Bin, I report from Watson et al. (1990 - p. 403)(1) following your notation:
Ho: mE = mC Ha: mE <> mC t=(mE - XmC)/SQRT[(s2E/nE)+(s2C/nC)] The confidence interval is defined as [-t(alpha/2),v;+t(alpha/2),v] where v=A/B, and A=[(s2E/nE)+(s2C/nC)]2 B=[(s2E/nE)2/(nE-1)]+[(s2C/nC)2/(nC-1)] HTH, Luca Mr. Luca MEYER Consumer research advisor http://lucameyer.com/ (1) Watson, C.J, Billingsley, P., Croft, D.J. & Huntsberger, D.V. (1990) "Statistics for Management and Economics" 4th ed. Allyn and Bacon > -----Messaggio originale----- > Da: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Per conto di Bin > Inviato: luned� 13 ottobre 2003 6.09 > A: [EMAIL PROTECTED] > Oggetto: [edstat] help for t-Test of Difference of Means with > different sapmle sizes > > > I found the following formula > ************ > Let E be the experimental condition and let C be the control > condition. Let m be the means, s the standard deviations, and > n be the sample size. Then t = (mE - mC) / SQRT[(s2E + s2C) / n ] > for the case of equal sample sizes, n, in both groups. See Levin > (1999: 25) for the formula when sample sizes differ. > ************ > in my case, the sample sizes is different, But I can not find > the formula. Could you please tell me the formula for t-Test > of Difference of Means with different sapmle sizes? > > Thanks. > . > . ================================================================= > Instructions for joining and leaving this list, remarks about > the problem of INAPPROPRIATE MESSAGES, and archives are available at: > . http://jse.stat.ncsu.edu/ . > ================================================================= > . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
