Data were analyzed on sentencing according to Pennsylvania guidelines: 120,300 cases from 1989 to 1992. The Pennsylvania data offer some of the richest information available in the country for analyzing judges' decisions regarding whether to imprison and the length of prison term. In addition to age, the independent variables used in the analysis were a combination of legally prescribed variables, offender characteristics, and contextual factors. The findings not only confirm the conventional wisdom that greater leniency will be extended to older offenders but also clarify the overall relationship between age and sentencing. The peak ages for receiving the harshest sentences were ages 21-25, followed by ages 26-29. Very young offenders -- those between ages 18 and 20 -- received sentences similar to those of offenders in their 30's and early 40's; offenders in their 50's and especially their 60's received the most lenient sentences. These findings are consistent with the "age bias" model, which emphasizes the relationship between judges' sentencing concerns (blameworthiness, protection of the community, and practicality or organizational demands) and age- linked attributions regarding the levels of dangerousness, propensity for crime, and ability to serve time in prison. The authors discuss whether or not age disparities in sentencing are warranted from a policy perspective. 3 tables, 1 figure, and 31 references