Multilevel model use level1 parameter... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Sihong Liu posted on Tuesday, February 27, 2018 - 10:02 am
I am running a multilevel mediation model.Level 2 is mother-child dyad (variable X, Y). For each dyad, we collected their M variable across 20 time-points (Level 1). The association between two variables in the level 1 (associations for 20 Mmother and Mchild for each dyad) is used as the mediator of X and Y at level 2. Below is the syntax I used for the analysis:

Cluster=ID;
WITHIN=Mp Mc ;
BETWEEN=X Y;


ANALYSIS:

TYPE=TWOLEVEL;
TYPE = RANDOM;
ALGORITHM=INTEGRATION;

MODEL:
%WITHIN%

SYC|Mp ON Mc;

%BETWEEN%
SYC on X;
Y ON SYC X;

However, Mplus keeps giving me warning "SERIOUS PROBLEM IN THE OPTIMIZATION WHEN COMPUTING THE POSTERIOR
DISTRIBUTION. CHANGE YOUR MODEL AND/OR STARTING VALUES.". And I find that the problem is on "Y ON SYC X;".

I wonder what should I do to fix this problem?

Thank you!
 Bengt O. Muthen posted on Tuesday, February 27, 2018 - 1:44 pm
Your random slope statement

%WITHIN%

SYC|Mp ON Mc;

does not make sense together with saying that Mp is a Within variable. There is a contradiction in that a Within variable has no between variation while the random slope says that its DV does have between variation. So delete Mp from the Within list. You may also want to use Define to group-mean center Mc as recommended in the literature.

Also request Tech8 and make sure the iterations look good in line with the FAQ on our website:

TECH8 – negative ABS changes
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: