STATISTC 111 Chapter 5: Chapter 5 Notes
Document Summary
Explanatory variables response variablesand treatments ifthere is a casualrelationship it occurs in aparticulardirection. Explanatoryvariables versus responsevariables explanatoryvariable onethatattemptstoexplain or is purported to cause differences in a responsevariable or outcome variable goal is torelateone or moreexplanatoryvariables toeachresponse variable examinerelationships in whichthere is no conceivabledirectcasualconnection in eitherdirect. Interactingvariableandeffect modifiers effectmodifiermodifiestheeffectof the explanatory variable on theoutcome interactionoccurswhentherelationshipofone ofthemtotheresponsedepends ontheother one. Experimentalunits subjectsand volunteers experimentalunitssmallestbasicobjectstowhichwe can assigndifferenttreatments in a randomized experiment observationalunitsobjects or peoplemeasured in a study participants subjectcommonlyusedwhenthe observational units are people volunteersparticipants in studies. Reducingand controlling naturalvariabilityand systematic bias block design blockeach driver randomassignmentstotreatmentsisusedtoreduceunknownsystematicbiasesdueto confounding variablesthatmightotherwiseexistbetweentreatmentgroups matchedpairsrepeatedmeasuresandblocksareusedtoreduceknownsourcesofnaturalvariability inthe responsevariablesothatdifferencesdueto theexplanatoryvariablecanbedetected more easily. Potential complications confoundingvariables interactingvariables placebohawthorneandexperimentereffects lackofecologicalvalidity andconsequentgeneralization ecologicalvalidityvariableshavebeenremovedfromtheirnaturalsettingandaremeasured in the laboratory or in someotherartificial setting. 5 4 designing a good observational study casecontrolstudycaseswho have a particularattribute or conditionare compared with controls who do not retrospective participants areaskedto recallpastevents prospectiveparticipantsarefollowedintothe futureandevents arerecorded.