Getting ANCOVA results with a regression

Hello !

My outcome variable is a change variable that can take positive and negative values. My ancova has 1 factor (with 3 categories) + 1 continous covariate + a categorical noise variable. Only my outcome variable has missing data.

I have to do a sensibility analysis comparing results from a complete case ANCOVA analysis and a regression analysis after multiple imputation of missing data. I'm using SAS for the analyses.

How do I compare my complete case analysis ANCOVA to regression results using my imputed datasets? The aims remains to compare 3 means so what I'm really asking is how to compare regression results to mean comparison ANCOVA results.


Fortran must die
Are you asking what software to do it in? The regression is going to give you slopes between your predictors and the DV. I don't think that is different than what ANCOVA generates although I have not worked with ANCOVA in many years.