Sylvain Cl�ment wrote:
If your data is in this format:
Subject Condition Measure1 Measure2 Measure3 Measure4
S1 C1 y y y
y
S2 C1 y y y
y
S3 C1 y y y
y
S4 C1 y y y
y
S5 C1 y y y
y
S6 C2 y y y
y
etc.
Then you probably have a mixed design with one fixed between factor and one
within factor (the four measures). In this design subjects would be random. (I
think this accords with b).
The only likely alternative I can think of is where experimental condition is
also within.
The random factor is the one which is being sampled from a larger population.
So subjects must be random (unless you are only interest in these specific subjects).
Thom
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
