Вы находитесь на странице: 1из 4

Arewewitnessingadishonestelection?

Abetweenstatecomparisonbasedontheusedvotingprocedures

ofthe2016DemocraticPartyPrimaryfor

thePresidencyoftheUnitedStatesofAmerica

AxelGeijsel
TilburgUniversityTheNetherlands

RodolfoCortesBarragan
StanfordUniversityU.S.A.

June7,2016

Youcanfoolsomeofthepeopleallofthetime,andallofthepeoplesomeofthetime,butyou
cannotfoolallofthepeopleallofthetime.
AbrahamLincoln

Noonehasyetfiguredoutastraightforwardmethodofensuringthatoneofthemostrevered
democraticinstitutionsinthiscase,electingaU.S.presidentcanbedoublecheckedfor
fraud,particularlywhenpaperlessevotingsystemsareused.LarryGreenemeier,
Scientific
American

SummaryStatement

Giventhestakesintheoutcomeofthe

Americanpresidentialelections,ensuringthe
integrityoftheelectoralprocessisoftheutmostimportance.Aretheresultswearewitnessing
inthe2016primaryelectionstrustworthy?WhileDonaldTrumpenjoyedaclearandearlyedge
overhisRepublicanrivals,theDemocraticcontestbetweenformerSecretaryofStateHillary
ClintonandSenatorBernardSandershasbeenfarmorecompetitive.Atpresent,Secretary
ClintonenjoysanapparentadvantageoverSanders.Isthisclaimedadvantagelegitimate?We
contendthatitisnot,andsuggestanexplanationfortheadvantage:Statesthatareatriskfor
electionfraudin2016systematicallyandoverwhelminglyfavorSecretaryClinton.Weprovide
convergingevidenceforthisclaim.

First,weshowthatitispossibletodetectirregularitiesinthe2016DemocraticPrimaries
bycomparingthestatesthathavehardpaperevidenceofalltheplacedvotestostatesthatdo
nothavethishardpaperevidence.Second,wecomparethefinalresultsin2016tothe
discrepantexitpolls.Furthermore,weshowthatnosuchirregularitiesoccurredinthe2008
competitiveelectioncycleinvolvingSecretaryClintonagainstPresidentObama.Assuch,we
findthatinstateswhereinvotingfraudhasthehighestpotentialtooccur,systematiceffortsmay
havetakenplacetoprovideSecretaryClintonwithanexaggeratedmarginofsupport.

Differentoutcomesinprimarystateswithpapertrailsandwithoutpapertrails

Dataprocurement
:Giventhepotentialthattheunderlyingvotingnumberhasbeencorrupted,
wehadtorestrictouranalysistoaproxy:thepercentageofdelegateswonbySecretaryClinton
andSenatorSanders.Togroupstatesaccordingtotheaccountabilityofthevote,weused
Ballotpediaandcreatedtwogroups.First,thereare18statesthatfeaturevotingprocedures
whereintheaccuracyofelectoralresultsofaprimaryballotvotearebackedbyapapertrail.
Second,thereare13statesthatdonothavesuchapapertrail.

Analysis
:
The[
data
]showastatisticallysignificantdifferencebetweenthegroups.States
withoutpapertrailsyieldedhighersupportforSecretaryClinton,(M
=65.13%,SD=
nopapertrail
no
=10.41%)thanstateswithpapertrails(M
=48.53%,SD=
=16.00%),t(29)
papertrail
papertrail
papertrail
=3.21,P=0.003,d=1.19[Figure1].Assuch,thepotentialforelectionfraudinvoting
proceduresisstronglyrelatedtoenhancedelectoraloutcomesforSecretaryClinton.Inthe
Appendix,weshowthatthisrelationshipholdsevenaboveandbeyondalternativeexplanations,
includingtheprevailingpoliticalideologyandthechangesinsupportovertime.

Supplementalanalysisoncaucusstates:
Doesthepatternseeninballotstatesoccurincaucus
states?Bytheverynatureofcaucusingprocedures,caucusresultsaregenerallythoughttobe
moretrustworthy.However,inthecurrentDemocraticcaucusingcycle,IowaandNevadahad
caucuseswidelyallegedtohaveinvolvedaconsiderablelevelofvotersuppressionand
potentialfraud.Weexaminedthe[
data
]andfoundthatthesetwostateshadfarhighersupport
forSecretaryClinton,[M
=54.71%,SD=
=3.44%]thantheothercaucus
fraudallegations
fraudallegation

states,[M
=31.61%,SD=
=9.98%],t
(11)=3.13,P=
nofraudallegations
nofraudallegations
independentmeans
0.009,d=3.10.

Anomaliesexistbetweenexitpollsandfinalresults

Dataprocurement
:Weobtainedexitpolldatafroma
database
keptbyanexpertonthe
Americanelections.

Analysis
:Ontheoverall,aretheexitpollsdifferentfromthefinalresults?Yestheyare.The
data
showlowersupportforSecretaryClintoninexitpollsthanthefinalresultswouldsuggest,[M
exit
=54.38%,SD=
=13.95%M
=57.52%,SD=
=13.87%],t
(23)=3.49,P=
exit
final
final
dependentmeans
0.002,d=0.71.

Whileaneffectsizeof0.71isquitesubstantial,andsuggestsaconsiderable
differencebetweenexitpollsandoutcomes,weexpectedthatthisdifferencewouldbeeven
moreexaggeratedinstateswithoutpapervotingtrails.Indeed,theeffectsizeinstateswithout
papervotingtrailsisconsiderablylarger:1.50,andyieldsmoreexaggeratedsupportforthe
Secretaryinthehoursfollowingtheexitpolls[M
=62.93%,SD=
=8.80%M
=65.68%,
exit
exit
final
SD=
=9.52%],t
(9)=4.68,P<0.001.Incontrast,theeffectsizeismuchsmaller
final
dependentmeans
instateswithpapertrails,[M
=48.28%,SD=
=13.94%M
=51.69%,SD=
=
exit
exit
final
final
13.77%],t
(13)=2.27,P=0.04,d=0.58.
dependentmeans

Irregularitiesareuniqueto2016

ToshowthatthepatternofvotesmaysuggestasystematicefforttoundercutSenatorSanders,
wemustshowthatnosuchpatternswereinplaceinsimilarelections.GiventhatSecretary
ClintonlosttoPresidentObamain2008,theirdataisanaturalcontrolandthebestpossible
pointofcomparisonforthe2016data.Thus,aswedidfor2016,wetabulatedthepercentageof
delegateswonineachstateby(thenSenator)HillaryClinton.The
data
showthat,contraryto
the2016data,thereisnoevidencethatprimarystateswithoutpapertrailsfavoredSenator
Clintonin2008,P=0.38.Assuch,thepatternsof2016aredifferentfromtheirbestpointof
comparison.

Conclusion

Arewewitnessingadishonestelection?Ourfirstanalysisshowedthatstateswhereinthevoting
outcomesaredifficulttoverifyshowfargreatersupportforSecretaryClinton.Second,our
examinationofexitpollingsuggestedlargedifferencesbetweentherespondentsthattookthe
exitpollsandtheclaimedvotersinthefinaltally.Beyondthesepoints,theseirregularpatterns
ofresultsdidnotexistin2008.Assuch,asawhole,thesedatasuggestthatelectionfraudis
occurringinthe2016DemocraticPartyPresidentialPrimaryelection.Thisfraudhas
overwhelminglybenefitedSecretaryClintonattheexpenseofSenatorSanders.


Figure1.PercentofsupportforClintonandSandersbystatevotingpapertrail
status.

Appendix,SupplementalAnalyses,andReferences

Вам также может понравиться