The California Department of Corrections (CDC) currently houses more inmates than any state corrections department in the country (Harrison and Karlberg, 2003). There are approximately 160,000 inmates in 33 institutions, 16 community corrections facilities, 41 camps, and 8 prisoner mother facilities. A key issue for correction officials and a range of stakeholders is how best to maintain order and safety in a cost-effective manner. Every year, approximately 15% of the inmates engage in some form of ”misconduct” that can range from failing to report for a work assignment to insubordination to possession of narcotics to an assault on a guard or fellow prisoner.

The California Department of Corrections (CDC) currently houses more inmates than any state corrections department in the country (Harrison and Karlberg, 2003). There are approximately 160,000 inmates in 33 institutions, 16 community corrections facilities, 41 camps, and 8 prisoner mother facilities. A key issue for correction officials and a range of stakeholders is how best to maintain order and safety in a cost-effective manner. Every year, approximately 15% of the inmates engage in some form of "misconduct" that can range from failing to report for a work assignment to insubordination to possession of narcotics to an assault on a guard or fellow prisoner.

In this paper, we attempt to forecast which prison inmates are likely to engage in very serious misconduct while incarcerated. Such misconduct would usually be a ma jor felony if committed outside of prison: drug trafficking, assault, rape, attempted murder and other crimes. The binary response variable is problematic because it is highly unbalanced. Using data from nearly 10,000 inmates held in facilities operated by the California Department of Corrections, we show that several popular classification procedures do no better than the marginal distribution unless the data are weighted in a fashion that compensates for the lack of balance. Then, random forests per- forms reasonably well, and better than CART or logistic regression. Although less than 3% of the inmates studied over 24 months were reported for very serious misconduct, we are able to correctly forecast such behavior about half the time.