Statics and probability snow ball sampling

what is snow ball sampling, methods, application field and advantages and disadvantages

1. 1
Statistics&Probability
SnowballSampling

2. 1
Whatissnowballsampling?
Snowballsamplingusesasmallpoolofinitial
informantstonominate,throughtheirsocialnetworks,otherparticipantswhomeetthe
eligibilitycriteriaandcouldpotentiallycontributetoaspecificstudy.Theterm
"snowballsampling"reflectsananalogytoasnowballincreasinginsizeasitrolls
downhill.
Method:
1.Draftaparticipationprogram(likelytobesubjecttochange,butindicative).
2.Approachstakeholdersandaskforcontacts.
3.Gaincontactsandaskthemtoparticipate.
4.Communityissuesgroupsmayemergethatcanbeincludedintheparticipation
program.
5.Continuethesnowballingwithcontactstogainmorestakeholdersifnecessary.
6.Ensureadiversityofcontactsbywideningtheprofileofpersonsinvolvedinthe
snowballingexercise.
Applicationsofsnowballsampling:
Theparticipantsarelikelytoknowotherswhosharethecharacteristicsthatmakethem
eligibleforinclusioninthestudy.
Snowballsamplingisquitesuitabletousewhenmembers
ofapopulationarehiddenanddifficulttolocate(e.g.samplesofthehomeless
orusersofillegaldrugs)andthesemembersarecloselyconnected(e.g.
organizedcrime,sharingsimilarinterests,involvementinthesamegroupsthat
arerelevanttotheprojectathand)
SnowballSampling

3. 2
Applicationfield:
1.Socialcomputing:
Snowballsamplingcanbeperceivedasanevaluation
samplinginthesocialcomputingfield.Forexample,intheinterviewphase,
snowballsamplingcanbeusedtoreachhard-to-reachpopulations.
Participantsorinformantswithwhomcontacthasalreadybeenmadecanuse
theirsocialnetworkstorefertheresearchertootherpeoplewhocould
potentiallyparticipateinorcontributetothestudy.
2.Conflictenvironment:
Ithasbeenobservedthatconductingresearchin
conflictenvironmentischallengingduetomistrustandsuspicion.Aconflict
environment,wherepeopleorgroupsthinkstheirneedsandgoalare
contradictorytothegoalsandorneedsofotherpeopleorgroup.These
conflictsamonggroupsorpeopleincludethedifferencestoclaimtheareaof
territory,resources,trade,civilandreligiousrightsthatcauseconsiderable
misunderstandingandheightenthedisagreementsthatleadtoanenvironment
withlackoftrustandsuspicion.Inconflictenvironment,theentirepopulation
ismarginalizedtosomeextentratherthanaspecificgroupofpeopleand
makesitveryhardforinvestigatorstoreachthestudysubjectstoconductthe
research.Forexample,athreateningpoliticalenvironmentunderauthoritarian
regimecreatesobstaclesfortheinvestigatorstoconducttheresearch.
Snowballsamplinghasdemonstratedasasecondbestmethodinconducting
researchinconflictenvironmentslike,inthecontextoftheIsraelandArab
Conflict.Snowballsamplingallowstheinvestigatorstoapproachthe
marginalizedpopulationatcognitiveandemotionallevelandenrollthemin
study.Snowballsamplingaddresstheconditionsoflackoftrustthatarises
duetouncertaintyaboutthefuturethroughtrace-linkingmethodology.
3.Expertinformationcollection:
Snowballsamplingcanbeusedtoidentify
expertsinacertainfieldsuchasmedicine,manufacturingprocesses,or
customerrelationmethods,andgatherprofessionalandvaluableknowledge.
Forinstance,3Mcalledinspecialistsfromallfieldsthatrelatedtohowa
surgicaldrapecouldbeappliedtothebodyusingsnowballsampling.Every
involvedexpertcansuggestanotherexpertwhotheymayknowcouldoffer
moreinformation.

4. 3
AdvantagesandDisadvantages
Advantages:
I.Locatehiddenpopulations:
Itispossibleforthesurveyorstoincludepeopleinthe
surveythattheywouldnothaveknownbut,throughtheuseofsocialnetwork
.
II.Locatingpeopleofaspecificpopulation:
Therearenolistsorotherobvioussourcesforlocating
membersofthepopulation(e.g.thehomeless,usersofillegaldrugs).The
investigatorsusepreviouscontactandcommunicationwithsubjectsthen,the
investigatorsareabletogainaccessandcooperationfromnewsubjects.Thekey
ingainingaccessanddocumentingthecooperationofsubjectsistrust.Thisis
achievedthatinvestigatorsactingoodfaithandestablishgoodworking
relationshipwiththesubjects.
III.Methodology:
Assubjectsareusedtolocatethehiddenpopulation,the
researcherinvestslessmoneyandtimeinsampling.Snowballsamplingmethod
doesnotrequirecomplexplanningandthestaffingrequiredisconsiderably
smallerincomparisontoothersamplingmethods.
SnowballSamplingcanusein
bothalternativeorcomplementaryresearchmethodology.Asanalternative
methodology,whenotherresearchmethodscannotbeemployed,dueto
challengingcircumstancingandwhenrandomsamplingisnotpossible.As
complementarymethodologywithotherresearchmethodstoboostthequality
andefficiencyofresearchconductandtominimizethesamplingbiaslikequota
sampling.
Disadvantages:
1.Communitybias:
Thefirstparticipantswillhaveastrongimpactonthesample.Snowball
samplingisinexactandcanproducevariedandinaccurateresults.Themethodisheavily
reliantontheskilloftheindividualconductingtheactualsampling,andthatindividual's
abilitytoverticallynetworkandfindanappropriatesample.Tobesuccessfulrequires

5. 4
previouscontactswithinthetargetareas,andtheabilitytokeeptheinformationflowgoing
throughoutthetargetgroup.
2.Non-random:
Snowballsamplingcontravenesmanyoftheassumptionssupporting
conventionalnotionsofrandomselectionandrepresentativeness.However,social
systemsarebeyondresearchers'abilitytorecruitrandomly.Snowballsamplingis
inevitableinsocialsystems.
3.Unknownsamplingpopulationsize:
Thereisnowaytoknowthetotalsizeoftheoverallpopulation.
4.Anchoring:
Anotherdisadvantageofsnowballsamplingisthelackofdefinite
knowledgeastowhetherornotthesampleisanaccuratereadingofthetargetpopulation.
Bytargetingonlyafewselectpeople,itisnotalwaysindicativeoftheactualtrends
withintheresultgroup.Identifyingtheappropriatepersontoconductthesampling,as
wellaslocatingthecorrecttargetsisatime-consumingprocesssuchthatthebenefits
onlyslightlyoutweighthecosts.
5.Lackofcontroloversamplingmethod:
Asthesubjectslocatethehiddenpopulation,
theresearchhasverylittlecontroloverthesamplingmethod,whichbecomesmainly
dependentontheoriginalandsubsequentsubjects,whomayaddtotheknownsampling
poolusingamethodoutsideoftheresearcher'scontrol.