Subjects for Random command in MIXED

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Subjects for Random command in MIXED

Pellauella
Hello people,

I have a question about how to define the subject for say a random intercept in MIXED but actually the same goes for GENLINMIXED.
Assume we have patients nested in centers and we want to the effect of genotype A vs B on a disease score.

Now I want to adjust for center by adding a random intercept. However although this seems perfectly sufficient to me, I found that when using the GUI SPSS suggests Center, as well as Center*Patient_ID as the subject for the random effect. I guess this is due to the nested design and defining the data structure with both patients and centers. I don't understand the difference for the two options, though. Conceptually a interaction doesn't make sense or does it? They produce different results whatsoever.

I guess I just dont get the concept here. So basically the question is: What is the difference between these two statements (conceptually):

/RANDOM=INTERCEPT | SUBJECT(Center_ID*mnppid)

and

/RANDOM=INTERCEPT | SUBJECT(Center_ID)

=====================
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
Reply | Threaded
Open this post in threaded view
|

Re: Subjects for Random command in MIXED

Maguin, Eugene
You have patients nested within centers. Each patient has a disease score and an A-B genotype code. You expect that the average disease score may (will) differ across centers, which implies that the ICC will be greater than .00. So, the subject is center_id.  To get an idea of how this stacks up, look at example 6, I think it is, in the syntax reference documentation for the mixed command.
Gene Maguin



-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Pachen
Sent: Wednesday, May 16, 2018 2:36 AM
To: [hidden email]
Subject: Subjects for Random command in MIXED

Hello people,

I have a question about how to define the subject for say a random intercept in MIXED but actually the same goes for GENLINMIXED.
Assume we have patients nested in centers and we want to the effect of genotype A vs B on a disease score.

Now I want to adjust for center by adding a random intercept. However although this seems perfectly sufficient to me, I found that when using the GUI SPSS suggests Center, as well as Center*Patient_ID as the subject for the random effect. I guess this is due to the nested design and defining the data structure with both patients and centers. I don't understand the difference for the two options, though. Conceptually a interaction doesn't make sense or does it? They produce different results whatsoever.

I guess I just dont get the concept here. So basically the question is: What is the difference between these two statements (conceptually):

/RANDOM=INTERCEPT | SUBJECT(Center_ID*mnppid)

and

/RANDOM=INTERCEPT | SUBJECT(Center_ID)

=====================
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

=====================
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