Saturated mediation model?

10 views
Skip to first unread message

Maria Eduarda Guimaraes

unread,
Feb 19, 2026, 5:42:16 AM (yesterday) Feb 19
to lavaan
Hello all, 
I've been trying to run a mediation model using lavaan. My variables are observables, so I use continuous scores for the model. 
it looks like this: 

model <- '
#a path
M ~ a*X

#b path
Y ~ b*M + cprime*X

#Indirect effect
ind := a*b

#total effect
total := cprime + ind

#direct effect 
dir := Y ~ c*X
'

however, the model fit parameters are so weird. 
example: 

Model Test Baseline Model: Test statistic 84.589 
Degrees of freedom 3 P-value 0.000 
User Model versus Baseline Model: 
Comparative Fit Index (CFI) 1.000 
Tucker-Lewis Index (TLI) 1.000 
Loglikelihood and Information Criteria:
 Loglikelihood user model (H0) -2666.671 
Loglikelihood unrestricted model (H1) -2666.671 
Akaike (AIC) 5343.342
Bayesian (BIC) 5363.448 
Sample-size adjusted Bayesian (SABIC) 5347.581 
Root Mean Square Error of Approximation: RMSEA 0.000
90 Percent confidence interval - lower 0.000 90 
Percent confidence interval - upper 0.000 P-value H_0: 
RMSEA <= 0.050 NA P-value H_0: 
RMSEA >= 0.080 NA


What that could mean? how would I fix it? I tried different estimators, I did use bootstrap, the variables are expressively correlated... 

Thank you very much in advance.

Edward Rigdon

unread,
Feb 19, 2026, 7:13:27 AM (23 hours ago) Feb 19
to lav...@googlegroups.com
There is nothing unusual (that I can see) in these results. Your
mediation model has 3 observed variables. The model is saturated, so
it produces a chi-square of 0 with 0 degrees of freedom. The
"baseline" or "null" model specifies the diagonal elements of the
covariance matrix free and the off-diagonal elements constrained to 0,
so it has 3 degrees of freedom. The fit indices all show no lack of
fit in your ("user") model. What you are seeing is what you should
see, in a situation where the user model is saturated.
> --
> You received this message because you are subscribed to the Google Groups "lavaan" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/lavaan/577dc517-2788-45d4-be7c-9b2b57b5d354n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages