Structural Equation Modeling

Structural Equation Modelling (SEM) is a statistical model used to test the causal relationships between latent variables (unobserved variables). There are two main parts in SEM: (1) Measurement model(s) and (2) Full model. SEM is widely used in supply chain research mainly to tested theories.

Theoretical Background

SEM has been developed from “Path Analysis” and “Factor Analysis”. Hence SEM can validates the model that include more than one set of causal relationships of the factors that can not directly measured.

Software for SEM
(Ranked by my experience)

1. R
   Open-access and free statistical programming available in PC, Mac and Linux.
   Now there are three main packages in R that can run SEM.
   Read more in “Structural Equation Modelling with R“.
2. MplusAdvanced but user-friendly software developed by one of the main SEM gurus Bengt O. Muthén  (UCLA) & Linda Muthén. Available in PC.Read more in “Structural Equation Modelling with Mplus”.
3. AMOS
   Graphically user-friendly software by SPSS. Only available in PC.
   Read more in “Structural Equation Modelling with AMOS.
4. LISREL
   The first software for running SEM developed by the legend Karl Gustav Jöreskog (Uppsala, Sweden) and Dag Sörbom. Only available in PC.
   Read more in “Structural  Equation Modelling with Lisrel”

My Research using SEM

1. Impacts of Supply Chain Collaboration
   My PhD thesis at Cardiff University UK.
2. Image of Thailand: as a Tourism Destination
   A collaborative work with Prof. Mingsarn Kaosa-ard (Chiang Mai University)
3. Performance Measurement of Supply Chain Emergency Response: A case of 2011 Thailand Floods
   A collaborative work with Dr.Thanawut Limpanitgul (Kasetsart University, Thailand)