I'm doing some logistic regression analysis in EG (4.1), and was just wondering if anyone knew the difference between using a code module that calls PROC LOGISTIC and using the EG window-based method. I know the EG method just generates a PROC LOGISTIC call, but from a quick look at the help files it seems that there are a few options that aren't incorporated into the GUI. Is there a concise list of which options are only available if I write the code myself?
The EG GUI pretty much follows the 80/20 rule for the statistical procedures. You can do 80% with the point and click interface and you need to add code for the remaining 20%. A few of the very complex procedures (MIXED, for instance) have much less than the 80% functionality in EG.
The EG developers did put the hooks in so it is easy to add most of the functionality that is not in the point-n-click interface. From within the task, click "preview code" and then click 'insert code.'
Of course, I forgot about the code hooks. Of course, it still requires you to know or be able to learn all the bits of PROC LOGISTIC anyway, but that's not necessarily a bad thing (after all, it's good practice to know how your code is doing what it's doing, instead of just blindly hoping that it will do it right).