Previous research has consistently revealed prior record to be a significant factor in the decision to incarcerate. However, little attention has focused on the multidimensional nature of prior record and how the importance given to these dimensions by judges vary in relation to the current offense type and sentencing jurisdiction. The present analysis noted that prior record includes dimensions related to risk and stakes as measured by prior offense type, aggregate prior events, and prior disposition. The study used logistic regression to model the incarceration decision on a random sample of 1,073 felony cases sentenced in 1990 in a nonguideline State operating under a modified desert rationale. The data compared urban, suburban, and rural counties. Results revealed that the effects of prior offense type, aggregate prior events, prior disposition, and other variables depended on the current offense type and sentencing jurisdiction. Prior offense type and prior record had a greater effect on nonviolent offenders than on violent offenders. The prior record dimensions did not significantly affect the probability of incarceration for violent offenders. Prior serious offending significantly increases the risk of incarceration in urban settings, but not in suburban and rural counties. The analysis concluded that prior record has several dimensions and that these dimensions and jurisdictional effects influence the probability of incarceration. Tables, notes, and 50 references (Author abstract modified)