Leymon 14
affect of guidelines will be instantaneous for some portion of imprisonment rate the year following the
institution of guidelines, but not all individuals at any given year will have been sentenced under
guidelines and thus be “caught up” by its effect, as some of the rate will be individuals who are still in
prison under the old sentencing procedure. Over time, a greater proportion of those in prison will have
been “caught up” by sentencing guidelines.
8
This increase from the effect of guidelines is likely to be
slow at first, but have an increasing impact over time and then at some point plateau. At the point the
effects of guidelines level off, only a small portion of the imprisonment rate will be individuals’ who were
sentenced to long sentences under indeterminate sentencing. To address this issue, the variables for
presumptive and voluntary guidelines were implemented using a procedure to capture a logged growth
curve in imprisonment growth.
9
These variables serve to approximate the logarithmic growth curve
expected in the analysis.
The six dependent variables represent various rates of imprisonment. Each of these incarceration
rates can be broken down into three components. Any count of this type is a function of new individuals
entering the system since the last count, plus those who are still in the system from before the last count,
minus those who have left the system since last count, which produces the new count. Taking this into
consideration, it is clear that new commitments and parole violators returned should have a significant
positive impact on incarceration rates. Guidelines were created specifically to target sentencing and the
inclusion of controls for new commitments and parole violators returned to prison will help limit the
dependent variables to a measure of those who have stayed in the system minus those who have left.
Their inclusion in some of the models will serve to create a more stringent test of guidelines as a measure
of “time served” (creating a quasi-measure of time served).
10
8
Under the time frames of the analysis in this study it is highly unlikely that all of the rate of imprisonment will ever be
“explained” by sentencing guidelines as some individuals will have sentences that originated prior to 1974 and have
sentences longer than 30 years.
9
Each state that had implemented one type of the guidelines was coded as a 1 the first year following implementation and
each subsequent year was given a one integer increase (2, 3, 4…n, ext). These codes were then logged to create the
logarithmic growth variable. States with indeterminate sentencing were coded to zero.
10
The quasi-measure of time served is the rate of those who stay in prison minus those who have left. This measure does
not distinguish from “time served” for those under sentencing guidelines or indeterminate sentencing. The measure is
only a rate of all “time served” across all prisoners.
Convention