Page 1 of 1
Effect size f2 and predictive relevance Q2 for complex model
Posted: Sun May 12, 2013 9:24 am
by hudapg
I need urgent help plz.
I have a complex model where B,C,D are directed to A and E,F,G,H are directed to C. I want to calculate the f2 & Q2 for the DV A. I know how to do the F2 & Q2 for B to A & D to A, but when I want to do C to A shall I remove C and its IVs or only C and connect its IVs to A??
Posted: Mon May 13, 2013 6:15 am
by Hengkov
Hello,
SmartPLS difficult computing f^2 and Q^2 for complex model. But in your case, if you already know f^2 B and D, you just compute R^2 A - (f^2 B + D) = f^2 C OK (It's just my trick). ;-)
Best Regards,
Hengky
Posted: Mon May 13, 2013 10:55 am
by hudapg
Dear Hengky:
Thanks for your kind reply.. Indeed I have sent an email to a professor in one of the universities around and who is an expert in PLS. He kindly advised that I should in my case delete C with its IVs because I need to know its impact, which originally is affected by its IVs..
I managed to do it well based on his advice.
Thanks anyway for your help.
Posted: Mon May 13, 2013 12:59 pm
by Hengkov
Hi,
I know must remove C, but I give you a simply and quickly approach.
You know, sum f^2 BCD = R^2 A. Effect size is partial variance and R-square is total variance. Now, you know f^2 B & D and R^2 A,
for compute f^2 C very simple: R^2 A - (f^2 B + D) done!
I just explain for your understand My suggest, but if you choice and follow the professor advice, up to you. The result will be same OK.
Best Regards,
Hengky
Re: Effect size f2 and predictive relevance Q2 for complex m
Posted: Sun Nov 30, 2014 11:43 am
by davetol
Please, I am a doctoral candidate. i am presently analysing my data. i have 1IV, 2mediators and 1DV and the are all second order constructs(reflective-reflective). What are those things that i need to report in my study starting from missing data? Thanks.