From what I have seen in the thermogravimetric analysis of polyolefins there is very little residual material. This is due to it's chemical structure [(CH2CHR)n] not forming large secondary products that would not be volatile (i.e., solid residue instead of evolved gases).
Now, consider carbon nanotubes, much like carbon fibers they leave residual carbon when heat treated in inert environments. My speculation is that increasing the CNT concentration would increase the char yield of the nano composite in a linear fashion.
This stems from my belief that the polyolefins would immediately form volatiles after the polymer undergoes chain scission. This would elute gases from the system prior to the formation of the higher molecular weight residual char.
If you compare CF composites with various matrix materials, the effect of matrix polymer will be prevalent. If, however, compare a narrow range of matrix materials and change, e.g. the sizing of the CF or the coupling agent, there may be differences. It cannot be granted, however, you have to prove by independent measurements that there is a correlation between the two.