interpreting growth model fitmeasures with "MLR"

[gurkan cuvitoglu]

Hi, i m new to these topics. i m running a basic lgm with 3-4 occasions, unconditional, without missing, n=1055, non-normal data so i m running the "MLR" estimator. 

i m testing effect of sample size (50-100-150-250) combined with different occasions to model fit and statistical power. i m using lsay data. cohort2 math7-8-9-10.

Which one to use: chisq, chisq.scaled, baseline.chisq or baseline.scaled.chisq.

i am confused because of this:

lavaan:::print.fit.measures(fitMeasures(fit)) 

Model test baseline model:

  Minimum Function Test Statistic               32.016      23.554

  Degrees of freedom                                 3                3

  P-value                                                   0.000            0.000

here lavaan giving measures of baseline. 

But i was thinking to use chisq.scaled. which one i have to use?

i want to ask three more question:

1. i am going to report robust fit indices. is it what to be done?

2. fmin 

   0.003  is this important? how to interpret?

3. i have been reading for weeks about model fit indices. my model is simple as wrote. but every researcher has his/her favourite incices. 

Kenyy says CFI and TLI not suitable with LGM but Muthens, Bollen etc. using these. a

because my data is non-normal i m using mlr but McDonald suggesting to use mfi with small sample size.

so i think to report these:

CFI

Mc/MFI

RMSEA/ RMSEA CI and close fit

SRMR

nullRMSEA if <0.158

power using -> findRMSEApower(rmsea0, rmseaA , alpha, n, df ) (semTools in R)

and 

reliability(fit) for i, s and total (semTools)

Do these choices make sense? (non-normal and small sample size)

thanks already

[Terrence Jorgensen]

I don't see any questions specifically about lavaan. Have you posted your question on the general forum SEMNET?

http://www2.gsu.edu/~mkteer/semnet.html

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