Anderson and Gerbing (1988)'s method What is baseline model

[Daniela L]

Hello 

I am following the method in a published paper to perform SEM to examine a causal model built based on a theoretical framework developed by previous research.

I am confused about the method they used. The study in this paper followed  Anderson and Gerbing (1988)'s method. The baseline model first being constructed to confirm the stability of the latent variables; then they specified the causal model, which has all of the proposed paths whether they are significant or not (full model). Then the trimmed model is generated by eliminating the least non-significant path only the significant paths remaining in the model.

I understand "full model",  which seems like the proposed theoretical model, and the following trimmed model.  However, I don't understand the "baseline model" in this method. What it looks like? 

I found some studies performed CFA first to test the measurement model first, then performed SEM to test the structural model.  I am unsure which method to take. 

Thank you very much.

Thanks

[Terrence Jorgensen]

I found some studies performed CFA first to test the measurement model first, then performed SEM to test the structural model. 

That is what Anderson & Gerbing (1988) advocated.  The idea is that the latent structure is unconstrained in a CFA, analogous to the saturated "h1" model that simply estimates all covariances among observed variables.  Just as a model's chi-squared is a LRT that compares the theoretical "h0" model to the saturated h1 model, the 2-step approach tries to achieve the same by starting with a CFA (saturated latent spaces: freely estimate all covariances) and comparing an SEM with constraints on the latent structure using a LRT (chi-squared difference).

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

[Daniela L]

Thank you so much for the explanation. I attached our hypothesized model based on theories.

 SEM-Page-4.png

The code for my baseline model is as below, I covered all variables and paths in the baseline model.

baseline_model<-
  '
  # measurement 
  V =~v1 + v2 + v3 + v4
  C =~ c1 + c2 + c3 + c4
 
  # regressions
  H ~ age + sex
  V ~   H + age + gender
  C ~   H + age + gender + V
  R ~   H + age + gender + V + C + H  
  E ~   H + age + gender + V + C + H    
 
  #residual variance
  c1~~ c2
  c1~~ c3
  c1~~ c4

  c2 ~~ c3

  c2 ~~ c4

  c3 ~~ c4

  v1~~ v2

  v1~~ v3
  v1~~ v4

  v2 ~~ v3

  v2 ~~ v4

  v3 ~~ v4

  E~~R

  Age ~~ Gender

  '

  Would you mind checking the baseline model I drafted, I'd appreciate your help.   

Terrence Jorgensen <tjorge...@gmail.com> 于2021年2月22日周一 下午4:00写道：

--
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 on the web visit https://groups.google.com/d/msgid/lavaan/00320463-1dd5-40fa-bc72-e4a52894da82n%40googlegroups.com.