On Mon, 28 Oct 2002 15:23:08 -0800, "frank koegler" <[EMAIL PROTECTED]> wrote:
> Two small groups of unequal numbers of subjects eating different diets are > tested (n = 4, 6). > > Both groups have a hormone measurements taken five times throughout the day > on a day of fasting and a day of regular eating. > > I am requesting help on setting up repeated measures analysis with Systat. > Data is arranged in rectangular indexed format (columns: diet A/B, > fasted/fed, subject, time, hormone level). I'm not sure what you are describing. For simplicity, consider what you put into an analysis if you don't have hypotheses about different changes that matter during the day: [ score up the average hormone level for each day; these are labeled here as Afast, ... ] Grp Subj Afast Afed Bfast Bfed 2 groups with N=4, 6. 10 subjects total. Here are two within-factors, A/B and fast/fed. If you use 5 scores for each day instead of their averages, then you have a 3rd within-factor. (It seems to me that that one might be either obvious, or trivial; but I know nothing about it. It is easier to look at the overall analysis without it.) > > Systat has instructed me to the "repeated measures" options, however, I'm > clear on how to proceed. If you use a program set up for repeated measures, then I presume that it will give you the right SS for various tests. If you have to perform this as the multi-way ANOVA, then you do have to figure out, carefully, what terms need to be used as which error terms.. Hope this helps. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
