NA st dev of some parameters in multi-level SEM

[Anyu Liu]

Dear All,

Any one can help to fix the following issue? 

When I run the following code with the attached data set, 

model<-'

level: 1

AW=~A1+A2+A3+A4+A5+A6+A7+A8

NW=~N1+N2+N3+N4+N5+N6+N7+N8+N9

GW=~G1+G2+G3+G4

NW~AW+1

GW~NW+1

level:2

AB=~A1+A2+A3+A5+A7+A8

NB=~N1+N2+N3+N5+N6+N8+N9

GB=~G1+G2+G3+G4

NB~AB+1

GB~NB+B+1

B~Q+1'

fit <- sem(model = model, data = aa, cluster = "Hotel")

if I use the default estimation method, I got the following warning

Warning message:

In lavaan::lavaan(model = model, data = aa, cluster = "Hotel", model.type = "sem",  :

  lavaan WARNING: the optimizer warns that a solution has NOT been found!

And all the Std.Errs are zero.

If I swapped to EM estimation

fit <- sem(model = model, data = aa, cluster = "Hotel",verbose = TRUE, optim.method = "em", em.iter.max = 200000),

The warnings are as follows

Warning messages:

1: In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats,  :

  lavaan WARNING:

    The variance-covariance matrix of the estimated parameters (vcov)

    does not appear to be positive definite! The smallest eigenvalue

    (= -9.831438e+05) is smaller than zero. This may be a symptom that

    the model is not identified.

2: In lav_object_post_check(object) :

  lavaan WARNING: some estimated ov variances are negative

3: In lavaan::lavaan(model = model, data = aa, cluster = "Hotel", verbose = TRUE,  :

  lavaan WARNING: not all elements of the gradient are (near) zero;

                  the optimizer may not have found a local solution;

                  use lavInspect(fit, "optim.gradient") to investigate

And the Std Err of some intercepts in the 2nd level are NA.

Would you please help to explain what are the reasons and how to fix this? Thank you very mcuh

Regards

Anyu

[Terrence Jorgensen]

And the Std Err of some intercepts in the 2nd level are NA.

Why are you estimating intercepts of latent factors?  Even with single-level models, they are not identified unless you fix an indicator's intercept to zero.  I would just let the factor (NW, NB, GW, and GB) intercepts remain fixed to zero for identification, which is lavaan's default behavior.

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen

[Anyu Liu]

Hi Terrence,

Thanks for your reply. After I removed the estimation of intercepts, I got the same warnings. There are four NAs in the Std Errors in the 2nd level.

Would you please have a further look at it? The code is as follows

model<-'

level: 1

AW=~A1+A2+A3+A4+A5+A6+A7+A8

NW=~N1+N2+N3+N4+N5+N6+N7+N8+N9

GW=~G1+G2+G3+G4

NW~AW

GW~NW

level:2

AB=~A1+A2+A3+A4+A5+A6+A7+A8

NB=~N1+N2+N3+N4+N5+N6+N7+N8+N9

GB=~G1+G2+G3+G4

NB~AB

GB~NB+B

B~Q

'

fit <- sem(model = model, data = aa, cluster = "Hotel",verbose = TRUE, optim.method = "em", em.iter.max = 200000)

Thanks.

Regards

Anyu

[Anyu Liu]

Anyone could help on this? Thanks!