I am struggling to find a way to run a factor analysis on my data in SAS. I have multiple constructs and several covariates that I would like to put into a structural model with an observed outcome that is zero-inflated. The outcome is a composite of three observed behaviors with success measured as a percentage of total occurrences for each. The composite variable ranges from 0 to 300 with half the values being zero. I cannot transform the data to achieve normality in this case do to the nature of the distribution.
I see nothing in the documentation for proc calis discussing link functions or zeromodels. Does anyone have a suggestion for how to approach this analysis?
Thank you.
