I work in developmental psychology and I am currently trying to run a path model that includes both cognitive assessments as well as observational data. The 4 latent variables that are made up of standardized cognitive tasks have outer loadings that would be expected, but the other 3 latent variables that are based on the observational coding of children's behaviour are coming out strange. Every item on the coding system is loading as .999 or 1.000 across all 3 of the latent variables. However, when I enter these items into a principal components analysis in SPSS, the loadings are of varying sizes and the 3 subscales are evident.
I have set up the model to be that the 4 standardized task latent variables are the exogenous predictors of the 3 endogenous, observational latent variables.
As I'm new to this PLS modeling, I am not sure what this result might be trying to tell me. I believe that I am somehow inputting my data incorrectly, but I cannot figure out why. If you have any idea why I might be getting such odd loadings, I would be happy to hear suggestions. Thanks!