Dear List,
We have both the same Patients (PT) and Controls (CL) scanned at
Baseline (BL) and Follow-up (FU).
On basis of:
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0801&L=SPM&P=R30874
We have modeled different scanner sequence/software levels as
different groups, testing using average correlations.
Now we are facing these options:
1) model PTBL, PTFU and CLBL, CLFU together with sequence/software as groups.
This will give us two factors: Subject and Groups
We have considered the following settings:
a)
Factor Independent Variance
----------------------------------------------------------
Subject yes equal
Group yes unequal
b)
Factor Independent Variance
----------------------------------------------------------
Subject yes equal
Group no unequal
(This gives a warning: Warning: Matrix is close to singular or badly
scaled. Results may be inaccurate. )
Group could be set to 'no' because it is a within subject design
(repeated measures).
However: not all the groups are dependent due to different
sequence/software, requiring the setting to 'yes' in 1a).
2) Model PTBL, PTFU and CLBL, CLFU as being a 'Condition'. This would
give the settings:
Factor Independent Variance
---------------------------------------------------------
Subject yes equal
Group yes unequal
Condition no unequal
Which of the options would you recommend?
Thank you for your advice.
Kind regards,
Simone.
--
Dr. A.A.T. Simone Reinders, MSc PhD
King's College London
Institute of Psychiatry (IoP)
Box P063, De Crespigny Park
London SE5 8AF
United Kingdom
|