Recidivism risk is usually completed by a forensic psychologist using one of many recidivism tools, a knowledge of research on the subject, and a comprehensive review of all available information on the individual in question (e.g., clinical interview, behavioral observations, record review, collateral contacts, possibly psychological testing, etc.). Recidivism tools consider a number of factors, almost all of which (depending on the test) are historical in nature. Many of the factors relate to offense characteristics. Sex offense recidivism measures particularly focus on offense characteristics, general offending history, and response to supervision.
You make a great point about bias, @Julianna Kirschner. The research on that subject is quite interesting. One could perhaps counter that, if the individual was returning to an overpoliced area following release from incarceration, the risk of arrest for recidivism would remain accurately high, justifying the use of the measure. Someone might argue then that the measure could simply be accurately reflecting bias which is already inherent to society rather than being biased itself. The research the measure was based on was, in large part, on those individuals in over-policed areas and therefore accurately reflects their risk. That being said, research also shows evaluators can be biased in their ratings on recidivism measures even outside of that, so you are quite right about the risks of bias.
Recidivism risk is usually completed by a forensic psychologist using one of many recidivism tools, a knowledge of research on the subject, and a comprehensive review of all available information on the individual in question (e.g., clinical interview, behavioral observations, record review, collateral contacts, possibly psychological testing, etc.). Recidivism tools consider a number of factors, almost all of which (depending on the test) are historical in nature. Many of the factors relate to offense characteristics. Sex offense recidivism measures particularly focus on offense characteristics, general offending history, and response to supervision.
You make a great point about bias, @Julianna Kirschner. The research on that subject is quite interesting. One could perhaps counter that, if the individual was returning to an overpoliced area following release from incarceration, the risk of arrest for recidivism would remain accurately high, justifying the use of the measure. Someone might argue then that the measure could simply be accurately reflecting bias which is already inherent to society rather than being biased itself. The research the measure was based on was, in large part, on those individuals in over-policed areas and therefore accurately reflects their risk. That being said, research also shows evaluators can be biased in their ratings on recidivism measures even outside of that, so you are quite right about the risks of bias.