I suspect I don't fully understand something in the question, because most SEM software does not require that you interpret matrices. If you do have to interpret matrices, then the interpretation depends on the matrix representation that you are using. re you using McArdle McDonald (RAM matrix specification?) or LISREL or some other matrix specification?

To answer what to report:

At a minimum:

1. Discuss how you arrived at the model. Did you create it from prior research and find that it fits well, or did you run SEM repeatedly with minor modifications to achieve fit.

2. Mention your software program and estimation method

3. Give all loadings and covariance terms estimated by the model

4. Give some fit statistics: CFI, RMSEA, SRMR (chi-square maybe also)

Other good things to report (that are usually not reported and therefore arguably optional in most journals):

5. Assumption testing that was done (normality)

6. Handling of missing data

Hi, I'm not sure if I get your question right. If you want to report the results of the measurement model (...) look at CFA-studies (e.g. www.researchgate.net/publication/243962411_Parental_psychological_control_dimensions_Connections_with_Russian_preschoolers%27_physical_and_relational_aggression/file/5046351d1cdff33310.pdf).

Furthermore, there are some recommenations what to report in CFA-Studies (e.g. people.cehd.tamu.edu/~okwok/epsy651R/Articles/Jackson2009.pdf; or the chapter of Boomsma, Hoyle and Panter (2012; The Structural euqation Modeling research report) in the Handbook of structural equation modeling.

Christoph

Article Parental psychological control dimensions: Connections with ...

I suspect I don't fully understand something in the question, because most SEM software does not require that you interpret matrices. If you do have to interpret matrices, then the interpretation depends on the matrix representation that you are using. re you using McArdle McDonald (RAM matrix specification?) or LISREL or some other matrix specification?

To answer what to report:

At a minimum:

1. Discuss how you arrived at the model. Did you create it from prior research and find that it fits well, or did you run SEM repeatedly with minor modifications to achieve fit.

2. Mention your software program and estimation method

3. Give all loadings and covariance terms estimated by the model

4. Give some fit statistics: CFI, RMSEA, SRMR (chi-square maybe also)

Other good things to report (that are usually not reported and therefore arguably optional in most journals):

5. Assumption testing that was done (normality)

6. Handling of missing data

Hi,

depending on which stat program you used - Barbara Byrnes SEM books should be helpfull to you.