Page 1 of 1

Reflective-Formative Validity

Posted: Sat Aug 17, 2019 8:09 pm
by cbakhan
I have developed reflective-formative construct. I have to check its validity. Initially, at first order, all the weights, factor loadings, AVE are appropriate. When I calculate the second-order by using the latent score and place the latent constructs on second-order formative. They come out to be insignificant. However, the factor loadings are significant for all the latent constructs of second order. Should I retain the insignificant latent second-order constructs? (Hair et al, 2016) suggests retaining the insignificant indicators. Any help will be appreciated.

Re: Reflective-Formative Validity

Posted: Wed Aug 21, 2019 4:08 pm
by cringle
Please take a look at this article which explains how to specify, estimate, and validate higher-order constructs in PLS-SEM: https://www.sciencedirect.com/science/a ... 8219301223