I am investigating the impact of a factor on 4 Dependent variables (DVs). Two DVs are following normal distributions, the rest two aren't. One of them is "count" data with the majority being 0s and it follows a Poisson distribution. The other is a "time length" data measured by "days". I know that I can use Poisson regression( or zero-inflated, negative binomial) for univariate analysis on the two variables that do not follow normal distribution. However, I need to do the multivariate analysis first to control the type I error. My question is :" What SAS program that I can use for Multivariate analysis with DVs like these? Is there a generalized method for all types of distributions?"
Apologies that this is not SAS but you may want to look at an earlier response of mine
https://www.researchgate.net/post/Can_you_share_any_examples_of_the_use_of_multivariate_regression
and consider MLwiN's capabilities - see
Multivariate mixed response models
at
http://www.bristol.ac.uk/cmm/software/mlwin/mlwin-resources.html#multipleresp
- Badges
- Science method
- Similar topics
- Methodology
- Research Methodology
- Qualitative Research