I think it is fairly traditional (although certainly not universal) that standard error bars are used for samples (in accordance with central limit theorem). Standard error = standard deviation divided by square-root of sample size. As sample size increases, your "guess" for group means will shrink (i.e., larger samples tend to be "more representative" of an unknwn population than smaller samples).
Robert Herdegen wrote: > TIPSters: > > I'm working with a student on preparing graphs for his senior thesis. He > plans to include error bars with each of the points on his figures, but I'm > not sure whether there is a convention for *what* is to be displayed in > error bars: standard deviation? standard error? confidence intervals? The > values are all different, of course, and each shows something a little > different. Are there any standards for what one should show in error bars? > > Bob Herdegen > > ******************************************** > Robert T. Herdegen III > Elliott Professor of Psychology and Chairman > Department of Psychology > Hampden-Sydney College > Hampden-Sydney, VA 23943 > 434-223-6166 > [EMAIL PROTECTED] > ******************************************** > > --- > You are currently subscribed to tips as: [EMAIL PROTECTED] > To unsubscribe send a blank email to [EMAIL PROTECTED] -- |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Steven M. Specht, Ph.D. Associate Professor of Psychology Department of Psychology Utica College Utica, NY 13502 (315) 792-3171 "unanswered questions are less dangerous than unquestioned answers" --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED]
