Almost correct.

DISP=(NEW,PASS,CATLG),
The results depend on the data set not the run.
If the data set currently exists, it would then depend on where it was being allocated. If the allocation was to the same dasd as the current one, you would get a jcl error. If it were being allocated to a different dasd, your job could run. If the step runs sucessfully, the data set disposition would be determined by subsequent steps with the default being delete. If the step abends, the data set would be cataloged unless there was a previously cataloged data set by that name in which case it would be kept and you would get a not catlg 2.