Dear all,
How can I create a second order construct in SPSS?
I want to create a higher order construct (A) based on three lower constructs (B,C and D).The lower order ones are reflective and the higher order is formative.
My goal is to regress A on some dependent variables.
In this case, A = Mean(B*C*D) is not correct approach to construct A , is it?
Please advise me an appropriate way to construct the second order for both reflective-formative and reflective-reflective constructs (if there are any differences).
Anyway, I read this paper below and I understand how to do it on SmartPLS. But I just wondered how I can do this on SPSS.
Sarstedt et al. (2019)
How to Specify, Estimate, and Validate Higher-Order Constructs in PLS-SEM
Regards!
Ted
Ted your goal doesn't make any sense at all. You regress y on x not v.v. David Booth
@Ted How can a second order latent construct thas is made-up of three reflective first order constructs be a formative one? Moreover, SPSS does not work with latent variables and it has extention package known as AMOS for doing confirmatory factor analysis and SEM with first and higher order latent constructs.
Dear Ted Shimada you already have read the most authentic source in this regard so just follow Sarstedt et al. (2019) by using repeated indicators approach for both reflective-reflective and reflective-formative hierarchical constructs in PLS.
However, if you are required to use SPSS for some special hypotheses testing (such as moderated-mediation through PROCESS), you can extract Latent Variable Scores from PLS. Tq
Dear Long She
Thank you so much for you replies!!
I was asked to use SPSS if I could rather than using SmartPLS or AMOS. So, I wanted to know how I could do it.
I guess the way Muhammad Zia Aslam suggested should work.
Dear David Eugene Booth, you are right! Sorry for my typo and thank you for correcting it!
Dear Imran Anwar,
Thank you for your reply. I was referring to so-called Type2 (reflective-formative) higher order construct (Sarstedt et al., 2019) as per image below.
In any case, I have a feeling that SPSS may be not capable of handling a complex model based on second order constructs. I will consider using SmartPLS or AMOS.
I am agreed with Dr. Muhammad Zia Aslam and others answered well.
Best of Luck Ted Totsuka
Hello, i have the same problem. I need to use a second-order formative variable to make a mediation test on SPSS. how can we calculate an index for this variable?
Thank you so much for your help!
- Badges
- Science topic
- Similar topics
- Psychology