SSV Supporting Statement Part B

B. Statistical Methods

1. Description of Universe

The SSV collects information on incidents of sexual violence that occur in correctional facilities. BJS estimates that there are 8,087 facilities covered by the Act as described in table 2:

Table 2. Estimated number of facilities covered by the

Prison Rape Elimination Act

Facility type	Number of facilities
Total	8,087
Prisons
Public - Federal	114
Public – State	1,304
Private	415
Local Jails
Public	3,234
Private	37
Juvenile facilities
Public	1,152
Private	1,658
Other facilities
Indian country jails	83
Military-operated	59
ICE-operated	31

The Act requires that BJS collect data from a sample of at least 10% of correctional facilities covered under PREA. Because of the low numbers of reported sexual victimizations to correctional authorities and the centralized authority at the jurisdiction level that governs responses to the BJS surveys, the SSV elected to conduct a complete enumeration at the system level – including the federal bureau of prisons, all state departments of corrections, all state juvenile justice systems, the each branch of the U.S. military. In the case of each of these systems, information systems are centralized and the authority to report is centralized. Moreover, this annual enumeration minimizes burden on the respondents (rather than selecting a sample from the more the more than 2,000 facilities operated by these systems).

In the case of the Bureau of Immigration and Customs Enforcement (ICE), a form is sent to each facility operated by ICE (31), as the system is largely decentralized and headquarters is unable to report for all facilities combined.

Finally, for local jail jurisdictions, private prisons, private jails, and local and private juvenile facilities, which are completely decentralized, a sample of each is drawn based on the most current frame available. For type of sample facility/jurisdiction, a detailed description (including coefficients of variation and variance estimates) has been provided (See Attachment 3).

A brief discussion is provided, below:

Federal and state prisons

Each year, the Bureau of Prisons and the 50 state departments of correction submit information based on reports of sexual assault among all state/federal operated facilities for a complete enumeration without the need for sampling. 

Private prisons

The private prison sample draws from the most recent Census of State and Federal

Adult Correctional Facilities (SFACF), conducted every 5-6 years by BJS. The 2005 CSFACF will be used for the 2010 and 2011 samples. The next CSFACF is scheduled for 2011 and will be used for the 2012 sample. Each year, as the CSFACF ages, it is adjusted for any openings and closings as they are discovered, via internet searches of private correctional companies.

In 2010, a sample of 125 privately operated state and federal prison facilities is expected to be drawn to produce a sample of the 417 private prisons identified by the 2005 CSFACF. A 30% sample of facilities is needed (as opposed to the 10% mandated by PREA) to create more precise estimates of sexual victimization in private facilities. (See Attachment 3.)

As with previous samples, facilities will be ranked by average daily population (ADP) in the 12-month period ending June 30, 2005. Facilities with ADPs greater than 488 inmates (n=71) will be selected with certainty because of their size. The remaining facilities will be sorted by region (i.e., Northeast, Midwest, South, or West), state, and ADP, and sampled systematically with probability proportional to size.

Public jails

The public jail sample for 2010 will be drawn from the 2009 Deaths in Custody collection (using the Annual Summary on Inmates under Jail Supervision). The Deaths in Custody file for the preceding year of collection is been used as the sampling frame for SSV.

In 2010, a sample of 700 publicly-operated jail facilities will be selected. As with private state and federal prisons, a sample greater than 10% is needed to provide more precise estimates of sexual victimization. Although the 2010 sample has not been selected, we expect the sample to follow the 2009 sampling procedures:

In 2009, the largest jail jurisdictions in 45 states and the District of Columbia were selected to meet the PREA requirement that at least one jail per state is selected each year. Another 128 jail jurisdictions with ADPs greater than or equal to 1,000 inmates were selected with certainty. The remaining 2,693 jail jurisdictions on the frame were then grouped into three strata. The first stratum contained 1,489 jails with an ADP of 85 or fewer inmates; the second stratum included 770 jails with an ADP of 86 to 268 inmates; and the third stratum included 434 jails with an ADP of 269 to 999 inmates. The cumulative sqrt(f(y)) method was used to determine noncertainty stratum boundaries (Cochran, Sampling Techniques, 1997 edition, p. 129). Jail jurisdictions in these three strata were sorted by region, state, and ADP and selected systematically with probability proportional to their size, resulting in 99 selections from stratum 1, 317 from stratum 2, and 110 from stratum three.

Private jails

Similar to the public jail sample, the private jail sample in 2010 will be drawn from the most recent Deaths in Custody file (2009). Like the public jail sample, the Deaths in Custody file will also be used to create the 2010, 2011, and 2012 sampling frames.

In 2009, a sample of 15 privately operated jails was selected based on data reported in the 2008 Deaths in Custody Annual Summary on Inmates under Jail Supervision. As in prior years, the 41 private facilities on the sampling frame were sorted by region, state, and ADP, and 5 jails were systematically sampled with probability proportional to size.

Given the large standard errors, estimates for private jails are combined with public jails. The separate sample is used to ensure inclusion of private jails in the SSV; in addition, summary counts of reported allegations and substantiated incidents for each sampled private jail are listed in the annual BJS report. This meets the reporting requirement under PREA to reveal counts by facility name.

Other prisons and jails

BJS collects a report from the main branches of the military each year to cover all facilities run by the Army, Navy, Coast Guard, Marines, and Air Force. Similarly, all facilities operated by the Bureau of Immigration and Customs Enforcement (ICE) are included each year. The ICE website is used to create the sampling frame. BJS intends to use the same procedures for 2010-12.

Jails in Indian country are sampled each year using the Annual Survey of Jails in Indian Country (SJIC) to create the sampling frame. The SJIC includes all known Indian country correctional facilities operated by tribal authorities or the Bureau of Indian Affairs, U.S. Department of the Interior. In 2009, 15 of the 63 jails in Indian country that housed adult inmates were selected. Two were sampled with certainty (with an ADP of over 100), while the remaining 13 were sorted by state and ADP and sampled with probability proportionate to size. BJS intends to use the same sampling procedures for 2010-12.

Private and local juvenile facilities

For private and local juvenile facilities in the 2010 sample, BJS will use the prior-year Census of Juveniles in Residential Placement (2009), conducted by the Census Bureau for the Office of Juvenile Justice and Delinquency. For the 2011 sample, BJS will use the 2010 Juvenile Residential Facility Census (JRFC), also conducted by the Census Bureau in years in which the CJRP is not conducted.

The 2010 SSV will include all state-operated juvenile residential placement facilities used to house juveniles and youthful offenders, regardless of age or reason for placement. As defined in the JRFC, residential placement facilities included detention centers, training schools, long-term secure facilities; reception or diagnostic centers; group homes or halfway houses; boot camps; ranches; forestry camps, wilderness or marine programs, or farms; runaway or homeless shelters; and residential treatment centers for juveniles. In the 2009 SSV all states and the District of Columbia operated a total of 473 juvenile facilities, and data for these facilities were collected from 51 central reporters. BJS intends to employ the same procedures for 2010-12.

Based on the sample design in the 2009 SSV, non-state, non-tribal juvenile facilities will be sampled in 2010-12. In 2009,

1. 36 locally operated facilities were sampled with certainty because they were the largest in their respective states, and 1 additional facility was included in the sample based on its size alone.

2. 19 locally operated facilities that were not the largest in their respective states were sampled with probabilities proportionate to size from 2 strata based on commitment status only.

3. 51 privately operated facilities were sampled with certainty as the largest in their respective states, and 1 additional facility was included in the sample based on its size alone.

4. 74 other privately operated facilities were sampled with probabilities proportionate to size from 2 strata defined by commitment status only.

5. 148 non-state detention facilities were sampled. Thirty-five facilities were included in the sample based on their sizes alone, while the remaining 94 facilities were stratified by region and selected with probability proportionate to size.

Finally, the 2009 sample also included the 19 tribal juvenile facilities identified in the 2008 Survey of Jails in Indian Country, bringing the total number of non-state facilities in the 2009 sample to 349.

See Attachment 3 for further details for the 2009 sample and expected corresponding coefficients of variation.

2. Data Collection Procedures

Each system/facility will receive a letter (attached) and a copy of the appropriate SSV form(s) following OMB approval (May 2011 for adults and June 2011 for youth). The U.S. Census Bureau will provide each respondent with the appropriate version of the SSV form. Respondents will transcribe data from their administrative records to the standardized forms or submit their data on the Web. The Census Bureau will follow up with a fax, and then phone calls and emails after the expiration of the return date on the form. Typically, Census is successful in obtaining over 98% of the responses. The remaining non-respondents are forwarded to BJS, and BJS follows up with personal calls to obtain the information. BJS has been successful in this endeavor. (See response rate discussion below).  

3. Methods to Maximize Response

Every effort is being made to make the survey materials clear and straightforward. The SSV questionnaire has been designed to make collection of the data as concise and easy for the respondents as possible. Some examples include uniform definitions of terms and concepts as well as counting rules for items to be reported. Additionally, the SSV uses some questions that have been used previously in other surveys and are known to be easily reported by most respondents.

Over the course of collection, Census Bureau staff learned that state reporters were including incidents that occurred in private- and/or locally-operated prisons in their counts of sexual victimization even though the survey instructions state that these incidents should be excluded. Since private and locally operated prisons are sampled separately from state prison systems, the inclusion of incidents occurring in these facilities by state systems leads to synthetically inflated estimates.

To correct for this issue, two follow-up questions were added to the 2009 SSV summary state adult and juvenile forms to minimize the burden associated with the Census Bureau making follow-up calls to all 50 state reporters:

13. Did any of the allegations reported in Items 2, 5, 8, or 11 occur in a privately operated facility?

Yes

No

14. Did any of the allegations reported in Items 2, 5, 8, or 11 occur in a facility operated and administered by local government?

Yes

No

Only when a state responded “yes” to either question did the Census Bureau followed up to determine how many incidents occurred in non-applicable facilities. The erroneous inclusions were then subtracted from the state totals. In 2009, five state adult prison systems and three state juvenile systems responded that they included allegations that occurred in privately operated facilities (against survey instructions). In addition, seven state adult prison systems and eight state juvenile systems responded they included incidents that occurred in a facility operated and administered by local government.

Table 3 (page 17) summarizes reporting capabilities of correctional authorities, by type of reported assault for collection years 2004 - 2008. This table tracks significant improved capacity to report data using uniform definitions and survey categories among State and Federal prison systems. However, local jail authorities show lower levels of conformity to data standards. Due to sampling, approximately one-third of jails selected each year are receiving the summary form containing the definitions and reporting criteria for the first time. However, reporting capabilities have improved as the survey includes an increasing percentage of jurisdictions that have previously been sampled.

Table 3. Reporting capabilities of correctional authorities to provide data, by type of assault, 2004-08

Overall, the data demonstrate an increasing capacity to provide data. Measures of nonconsensual sexual acts have the highest percentage of partial reporting (6% in jails), prisons and jails report that the they are most unable to report on incidents of staff sexual misconduct (2%); some 2% of jails are unable to report abusive sexual contacts, as well. BJS believes that such low levels do not impair the ability to draw accurate estimates of sexual violence by type. BJS reports these deviations from the reporting criteria, but does not introduce adjustments to account for these deviations. Instead, BJS provides detailed appendix tables containing notes deviations for applicable facilities/jurisdictions.

The response rate varies according to the type of correctional jurisdiction. All 50 State Departments of Correction and the Federal Bureau of Prisons have participated in the 2004 - 2008 collections (100%). For the remaining entities, responses have been as follows: 1 adult and 3 juvenile non-responses in 2004; 0 adult and 6 juvenile non-respondents in 2005; 5 adult and 7 juvenile non-respondents in 2006, 10 adult non-respondents in 2007, and 12 adult non-respondents in 2008.

Table 4 (page 18) summarizes item non-response on the SSV incident report. Overall, these data show low non-response rates for critical items. Item non-response rates greater than 10% are observed for the following items: time of occurrence of staff-on-inmate victimization (item 4; 13.7%), age of the second victim in inmate-on-inmate sexual victimization (item 10; 16.2%), race of the second victim in staff-on-inmate sexual victimization (item 11; 10.8%), age of the second perpetrator of inmate-on-inmate victimization (item 22; 17.2%), and race of the second perpetrator of inmate-on-inmate victimization (item 23; 12.6%).

While some of these rates may be considered high, the need to track these characteristics outweighs deletion of these items from the SSV. Moreover, based on data from SSV 2007-08 in prisons and jails, most substantiated incidents of inmate-on-inmate victimization involve a single victim (96%) or a single perpetrator (88%); and most incidents of staff sexual misconduct/harassment involve a single victim (91%) or a single perpetrator (96%). These item specific non-response rates reflect an absence of information, even following completion of an investigation.

4. Testing Procedures

BJS and Census have established that jurisdictions are able to respond to the SSV. Items will not change from the current collection.

5. Contacts for Statistical Aspects of Data Collection

BJS takes responsibility for the overall design and management of the survey, including sampling procedures, development of the questionnaires, and the analysis and publication of the data. The BJS contacts are:

Paul Guerino, Statistician

Bureau of Justice Statistics

810 Seventh St., N.W.

Washington, DC 20531

(202) 307-0349

Paul.Guerino@usdoj.gov

Allen Beck, Senior Statistical Advisor

Bureau of Justice Statistics

810 Seventh St., N.W.

Washington, DC 20531

(202) 307-0349

Allen.Beck@usdoj.gov

The Governments Division at the Census Bureau is the collection agent and is responsible for the collection of all data. The contact at the Census Bureau is

Stephen G. Simoncini, Chief

Criminal Justice Statistics Branch

Governments Division

U.S. Census Bureau

Washington, DC 20233-6800

(301) 763-7375

Stephen.G.Simoncini@census.gov