THE MATH BEHIND THE COST ESTIMATE - MODEL DEVELOPED WITH STATE CORRECTIONS DATA

To estimate the financial impact of Wisconsin's truth-in-sentencing
law, the Journal Sentinel used a mathematical model developed for this
project by David L. Munro, associate professor of information systems
at the University of Wisconsin-Whitewater.

As with any model, we began with data and made some assumptions. This
model draws on data from a Department of Corrections database with
more than 123,000 records from 1990 through June 30, 2004.

The assumptions begin with a baseline - that is, an estimate of how
much prison and parole time would have been served each year if truth
in sentencing had never been enacted. To create that baseline, we used
data on the offenders who were released from prison in 2000, the last
year in which all of the inmates released had been sentenced under the
old system.

We computed the average number of months they had served in prison -
with the remainder of their sentences served on parole - for each of
four broad classifications of felonies: sexual assault, drug crimes,
assault and property/other crimes. In essence, the baseline assumes
that sentencing and parole trends would remain similar to those that
affected offenders released in 2000.

From 2000 through the first half of 2004, the model uses data from the
Department of Corrections database on actual prison and extended
supervision terms given under truth in sentencing.

After that, because there is no way to predict whether crime will go
up or down from current levels, the model assumes that each year's
group of new offenders through 2025 will be identical to those in the
first half of 2004.

The term given to each offender under truth in sentencing was compared
with the average time that would have been served under the old
system, based on data from the offenders released in 2000. The
difference was calculated as an estimated number of additional days of
incarceration and extended supervision.

Finally, those days were multiplied by the Department of Corrections'
current daily cost of $76.90 per person for prison and $5.58 for
extended supervision, resulting in an estimated additional cost per
year attributable to truth in sentencing.

We evaluated only those offenders who were admitted to prison for new
crimes; violations of parole or extended supervision were excluded
both from our evaluations of past data and our projections of future
offenders.

Most offenders serve their entire term, but some will have their
sentences reduced, such as through an appeal or placement in an
alternative program such as boot camp. Our model assumes that future
felons will serve out their terms following the trends established in
the first 41/2 years of truth in sentencing.

We did not adjust for:

* The effect of inflation.

* Additional sentences that some inmates incur after arriving in
prison.

* Additional health costs of aging inmates.

* The fact that some inmates will die before completing their
sentences.

* A small percentage of inmates who were omitted because of data
quality problems.

The Department of Corrections produced a study in 2001 using some of
the same methods to estimate the cost of extra prison time that would
be served by a limited sample of truth-in-sentencing inmates. The
estimate for just those 752 inmates was $26.2 million over eight years.