Multi-level GM and distal outcomes PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Amber Fahey posted on Friday, March 09, 2018 - 9:01 am
I have successfully run a conditional model following the UG ex. 9.16. Now I would like to add both continuos and categorical distal outcomes; however, I cannot find any examples for this. I ran a test-run with the following input, including only one continuous distal outcome (SWLSTOT) and I received an output with no alarming warnings. Can you confirm this is the correct way to do this? And how would my input change if I wanted to include categorical distal outcomes? Or, would I have to dichotomize my categorical v.'s for it to work? Thank you!

USEVARIABLE = ID day2inpt AGE FIMTOTA EduYears PTA1 SWLSTOT cog day;

CLUSTER = ID;

WITHIN = DAY;

BETWEEN = day2inpt AGE FIMTOTA EduYears PTA1 SWLSTOT;

DEFINE: center day2inpt AGE FIMTOTA
EduYears PTA1 (grandmean);

ANALYSIS: TYPE = TWOLEVEL RANDOM;

MODEL: %WITHIN%

s | COG ON DAY;

%BETWEEN%

COG S ON day2inpt AGE FIMTOTA EduYears PTA1;

SWLSTOT ON COG S day2inpt AGE FIMTOTA EduYears PTA1;

COG WITH S;
 Bengt O. Muthen posted on Friday, March 09, 2018 - 1:01 pm
This input looks fine. A categorical distal is handled the same way, you just put it on the Categorical= list in the Variable command.
 Amber Fahey posted on Friday, March 09, 2018 - 1:12 pm
Thank you for the quick feedback. I had been told that my time-invariant categorical covaraites needed to be dichotomized. Is that true, or could I also have simply used the Categorical = command?
 Bengt O. Muthen posted on Friday, March 09, 2018 - 1:18 pm
Covariates don't need to be dichotomized unless they are nominal in which case you split them up into dummies. You don't use Categorical= for covariates, only for DVs.
 Amber Fahey posted on Thursday, April 05, 2018 - 9:41 am
Thank you again. I have one more related question. If I include a predictor/covariate variable that has been split into dummies, should I center those on their granmean, or does that not make sense, e.g., gender (0,1).
 Bengt O. Muthen posted on Thursday, April 05, 2018 - 2:29 pm
I would just use 0/1.
 Amber Fahey posted on Friday, April 06, 2018 - 1:26 pm
Thank you, Now that I have added all of them to the model, and specified some of them as categorical. I am getting this error: TYPE=TWOLEVEL requires specification for the CLUSTER option. I can't figure out why I was working before, and now it is not.

A sample of my input is:

VARIABLE:NAMES ARE ID Empfu Trans SRS SWLST FIMfu DRSPI PARSum day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1 olog day;

CATEGORICAL = Empfu Trans SRS SWLST

CLUSTER = ID;

WITHIN = DAY;

BETWEEN = Empfu Trans SRS SWLST FIMfu DRSPI PARSum day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1;

DEFINE: center day2inpt AGE FIMTOTA
TFCDays EduYears PTA1 (grandmean);

ANALYSIS: TYPE = TWOLEVEL RANDOM;

MODEL: %WITHIN%

s | OLOG ON DAY;


%BETWEEN%

OLOG S ON day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1;

Empfu ON OLOG S day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre
PTA1;
Trans ON OLOG S day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre
PTA1;

SRS ON OLOG S day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1;

SWLST ON OLOG S day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1;

FIMfu ON OLOG S day2inpt AGE FIMTOTA TFCDays EduYears Sex Racec Emppre PTA1;

OLOG WITH S;
 Bengt O. Muthen posted on Friday, April 06, 2018 - 2:27 pm
Note that you are missing a semi colon after this statement:

CATEGORICAL = Empfu Trans SRS SWLST

If this doesn't help, we need to see your full output showing the error - send to Support along with your license number.
 Amber Fahey posted on Saturday, April 07, 2018 - 11:03 am
Oh my goodness, I am so sorry for wasting your time with that!

I did it get it to run now; however, I am receiving the following error:
ERROR in MODEL command
Observed variable on the right-hand side of a between-level ON statement must be a BETWEEN variable. Problem with: OLOG

Olog is my intercept- This is a little strange because when I ran this model before with just one distal outcome, I did not receive this error at all, and I was able to regress the distal outcome on the intercept of Olog. Am I missing something? Why would I be getting this now and not with a more simple model? And is there some other way to regress my distal outcomes on the intercept?

And of course if I then include Olog on the between statement, I receive an error message saying a within variable cannot also be a between variable.

Thank you again!
 Bengt O. Muthen posted on Sunday, April 08, 2018 - 5:34 pm
Please send the output, your data set, and your license number to support@statmodel.com.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: