Not sure quite what you're asking.  You have three days, and 7 time
points within each day;   by "individual time points" do you mean the
latter, or do you mean a "time point" as one out of 21 possible times?
(The latter question won't make sense if, as re-reading your post seems
to indicate, the different days are indicators of different treatments.)

Are we to assume that you found an interaction between "treatment" and
"time"?  If you did not, there would be no obvious point to asking
questions of each time point separately.

In any case, it might make better sense to model "time" as a covariate
(rather than as a categorical variable with seven values) and try
establishing a useful functional form for "folate concentration" in
terms of "time".  (The relationship might be linear in the range of your
observations, which would simplify your overall problem;  or it might
not, which would make things more interesting.)  Then you can ask about
heterogenity of regression slopes (which would show up as significant
interaction between covariate and group), and whether you appear to have
the same functional form in the several groups.

On 26 Feb 2003, Judith wrote (edited):

> I fed 16 Ss three different meals (one was a control) on three
> seperate days and measured their folate concentrations at 7 time
> points on each study day.  [In] a repeated measures anova in SPSS 10.0
> using treatment and time as my within subject variables[,] I can't
> seem to generate contrasts at individual time points or ["for"?]
> post hoc analysis.  [Is] there an optimal way to enter the data into
> SPSS to do these analyses?

 -----------------------------------------------------------------------
 Donald F. Burrill                                            [EMAIL PROTECTED]
 56 Sebbins Pond Drive, Bedford, NH 03110                 (603) 626-0816

.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to