Hi,
the model I want to calculate with SmartPLS is a second-order-construct with a latent variable consisting of four dimensions in the form of latent variables. The problem I'm facing is that for one of the four dimensions the sample size is 25% smaller than for all the other LVs in my model. The reason is that 25% of the interviewees had no experience with this dimension and therefore, couldn't assess it. How can I deal with this problem in SmartPLS?
Thanks in advance,
Friederike