I am a Ph.D. candidate (Information Systems) and a beginner in statistics and SEM. I have several questions in regard to my SEM model and analysis for my dissertation study and would appreciate advice on that.
My SEM model:
------------------------------
G1-G5 (latent variables) => GOV (latent variable)
Each of G vars has one or more indicators (measured vars)
OS1-OS3 (latent variables) => SPON (latent variable)
Each of OS vars has one or more indicators (measured vars)
Main hypothesized paths:
GOV => FS (dependent variable (DV), latent)
SPON => FS
FS = FS1-FS5 (FS construct's components, latent variables)
Each of FS vars has one or more indicators (measured vars)
Three control variables (latent, but may simplify to measured)
Legend: GOV represents governance, SPON - organizational sponsorship, FS - FLOSS Success.
Context: free/libre open source software (FLOSS) development
------------------------------
Questions:
1. If leftmost part of my model (G => GOV, OS => SPON) represents MIMIC model (Schumacker & Lomax, 2010), how do I need to handle this?
2. Is there any special way to handle SEM models with multiple levels of latent variables, such as [measured -> latent => latent => latent (DV) = latent