Hi Benjamin,

Your model is very extranger. Why all constructs are formatives? I presume those will be reflexives.The constructs are measured of likert scales? If Yes, the directions of arrows in model should be inverses! Confime it, please.

Best regards

Dirceu

Hi Dirceu,

Thank you for responding to my question. 

1. I presumed that all my constructs are formative as i thought the variables/measurements will cause the increase/decrease of the construct. 

2. Yes my questionnaires are all likert scale (1 to 5). Sorry to ask my i know the reason why questionnaires are all reflective construct?

Hope to hear from you soon.

Best Regards,

Benjamin Chua

Hi Benjamin,

Yes, the likert scales have always reflexive constructs,, due the internal idea of the construction "of the construct": We have first the construct idea and in progress we created the variables that mesurement it. The observad variables are a reflexion of the construct. We expect that the variables in same construct have "strong" inter-correlation. See Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York; McGraw-Hill.

The fomative idea is different: we have a set of variables without inter-correlation which together form a concept.

Now, about the first question; "How can i combine more than 1 item-scale into one variable in SmartPLS?"

if i understood your question, you can use the, for exemple, the exploratory factor analysis in SPSS software. Select the variables which you wish creteated a "packet" and in the option "extraction" choose "numbers of factors to extract = 1 and in

"options" choose "save as variable" - "regression". In data bank will appear a new variable that correpond to packet generated. And you can use this new variable in a construct like a commun variable.

Please, see if i aided you!

Best Regards,

Dirceu

Dear Dirceu,

Thank you for your comprehensive reply. I have attached my new framework. With all the variables being reflective. Is it correct for me to do it this way?

After i did the factor analysis in SPSS and extracted the common variable, i can import this variable into SmartPLS to do my Structural Equation Modelling? 

Sorry to ask this, the common factor outcome is it like a combine variable to link to the construct?

Thanks Prof.

Best Regards,

Benjamin Chua