Three principal objectives were defined:
- Develop a simulation model that can support two sentencing models simultaneously. As a voluntary system, some judges may embrace the sentencing standards while others will continue sentencing under the legacy system.
- Develop a simulation model that accommodates Virginia-style judicial worksheet sentence recommendations where worksheet points and scores are translated into structured sentence recommendations.
- Identify the optimal mix of prison/non-prison recommendations, worksheet factor points, and sentencing ranges to guarantee that violent and sex offenders spend more time in prison, while also making the overall system bed-space neutral.
Microsimulation is designed to mimic the flow of offender populations over the course of a specified time frame. This is achieved by culling historical data and reviewing trends in the criminal justice system, while adjusting the underlying assumptions of the model. Microsimulation enables users to test "what if" scenarios by altering actual or proposed policy and practice changes that influence the path of individuals through the criminal justice system. The Alabama Sentencing Commission's decision to use a mircosimulation model to project correctional populations was based on the model's flexibility in incorporating anticipated changes; the Commission's access to accurate, detailed individual offender records; and the ability to incorporate core assumptions.
The development of the simulation model was undertaken in a three-stage process. The first stage involved the development of a baseline projection of current practices for later comparison with projections made following implementation of the sentencing standards. The second stage incorporated the initial sentencing standards into the simulation model; and the third stage integrated disparate modules together into a user friendly model interface. Banks, Carson, and Nelson (1996) recommend a specific algorithm to follow when building simulation models, which has become the de facto industry standard to design and build simulation models. This algorithm served as the outline for the development of the Alabama Sentencing Simulation Model. Refer to the Final Report beginning on page 27 for a discussion on each step in the design process.
Mode of Data Collection:
The following agency databases were selected to develop the model:
- The Administrative Office of Courts (AOC) maintains court records for all felony convictions including filing, disposition, and sentencing information;
- The Alabama Criminal Justice Information Center (CJIC) is the Alabama state agency responsible for gathering and providing critical information for law enforcement and the criminal justice community. The Commission specifically obtains arrest records of offenders in Alabama;
- The Alabama Department of Corrections (ADOC) maintains data on the felony offenders who are admitted to prison, actively serving a prison sentence, and released from prison;
- The Alabama Board of Pardons and Paroles maintains data on paroles, pardons, restoration of voting rights, pre-sentence, pre-probation, youthful offender, and other investigations and reports provided to the sentencing court;
- The Sentencing Commission relied on existing data maintained by various agency information systems, as well as initiating a number of data collection projects to fill gaps in Alabama's existing records system. This included collection and analysis of defendant pre-sentence investigation reports and surveys of county jails, community corrections programs, and drug court programs.
Description of Variables:
Dataset 1 (General Information Table: 35 variables and 189,570 cases) includes information on the inmate's race, education level, family status, military service, and employment status.
Dataset 2 (Inmate Table: 93 variables and 76,718 cases) includes information on the inmate's incarceration including admission type, total sentence length to serve, minimum release date, maximum release date, good time credit, and institutional placement.
Dataset 3 (Initial Sentence Table: 44 variables and 369,908 cases) includes a record for each sentence per incarceration per inmate. For each sentence record, this dataset includes the specific offense information for each incarceration, sentence length for each incarceration, county of conviction, and if this inmate is a habitual offender.
Dataset 4 (Transfer Leave Table: 31 variables and 991,266 cases) contains a record for every movement the inmate makes within the correctional system.
Dataset 5 (AOC Cohort Data: 128 variables and 74,691 cases) includes arrest date, filing date, indictment date, offense literal, offense classification, court action and action date, sentence date, begin date of the sentence, sentence imposed, probation imposed, and court ordered programs.
Dataset 6 (Arrest Records Data: 14 variables and 64,281 cases) includes offender sex, race, National Crime Information Center (NCIC) offense code at arrest and at conviction, arrest data, charged date, offense literal, and disposition code at conviction.
Dataset 7 (Pardons and Paroles Data: 36 variables and 232,183 cases) includes offender race and sex, youthful offender status, sex offender status, conviction offense, status code, prison sentence length, probation sentence length, supervision level, date of sentence and date of probation, date and totals of fees paid.
Dataset 8 (Personal Crime Sentencing Worksheet Data: 93 variables and 2,569 cases), Dataset 9 (Drug Crime Sentencing Worksheet Data: 105 variables and 1,956 cases), and Dataset 10 (Property Crime Sentencing Worksheet Data: 92 variables and 1,093 cases) include variables on the most serious offense, degree of the offender's participation, gang related activity, weapons used, number of victims and victims' injuries, property taken, drug types and amounts, number of prior convictions, prior probation sentences and revocations, number of parole revocations, number of prior incarcerations, offenders marital status, highest grade completed, employment and legal status, history of drug, alcohol, mental health or domestic violence problems, and history of treatment for drug, alcohol, mental health or domestic violence problems.
Presence of Common Scales:
Extent of Processing: ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of
disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major
statistical software formats as well as standard codebooks to accompany the data. In addition to
these procedures, ICPSR performed the following processing steps for this data collection:
Standardized missing values.
Checked for undocumented or out-of-range codes.