The analytic technique was chosen to overcome the limitations of the techniques used in previous research on escalation and specialization. The research used data from the Michigan Felony Offenders Study published by Blumstein and Cohen in 1988. It included all 5,507 cases with five or more arrests. It also used the 10 offense classifications used by Blumstein and Cohen. The analysis split the sample by race; participants included 2,781 white offenders and 2,726 black offenders. Results revealed a much higher degree of criminal offense specialization than had been indicated using previous techniques. For all pairs of offenses except homicide-rape, offenders were much more likely to repeat the same offense than to switch offenses. Results also revealed similarities and differences in black and white offending patterns than cannot be observed with the other methods. The research demonstrated how the models used can address questions that the other approaches cannot. Tables and 32 references (Author summary modified)