效应量的意义及测定方法

(,250014):,d,r2pb,22,:;d;r2pb;2;2:B841.7:A:1003-5184(2003)02-0039-0611.1,,,,F,F,(thesizeofaneffect,SE)[1],CohenCohen,,,0.05,0.05,;,0.01,,,[2],,4,15,8,,135,,,,(1)1ttdfp1548.01.37280.193548.02.09680.0513548.04.112680.0001Chow(1988):,,,,,,,,,,2386PSYCHOLOGICALEXPLORATION20032©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.[3],t,t,,,,,1994,,,[4]1.2,Cohen(1988),,d(1)(point-biserialcorrelationcoefficient)r2pb:d=0.2();d=0.5();d=0.8()r2pb=0.010();r2pb=0.059();r2pb=0.138(),22;2,22,Cohen(1982,1988):6%,;6%16%,;16%,[5]2dCohen(EffectSize)(PercentileStanding)(PercentofNonoverlap)2.097.781.1%1.997.179.4%1.694.573.1%1.491.968.1%1.28862.2%1.08455.4%0.87947.4%0.77643.0%0.67338.2%0.56933.0%0.46627.4%0.36221.3%0.25814.7%0.1547.7%0.0500%2,0.879;47.4%(AnESof0.8indicatesthatthemeanofthetreatedgroupisatthe79thpercentileoftheuntreatedgroup.AnESof0.8indicatesanonoverlapof47.4%inthetwodistributions.),,,:r2pb2222.1ddt,d(1)d=…x-…xs(1)Rosenthal(1994)345,,0.70Cohen,0.700.5,042003©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.[6]2.2r2pbr2pb:r2pb=(t)2(t)2+df(2)2,,df=n1+n2-2,n1n2;df=n-1,n,n1=17,n2=15,,3312…x2320n1715s97.53,(t)=22.94,p0.05(2),r2pb=(t)2(t)2+df=2.9422.942+30=0.22,df=17+15-2=30,22%,5,;t,,t=2.88,p0.052,r2pb=(t)2(t)2+df=2.8822.882+4=0.67,df=5-1=4,0.67,:67%,2.322(eta-squared),(eta),,0.001.00,,2,r[7](coefficientofdetermination),,0.64,20.90,90%;64%,26%(0.900.64)[8]2,2,,2,,222.3.12t,,,44SourceSSdfMeanSquareFpEtaSquared22.500122.5009.0000.0170.52920.00082.50042.500943,2:2=SSbSSt(3)2=SSbSSt=22.542.5=0.52953%,2r2pb2.3.22(one-wayANOVA),,;142©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.=SSbSSt=529.4021170.235=0.4520.452,,,45%2.3.32(two-wayANOVA),:A;B;::2=SSSSt(4),(A)(B),,66(A)(B)SourceSSdfMeanSquareFp293.389193.38911.3580.0060.23117.444258.7227.1420.0090.28102.778251.3896.2500.0140.2598.667128.222412.278174,2A=SSASSt=117.45412.28=0.28,2B=SSBSSt=93.39412.28=0.23,2AB=SSABSSt=102.77412.28=0.25,28%;23%;25%,,,,,2.3.42(ranksumstest)75392759748018908600511051059035954605694196R1738,,5,7,,,6R=17;56Rexp=27.51727.57,Z=-2.19Z-2.196,2=0.53,:53%6Rexp=n1(N+1)2(5)Z=6R-6Rexp(n1)(n2)(N+1)12(6)2=Z2N-1(7)2.3.52-H(Kruskai-WallisHtest),,,8242003©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.[9]H4.94,8:2=HN-1(8)2=HN-1=4.9418-1=0.29,N=1829%2.3.622(Fried2man2test),7,99711232123321412351236123712368151992[10],2=8.869:2=2(N)(K)-1(9)2=2(N)(K)-1=8.8673-1=0.443,44.3%2.42,2,;2(omegasquared),,22,2,10:2=SSb-dfbMSWSSt+MSW(10)5,2=SSb-dfbMSWSSt+MSW=529.402-245.7741170.235+45.774=0.3720.37,,,37%Cohen,2=0.37=37%16%,33.1,,(powerofstatisticstest),,,,,3.2,30,,,,,,,,t,rpb,r2pb,d342©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.[1]ShaughnessyJJ,ZechmeisterEB.ResearchMethodsinPsychology[M].TheMcGraw-HillCompanies,Inc.1997.228-229.[2]CohenJA.PowerPrimer[M].PsychologyBulletin,1992.112,155-159.LeongFTL,AustinJT.ThePsychologyResearchHandbook[M].SagePublications,Inc.1996.210-219.[3]ChowSL.Significancetestoreffectsize[M].Psycholo2gyBulletin,1988.103,105-110.[4]HeimanGaryW.BasicStatisticsfortheBehavioralSci2ences[M].ThirdEdition,HoughtonMifflinCompany,2000.331-332.[5]CohenJ.Statisticalpowerandanalysisforthebehavioralsciences[M].Seconddition,Hillsdalee,NJ:Erlbaum,1988.[6]RosenthalR.Interpersonalexpectancyeffects:A30-yearperspective[J].CurrentDirectionsinPsychologicalScience,1994b,(5):127-134.[7]VockellEL.EducationalResearch[M].MacmillanPub2lishingCo.Inc.1983.210-213.[8]JaegerRichardM.Statistics[M].SecondEdition,SagePublications,Inc.1993.71.[9]SnodgrassJG,Levy-BergerG,HaydonM.HumanExperimentalPsychology[M].OxfordUniversityPress,1985.395.[10].[M].:,1993.423.SignificanceofStudyoftheSizeofEffectanditsTestingMethodsQuanChaolu(SchoolofEducationScience,ShandongTeachersUniversity,Jinan250014)Abstract:Inthepapertheauthorexplainsthemeaningsofthesizeofaneffectinpsychologicalresearchandthestandardtoevaluateit,especiallythemethodstotestsuchthesizeofaneffectasd,r2pb,2and2.Finallytheauthorputsforwardssomemeasurestoraisethesizeofeffectandmakesomecommentsonthestudyofthesizeofaneffect.Keywords:thesizeofaneffect;d;r2pb;2;2(25)AStudyontheEffectsoftheDefinitionoftheTextStructureonTextRevisionHuangJiehua1,MoLei2(1.ShantouUniversity,Shantou515063;2.SouthChinaNormalUniversity,Guangzhou510631)Abstract:Thisstudywasdesignedtoinvestigatetheeffectsofthedefinitionofthetextstructureontextrevision.Studentsingradetwofromseniormiddleschool,identifiedashigh-abilityandlow-abilitywriters,respectivelyrevisedtextswithorwithoutorganiza2tionalsignals.Theresultsindicatedtherewassignificantlyleveldifferenceonthequalityoftextrevision.Textwithorganizationalsig2nalscontributedtorevisersconstructingtheglobalmodelofthetextdurin