Thanks - that sounds reasonable. But how do I use the saveold option?
----- Original Message ----
From: Jeph Herrin <junk@spandrel.net>
To: statalist@hsphsun2.harvard.edu
Sent: Wednesday, March 11, 2009 2:35:06 PM
Subject: Re: st: hlm utility in Stata10
It is much faster, and I often switch to HLM for large datasets.
However, I usually find it easier to work directly with HLM than
through -hlm-. Even for bootstrapping a Stata dataset, I usually
use -saveold- and then use HLM command line calls from within Stata.
hth,
Jeph
Grigore Pop-Eleches wrote:
You're right that xtmixed does much of the same stuff but it runs
excruciatingly slow on my computer (since I am using a very large
dataset) and I had heard that hlm is significantly faster.
----- Original Message ---- From: Stas Kolenikov <skolenik@gmail.com>
To: statalist@hsphsun2.harvard.edu Sent: Wednesday, March 11, 2009
11:41:42 AM Subject: Re: st: hlm utility in Stata10
Have you tried -xtmixed-? My understanding is that it covers most of what you can imagine with hierarchical linear models, although the terminology might be different (biostatistics vs. social sciences).
On 3/11/09, Grigore Pop-Eleches <grigopop@yahoo.com> wrote:
Hello,
I was wondering if you'd be kind enough to post the following
question on Statalist.
"I have been trying to run Sean Reardon's hlm utility in Stata 10,
and even though I have successfully installed HLM 6.0 and the hlm
module, and set the path name, I keep getting an error message in
Stata saying "file HLM2MDM.STS not found" and in a separate window
another message that HLM does not recognize my version of Stata.
When running the mkmdm command, Stata creates a new .dta file, as
well as a .mdmt and a .mdmvars file but not a .mdm file.
Does anyone have an idea what could be going on? Does hlm only run
with Stata 8.2 but not later versions? Or am I doing something
wrong?
Thanks,
Grigore Pop-Eleches Politics Dep't - Princeton University