More than likely the model is just too complex. You should be running your model in stages anyway, so figure out which variable set entered is causing this. More than likely you have put too many parameters into the random coefficients section than can reasonably be estimated. In my experience, its often difficult to even fully theorize the benefit of all the different random coefficients that people try to estimate, so its completely possible that this isn't adding any useful information. Make sure you aren't including interaction terms in the random effects section, this are typically hard to justify estimating anyway.
It's rare that I find a correctly estimated model won't run do to memory, and I've run these analysis on very large data sets with very large numbers of variables. In nearly every case, I've found that the model failed to run (due to this same memory issue) not because my system was under speced, or SPSS inefficient, but because I was attempting to massively overestimate my model with little or no theoretical reason for doing so. This was also a major warning I received in the courses I took on MLM, and so I have a strong suspicion this is the issue.
I will say that I've had a few models fail to run in SPSS, but run in other programs. This isn't to say the final model was a good one, just one that I was able to run and have converge. In all of those cases, the extra random effects added nothing useful to the model and clouded interpretation greatly.
Matthew J Poes
Research Data Specialist
Center for Prevention Research and Development
University of Illinois
510 Devonshire Dr.
Champaign, IL 61820
Phone: 217-265-4576
email:
[hidden email]
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of DG
Sent: Tuesday, July 10, 2012 5:53 AM
To:
[hidden email]
Subject: Insufficient Memory when running a 3 level model.
Hi,
i'm attempting to run a 3 level model on a large number of cases, and it breaks down due an "insufficient memory", as shown in the message.
"Warnings
Insufficient memory to estimate the model parameters.
This command is not executed".
I've tried both ML and REML methods, but this persists. Does anyone have any deas? is a way to reallocate memory when running SPSS that offers a partial work around?
many thanks,
--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Insufficient-Memory-when-running-a-3-level-model-tp5714102.htmlSent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
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
Locked
