The Proliferation of Actuarial
Methods in Punishing and Policing
The academic status of parole-prediction models changed rapidly in the last decades of the twentieth century: at an increasing rate, more and more states began using risk-assessment instruments to determine whether inmates should be released on parole. The trend is visually dramatic and is well illustrated in figure 1.1.
By 2004, twenty-eight states were using a risk-assessment tool as part of their parole decision-making process. Illinois was the only state to use an actuarial instrument during the 1930s, '40s, and '50s, and Ohio experimented with them in the 1960s, but many other states turned to prediction tools after California and the federal government adopted them in the 1970s. Georgia, Iowa, Tennessee, South Carolina, Alabama, and Florida began using risk-assessment instruments in the late 1970s and early 1980s, and other states—including Connecticut, Michigan, New Jersey, Ohio, Pennsylvania, and Texas—followed suit. Marylynne Hunt-Dorta's 2004 survey and the time-line shown in figure 1.1—which are corroborated by other surveys conducted in 1939, 1961, and 19941—reflect an exponential increase in the use of actuarial methods at the turn of the twenty-first century.
What is especially remarkable about this exponential trend is that it coincided with the relative demise of parole. The number of states offering parole declined steadily from about 44 in 1979 to about 32 in 2003. Despite this decline, the number of states withparole that used an actuarial method increased from about 1 in 1979 to 23 in 2004.2In other words, of the 28 jurisdictions using prediction instruments in 2004, 5 no longer had parole for new convictions but used an actuarial method to determine