MetaX,(,510650):Meta,,3:;Meta:Q-332:A:100828873(2001)01,220081204Meta,,2090,Meta[1],[2]MetaMetaMeta,Meta,(),,,,3:;22;t-F-31,,,,,111Glass=(Xe-Xc)/Sc(1)V=Nc+NeNcNe+22(Nc-1)(2),Xe,Xc,Sc,Ne,NcGlass[3],,X:(39899370);(K2951-B1-110);(980952);(003031):2001202212;:(1973-),,.201,220016EcologicScienceVol120No11,2Jun12001,,Meta,112HedgesgHedgesGlass[4],S,Hedgesgg=(Xe-Xc)/S(3)S=(Ne-1)(Se)2+(Nc-1)(Sc)2Ne+Nc-2(4)Vg=Nc+NeNcNe+g22(Nc+Ne-2)(5),Se,113HedgesdHedgesgGlass,,,Hedges1985J[5],Hedgesdd=(Xe-Xc)SJ(6)J=1-34(Nc+Ne-2)-1(7)Vd=Nc+NeNcNe+d22(Nc+Ne)(8)5,d114(responseratio,lnR)Hedges1999,(lnR)[6],:lnR=lnXe/Xc=lnXe-lnXc(9)VlnR=Se2NeXe2+Sc2NcXc2(10)lnRHedgesd,,4-+,0,;0,;0222,22(1),2820122ABA+BCDC+DNt=A+CNc=B+DN=A+B+C+DPt=A/Nt(11)Pc=B/Nc(12)PtPc,2211(ratedifference,RD)[7],:RD=Pt-Pc(13)VRD=Pt(1-Pt)Nt+Pc(1-Pc)Nc(14),PtPc,PtPc0,1,RD,015,RD,[8]212(relativerate,RR),RD[7]RR=Pt/Pc(15)VRR=(1-Pt)NtPt+(1-Pc)NcPc(16)RR=0,;0RR1,;RR1,3Meta,,,F-t-2,[9],2(r)1)tF2r=t2t2+dfr=FF+dfr=2Ne+Nc1);t-,df=Nc+Ne-2;F-,df=Nc+Ne-1Fisherz-[10]381,2:Metaz=12ln1+r1-r(17)Vz=1(Ne+Nc)-3(18)z-+,,,0,HedgesdlnR,Meta,,,CO2Meta,lnRd,lnR:[1],.Meta:[J].,1999,18(6):65-70.[2],.Meta[J].,1999,19(4):448-452.[3]GLASSGV.Primary,secondaryandmeta-analysisofresearch[J].EducationalResearcher,1976,5:3-8.[4]HEDGESLV.DistributiontheoryforGlass’sestimatorofeffectsizeandrelatedestimators[J].JEducationalStatistics,1981,6:107-128.[5]HEDGESLV,OLKINI.StatisticalmethodsforMeta-analysis[M].NewYork:AcademicPress,1985.[6]ROSENBERGMS,ADAMSDC,GUREVITCHJC.MetaWin:statisticalsoftwareforMeta-analysiswithresam2plingtests[M].Verion1.0.SinauerAssociates,SunderlandMassachusetts.1997.[7]NORMANDSLT.Meta-analysis:formulating,evaluating,combining,andreporting[J].StatisticsinMedicine,1999,18:321-359.[8]FRIEDMANH.Magnitudeofexperimentaleffectandatableforitsrapidestimation[J].PsychologicalBulletin,1968,70:245-251.[9]FLEISSJL.Measuresofeffectsizeforcategoricaldata[A].In:COOPERH,HEDGESLV,eds.Thehandbookofresearchdynthesis[M].NewYork:RussellSageFoundation,1994.125-260.[10]FISHERRA.Statisticalmethodsforresearchworders[M].2nded.London:OloverandBoyd,1932.IntroductiontoConnonlyUsedEffectSizesinMeta-analysisZHENGFeng-ying,PENGShao-lin(SouthChinaInstituteofBotany,AcademiaSinica,Guangzhou510650,China)Abstract:EffectsizeisthestatisticinquantitativeMeta-analysis,thecalculationofwhichdependsonprimarilythedataavailablefromtheoriginalstudies.Becausethedatafromoriginalstudiesarepre2sentedindifferentforms,methodsforcalculatingeffectsizesweredevelopedforthesedifferenttypesofprimarydata.Theseeffectsizessuitedtothreedifferenttypesofdatawereintroducedandcompared.Keywords:effectsize;Meta-analysis4820