Jim,
I may be quite off in right field on this but I'm not sure how much you gain
from a multilevel analysis with your data. Not counting aggregate level 1
terms, you seem to have only one level 2 predictor: survey year. More
important, you have only 15 level 2 data points which is generally regarded
as two few. Joop Hox among others have written on this; however, I don't
have specific citations to offer. One thing you should do is check the ICCs.
What is the argument that persons in one survey year are more alike to each
other than they are to persons on another survey year?
I am thinking that your data could be analyzed as a an ordinary regression
problem where survey year is another predictor (or a set of contrast terms
for survey year).
Gene Maguin
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
jmdpulido
Sent: Tuesday, June 14, 2011 9:19 AM
To:
[hidden email]
Subject: Different Surveys and Multilevel Models (aka mixed models)
Dear all...
I have a series of surveys about the same topic. It is not a panel
(longitudinal study) as the individuals change from survey to survey. I have
about 15 surveys (from 1995 to 2010).
In order to analyse how the individuals' characteristics affect the result
on my dependent variable and to analyse some "contextual time effects".
I am thinking in running a multilevel regression (aka, mixed model),
defining as level 1 the individuals and as level 2 the survey (year) each
individual belongs to.
Then, I am thinking in analysing how level 2 residuals (after controlling
for all relevant individual characteristics) change over time (and if I can
explain this level-2 residuals using contextual effects or a time trend).
Nevertheless, as my data is not a real panel but a typical case of
multilevell (each level-1 individual belongs to only one level-2 unit (the
survey)), I am not sure if this approach is right.
What do you think?
--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Different-Surveys-and-Multilevel-Models-aka-mixed-models-tp4487803p4487803.html
Sent 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
