Thanks a lot!!! This helped me a lot!! I was confused about how they model the constructed latent variable for path analysis. Your line that they consider the average of each indicator cleared all the doubts!!! Thanks Again!!
Thank you very much for the explanation. I still wonder how to deal with path analysis when there is a factor that acts as a second order. Should it be composited twice; that is, 1) composited as each item and then 2) these each items will be comopsited as one factor so that it can all represent as one factor in path analysis? Best regards, Peeraya
If you have a higher order construct, the first order constructs would be composite variables and they would have a relationship to the higher order construct. You would not want to combine multiple first order dimensions into one composite variable. That would be inappropriate. Hope that helps.
Absolutely. You can cite my book: Collier, J.E. (2020). Applied Structural Equation Modeling Using AMOS: Basic to Advanced Techniques (1st ed.). Routledge. doi.org/10.4324/9781003018414
I recently run analysis of my data. Hypotheses look good, r square looks good; but model fits are not good. Please what would you suggest for me to do in order to get my paper published
You most likely have a measurement issue with one of your unobservable constructs. That is what is causing the model fit issues..you can run modification indices which should help solve some of those issues.
Thank you so much for this very informative video. Would love to see a video on the interpretation to each SEM approach. 👍
Thanks a lot!!! This helped me a lot!! I was confused about how they model the constructed latent variable for path analysis. Your line that they consider the average of each indicator cleared all the doubts!!! Thanks Again!!
Glad it helped. Thanks
Great explanations! Thank you for the video.
Thank you very much for the explanation.
I still wonder how to deal with path analysis when there is a factor that acts as a second order.
Should it be composited twice; that is, 1) composited as each item and then 2) these each items will be comopsited as one factor so that it can all represent as one factor in path analysis?
Best regards,
Peeraya
If you have a higher order construct, the first order constructs would be composite variables and they would have a relationship to the higher order construct. You would not want to combine multiple first order dimensions into one composite variable. That would be inappropriate. Hope that helps.
@@joelcollier9387 Thank you very much for the answer. I am grateful for your help.😊
Thanks for the information. Do you have any reference for this preference of Full Structural Model over Path model? Thanks. Waiting for your reply..
Absolutely. You can cite my book: Collier, J.E. (2020). Applied Structural Equation Modeling Using AMOS: Basic to Advanced Techniques (1st ed.). Routledge. doi.org/10.4324/9781003018414
I recently run analysis of my data. Hypotheses look good, r square looks good; but model fits are not good. Please what would you suggest for me to do in order to get my paper published
You most likely have a measurement issue with one of your unobservable constructs. That is what is causing the model fit issues..you can run modification indices which should help solve some of those issues.