This is an obvious point, and it's something that I asked about at JPL. I don't remember exactly what the answer was now.

I think it goes back to the issue with the scanning platform on Voyager 2 which caused a lot of hassle when it got jammed. It was decided that rotating attitude using reaction wheels as the primary solution was better then doing it as a secondary solution with the scan platform stuck pointing in an unfavourable direction. Clearly the risk in having a scan platform is completely obviated if you don't have one.

As an engineer I would also throw in the possibility that the required pointing tolerances would be easier to achieve using large reaction wheels rather then small motors after having been flying in space for years and years. The higher the magnification and resolution required the tighter the pointing and moving scanning tolerance will be required as the spacecraft moves past the target objects.

All this is speculation and guesswork.