The incarceration of criminal offenders in the United States has reached epidemic proportions. One way to scale back the prison population is by using empirical risk assessment methods to apportion prison sentences based on the likelihood of the offender recidivating, so-called "evidence-based sentencing." This practice has been denounced by some legal scholars, who claim that the use of certain empirically relevant risk factors--including gender, age, and race--is plainly immoral. This study tested whether lay individuals share their sentiment. More than 600 participants weighted to be representative of the United States population were asked about the extent to which they would support imposing shorter sentences for old versus young offenders, female versus male offenders, and white versus black offenders, all else being equal. The results indicate that very few participants (<3%) had no settled opinion about using evidence-based sentencing, and approximately half were unequivocally opposed to the practice. Whereas more than 3-quarters of participants were against using race to determine prison sentences, almost half were open to the possibility of using gender and more than 3-quarters of the participants were open to the possibility of using age to determine prison sentences. Individual differences as a function of participants' own demographic characteristics, or of their belief in "just deserts" as the primary purpose of sentencing, or of their political outlook, were either inconsistently or meagerly related to these findings. The profoundly disparate views held by the general public regarding the use of specific risk factors do not bode well for the use of demographic risk factors in sentencing as a way to roll back mass incarceration.