Memo to OMB on Pre-Test of Jail Administrators

U.S. Department of Justice

Office of Justice Programs

Bureau of Justice Statistics

Washington, D.C. 20531

MEMORANDUM TO: Jennifer Park

Office of Statistical Policy and Planning

Office of Management and Budget

THROUGH: Melody Braswell, Justice Management Division

Jeri M. Mulrow, Acting Director, Bureau of Justice Statistics (BJS)

Gerard F. Ramker, Deputy Director, BJS

Elizabeth Ann Carson, Acting Corrections Unit Chief, BJS

FROM: Todd D. Minton, Statistician, BJS

SUBJECT: BJS request for OMB clearance to pre-test an online survey to jail administrators to assess facility capabilities to provide administrative data in support of the design and testing of the Survey of Inmates in Local Jails (SILJ), through the generic clearance agreement granted to BJS (OMB Number 1121-0339)

DATE: October 26, 2017

Introduction

The Bureau of Justice Statistics (BJS) is requesting a generic clearance to pre-test a one-time Survey of Jail Administrative Records (SJAR), an online survey to assess the feasibility of collecting and then using jail administrative data to support the next Survey of Inmates in Local Jails (SILJ). The SILJ was last fielded in 2002 and BJS is in the process of redesigning the survey with the intention of collecting data in 2020 or 2021. The SJAR will ask about administrative data currently available from local jail administrators, and the willingness and ability of jail administrators to provide these data in support of the SILJ program. The study will further support BJS’s goal of using administrative data to assess the reliability of some measures provided by self-report and reduce overall burden by exploring whether survey questions could be replaced with data from administrative records. The SJAR will also support BJS’s efforts to utilize administrative records as one way to expand and enhance its data collection portfolio and its statistical reporting practices recommended by the Committee on National Statistics of the National Academies in its report to BJS titled Ensuring the Quality, Credibility, and Relevance of U.S. Justice Statistics.¹ Lastly, this study supports the Federal Government’s efforts to leverage administrative data for research and evaluation purposes.²

The 2002 SILJ asked confined persons to self-report on ten different categories of information. Three of the ten sections consisted of questions related to an inmate’s current charges and any criminal justice status they may have been on at the time of confinement (97 questions), as well as questions related to pre-trial release and trial (17 questions) and their current sentence (65 questions). As has been explored with other national surveys (e.g., The American Community Survey, the Medical Expenditure Panel Survey) and in social policy research (e.g. The Reemployment and Eligibility Assessment Initiative, Sector-Focused Career Centers) administrative data may be used to replace or improve data quality, as well as reduce the burden associated with collecting this data.^3,4,5,6,7,8 Similarly, BJS’s 2016 Survey of Prison Inmates (SPI) relies on using administrative records to supplement inmate self-report data as a way to expand the breadth and depth of substantive issues that can be addressed through the study and to conduct future studies of prisoners.

The primary objective of the SJAR is to explore the availability of administrative data and the feasibility of using it to strengthen the survey by improving data quality and reducing the burden associated with collection for the next iteration of the SILJ. A secondary objective, which would also support efforts to improve estimation procedures, is to confirm access to information on each person confined to jail during the data collection period. While both sets of information are integral to informing potential methodological changes to SILJ, as reflected below, these issues have not been fully explored at the national level. Given the exploratory nature of the SJAR, a pre-test is necessary to collect feedback from a sample of jail administrators to ensure the questionnaire is designed to achieve the goals of the SJAR prior to full implementation. Prior to making a decision on whether some SILJ questions can be removed and obtained through administrative records, BJS plans to request OMB clearance to conduct a separate pretest of jail inmates to gauge their willingness to consent to record linkage.

Jails are responsible for confining defendants as they move through the adjudication process, from booking through arraignment to sentencing and, in some cases, incarceration after sentencing. Adequate monitoring of this movement requires that most jails maintain an inmate population management process that records the admission, processing, and release of inmates, which is one of the standards promoted by the American Correctional Association.⁹ While the over 3,000 jails across the United States use different systems to implement this standard, there are certain variables these systems are likely to maintain to support the basic functions of a jail, that is, to track inmate admissions and releases so that an accurate count is available at any point in time, as well as to track the movement of each person through the criminal justice process to ensure constitutional rights are not being violated. Regardless of the systems used, jails should be able to provide information on the identity of who has been admitted to the jail and why, who is still in the jail and why, and who has been released from the jail and why. Less clear are the specific data elements that are collected and/or maintained by jails.

BJS recognizes that having access to data and sharing data are two different things. BJS anticipates that some jails’ information management systems will not be linked to other components of the criminal justice system, (for example courts) and therefore, may not have access to some data, while others may be integrated with or linked to information through a common web portal that would support data sharing. However, because data may have different owners, each with their own policies, processes, and data sharing agreements, the extent to which available data may be accessed and shared may be limited.¹⁰ Even if the data can technically be made available, there may be legal or other restrictions that impede administrators’ ability to provide the data. Conducting a feasibility study to assess the availability of administrative records as well as the privacy concerns around accessing the data will allow us to fully understand the data systems and applicable privacy rules and regulations, and is consistent with best practice for researchers planning to use administrative data. ¹¹

Additionally, while much of the arrest and conviction data of interest is in the public domain, the provision of some data, data pertaining to unconvicted populations that may be part of the SILJ sample, for example, may be more restricted, particularly to noncriminal justice agencies and private entities.^12,13 Therefore, we must also measure the ability and willingness of jail administrators to provide the data to BJS in electronic or paper format.

Although there are third party jail data aggregators, e.g., APPRISS, that may be able to provide some of these data for some of the participating jails for a fee, BJS wants to explore receiving data directly from facilities and access the data quality to support the future SILJ.¹⁴

BJS has evidence that inmate-level administrative data is collected and may be made available. In 1999, Beck conducted a national assessment of reporting capabilities and developed an inventory of data elements for State and Federal Corrections Systems and found that all 50 State departments of corrections, some of which include jails, maintain data related to describing, committing, managing, and supervising offenders, however the exact variables and each department’s capacity to report on and share those data varied. The format in which the data resided varied as well, with some departments maintaining electronic records for some variables and paper for others.¹⁵ Beginning in 1999, administrators of the Arrestee Drug Abuse Monitoring (ADAM) survey were able to successfully obtain administrative data that included inmate demographic information, offense information, and booking date and time from thirty-five county jails that were participating program.¹⁶ More recently, the NYC Department of Corrections (DOC) and the Vera Institute were able to use administrative data maintained by the NYC jails, such as admission age, current charge, number of prior DOC admissions, and recent DOC admissions, to develop a risk assessment tool for recidivism among jail inmates.¹⁷

Researchers in Connecticut were able to combine data from Connecticut Departments of Mental Health, Social Services, Public Safety, Corrections, and the Division of Court Supported Services to examine costs associated with criminal justice involvement. In the first study, the Department of Public Safety provided data on arrests for individuals who were later convicted of an offense, including dates of arrest, statutory charges, and offense class, while the Connecticut Department of Correction provided data on incarceration days and parole days. The Court Support Services Division of the Judicial Branch provided data on probation days.¹⁸ For the second study, criminal justice records included arrests, incarceration, parole, probation, jail diversion program participation, evaluations for competency to stand trial, and records for individuals found not guilty by reason of insanity.¹⁹ In Florida, researchers matched and merged administrative data from the mental health and criminal justice systems of Miami-Dade County and Pinellas County to study the effects of treatment participation and medication adherence on criminal justice costs by looking at arrest and incarceration days for over 1,300 individuals.²⁰

These studies demonstrate that many jails, particularly large facilities, have some capacity to provide administrative data. The capacity of smaller facilities remains unknown. This study will allow us to update our understanding of data that can be provided and the extent to which it corresponds with information collected through SILJ.

Feedback obtained from participants at the National Jail Research Network meeting in June 2016 expressed a willingness on behalf of jail administrators to provide administrative data as a means to improve the quality of the data and reduce respondent burden. Conversations with individual members supported the notion that a request to provide administrative data related to offense, pretrial activities, and sentencing information would not be a challenge, especially for larger agencies that are more likely to have information management systems that are linked to or, at least retain, court information. On the other hand, participants felt that a request for personal identifiers is likely to be more of a challenge because of restrictions around sharing personally identifiable information, but worth exploring. They also emphasized that there is no standardization among jail data systems and that, even if BJS obtains the records, BJS should anticipate a range in terms of quality and format. For these reasons, it is important that BJS assesses feasibility across a range of jails that may be more or less capable or willing to provide administrative data.

Request for Pre-test and Follow-up Interviews

For the pre-test, BJS will administer the SJAR to 40 local jails to measure the feasibility of providing administrative data and the burden to complete the survey. BJS will follow-up with interviews with all survey respondents to assess how respondents interpreted the survey questions and instructions, as well as their understanding of the meaning of the survey questions and the time it took for them to complete the survey. These jails are located across the U.S. and will be sampled based on jail size.

The pre-test and follow-up interviews will-

• Identify technical or other difficulties jail administrators had when trying to complete the SJAR.

• Confirm the SJAR reflects the full range of potential challenges jail administrators may face when trying to provide data in conjunction with participation in the SILJ.

• Inform revisions to the SJAR to improve question order, terminology, and response options.

• Inform estimates on response time and anticipated response.

Respondent Universe and Sample Design

The Census of Jails has been conducted ten times by BJS since 1970, collecting aggregate data on jail capacity and jail inmate populations from all jails that hold inmates 72 hours or more, which includes approximately 3,000 jails. In 2013, the Census of Jails was collected jointly with the Deaths in Custody Reporting Program (DCRP) and will be used as the sampling frame for the 40 facilities in the pre-test as reflected in table 1.

We anticipate that the types of administrative data available will vary based on the size of the facility, so we will stratify the sample based on the facilities’ average daily population, splitting the jails into three strata by size, and select an equal number of jails from each stratum. While we understand that smaller jails may have more difficulty providing administrative data than larger jails due to limited funds available for technology, the larger jails represent a greater proportion of the 2002 SILJ sample, so it is as important to confirm larger jails’ ability and willingness to provide administrative data.

Table 1. Respondent Universe for the SJAR Pre-test
Jail Size (Average Daily Population)	Number of Responding Jails in the 2013 Census of Jails	Number Sampled for Pre-test
1-40	935	13
41-200	1,156	13
201 or more	833	14
Total	2,924	40

To increase awareness among jail administrators of the forthcoming pre-test, BJS and Abt plan to work with the American Jail Association (AJA) to market the study. BJS will utilize AJA’s communication outlets (e.g. AJAlert weekly e-newsletter newsletter, article in the American Jails magazine) to help implement the SJAR pre-test. BJS also plans to seek a letter of support from the AJA Executive Director. This letter will be included in our outreach to the selected jails.

Information to be collected

Instrument. The survey questionnaire is provided as Attachment A and selected screenshots of the online version of this instrument are shown in Attachment B. The survey begins with an introduction, and then is organized into three substantive sections:

Section 1. Data Elements, asks about specific data elements from six different categories of administrative data including individual inmate characteristics, current commitment, criminal justice status, current offense, pretrial and trial information, and sentencing information.

Section 2. Obtaining Administrative Data, asks about process and procedures for obtaining administrative as well as costs associated with providing the data.

Section 3. Survey Administration, asks about specific aspects of SILJ survey administration to inform planning for the full administration, including the roster and contact information for discussing these items.

In preparation for the pre-test, BJS and Abt Associates consulted the following experts.

• Robert Kasabian, Executive Director, American Jails Association

• Wayne Dicky, Jail Administrator, Brazos County Office of Sheriff, Bryan, TX.

• Laura van der Lugt, Director of Research & Innovation, Suffolk County Sheriff's Department, Boston, MA.

• Bob Balkema, Lieutenant, Kirkland Police Department, Kirkland, WA.

This consultation served to gauge respondent burden, question language and ordering, and the functionality of the web-based survey instrument. BJS revised the online instrument in response to their findings.

Interview. The semi-structured interview protocol is provided as Attachment C. The protocol begins with an introduction, and then is organized to review the three substantive sections of the SJAR as well as general questions related to providing their responses.

Design of the Feasibility Study

Pre-test Design

The survey will be sent to the administrative contact for the jails, which is typically the administrator or sheriff of the jail. He or she may opt to delegate all or some of the items to colleagues within the agency who may be in a better position to provide the information requested. The study’s invitation letter, Attachment D, explains how respondents may share the survey with others in their agency.

Pre-test Data Collection Procedures and Respondent Follow-up Interview Plan

The survey will be self-administered on the web. The survey questionnaire (Attachment A) has been programmed into Fluid Surveys software, a data collection tool that automates the skip patterns built into the instrument and collects categorical responses in addition to longer text responses. The frequent use of structured response categories throughout the instrument will minimize response time, and the software will be programmed so that respondents can skip questions that are not applicable based on prior responses. The software also allows respondents to complete the survey in multiple sessions. Additionally, respondents have the discretion to delegate portions of the survey to others in their agencies by sharing the agency-specific link to the survey, allowing multiple respondents to input their individual contributions before submitting a final response for the agency.

Abt Associates will field the pre-test (questionnaire and follow-up interviews) using the following recruitment strategy with periodic reminders to ensure survey completion. Abt has successfully used this strategy in other similar projects to achieve high response rates.

• To increase awareness among jail administrators of the forthcoming pre-test, BJS and Abt Associates staff will work with staff from the American Jail Association to market the feasibility study. Additionally, BJS will submit an announcement for the SJAR in the American Jails Magazine.

• To solicit participation for this pre-test, Abt Associates staff will conduct outreach by mail using an initial hard-copy invitation letter (Attachments D) customized to each respondent from Jeri Mulrow, BJS’s Acting Director. This letter describes the purpose of the SJAR pre-test, gives a summary of the information to be collected through the online survey and follow-up that will be conducted through the interview. The invitation letter also specifies the estimated time to complete the survey and participate in the interview, noting that participation in both is voluntary, but emphasizing the importance of complete participation to benefit the field.

• An initial email invitation (Attachment E) containing much of the same information will be sent to respondents within one week of the hard-copy letter to ensure that the letters have been received. The email will facilitate easy access by including a clickable link to the online survey.

• Email reminders to non-respondents (Attachment F) will begin two weeks following the email invitation. These emails will be sent approximately weekly until the survey is completed or the respondent actively declines participation by notifying the project team. Each email will include a clickable link to the online survey along with the respondent’s unique username for logging in to the survey.

• Telephone calls to non-respondents will begin five weeks after the email invitation, using the script provided as Attachment G. Similar follow-up calls will be made approximately weekly until the survey is completed or the respondent actively declines participation. In the course of these telephone calls, project staff will provide assistance with the survey instrument as needed. For example, respondents may wish to receive a PDF copy of the instrument for reference as they complete the online survey. In rare cases and at the request of the respondent, project staff may administer the survey by telephone and record the responses in Fluid Survey and conduct the interview at the same time.

• Upon completion of the survey, respondents will receive a phone call from Abt Associates staff requesting participation in an interview, requesting preferred time and date for the interview, using the script provided as Attachment H. Abt Associates staff will follow any voice mail messages left with respondents with an email providing the same information left on the voice mail and attach a PDF-version of the instrument. The emails will be sent approximately weekly until the interview is scheduled or the respondent actively declines participation by notifying the project team.

• Telephone calls to non-respondents will begin one week after initial attempts to schedule the interview, and continue approximately weekly until the interview is completed or the respondent actively declines participation.

All data collection and recruitment protocols for this work have been approved by Abt’s Institutional Review Board (IRB).

Burden Hours Estimated for Pre-test and Follow-up Interview

BJS estimates the total respondent burden as 57 hours. The survey will be administered to 40 respondents with an average self-administration time per respondent of 25 minutes and an additional 60 minutes for follow-up interview. Project staff from both BJS, Abt, as well as four outside experts reviewed and tested the survey for quality assurance and burden and reported completion times averaging about 25 minutes.

Table 2. Summary of burden hours for the Survey of Jails Administrative Records Pre-test and follow-up Interview
Reporting mode	Purpose of contact	Number of data providers	Average reporting time	Total burden hours
Online	Survey of Jail Administrative Records pre-test	40	25 min	17
Email and telephone	Respondent follow-up interview	40	60 min	40
Total				57

Efforts to Identify Duplication

Abt conducted a literature review and environmental scan to identify any additional potential duplication efforts currently taking place in the field. None were identified.

Informed Consent and Data Confidentiality

The surveys are not intended to collect information on individuals or ask for information that would otherwise be considered sensitive in nature. As such, the activities associated with this task are not considered human subjects research. The beginning of the survey contains an informed consent specifying that participation in the survey is voluntary and that respondents may decline to answer any and all questions and may stop their participation at any time. The SJAR data will be used to make changes to the SILJ collection. Individual jail responses collected from SJAR will not be published or released, but BJS plans to produce a methodological report that includes information on the survey design, survey procedures, and data quality issues.

Data Security

Information collected from the surveys will be stored on Abt Associates’ computer network that resides behind Abt’s firewall. Precautions will be taken to protect personal information, i.e., respondents’ contact information, by maintaining responses on a password-protected computer for analysis. Because the survey elicits factual information about jail capabilities and data systems and the only human subjects data collected is name and contact information for any follow-up questions, Abt’s IRB has determined that the data collection does not involve human subjects research (Attachment I).

Contact Information

Questions regarding any aspect of this project can be directed to:

Todd Minton

Statistician

Bureau of Justice Statistics

U.S. Department of Justice

810 7th Street NW

Washington, DC 20531

Office Phone: 202-305-9630

E-Mail: [email protected]

Attachments

1. Paper version on the Survey of Jail Administrative Records

2. Selected screenshots from web-based survey data collection tool

3. Interview Protocol

4. Initial invitation letter

5. Initial invitation email

6. Follow-up email reminder

7. Follow-up telephone call script

8. Interview scheduling script

9. Letter of project approval from Abt IRB

10. SILJ fact sheet

11. SJAR fact sheet

1https://www.nap.edu/catalog/12671/ensuring-the-quality-credibility-and-relevance-of-us-justice-statistics

2OMB. (2014, February). Guidance for Providing and Using Administrative Data for Statistical Purposes. http://www.thecre.com/oira_news/wp-content/uploads/2014/02/m-14-06.pdf

3U.S. Census Bureau, American Community Survey Office. Evaluating the Availability and Suitability of External Data Sources for Use in the American Community Survey. July 24, 2015.

4Agency for Healthcare Research and Quality, US Department of Health and Human Services. Medical Expenditure Panel Survey. https://meps.ahrq.gov/mepsweb/index.jsp.

5Gasper, J., & Henderson, K. (2014). Sector-Focused Career Centers Evaluation: Effects on Employment and Earnings after One Year. Prepared for NYC Center for Economic Opportunity by Westat, http://www.nyc.gov/html/ceo/downloads/pdf/CEO-Sector_Based_Approaches_Evaluation_Report-2014_final.pdf.

6Michaelides, M., Benus, J., Poe-Yamagata, E., Shen, T., Bill, N., Carrington, H., & Tirumalasetti, D. (2012). Impact of the Reemployment and Eligibility Assessment (REA) Initiative. Prepared for the U.S. Department of Labor by IMPAQ International.

7Burghardt, P., McConnell, S., Schochet, P. Z. (2006). National Job Corps Study and Longer-Term Follow-up Study: Impact and Benefit-Cost Findings Using Survey and Summary Earnings Records Data. Prepared for the U.S. Department of Labor, Employment and Training Administration by Mathematica Policy Research.

8Kleinman, R, Liu, A., Mastri, A., Reed, D., Reed, D., Sattar, S., & Ziegler, J. (2012). An Effectiveness Assessment and Cost-Benefit Analysis of Registered Apprenticeship in 10 States. Prepared for the U.S. Department of Labor, Employment and Training Administration by Mathematica Policy Research (which uses a variety of Federal and state administrative data, including Unemployment Insurance wage records, and data from the Current Population Survey).

9National Core Jail Standards, American Correctional Association, June 2010.

10OMB. (2016, July). Using Administrative and Survey Data to Build Evidence. https://www.whitehouse.gov/sites/default/files/omb/mgmt-gpra/using_administrative_and_survey_data_to_build_evidence_0.pdf

11Using administrative data in social policy research, OPRE Report #2016-62, Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.

12Bureau of Justice Statistics. (2003). Compendium of State Privacy and Security Legislation: 2002 Overview. https://www.bjs.gov/content/pub/pdf/cspsl02.pdf

13The 2002 Compendium of State Privacy and Security Legislation indicates that all States have enacted legislation governing at least the dissemination of criminal history records. Although the approaches differ considerably, virtually all States have followed the lead of the DOJ regulations in distinguishing between information referring to convictions and current arrests, on the one hand, and nonconviction data on the other (information referring to cases without recorded dispositions or with dispositions favorable to the accused). Most States have placed stricter limits on the release of nonconviction data for noncriminal justice purposes, such as background screening for employment and licensing purposes.

14APPRISS is a private software company that has linked booking records from over 2000 facilities in 48 states and makes them available to law enforcement and correctional agencies via a secure internet connection.

15Beck, Allen. State and federal corrections information systems: An inventory of data elements and an assessment of reporting capabilities. DIANE Publishing, 1999.

16United States Department of Justice. Office of Justice Programs. National Institute of Justice. Arrestee Drug Abuse Monitoring (ADAM) Program in the United States, 2000. ICPSR03270-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2001. http://doi.org/10.3886/ICPSR03270.v1

17Wei, Q. & Parsons, J. (2012). Using Administrative Data to Prioritize Jail Reentry Services Findings from the Comprehensive Transition Planning Project. Vera Institute Research Brief. http://archive.vera.org/sites/default/files/resources/downloads/CTPP-research_brief.pdf

18Swanson et al., (2013). Costs of Criminal Justice Involvement Among Persons With Serious Mental Illness in Connecticut. Psychiatric Services, 64, 630-637.

19Robertson, A. G., et al. (2015). Influence of criminal justice involvement and psychiatric diagnoses on treatment costs among adults with serious mental illness. Psychiatric Services, 66, 907-909.

20Robertson, A.G., et al. (2014). Treatment Participation and Medication Adherence: Effects on Criminal Justice

Costs of Persons With Mental Illness. Psychiatric Services, 65, 1189-1191.