1

I am still a beginner in SEM though I know how to write the basic script for SEM in Lavaan R. Background on the data: It's a survey with 308 observations, 2 variables with a scale from 1 to 10, and 10 more with a scale from 1 to 5. I defined my model, then ran the cfa function, and then the summary function. I identified standardized factor loadings, but I don't know what to do next? From the research papers I've read, I realized I need to calculate t-values and find a "critical t-value". However, I don't know how to do that and in the summary table I only have z-values (my question might be stupid but I don't have a background in statistics) Second, dos a latent variable have an rsqaure? I ran the summary(model1.fit, rsqaure=TRUE) for the model but didn't get an R-square for only one of the 2 latent variables I have.

Thank you for your help!

1 Answers1

0

You can not perform CFA unless you are testing an established model. If you are trying for a new model, you have to perform the EFA at first to identify the factor structure and then the CFA with the identified structure. Assuming that you are testing an established model, in short, once you are done with identifying the factor loadings, you have to check the model fit indices such as CMIN/df,CFI, SRMR, RMSEA and PClose, and make the necessary manipulations in your model like identifying the covariating error terms. I suggest you to through CFA by James Gaskin. Though this is for AMOS, you can have a clear idea of what needs to be done for having a perfect model.

Mohanasundaram
  • 2,889
  • 1
  • 8
  • 18
  • Thank you very much for your help! I am indeed testing an established model. I have factor loadings above 0.3, and I have CFI and TLI >0.9 and SRMR and RMSEA <0.1 The only question left is: How can I assess the reliability of my hypotheses? Do I need to look at z-values and see if they are above the critical z-value at a confidence level of 95%? Thank you very much for your help! – cookiemonster1697 Apr 24 '20 at 09:04