THEEMERGENCEOFLONEWOLFTERRORISM:PATTERNSOFBEHAVIORANDIMPLICATIONSFORINTERVENTION
BrentSmith,JeffGruenewald,PaxtonRoberts,andKellyDamphousse
ABSTRACT
Purpose–Inthischapter,weexamineseveralattributesoflonewolfterroristsandhowtheiractivitiesaretemporallyandgeospatiallypatterned.Inparticular,wedemonstratehowprecursorbehaviorsandattackcharacteristicsoflonewolvesaresimilaranddifferentcomparedtothoseofgroup-basedterrorists.
Design–BasedondatadrawnfromtheAmericanTerrorismStudy(ATS),weexamine476federalterrorism“indictees”linkedto264incidents.Threetypesoflonersareidentifiedbasedongroupaffiliationsandlevelsofassistanceinpreparingforandexecutingterroristattacks.Aseriesofanalysescomparativelyexaminelonerswhohadnoassistanceandthoseactorsthatdid.
Originality/Value–Studiesonlonewolfterrorismremainfewandmanyareplaguedbymethodologicalconceptuallimitations.ThecurrentstudyaddstothisgrowingliteraturebyrelyingonlonewolfterrorismdatarecentlymadeavailablebytheAmericanTerrorismStudy(ATS).OurfindingsarevaluableformembersofthelawenforcementandintelligencecommunitiesresponsiblefortheearlydetectionandpreventionoflonewolfterrorismintheUnitedStates.
Findings–Theresultsofthisstudysuggestthatlonewolfterroristsaremoreeducatedandsociallyisolatedthangroup-basedactors.Lonewolvesalsoengageinlessprecursoractivitiesthangroupactors,butarewillingtotravelgreaterdistancestoprepareforandexecuteattacks.Explanationsforwhylonewolvesareableto“survive”longerthanterroristgroupsbyavoidingarrestmayinpartstemfromtheirabilitytotemporallyandgeospatiallypositiontheirplanningandpreparatoryactivities.
Keywords:Americanterrorism;lonewolf;loneactor;precursorconduct;pre-incidentindicators;leaderlessresistance
INTRODUCTION
TheemergenceoflonewolfterrorismisoneofthemostfrequentlymentionedissuesamongthosetaskedwithcounteringthethreatofterrorisminAmerica.Despitethis,itisalsooneofthemostfrequentlymisunderstoodconceptsinthecriminologicalliterature,difficultto
defineconceptually,andevenmoreproblematictomeasureempirically.Whencompoundedwiththenotionthatlonewolvesareespeciallydifficulttodetect,morelikelytoengagein“lowexpenseandhighsuccess”targeting(Zierhoffer,2014),andlesslikelytobecaughtusingtraditionalcounterintelligencemethodsthanothertypesofterrorists(BakkerdeGraaf,2011;Hewitt,2014),theproblemof“lonewolf”terrorismbecomesevenmoreexasperating.
Conceptually,“lonewolf”terrorismhasgonebyanumberofdifferentnames,suchas“loneavengers”(Stern,2003)or“free-lanceterrorists”(Hewitt,2003). Liketheterm“terrorism,”boththosewhoareresponsibleforrespondingtoitandthosewhoacademicallystudyit,havehadadifficulttimearrivingataconsensusregardingwhatconstitutesa“lonewolf.”Consequently,effortstooperationalizetheconceptof “lonewolf”haveresultedinawidevarietyofnamesorvariablesandmultiplewaysofmeasuringthem.Notsurprisingly,the
resultingliteraturereflectsanimpressivearrayofconflictingpatternsanddemographictraitsthatsupposedlyrevealthecharacteristicsof“lonewolves.”Unfortunately,thisgrowingbodyofliteratureissoabsentofaconsensusondefiningthetermthatwehavefewmethodologicallyrigorouspiecesfromwhichtoevenidentifythenatureandextentoftheproblem.
Inthesectionsthatfollow,wewillexaminewhythe“lonewolf”strategyemerged,theoriginsoftheconceptofthe“lonewolf,”andeffortstofurtherrefinethedefinition.Wewillalsoreviewrecentempiricaleffortstostudythephenomenonandhowthesevaryingdefinitionsof“lonewolf”haveresultedinsomewhatcontradictoryfindingsandpatternsofbehavior.Wewillalsoenterintothisfraybyanalyzingdatafromadifferentsource,theAmericanTerrorismStudy(ATS)toassessthenatureandextentof“loneactor”terrorisminAmericaasmeasuredthroughtheuseofFBI“officiallydesignated”actsofterrorismoverthepastthirtyyearsorso. Finally,
wehopethatourfindingswillprovidesomeclaritytotheoften-confusingbodyofliteratureon
“lonewolf”terrorism.
BACKGROUNDONTHEORIGINSOFLONEACTORTERRORISM
Although“lonewolf”orloneactorterrorismbyAlQaedaadherentsisthoughttobeoneofthegreatestthreatstoAmericansecurity,andsomemaybelievethatthisisauniquelyAlQaedaphenomenon,theevolutionof“lonewolf”orloneactorterrorismhasanextensivehistoryintheUnitedStates.Infact,itisdifficulttofullyappreciatetheuseofthistacticwithoutunderstandingitsevolutioninAmericanterrorism.Thestorybeginsseveraldecadesbeforethe
9/11attacks.
Inthelate1950s,FidelCastroinstigatedrevoltinCubabyemployingaruralrevolutionarymodel.InspiredbytheArabrevoltagainstFrenchruleinAlgeriain1954,CastromodeledhisstrategyfromtheruralmovementemployedinitiallybytheNationalLiberationFrontinAlgeria.Castro’sruralrevolutionarymodelhadfourmajorcharacteristics.First,itutilizedatraditionalmilitaryhierarchicalstructurecomposedofatraditionalcommandandcontrolstructure.Second,thisconventionalguerillaarmyfocusedonthetraditionalmilitaryeffortstocaptureandholdterrain.InCastro’scase,thisinvolvedextortionandterrorisminadditiontotraditionalmilitaryoperationsashisarmyobtainedcontrolofruralCuba,whileisolatingtheurbanareassurroundingHavana.Third,aspartoftheefforttocaptureandholdterrain,Castrocreated“fixedcompounds”suchasmilitaryhospitalsandtrainingcampsasheadvancedonHavana.Finally,hetookadvantageofasystemofnationalandinternationalnetworkingtofinance,arm,andfeedhisgrowingmilitarycomponent.
WiththesuccessfuloverthrowoftheBatistaregimein1959,Castrosoughttoimportthemodeltoothercountries,primarilythroughwhatbecameknownastheTri-ContinentalConferences(Sterling,1981). However,whenCheGuevaratriedtoimplementtheruralrevolutionarymodelinBoliviain1967,hewasquicklycapturedandhisarmydisintegrated.Theintelligencegatheringcapabilitiesoftargetedgovernmentshadimproveddramaticallysincethelate1950sandtheuseoffixedcompoundsandahierarchicalmilitarystructureresultedinabysmalfailure.Theextremeleftadaptedalmostovernightwiththeintroductionofanurbancellularmodel.IntheWesternHemisphere,thepublicationofCarlosMarighella’sMini-ManualoftheUrbanGuerillain1969ledtothewholesaleadoptionofthecellularapproachbyawiderangeofAmericanleftistextremists.Throughoutthe1970sand1980s,thestrategywas
employedbytheUnitedFreedomFront,theWeatherUnderground,theSLA,theMay19th
CommunistOrganization,andscoresofotherviolentleftistgroups.
ExtremerightwinggroupsinAmerica,however,failedtolearnthelessonsofCheGuevara’sdefeatinBolivia.Whentheyturnedviolentintheearly1980s,theyimmediatelyadoptedastrategysimilartoCastro’soldruralrevolutionarymodel.AlmostalloftherightwingterroristgroupsindictedundertheFBI’sCounterterrorismPrograminthe1980semployedfixedcompounds,nationalnetworking,andahierarchicalstructuretosomeextent(Smith,1994).
Fromtheanti-taxgroupsliketheArizonaPatriotsandtheSheriff’sPosseCommitatustoChristianIdentitygroupsliketheAryanNationsandCovenant,SwordandArmoftheLord,allsoughtto“captureandholdterrain”byinitiallycreatinglocalcompounds,stakingouttownships,ormappingtheboundariesoftheirproposednew“sovereignstates.”
SimilartoGuevara’sdemise,withinthreeyears,theFBIhadidentifiedthelocationsofviolentrightwingcompoundsintheUnitedStates,indicted,andcapturedorkilledtheleading
membersofoverhalfadozengroups.Theso-called“warin‘84”merelyledto“arrestsin‘85”
astheFBIhadoneofitsmostproductivecounterterrorismeffortsofalltime.Allbutthreeoftheleadingfiguresoftheextremerighthadbeenconvictedonvariousfederalchargesby1987. Tocompletethecampaign,U.S.AttorneysindictedLouisBeam,RichardButler,andeightothersassociatedwiththeOrder,AryanNations,andtheCovenant,Sword,andArmoftheLordonseditiousconspiracychargesinfederaldistrictcourtinFortSmith,Arkansas.Alleleven
indicteeswereacquittedofallchargesin1988. Whilethecaseitselfisworthyofadditionalcomment,forthepurposesofthisarticle,thelessonslearned,particularlybyBeam,areofparticularinterest.
Overthenextfouryears,Beamconcentratedondevelopingastrategytominimizethecivilandcriminalliabilityofgroupleaders(DamphousseSmith,2004). ThesiegeatRubyRidge,Idahoin1992providedthecatalystforBeamtoadvancehisstrategy(Kaplan,1997). AtahastilycalledmeetingofextremerightgroupleadersinEstesPark,Coloradothatsummer,he
publiclycalledfortheimplementationof“leaderlessresistance”(Beam,1992),anuncoordinatedviolencemodelthatessentiallybypassedthecellularapproachadoptedbyleftists.Althoughsomescholarshavesuggestedthatthe“lonewolf”conceptbeganinthepost-9/11eraandislargelyabyproductoftheemergenceofAlQaeda(e.g.,Barnes,2013),theterm“lonewolf”emergedinthemid-1990sandwasusedtodescribethethreatofrightwingterrorismintheUnitedStates.ItbecameparticularlyprominentamongcounterterrorismofficialssoonaftertheOklahomaCitybombingin1995,asmanycametobelievethatMcVeigh’sactionsmayhave
representedanearlyexampleofleaderlessresistanceandlonewolfbehavior.1 Although
leaderlessresistanceisthemostfamousoftheseuncoordinatedviolencemodels,itisnottheonlyone.
Themid-1990sprovedtobeaturningpointinthestrategicplanningofterroristgroupsworldwide.Inadditiontotheadvocacyofleaderlessresistancebytheextremeright,environmentalgroupsbeganusingthenewlycreatedWorldWideWebasbotharecruitingtoolandasawaytodisseminateinformationaboutstrategictargets(Joosse,2007). BoththeEarthLiberationFrontandtheAnimalLiberationFrontproducedsomeofthemostadvancedwebsitesontheInternetinthemid-1990s.ThestructureofthemovementandtheiruseoftheWeballowedleadingmovementfigurestoindirectlyidentifypotentialtargetstoothermembersthrougharticlesandposts,whilerecordingthesuccessful“directactions”madebymembersforsupposedlyinformationalpurposesonly.
Similarly,Islamicextremists,particularlythoseassociatedwithBinLaden’sAlQaedamovement,used“fatwas”asacalltoactionthatwouldmakeestablishingcriminalliabilityextremelydifficultsolelyonthebasisoftheissuedproclamation.BinLaden’sfamous1998fatwato“killtheAmericans”istheclassicexample:
Onthatbasis,andincompliancewithGod'sorder,weissuethefollowingfatwatoallMuslims:TherulingtokilltheAmericansandtheirallies--civiliansandmilitary--isanindividualdutyforeveryMuslimwhocandoitinanycountryinwhichitispossibletodoit….[E]veryMuslimwhobelievesinGodandwishestoberewardedtocomplywithGod'sordertokilltheAmericansandplundertheirmoneywhereverandwhenevertheyfindit(ascitedinRanstorp,1998,p.329).
AlthoughAlQaedaandsubsequentSunniextremistgroupslikeISILrepresentahybridmodelinwhichcomponentsoftheruralrevolutionarymodel(creatinganIslamicStatethroughthe
holdingofterrain),thecellularmodel(e.g.,the9/11attackers),andanuncoordinatedviolencemodelfeaturinglonewolves(e.g.,the2009NidalHasanattackatFortHood,TX),thegreatestconcernamonghomelandsecurityexpertshereistheUnitedStatesinthelastofthesestrategic
andtacticalapproaches(Borum,Fein,Vossefuil,2012;Gruenewald,Chermak,Freilich,
2013b;Zierhoffer,2014).
Whileourconsiderationoftheoriginsofthe“lonewolf”conceptmaybebroaderthanotherrelevantstudies,webelieveditnecessarytoplacetheuseoftheterminthehistoricalcontextofAmericanterrorism.Doingsoseemsespeciallyimportant,asasignificantbodyofliteratureonloneactorterrorismhasbeguntoemergeoverthepastdecade.Althoughallofthisliteraturehassurelycontributedtoourunderstandingoftheconcept,thetermneverthelessremainsconceptuallyconfusingand,consequently,difficulttomeasurewithanydegreeofaccuracy(SpaaijHamm,2015). Inthefollowingsection,wereviewhighlightsofsomeofthisburgeoningliterature.
REVIEWOFTHELITERATURE
Inarecentreviewofresearchonlonewolfterrorism,SpaaijHamm(2015)recountedsixteendifferentmonikersusedfortheterm“lonewolf.”Somescholarshaveusedthesetermsinterchangeablythoughtheseseeminglysynonymoustermsmayactuallyrefertoconceptuallyuniqueformsofterroristoffending.Aconsensusdefinitionoflonewolfterrorismcontinuestoeluderesearchers;however,severalconceptualdimensionsofthisphenomenonhavebeguntoemerge.Thefirst,andarguablyleastproblematic,dimensionoflonewolfterrorismismotivation.WhileSimon(2013)suggestedthatconventional(ornon-ideological)criminalsbeincludedasonetypeoflonewolf(Simon,2013),mostotherlonewolfdefinitionsnecessitateoffenderstohavepolitical,social,orotherideologicalobjectives.Asecondfuzzierconceptualdimensioninvolvesterrorists’affiliationswithgroupsthatareorganizedbyhierarchicalcommandandcontrolorganizationalstructures.Paststudieshavevariedinsubtlebutimportant
waysinregardstohowgroupaffiliationismeasured.Forexample,Gruenewaldetal.,(2013a)andPantucci(2011)conceptualizegroupaffiliationasdirect(orfirst-hand)interactionwithotherlike-mindedextremists,whileothersconsidergroupinfluenceintermsofautonomousdecision-making,ortheextanttowhichterroristsreceivedirectionfromgroupleaders(Borumetal.,
2012). AccordingtoHewitt(2003)andotherterrorismexperts,lonewolvescanbemembersofterroristgroupsaslongastheyarenotactingundertheordersofterroristleaders.Thethirdkeydefinitionaldimensionoflonewolfterrorismreferstotheextentthatterroristsoperatealone.Whilemeasuringthe“aloneness”ofterroristsmayseemstraightforward,empiricalconsiderationsofoffendingarrangementscanquicklybecomemuddled.Forinstance,somedefinitionalschemaallowformultiple“loneactors”tobecategorizedaslonewolf“packs”or“isolateddyads”(Gilletal.,2014;Gruenewaldetal.,2013a;Pantucci,2011),whileotherresearchersdrawaharderlineonthenecessityofaloneness(SpaaijHamm,2015). Relatedly,thereremainsconceptualambiguityinregardstothespecificactivitiesterroristsmustengageinalonetobeconsideredalonewolf.Borumetal.(2012)refertooffenderswhoinitiate,plan,preparefor,andexecuteanattackwithoutdirectassistancefromanyotherpersonas“solo”offenders,whilethosewhoreceivedirectassistancefromoneortwoothersinaccomplishingtheseactivitiesareconsidered“lone”offenders.Itispossible,however,thatterroristsmayreceivedirectassistanceinsomestagesoftheterroristcycle,whilereceivingnoassistanceinotherstages.Itremainsunclearastowhetherornotaterroristwhoexecutesanattackby
himselfcanbeconsideredalonewolfifhehadassistanceinbuildingthebomb.Addressingthisquestionisanimportantnextstepinsortingthroughtheconceptualambiguitiesoflonewolfdefinitions.
Thoughthenumberofempiricalstudiesoflonewolfterrorismremainrelativelyfew,conceptualandmethodologicaldilemmashavenotkeptallresearchersfromstudyingthisimportanttopic.Justasthenumberofacademicstudiesofterrorismhasgrownexponentially,thenumberoflonewolfstudieshasalsoincreasedinthepastfewyears.Itisfromthissmallbutgrowingliteraturethatoffenderandincidentpatternsoflonewolfterrorismbegintoemerge.Below,wereviewsomeofthesepatterns,andthensuggesthowthecurrentstudyfillssomeremaininggapsintheextantliterature.
Severalstudieshaveconcludedthatlonewolfoffendingisontherise.AstudybySpaaij(2010),forinstance,concludedthatwhilelonewolfoffendinghasremainedstableinEurope,thistypeofterrorismhadincreaseddramaticallyintheUnitedStates.AnotherearlierstudyfoundthatunaffiliatedindividualskilledoneinsixterrorismvictimsintheUnitedStates,andwhenincludingtheOklahomaCitybombing,themajorityofterrorism-relateddeathsbetween
1978and2001wereattackscommittedby“unaffiliated”individuals(Hewitt,2003). Stillotherstudieshaveshownthattrendsinlonewolfoffendingmaybeterrorismmovementspecific.ThoughlonewolfterrorismbyIslamicextremistsmaybeincreasingafter9/11,far-rightlonewolfterrorismmayhavepeakedduringthe1990s(Gruenewaldetal.,2013a;seealsoHewitt,
2003;Spaaij,2010).
Asocialanddemographic“profile”oflonewolveshasalsobeguntoemergefromrecentstudies.Inregardstoage,studieshavefoundthatlonewolvesaretypicallyintheirmidtolate
30s(Gilletal.,2014). Agesofoffendersmayvary,however,bythe“type”oflonewolf.Forexample,far-rightlonewolveswhooperateinsmallcells,or“wolfpacks,”havebeenshowntobe,onaverage,muchyounger(intheirearly20s)(Gruenewaldetal.,2013a).Thedisproportionatenumberofmaleoffendersisgenerallyoneofthemostconsistentfindings
acrossterrorismstudiesandextendstolonewolvesaswell.Infact,Gilletal.(2014)foundthatnearly97%ofthe119lonewolvestheyexaminedweremale.Thoughfewstudieshave
measuredterrorists’educationalattainmentstatuses,recentevidenceindicatesthatapproximately
75percentoflonewolveshaveatleastsomecollegeexperience.Fourteenpercentoflonewolvesactuallyearnedgraduatedegrees.Despitetheirrelativelyprivilegedsocialpositions,lonewolveshaveoftenbeendescribedassocially“isolated,”“withdrawn,”“awkward,”and“inept”(MoskalenkoMcCauley,2011;Nijboer,2012;Spaaij,2010). Asevidence,Gruenewaldetal.,(2013b)foundthatfar-right“loners”weremorelikelytobedivorced,separated,orwidowedthanothertypesoffar-rightterrorists.Inaddition,onestudyfoundthatlessthan25percentoflonewolveshadbeenmarried,relativelylesscomparedtoothertypesofterroristsoperatingaroundtheworld(Gilletal.,2014).
Inadditiontooffenders’backgroundcharacteristics,wesuggestthatobservinghowterrorismevent“cycles”aregeospatiallyandtemporallypatternedisanimportantnextstepinresearchonlonewolfterrorism.Thisinvolvesnotonlyexaminingcompletedterroristacts,butalsothepre-incidentactivitiesoflonewolvesandthoseplannedbutfailedorfoiledplotsthatnevercometofruition.Gilletal.(2014)havealreadybeguntoexamineseveralofthesequestions,whichhaveimportantimplicationsforthelawenforcementcommunitiesresponsibleforpreventingandinvestigatingterrorism.Basedonopen-sourcedatathattheycollected,Gilletal.(2014)observedseveraltypesofpre-incidentactivitiesthatoccurredwithintheplanningandpreparatorystagesofterroristcycles,suchasrecruitmentandtheassemblyofexplosivedevices.Theyfoundthatoverhalfofthesepre-incidentactivitiesoccurredwithinoneyearoftheattacks.Thoughtheysuggestedthatlonewolveswerenotnecessarilyimpulsive,therewasoftenvery
littletimebetweenthelonewolves’choicestouseviolenceandthecommissionofattacks(Gilletal.,2014).
Muchhasbeenlearnedaboutlonewolvesinthelastseveralyears,butlimitationsofavailableterrorismdatacontinuetoplagueresearchers(Silke,2001). WhiletheAmericanTerrorismStudyhasmaintaineddataonkeyterrorismactivitiesforover25years,ithasonlybeenrecentlythatadequatedatahavebecomeavailabletoconceptuallydistinguishbetweendifferentformsoflonewolfterrorism.Basedonthesedata,theremainderofthisessayseekstocontributetothegrowinglonewolfliteratureinseveralways.First,acategorizationschemeof
lonewolfoffendingisofferedthatconsiderstheextenttowhichoffendersreceivedassistanceforplanningandpreparingforattacksseparatelyfromassistancereceivedintheexecutionof
attacks.Second,theATSprovidesusaccesstotheonlyknownsourceofinformationonthousandsofpre-incidentactivitiesassociatedwithU.S.federalterrorismcases.Increasingourunderstandingoftheseprecursoractivitiescanpotentiallyaidlawenforcementinpreventingseeminglyelusiveformsofterroristoffending.Finally,thisstudyextendstheworkofGilletal.(2014)byexploringtemporalandgeospatialpatternsoflonewolfterroristsincomparisontogroup-basedterroristactors.Inparticular,theextentthatlonewolvestraveltocommittheiracts,engageinpreparatorybehaviors,and“survive”beforebeingdetectedbylawenforcementisexamined.
METHODOLOGY
DataforouranalysisofloneactorterrorismintheUnitedStateswasdrawnfromtheAmericanTerrorismStudy(ATS).TheATSwascreatedin1987incollaborationwiththeFBI’sTerroristResearchandAnalyticalCenter.Atthattime,theFBIprovidedthenamesandcase
numbersofpersonsindictedinfederaldistrictcourtsfrom1980-1987asaresultofanofficialFBIinvestigationfor“terrorismorterrorism-relatedactivities.”Fortheperiod1988to2004,theFBIprovidedthenamesofterrorismdefendantsthroughsponsorshipwitheithertheU.S.HouseofRepresentativesJudiciarySubcommitteeonCrimeortheSenateJudiciaryCommittee.Since
2004,thenameofterrorismdefendantsandtheirrespectivecourtcasenumbershavebeenprovidedbyrequeststhroughCongressionalrepresentativesintheStateofArkansas.
CourtdocumentsforeachcaseareretrievedthroughthefederalPublicAccesstoCourtElectronicRecords(PACER)system.IftheinformationisnotavailableonPACER,ATSpersonnelvisitthedistrictcourtswherethecaseswerefiled,extracttheneededdocuments,andcopythemintotheATSOraclerelationaldatabase.Oncethecourtdocumentsareobtained,theyarereadandquantifiablevariablesareextractedandcodedforentryintothedatabase.ATSprogrammanagersthenreviewthecodeddataforaccuracybeforefinalentry.Quantitativedataoneachcaseincludesprimarily:1)demographicindividualandgroupinformation,2)legalinformationabouteachcount/caseanditsoutcome,and3)anyinformationrelatedtotheprecursorconductoftheterroristspriortocommissionofaterroristincident.Informationontheprecursordataincludes“time-stamping”thedatesoftheseeventsaswellasgeo-codingtheaddresseswheretheseeventstookplace.
Forthecurrentanalysis,weexamined476discrete“indictees”whowerelinkedto264preventedorcompletedactsofterrorismfrom1980-present.Thesepersonswereindictedforover3,000federalcountsandwereassociatedwith1,788knownprecursorbehaviorsatover
1,100addresses.BuildingontheworkofPantucci(2011),Borumetal.(2012),andGruenewaldetal.(2013a),wechosetoexaminethepatternsofbehavioroftheseindicteesbycreatingaparticipatorytypologybasedonthreebasicelements:1)wasthepersonaffiliatedwithagroup
ormovement;2)didtheindividualhavehelpcommittinganyoftheknownprecursorbehaviorsidentifiedincourtrecords;and3)didthepersonhavehelpcommittingtheplannedorcompletedincident.Whilewerecognizethatthisisarathersimpleapproachtoexaminingtheissueoflonewolfbehavior,wechosetodosobecausethethreecategoriesmaybeoperationalizedandreadilyquantified.Assuch,wehopedtoavoidsomeofthemethodologicalproblemsidentifiedby
SpaaijandHamm(2015).
Figure1.ParticipatoryTypology
AsseeninFigure1,theterroristswerecategorizedinthetypologybasedontheextenttowhichtheyscoredpositiveoneachofthesethreeoperationalizedvariables.Personswhowerenotaffiliatedwithaknownterroristgroup,hadnohelpcommittinganyoftheprecursoracts,andwhohadnohelpcommittingtheterroristincidentwerecategorizedasloners.AnexampleinoursamplewouldbeTedKaczynski,theUnabomber.PersonswherewereaffiliatedwithagroupknowntohavebeeninvolvedinterrorismintheUnitedStates,butwhohadnohelpineitherpreparingorcarryingouttheterrorismincidentforwhichtheywereindictedwerelabeledas
affiliatedloners.BufordFurrow,theAryanNationsaffiliatewhoopenedfireinaLosAngeles
Jewishcommunitycenterin1999,isoneexamplefromourdatasetwhofitsthiscategorization.
Thosewhowereaffiliatedwithagroupandwhohadhelppreparingfortheincident,butwhocarriedouttheterrorismincidentbythemselveswerecategorizedasloneconspirators.Whilethetermloneconspiratorseemslikeanoxymoron(aconspiracyrequiresmorethanoneperson,otherwiseitcannotmeettheessentialelementforthelegalrequirementsunderthisparticularinchoatecrime).Inoursituation,however,thetermisquiteappropriatewhenplacedinthecontextofthetypology.TimothyMcVeigh,theOklahomaCitybomberwhoenlistedtheaidofTerryNichols,butwhocarriedouttheattackalone,fitsthiscategory.Finally,thosewhoexhibitedallofthecharacteristics(groupaffiliation,precursorhelp,andhelpcommittingtheincident)wereidentifiedascells/groups.Althoughallofthe9/11participantswhoactuallycarriedouttheattackdiedintheincidentandwerethereforeneverindicted,theyserveasagoodexampleofacell/group.
Samplesizehasbeenaseriousmethodologicalprobleminpreviousexaminationsoflonewolfterrorism.Therefore,forthecurrentanalysis,wehavechosentocollapseourtypologyintotwosimplecategories–thosewhohadhelpandthosewhodidnot.Lonersandaffiliatedlonersfallintothefirstcategory,whileloneconspiratorsandcells/groupsfallintothecategoryin
whichtheparticipant(s)hadhelp.Wecomparedthesetwogroups(thosewhohadhelpandthosewhodidnot)onfivemajorissues:1)successfulcompletionoftheterroristincident,2)volume
ofprecursoractivity,3)lengthoftheplanningcycle,4)spatialvariances,and5)lengthoftheindividualterrorist’s“lifecycle.”
RESULTS
Beforepresentingtheresultsofouranalysesontheissuesabove,somediscussionofthecharacteristicsofthesampleisappropriate.Asnotedearlier,thesampleincludes476discreteindividualswhowereinvolvedin264“officiallydesignated”terrorismincidentsintheUnitedStates.However,sixteenoftheseindividualswereinvolvedinmultipleterrorismincidentsasbotha“loner”andaspartofa“cell/group.”Consequently,forouranalysis,thesesixteen
personsarecountedmorethanonce,renderingananalysisofthebehaviorof492individuals.Aspreviousscholarshavenoted,lonewolfactivityisrelativelyrare,evenamongterrorists.Onlyeightpercent(n=37)ofthe492personscommittedterrorismincidentswithoutanyhelpinplanning,preparing,orcarryingouttheincident.
Table1.DemographicsofLonersvs.Cells/Groups
LonersCells/Groups
Perpetrators8%
n=37
Gender27%female**
n=37
AverageAge34yearsoldn=35
CollegeEducation89%attendedcollege***(n=28)
MaritalStatus19%hadbeenmarried***
n=32
92%
n=455
12%female**
n=455
36yearsoldn=391
59%attendedcollege***
n=324
56%hadbeenmarried***
n=294
**p<.01,***p<.001
Onaverage,lonerswereonlyslightlyyoungerthancell/groupparticipants,34yearscomparedto36yearsofage,respectively.Thisfindingisremarkablyconsistentwithotherscholars(Gilletal.,2014;Gruenewaldetal.,2013). However,ourmeasureofagewas“ageat
indictment”andasweshalllatersee,lonersparticipateinterrorismformuchlongerperiodsoftimebeforearrestandindictmentthandocell/groupparticipants.Ifwehadused“ageatfirstpreparatoryconduct,”thereislittledoubtthatlonerswouldhavebeensignificantlyyoungerthancell/groupparticipants.Wethencomparedlonerstocell/groupmembersonthreeotherdemographictraits:gender,education,andmaritalstatus.Lonersweresignificantlydifferentfromcell/groupparticipantsoneachofthesevariables.Lonersweremuchlesslikelytobemarried(19%comparedto56%),afindingthatisconsistentwithmostofthelonewolfliteratureandwhichsupportsthenotionthatlonerstendtobesociallyisolated(Nijboer,2012). Theyalsotendedtobemuchbettereducatedthancell/groupparticipants.Nearlynineoutoften(89%)ofthelonershadattendedcollegetosomeextentcomparedwith59%ofcell/groupmembers.Ourdemographicfindingsdifferedsignificantlyfromotherscholarsononetrait(Gilletal,2014).Althoughterrorismisanoverwhelminglymale“occupation,”wefoundthatlonerswerecomprisedofasignificantlylargerpercentageoffemalesthanwerecell/groupparticipants.Approximatelyonefourthofthelonersinthesamplewerefemales,aninterestingandsingularlydifferentfindingfromotherscholars.Thisdifferencecanbeattributedtothenumberoffemaleswhofallintoouraffiliatedlonercategory.Thesewerepredominatelywomenwhoaffiliated
withenvironmentalextremistgroups,butwhopreparedandcommitted“ecotage”activitieswithnoevidenceofhelpinpreparingorcommittingtheincident.
Table2.IncidentCompletionsbyLonersvs.Cells/Groups
AttackCompleted
Successfully
AttackNot
CompletedSuccessfully
LonerIncidents
n=6534%66%
Cell/GroupIncidents
n=19944%56%
AllIncidents
n=26442%58%
p<.10
Weexaminedwhethertheincidentwassuccessfullycompletedbycomparingincidentswheretheattackwascarriedoutandknownobjectivesweremetwiththoseinwhichtheincidentfailedand/orwasfoiled.Althoughwefoundaslightdifferencebetweenthefailureratesof
lonersandcells/groups,thesedifferenceswerenotsignificant.ThisisincontrasttotherecentworkofGilletal.(2014)whofoundthatterroristsoperating“withoutcommandandcontrollinks”weresignificantlymorelikelytosuccessfullyexecuteanattackthanthoseoperatingundersomeformofcommandandcontrol.Itshouldbenoted,however,thatGillandhiscolleagues’sampleincludesbothAmericanandEuropeanterrorists,whileoursampleislimitedtoAmericanterrorism.
Table3.PrecursorActivitiesPerIncident.
Precursor
ActsIncidents
PrecursorActs
PerIncident
LonerIncidents162652.49
Cell/GroupIncidents1,6751998.42
AllIncidents1,8372646.96
p<.001
Next,weexaminedthevolumeofprecursorconductspecificallyassociatedwiththeplanningandpreparationofanincident(Table3). Asonemightexpect,wefoundasignificantdifferencebetweenlonersandcells/groupsonthismeasure.Membersofcells/groupsengagedinoverthreetimesasmanypreparatorybehaviorsperincidentthanloners(8.4comparedto2.5,respectively).Forlawenforcementplanning,thelesservolumeofactivitypriortoanincidentisoneofthemajorreasonslonewolftacticshavebecomeofsuchconcern(Hewitt,2014). Toour
knowledge,however,thisisthefirstempiricalevidenceconfirmingthispattern.AlsoofparticularnoteinTable3istheratioofincidentstogroupsize.Althoughnotcalculatedexplicitly,loners(whocompriseonly8%ofthesample)accountedforone-fourthoftheterrorismincidents(65of264). Thisdisproportionatenumberofincidentsshouldbesubstantialcauseforconcerntocounterterrorismofficials.However,itshouldalsobenoted,thatthefewerthenumberofprecursoractivities,thegreaterthelikelihoodthattheplannedincidentisoflessermagnitude.ThispostulateisbeingtestedcurrentlywithfundingfromtheNationalInstituteof
Justice.2
Table4.LengthofPlanningCycleforAttacks.
TimeSpanfromFirstPrecursorActtoIncident
Lessthan3months3monthsto1yearLongerthan1year
LonerIncidents
n=3033%33%33%
Cell/GroupIncidents
n=15433%31%36%
AllIncidents
n=18433%31%36%
ns
Wewerealsointerestedinassessingwhetherincidentsinvolvinglonersrequiredashorterplanningcycle.Sinceourearlierfindingindicatedthattheyengagedinsubstantiallyfewerprecursoractivitiesinpreparationforanincident,itwaslogicaltohypothesizethattheir
planningcyclewouldbeshortened.Interestingly,thiswasnotthecase.Usingincrementsof“lessthanthreemonths,”“threemonthstooneyear,”and“longerthanoneyear,”Table4demonstratesthealmostidenticalpatternsamongthetwogroupsrelativetotheoveralltemporaldimensionsoftheplanningprocess.
Sincetemporalpatternsmaybeshapedbyspatiallimitationsorconstraints,wewantedtocomparethebehavioroflonersandcells/groupsontwoissues:1)howfartheylivedfromwheretheycommittedtheirpreparatoryactivities,and2)howfartheylivedfromtheselectedterrorismincidentlocation.Theseanalyseswereconductedthroughaseriesof“rosediagrams”inwhichtheresidencesoftheterroristsareplacedinthecenterofaseriesofconcentriccirclesrepresentingdistancesfromtheresidence.Thenthelocationofeithertheprecursorconductor
theterroristincidentlocation(dependingonwhatisbeinganalyzed)islocatedontherosediagramusingbothadistancemeasurementandanazimuth.Thisstrategyrendersavisualdepictionofboththedistancesanddirectionfromtheindependenttothedependentvariable,inthiscase,residencelocationstoeitherthelocationsofprecursoractivitiesorlocationsoftheincidents.3
Table5.LinearDistancesfromResidencestoPrecursorActivities.
Loners / Cells/Groups / AllPerpetratorsAverageDistance / 454miles / 444miles / 445miles
MedianDistance / 170miles / 79miles / 88miles
StandardDeviation / 638 / 669 / 666
MinimumDistance / 0miles / 0miles / 0miles
MaximumDistance / 2,571miles / 2,696miles / 2,696miles
Forthecurrentanalysis,thepatternsaredemonstratednumericallyinTables5and6. Toavoidtheimpactofoutliers,theuseof“mediandistances”arethemostmeaningful.Mediandistancesindicatethathalfofthebehaviorsoccurredcloserthanthemedian,whiletheotherhalfoccurredfartherthanthemediandistance.Table5revealsthatloners,onaverage,committedtheirprecursorconductovertwiceasfarfromtheirplacesofresidencethandidmembersof
cells/groups(170and79miles,respectively).Overtwo-fifths(41%)ofthepreparatorybehaviorcommittedbycell/groupmembersoccurredwithinthirtymilesoftheirresidences;lessthanone-fourth(23%)oflonerscommittedpreparatoryactswithinthesamethirty-mileradius(tablenotshown).
Table6.LinearDistancesfromResidencestoIncidents.
Loners / Cells/Groups / AllPerpetratorsAverageDistance / 604miles / 374miles / 403miles
MedianDistance / 328miles / 118miles / 132miles
StandardDeviation / 685 / 543 / 567
MinimumDistance / 0.57miles / 0.32miles / 0.32miles
MaximumDistance / 2,563miles / 2,696miles / 2,696miles
Thesamepatternemergeswhenexaminingthedistancesbetweenresidenceoftheterroristandincidentlocation(seeTable6). Lonerstraveled,onaverage,almostthreetimesfurtherfromtheirplacesofresidencetostriketargetsthandidthemembersofcells/groups(328milescomparedto118miles,respectively).Ananalysisoftheconcentricpatternsassociatedwiththisdatarevealedthatoverone-third(37%)ofcell/groupmemberslivedwithinthirtymilesoftheincidentlocationcomparedwithlessthanone-fifth(18%)oflonerswholivedthatclosetothetarget.Fifty-fivepercentoflonersresidedgreaterthan270milesfromtheincidentlocationcomparedwithonly38%ofcell/groupmembers(tablenotshown). Allofthesepatternssuggestthatlonersgotomuchgreaterlengthstodisguisetheiractivitiesandavoiddetectionthandomembersofcellsorgroups. Insomeways,theydefythelogicoftimeandspacethatsuggeststhatgreaterdistancesrequirelongerperiodsoftimeforplanning,preparation,andexecution.Whilethatgeneralpremiseappearstobetrueofcell/groupaffiliatedterrorists,thepatternisnot
evidencedinthebehaviorofloners–theytravelgreaterdistancestoprepareforanincidentandtheylivesubstantiallygreaterdistancesfromtheirchosentargetlocations,yettheyexhibitaplanningprocessthatisofessentiallythesamelengthasterroristsoperatingunderacommandandcontrolstructurewithmultiplemembers.Thesebehaviors,combinedwiththeirrelativeisolation(asevidencedbytheirlowermarriagerates)andotherindicatorssuggestedbyrecentscholarship(Gilletal.,2014;Hewitt,2014;Becker,2014)allsuggestthatlonersdoindeedposeadifficulttaskforlawenforcementearlyintervention.
Table7.LengthofTerrorist’sLifeCycle
PercentofTerroristPrecursorActsbyTimePeriod
Lessthan3months3monthsto1yearLongerthan1year
Loners
n=248%13%79%
Cells/Groups
n=27922%28%50%
AllPerpetrators
n=30321%26%53%
p<.001
Giventhesefindings,wewonderedifthebehaviorsoflonersenabledthemtosurvivelongerthanmembersofcells/groupsbeforebeingarrestedorkilled.Thedifferences,reportedinTable7,werehighlysignificant.Proportionally,nearlythreetimesasmanycell/groupmemberswerearrestedwithinthreemonthsoftheirfirstprecursoractivitycomparedtoloners(22%and
8%,respectively).Similarly,three-fifths(60%)ofcell/groupmembersinthedatasetwerearrestedwithinoneyearoftheirfirstknownprecursoractivity.Incontrast,onlyaboutone-fifth(21%)oflonerswerearrestedduringthefirstyearfollowingengagementinprecursorconduct.Ananalysisof“longevity”measuredasacontinuousvariablerangingfromdateoffirstknown
precursoracttodateofindictmentrevealedevenmoreastoundingresults.Onaverage,cell/groupparticipants“survived”aboutoneyear(medianof370days)beforebeingindicted,whilelonershadanaverage“lifespan”fromfirstprecursortoindictment(orarrest,ifnotcapturedpriortoindictment)ofover1,900days–aperiodofoverfiveyears.Whilesomeoftheseloners,exemplifiedbyterroristslikeOlympicParkandabortionclinicbomberEricRudolph,desistedfromterroristsandlived“underground”foryearsaftercommittingtheirlast
actofterrorism(Vollers,2006),theirabilitytoavoidcapturefuelstheimaginationofindividualsradicalizingtowardviolence.
CONCLUSIONS
Despitealackofconsensusaboutwhatconstitutesa“lonewolf,”ourunderstandingofterrorismcommittedbylonershasincreasedinthepastseveralyears.Likeothers,wesuggestthatasharedterminologyneedstoconsiderthreekeyconceptualdimensionsofloneactorterrorism,includingideologicalmotive,groupaffiliation,andthedegreetowhichterroristsreceiveassistanceinpreparingandexecutingterroristattacks.Itwasbasedonthesedimensionsthatthisstudysoughttoextendpreviouscategorizationschemesbydistinguishingbetweenloneactorterroristswhohadnohelppreparingforattacksandexecutedthemalonefromthoselonerswhoreceivedassistancepreparingforattacksbutexecutedthemalone.Asisoftenthecasewithterrorismresearch,ourlimitedsamplesizeprecludedusfrommakingstatisticalcomparisonsacrosseachofourthreetypesoflonewolfterrorisminrelationtogroup-basedterrorism.Wedid,however,statisticallyexaminehowcasesinvolvinglonewolves,whichincludedboth“loners’and“unaffiliatedloners”whohadnohelpinpreparingfororexecutingattacks,comparedtoallotherterrorismcasesinvolvingmultipleconspirators.Theresultsofourstudy
suggestedthatthemajorityoflonewolveswerepredominatelymale,slightlyyounger,andsignificantlymoreeducatedthangroup-basedterrorists.Ouranalysisalsosupportedpreviousresearchbyfindingthatlonewolveswere,onaverage,moresociallyisolatedthangroup-basedactors.Unfortunately,theincreasedisolationoftheseterroristsisoneofthemostproblematicsymptomsofthelonewolf“syndrome”forlawenforcementtocounter.Fewersocialtiesconceivablyreducetheopportunitiesforfamily,friends,employers,andothersknowntoterroriststodetectsuspiciousbehaviors,andopportunitiesforthepublicto“seesomething,saysomething”arereduced.
Alsoperplexingtothelawenforcementandintelligencecommunityistheassumptionthatthatlonewolvesengageinfewerprecursor,orplanningandpreparatory,activitiesthatalsoreduceopportunitiesforinterdictionbylawenforcement.WiththeexceptionsofGilletal.(2014)andthecurrentstudy,therehasbeenalackofempiricalresearchonthedifferencesinprecursorbehaviorsoflonewolvesandgroup-basedterrorists.BasedondatamadeavailablebytheAmericanTerrorismStudy,wenotonlyconfirmedconventionalwisdomthatlonewolvesengageinlessdetectableprecursoractivitiesbutwealsofoundthattheseactivitiesaresimilarlytemporallyspacedandincreasinglygeospatiallydistancedcomparedtogroup-basedterrorismcases.Aseachprecursoractivityrepresentsapotentialopportunityforcriminalandsuspiciousbehaviorstobedetected,lonewolvesprovidefewerandmoredistanced“dots”thatcanbeconnectedbyofficialspriortoterrorismattacks.
Althoughourfindingssupporttheemergingnarrativethatdepictslonewolvesasaformidablechallengeandsignificantthreattohomelandsecurity,itisalsoimportantto
rememberthatnoterroristoperateswithinasocialvacuum.Lonewolvesareuniquefromgroup-basedterroristsinseveralways;however,theydoplan,prepare,andmayevensolicitdirectand
virtualformsofassistancepriortocommittingtheircrimes.Itisclearthatlonewolvestendtobemoreisolatedthanotherterroristsinregardstopersonalandintimaterelationships,buttheseindividualsareoftenemployed,attendingclasses,membersofthemilitary,ormembersofonline(orInternet)communities.Itisuptoterrorismresearcherstolearnmoreaboutthepatternednatureofcriminalandnon-criminalprecursoractivitiesoflonewolves,whilealsobeingcarefulnottoassumehomogeneityinprecursoractivitiesacrossloners,unaffiliatedloners,loneconspirators,andgroup-basedactors.Ideally,membersofthelawenforcementandintelligencecommunitieswillultimatelybeabletouseevidence-basedfindingssuchastheonespresentedinthisstudytoinformterrorismpreventionpoliciesandpractices
REFERENCES
Bakker,E.,deGraaf,B.(2011).Preventinglonewolfterrorism:SomeCTapproaches.
PerspectivesonTerrorism,5(5-6),Retrievedfrom
Barnes,B.(2012).Confrontingtheone-manwolfpack:Adaptinglawenforcementandprosecutionresponsestothethreatoflonewolfterrorism.BostonUniversityLawReview,
92,1613–1662.
Beam,L.1992.Leaderlessresistance.TheSeditionist,12.Retrievedfrom
Becker,M.(2014).ExplaininglonewolftargetselectionintheUnitedStates.StudiesinConflict
Terrorism,37(11),959-978.
Borum,R.,Fein,R.,Vossekuil,B.(2012).Adimensionalapproachtoanalyzingloneoffenderterrorism.AggressionandViolentBehavior,17,389-396.
Damphousse,K.,Smith,B.(2004). Terrorismandempiricaltesting:Usingindictmentdatatoassesschangesinterroristconduct.InM.Deflem(Ed.),Terrorismandcounterterrorism:Criminologicalperspectives(pp.75-90).NewYork:Elsevier.
Gill,P.,Horgan,J.,Deckert,P.(2014).Bombingalone:Tracingthemotivationsandantecedentbehaviorsoflone-actorterrorists.JournalofForensicSciences,59(2),425-
435.
Gruenewald,J.,Chermak,S.,Freilich,J.D.(2013b).Distinguishing“loner”attacksfromotherdomesticextremistviolence:Acomparisonofhomicideincidentandoffendercharacteristics.CriminologyPublicPolicy,12(1),65-91.
Hewitt,C.(2003).UnderstandingterrorisminAmerica:FromtheKlantoAlQaeda.NewYork:Routledge.
Hewitt,C.(2014).LawenforcementtacticsandtheireffectivenessindealingwithAmericanterrorism:Organizations,autonomouscells,andlonewolves.TerrorismPoliticalViolence,26(1),58-68.
Joosse,P.(2007).Leaderlessresistanceandideologicalinclusion:ThecaseoftheEarth
LiberationFront.TerrorismPoliticalViolence,19(3),351-368.
Moskalenko,S.,McCauley,C.(2011).Thepsychologyoflone-wolfterrorism.Counseling
PsychologyQuarterly,24,115–126.
Nijboer,M.(2012).Areviewoflonewolfterrorism:Theneedforadifferentapproach.SocialCosmos,3.Retrievedfrom
Pantucci,R.(2011).Atypologyoflonewolves:PreliminaryanalysisofloneIslamistterrorists.
UnitedKingdom.TheInternationalCentrefortheStudyofRadicalisationandPoliticalViolence.Retrievedfrom
Ranstorp,M.(1998).InterpretingthebroadercontextandmeaningofBin-Laden’sfatwa.StudiesinConflictTerrorism,21(4),321-330.
Silke,A.(2001).Thedevilyouknow:Continuingproblemswithresearchonterrorism.
TerrorismPoliticalViolence,13(4),1-14.
Simon,J.D.(2013).Lonewolfterrorism:Understandingthegrowingthreat.Amherst,NY:PrometheusBooks.
Smith,B.(1994). TerrorisminAmerica:Pipebombsandpipedreams.Albany,NY:State
UniversityofNewYorkPress.
Spaaij,R.(2010).Theenigmaoflonewolfterrorism:Anassessment.StudiesinConflictTerrorism,33,854–870.
Spaaij,R.,Hamm,M.S.(2015).Keyissuesandresearchagendasinlonewolfterrorism.
StudiesinConflictTerrorism,38(3),167-178.
Sterling,C.(1981).Theterrornetwork:Thesecretwarofinternationalterrorism.NewYork:Holt,Rinehart,andWinston.
Stern,J.(2003).TerrorinthenameofGod:Whyreligiousmilitantskill.NewYork:Harper
Collins.
Vollers,M.(2006).Lonewolf:EricRudolphandthelegacyofAmericanterror.NewYork:HarperPerennial.
NOTES
1Personalconversationsandmeetingsin1995-96withmembersofJTTF’sindifferentFBIfieldofficesfollowingthebombingoftheMurrahFederalBuilding.OneoftheauthorsofthisarticlewascriticizedextensivelyinanarticlebyWilliamPierceonhiswebsitefollowingtheauthor’stestimonybeforetheHouseJudiciarySubcommitteeonCrimeaftertheOklahomaCity
Bombing.Concernsthatmentionoftheauthorbynamemightbesuggestiveofatargetforalone
wolfresultedinincreasedsecuritybylawenforcement–thepointbeingthattheterm“lonewolf”wasusedextensivelyevenpriortoAlQaeda’semergenceasaglobalthreat.
2SequencingTerrorists’PrecursorBehaviors:ACrime-SpecificAnalysis.NationalInstituteof
Justice,Grant#2013-ZA-BX-0001.
3Therosediagramsarenotshowhere,butexamplesmaybeseeninsomeofourpreliminaryworkpublishedintheNIJJournal,IssueNumber260,pp.2-6.