On Fri, 25 Oct 2002, Allyson Rosen wrote: > You've understood me perfectly. I want the interaction and there will > be 6 values graphed. Now I understand why the multiplication by 2 > (which is unnecessary). > > To restate what you said, if I want an error bar (1 SE) for the > interaction it's simply: > > Sqrt(MSEinteraction/18) > > Correct?
Yes. That's the distance from the mean to one end of the error bar; the bar itself will be twice as long (for "mean +/- 1 SE"). > There will be six bars and each bar represents 18 subjects. > > Allyson Looks OK. I've been assuming that the interaction effect was significant, else you'd want to plot only the significant main effect(s), I should think; unless there were some point to illustrating the degree of lack of interaction (so to speak, and rather sloppily). (This might be the case, for instance, if you had expected to see an interaction, but it turned out NOT to be significant, possibly due to small sample sizes or possibly due to unexpectedly large error variance(s); and it were desired to depict what the interaction would look like (more or less, anyway) if only one had had more Ss or better-controlled variation. This would be fairly thin logical ice for skating on, but might sometimes be worth a conjecture or two in designing a subsequent experiment.) ----------------------------------------------------------------------- Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 [was: 184 Nashua Road, Bedford, NH 03110 (603) 471-7128] . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
