I have a question for multilevel mediation analysis(2-2-1)

[Jin-yong Kim]

Hi all,

I am trying to perform a multilevel mediation analysis for type 2-2-1 using ml_mediation command.
However, the problem occurs on the equation 2, testing M(level 2) on X(level 2).
It runs iteration 16,000 and Massage says like this "Iteration 16000: log restricted-likelihood = 5592.2807 (not concave), convergence not achieved" after all.

Did I do something wrong? Any suggestions? and/or possible solutions (if available)?
Thank you in advance.

Best regards,

Jin-yong Kim

Command and data scheme are below,

------------------------------------------------------------------------------------------------------------

<Command> ml_mediation, dv(y) mv(m) iv(x) l2id(SQ4) cv(C1,C2,C3,C4)

<data scheme>
- DV : y(level 1, numeric)
- IV : x(level 2, numeric)
- MV : m(level 2, numeric)
- CV : C1, C2, C3, C4(C1, C2 are level 1 variables, C3, C4 are level 2 variables, all numeric)
- Clustering variable : SQ4(24 groups)

STATA version : STATASE 12

----------------------------------------------------------------------------------------------------------------------------