實驗設計之基本概念及實際應用

《實驗設計之基本概念及實際應用》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、2010/6/152010SummerSchoolforCognitiveNeuroscience,Taiwan實驗設計之基本概念及實際應用張智宏助理教授中央大學認知神經科學研究所遠流出版社12010/6/15DefinitionofanexperimentTypesofexperimentaldesignIndependentgroups▪Definition▪ThreatstointernalvalidityandwaystoovercomethemWithin‐subjects(repeated‐measures)▪Definition▪Threatstointe

2、rnalvalidityandwaystoovercomethemWhenwesayAcausesB,basicallywearesayingthat:AandBco‐varyChangesinAprecedeschangesinBThereisnootherexplanationforchangesinBexceptforchangesinA22010/6/15Test1:60Test2:9032010/6/15DefinitionDependentvariableIndependentvariableManipulation/randomization/se

3、lectionBehavioralmeasuresReactiontime,errorrate,d’,spatialerrorsPhysiologicalmeasuresHeartrate,bloodpressure,respirationrate,cortisollevelBOLDsignal,EEG,EPR,MEG42010/6/15EventTrialBlockConditionSessionTotesthypothesesaboutacausalrelationshipTodecidetheeffectofatreatment52010/6/15

4、Theextenttowhichtheobservedeffectscanbeattributedtothemanipulationsoftheindependentvariables.IndependentDependentvaaaberiablevaaaberiableConfoundingvariableSchedulingofobservationsThechoiceoftreatmentsandcomparisonsSelectionofobservationsandmeasuresSelectionofparticipantsThemannerofass

5、igningunitstotreatments62010/6/15IndependentggproupsdesignWithin‐subjectsdesignComplexdesignSubjectassignmentDifferentgroupofsubjectsareassignedtodifferentconditionsasdefinedbytheindependentvariable.BalancingconfoundingfactorsIntelligenceMotivationConscientiousnessGenderMajorinsch

6、ool72010/6/15Group1:60Group2:90RandomggproupsdesignMatchedgroupsdesignNaturalgroupsdesign82010/6/15AllrespectbeingcomparableamongggproupsbutonlythetreatmentreceiveddiffersComparablegroupformedbyrandomassignmentStage1Stage2Stage3RandomlyAssignedTreatment1ObservationGroup1RandomlyAssigned

7、Treatment2ObservationGroup292010/6/15AhiAchieveddbyanyprocedureththtatassignsunitstoconditionsbasedonlychanceEx.TossingcoinsordicesHasdifferentpurposefromrandomsamplingIntactggproupsareusedExtraneousvariablesarenotcontrolledSelectivesubjectlo