mixed model interpretation

[P.van Groenestijn]

Dear List members,

I am using 5 models within the multi level analysis. For each model I
wrote down the U0j (see below).
I expected a decrease of this value, when adding more variables, but
this is not the case!

These are the results:/

/
/



Model



U0j

intercept (individuall-level variance)



0



Model 0 + random variance on contextual (country) level



1



0,044

Model 1 + variables individual level



2



0,049

Model 2 + interactions individual level



3



0,049

Model 2 + variables contextual level



4



0,039

Model 4 + interactions contextuael level



5



0,042

/
As you see, U0j decreases at model 2, when adding level 1 variables to
model 1.
Futhermore, there is also a increase in U0j, when  I add the cross-level
interactions to model 4.

I used the following to calculate model 1 and 2:

/Model 1./
/mixed clt_torture/
/  /random intercept | subject(country_yr) covtype(un)/
/  /print solution g testcov/
/  /method ml./
/
/
/*model 2./
/mixed clt_torture with zthreat zanx sex zage0 zhealth zincome zsoccap
zedulvl zgov1 zgov2 zgov3 zsoctrust active ztv znewspaper /
/   /fixed  zthreat zanx sex zage0 zhealth zincome zsoccap zedulvl zgov1
zgov2 zgov3 zsoctrust active ztv znewspaper /
/  /random intercept | subject(country_yr) covtype(un)/
/  /print solution g testcov/
/  /method ml./
/
/How can the increase of U0j be explained?

Thanks!
Pieter van Groenestijn
University of Nijmegen


--
RadboudUniversiteit Nijmegen
Faculteit Sociale Wetenschappen
Research Technische OndersteuningsGroep
Thomas van Aquinostraat 4.00.51
tel: [phone]
fax: [phone]
email: [hidden email]
hp: http://www.ru.nl/fsw/rtog/

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