore,thescaleisquitereliable,justasIwouldexpectfromacarefullycraftedmultiple-itemmeasure(Ansolabehere,Rodden&Snyder2008).Indeed,theMolenaar-Sijtsmareliabilityestimateforthescaleis0.707,whichindicatesthatmorethan70%ofthevarianceintheobservedscalescoresissharedwiththetrue{butunobserved{distributionoflocationsalongthelatentcontinuumofsuspiciousthinking3.Finally,adiagnosticexaminationofallassumptionsoftheprobabilisticcumulativemodelrevealednocantmodelviolations.Allevidencetaketogether,Icancomfortablyconcludethattheitemsformastrongunidimensional,cumulativescale4.Beforesubjectingthissuspicionscaletocriterionvalidationviamultivariateanalysis,Iconsiderthepropertiesofthescalealone.First,Iconsiderthey"orderingofthefoursuspicionitemsalongthelatentsuspicioncontinuum,asoperationalizedbythesuspicionscale.Itemyissimplyanestimateofhowsuspiciousone(onaverage)mustbetoagreetothesuspiciouspropositioninagivenitem.Thus,amoreitemisonethat3TheMolenaar-SijtsmacotisamoreappropriateestimateofthereliabilityofacumulativescalethanisCronbach'salpha.Thatsaid,theCronbach'salphaisalso0.70.4Bothandexploratoryfactoranalysesusingt(indeed,lessstringent)as-sumptionsthanthecumulativemodelemployedhere,thatresponsestothefoursuspicionitemscanbeaccuratelyconceptualizedasproductsofasinglelatentdimensionorfactor.Resultsoftheseanalyses,aswellasamatrixofPearsoncorrelationsbetweenpairsofthesuspicionitems,canbefoundintheAppendixofthischapter.51Figure3.2:Distributionofsuspicionscalescores.requiresamoresuspiciousresponse(i.e.,a\somewhat"or\strongly"agreeresponse)fromanindividual.Theorderoftheitems,fromlowestonthescale(leastsuspicious)tohighest(mostsuspicious),isasfollows:1)politicianslie,2)outsideinterests,3)noaccidents,and4)secretdesigns.Notonlydothedataconformtoacumulativepattern,buttheorderhypothesizedabovealsoappearstohold.Next,Iconsiderthedistributionofthescale.Thescaleisoperationalizedasanadditiveindexofthefoursuspicionitems,whichthemselveshavefourresponsecategories,resultinginascalethatrangesfrom0-12.ThedistributionofthescaleisdepictedviaahistograminFigure1.Perhapsthemostinterestingaspectofthescale'sdistributionisitsnegativeskew.Agreatdealmoreoftherespondentsarelocatedabovethescale'smidpoint(6.5)thanbelowit.Indeed,themeanofthescaleis9.08andthemedianscalescoreis9,bothofwhichareaboutthreequartersofthewayupthescale.Furthermore,norespondentsregisteredthe52leastsuspiciousattitudeacrossallquestions,while15%ofrespondentsregisteredthemostsuspiciousattitudeacrossallquestions.Ofcourse,Idonot(andindeedcannot)possesssomeobjectivecriterionagainstwhichtocomparethisscale,soIcanneverbecompletelysurethatIhavecapturedthefullbreadthofsuspiciousthinkinginthemasspublic.However,Icanbequitesurethatmostpeopleatleastsomewhatagreetoeachofthesuspicionitems,andanon-trivialproportionstronglyagreetomostsuspicionitems5.Thisisparticularlynoteworthygiventhefairlyabsolutisttermsinwhichthequestionswerephrased.Thelargeproportionofrespondentsprovidingeanswerstothesequestionssuggests{quitesurprisingly{thatneitherconcernsaboutsocialdesirabilitynorreservationsaboutthestrengthofthestatementsbeingrespondedtoappeartohavenegativelyresponses.Thus,IhaveempiricalevidencethatamuchgreaterproportionoftheAmericanmasspublicispronetohighlysuspiciousthinkingwhenitcomestothegovernmentandpoliticsthanIwouldhavethoughthadIconsideredonlythemarginaldistributionsofresponsestomostquestionsaboutbeliefsinindividualspconspiracies6.3.4.1ValidatingtheSuspicionScaleBeforecompletingmyassessmentofthevalidityofthesuspicionscaleandturningtowardaninvestigationoftheattitudinalandbehavioralconsequencesofsuspiciousthinking,Iestablishthecriterionvalidityofthescalebyexaminingrelationshipsbetweensuspiciousthinkingandtheoreticallyrelatedattitudes,predispositions,andcharacteristics.Thoughverylittleworkhasexaminedtheindividual-levelcorrelatesofaconspiracyscalethatisnotcomprisedofresponsestoquestionsaboutspconspiracies,Idohaveagooddealofliteraturetoguidemeinchoosingasubstantialsetofcriterionvariables.Ofcourse,aquestionremainsofwhethersymbolicpoliticalpredispositions,suchaspartyidenandliberal-conservativeself-idenarerelatedtothegeneralsuspicion5Indeed,morepeoplestronglyagreetoallitemsthanstronglyand/orweaklydisagreetoallitems.6Foracomparativeanalysisofthedistributionsofthespconspiraciesfoundinthe2012AmericanNationalElectionStudy,seetheChapter2.53traitcapturedbythesuspicionscale.Accordingtothedistinctionbetweenconspiratorialthinkingandpartisanbias/motivatedreasoningmadeabove,andthebetweenameasurecomprisedofresponsestoquestionsaboutspconspiraciesratherthanamoregeneralsuspiciousstyleofthinking,Iwouldnotexpectpartisanidentoberelatedtothescale.AlthoughonecouldsayIamover-specifyingthemodelbyincludingthisvariable,Idosoforthegoodreasonoftestingthisimportantproposition.Whenitcomestoideology,however,itcouldbethecasethatconservativesaremorepronetosuspiciousthinking,particularlybecauseitispartoftheconservativemantratodistrustgovernment.Ialsoconsiderwhethertrustinvariousinstitutionsisrelatedtosuspicion.Morespecif-ically,Ihypothesizethatpoliticaltrustandsuspicionarenegativelyrelatedsuchthatastrustininstitutionsdecreases,thepropensitytowardsuspiciousthinkingincreases.Thoughdistrustislikelycloselyrelatedtoconspiratorialthinking,Idonotbelievethattheyarethesamegeneralconstruct.Putsimply,whiledistrustmaybealogically-necessarystepto-wardconspiratorialthinking,itisnotatone.Asmyquestionsthemselvessuggest,itwouldbetosaythatsomeonewhoagreedthattheycouldseesecretdesignsandthatnothinghappensbyaccidenttobemerelydistrustfulofgovernment.And,conversely,individualscanhavereasonsfordistrustinggovernmentthatareunrelatedtoconspiratorialthinking,suchasbadexperienceswheninteractingwithgovernment(Weaver&Lerman2010).Lastly,traditionalmeasuresoftrustarealsopronetothesamepartisancontaminationsthatquestionsaboutbeliefsinspconspiracytheoriesare,furtherdis-pellinganyquestionabouttcollinearitybetweentrustandconspiratorialthinkingwithregardtotheoryormeasurement.Next,Iconsiderthateducationalattainment,andknowledgeandinterestinpoliticsmayberelatedtosuspicion.Iexpectthateducationandsuspicionarenegativelyrelatedsuchthataseducationalattainmentincreases,thepropensitytothinksuspiciouslydecreases.Thisislargelyduetomyexpectationthathigherlevelsofeducationarenegativelyrelatedtotheerrorsinreasoningthattheoreticallyunderwritethescale(Brotherton,French&54Pickering2013).Iapplythesamereasoningwhenitcomestoknowledgeaboutpolitics.Itislessclearhowinterestinpoliticsmayberelatedtosuspicion,however.Onetheonehand,onecouldeasilyconceiveofahighlysuspiciousindividualwhoisalsohighlyinterestedinpolitics.Suchanindividualwouldbemoreinterestedinpoliticsbecauseofaneedtomonitorthegoverningelite.Ontheotherhand,onecouldconceiveofanindividualwhoissosuspiciousofauthoritiesthattheyrejectanyandallinformationthyespouse.Iremainambivalentonthisissue.Finally,Iexaminewhethergeneralreligiosityisrelatedtosuspiciousthinking.Thoughothershavefoundthat\supernatural"or\paranormal"thinkingisrelatedtobeliefinpar-ticularconspiracies(Oliver&Wood2014),Idonotnecessarilyexpecttosuchacon-nection.Religiousindividualsmay,indeed,bepronetobelievingin\unseenforces"andinthe\struggleforgoodandevil,"butconspiratorialthinkingshouldbedistinctfromescha-tologicalnarratives:onecanbesuspiciouswithoutbeingreligious.Infact,bynotincludingreligiouslysymbolicquestionsinthescale,Iwillbeabletodeterminetheproperrelationshipbetweenreligiosityandsuspicion.TotesttherelationshipsoutlinedaboveIregresstheindividualsuspicionscalescoresonindicatorsofallvariablespreviouslydescribed,inadditiontocustomarycontrolsforage,race,andgender.Iconsidertheofthevariablesintheordertheywereoutlinedabove.First,Inotethattheslopecotassociatedwiththepartyidenvariableisnotstatisticallytatconventionallevels.ThisisasIexpectedsinceIseenoreasonwhyasuspiciousmindsetshouldcharacterizemembersofonepoliticalpartymoresothanmembersofanother.Ideology,ontheotherhand,istlyrelatedtosuspiciousthinking.Indeed,theideologyvariableiscodedsuchthatpositivevaluesindicateprogressivelystrongerconservativeidentions.Sincethecotestimateispositive,itdoesappearthatconservativesarepronetohigherlevelsofsuspiciousthinkingthanareliberals.Consideringthehypothesizedrelationshipbetweentrustingovernmentalinstitutionsandsuspicion,Iastatisticallytnegativecotestimate.Ashypothesized,I55Table3.1:OLSregressionvalidatingthesuspicionscale.OLSStandardCotErrorPartisanship0.0140.056Ideology0.1710.068Trust-0.1840.027Religiosity0.0420.035Knowledge-0.2640.117Interest0.2580.126Education-0.2570.061Age0.0100.006Female-0.0870.180Black-0.2980.311Latino-0.1610.399Intercept10.3340.469R20.129n613p<0:05withrespecttoatwo-tailedtest.thatbotheducationandknowledgeaboutpoliticsaretlynegativelyrelatedtosuspiciousthinking.Whenitcomestointerestinpolitics,Iastatisticallytpositiverelationshipsuchthatmoreinterestedindividualsare,onaverage,moresuspicious.ContrarytoOliverandWood(2014),Idonotthatreligiousindividualsaremorepronetobeingsuspiciousofthegovernmentandpoliticians,comyexpectation.Finally,Idonotanytage-,gender-,orrace-basedinlevelofsuspicion,controllingforotherfactors.Takentogether,theserelationshipsprovidestrongevidencethatthescaleIamlabelingsuspiciousthinkingcapturesthelatentpredispositionwhichinformsconspiracybeliefs.Asanevenstrictertestofthevalidityofthescale,Inextconsidertherelationshipbetweenthesuspicionscaleandthreetconspiratorialattitudes7.7AmatrixofPearsonpairwisecorrelationsbetweenresponsestotheconspiracyitemscanbefoundintheAppendixtothischapter.563.4.2SuspicionandSpConspiracyBeliefsIbeginwithconspiratorialbeliefsaboutthelinkbetweenchildhoodvaccinationsandautism.Sincethesuspicionscalewasdesignedtocapturethelatenttraitthatinformsmorespe-conspiracybeliefs,Iexpecttoobserveastrongrelationshipbetweenthetwovariables.Roughly28%ofrespondentsinmysampleeitherstronglyorsomewhatagreedthatchild-hoodvaccinationsarelinkedtoautism.Thisisanon-trivialproportionofthemasspublic.Ialsonotethatthisiscongruentwiththeresultsofothernationalpublicopinionpolls.Mostrecently,a2014HarrisPollfoundthat29%ofalladultsbelievethatvaccinationscancauseautism(League2014).Thus,theproportionofbelieversIreportappearstobeonparwithotheravailableinformationaboutpublicbeliefsonthistopic.Iuseabinarylogisticregressionmodeltoregressbeliefsinthelinkbetweenvaccinationsandautismonthesuspicionscale,liberal-conservativeideology,trustingovernment,interestinpolitics,educationalattainment,andsocio-demographiccharacteristics.Sincemycentralfocusisonthesubstantiveofthesuspicionscale,Iturntoavisualizationofthekeyrelationshipviaapredictedprobabilityplot;however,thefullresultsofthisanalysisarepresentedinthecolumnofTable3.2.ThepanelofFigure2belowdepictsthechangeinthepredictedprobabilityofregisteringabeliefinthevaccine-autismlinkacrosstherangeofthesuspicionscale,holdingothervariablesconstant.Thisrelationshipispositive,suchthatthemoresuspiciousoneis,themorelikelysheistobelievethatvaccinationsforsuchdiseasesasmeasles,mumps,andrubellacancauseautisminchildren.Theprobabilityofbelievinginthevaccine-autismlinkfortheleastsuspiciousindividualsisapproximately0.20,whileitmorethandoublesto0.48forthemostsuspiciousindividuals.Thisisaremarkableparticularlyinlightofthefactthatneitherideologynorpartisanshipistlyrelatedtobeliefinthevaccine-autismlink.Indeed,suspicionisthebestpredictorinthismodel.Next,Iconsiderwhethersuspicionisrelatedtobeliefsaboutthesafenessofgeneticallymofoods.Muchlikewithbeliefsaboutthevaccine-autismlink,agreatdealofcon-57Table3.2:Regressionsofspconspiratorialbeliefsonsuspicionscale.Vaccine-DisasterAutismGMOsReliefSuspicion0.1260.1650.164(0.042)(0.055)(0.044)Partisanship0.087-0.0530.122(0.057)(0.075)(0.062)Ideology0.081-0.0970.234(0.068)(0.091)(0.074)Trust0.063-0.150-0.111(0.028)(0.037)(0.030)Sophistication-0.3010.007-0.078(0.007)(0.160)(0.127)Education-0.167-0.351-0.008(0.062)(0.084)(0.065)Age0.0030.004-0.006(0.006)(0.008)(0.006)Female-0.4020.704-0.240(0.180)(0.236)(0.183)Black0.9010.5980.028(0.209)(0.522)(0.397)Latino1.1820.494-0.583(0.419)(0.482)(0.430)Intercept-1.2720.862(0.610)(0.803)Cutpoint1-1.532(0.638)Cutpoint20.516(0.634)Pseudo-R20.0870.1530.085n615435485Column1&2arebinarylogits,Column3isordered.logit.p<0:05withrespecttoatwo-tailedtest.58troversyintheformofmisinformationandthepromulgationofconspiracytheorieshassurroundedanationalconcernaboutGMOs(e.g.,Kimbrell2015,Nestle2015).Approxi-mately68%ofrespondentsbelievethatgeneticallymofoodsaredangerousbecausetheyareunnaturalandcontrolledbycorporateinterests.ThisestimateiscomparabletothatgarneredbyaJune2015PewSurveythat57%ofadultsthinkthatitisgenerallyunsafetoeatgeneticallymofoods(Center2015a).InordertoinvestigatetherelationshipbetweensuspicionandconspiratorialbeliefsaboutGMOs,Iestimateabinarylogisticregressioninthesameformasthatforthevaccine-autismmodel.ThefullresultsofthisanalysiscanbefoundinthesecondcolumnofTable2.Onceagain,IfocusonthesubstantiverelationshipbetweensuspicionandbeliefsaboutGMOsviaapredictedprobabilityplot.ThesecondpanelofFigure2depictsapositiverelationshipbetweensuspicionandtheprobabilityofregisteringasuspiciousbeliefaboutGMOs:theleastsuspiciousindividualsregisteredaconspiratorialbeliefwithaprobabilityof0.45,whichnearlydoublesto0.82whenconsideringthemostsuspiciousindividuals.Aswiththepreviousanalysis,neitherideologynorpartisanshipexerttonbeliefsaboutGMOsaftercontrollingforotherfactors.Onceagain,therefore,Ihavestrongsupportfortheindependentofsuspiciousnessonsubsequentpoliticalattitudes.Toexaminetheofsuspiciononconspiratorialattitudesthataremoreexplicitlypartisan{orprovideaconservativeexaminationofthepervasivenessofsuspiciousthinking{Iconsiderlasttheextenttowhichsuspicionisrelatedtothebeliefthatpresidentsaremotivatedbypoliticalinterestswhenallocatingdisasterrelief.AlthoughIdonotproberespondentsabouttheirbeliefsinaspconspiracytheory,variousconspiracieshavesurroundedrecentnaturaldisasterreliefonthepartofthepresidentandhisadminis-tration(Mullouy2003,Center2015b).Giventhatthisquestionexplicitlymentionsamajorpoliticalactor,Iexpectthatpartisanshipandideologywillrelatetotheeresponseofpoliticalmotivesdrivingdisasterrelief;however,IneverthelesspositthatsuspicionwillstillexertanindependentRegardlessoftheexactconsiderationsthatcometomind59Figure3.3:Predictedprobabilityofregisteringaspconspiratorialattitudeovertherangeofthesuspicionscale.whenconfrontedwiththisquestion,approximately42%ofrespondentsbelievethatthepres-identisgreatlybypoliticalconcerns,suggestingagaintheubiquityofsuspiciousthinkinginthemasspublic.Itestwhethersuspicionthisbeliefaftercontrollingforpartisanshipandide-ologyviaanorderedlogisticregressionanalysis,presentedinthethirdcolumnofTable2.Thepredictedprobabilityofregisteringthemostconspiratorialattituderangesfrom0.16fortheleastsuspiciousindividualsto0.51forthemostsuspiciousrespondents.ThisisthestrongestIhaveuncoveredthusfar;indeed,themostsuspiciousindividualsare,onaverage,morethanthreetimesmorelikelytoregisterthemostconspiratorialattitude,con-trollingforpartisanandideologicalconsiderations.Takentogether,Istrongsupportfortheargumentthatthepropensitytoviewtheworldthroughthelensofsuspicionisstronglyrelatedtospconspiratorialbeliefs8.Inotherwords,thesuspicionscalenicelycaptures8Theobjectofthischapterisnottodispelpossiblemisperceptionsregardingtheveracityofthesespbeliefs.Myobjecthereissimplytoexamineandexplaintheextenttowhichindividualsacceptthesebeliefs,hypothesizingthatthroughtheseconspiracybeliefsIcanunearthandvalidatetheindependentroleofsuspicion.Thechoiceofthesecontroversialtopicsdoesnotdiminishtheconsistentregardingtheroleofsuspicion.Determiningwhatisandwhatisnotaconspiracybeliefisultandwellbeyondthescopeofthisproject.However,Ifocusontheseparticularbeliefsbecauseinpublicdiscoursetheyareconsideredtobe\conspiratorial,"andthisprojectempiricallyexamineswhat\conspiratorial"entails.60thegeneralpredispositionwhichinformsspconspiracybeliefs.3.4.3RobustnessCheckAlthoughIhavenoparticularreasontobelievethateitherthesampleorthespmea-surementstrategyemployedinthepreviousanalysesisproblematicinanyway,Ipresentinthissectiontheresultsofanalysessimilartothoseabovewithslightlytoperational-izationsofthesuspicionconstruct.SincetheoperationalizationIpresentaboveiscompletelyunique,Ibelieveitisausefulexercisetoconsidertheresultsofsimilaranalysesusingslightlyalteredmeasurementstrategiesandt{albeitlessrepresentative{samples.Moresplly,Ialterboththewordingofthesuspicionstatementsandtheresponsecategoriesprovidedtorespondents9.Ratherthanrestrictrespondentstoonlydirectionalchoices(i.e.,(dis)agreeingwithsomelevelofconviction),Iincludea\neitheragree/nordisagree"midpoint,providinga5-point,ratherthan4-point,setofresponsecategories.Thealteredstatementwordingsareafollows:\Innationalpolitics,eventsneveroccurbyaccident."\Politiciansrarelylie."\Unseenpatternsandsecretactivitiescanbefoundeverywhereinpolitics."\Governmentinstitutionsarelargelycontrolledbyeliteoutsideinterests."Notethatinadditiontoslightlyalteringthewordingsincontent(though,themainideasshouldbeidentical),Ireversedthedirectionofthe\politicianslie"and\noaccidents"statementssuchthatmoresuspiciousindividualsshouldmorelikelytodisagreewiththe9Ialsotestedthefollowingstatement:\politicsisaboutgoodversusevil."AccordingtoOliverandWood(2014),conspiracytheoristsarecharacterizedbyageneralbeliefinunseen,religiousforces{angels,theDevil,andghostsinparticular.Thoughsomeimaginationisrequiredto\seepatterns,designs,andsecretactivitieseverywhereonceyouknowwheretolook,"itismybeliefthatprimarilyreligiousideasaretthanmoregeneralconspiratorialideas,altogether.ContrarytoOliverandWood'sclaims,thereisnothingmagical,mystical,orreligiousaboutaircraftvaportrails,presidentialassassinations,orglobalwarming.TheincongruencebetweentheresultspresentedhereandthosepresentedbyOliverandWoodisaprimeexampleoftheinferentialproblemscausedbytheresearchers'selectionofspconspiracytheoriesaboutwhichtoinquire.61statements.AlthoughIuncoverednoevidenceofresponsesetbiasintheCCESsample,Iprovideforthepossibilityofsuchabiasbyalteringmymeasurementstrategyinthisway.ThedataemployedinthisstudycomesfromsameMTurksampleof1543U.S.adultsemployedinthepreviouschapter.Informationinthe\Study2"sectionofthepreviouschapter,particularlyTable4,includesmoredetailsaboutthedemographiccharacteristicsofthesample.Figure4depictsthedistributionsofresponsestothealternatesuspicionquestions.Al-thoughthedistributionofbeliefsregardingthe\politicianslie"statementisstronglynega-tivelyskewedasitisintheCCESdatapresentedabove,therestofthedistributionsexhibitvaryingIndeed,thedistributionsofthe\outsideinterests"and\secretdesigns"responsesarenowslightlypositivelyskewed,whilethe\noaccidents"distributionrevealsanaggregateambivalenceabouttheprevalenceoftrueaccidentsinpolitics.Inotherwordsthereistlymoredisagreementwiththesestatementsinthecaseofatsample,questionwordings,andresponsecategories,thanindatapresentedabove.Responsestothequestionsdo,however,stillloadhighlyonasinglelatentfactor.Anexploratoryfactoranalysisrevealsasinglefactoraccountsforapproximately85%ofthevarianceoftheindividualitemresponses(similartothe88%intheCCESdata),andafactoranalysissupportsthenotionofasinglelatenttraitbeingcapturedbytheseitems.Sincetheitemsstillserveasindicatorsofasingleconstruct,Iamcotthatthealterationofquestionwordingsorresponsecategoriesdidnotsubstantivelyaltertheinterpretationsoftheirmeaningonthepartoftherespondent.ItismorelikelythatthemoreeducatedrespondentsintheMTurksamplerecognizedtheconspiratorialelementofthesequestionsandoptedtoprovidethelessconspiratorialresponse(though,fewpeoplestronglydisagreedwithanygivenquestion).Unsurprisingly,giventhedistributionsoftheindividualalternatesuspicionitems,theMokkenscalecomprisedoftheseitemsismorenormallydistributedthanthenegativelyskewedonepresentedabove.ThisdistributionisdepictedinFigure5.Sincethereare562Figure3.4:Distributionofresponsestoalternatesuspicionquestions.responsecategoriesavailableinthealternatequestions,thescaleinthisdatarangesfrom0-1610.Themeanandmedianscalescoresare7.82and8,respectively.Eventhoughthesedistributionslowerlevelsofsuspicionthandothedistributionspresentedabove,theseshouldnotbeover-interpreted;indeed,moretestsmustbedonetodecipherwhatmaybectingobservedresponses.Afterall,therewerefourchangesmovingfromtheCCEStotheMTurkdata:questionwording,directionality,avail-ableresponsecategories,andsurveysample.ThereisalsoreasonforoptimismdespitetheobservedTheofsuchreasonsisthatanormaldistributionstillindicatesthatasubstantialproportionofindividualsholdsuspiciousbeliefs.ItisnotthecasethattheMTurkdatarevealsthatindividualsaresimplynotsuspicious,butthattheymaybelesssuspiciousthantheCCESdatamightleadustobelieve.And,thoughrespondentsmightbe10Theresponsecategoriesarearrangedsuchthatthemoresuspiciousattitudesareassignedhighervalues.Thentheresponsecategoriesareassignedavaluefrom0-4andsubsequentlysummed.63Figure3.5:Distributionofsuspicionscalescores.morewillingtomoderatestatedsuspiciousbeliefswhenprovidedaneutralcategory,theytendtowardthemoresuspiciousbeliefswhenneutralityisnotanoption.ThismaybeanindicatorofasocialdesirabilitythatisattenuatingstatedsuspiciousbeliefsintheMTurkdata.Finally,thescaleconstructedoftheMTurkdatapredictsbeliefsinspconspiracytheoriesquitewell.Table3includesthreeregressionsofspconspiracybeliefsontheMTurksuspicionscaleonahostoncontrolvariables.EachofthequestionsisaskedinthesameformatastheANESconspiracyquestionswherebyindividualscancanregisterdirectional,probabilisticbeliefsinthe\birther,"\truther,"andJFKassassinationconspiracytheories11.IincludethesamecontrolvariablesinthesemodelsasinthemodelspresentedinTable2above.OLSwasusedtoestimatethecoients12.11Thevaccine-autismregressionfromabovealsoreplicatesintheMTurksample,asdoesthecorrelationbetweensuspicionandtheperceivedproportionofcorruptgovernmentIomitthefullanalysestosavespace.12ThesamplesizesaremuchsmallerthaninthepreviousMTurk-basedanalysesbecauseeachrespondent64Table3.3:Regressionsofspconspiratorialbeliefsonalternatesuspicionscale(MTurk).TrutherJFKBirtherSuspicion0.0820.2010.0710.1820.0290.084(0.022)(0.022)(0.017)Partisanship0.0070.016-0.019-0.0440.0570.151(0.044)(0.041)(0.037)Ideology-0.052-0.1200.0420.1040.1470.405(0.043)(0.039)(0.035)Trust-0.151-0.140-0.149-0.144-0.066-0.072(0.058)(0.058)(0.045)Knowledge-0.160-0.210-0.051-0.073-0.056-0.093(0.041)(0.038)(0.030)Authoritarianism-0.020-0.045-0.007-0.0170.0240.062(0.025)(0.024)(0.019)Education-0.130-0.1290.0030.003-0.033-0.036(0.054)(0.055)(0.043)Age-0.008-0.1110.0090.1230.0000.003(0.004)(0.004)(0.003)Female-0.063-0.0740.147(0.093)(0.095)(0.073)Black0.2030.2070.128(0.176)(0.169)(0.188)Latino0.1820.378-0.167(0.223)(0.194)(0.160)Intercept3.4052.2261.859(0.392)(0.399)(0.307)R20.2080.1160.393n322341304OLScotsw/standarderrorsinparenthesesincolumns1,3,and4.Standardizedcotsincolumns2,4,and6.=p<0:05withrespecttoatwo-tailedtest.65Suspicionisastatisticallyt,substantivelystrongpredictorof\truther"andJFKassassinationbeliefs.Accordingtostandardizedestimates,suspicionisthestrongestpredictor13ofJFK-relatedconspiracybeliefs,and,moreorless,tied{withknowledge{forstrongestpredictorof\truther"beliefs14.UnsurprisinglygiventheanalysesinChapter2,neitherpartisanshipnorideologicalself-idenarestatisticallytpredictorsofthesetwoconspiracybeliefs.Conversely,suspicionisnotastatisticallytpredictorof\birther"beliefs15,whileideologicalself-idenaretightlyrelatedtosuchbeliefs.Thus,thesuspicionscalebehavespreciselyaswewouldexpectgivenboththeanalysespresentedearlyinthischapterandthosecomveyedinChapter2.3.5ConclusionGiventhatsuspicion/conspiracismare{atleast,seemingly{increasinglyimportantfeaturesofcontemporaryAmericanpoliticalculture,howshouldpoliticalscientistsunderstandthis\paranoidstyle"?Manyprevioussuggestingthatconspiratorialthinkingistheconsequenceofsomeminoritystatusaretheproductofstudiesemployingresponsestoquestionsaboutbeliefsinspconspiracytheoriesthatseverelycomplicateinferencesaboutconspiratorialthinking.Thisgeneralmeasurementstrategyis,however,pronetocapturingpartisanbiases,andbeingbysocialdesirabilityconcernsandothertypesofbiasesduetothecontentandcontextassociatedwithspconspiracytheories(thetypesofwhichvarytremendouslyfromstudytostudy).Inlightofthesemeasurementconcerns,Irecognizedthatanysubstantialstepforwardinexaminingconspiratorialthinkinginthemasspublicmustbeginwithabandoningtheproblematicmeasurementstrategyinfavorofdevelopinganewonethatcapturesthelatentonlyreceivedoneofthethreespconspiracybeliefquestions.13Thepairwisecorrelationis0.237.14Thepairwisecorrelationis0.316.15Thereis,however,astatisticallytcorrelationbetweensuspicionand\birther"beliefs(0.137).Thecorrelationbetween\birther"beliefsandpartisanship(0.476)andideologicalself-iden(0.573)aresimplymuchstronger.66suspiciontraitthatpromotesconspiracybeliefsandshapesone'soutlookofthepoliticallandscape.Bysynthesizingtherobustworkonthephilosophyofconspiratorialthinking,Iwasabletoconstructavalidmeasureofthegeneralpropensitytowardconspiratorialthinkingthatislargelyfreeofthemajorconcernsoverinherentmeasurementerrorthathaveplaguedpreviousstudies.Moresp,Idemonstratedthatthescalewasbothunidimensional,whichoperationalizesthemonologicalpropertyoftheconspiracistbeliefsystemhighlightedbymanytheorists,andcumulative,thehypothesizedorderingoftheparticularitemswithinthescale.Themarginaldistributionofthesuspicionscalealsorevealedthatatproportionofthepopulationispronetohighlevelsofsuspiciousthinking.Indeed,itappearsthatthetendencytowardconspiratorialthinkingisquiteprevalentinAmericanpoliticalculture.Finally,Ifoundthatsuspicionwaspositivelyrelatedtovariousspconspiratorialbeliefswithfocirangingfrompublichealth,topresidentialdecision-making,toassassina-tions.Thestrengthoftherelationshipbetweensuspicionandspconspiratorialbeliefsispotentiallyalarminggiventhehighproportionofthepopulationthatappearstobeengagedinhighlevelsofsuspiciousthought.Ofcourse,thegeneralpredispositiontowardsuspicioncertainlydoesnotguaranteeindividualendorsementinall(orevenmost)spconspiracytheories;but,thepsychologicalgroundworkforthesebeliefsisnonethelesswell-established.AlthoughthegeneralmeasureofconspiratorialthinkingthatIpresentaboveisasignif-icantimprovementuponpreviousattempts,muchimportantworkremainstobedone.Forexample,whilemyfocushasbeenexclusivelyonAmericanpoliticalbeliefs,thereislittlereasontobelievethatsuspicionisauniquelyAmericanphenomenon.Ofcourse,thisisatestablepropositionthatthemeasureofgeneralsuspicionisuniquelysuitedtotest.Inaddi-tion,aswithanynewscale,Ithinkitimportanttocontinuetotestandthemechanicsofthescale.Moreover,whilesuspicionisaofourpoliticalthinking,cross-sectionaldatacannottellusifweareinapoliticalenvironmentcharacterizedbyparticularlyhighorlowlevelsofconspiratorialthinking.67APPENDIX68Variablecoding/wordingforCCESvariablesDoyouagreeordisagreewiththefollowingstatements?0-3,where0-Stronglyagree,3-Stronglydisagree(a.)Thereisalinkbetweenchildhoodvaccinationsandautism.(UOG309_C)(b.)Governmentinstitutionsarecontrolledlargelybyeliteoutsideinterests.(UOG309_E)(c.)Innationalpolitics,nothinghappensbyaccident.(UOG309_I)(d.)Nothingisasitseems.Politiciansoftenlie,blameandotherwaystolookinnocent.(UOG309_L)(e.)Innationalpolitics,youcanseepatterns,designsandsecretactivitieseverywhereonceyouknowwheretolook.(UOG309_N)Howmuchtrustdoyouhaveineachofthefollowinginstitutionstodotherightthing?0-4,where0-noneatall,4-Agreatdeal(a.)TheCIA(UOG308_A)(b.)Themilitary(UOG308_B)(c.)TheSupremeCourt(UOG308_C)(d.)Bigbusiness(UOG308_D)Whichstatementcomesclosertoyourview?(UOG402)0.Geneticallymofoodsareapositivedevelopmentbecausetheyarecheaper,nutritional,andmoreresistanttoweatherconditionsandpests.1.Geneticallymofoodsaredangerousbecausetheyareunnatural,notashealthyasorganicfoods,andarecontrolledbycorporateinterests.69Eachpresidentsincethe1950shasbeencalledupontoprovidedisasterrelieftostatesbysomesortofnaturaldisaster(suchasanearthquakeortornado).Whichofthefollowingstatementsmostcloselyyourviewofthisprocess?(UOG408)0.Presidentsarenotbypoliticalmotiveswhenallocatingdisasterrelief.1.Presidentsareonlysomewhatbypoliticalmotiveswhenallocatingdisasterrelief.2.Presidentsaregreatlybypoliticalmotiveswhenallocatingdisasterrelief.WhatisyourviewoftheTeaParty?(CC424)0.Verynegative1.Somewhatnegative2.Neutral3.Somewhatpositive4.Verypositive7-pointPartyID(pid7)-3-3,where-3strongDemocrat,3strongRepublicanStrengthofpartisanidencations(pid7)0.Independent1.Leaner2.Weak3.Strong7-pointIdeology(CC334A)-3-3,where-3extremelyliberal,3extremelyconservative70Importanceofreligion(pew_religimp)0-3,where0notatallimportant,3veryimportantChurchattendance(pew_churatd)0-5,where0never,5morethanonceaweekPoliticalInterest(newsint)0-3,where0Hardlyatall,3MostofthetimeDuringthepastyear,didyouattendpoliticalmeetings?(CC417a_1)0.No1.YesDuringthepastyear,didyouputupapoliticalsign?(CC417a_2)0.No1.YesDuringthepastyear,didyouworkforacandidateorcampaign?(CC417a_3)0.No1.YesDuringthepastyear,didyoudonatemoneytoacandidate,campaign,orpoliticalorganization?(CC417a_4)0.No1.YesKnowledge:whichpartycontrolstheU.S.HouseofRepresentatives?(CC14_309a)0.Incorrect711.CorrectKnowledge:whichpartycontrolstheU.S.Senate?(CC14_309b)0.Incorrect1.CorrectEducation(educ)0-5,where0nohighschool,5postgradBlack(race)0.Non-black/AfricanAmerican1.Black/AfricanAmericanLatino(race)0.Non-Latino1.LatinoAge(birthyr)AgeinyearsFemale(gender)0.Male1.Female72CorrelationsBetweenSuspicionItemsTable3.4:Correlationsbetweensuspicionitems,2014CCES.(1)(2)(3)(4)(1)PoliticiansLie1.000(2)OutsideInterests0.4261.000(3)NoAccidents0.3270.2951.000(4)SecretDesigns0.3840.3260.4021.000EntriesarePearsoncorrelationcoets.Alltatp<0:05level.73ExploratoryFactorAnalysisofSuspicionItemsTable3.5:ExploratoryfactoranalysisofCCESsuspicionitems.Factor1Factor2Factor3UniquenessFactorLoadings:PoliticiansLie0.663-0.195-0.0900.514OutsideInterests0.605-0.2030.1050.582NoAccidents0.5640.2220.1010.623SecretDesigns0.6460.198-0.0940.535Eigenvalue:1.5410.1680.038VarianceAccountedFor:88.24%9.59%2.17%Note:Methodofestimationisiteratedprincipalfactors.n=707.Table3.6:ExploratoryfactoranalysisofMTurksuspicionitems.Factor1Factor2Factor3UniquenessFactorLoadings:PoliticiansLie-0.4170.3170.0220.726OutsideInterests0.669-0.0970.0710.538NoAccidents0.4360.3150.0190.710SecretDesigns0.6870.087-0.0680.516Eigenvalue:1.2830.2170.011VarianceAccountedFor:84.96%14.35%0.70%Note:Methodofestimationisiteratedprincipalfactors.n=1;479.74FactorAnalysisofSuspicionItemsTable3.7:factoranalysisofCCESsuspicionitems.StandardizedStandardEstimateErrorFactorLoadingsPoliticiansLie0.6530.034OutsideInterests0.5960.035NoAccidents0.5450.037SecretDesigns0.6270.035FitStatistics˜2(7df),p-value13.14,0.001RMSEA0.089Prob(RMSEA0.05)0.060SRMR0.023CFI0.976TLI0.927n707Note:=p<0:001withrespecttoatwo-tailedtest.75Table3.8:factoranalysisofMTurksuspicionitems.StandardizedStandardEstimateErrorFactorLoadingsPoliticiansLie-0.3930.029OutsideInterests0.6630.030NoAccidents0.4140.028SecretDesigns0.6910.030FitStatistics˜2(7df),p-value27.35,0.000RMSEA0.093Prob(RMSEA0.05)0.009SRMR0.028CFI0.964TLI0.891n1479Note:=p<0:001withrespecttoatwo-tailedtest.76CharacteristicsofSpConspiracyBeliefQuestionson2012ANESThefollowinghistograms(FigureA1)depictthedistributionsofquestionsaboutthebirth-placeofBarackObama(labeled\birther"),theinclusionofaprovisionauthorizingthecreationofpanelstomakeend-of-lifedecisionsforpeopleonMedicareinthe2010ableCareAct(labeled\deathpanels"),andtheamountofrelated-knowledgethefederalgovernmentpossessedpriortotheterroristattacksofSeptember11,2001(labeled\truther").Asmytheorywouldpredict,thedistributionsofbeliefsabouttheseconspiraciesarequitet.Aretheseduetovariationinstimuli(i.e.,Obama,healthcare,9/11)?Or,perhapspoliticalcontext(aRepublicanwasintheWhiteHouseduringthe9/11attacks,whileaDemocratusheredthroughhealthcarereform)?Thesearepreciselythequestionswemustattempttoanswerandcontrolforifweemploythisgeneralstrategyformeasuringconspiracytheorists.77Figure3.6:DistributionofANESconspiracyitems.78Unsurprisingly,giventhewidelyvaryingshapesofthedistributionsofresponsestotheANESconspiracyitems,thecovariatesofthesespbeliefsalsovarysubstantially.In-deed,partyidenpoliticalsophistication,thecognitiveneedtoevaluate,religiosity,age,andraceallexhibitvaryingwithrespecttostatisticalalone.Thesizesoftheconsistentlystatisticallyantvariables(i.e.,ideology,trust,author-itarianism)alsovaryacrossconspiracies.Finally,Ishouldnotethattheresponsestothefollowingconspiracydonoteventapasingleconspiratorialthinkingtrait(Milleretal.forth-coming),furtherdemonstratingthegreatvariabilityinthecharacteristicsofvariousspconspiracytheories.Itisquitelikelythatthedacrosstheseconspiracytheoriesaresmallerthanwemightobservewithothercomparisonsgiventheovertlypoliticalandtemporallyrecentnatureofthisset.Both\birther"and\deathpanel"beliefshavetodowiththesittingpresident,and\truth"beliefsareassociatedwithawell-known,andfairly\young,"politicalevent.IitwhollyplausiblethatbeliefsinconspiraciesaboutPrincessDiana,indrinkingwater,andArea51{allofwhichhavebeenregularlyemployedinrecentstudies{wouldhaveveryerentpredictorsgiventhevaryingculturalandpoliticalcontextwithwhichtheyareassociated.79Table3.9:OLSregressionsdemonstratingthecovariatesacrossspconspiracytheories.DeathBirtherPanelsTrutherPartisanship0.1510.1160.008(0.010)(0.012)(0.012)Ideology0.0500.078-0.057(0.015)(0.018)(0.017)Trust-0.156-0.070-0.123(0.020)(0.024)(0.023)Sophistication-0.052-0.016-0.138(0.024)(0.029)(0.028)Authoritarianism0.1180.1100.052(0.015)(0.018)(0.017)NeedtoEvaluate0.0570.0590.029(0.020)(0.024)(0.023)Religiosity0.0160.0330.006(0.008)(0.009)(0.009)Education-0.111-0.097-0.097(0.016)(0.019)(0.018)Age-0.000-0.006-0.004(0.001)(0.001)(0.001)Female-0.003-0.0740.005(0.034)(0.040)(0.039)Black-0.2510.0640.193(0.053)(0.064)(0.061)Latino-0.1090.0740.159(0.050)(0.061)(0.057)Intercept0.9111.4131.571(0.095)(0.113)(0.108)R20.3240.2000.081n219921252229OLScotswithstandarderrorsinparentheses.p<0:05withrespecttoatwo-tailedtest.80CorrelationsBetweenSpConspiracyBeliefsTable3.10:Correlationsbetweenbeliefsinspconspiracytheories,2014CCES.(1)(2)(3)(1)Vaccine-Autism1.000(2)GMOs0.1861.000(3)DisasterRelief0.1260.0781.000CellentriesarePearsoncorrelationcots.p<0:05levelwithrespecttoatwo-tailedtest.81Chapter4:TheAttitudinalandBehavioralofConspiratorialThinkingInthissubstantivechapter,Iconsiderthepotentialconsequencesofconspiratorialthinking.Whereweknowfairlylittleaboutthecharacteristicsoftheconspiratorialmindsetorthoseindividualspronetosuchapredisposition,weknowevenlessaboutthepracticalandnormativeconsequencesofconspiracism.Veryrecentwork{allwithinthelast3years{suggeststhatcertaintypesofconspiratorialbeliefsareassociatedwithdecreasedinten-tiontovaccinateone'schildren(Jolley&Douglas2014a,Nyhanetal.2014),reduceone'scarbonfootprint(Jolley&Douglas2014b),andparticipateincampaignactivities(Jolley&Douglas2014b,Uscinski&Parent2014,Uscinski,Klofstad&Atkinsonforthcoming),aswellaslowerlevelsofgovernmentaltrust(Einstein&Glick2015)andwillingnesstoacceptscien(Lewandowsky,Gignac&Oberauer2013).Takentogether,thereissomeevidencethatspconspiratorialbeliefscanaactionsrelatedtothetopicofthespe-conspiracybeingexamined,andthatconspiracybeliefsareassociatedwithlowerlevelsofengagementinpolitics.Thisdearthofempiricalevidencefortheconsequencesofconspiratorialthinking,whileacceptablegivenboththerelativeyouthoftheresearchareaandthemeasurementproblemsthatmakeinferencesproblematic,isstilltroubling.Unfortunately,wehavefairlylimited{andcertainlyinconclusive{evidencethatconspiratorialthinkingreallymattersforpoliticsorsocialbehaviors,ingeneral.Ifconspiracytheoristsare\odd"inmanyways,butfairlyinnocuouswhenitcomessocio-politicalmatters,itwouldbenaturalandwhollyappropriatetoquestionwhyweshoulddevoteresourcestounderstandingthem.InthischapterIdemonstratethreepotentialconsequencesofconspiratorialthinking.The82regardstheconditioningroleofconspiratorialthinkingwhenitcomestotherelation-shipbetweenpartisanself-idenionsandattitudesaboutthegeneralrolegovernmentandtheappropriatebalancebetweengovernmentservicesandspending,supportfortheTeaParty,andgeneralfeelingstowardthefederalgovernment.Ithatwhilepartisanpredis-positions,unsurprisingly,arestronglyrelatedtotheseattitudesandfeelingsonaverage,themarginalofpartisanshiponthesehighlypartisanorientationsisgreatlyattenuatedasconspiratorialthinkingincreases.Thisdemonstratesthepotentconditioningofconspiratorialthinkingontraditionalpoliticalthought{conspiracytheoristsdonotlinkpartisanorideologicalpredispositionstotheirattitudesaboutvariousaspectsofgovernmentandgovernance.Next,Iconsiderwhetherconspiratorialthinkingisassociatedwithapprovalratingsofmajorgovernmentalactorsandinstitutions:BarackObama,Congress,andtheSupremeCourt.Astheorywouldpredict,approvalofeachofthesepoliticalobjects{eachofwhichcontainsatpartisanvalenceorlevelofpoliticalcontent{decreasesasconspiratorialthinkingincreases.Thus,conspiratorialthinkingisfoundtotraverseanypartisantendenciesandresultinageneraldisapprovalofallbranchesofgovernment.Finally,Iexpandpreviousworkontherelationshipbetweenconspiracybeliefsandpoliti-calengagement.Morespcally,Iconsidertherelationshipbetweenconspiratorialthinkingandone'sstrengthofidenwiththetwomajorparties,supportforanti-establishmentideasliketheTeaParty,andcampaignactivitiesundertakenduringthe2012presidentialelection.InallcasesIthatconspiratorialthinkingispositivelyrelatedtoengagementinanti-establishmentideaslikesupportfortheTeaParty,andnegativelyrelatedtosupportforoneofthetwomajorpartiesandparticipationintraditionalcampaignactivities.4.1TheConditioningofConspiratorialThinkingBeyondservingasthelatentsourcesofbeliefsinspconspiracytheories,conspiratorialthinkingandpartisanshipmayinteractintways.Whilethestatisticalandsubstan-83tiveofpartisanshiponamajorityofvariablesofinteresttopoliticalscientistsarewell-documented,thereexistsnotheoryormodelofhowpartisanshipandconspiratorialthinkinginteractinsubsequentattitudesandbehaviors.Inotherwords,conspir-atorialthinkingandpartisanshiphaveneverbeenconsideredasindependentvariablesinamodelofsomepoliticalphenomenon.Thatthisisthecaseistothedetrimentofpoliticalbehaviorresearch.Whileweknowthatthesubstantiveofpartisanshipareasbroadastheyaresizable,wehavenoideahowpartisanship{theallencompassingsocialidentity,groupattachment,andperceptualscreen{mightinteractwithconspiratorialthinkingwhenitcomestoattitudesandbehaviorsotherthanbeliefsinspconspiracytheories,whichIdocumentedinChapter2.Althoughthereexistsnoliteraturethatcanguidemeonthistopicsp,well-establishedinformationonsomeoftheknowntraitsofconspiracytheoristsmayproveofuse.Mostimportantly,weknowthattheconspiratorial-mindedtendtobesocio-political\outsiders"whofundamentallydistrustestablishedinstitutionsandtheauthoritiesthatpromulgatethem(e.g.,Goertzel1994,McHoskey1995).And,ofcourse,itisnosecretthattheAmericanpoliticalsystemiscontrolledbytwopowerfulpoliticalparties.Sincetheconnectionbetweenthemajorparties(andthepoliticalactorsmostwidelyassociatedwiththem)andsystemiccontrolofmostpoliticalinstitutionsaresointimatelyconnected,itwouldcomeasnosurprisethattherelationshipbetweenpartisanshipandotherattitudesishighlycontingentonawillingnesstoaccept,oratleastengagewith,establishmentpolitics.Inotherwords,theofpartisanship,whereitevenexistsintraditionalforms,forthehighlyconspiratorialmaybesubstantiallyattenuated.Inlightofourcompletelackofunderstandingoftherelationshipbetweenconspiratorialthinkingandpartisanshipwhenitcomestoattitudesandbeliefsaboutthingsotherthanspconspiracytheories,Iinvestigatetheindependentandinteractiveofthesetwoconstructsonattitudesabouttheappropriateroleandreachofgovernment,generalfeelingstowardthefederalgovernment,andsupportforthenotoriouslyanti-establishment84TeaPartyusingdatafromthe2012ANESalongwiththefactormodelofconspiracybeliefspresentedinChapter2.Whilethechoicesetofspattitudestoinvestigateisquitelarge,Iconsidergeneralquestionsaboutattitudesaboutgovernmentbothbecausetheyareattitudesaboutfundamentalpoliticalideasandinstitutions,andbecausepartisanshipisknowntobestronglyrelatedtothem,establishingthisanalysisasaconservativetestofsorts.Althoughthisportionofmyanalysisislargelyexploratory,Iexpect,giventhepresumablyorthogonalnatureoftherelationshipbetweenconspiratorialthinkingandpartisanship1,thatconspiracismheavilymoderatesthetiebetweenpartisanpredispositionsandtheseattitudesaboutgovernment.Morespec,Ihypothesizethatasthelevelofconspiratorialthinkingincreases,therelationshipbetweenpartisanshipandattitudesaboutthescopeandroleofthegovernmentdissolves.InthefollowingsetofanalysesIemployfourtdependentvariablestoexaminetheinteractiverelationshipbetweenconspiratorialthinkingandpartisanship.Feelingstowardnationalgovernmentareoperationalizedverysimplyviathefamiliarfeelingthermometer.Higherscoresonthisitemcorrespondtomorepositivefeelingstowardthegovernmentandlowerscorescorrespondtomorenegativefeelings.Attitudesaboutthescopeandroleofgovernmentareoperationalizedviaanadditiveindexofattitudesabout1)thesizeofgov-ernment,2)preferenceforgovernmentalvs.freemarketsolutionstoeconomicproblems,3)whetherthegovernmentshouldbedoingmoreorless,ingeneral,and4)howmuchgov-ernmentregulationisgoodforsociety2.Highervaluesindicatemoreconservativeresponses.Attitudesaboutthegovernmentspendingandservicesaremeasuredviathefamiliarseven-pointitemwhichasksrespondentstoplacethemselvesonscalerangingfromthereductionofgovernmentservicesandspending(lowvalues)totheincreaseingovernmentspendingandservices(highvalues).Finally,supportfortheTeaPartyismeasuredviaaseven-point1ThisassumptionoforthogonalityenjoyssomeempiricalsupportinChapter2.Indeed,thebetter-factoranalysismodelofthelatentsourcesofspconspiracybeliefsrestrictedthepartisanshipandconspiratorialthinkingfactorstobeuncorrelated.2Theseitemsscaletogetherverywell.Anexploratoryfactoranalysisoftheitemsresultedinaverylargeeigenvalueforthefactor,whichaccountedforapproximately94%ofthevarianceinthefouritems.Cronbach'salphais0.72.85scalerangingfromstrongopposition(lowvalues)tostrongsupport(highvalues).Eachofthesevariableshasbeenre-scaledsuchthattheyrangefrom0-1.Inordertoempiricallytestboththeindependentandinteractiveofconspirato-rialthinkingandpartisanship3oneachoftheseattitudesaboutgovernment,IestimatedaseriesofOLSregressions4.First,Iestimatedfourmodelswheretheofconspiratorialthinkingandpartisanship,aswellasallrelevantcontrols,exertonlyadditiveoneachofthefourdependentvariables.ForeachmodelIincludecontrolsfortrustingovernment,ideologicalself-idenpoliticalsophistication(acombinationofinterest,campaignactivity,andknowledge)5,age,education,income,race,andgender.Ialsocontrolforracialresentment6duetoitsknownrelationshipwithTeaPartysupport(Barreto,Cooper,Gon-zalez,Towler&Parker2012,Parker&Barreto2013),governmentspendingattitudes(e.g.,Jacoby2000),andfeelingstowardBarackObama(e.g.,Kam&Kinder2012),aswellasreligiosity7andauthoritarianism8.ThoughIammoreinterestedinamodelwherethe3Thepartisanshipvariableincludedinthesemodelsisthefactorscorederivedfromthefactormodelpresentedabove.Employinginsteadthetraditionalpartyidenmeasure(aseven-pointscaleconstructedviarespondentanswersto2questions)doesnotsubstantivelyaltertheresults.Sincemultiple-itemmeasuresprovidemorereliablemeasuresofinherentlyerror-proneconcepts(Ansolabehere,Rodden&Snyder2008),Ielectedtopresentmodelswithmultipleindicatoroperationalizationofpartyiden4WhileIcouldhavesimultaneouslyestimatedafull\structural"modelwherebythefac-toranalysispresentedaboveisintegratedintoasetofequationspredictingeachofthesefourindependentvariables,onlyspecializedstructuralequationmodeling(SEM)software(suchasMPLus)iscapableofcal-culatinginteractiontsbetweenlatentvariablesthatareestimatedwithinthesameequation.Therefore,IwouldbeunabletotestinteractivehypothesesinafullSEMframeworkusingmostsoftware(i.e.,Stata,AMOS,andR).Sincethistypeofmodelingisstillbeingandoutsidethegripsofmostpoliticalsciencepractitioners,Ielectedtoestimatethemodelsina(programmatically)widely-availableandeasily-interpretableOLSframework.SincemeasurementerrorinthelatentvariablesisalreadyaccountedforintheCFAframeworktoamuchbetterextentthanitwouldhavebeenhadIsimplyconstructedanadditiveindexalaMilleretal.forthcoming)oremployedtheconspiracyquestionsindividually(Ansolabehere,Rodden&Snyder2008),Ihavelittleworrythatthisapproachposesathreattothevalidityofourinferences.Indeed,theuseoffactorscores(orsimilarlatentvariablescoressuchasthoseproducedfromincreasinglypopularitemresponsetheory,or\idealpoint,"models)insubsequentanalysesiscommonplace.5Cronbach'salphais0.72.6RacialresentmentismeasuredviaanadditiveindexofthecommonfourquestionsfoundontheANES.Cronbach'salphais0.80.ExactquestionswordingsappearintheAppendix.7Religiosityisoperationalizedbyaadditiveindexoftwovariablescapturinghowoftenoneattendschurchandhowimportantreligionisinone'slife.ExactquestionswordingsappearintheAppendix.8AuthoritarianismisoperationalizedbyanadditiveindexoffourquestionsregularlyappearingontheANESspdesignedtocapturetheconstruct.Cronbach'salphais0.61.ExactquestionswordingsappearintheAppendix.86ofpartisanshipisconditionedonconspiratorialthinking,implyinganinteractivemodel,Iestimatetheadditivemodelstodemonstratejusthowsubstantialtheindependentofpartisanshipisoneachoftheseattitudesaboutgovernment9.Inordertofacilitatetheinterpretationofthemagnitudeofthestandardizedcots10resultingfromeachofthesemodelsispresentedinTable4.InterestedreaderswillatableofunstandardizedcotsintheAppendix.AcursoryinspectionofthestandardizedcotsinTable4revealsthatpartisan-shiphasthestrongestlinearrelationshipwithattitudesabouttheTeaParty,theroleofgovernment,andgovernmentspendingandservicesbyawidemargin.Ineachofthesecasesthestandardizedcotforpartisanshipisatleastdoublethatofthenextlargeststandardizedcot,whichwasideologicalself-idens(withracialresentmentinaclosethird).Onlywhenitcomestofeelingsaboutthefederalgovernmentistheofpartisanshiptrumpedbyanothervariable.Inthiscase,generalizedtrustingovernmentsharesastrongerlinearrelationshipwithfeelingstowardthefederalgovernmentthandoespartisanship,thoughthebetweenthesestandardizedcotsismuchsmallerthanisthedbetweenthestandardizedestimatesforpartisanshipandthenextmostsubstantiallytpredictor.Inadditiontopartisanshipbeingundeniablystronglyrelatedtoeachofthefourattitudesaboutgovernment,weseethatconspiratorialthinkingisnon-triviallyrelatedtothreeofthefourdependentvariables.Indeed,conspiratorialthinkingisstatisticallytlyrelatedtoattitudesabouttheTeaPartyandgovernmentspendingandservices,aswellasfeelingsaboutthefederalgovernment.Moresp,anincreaseinconspiratorialthinkingis9Icheckedforpotentialesintheofconspiratorialthinkingacrosssurveymodesbyinter-actingtheconspiratorialthinkingvariablewithadummyvariableindicatingwhethertherespondentwasintheinternetorface-to-facesample.Ididthesameinthemodelsbelow,specifyingathree-wayinteractionbetweenconspiratorialthinking,partisanship,andsurveymode.InnoinstancedidIndainthereportedrelationshipsbysurveymode.10Althoughpresentedinthetableabovesothatstatisticalcanbegleaned,werefrainfrominterpretingthestandardizedcotsassociatedwithdummyvariables(i.e.,thedichotomousopera-tionalizationsofgenderandrace).Suchcotsarenoteasilyinterpretablesinceastandarddeviationchangeonadichotomousvariablehasnosubstantivemeaning.87Table4.1:OLSregressionsdemonstratingtheindependentadditiveofpartisanshipandconspiratorialthinking(ANES).TeaPartyRoleofSpending&FederalGov.SupportGovernmentServicesThermometerConspiratorialThinking0.0360.0220.054-0.047Partisanship0.4700.382-0.344-0.263Ideology0.2010.168-0.187-0.031Trust0.026-0.1270.0760.381RacialResentment0.1510.136-0.142-0.053Sophistication0.0450.020-0.029-0.071Education-0.0240.063-0.070-0.071Age-0.045-0.0250.0120.010Income-0.0690.014-0.076-0.027Female0.014-0.0250.0520.027Black-0.005-0.1020.0460.138Latino0.042-0.0470.0630.122Religiosity0.007Authoritarianism0.019R20.5100.4710.4250.434n2032216520372199StandardizedOLSregressioncots.p<0:05withrespecttoatwo-tailedtest.88associatedwithanincreaseinsupportfortheTeaParty,adecreaseinsupportforgovernmentspendingaservices,andmorenegativefeelingstowardthefederalgovernment.Theserelationshipsmakesense.TheTeaPartyhastwocharacteristicsthatwouldbeattractivetotheconspiratorially-minded:1)ananti-establishmentfeelpromotedbytheinsurgentriseofthe\party,"and2)anexplicitideologicalgoaloflimitedgovernmentalinvolvementinprivatelife.Therelationshipbetweenconspiratorialthinkingandadesireforlessgovernmentalspendingandservicescanalsobeexplainedadesiretolimitthereachofgovernmentinthelivesofprivateindividuals.Finally,conspiracytheorists{giventheirpropensitytodistrustformal,powerfulinstitutionslikegovernments{shouldbelesslikelytoexpresspositiveemotionstowardtheU.S.federalgovernment.Havingestablishedthesubstantiveadditiveofconspiratorialthinkingandparti-sanship11,Iconsiderwhethertheofpartisanshiponattitudesandfeelingsaboutthegovernmentandanti-establishmentpartiesismoderatedbyconspiratorialthinking.Inordertotestthisproposition,Iestimatedaseriesofmodelsthateachincludealloftheindepen-dentvariablesincludedintheadditivemodels,inadditiontoamultiplicativetermdesignedtocapturetheinteractionbetweenpartisanshipandconspiratorialthinking.TheresultsoftheseanalysesarepresentedinTable4.Whenmultiplicativetermsandalloftheirconstitutivetermsareincludedinagivenregressionmodel,theassociatedcotsbecomerathertosubstantivelyinterpret.Iwilladdressthisissuebelow.First,however,Iconsiderissuesofstatisticalwithrespecttomykeyindependentvariablesaswellatthecontrols.Mostimportantly,thecotassociatedwiththeconspiratorialthinking{partisanshipinteractiontermisstatisticallytatconventionallevelsineachofthefourmodels.Furthermore,theconstitutivetermsarestatisticallytinallfourmodelsaswell,withoneexception.Unsurprisingly,ideologicalself-idenandtrustingovernmentarestatistically11Ielectedtosavespacebyignoringanextendeddiscussionofthecontrolvariablesintheadditivemodelssincethesemodelsaretheoretically,insomesense,misspsinceIbelievethatthereisaninteractionbetweenconspiratorialthinkingandpartisanshipinreality.Thus,thesevariablesarediscussedatgreaterlengthbelowinthecontextofthe\true"interactivemodels.89Table4.2:OLSregressionsdemonstratingthetsofpartisanshipconditionaloncon-spiratorialthinking(ANES).TeaPartyRoleofSpending&FederalGov.SupportGovernmentServicesThermometerConspiratorialThinking0.4540.221-0.056-0.160(0.058)(0.045)(0.050)(0.044)Partisanship1.0130.592-0.506-0.365(0.056)(0.043)(0.049)(0.043)ConspiratorialThinking-0.954-0.4740.3220.245Partisanship(0.121)(0.094)(0.106)(0.093)Trust0.009-0.0390.0240.110(0.006)(0.005)(0.006)(0.005)Ideology0.0420.027-0.032-0.004(0.005)(0.004)(0.004)(0.003)RacialResentment0.0110.008-0.009-0.003(0.002)(0.001)(0.001)(0.001)Sophistication0.0160.005-0.009-0.021(0.007)(0.006)(0.006)(0.006)Education-0.0070.014-0.016-0.015(0.005)(0.004)(0.005)(0.004)Age-0.000-0.0000.0000.000(0.000)(0.000)(0.000)(0.000)Income-0.0030.000-0.002-0.000(0.001)(0.001)(0.001)(0.001)Female0.008-0.0130.0280.013(0.012)(0.008)(0.009)(0.008)Black-0.022-0.0810.0410.100(0.018)(0.013)(0.015)(0.013)Latino0.031-0.0380.0510.086(0.016)(0.012)(0.014)(0.012)Religiosity0.001(0.002)Authoritarianism0.006(0.005)Intercept-0.1070.2090.7540.471(0.040)(0.031)(0.034)(0.030)R20.5250.4770.4270.436n2032216520372199OLSregressioncotswithstandarderrorsinparentheses.p<0:05withrespecttoatwo-tailedtest.90tpredictorsofattitudesaboutthegovernmentinmostmodels.ConservativestendtosupporttheTeaParty,favorareducedscopeofgovernment,andpreferlessspendingongovernmentalservicestoagreaterextentthandoliberals.Conversely,themoreindividualstrustthegovernment,themorepositivefeelingstheyhavetowardthenationalgovernment,andthemoretheypreferanexpandedscopeofgovernmentandspendingongovernmentalservices.Theestimatesassociatedwitheducation,gender,andracearestatisticallysignif-icantineachmodelwithoneexceptionforgender.Women,blacks,andhispanicsalltakemoreliberalpositionswhenitcomestotheroleofgovernment,preferencesaboutgovernmentspendingandservices,andTeaPartysupport(exceptwomen,inthisinstance),andtendtolikethefederalgovernmentmorethanmalesandwhites,respectively.Finally,racialre-sentmentremainsastatisticallytpredictorofTeaPartysupportintheinteractivemodels,wherebyanincreaseinracialresentmentisassociatedwithanincreaseinsupportfortheTeaParty,aspreviousworkwouldsuggest(Arceneaux&Nicholson2012,Barretoetal.2012).Inordertoaidintheinterpretationoftheinteractioncapturedbythecotsonthemultiplicativetermsincludedinthesemodels,IvisuallypresentthemarginalofpartisanshipacrossthevaluesofconspiratorialthinkingwithrespecttoeachofthefourdependentvariablesinFigure3.Oneimportantcharacteristicispresentineachofthefourpanels:asthelevelofconspiratorialthinkingincreases,themarginalofpartisanshipdecreases.Infact,ineverycasethemarginalofpartisanshipatthehighestlevelsofconspiratorialthinkingapproaches0.Thenatureofthisisparticularlyremarkablegiventhewidearrayofvariablesthatpartisanshipisknowntobestronglyrelatedto.EveninthemodelsIpresenthere,partisanshipisthestrongestpredictorofTeaPartysupport,attitudesabouttheroleofgovernment,andpreferencesaboutgovernmentspendinginser-vices,whileitissecondinpredictivestrengthtotrustwhenitcomestogeneralfeelingsaboutthefederalgovernment.Thattherobustofpartisanshiparereducedsodramaticallyasthelevelofconspir-91Figure4.1:Marginalofpartisanshipconditionalonconspiratorialthinking(ANES).92atorialthinkingincreasesisnotonlystatisticallyimpressive,butsubstantivelyimportantforanumberofreasons.First,itunderscoresthenatureofaconspiratorialmindsetasbeingonelargelycharacterizedbyageneralsuspiciontowardanddistrustofestablishmentures,institutions,andauthoritiessuchaspoliticalparties(andthosethatrepresentthem).Ratherthanbeingcorrelatedwith(Republican)partisanshipasananalysisemployingbe-liefsinspconspiracytheoriesaboutpopularpoliticallikeBarackObamamightsuggest,conspiratorialthinkingisreallyuncorrelatedwithpartisanshipingeneral.Indeed,tobeconspiratorialistrulytoeschewidentwithauthorities,and,bytoeschewaconnectionbetweentheseidenandotherpoliticalattitudes.Second,thesehighlightaneedtoconsiderthecorrelatesofwhatarewidelyas-sumedtobepartisanattitudeswhenpartisanshipisnotanimportantpredispositiontotheindividual.Recentworkontheriseoftheconspiracytheories(Brotherton2015,Oliver&Wood2014),increasinglevelsofdistrust(Hetherington&Rudolph2015),andthenatureofpoliticalindependence(Klar&Krupnikov2016),takentogether,shouldpromptare-considerationoftheimportanceofseeminglyapoliticalanddisengagedindividuals.Indeed,forsome,ariseingeneralpoliticaldivisivenessmaynotpromptpolarizationalongpartisanlines,butratherapsychologicaldistancingfromstatusquopoliticsaltogether.Althoughtherecouldbemanyreasonsforandoutcomesassociatedwithsuchaphenomenon,itseemstomethatconspiratorialthinkingiscertainlyaprimecandidateonbothfronts.4.2onApprovalofActorsandInstitutionsNextIconsidertherelationshipbetweenconspiratorialthinkingandapprovalofObama,Congress,andtheSupremeCourt.Sinceconspiracyideationiswidelycharacterizedbyanintensesuspiciontowardformalpoliticalinstitutions,wemightexpectthatthemorecon-spiratorialoneis,thelesstheywouldapproveofthe\job"beingdonebythoseinstitutions,ortheactorsthatrepresentthem.Thoughthiswouldmakesensetheoretically,onemaystillwonderwhylowapprovalreallymatters.Indeed,therelativelevelofapprovalofObama93isnearlycompletelydeterminedbypartisanandideologicalself-idenandmostAmericansseemtodisapproveofthejobCongresshasbeendoingregardlessofpartisanandideologicalpredispositions.Thatconspiratorialthinkingmightunilaterallydecreaseinstitutionalapprovalacrossbranchesis,however,quitepowerful.Ifconspiratorialthinkingweretosystematicallyrelatetoapprovalofeachoftheinstitutions,wewouldhavesomeroughindicationthatitisca-pableofcuttingacrosspartisanlines(inthecaseofObamaapproval),apopularsentimentregardingthecompetenceofCongress,andthesupposed\apolitical"natureoftheSupremeCourt.Inotherwords,itwouldbeatestamenttotheoverarchingstatisticalandsubstan-tivepowerofconspiratorialthinking{apowerthatisnotcurrentlybeingcapturedbyevenrelatedconstructssuchasinstitutional(dis)trust.Morethansomegeneraldistrust,individ-ualorientationstowardmajorgovernmentalinstitutionscanbectedbyafundamentallysuspiciousorientationtowardthedesignsoftheseinstitutionsandthemotivationsoftheactorsthatoperatewithinthem.Totestthissetofhypotheses,IshiftfromtheANESdatatotheCCESdatawhichincludesthebettermeasureofconspiratorialthinkingoutlinedinthepreviouschapter12.Ichosetoemployquestionsaboutthejobapprovaloftheinstitutionsthatrepresentthethreebranchesofgovernment:theexecutive(Obama),legislative(Congress),andjudicial(SupremeCourt).Thequestionssimplyaskedrespondentstorate,usinga4-pointLikert-typesetofresponsecategories,towhatdegreetheyapprovedordisapprovedofthejobbeingdonebyeachoftheseinstitutions.Inadditiontothesuspicionscale,Iincludeahostofcon-trolvariablesincludinginstitutionaltrust,partisanship,ideologicalself-idenknowl-edge,educationalattainment,religiosity,income,age,gender,andrace.Allofthesevariablesaremeasuredasoutlinedinthepreviouschapter.Sincethedependent\approval"questionshavefourdiscreteresponsecategories,eitherOLSorMLE(i.e.,ordinallogit/probit)could12AlthoughIampresentingtheresultsofanalysesusingthesuspicionquestionsonthe2014CCES,thesameresultsareproducedforapprovalofObamaandCongressusingtheconspiratorialthinkingfactorfromChapter2andthe2012ANESdata(SCOTUSapprovalwasnotonthesurvey).94beusedestimatethemodels.Sincetherewasnosubstantiveintheconclusionsgarneredfromthetwoprocedures,IpresenttheOLSestimatesbelow13.Table3includesboththe\raw"estimates(incolumns1,3,and5),aswellasstandardizedestimates(incolumns2,4,and6)14.Conspiratorialthinkingisastatisticallytpredictorofeachofthejobapprovalresponses.Moresp,asIhypothesized,anincreaseinthelevelofconspiratorialthinkingisassociatewithadecreaseinjobapprovalassociatedwitheachoftheinstitutions.IntheObamajobapprovalmodel,conspiratorialthinkingexertsastatisticallytinthefaceofsubstantivelylargeofpartisanandideologicalpredispositions.IntheCongressandSupremeCourtapprovalmodels,conspiratorialthinkingisstatisticallytdespitethelowvarianceoftheapprovalofCongressandtheof(explicitly)institutionaltrust.Thoughthetrustmeasureexerts,accordingtothestandardizedesti-mates,thelargestonCongressionalandSCOTUSapprovalattitudes,weshouldex-pectasmuchfromamulti-itemmeasurecomprisedofattitudesaboutindividualinstitutionsincludingthemilitary,bigbusiness,andtheSupremeCourt,itself.Assuch,thestatisticallytrelationshipbetweenconspiratorialthinkingandthesetwoinstitutions(certainlytheSupremeCourt)servesasaveryconservativetestofboththeindependentofconspiratorialthinking,andthebetweenconspiratorialthinkingand(dis)trust.4.3onPoliticalEngagementFinally,Iturntowardaconsiderationofthepoliticalengagement-relatedconsequencesofconspiratorialthinking.Accordingtobasicnormativetheoryregardingthenecessarycriteriathatmustbemetbyaliberaldemocracy,citizenengagementinpoliticsisimperative.Indeed,13Thetheoreticalmodelcouldalsobeoperationalizedviaaseeminglyunrelatedregression(SUR)sincetheequationerrorsarelikelycorrelated.AlthoughtheSURmodeldoesandallpairsofequationerrorsaretlycorrelated,neitherthestatisticalnorsubstantiveoftheindependentvariableschangeswiththisestimation.14NotethatIexcludestandardizedcotsfordichotomousvariablessincesuchquantitiesarenoteasilyinterpretable.Sincedichotomousvariablesonlytakeatwovalues(inallofthecasespresentedhere,a0ora1),astandarddeviationchangeismeaningless.95Table4.3:Independentofsuspicion(CCES)onapprovalofgovernmentalactorsandinstitutions.ApprovalofApprovalofApprovalofObamaCongressSCOTUSSuspicion-0.039-0.073-0.032-0.095-0.036-0.090(0.015)(0.014)(0.017)Partisanship-0.240-0.4520.0220.065-0.026-0.065(0.021)(0.019)(0.023)Ideology-0.212-0.3370.0290.073-0.260-0.055(0.026)(0.023)(0.028)Trust0.0040.0140.0640.2920.1180.444(0.010)(0.009)(0.011)Knowledge0.0400.029-0.178-0.202-0.102-0.094(0.042)(0.038)(0.046)Education0.0260.0320.0360.0700.0030.004(0.025)(0.022)(0.027)Religiosity0.0100.022-0.002-0.006-0.036-0.107(0.013)(0.012)(0.014)Income-0.015-0.040-0.010-0.044-0.001-0.003(0.011)(0.010)(0.012)Age0.0000.005-0.008-0.167-0.008-0.144(0.002)(0.002)(0.002)Female-0.136-0.054-0.025(0.067)(0.060)(0.072)Black0.7500.2650.248(0.117)(0.104)(0.125)Latino0.0010.160-0.056(0.139)(0.124)(0.147)Intercept1.3751.0301.170(0.229)(0.205)(0.246)R20.6150.2290.264N539533526OLScotsw/standarderrorsinparenthesesincolumn.Standardizedcotsinsecondcolumn.p<0:05withrespecttoatwo-tailedtest.96engagementonthepartofthegovernedhelps\check"thegovernors{itfostersdirectcitizeninvolvementinthegoverningprocess,andisawayforcitizenstoregistertheiropinions(DelliCarpini&Keeter1996).Aconspiratorialworldviewmay,however,discouragethesekindsofbehaviorsandoutcomes.Themostconspiratorialindividualsagreestronglywithpropositionssuchasgovernmentalinstitutionsbeingcontrolledbyoutsideinterestsandsecretpatternsanddesignsbeingallaroundus.Asaninherentdistrustofestablishedinstitutionsandthosewhogovernthemaccompaniesaconspiratorialworldview,Imightexpectthatmoreconspiratorialindividualsarelesslikelytoengageinestablishmentpoliticsintraditionalways,andperhapsmorelikelytoengageinanti-establishmentcauses.Iconsiderthreeformsof(dis)engagementinpolitics.Participationincampaignactivi-tiesisaparticularlyimportantformofengagementsinceitisoneofthemajor(non-voting)behavioralactsinwhichcitizensdirectlydealwithquestionsaboutwhatisbestforthecountry,whattheirvaluesarewhenitcomestopublicpolicy,and,ultimately,whoshouldgovern.Furthermore,previousexperimentalresearchinvestigatingbeliefsinspconspir-acytheoriesdemonstratesthattheremaybearelationshipbetweensuspicionandintentiontoparticipateinpoliticsinthefuturevia,forexample,votingordonatingtocampaigns(Jolley&Douglas2014b).Thus,Iinvestigatewhetherconspiratorialthinkingisrelatedtoparticipationincampaignactivities.Inadditiontoengagementinthephysicalformofparticipation,Iconsiderpsychologicalengagementwithmainstreampolitics.Inparticular,Iinvestigatewetherthestrengthofindividualattachmentstooneofthetwomajorpartiesorsupportfortheanti-establishmentTeaPartyarerelatedtoasuspiciousmindset.Sinceconspiratorialthinkingcanbegenerallycharacterizedasapredispositiontoviewestablishedinstitutionsandtheactorsthatcomprisethemwithsuspicion,Iexpectthathighlysuspi-ciousindividualsarelesslikelytoidentifywithestablishmentpartiesverystrongly,andmorelikelytosupportanti-establishmentpartiessuchastheTeaParty.Beginningwiththerelationshipbetweencampaignactivitiesandconspiratorialthinking,thecampaignactivityvariableIemployinthisanalysisisanadditiveindexofresponsesto97questionsaskingrespondentswhetherornotthey1)donatedmoneytoapoliticalcandidate,campaign,ororganization,2)displayedapoliticalsignofsomesort,3)attendedlocalpoliticalmeetings,and/or4)workedforacandidateorcampaign.Strengthofpartisanidenisoperationalizedviathefoldedseven-pointpartyidenscale.Assuch,thefourpointstrengthvariabletakesonavalueof1iftherespondentidenasan\independent,"avalueof2iftheyidentifyasapartisanleaner,a3iftheyidentifyasaweakpartisan,andavalueof4iftheyidentifyasastrongpartisan.Lastly,TeaPartysupportismeasuredviaae-pointLikert-typeitemwithresponsecategoriesrangingfromverynegativetoverypositive.Aswiththepreviousmultivariateanalyses,Iincludetrust,education,religiosity,knowl-edge,interest,anddemographiccontrolsinthesemodels.Iadditionallyincludestrengthofpartisanandideologicalpredispositionsinthecampaignactivitymodel(e.g.,Verba,Schlozman&Brady1995),campaignactivityinthepartisanstrengthmodel(tocontrolforthepotentialofendogeneityinthatrelationship),andideologicalandpartisanself-idenintheTeaPartysupportmodel(e.g.,Arceneaux&Nicholson2012,Barretoetal.2012,Cho,Gimpel&Shaw2012,Crawford&Xhambazi2015,Parker&Barreto2013).Bothunstandardizedandstandardized15modelestimates{allofwhichwereproducedusingOLS{arepresentedinTable3.Beginningwiththecampaignactivitymodel,Ithatconspiratorialthinkingissig-tlynegativelyrelatedtocampaignactivity,holdingothervariablesconstant.Thatis,asone'slevelofconspiratorialthinkingincreases,herpropensitytoparticipateintradi-tionalcampaignactivities,onaverage,decreases.Moresplly,increasingthelevelofconspiratorialthinkingfromthelowestscalescoretothehighestscalescoreproduces,onaverage,adecreaseof0.552activitiesonthecampaignactivityindex.Whilethismayseemsmall,mostpeopleparticipateinzerocampaignactivitiesandapproximately18.63%participateinjustone.If,then,highlevelsofconspiratorialthinkingectindividualseven15Onceagain,Iexcludestandardizedcoientsfordichotomousvariables.98Table4.4:Independentofsuspicion(CCES)onpoliticalengagement.CampaignPartisanTeaPartyActivityStrengthSupportSuspicion-0.046-0.100-0.065-0.1290.0480.068(0.020)(0.022)(0.023)Knowledge0.1180.1000.0960.075-0.019-0.010(0.054)(0.062)(0.064)Interest0.3020.2360.0690.0500.0340.016(0.060)(0.068)(0.072)Education0.0560.0830.0100.013-0.024-0.023(0.029)(0.033)(0.033)Trust-0.006-0.0200.0040.0110.0510.111(0013)(0.014)(0.015)Religiosity0.0410.1130.0470.1180.01730.031(0.016)(0.018)(0.019)Age0.0040.0570.0040.062-0.005-0.053(0.003)(0.003)(0.003)Female0.0360.347-0.036(0.085)(0.095)(0.095)Black-0.2230.340-0.531(0.163)(0.184)(0.196)Latino0.1730.2780.206(0.182)(0.206)(0.212)PartisanStrength-0.000-0.000(0.038)CampaignActivity-0.000-0.000(0.049)Ideology0.3160.392(0.038)Partisanship0.2400.345(0.031)Intercept-0.3121.5231.004(0.297)(0.330)(0.371)R20.1380.0830.588N549549484OLScotsw/standarderrorsinparenthesesincolumn.Standardizedcotsinsecondcolumn.p<0:05withrespecttoatwo-tailedtest.99slightly,thenitcouldmakeasubstantialenceindrivingindividualspredisposedtoengageinjustoneformofcampaignactivitytojointhemanyoftheirpeersandabandontheenterprisealtogether.Whatismore,conspiratorialthinkingisexactlytiedwithknowl-edgeforthesecondstrongestoncampaignactivityaccordingtothemagnitudeofthestandardizedcots,behindonlystatedinterestinpolitics.Theseeminglysmallis,then,nothingatwhichto{infact,itisquitesubstantialrelativetomostknownpredictorsofcampaignactivityandwhenconsideredinconjunctionwithinformationaboutabsolutelevelsofcampaignparticipation.WhenitcomestostrengthofpartisanidenIalsoobserveastatisticallycantofconspiratorialthinking.Ashypothesized,conspiratorialthinkingisnegativelyrelatedtoone'sstrengthofidenwithmainstreampoliticalparties.Moreover,thisisthestrongestrelationshippresented,slightlyedgingoutreligiosity,andevenknownpredic-torsofpartisanstrengthsuchasinterestinpoliticsandparticipationincampaignactivity.Varyingtheconspiratorialthinkingscalefromitslowesttoitshighestvalueproduces,onaverage,a0.78unitdecreasealongthefour-pointpartisanstrengthmeasure.Again,thisisquitearemarkablegiventhe\short"scaleofthepartisanstrengthvariable,andrelativetootherpredictorsinthemodel.Finally,IobservethatconspiratorialthinkingistlyandpositivelyrelatedtoTeaPartysupport.Ashypothesized,asconspiratorialthinkingincreases,sotoodoesthelevelofsupportonehasfortheTeaParty,onaverage.AsImightexpect,bothliberal-conservativeideologicalandpartisanself-idenarestronglyrelatedtoTeaPartysupportsuchthatconservativesandRepublicansaremorelikelytosupporttheTeaParty.Furthermore,trustisstronglyrelatedtoTeaPartysupport,suchthatastrustininstitutionsincreases,supportfortheTeaPartyincreases.Atstglance,thismayseemsurprising.Intuitively,lowtrustindividualsshouldbemorelikelytosupporttheTeaPartythanhightrustindividuals.However,thetrustvariableiscomprisedofresponsestoquestionsabouttrustintheCIA,military,bigbusiness,andSupremeCourt.SinceTeaPartymembers100areunabashedsupportersofthefreemarket(bigbusiness)andfavorfederalspendingondefense(CIA,military)morethananysortofdomesticspending,thisresultmakessense.Regardlessoftheoftrust,conspiratorialthinkingisalsosubstantivelyrelatedtoTeaPartysupport;indeed,itexertsanslightlysmallerthanthatoftrust16.4.4ConclusionInthischapterIfurtheredthesparseliteratureonthepotentialtsofconspiratorialthinkingonanumberofnormativelyimportantattitudesandbehaviors.InthesectionofthechapterIshedsomelightonthenatureoftheinteractionbetweenconspiratorialthink-ingandperhapsthemostimportantvariableinpoliticalbehaviorresearch{partisanship{whenitcomestofundamentalattitudesabouttheroleofgovernmentinsociety,generalfeel-ingstowardthefederalgovernment,andsupportfortheTeaParty.Althoughwetypicallyunderstandthesekindsofattitudestolargelybetheproductof(or,atleast,tightlyassoci-atedwith)partisanshipandideology,weknowfairlylittleabouttheseattitudesoutsideofthecontextofpartisanshipandliberal-conservativeideology.Thattheectofpartisanshipontheseattitudesissogreatlydiminishedamongthemostconspiratorially-mindedindi-vidualsshouldpromptresearcherstoreconsiderthepsychologicalroleofconspiracismforpartisansandnon-partisansalike,aswellascontemplatetheconditioningrolethatconspir-atorialthinkingmayplayinmanyothermoretraditionalrelationshipsbetweenbehavioralvariablesofinterest.Next,Idemonstratedtherelationshipbetweenapprovalofgovernmentalinstitutionsandconspiratorialthinking.Astheorywouldsuggest,butscholarshaveyettodemonstrate,thereexistsatightrelationshipbetweentheconspiratorialmindsetandafundamentaldis-approvalofgovernmentalinstitutions.Indeed,tobehighlysuspiciousofthegovernmentis16OnlytheresultsoftheTeaPartysupportmodelweresupportedbythe2012ANESdataand(admit-tedlylessthanideal)operationalizationofconspiratorialthinking.ThepartisanstrengthandTeaPartysupportmodelsaresupportedbythealternatesuspicionmeasurementstrategyemployedintheMTurkdata(questionsaboutcampaignactivitieswherenotinthesurvey).Takentogether,therethepresentedinthischapterarequiterobusttovaryingmeasurementstrategiesandsamples.101toinherentlydistrustanddisapproveofanysortofgovernmentalestablishmentorFinally,Ibothandextendedpreviousworkthathasrevealedanegativerela-tionshipbetweenconspiratorialthinkingandpoliticalparticipation.Morethanasigntdecreaseinparticipationincampaignactivities,moreconspiratorialindividualsarelesslikelytoidentifywithamajorpartyandmorelikelytoidentifywithinsurgentpartiesliketheTeaParty.Thus,Ithatconspiratorialthinkersarelesslikelytoengageinestablishmentpoliticsbothbehaviorallyandwithrespecttofundamentalgroupidentities.Inapoliticalenvironmentwheresocio-politicalidentitieshavebeenfoundtoeverythingfromattitudesaboutracialgroups(Kam&Kinder2012)toindividualtivepolarization(e.g.,Iyengar,Sood&Lelkes2012,Mason2015),thisofconspiratorialthinkingcanhavesubstantialconsequencesforwhichandhowindividualsengageinpolitics.Beyondthoseconsideredhere,Iurgeotherstoinvestigatetherelationshipsbetweenothervariablessuchasideology,individual-levelpolarization,andvalueorientationsandconspir-atorialthinking.Ofcourse,wemustalsounderstandtheofconspiratorialthinkingonimportantpoliticalbehaviorssuchascampaignparticipationandvoting,andgeneralorientationstowardpoliticssuchas,anomie,andalienation.Withoutamorerobustunderstandingofthenatureofconspiratorialthinking,wewillnotbeabletounderstandhowmanyconspiracytheoristsreallyexist,tiatebetween\true"conspiracytheoristsandpartisanmotivatereasoners,orexplainthemoderndayriseinpopularconspiracybeliefs,especiallyamongmassivelypopularpresidentialcandidatesandtheirfollowers.102APPENDIX103Variablecoding/wordingfornewANESvariablesWherewouldyouplaceYOURSELFonthisscale,orhaven'tyouthoughtmuchaboutthis?(spsrvpr_ssself)7-pointscale;0{Governmentshouldprovidemanyfewerservices,reducespending,1{Governmentshouldprovidemanyservices,increasespendingalot.Whichofthetwostatementscomesclosertoyourview?1.Themainreasongov-ernmenthasbecomebiggerovertheyearsisbecauseithasgotteninvolvedinthingsthatpeopleshoulddoforthemselves.2.Governmenthasbecomebiggerbecausetheproblemswefacehavebecomebigger.(govrole_big)0.Governmentbiggerbecauseproblemsbigger1.Governmentbiggerbecauseit'sinvolvedinthingspeopleshouldhandlethemselvesWhichofthetwostatementscomesclosertoyourview?1.Weneedastronggovern-menttohandletoday'scomplexeconomicproblems.2.Thefreemarketcanhandletheseproblemswithoutgovernmentbeinginvolved.(govrole_market)0.Needastronggovernmenttohandlecomplexeconomicproblems1.FreemarketcanhandlewithoutgovernmentinvolvementWhichofthetwostatementscomesclosertoyourview?1.Thelessgovernment,thebetter.2.Therearemorethingsthatgovernmentshouldbedoing.(govrole_lessmore)0.Morethingsgovernmentshouldbedoing1.ThelessgovernmentthebetterHowmuchgovernmentregulationofbusinessisgoodforsociety?(govrole_regbus)0.Noneatall1041.Alittle2.Amoderateamount3.Alot4.AgreatdealNationalgovernmentfeelingthermometer(ftgr_fedgov)0-100,where0correspondsto\cold"feelings,100correspondsto\hot"(veryfavorable)TeaPartyfeelingthermometer(tea_supp_x)0-100Irish,Italians,Jewishandmanyotherminoritiesovercameprejudiceandworkedtheirwayup.Blacksshoulddothesamewithoutanyspecialfavors.0.Disagreestrongly1.Disagreesomewhat2.Neitheragreenordisagree3.Agreesomewhat4.AgreeStronglyGenerationsofslaveryanddiscriminationhavecreatedconditionsthatmakeitforblackstoworktheirwayoutofthelowerclass.0.Agreestrongly1.Agreesomewhat2.Neitheragreenordisagree3.Disagreesomewhat4.DisagreeStrongly105Overthepastfewyears,blackshavegottenlessthantheydeserve.0.Agreestrongly1.Agreesomewhat2.Neitheragreenordisagree3.Disagreesomewhat4.DisagreeStronglyIt'sreallyamatterofsomepeoplenottryinghardenough;ifblackswouldonlytryhardertheycouldbejustaswellaswhites.0.Disagreestrongly1.Disagreesomewhat2.Neitheragreenordisagree3.Agreesomewhat4.AgreeStronglyBlack(dem_raceeth_x)0.Non-black/AfricanAmerican1.Black/AfricanAmericanLatino(dem_raceeth_x)0.Non-Latino1.LatinoFemale(gender_respondent_x)0.Male1061.FemaleIncome(inc_incgroup_pre)0-27,0=Under$5,000;27=$250,000ormore107Variablecoding/wordingfornewCCESvariablesDoyouapproveofthewayeachisdoingtheirjob...[PresidentObama,TheU.S.Congress,TheU.S.SupremeCourt]{Stronglydisapprove{Somewhatdisapprove{Somewhatapprove{Stronglyapprove108AdditiveANESRegressionModelsTable4.5:OLSregressionsdemonstratingtheindependentadditiveofpartisanshipandconspiratorialthinking.TeaPartyRoleofSpending&FederalGov.SupportGovernmentServicesThermometerConspiratorialThinking0.0630.0300.076-0.061(0.030)(0.023)(0.026)(0.023)Partisanship0.6530.412-0.384-0.272(0.032)(0.025)(0.028)(0.024)Trust0.010-0.0380.0240.109(0.007)(0.005)(0.006)(0.005)Ideology0.0470.029-0.034-0.005(0.005)(0.003)(0.004)(0.003)RacialResentment0.0120.009-0.009-0.003(0.002)(0.001)(0.001)(0.001)Sophistication0.0190.007-0.010-0.022(0.007)(0.006)(0.006)(0.006)Education-0.0070.014-0.016-0.015(0.005)(0.004)(0.005)(0.004)Age-0.001-0.0000.0000.000(0.000)(0.000)(0.000)(0.000)Income-0.0030.000-0.003-0.001(0.001)(0.001)(0.001)(0.001)Female0.010-0.0130.0280.013(0.011)(0.008)(0.009)(0.008)Black0.005-0.0740.0360.096(0.018)(0.013)(0.015)(0.013)Latino0.041-0.0340.0480.083(0.016)(0.012)(0.014)(0.012)Religiosity0.001(0.002)Authoritarianism0.005(0.005)Intercept0.0370.2790.7070.435(0.036)(0.027)(0.031)(0.027)R20.5100.4710.4250.434n2032216520372199OLSregressioncoetswithstandarderrorsinparentheses.p<0:05withrespecttoatwo-tailedtest.109Chapter5:ConclusionInthepreviouspagesofthismanuscriptIhopetohaveelucidatedanumberofimportantcharacteristicsofconspiracytheoriesandthosepredisposedtobelievethem.Perhapsoneofthemostimportantconclusionstobemadeisneithernovelnorsptothistopic:forbetterorworse,measurementcolorsourviewofthetopicswestudy(Jacoby1999).Inthiscase,wesawthatquestionsaboutbeliefsincertaintypesofconspiracytheories{particularlyonesinvolvinghighlyvisible,partisanobjects{caneasilyleadinterestedresearchersintoerroneouslybelievingnotonlythatconspiracytheoristsarepartisansinsheep'sclothing,butthatsptypesofpartisans{Republicans{aremorepronetoconspiratorialthought.Thoughitislikelytruethatpartisansuseconspiracytheories,andeventhe\conspiracytheorist"label,todiscreditandregisterdisapprovaloftheirpoliticalopponents,tohaveatrulyconspiratorialmindsetissomethingtaltogether.Onlybyusingtheorytoproperlyoperationalizeconspiratorialthinkingwereweabletomoreaccuratelydecipherthequalitiesandcharacteristicsoftheindividualswhoviewthepoliticalworldinthisway.Indeed,Ifoundthat{contrarytoagreatdealofpreviouswork{conspiratorialthinkingisapredispositionthatisnegativelyrelatedtopartisanship.Individualswhonotonlyfailtoplacetrustinpeopleofsocio-politicalandeconomicpowerbutareconvincedthatsuchpeopleareconstantlyworkingtoshlyusurppowerfromandmislead"thepeople"activelyrejectestablishmentpoliticsandthepartiesthatgovernit.Theseindividualseschewpartisantiesandpossessattitudesaboutandevaluationsofthegovernmentthatareunconnectedtoleft/rightpartisanship.Furthermore,farfrombeingadiscreteconceptwherepeopleeitherareorarenotcon-110spiracytheorists,Iarguethatthegeneralconspiratorial{or,suspicious{mindsetisbetterconceivedofasacontinuumwhereindividualscanpossessmoreorlessofthepredisposition.AlthoughthemeasureofconspiratorialthinkingproposedinChapter3isadmittedlyasteptowardthegoalitisdesignedtoachieve,andthedistributionsofsuspiciousthoughtshouldbetakenastentativetosomedegree,thetheoryandvalidityassessmentIpresentalsosuggeststhatpeoplearemoreconspiratorialthanconventionalwisdomposits.Socialdesir-abilityconcernssurroundingthe\conspiracytheorist"labelcoupledwiththeotherproblemsassociatedwiththestrategyofmeasuringconspiratorialthinkingviaquestionsaboutbeliefsinspconspiracytheorieswouldhaveleadmetowhollytand,asarguedhere,misguidedconclusions.Properlyconceivedofasacontinuouslatentvariable,wealsoobservedthatincreasesintheindividuallevelofconspiratorialthinkingcoincidewithbeliefsinavarietyofspe-conspiracytheoriesanddecreasesinengagementwithandparticipationinestablish-ment,partisanpolitics.Moresp,Idemonstratedthatconspiratorialthinkingwaspositivelyrelatedto\birther"and\truther"beliefs,beliefsaboutLeeHarveyOswald'sco-conspirators,thelinkbetweenchildhoodvaccinesandautism,thecorporation-induceddangerofgenetically-mofoods,andtheroleofpoliticalmotivesinpresidentialallo-cationsofdisasterrelieffunds.Thatthegeneralmeasureofconspiratorialthinking{evenoperationalizedslightlytly,andmeasuredacrosssamples{issostronglyre-latedtospconspiratorialbeliefsstronglysupportsthevalidityoftheoperationalizationsIemploy.Perhapsofmorenormativeimportarethepotentialtsofconspiratorialthinkingonengagementwiththepoliticalworld.Iobservedthatasthelevelofconspiratorialthinkingincreasedindividualswere,onaverage,lesslikelytoidentifywithoneofthemajorpoliticalpartiesorparticipateintraditionalcampaignactivities,andmorelikelytosupporttheTeaParty{afactionoftheRepublicanPartypropelledbyadissatisfactionwithestablishmentpolicies.Higherlevelsofconspiratorialthinkingalsocoincidewithgreatlyattenuatedrela-111tionshipsbetweenpartisanpredispositionsandattitudesabouttheroleofthegovernmentinsociety,theproperbalancebetweengovernmentprovidedservicesandgovernmentspending,generalfeelingstowardthegovernment,andsupportfortheTeaParty.Extrapolatingfur-ther,itappearsthatconspiratorialthinkingunderminestheconceptofaleft/rightpoliticalorientationorganizing,indeedconstraining,attitudesandfeelingsaboutcommonpoliticalobjectsandphenomena.5.1NormativeImplicationsConsideringthenormativeimplicationsofthesetheremaybesomereasonforcon-cern.Atfacevalue,thefactthattheconspiratorially-mindedarelesswillingtoengageinpoliticsposesaproblemfordemocratictheory,traditionallyunderstood.Indeed,oneofpo-liticalscience'sfoundationalgoals(certainlysincethe\behavioralrevolution")istoconsiderboththeoreticallyandempiricallytheroleofrepresentationingovernmentalsystems.Aspoliticalscientists,weareallinterestedinsomeshapeorforminunderstandinghowandwhetherindividualsarebeingrepresentedbythesocio-politicalinstitutionstheyhavecon-structedaroundthem.Thatanon-trivial{and,perhaps,quitelarge{sectofindividualsissohighlysuspiciousoftheirelectedrepresentativesthattheyalienatethemselvesfromthepoliticalsystemis,then,disconcerting.Fromthepointofviewoftheseindividuals,theyareneitherbeingproperlyrepresentedingovernmentalinstitutions,noraretheyareabletosuccessfullyexpresstheirpoliticalperspectivesthroughnon-partisan(left/right,thatis)means.Whethertheirperceptionis\factually"accurateornot,theseindividualsarealienatedfromthepoliticalsystembecauseofsuchperceptions1.Moretangiblepracticalconsequencesofconspiratorialthoughtalsoabound.Asnotedabove,believingthatchildhoodvaccinationsarelinkedtoautismcansuppressindividualwillingnesstovaccinateone'schildren(Jolley&Douglas2014a,Nyhanetal.2014).Conspir-1Thisalienationisnotunlikealienationthatpoliticalindependentsexpresswhentheychoosetoeschewpartisanlabels.RecentworkbyKlarandKrupnikov(Klar&Krupnikov2016)idenpoliticaldivisiveness(suchasthatbroadlyassociatedwithpolarization)asamechanismbywhichpeopleindividualswhoeschewpartisanlabelsaresubsequentlycastoutoftheAmericantwo-partypoliticalsystem.112acytheoriesaboutbirthcontrol(Bird&Bogart2003),HIV/AIDS(Bogart&Bird2006,Bird&Bogart2005),andhealthcarereform(Nyhan,&Ubel2013)canresultinsimilar(in)actionsbyindividuals,allofwhichcanveryplausiblyleadtoarapidresurgenceofheretofore\extinct"andwhollypreventablediseasesandgeneralhealthconditions.Beliefinseeminglyinnocuousconspiracytheoriessuchas\birtherism"or\trutherism"canalsohavepracticalconsequences.SuspiciontowardObamacould,forinstance,enhancepartisanpolarizationandallthenegativeconsequencesassociatedwithit,orevenfeedintofeelingsofracialresentmentandethnocentrismthatwidelycharacterizemanyofObama'sdetractors.Moregenerally,theprivate,free-willedsortofthinkingassociatedwiththeconspiratorialmindsetcanprecludeanysortofmeaningfulsocialinteractionorcooperation.Theseseeminglynegativeconsequencesofconspiratorialthoughtshouldnot,inmyopin-ion,overshadowthemanypositive{albeitlessobvious{ones.Oneofsuchpositiveconse-quencesistheheightenedlevelofscrutinythatconspiratorialindividualsapplytoinformation.Aprimaryywithpreciselyconspiracytheoriesisthatwecanrarelybecertainthattheyarenottrue.Conspiracytheories,therefore,becomelabeledassuchduetheplausibilityoftheircentralclaims,amongothercharacteristics.Asscientists,however,weareparticularlykeennottooutrightrejectimplausibleclaims,astheworldisaninherentlyuncertainplace.Insomeways,conspiracytheoristsplaytheroleofdevil'sadvo-cate,and,indeed,theyevendososuccessfullyattimes.Although,forinstance,Watergateisnowwidelyregardedasaproven,importanthistoricaleventof20thcenturyAmericanpolitics,atonepointitwasnomorethananimplausibleaccusationbeingpromotedbypoliticaladversaries.Similarsituationssurroundone-timeconspiraciesaboutCIAmindcontrolprograms(ProjectMKUltra)andclandestinetomanipulatethemassmediaanddisseminatepropagandaofvarioussorts(OperationMockingbird).Inotherwords,weshouldbecarefulinlabelingconspiracytheoriesandthosewhobelievethem\crazy"untilprovenfalsebeyondareasonabledoubt.Moregenerally,theheightenedscrutinyofinformationassociatedwithasuspicious113mindsetisapositiveandnecessarycomponentofafunctioningrepresentativedemocracy.Inaworldwithnowatchdogsorinstitutionalchecks,asweknow,humannatureoverrid-inglypromotesself-interest.\Thepeople,"theprinciples,mustholdtheirrepresentatives,theagents,accountable.Thoughthelevelofscrutinyappliedbythemostconspiratorialindividualsmaybehigherthanthatwhichisrequiredtosatisfythisbasicneed,thegeneralprincipleholds.Simplyput,conspiratorialthinkingoverlapswithsomeofthepsychologicalqualitiesthatwe{asindividualswhotheorizeaboutandempiricallyinvestigatethequalitiesofdemocracy{wishtheaverageindividualexhibited.Whiletinfoilhatsareprobablyworthyofsomemockery(ifnotsimplyfromthepointofviewofhighfashion),weshouldbecau-tiousinrejectingoutrightthemindsetthroughwhichconspiracistsinterpretandintegrateinformationabouttheworldaroundthem2.5.2DirectionsforFutureResearchGiventherelativeyouthofthe(empirical)conspiracytheoryliterature,directionsforfutureworkareplentiful.Mypersonalinterestsandperspectiveonthistopicleadmetoconsiderthreepossibleavenuesinparticular.Relatedtothepotentialpositiveconsequencesofcon-spiratorialthought,itismyopinionthatfutureworkshouldmoreactivelyconsiderwhatcomponentsoftheconspiratorialmindsetarenormativelygood.Experimentalandinter-activesurveydesignscould,forinstance,beleveragedtobetterassesshowconspiratorialindividualsseekout,process,andemploytheinformationtheyareconfrontedwith.Eachoftheseinformation-relatedstageshaspotentialforfurtherrevealingandscienknowledgeaboutthenatureoftheconspiratorialmindset.Weknow,forinstance,essen-tiallynothingaboutwhereconspiracytheoristsgettheirinformation,orwhichnewssourcestheytrust.And,although,wehavetheorizedabouthowwebsofconspiratorialbeliefsareconstructed,noneofsuchtheorieshavebeenivelytested.2Iencourageinterestedpartiestoseektheworkonmycolleague,StevenSmallpage,whois{asanormativepoliticaltheorist{completingadissertationonthefundamentalroleofpoliticalsuspicioninLocke'spoliticalthought.Steve'sworkwillsurelymoreclearlyandelegantlyelucidateallofthenormativecomponentsofsuspicionandtheconspiratorialmindset.114Relatedtoquestionsaboutinformationsources,weknowfairlylittleaboutconspiracytheoristsasagroup.Dohigh-levelconspiraciststhinkofthemselvesasmembersofacoherentgroup?Evenatthelevelofpartisanpolitics,doesstatedassociationwith\birther"beliefssignaltoco-partisans,forinstance,someuniquelevelofdisdainforBarackObama?And,asforthosepartisanswhoconvenientlychoosetoendorsespconspiracytheories,cantheuseofthe\conspiracytheorist"labelbemanipulated,muchlikecorrectanswerstoknowledgequestionscanbe?Attheheartofeachofthesequestionsliesaneedtobetterunderstandthesocialandpoliticalcomponentsofbothconspiratorialthinkingasageneraltraitandspconspiratorialbeliefs.And,ofcourse,itispreciselythesocio-politicalcomponentsofconspiracismthataremostpressingfromtheperspectiveofsocialscience.Finally,andmostsimply,futureworkshouldfocusonandreplicatingtheresultshere,particularlythoseexplicitlyrelatedtothemeasurementofconspiratorialthinking.Notimeseries,letalongpanel,dataexistonconspiratorialthinking.Assuch,wehavenosoundguessastowhetherornotconspiratorialthinking{asageneraltrait{hasreallyincreasedovertime,aspopularsentimentwouldsuggest.Thoughothershavedocumentedariseinthelevelofattentionthatconspiracytheorieshavegarneredinpopularmediaoverthepastseveraldecades(Uscinski&Parent2014),sucharisecouldbeduemoretothepartisanelementofspconspiracytheoriesorchangesinthequalityofnewmedia,ratherthananincreaseinconspiratorialthinking.Similarly,weshouldconsidertheabilityoftheproposedgeneralmeasureofconspiratorialthoughttotranslateacrosssocio-politicalenvironmentsandcultures,andthecomparativelevelofsuspicionintheU.S.versusothercountries.Iexpect,inyearstocome,toprovideanswerssomeofthesequestionsinmyownwork.115BIBLIOGRAPHY116BIBLIOGRAPHYAbalakina-Paap,Marina,WalterG.Stephan,TraciCraig&W.LarryGregory.1999.\BeliefinConspiracies."PoliticalPsychology20(3):637{647.Ansolabehere,Stephen,JonathanRodden&Jr.Snyder,JamesM.2008.\TheStrengthofIssues:UsingMultipleMeasurestoGaugePreferenceStability,IdeologicalConstraint,andIssueVoting."TheAmericanPoliticalScienceReview102(2):215{232.Arceneaux,Kevin&StephenP.Nicholson.2012.\WhoWantstoHaveaTeaParty?TheWho,What,andWhyoftheTeaPartyMovement."P.S.PoliticalScienceandPolitics45(4):700{710.Bailyn,Bernard.1992.TheIdeologicalOriginsoftheAmericanRevolution.?:BelknapPress.Barkun,Mark.2003.CultureofConspiracy:ApocalypticVisionsinContemporaryAmerica.Berkeley,CA:UniversityofCaliforniaPress.Barreto,MattA.,BetsyCooper,BenGonzalez,ChrisTowler&ChristopherParker.2012.\TheTeaPartyintheAgeofObama:MainstreamConservatismorOut-GroupAnxi-ety?"PoliticalPowerandSocialTheory22(1):105{137.Bartels,Larry.2002.\BeyondtheRunningTally:PartisanBiasinPoliticalPerceptions."PoliticalBehavior24:117{150.Berinsky,AdamJ.1999.\TheTwoFacesofPublicOpinion."AmericanJournalofPoliticalScience43(4):1209{1230.Bird,SherylThorburn&LauraM.Bogart.2003.\BirthControlConspiracyBeliefs,Per-ceivedDiscrimination,andContraceptionamongAfricanAmericans."JournalofHealthPsychology8:263{276.Bird,SherylThorburn&LauraM.Bogart.2005.\ConspiracyBeliefsAboutHIV/AIDSandBirthControlAmongAfricanAmericans:ImplicationsforthePreventionofHIV,OtherSTIs,andUnintendedPregnancy."JournalofSocialIssues61(1):109{126.Blank,JoshuaM.&DaronShaw.2015.\DoesPartisanshipShapeAttitudestowardScienceandPublicPolicy?TheCaseforIdeologyandReligion."TheANNALSoftheAmericanAcademyofPoliticalandSocialScience658:18{35.Bogart,LauraM.&SherylThorburnBird.2006.\RelationshipofAfricanAmericans'SociodemographicCharacteristicstoBeliefinConspiraciesaboutHIV/AIDSandBirthControl."JournaloftheNationalMedicalAssociation98:1144{1150.Bolsen,Toby,JamesN.Druckman&FayLomaxCook.2015.\Citizens',Scientists',andPol-icyAdvisors'BeliefsaboutGlobalWarming."TheANNALSoftheAmericanAcademyofPoliticalandSocialScience658:271{295.117Brotherton,Rob.2015.SuspiciousMinds:WhyWeBelieveConspiracyTheories.NewYork,NY:Bloomsbury.Brotherton,Rob&ChristopherC.French.2014.\BeliefinConspiracyTheoriesandSus-ceptibilitytotheConjunctionFallacy."AppliedCognitivePsychology28:238{248.Brotherton,Rob&ChristopherC.French.2015.\IntentionSeekers:ConspiracistIdeationandBiasedAttributionsofIntentionality."PloSONE10(5):e0124125.Brotherton,Rob,ChristopherC.French&AlanD.Pickering.2013.\MeasuringBeliefinConspiracyTheories:TheGenericConspiracistBeliefsScale."FrontiersinPsychology4:279.Browne,MichaelW.&RobertCudeck.1992.\AlternativeWaysofAssessingModelFit."SociologicalMethodsandResearch21(2):230{258.Bruder,Martin,Petere,NickNeave,NinaNouripanah&Roland2013.\Mea-suringIndividualinGenericBeliefsinConspiracyTheoriesAcrossCultures:ConspiracyMentalityQuestionnaire."FrontiersinPsychology4:225.Bullock,JohnG.,AlanS.Gerber,SethJ.Hill&GregoryA.Huber.forthcoming.\PartisanBiasinFactualBeliefsaboutPolitics."QuarterlyJournalofPoliticalScience.Campbell,Angus,PhilipE.Converse,WarrenE.Miller&DonaldE.Stokes.1960.TheAmericanVoter.NewYork:Wiley.Center,PewResearch.2015a.\Americans,PoliticsandScienceIssues.".RetrievedOctober9,2015.URL:http://www.pewinternet.orscience-and-politicsFINAL.pdfCenter,SouthernPovertyLaw.2015b.\AntigovernmentMovement.".RetrievedOctober9,2015.URL:https://www.splcenter.orology/antigovernmentCho,WendyK.Tam,JamesG.Gimpel&DaronW.Shaw.2012.\TheTeaPartyMovementandtheGeographyofCollectiveAction."QuarterlyJournalofPoliticalScience7:105{133.Crawford,JarretT.&EnedaXhambazi.2015.\PredictingPoliticalBiasesAgainsttheOccupyWallStreetandTeaPartyMovements."PoliticalPsychology36(1):111{121.Darwin,KarenM.,NickNeave&JoniHomes.2011.\BeliefinConspiracyTheories:TheRoleofParanormalBelief,ParanoidIdeation,andSchiztotypy."PersonalityandIndi-vidualences50(8):1289{1293.DelliCarpini,MichaelX.&ScottKeeter.1996.WhatAmericansKnowAboutPoliticsandWhyitMatters.NewHaven:YaleUniversityPress.118Douglas,KarenM.&RobbieSutton.2010.\TheHiddenImpactofConspiracyTheories:PerceivedandActualofTheoriesSurroundingtheDeathofPrincessDiana."JournalofSocialPsychology148(2):210{222.Edwards,AllenL.1957.TheSocialDesirabilityVariableinPersonalityAssessmentResearch.OakBrook,IL:DrydenPress.Einstein,KatherineLevine&DavidM.Glick.2015.\DoIThinkBLSDataareBS?TheConsequencesofConspiracyTheories."PoliticalBehavior37(3):679{701.Fenster,Mark.2008.ConspiracyTheories:SecrecyandPowerinAmericanCulture.Min-neapolis,MN:UniversityofMinnesotaPress.Goertzel,Ted.1994.\BeliefinConspiracyTheories."PoliticalPsychology15:731{742.Goertzel,Ted.2010.\ConspiracyTheoriesinScience."EMBOReports11(7):493{499.Hargrove,Thomas.2006.\ThirdofAmericansSuspect9-11GovernmentConspiracy.".RetrievedJune11,2015.URL:http://www.scrippsnews.com/911pollHartman,ToddK.&AdamJ.Newmark.2012.\MotivatedReasoning,PoliticalSophistica-tion,andAssociationsbetweenPresidentObamaandIslam."PS:PoliticalScience&Politics45(3):449{455.Hetherington,MarcJ.&ThomasJ.Rudolph.2015.WhyWashingtonWon'tWork:Polar-ization,PoliticalTrust,andtheGoverningCrisis.Chicago,IL:UniversityofChicagoPress.Hofstadter,Richard.2008.TheParanoidStyleinAmericanPolitics.VintageBoos.Hu,Li-tze&PeterM.Bentler.1999.CriteriaforFitIndexesinCovarianceStruc-tureAnalysis:ConventionalCriteriaVersusNewAlternatives."StructuralEquationModeling6(1):1{55.Roland&MartinBruder.2014.\Speaking(Un-)TruthtoPower:ConspiracyMen-talityasaGeneralisedPoliticalAttitude."EuropeanJournalofPersonality28(1):25{43.Iyengar,Shanto,GauravSood&YphtachLelkes.2012.NotIdeology:ASocialIdentityPerspectiveonPolarization."PublicOpinionQuarterly76(3):405{431.Jacoby,WilliamG.1999.\LevelsofMeasurementandPoliticalResearch:AnOptimisticView."AmericanJournalofPoliticalScience43:271{301.Jacoby,WilliamG.2000.\IssueFramingandPublicOpiniononGovernmentSpending."AmericanJournalofPoliticalScience44:750{767.Jolley,Daniel&KarenM.Douglas.2014a.\TheofAnti-VaccineConspiracyTheoriesonVaccinationIntentions."PLoSONE9(2):e89177.119Jolley,Daniel&KarenM.Douglas.2014b.\TheSocialConsequencesofConspiracism:ExposuretoConspiracyTheoriesDecreasesIntentionstoEngageinPoliticsandtoReduceOne'sCarbonFootprint."BritishJournalofPsychology105:35{56.Kam,CindyD.&DonaldR.Kinder.2012.\EthnocentrismasaShort-TermForceinthe2008AmericanPresidentialElection."AmericanJournalofPoliticalScience56(2):326{340.Keeley,BrianL.1999.\OfConspiracyTheories."JournalofPhilospohy96(3):109{126.Kimbrell,Andrew.2015.\GMOScienceDeniers:MonsantoandtheUSDA.".RetrievedOctober9,2015.URL:ost.com/andrew-kimbrell/gmo-science-deniers-monsanto-and-the-usdab6904606.htmlKlar,Samara&YannaKrupnikov.2016.IndependentPolitics:HowAmericanDisdainforPartiesLeadstoPoliticalInaction.Cambridge,MA:CambridgeUniversityPress.Kline,RexB.2011.PrinciplesandPracticeofStructuralEquationModeling,ThirdEdition.NewYork:TheGuilfordPress.Kraft,PatrickW.,MiltonLodge&CharlesS.Taber.2015.\WhyPeople`DontTrusttheEvidence':MotivatedReasoningandScienBeliefs."TheANNALSoftheAmericanAcademyofPoliticalandSocialScience658:121{133.Kunda,Ziva.1990.\TheCaseforMotivatedReasoning."PsychologicalBulletin108:480{498.League,NationalConsumers.2014.\Survey:OneThirdofAmericanParentsMistakenlyLinkVaccinestoAutism.".RetrievedJune10,2015.URL:www.nclnet.orgsurveyonethirdofamericanparentsmistakenlylinkvaccinestoautismLewandowsky,Stephan,GillesE.Gignac&KlausOberauer.2013.\TheRoleofConspiracistIdeationandWorldviewsinPredictingRejectionofScience."PLoSONE8(10):e75637.Lodge,Milton&CharlesS.Taber.2013.TheRationalizingVoter.NewYork,NY:CambridgeUniversityPress.Mason,Lilliana.2015.\`IDisrespectfullyAgree':ThetialofPartisanSortingonSocialandIssuePolarization."AmericanJournalofPoliticalScience59:128{145.McCright,AaronM.&RileyE.Dunlap.2011.\ThePoliticizationofGlobalClimateChangeandPolarizationintheAmericanPublicsViewsofGlobalWarming,2001-2010."Soci-ologicalQuarterly52(2):155{194.McHoskey,JohnW.1995.\CaseClosed?OntheJohnF.KennedyAssassination:BiasedAs-similationofEvidenceandAttitudePolarization."BasicandAppliedSocialPsychology17(3):395{409.Miller,JoanneM.,KyleL.Saunders&ChristinaE.Farhart.forthcoming.\ConspiracyEndorsementasMotivatedReasoning:TheModeratingRolesofPoliticalKnowledgeandTrust."AmericanJournalofPoliticalScience.120Mokken,R.J.1971.ATheoryandProcedureofScaleAnalysis.TheHague:Mouton.Moscovici,Serge.1987.TheConspiracyMentality.InChangingConceptionsofConspiracy,ed.F.Graumann&S.Moscovici.NewYork:SpringerPress.Mullouy,D.J.2003.FederalEmergencyManagementAgency.InConspiracyTheoriesinAmericanHistory:AnEncyclopediaVolumeI,ed.PeterKnight.SantaBarbara,CA:ABC-CLIOPress.Nestle,Marion.2015.\It'sTimetoRequireLabelsforGMOs.".RetrievedOctober9,2015.URL:http://time.com/3853160/require-labels-for-gmos/Newheiser,Anna-Kaisa,MiguelFarias&NicoleTausch.2011.\TheFunctionalNatureofConspiracyBeliefs:ExaminingtheUnderpinningsofBeliefintheDaVinciCodeConspiracy."PersonalityandIndividualences51:1007{1011.Nyhan,Brendan&Jason2010.\WhenCorrectionsFail:ThePersistenceofPoliticalMisperceptions."PoliticalBehavior32:303{330.Nyhan,Brendan,Jason&PeterA.Ubel.2013.\TheHazardsofCorrectingMythsAboutHealthCareReform."MedicalCare51:127{132.Nyhan,Brendan,JasonSeanRichey&GaryL.Freed.2014.eMessagesinVaccinePromotion:ARandomizedTrial."Pediatrics133:835{843.Oliver,J.Eric&ThomasJ.Wood.2014.\ConspiracyTheoriesandtheParanoidStyles(s)ofMassOpinion."AmericanJournalofPoliticalScience58(4):952{966.Parker,Christopher&MattA.Barreto.2013.ChangeTheyCan'tBelieveIn:TheTeaPartyandReactionaryPoliticsinAmerica.PrincetonUniversityPress.Pasek,Josh,TobiasH.Stark,JonA.Krosnick&TrevorTompson.2015.\WhatMotivatesaConspiracyTheory?BirtherBeliefs,Partisanship,Liberal-ConservativeIdeology,andAnti-BlackAttitudes."ElectoralStudies40:482{489.Pigden,Charles.2007.\ConspiracyTheoriesandtheConventionalWisdom."Episteme:AJournalofSocialEpistemology4(2):219{232.Robins,RobertS.&JeroldPost.1997.PoliticalParanoia:ThePsychopoliticsofHatred.NewHaven,CT:YaleUniversityPost.Swami,Viren,JakobPietschnig,UlrichS.Tran,IngoW.Nader,StefanStieger&MartinVoracek.2013.\LunarLies:TheImpactofInformationalFramingandIndividualDif-ferencesinShapingConspiracistBeliefsAbouttheMoonLandings."AppliedCognitivePsychology27:71{80.Swami,Viren,RebeccaColes,StefanStieger,JakobPietschnig,AdrianFurnham,SherryRehim&MartinVoracek.2011.\ConspiracistIdeationinBritainandAustria:Evi-denceofaMonologicalBeliefSystemandAssociationsBetweenIndividualPsychologi-calandReal-WorldandFictitiousConspiracyTheories."BritishJournalofPsychology102:443{463.121Swami,Viren,TomasChamorro-Premuzic&AdrianFurnham.2010.\UnansweredQues-tions:APreliminaryInvestigationofPersonalityandIndividualPredictorsof9/11ConspiracistBeliefs."AppliedCognitivePsychology24:749{761.Taber,CharlesS.&MiltonLodge.2006.\MotivatedSkepticismintheEvaluationofPoliticalBeliefs."AmericanJournalofPoliticalScience50(3):755{769.Uscinski,JosephE.,CaseyKlofstad&MatthewD.Atkinson.forthcoming.\WhatDrivesConspiratorialBeliefs?TheRoleofInformationalCuesandPredispositions."PoliticalResearchQuarterly.Uscinski,JosephE.&JosephM.Parent.2014.AmericanConspiracyTheories.NewYork,NY:OxfordUniversityPress.Verba,Sidney,KayLehmanSchlozman&HenryE.Brady.1995.VoiceandEquality:CivicVoluntarisminAmericanPolitics.Cambridge:HarvardUniversityPress.Weaver,VeslaM.&AmyE.Lerman.2010.\PoliticalConsequencesoftheCarceralState."AmericanPoliticalScienceReview104(4):817{833.Wood,Gordon.1993.TheRadicalismoftheAmericanRevolution.NewYork,NY:VintagePress.Wood,MichaelJ.,KarenM.Douglas&RobbieM.Sutton.2012.\DeadandAlive:BeliefsinContradictoryConspiracyTheories."SocialPsychologicalandPersonalityScience3(6):767{773.122