Additional Justification for change request

Additional explanations for change requests to the National Inmate Survey, OMB 1121-0311

General explanation

We learned substantially about the strengths and flaws of our instrumentation in conducting the first survey of sexual victimization in prisons and jails across the nation. There are some desired changes that require additional research and testing which we hope to implement in Year 3. The current proposed changes represent what we believe are non-substantive changes to the instrument, refined to collect better data, or elaborate minimally upon the existing data in order to better address important policy questions posed by stakeholders of the Prison Rape Elimination Act of 2003 (PREA). Changes in the paper and pencil interview (PAPI) instrument strategically mirror the data collected in the audio computer-assisted self-interview (ACASI) instrument more closely and enhance our ability to treat these cases in a similar way.

Reformatting Section B (B1-B27)

During the 2007 National Inmate Survey (NIS) we determined that this module was somewhat burdensome on some inmates who viewed the questions as personal, and on the interviewers who had to record the information into the computer. In order to allow inmates to answer the questions themselves (using ACASI), it was necessary to reword the questions and format them for self-administration. The resulting module includes fewer questions (taking into account skip patterns, inmates answer an average of 15 questions as opposed to an average of 32 question in the 2007 NIS), is estimated to take approximately half the time of the 2007 module, and inmates will be able answer the questions in complete privacy which will increase their willingness to report honestly.

We recognize these changes do not allow for the collection of specific offense, as in the first year. However, we had 50,000 unique literals to re-code as a result of the data-entry function in the previous instrument. This clustering of offense types still allows general categorization of offense and is much simpler to clean for analysis.

Following up on number of times

Methodological literature related to asking about number of times for an event or incident indicates that an open field should be avoided when possible, as respondents tend to estimate upwards, and load on zeroes and fives. In an effort to glean more information about serial victims, however, we would like to follow up with these questions to clarify the number of times. In our Year 1 analysis, we coded entries of 11 or more as 12 for a conservative estimate of incidence; the follow-up questions here would eliminate the need to estimate number of times for categories.

Obtaining data on the first and most recent victimization

Two questions that remain unanswered after Year 1 are when an inmate is most at risk for victimization and when the most recent victimization occurred. It is believed that inmates are most at-risk upon initial admission, particularly in jails, yet we have no specific evidence to support or refute this belief. Administrators would find this information very useful in sexual assault prevention efforts. Similarly, we are missing data for repeat victims on the most recent occurrence of sexual assault. It is unclear from Year 1 whether inmates experienced sexual assault recently, also a concern for prevention and reporting efforts for administrative staff.

Following up on position of staff involved in sexual assault

In our Survey of Sexual Violence (SSV) we collect information on substantiated incidents of sexual assault, including the position of staff involved. Of these incidents proven to have occurred, correctional officers are most often involved. These follow-up questions allow us to determine whether this holds true for allegations made by inmates in the NIS. Such information is important in terms of training staff regarding boundaries with inmates.

Pat-downs and strip-search follow-up questions

Administrators were concerned that inmates may be reporting inappropriate touching by staff which occurred in the course of a pat-down or strip search. In some facilities, such touching made up the majority of incidents alleged by inmates, and BJS was unable to make this important distinction. While the added questions will not allow us to determine whether a search was truly inappropriate, they will allow us to report a proportion of allegations in this category. This information allows administrators to revisit policies and procedures regarding searches and provide additional protections for both their staff and for inmates.

PAPI changes

Although few inmates participate in the survey using the PAPI, it is important that we are able to describe the characteristics of the incidents and possible correlates with victimization from each and every victim we run across in our sample. From Year 1 analysis we found that sexual orientation and previous sexual assault were highly correlated with victimization in the current facility.

Incentives

During the 2007 NIS, we did not provide incentives of any kind. This yielded a response rate of 72% in prisons and 67% in jails. We believe offering a small snack item as an incentive to inmates will maintain or boost the 2007 rates in light of the national trend in decreasing response. Maintaining response level is critical to NIS in order to provide Congress with facility sexual victimization rankings as mandated by PREA.

Our proposal is to offer each sampled inmate a snack item (eg., granola bar, crackers, etc.) that he or she must consume during the interview. The inmate will not be allowed to leave the interviewing room with the snack which will eliminate the possibility of it becoming currency within the facility. During our preliminary calls with sampled facilities, we will determine whether the facility is willing to have us provide incentives. If the facility is unwilling, we will not offer incentives at the facility. In facilities where an incentive is permitted, the inmates will be asked one additional question during the debriefing section (MINC1) to find out how important the incentive was in their decision to participate in the 2008 NIS. We will use this information to determine if incentives should be maintained in Year 3 of the survey.