Obviously, which dyno to use is a matter of choice, and if your existing data is based on chassis it's an advantage to continue.
However: in my limited experience, a chassis dyno requires careful control of factors that are absent (and therefore neutral) on an engine dyno. These have a considerable effect on power delivery to the wheels, and are generally harmless and may be ignored in back-to-back tests of the same chassis with motor mods, but (IMO) not as closely monitored or regulated as would be needed to make effective test comparisons between different chassis. It can be done, but requires more work and record keeping.
These include such seemingly minor factors as driveshaft/U-joint angle, type and size of differential (Ford 9" has far more power loss than GM, 7.5" ring gear vs. 9-3/4"), viscosity of hypoid lube and transmission fluid, torque converter slip, ATF temperature, tire diameter and compound, tire pressure, etc.
How the engine gets air with the hood closed also varies somewhat between chassis based on hood clearance, underhood air temperature, fan discharge pattern, etc.
