Thank you a lot for this wonderful video. I have two questions: In my thesis, the scale that I adopted is mainly from the literature, with a few items from my qualitative research. My question is: should I do an EFA or CFA? According to your video, since I already have items related to the latent variable, I need to perform only a CFA. Am I right? Also, if I opt for CFA, which option should I follow please: Option 1:performing only CB-SEM, to assess the measurement model (calculate - CB SEM algorithm), and the structural model (calculate - CB SEM Bootstrapping)using SmartPLS . Option 2: Do a CFA with SPSS by specifying the number of factors in data reduction (not AMOS), and then use PLS-SEM to assess the measurement model again (calculate - PLS SEM algorithm)and the structural model (calculate - PLS SEM Bootstrapping). Thank you.
The new version of SmartPLS can do CB-SEM, but much simpler than AMOS. The current limitation of SmartPLS CB-SEM is that it cannot do multigroup analysis.
Thanks for the video, James. I wonder how to conduct the analysis using a two-dimensional measure of socially desirable responding, such as the Balanced Inventory of Desirable Responding (BIDR-16; DOI: 10.1177/2158244015621113). Should I analyze each dimension separately, or would it be better to aggregate the items from both dimensions into a single factor? What would you recommend?
James, please I sincerely appreciate your efforts. I have a question, please. The literature is divided on the interaction of three variables, such as X, Y, and Z. The various perspectives are as follows: Some contend that the X-Z link is mediated by Y. According to others, X mediates the Y-Z path. Some contend that Y moderates the path X-Z. and some contend that X moderates the path Y-Z. Can I test each of these models on its own?
Sorry for the slow reply. Holiday weekend in the US :) If you mean three-way interaction, it is simply the dual moderation of two moderating variables on the path from X to Y. So, you have a main hypothesized effect (X-Y), then that path is moderated by both moderators.
I have a model with likert scale and categorical scale. The categorical scale as dependent varible. I'm confusing how to calculate the data. Could you please make the tutorial for this case 😢🙏
If your dependent variable is categorical with more than two categories, then you must create dummy variables for the outcome (or use a different method, like logistic regression or ANOVA). If your DV is binary, then you can include it as a single variable and just recognize that positive coefficients imply that positive values of the IV are more associated with the group coded with the higher values (e.g., 1 vs zero) and negative values are more associated with the lower coded group.
Thanks, great video. Would we now use that adjusted data file when running the structural model?
You make it look easy! Thanks.
Thank you a lot for this wonderful video. I have two questions:
In my thesis, the scale that I adopted is mainly from the literature, with a few items from my qualitative research. My question is: should I do an EFA or CFA? According to your video, since I already have items related to the latent variable, I need to perform only a CFA. Am I right?
Also, if I opt for CFA, which option should I follow please:
Option 1:performing only CB-SEM, to assess the measurement model (calculate - CB SEM algorithm), and the structural model (calculate - CB SEM Bootstrapping)using SmartPLS .
Option 2: Do a CFA with SPSS by specifying the number of factors in data reduction (not AMOS), and then use PLS-SEM to assess the measurement model again (calculate - PLS SEM algorithm)and the structural model (calculate - PLS SEM Bootstrapping).
Thank you.
Hi, thanks for this video. I have a question!
what are the capabilities of Smart PLS in compared to AMOS?
The new version of SmartPLS can do CB-SEM, but much simpler than AMOS. The current limitation of SmartPLS CB-SEM is that it cannot do multigroup analysis.
Thanks for the video, James. I wonder how to conduct the analysis using a two-dimensional measure of socially desirable responding, such as the Balanced Inventory of Desirable Responding (BIDR-16; DOI: 10.1177/2158244015621113). Should I analyze each dimension separately, or would it be better to aggregate the items from both dimensions into a single factor? What would you recommend?
I would separate it into its dimensions. If I recall correctly, those two dimensions are inversely correlated.
@@Gaskination Thank you!
James, please
I sincerely appreciate your efforts. I have a question, please.
The literature is divided on the interaction of three variables, such as X, Y, and Z. The various perspectives are as follows:
Some contend that the X-Z link is mediated by Y.
According to others, X mediates the Y-Z path.
Some contend that Y moderates the path X-Z.
and some contend that X moderates the path Y-Z.
Can I test each of these models on its own?
Sorry for the slow reply. Holiday weekend in the US :) If you mean three-way interaction, it is simply the dual moderation of two moderating variables on the path from X to Y. So, you have a main hypothesized effect (X-Y), then that path is moderated by both moderators.
I have a model with likert scale and categorical scale. The categorical scale as dependent varible. I'm confusing how to calculate the data. Could you please make the tutorial for this case 😢🙏
If your dependent variable is categorical with more than two categories, then you must create dummy variables for the outcome (or use a different method, like logistic regression or ANOVA). If your DV is binary, then you can include it as a single variable and just recognize that positive coefficients imply that positive values of the IV are more associated with the group coded with the higher values (e.g., 1 vs zero) and negative values are more associated with the lower coded group.
how to remove amos plugin?
From the AMOS software? You would have to navigate to the folder (it's a hidden folder) in app data, amos development.