Some of the concern could be that the prototypes might not reflect the behavior of real world devices. It sounds like the band being licensed itself would not overlap with GPS, but is close enough that imprecise tuning could spread out into those areas. So it’s possible that test/prototype devices could be built with high quality filters to keep the transmission narrow, then the real devices with more cost saving measures spread more. I would expect that the FCC could come back in if the spread is too big, but by then it’s pretty late.
My experience with testing and scientific inquiry is derived from understanding human behavior not the physics of rf propagation, however the basic principles are likely the same. Develop an idea, test it under ideal circumstances, and then generalize it to other settings and conditions.
The initial testing is likely done with up to date equipment in ideal settings, if it works then, then expand the program. How does it work under various weather conditions? What equipment works and what doesn't work? What happens if the system is loaded up with users? And probably many more questions and conditions that I can't conceive of.
The consequences of making a mistake need to be considered. If the system is going to just mess up a few garage door openers, its not all that critical. However, if it can interfere with essential gps functions, then that is a much more serious concern and deserves more scrutiny.
Lockheed-Martin has a large research and engineering facility down the road from me. A few years back they were working mobile radar units for the military. In their large parking lot there would often be a half dozen or so mobile radar units in operation. Friends who work there couldn't talk much about their work, but it was pretty clear the units were being tested. Conveniently, the facility is located beneath the final approach to the Syracuse airport.