Posted by
San K on
May 12, 2012; 9:45am
URL: http://spssx-discussion.165.s1.nabble.com/Linear-Mixed-Model-Analysis-tp5703558p5706908.html
> each group. These are sequential measurements so values close to each
> other in the sequence are highly correlated.
That is ok. Only the error shouldn't be correlated.
> 1) Are the 5 variables statistically different from each other at any
> of the 20 measurement points (for the n=20 group)
> 2) For each of the 5 variables, do their values change significantly
> over the 20 measurement points (and where)
> 3) Are there any overall differences between the two groups for the 5
> variables (e.g. variable 4 is higher in group 1 and variable 5 is
> higher in group 2)
> 4) If there are differences between groups, where are they within the
> 20 measurement points?
Do you really need MIXED model for this?
> If anyone has advice on the appropriate input commands for this
> analysis, I would appreciate your help. I am only familiar with using
> the dropdown menus, but I hopefully I can click the correct buttons
> given the proper command lines. Thanks.
You can always paste the syntax. There are few commands for panel
modelling in SPSS if you have the advanced module installed.
Regards,
San K
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD