HI,
I am running the following mixed model (see syntax below). I want to plot
the estimates over age (time) to compare the three category groups over time
with the dependent variable. The equation I use follows the syntax.
I am finding that my graphical depiction is not as expected and that the
results between two categories seem to be reversed.
I am wondering whether my logic for my equation is muddled and lies in
particular with the category number (1,2 or 3). As background I have 3
categories namely 1,2 or 3. IN this model 3 is my null and results are 0.
However for category 1 or 2, I then multiply my estimate by the category
number assigned, ie. for category 1 by 1, and category 2 by 2. Obviously 1
has no impact. So the issue lies with category 2. If I divide by 2 my
results align as expected.
Does this make sense to anyone, and if so please enlighten me.
MIXED
Dependent BY Sex Category WITH Age Age2
/FIXED = Sex Age Age2 Category Age*Sex Age2*Sex Age*Category
Age2*Category | SSTYPE(3)
/METHOD = REML
/PRINT = SOLUTION
/REPEATED = Time | SUBJECT(STUDYNO) COVTYPE(UN).
Dependent=intercept + sex + ((age)(cellage)) + ((age2)(cellage)(cellage)) +
((Category)(category 1,2 or 3)) + ((sex*age)(cellage))
+((sex*age2)(cellage)(cellage)) + ((age*category)(cellage)(category number
1,2 or 3)) + ((age2*category)(cellage)(cellage)(category number 1,2 or 3)).
Nb. Cellage is the age in time that the equation will be calculated for. In
my case at 0, 6,12,18 ..
Regards,
Paola
"Ours has become a time-poor society, fatigued by non-physical demands and
trying to compartmentalize daily living tasks. It is small wonder that
physical activity is discarded in this environment" p126 (Steinbeck, 2001)
P Please consider the environment before printing this email.
=====================
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
Locked
