BJS Response to OMB/OIRA Inquiry

4.1D Random Assignment to Data Linkage Timeframes

“Could you provide more background about why you are doing this?  Why is this important to the field test and/or overall survey?  What is the significance of the two timeframes?  Why would or wouldn’t a respondent want to agree to this linkage?”

• The 2016 Survey of Prison Inmates (SPI) was the first iteration of survey that had a data linkage element. It included a 5-year time limit in which participants consented to have their survey responses linked to employment and incomes records, criminal records, and government benefits records. The year limit proved to be a somewhat of a challenge logistically on the backend as the record linkage had to be completed within 5 years of survey being out of the field (late 2021). With this limit, it meant only about 2-3 years of subsequent data were available to be linked. We want to test a longer period of 10 years and see if this impacts consent. A longer follow-up would account for the operational difficultly of linking data and would include a potentially larger population to be linked (more potentially incarcerated people who will have been released after a 10-year period rather than 5-year period). If a sample member declines data linkage, the interviewer will note this refusal, but the survey can still be administered if the sample member agrees to be interviewed. About 98% of respondents consented to the record linkage in 2016, therefore we do anticipate respondents agreeing, but there is some concern that a decade of potential linkage could be off-putting to some, therefore we want to test it before the national implementation. As this is a field test, consent to data linkage will only be requested of respondents to test willingness; no linkage will actually be performed. After the interview, respondents who consented to the linkage will be told, “Earlier, I told you that your survey answers would be linked to other information about you. Thank you for agreeing to that part of the study, but we will not need to link your answers. It is very helpful to know that you would be willing to permit that.”

4.4 Burden Hours

“The burden for escorting incarcerated participants to/from the interview is accounted for, but the burden for going to/from the interview for the incarcerated respondent is not included.  Their only burden is during the interview.  The incarcerated participant also loses time going to/from the interview.”

• This was not included in the full OMB packages for the 2016 SPI, 2023/24 National Inmate Survey (NIS)-Prisons or 2024 Survey of Inmates in Local Jails (SILJ) national study materials nor have we included it in other generic clearance requests. I am not aware of any precedent for including this in the estimates from other similar surveys of incarcerated persons.

• If we are to assume that the time includes their movement to the interview location and from the interview location, this would be a total of 15 minutes. We came to this conclusion since the 30 mins for the staff member includes a return trip based on their retrieval of the person and back to an assigned space, therefore the time for the incarcerated person would be half as much.

“For virtual interviews, are facility staff present in any capacity either inside/outside the room to monitor the respondent as they participate in the interview?  If so, this would be extra burden incurred on the facility staff that should be added to burden hours.”

• There is not an expectation for facility staff to be present or actively monitor either the in-person or virtual interviews. For both conditions, the interview must occur outside of hearing range and screen outside of viewing range of other people (either incarcerated adults or facility staff). We will adhere to facility-specific security protocols, such as maintaining line of sight of the incarcerated person during the interview. This may be physically monitoring the location or using video surveillance. Alternatively, interviews may be conducted out of sight from facility staff, such as in a telehealth room or office. However, this is dependent on the facility and their protocols, not a feature or necessity of the field test. Facility staff are expected to continually monitor the safety and security of the population, therefore doing so in the context of the interview is not an additional burden to the staff.

“For in-person interviews, are facility staff present in any capacity either inside/outside the room during the interview or do the staff simply drop the respondent off and return an hour later to escort the respondent?   I assume that a staff member may need to be available for the duration of the interview in case the respondent needs to be excused early. If facility staff are present in any capacity for the duration of the interview, then their time would need to be included in the burden estimate.”

• Facility staff are not expected to be inside or outside the room during the interview. They are able to drop the person off at the location and return at the completion of the interview to retrieve the person and take them back to where they need to go.

• Note: some facilities may rely on a pass system whereby someone is granted a pass that is signed off on by the staff member at the location they are coming from and going to. This could be in use at the facilities that may impact the burden estimate. Because every facility is different and may have different security protocols, we are not including monitoring of any kind by staff since it is not a requirement of the data collection itself.

“What is the burden estimate for facility staff to conduct background checks for interviewers or to conduct facility security protocols?  This is not accounted for in the burden hours estimate.”

• Such tasks were not included as part of the burden estimates for the 2016 SPI or the 2023/24 NIS. The 2024 SILJ did include an estimate of 45 minutes associated with logistics planning for the jail facilities which included communicating security protocols and background checks. We could potentially anticipate a similar time commitment for facility staff in prisons.

“In section 9.1 you discuss conducting debriefing discussions with facilities about the overall process for conducting the interviews.  Please include some burden hour estimates for these discussions.”

• This would not necessarily be a scheduled meeting or set of questions, but rather a check after the completion of the interview in the facility about how things went and if they had any concerns, feedback, or thought about how the process could be improved. It would not be mandatory for a facility liaison to participate in this debriefing or expected that they take up an extended period of time to provide their thoughts. A conservative estimate may be 15 minutes of additional burden to debrief. This could also be conducted via email at a later date per facility staff preference.

9.1 Analysis Plan Summary

“How do you plan on measuring respondent engagement for your analysis?”

• Quantitatively, this can be measured in the number of break-offs for each mode type or the number of attempted break-offs that were managed by the interviewer. There may be some qualitative indicators we ask field interview staff to make note of during the interview. For example, if the respondent does not maintain eye contact, is fidgeting or rising from the chair, etc.

Frequently Asked Questions (from the Appendices)

“It may be helpful to briefly note in the What should I say to the incarcerated persons who have been randomly selected to participate? section why the facility shouldn’t provide any details to the incarcerated person about the study.  I don’t know if the facility understands the importance of being vague when discussing the study with the incarcerated person.”

• We will make the following edits and add text (shown in highlighted text below) to the existing FAQ:

What should I say to the incarcerated persons who have been randomly selected to participate?

When approaching those who have been randomly selected to participate in the SPI Field Test, please use the following script: “[Insert person’s name], our facility is participating in a special research study and one of the researchers would like to invite you to participate in an interview. An interviewer will meet with you [in-person/virtually by phone] and can tell you what it is about. Please come with me so I can introduce you to the interviewer.”

We request that facility staff refrain from providing additional information about the study to the sampled individuals. This helps achieve three goals: (1) it minimizes burden placed on staff (i.e., they do not need to learn details of the study design), (2) it helps ensure that sample members receive complete and accurate information, and (3) it helps interviewers establish rapport with sample members as they introduce the study.

If the person requests more information from facility staff, you can say: “I do not know much about the interview, but the interviewer will tell you what it is all about. If you come with me, I will introduce you, but you do not have to come with me or participate in the survey.”

Table 1: Burden Associated with Planned Testing – Updated Estimates in Red

Category of Respondent and Activity		Maximum number of respondents	Estimated burden (minutes)	Total burden (hours)
State and federal governments – provide points of contact for facilities that will be part of the sample		8	15	2
State and federal governments – provide preliminary and final rosters from which incarcerated individuals are sampled and then assigned to in-person or virtual interview conditions		30	30	15
State and federal governments – conduct logistical planning and background checks on interview personnel		30	45	23
In-person interviews	State and federal governments – staff escorts sampled participants	1,500	30	750
	Incarcerated individuals – participants escort and interview time	1,500	75	1,875
Virtual interviews	State and federal governments – staff escorts sampled participants	1,500	30	750
	Incarcerated individuals – participants escort and interview time	1,500	75	1,875
Total burden		6,030		5,290

Note: These estimates represent the number of facilities and inmates that will be sampled, not the number expected to participate. As such, these burden hours represent the maximum burden.