Prof., thank you for this video. I would like to ask how to perform convergent validity for a higher-order construct (reflective-formative)? You mentioned in this presentation that a disjoint 2-stage approach was used, and the assessment of the measurement model will include weights and significance, collinearity using VIFs, and full-collinearity VIFs for discriminant validity. How about the convergent validity of HOC? Thank you, Prof.
If you mean convergent validity = redundancy analysis for higher-order HOC, it does not need to assess and report for formative HOC. Because, in HOC level, we do not have missing items and error term, so no need to test redundancy analysis for second-order formative.
@@s.mostafarasoolimanesh7567 Dear Mostafa, thank you for your great videos.....I have six independent variables (all single item) 3 are continuous and 3 are dummy (yes/no). My dependent variable is continuous. I cannot use spss as I think my residuals violate normality. Can I test this model with warppls please ? Ned Kock says in his blog that warppls handles dichotomous variables, and it seems to be the case for single item variables too using the "Robust Path Analysis" algorithm... Could you please confirm my understanding is correct and that I can use Warppls in this instance? Many thanks for your help. It is greatly appreciated!
Prof., thank you for this video. I would like to ask how to perform convergent validity for a higher-order construct (reflective-formative)? You mentioned in this presentation that a disjoint 2-stage approach was used, and the assessment of the measurement model will include weights and significance, collinearity using VIFs, and full-collinearity VIFs for discriminant validity. How about the convergent validity of HOC? Thank you, Prof.
If you mean convergent validity = redundancy analysis for higher-order HOC, it does not need to assess and report for formative HOC. Because, in HOC level, we do not have missing items and error term, so no need to test redundancy analysis for second-order formative.
@@s.mostafarasoolimanesh7567 Dear Mostafa, thank you for your great videos.....I have six independent variables (all single item) 3 are continuous and 3 are dummy (yes/no). My dependent variable is continuous. I cannot use spss as I think my residuals violate normality. Can I test this model with warppls please ? Ned Kock says in his blog that warppls handles dichotomous variables, and it seems to be the case for single item variables too using the "Robust Path Analysis" algorithm... Could you please confirm my understanding is correct and that I can use Warppls in this instance? Many thanks for your help. It is greatly appreciated!