SupportingStatementPartB722

B. Collection of Information Employing Statistical Methods

1. Potential Respondent Universe

The NCRP collects information on prisoners entering and leaving State prisons and persons discharged from parole supervision. Since 1999, the NCRP also collects data on offenders in custody of State prisons at year-end. There are 51 respondents in the NCRP data collection universe including the department of corrections in each of the 50 states plus the California Division of Juvenile Justice. Federal data are separately obtained outside of the NCRP through the BJS Federal Justice Statistics Program. Currently 39 states provide data on prison admissions and releases, 29 provide information on releases from parole supervision, and 27 provide information on the yearend stock prison population, but these numbers are expected to increase as additional states are invited to participate (see item B.3. “Methods to Maximize Response”, below).

2. Procedures for Information Collection

For the automated participants, the Census Bureau provides electronic media allowing these agencies to use existing extraction computer programs to download the year=s set of inmate administrative records. Since 2007 all respondents have provided data in electronic format. No manual submissions are expected in the future.

3. Methods to Maximize Response

Efforts have been made to make the NCRP materials clear and simple to use. Both BJS and Census Bureau staff are available to assist respondents with the data collection. BJS has also offered financial assistance to respondents to update their internal information systems to develop new NCRP data extraction programs. BJS and Census Bureau staff maintain frequent contact with data providers and data users in an effort to improve data collection, reporting procedures, data analysis, and data presentation.

BJS has also taken other efforts to maximize response in the NCRP. First, BJS has coordinated the two divisions within the Census Bureau that collect data for NCRP, the National Prisoner Statistics, Summary of Sentenced Population Movement (NPS-1, OMB # 1121-0078), the Annual Parole Survey and Annual Probation Survey (P&P, OMB # 1121-0064), the Capital Punishment Report of Inmates Under Sentence of Death (OMB # 1121-0030) and Deaths in Custody (OMB # 1121-0249) so that respondents who provide data for more than one BJS collection know what collections to expect during the year, the time frames for the collections, and the points of contact. BJS will continue to work with Census staff to enhance their understanding of the collections so that they are capable of addresses data providers’ questions across multiple surveys.

Second, BJS has been taking a more comprehensive approach to its correctional data collections, taking advantage of the fact that some respondents who provide summary prisoner data to other BJS data collections such as NPS-1 might also be able to provide individual level data to NCRP. BJS staff have used and will continue to use opportunities associated with follow-up contacts on one survey to attempt to enlist respondents to participate in NCRP. As the result of a contact with the NPS-1 respondent in the Arizona Department of Corrections in June, 2009, for example, BJS expects Arizona to contribute data to NCRP for 2009. In addition, the P&P respondent for parole for the state of Georgia has described his state’s parole tracking information system to BJS staff, a system that could be used to generate NCRP extracts. BJS will contact him about participating in the NCRP-1C during 2009. BJS has as its goal to achieve full participation by each of the 50 state departments of corrections and the California Division of Juvenile Justice in NCRP.

Third, BJS has and will continue to work with the Justice Research and Statistics Association (JRSA), an association that supports the Statistical Analysis Centers. The State Justice Statistics Program for Statistical Analysis Centers is designed to maintain and enhance each state's capacity to address criminal justice issues through collection and analysis of data. The program provides limited funds to coordinate statistical activities with the state, conduct research as need to estimate impacts of legislative and policy changes, and serve a liaison role in assisting BJS to gather data from respondent agencies within their state. JRSA is funding several projects from the state Statistical Analysis Centers (SACs) that look at parole revocations. BJS worked with JRSA to design the solicitation, and in the solicitation, SACs are asked to provide an overview of the statewide parole databases that they are using for their research. Two of the applicants currently do not provide parole data to NCRP. Pending funding of the projects, BJS will be involved with the projects and will involve the SACs in encouraging states to provide NCRP data, at least on parole.

Fourth, at business meetings of the JRSA, BJS has had discussions with SAC directors about NCRP, its utility to the field, and methods to expand it. The general idea that comparative data by state on the characteristics of the prison populations would be informative to the field has been expressed, and SAC directors have expressed both their interest in NCRP and a willingness to work with potential data providers in nonparticipating NCRP states to encourage them to participate in NCRP.

4. Test of Procedures or Methods

The procedures and methods utilized in the NCRP have been revised as necessary since inception in 1983, and have been codified in an NCRP Processing Specification document. The procedures include reading data in the format provided by each respondent, along with the preparation of initial “item tallies”. Following conversion of the data to NCRP format, a second set of unedited item tallies are prepared for purposes of comparison to the initial item tallies so as to detect and correct any anomalies which might have been introduced in the conversion process.

The process of converting state supplied data to a standard NCRP format includes the conversion of each state’s unique offense codes to standard 3-digit BJS offenses codes so as to provide for a common set of codes across jurisdictions. Where an analyst has previously matched a BJS offense code to a state offense code, a computer program assigns the same BJS code for the new data collection year. If a previous match is not found, as happens when new laws are enacted, a preliminary “crosswalk” to the most appropriate BJS offense code is assigned by the Census Bureau contact who works most closely with that particular state. The crosswalk is made based on statutes provided by the state, and includes a conversation with the state respondent, if necessary. Preliminary crosswalks of previously unmatched offense codes are sent to BJS for further review and confirmation before continuing with processing of each state’s data.

Following the conversion of state data to NCRP format, the Census Bureau performs range checks and consistency edits, and generates “edited item tallies” which are compared to the earlier sets of tallies, as well as to the previous year’s NCRP data. Finally, the edited item tallies are sent back to the state that provided the data to ensure that the NCRP format adequately reflects the information provided.

NCRP processing specifications also provide for the Census Bureau to generate a standard set of 66 tables for each new data collection year, based on criteria provided by BJS. These criteria are objectively specified by a set of “if statements” which define, for example, which NCRP codes are to be used to define a prison admission type of “parole violator’ or of a “new court commitment” to prison. For each data collection year, BJS reviews the data and makes minor adjustments, based on frequency distributions prepared using preliminary data files provided to BJS by the Census Bureau. Frequency distributions are separately prepared for selected variables from each of the four file types, including prison admissions (NCRP-1A), prison releases (NCRP-1B), parole releases (NCRP-1C) and yearend prisoner population (NCRP-1D) data. BJS makes adjustments, as necessary, to accommodate anomalies which sometimes occur in the data. For example, while the if statements are designed to only include persons sentenced to more than one year in prison in the data tables, BJS will nevertheless allow observations without a valid sentence length for a particular state to be included if the state is not one of the six states that have a combined jail-prison system, as jail inmates are generally defined to be persons sentenced to a year or less. Using the criteria provided by BJS, the Census Bureau prepares a preliminary set of 66 tables for BJS review. This review consists of a comparison to summary counts of prisoner admissions, prisoner releases, and the total prison population available for the same calendar year from the National Prisoner Statistics program (NPS-1, OMB # 1121-0078), and NCRP tables for the previous year for the purpose of detecting unanticipated anomalies in the data. Following further investigation of any unexpected variations in the data, changes may be made before a final set of 66 tables is produced. A subset of the 15 most frequently requested tables are posted on the BJS website, with the remainder available by request.

As mentioned above under item A. 16.”Plans for publication, BJS plans to undertake methodological research investigating alternative methods for using NCRP data to generate national-level estimates of characteristics of the prison population. While NCRP now includes more than 90 percent of all persons admitted or released from prisons annually and 80 percent of prisoners in custody at year-end, many persons may treat the tables as nationally representative, simply because no suitable alternative exists. The intended methodological studies are expected to more rigorously address how NCRP data might be adjusted to generate representative national-level estimates, along with an associated error term. Simultaneously, BJS intends to recruit full participation to NCRP by all states that provide data to NPS-1 (refer to item B.3., above, “Methods to maximize response”).

5. Contact Information

The Corrections Statistics Unit at BJS is responsible for the overall design and management of the activities related to the NCRP collection including: data collection; data elements, definitions, and counting rules; and data analysis and dissemination. BJS contacts for the NCRP include:

William J. Sabol, Ph.D.

Chief, Corrections Statistics Unit

Bureau of Justice Statistics

810 Seventh Street, N.W.

Washington, DC 20531

(202) 514-1062

Thomas P. Bonczar

Statistician and NCRP Project Manager

Corrections Statistics Unit

Bureau of Justice Statistics

810 Seventh Street, N.W.

Washington, DC 20531

(202) 616-3615