非随机抽样和随机抽样的比较.pptx

《非随机抽样和随机抽样的比较.pptx》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、Ch.7SelectingsamplesTheneedtosampleOverviewofSamplingtechniquesProbabilitysamplingNon-probabilitysamplingDefinitionoftermsCensusCollectandanalysedatafromeverypossiblecaseorgroupmemberSamplingArangeofmethodsthatenableresearchertoreducetheamountofdatabyonlydatafromasubgrou

2、pratherthanallpossiblecasesorelementsPopulationThefullsetofcasesfromwhichasampleistakenFigure7.1Population,sampleandindividualcases1.TheneedtosampleBudgetconstraintspreventyoufromsurveyingtheentirepopulationTimeconstraintspreventyoufromsurveyingtheentirepopulationImpract

3、icabletosurveytheentirepopulationYouhavecollectedallthedatabutneedtheresultsquickly2.OverviewofsamplingtechniquesProbabilityorrepresentativesampling-Eachcasefrompopulationisknownandusuallyisequalforallcases-(surveyandexperimentalresearchstrategies)Non-probabilityorjudgem

4、entalsampling-Probabilityofeachcasefromthepopulationisunknown-Impossibletoanswerresearchquestionsortoaddressobjectivesthatrequireyoutomakestatisticalinferencesaboutthecharacteristicsofthepopulation-(casestudystrategy)非随机抽样和随机抽样的比较抽样方法作用抽样原则误差判断应用优缺点非随机抽样研究总体的局部现象非随机抽出样本，

5、主观性强不能计算和判断抽样误差可随时随地采用不够科学规范，但省钱、省事、灵活方便随机抽样以部分推断总体随机抽出样本，客观性强不能计算和判断抽样误差只能定期采用科学规范，但费时、费钱、不够灵活方便Figure7.2Samplingtechniques随机抽样非随机抽样简单随机抽样系统抽样分层抽样分群抽样多步抽样配额抽样雪球抽样便利抽样自选抽样目的抽样极端抽样同质抽样不均匀抽样典型抽样关键抽样3.ProbabilitysamplingProcessofprobabilitysampling(2)decideonasuitablesample

6、size(3)selectthemostappropriatesamplingtechniqueandselectthesample(4)checkthesampleisrepresentativeofthepopulation(1)identifyasuitablesamplingframebasedonyourresearchquestionsorobjectivesSampleframeAcompletelistofallthecasesinthepopulationfromwhichyoursamplewillbedrawnSa

7、mplesizethenumberofcasesusedfortheresearchanalysisStatisticalinferenceaprobableconclusionaboutapopulationonthebasisofdataofsampleLawoflargenumberLargersamplesizecanbetterrepresentthepopulationthanSmallersamplesizeHowtochoosethesamplesize?Theconfidenceyouneedtohaveinyourd

8、ata:thelevelofcertaintythatthesamplecanrepresentthetotalpopulation(confidence↑samplesize↓)Themarginofer