The study does not indicate whether the crimes are repeated by the same person. IOW if you had a 100 black males and 16 assaults were committed by black males, it would be significant if those assaults were carried out by only 6 black males.
Good point-the statistics stored in the database do not allow that kind of information to be extracted (for obvious reasons)...but it does point directly to exactly what you said.
If only 2-3% of the population at most is committing nearly half the violent crime, then by necessity, it HAS to be the same people committing crimes over and over again.
That's a good point.