I am currently working on developing a predictive model for childhood developmental outcomes. I am using plssem to develop this predictive model and encountered two issues. The first one is that I have an observed outcome variable as a mediator of the latent outcome variable and confused how to fit this as the command looks like only for latent outcome variables. The second most important problem I have faced is that I could not assess the fitness of the structural model (I can't produce indices which help to assess fitness of a structural model such as f2 , Q2 , R2) and any statistics that I can use to assess the out of sample prediction of my predictive model. I am wondering and waiting your suggestion on these issues.