Document

Supporting Statement B

ICR 202611-1121-001 · OMB 1121-0363 · Object 175045500.

Document Viewer [docx]

Status: Original and derived artifacts are available for this document.

Download: docx | pdf | html

Primary: docxSource: application/vnd.openxmlformats-officedocument.wordprocessingml.document
Loading document viewer…

Document Metadata

Record metadata
application/vnd.openxmlformats-officedocument.wordprocessingml.document
Supporting Statement B
Oudekerk, Barbara Ann
Writer
2026-10-28
2026-11-17
complete

Extracted Text


B. COLLECTION OF INFORMATION EMPLOYING STATISTICAL METHODS
1.	Universe and Respondent Selection
Universe
The sample universe for the 2026 National Survey of Victim Service Providers (NSVSP) is all entities nationwide (1) identified through the 2023 National Census of Victim Service Providers (NCVSP) as eligible victim service providers (VSPs) and (2) not identified as ineligible during the NCVSP frame verification outreach that will occur approximately 4 months prior to the launch of the 2026 NSVSP (OMB No. 1121-0339; exp. 01/31/2028).  Eligible VSPs are those organizations where victim services are their primary function, or they have designated victim services staff or programs and have provided direct services to victims within the previous six months.
Frame
The 2023 NCVSP identified 14,266 VSPs in the 50 states, the District of Columbia, and the U.S. territories of Guam, the Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands.1 The 2026 NSVSP sampling frame will be constructed from the 2023 NCVSP data file, which includes 13,951 eligible VSPs prior to the frame verification operation.
Within the 2023 NCVSP data, VSPs are categorized into seven VSP types: tribal government, tribal coalition, or other tribal organization; campus organization or other educational institution (public or private); hospital, medical or emergency facility (public or private); government agency; nonprofit or faith-based organization (501(c)(3) status); for-profit organization; and informal organization (such as a survivor network). Keeping consistent with the 2019 NSVSP, VSPs categorized as for-profit (n = 73) or informal organizations (n = 61) will be excluded from the sampling frame. VSPs located in the aforementioned U.S. territories (n = 179) and those missing a value for VSP type (n = 2) will also be excluded from the frame.2 
Prior to drawing the sample, BJS will conduct the frame verification outreach and update the frame accordingly. This outreach was previously approved by OMB through BJS’s generic clearance agreement (OMB No. 1121-0339; exp. 01/31/2028). VSPs will be asked to provide updated information about how their organization is structured to provide services to victims of crime or abuse and contact information for the person completing the survey. The frame verification outreach will identify VSPs that are no longer eligible for the NSVSP (e.g., because they no longer have a designated victim services program or they have closed). The frame verification outreach will improve the quality of the NSVSP sample by reducing ineligibility rates and updating organizations’ contact information. It will also raise awareness of the NSVSP in advance of its launch. 


Sampling
The 2026 design will produce estimates at desired precision levels for the nation as a whole and for key subgroups including U.S. census region (Northeast, Midwest, South, and West) and VSP type (government and nongovernment). Government will include organizations categorized as tribal government, tribal coalitions, or other tribal organizations and non-tribal government agencies on the 2023 NCVSP. Nongovernment will include organizations categorized as nonprofit or faith-based organizations, campus organizations, and hospital, medical, or emergency facilities on the 2023 NCVSP. The sampling design will stratify VSPs by census region and type by allocating sufficient sample sizes to strata defined along these two dimensions. 
The targeted number of VSPs to be selected for the 2026 NSVSP will decrease from 7,237 in the 2019 NSVSP to a maximum of approximately 6,500 VSPs across an initial and reserve sample in the 2026 collection. The 2026 sampling design aims to improve the precision of national and subgroup estimates compared to the 2019 collection and assumes a higher response rate (80%) than what was assumed for the prior NSVSP sample (70%) due to improvements in survey methodology that will be implemented. Similar to the 2019 collection, the 2026 collection assumes a 15% ineligibility rate given that more than two years will have passed since the sampling frame was developed from the 2023 NCVSP to the launch of the frame verification outreach. 
The goal for the 2026 sample is to be able to produce a margin of error (MOE) of 0.05 (±5%) for an estimated proportion of 50% at the subgroup level and an MOE of 0.02 (±2%) for an estimated proportion of 50% at the national level. An estimated percentage of 50% is used conservatively as it leads to the highest value of the variances and standard errors. 
The initial stratification by census region and VSP type will improve the efficiency of overall sample estimates since the sample variance is decreased as the stratum means diverge and there is within-stratum homogeneity. Samples will be drawn independently from each primary stratum to support precise overall estimates. Table 1 summarizes the sample size allocation to primary strata in the initial sample. It also presents the precision expected for stratum-level estimates. To meet the 3,579 minimum eligible completes (MEC) needed for national and key subgroup estimates, and assuming an 80% response rate and 85% eligibility rate, a sample of approximately 5,266 entities will be selected from this national roster of VSPs. 









Table 1.  Sample allocation of VSPs to strata and expected precision of stratum-level estimates
Strata3
Number of VSPs
Proportion of Total VSPs
 Sample Size1
Minimum Eligible Completes (MEC)2
Sampling Rate of Eligible Completes
Assumed
MOE
Northeast Nongovernment
1,333
0.10
642
436
0.33
0.05
Northeast Government
759
0.05
549
373
0.49
0.05
Midwest Nongovernment
2,224
0.16
703
478
0.21
0.05
Midwest Government
1,293
0.09
637
433
0.33
0.05
South Nongovernment
3,225
0.23
737
501
0.16
0.05
South Government
2,237
0.16
705
479
0.21
0.05
West Nongovernment
1,733
0.12
675
459
0.26
0.05
West Government
1,147
0.08
618
420
0.37
0.05
Frame
13,951
 
5,266
3,579
0.26
 0.02
1 Sample Size = RoundUp(MEC/(.80*.85),0)
2 MEC = (N * (z^2) * (p*(1-p))) / ((MOE^2) * (N-1) + (z^2) * (p*(1-p))) * DEFF, where N = Frequency, z = 1.96, p =.50, DEFF = 1.46
3 Government = Tribal and non-tribal government; Nongovernment = nonprofit/faith-based, university or college campus, and hospital.

Compared to the large number of primary strata (39) used in the 2019 NSVSP, which was designed to produce national estimates and sub-national estimates for census region and the 14 U.S. states with the largest number of VSPs, the reduced number of primary strata (8) in the 2026 collection will improve the precision of national and subgroup estimates. The 2026 collection is not designed to produce any state-level estimates.3 
The 2026 NSVSP sampling design will also introduce another stratification dimension to improve the efficiency of the survey design. The sub-stratification by a measure of the predicted likelihood of response will lead to improved response rates as well as eligibility rates. This sub-stratification will allow the oversampling of VSPs (substrata) that are more likely to participate in the survey.
After first allocating the sample to the 8 primary strata defined by region and type, within each primary stratum, VSPs that are more likely to respond to the survey will be oversampled. The substrata have been constructed by analyzing the 2023 NCVSP and classifying all VSP units in the frame into groups (clusters) based on a “response likelihood index.” This index will use information on the number of contact attempts needed to reach a final disposition for both respondents and non-respondents in the 2023 NCVSP and for the frame verification outreach form. Total contact attempts are broken into quantiles to create the clusters. The numbers of VSPs in the frame by strata and defined cluster are included in Table 2. These groups may also contribute to the identification of VSPs that would benefit from additional support during the 2026 NSVSP data collection. Once contact attempts from the frame update are available, alternative possible cluster structures will be evaluated based on an updated likelihood of response index. The updated likelihood index will strongly consider the responses obtained in the frame verification survey because these are more recent contact attempts.
Table 2.  Sub-stratification of VSPs by predicted likelihood of response
 
Likelihood of Response1
Strata
Highest
Medium-High
Medium-Low
Lowest
Total
Northeast Nongovernment
393
210
246
484
1,333
Northeast Government
86
79
96
498
759
Midwest Nongovernment
743
389
391
701
2,224
Midwest Government
219
178
198
698
1,293
South Nongovernment
1,045
457
621
1,102
3,225
South Government
456
340
308
1,133
2,237
West Nongovernment
528
272
335
598
1,733
West Government
260
199
186
502
1,147
Total
3,730
2,124
2,381
5,716
13,9512
1Likelihood of response to the 2026 NSVSP: Highest (3 median number of contact attempts in the 2023 NCVSP), Medium-High (6), Medium-Low (13), Lowest (no response/uncontacted).
2 Equal to population by strata in Table 1.

The allocation of sampled VSPs to substrata (Table 3) reflects the aim of sampling at higher rates, within each primary stratum, those substrata where VSPs are predicted to have a higher propensity to respond. Table 3 will be updated after the frame verification to reflect additional information on response likelihood. Unequal sampling rates are expected to lead to higher response rates overall while also limiting unequal weighting effects on the variance. This sampling design presents an optimum balance between these two conflicting objectives. 









Table 3.  Allocation of sampled VSPs to substrata
 
Strata
Likelihood of Response1

Highest
Medium-High
Medium-Low
Low
Total
Northeast Nongovernment
289
193
96
64
642
Northeast Government
86
79
96
288
549
Midwest Nongovernment
316
211
106
70
703
Midwest Government
219
178
136
104
637
South Nongovernment
332
221
110
74
737
South Government
317
212
106
70
705
West Nongovernment
304
203
101
67
675
West Government
260
191
99
68
618
Total
2,123
1,488
850
805
5,2662
1Likelihood of response to the 2026 NSVSP: Highest (3 median number of contact attempts in the 2023 NCVSP), Medium-High (6), Medium-Low (13), Lowest (no response/uncontacted).
2Equal to sample size in Table 1.

Reserve Sample
An 80% response rate is greater than both the final 2019 NSVSP rate (57%) and 2023 NCVSP rate (62%). By conducting the frame update and subsequently oversampling VSPs that are more likely to respond to the survey based on their past behavior, a higher response rate is anticipated in the 2026 collection. However, in the event that the 80% target with the initial sample is not met, an additional reserve sample of VSPs (n = 1,216) will be released that can be used to reach the MEC (n = 3,579) needed to produce national and subgroup estimates at the targeted level of precision.

The reserve sample will be drawn at the same time as the initial sample. By preparing a reserve sample in advance, the survey process can maintain the desired level of statistical power and precision for national and subgroup estimates without significant delays, thereby supporting the overall reliability of the 2026 NSVSP data. The reserve sample will be released if response rates for the initial sample are 65% or lower at the time of the initial survey due date (approximately 5 weeks after data collection begins).

The reserve sample assumes a 65% response rate, which is more aligned with response rates of previous VSSRP data collections. Assuming an 85% eligibility rate, a total of 1,216 VSPs will be in the reserve sample. The total sample size across the initial and reserve samples is 6,482 VSPs (5,266 in the initial sample and 1,216 in the reserve). Table 4 displays the initial and reserve sample allocation to strata.  



Table 4.  Allocation of sampled VSPs to substrata, initial and reserve samples
Strata
Number of VSPs
Proportion of Total
Initial Sample Size1
Reserve Sample Size2
Total Sample (initial + reserve)
Minimum Eligible Completes (MEC)3
Northeast Nongovernment
1,333
0.10
642
148
790
436
Northeast Government
759
0.05
549
127
676
373
Midwest Nongovernment
2,224
0.16
703
163
866
478
Midwest Government
1,293
0.09
637
147
784
433
South Nongovernment
3,225
0.23
737
170
907
501
South Government
2,237
0.16
705
162
867
479
West Nongovernment
1,733
0.12
675
156
831
459
West Government
1,147
0.08
618
143
761
420
Frame Total
13,951
 
5,266
1,216
6,482
3,579
1Initial Sample Size = RoundUp(MEC/(.80*.85),0)
2Reserve Sample Size = RoundUp(MEC/(.65*.85),0) – Initial Sample
3 See Table 1 for details of MEC calculation

Drawing a reserve sample to address low unit response rates is preferred to other methods, such as fielding a shortened critical items survey. Fielding a shortened survey late in a collection typically only improves response rates a small amount. A shortened survey was used in the 2023 NCVSP, where almost 90% of respondents to the NCVSP completed the full instrument. For a census collection, like the NCVSP, where the entire population is already being invited to participate and financial incentives are not an option, a critical item survey is one of the few options for increasing response rates. However, for a sample collection where a large percentage of the population has not yet been asked to complete a survey, drawing and fielding a reserve sample is a more efficient means of obtaining the number of completes (MEC in Table 4) needed for the desired national and subgroup estimates. Running a second collection wave with a reserve sample does not incur the additional programming, testing, or training costs associated with fielding a separate critical items instrument. Also, a second wave often produces more timely data. The NCVSP critical items survey was fielded for 2 months before the end of 2023 collection; the proposed NSVSP reserve sample would only add an estimated 6 weeks to the total collection timeline.  
BJS has, however, identified the items that need to be answered in order to be considered a completed survey, prioritizing items that enable proper organization classification and facilitate comparisons with the 2019 NSVSP (see #3 for additional information). BJS will focus on obtaining answers from VSPs for these items during nonresponse follow-up efforts if unit response rates are problematic. 
2. Procedures for Collecting Information
The 2026 NSVSP will be a multimode data collection effort, primarily administered as a web survey. The instrument (see Attachment 1) will be programmed as a web survey but VSPs will also have the option to complete the survey by telephone in subsequent nonresponse follow-up efforts via the NSVSP team and computer-assisted telephone interviewing (CATI). Web-based data collection is the preferred mode to increase response rates, expedite the data collection process, simplify data verification, and facilitate report preparation. In the 2023 NCVSP, about 75% of respondents completed the survey by web, with the remaining respondents completing through a combination of web and phone methodologies. This suggests that most VSPs are capable of and even prefer to complete the survey online but that response rates can be increased by offering a different mode of response as part of the nonresponse follow-up procedures. 
To further increase response rates, VSPs will be classified into different tiers based on anticipated needs/support to complete the survey. Tier assignments will be based on several criteria, including:
    • Likelihood to respond, based on response or nonresponse to the 2023 NCVSP and the frame verification outreach form, 
    • Status as a federally funded victim service organization, gathered from grantee lists from the Office for Victims of Crime (OVC) and the Office on Violence against Women (OVW),
    • Whether the VSP has a case management system, gathered from the frame verification outreach form,
    • Whether the VSP is known to respond to mass violence incidents (the special topic module in the 2026 NSVSP), based on ICF’s knowledge of the field, and
    • Whether the VSP indicated on the frame verification outreach form that they were interested in receiving assistance from ICF if they were selected for the survey.

Based on this information, VSPs considered higher risk for nonresponse to the NSVSP will be assigned to a Tier 2 or Tier 3 group, where they will be offered additional resources and follow-ups from victim services experts on the NSVSP data collection team in the administration materials they receive. VSPs not considered to be high risk for nonresponse will be assigned to the Tier 1 group, which will not be offered these resources in the administration materials they receive. Because some of the tier criteria are based on ICF’s knowledge of the VSP field, assignment into tiers will be qualitative in nature. Tier assignments may shift during data collection as more is learned about the VSP.  
Survey Invitations and Due Date Reminders
Mailed and emailed invitations to complete the 2026 NSVSP will be sent to all VSPs that are selected for the final sample and have active contact information. Table 5 outlines the contacts and events and provides a high-level description of each one. Initial mailed invitations (see Attachment 6) will be followed by an emailed invitation (see Attachment 7) approximately one week later. The invitations will emphasize the significance of the data collection, identify the organizations sponsoring and conducting the survey, and provide an estimated time needed to complete the survey. The invitations will also include a URL and password to complete the web questionnaire. The mailed invitation will include a QR code that, when scanned, takes the respondent to the NSVSP website. The NSVSP website will include links to the FAQs, videos, and other training materials prepared by technical specialists experienced in supporting VSP data reporting.

The invitations will also announce a survey due date approximately 5 weeks after the initial mailed invitation is sent. Announcing a future due date is common practice for surveys of establishments where organizations are asked to complete a questionnaire where they need to query the organization’s administrative records and potentially consult with colleagues. The 5 weeks give individuals an opportunity to plan time to complete the questionnaire, potentially over several days between work obligations, without feeling pressured to complete the survey immediately. There will be 2 email reminders of the due date between the initial mailed and emailed invitations and the announced due date (see Attachments 8 and 9). These emails will be as short as possible, remind VSPs of the upcoming due date, point them to additional resources, and offer assistance, if needed. In these emails, VSPs identified as being at higher risk for nonresponse (Tiers 2 and 3) will see language inviting them to work with a member of the NSVSP data collection team to help them complete the survey. This assistance will include either a short group session focused on pulling needed data from common case management systems or an individual 1-hour support session. This will help to ease burden prior to the due date and prepare respondents for similar future collections. VSPs that have not been identified as high risk (Tier 1) will receive versions of these emails that omit this language. 
Nonresponse Follow-Up
After the announced survey due date has passed, outreach will shift to highlighting an extended due date and offering additional assistance to all nonresponding VSPs. The first nonresponse follow-up email (see Attachment 10) will be sent within a few days of the missed due date and will encourage VSPs to complete their questionnaire by a new, extended due date. This will be the first email where VSPs that were previously considered low risk for nonresponse will be invited to schedule appointments for support. 
Approximately two weeks after the first nonresponse email is sent, a nonresponse letter will be mailed via USPS Priority Mail to all VSPs who have not yet completed the survey (see Attachment 11). Like the invitation letter, the nonresponse letter will include the survey URL and a scannable QR code. A second nonresponse follow-up email (see Attachment 12) will be sent approximately two days after the letters are mailed so that the emails arrive at approximately the same time as the letters. The letter and email will remind VSPs of the extended due date and offer resources for live support to complete the survey. 
Three days after the extended due date has passed, an email will be sent to all VSPs in sample stratum that have not yet reached the target response rate for that stratum (see Attachment 13). This email will inform VSPs that they will begin receiving phone calls over the next few weeks asking them to complete the survey. 
Approximately four days after the most recent email, follow-up with nonresponding VSPs will move to phone calls (see Attachment 14 for the CATI script). VSPs considered low risk for nonresponse based on multiple factors, including being federally funded VSPs and past engagement with the data collection team (e.g., responding to the frame verification outreach or contacting the survey helpdesk), will be contacted through trained CATI interviewers. CATI interviewers will remind VSPs to complete the web survey and offer to complete the questionnaire with them over the phone immediately or during a scheduled appointment. Experienced phone interviewers will be trained on the questionnaire, support materials, and VSP characteristics. 
VSPs identified as being at higher risk for nonresponse will be contacted by the NSVSP data collection team directly. These team members will complete additional custom outreach calls to VSPs identified as high risk for nonresponse based on previously collected information (e.g., do not have case management system, are new to the field) and those who have previously refused to participate in the 2019 NSVSP.
After data collection ends, all responding VSPs will receive a thank-you email (see Attachment 15) that explains how they can use the data in their work and sign up for news about data and publication releases. The reserve sample (if released) will receive the same contact protocol as described above. In addition to the above planned outreach, BJS may conduct additional outreach activities or modify the above communication strategies on an as-needed basis to increase response rates and improve overall data quality. 
Table 5. Data collection protocol and event descriptions

Data Quality Monitoring
Respondents completing the survey via the web instrument will enter their responses directly into the online instrument. The survey will be programmed with internal range and logic checks and custom validations to enforce data quality at the point of collection. 
To confirm that editing rules are being followed, the data management team will review frequencies for the entered data throughout data collection. Any issues will be investigated and resolved. Throughout the data collection period, project staff will conduct regular data frequency reviews to evaluate the quality and completeness of data captured in both the data entry (through trained interviewers via phone) and web modes.
3. Methods to Maximize Response
The 2026 NSVSP will implement best practices in establishment survey research, lessons learned from prior VSSRP collections, and insights from cognitive testing to maximize response rates for the 2026 NSVSP. Key approaches are described below. 
Sample Stratification by Likelihood to Respond
As described in #1 (Universe and Respondent Selection), the 2026 NSVSP sampling design includes sub-stratification by a measure of the predicted likelihood of response. This sub-stratification (described above in more detail) will allow for the oversampling of VSPs (substrata) that are more likely to participate in the survey, which will lead to improved response rates as well as eligibility rates. 
Reduced Survey Length and Improved Design Elements
The 2026 NSVSP instrument represents a major overhaul of the 2019 instrument and was revised with the end goals of minimizing burden on VSPs and maximizing response rates. The 2026 NSVSP instrument and the procedures to collect, clean, and analyze the data have been developed based on technological advances that enhance data quality and minimize burden to survey participants and researchers. Based on the 2023 NCVSP results, it is expected that the majority of VSPs will complete the survey via web. As discussed in more detail in Supporting Statement A, the web-based survey capabilities are designed to assist respondents in completing their questionnaires by providing a high-quality user experience and including features that reduce respondent burden and ensure complete and accurate data. In addition, the web survey has a user-friendly interface, is easy to share, and allows for easy navigation to specific sections in the event that multiple people at the VSP need to collaborate to complete the survey. 
In addition to administration issues, the research team spent significant attention on the NSVSP survey length and content. The project team worked to balance the many data priorities identified by the field with the need to keep the survey to 25 minutes or less. When possible, item wording or content was selected to be similar to wording or content familiar to large parts of the VSP field that are subject to federal grant reporting requirements on performance measures (for example, VSPs with OVC funding report numbers of victims served by crime type each year). The NSVSP is intended to be a more detailed, longer follow-up to the NCVSP but was kept to an average of 25 minutes in light of the response rate challenges BJS had with the 2023 NCVSP. 
Further, the instrument itself includes several design elements intended to increase the ease of reading and understanding the questionnaire. First, related questions are grouped together in topical sections. In addition, the survey instrument begins with the most salient items. Questions and instructions are presented in a consistent manner on each page in order to allow respondents to comprehend question items more readily. Because we expect that some items may need to be answered by different individuals within the organization, the web survey features navigational tools that allow a survey respondent to easily see which sections have been completed and which are still outstanding. The web survey will also feature a link to the FAQs and a glossary of terms, so that no matter what question is being answered, the survey respondent can easily find this information. A helpdesk will be staffed to provide assistance by phone and email to all respondents during normal business hours (Eastern Time) and will be available to all respondents through a toll-free number.
Contact Strategies Designed for Establishments
The NSVSP contact materials utilize best practices for establishment survey research by asking VSPs to complete the survey by a target due date, describing to respondents what information they will need to complete it, and encouraging them to share the survey with their colleagues, as needed. This approach allows VSPs to make the necessary plans to ensure a quick and efficient survey completion experience.
In addition to the above, the contact materials use plain language, describe how the data will be used, highlight key findings from the 2019 collection, and provide links to BJS-produced infographics. These design features were added as a result of cognitive testing, where VSPs expressed frustration with overly formal language and a lack of information about how the data they spend time providing are actually used. BJS’s expert panelists also recommended that BJS be more explicit in explaining in outreach materials how the VSSRP data are used.
Although the web will be emphasized as the preferred mode of survey completion, data collected over the phone during the nonresponse follow-up phase of the fielding will be collected using the same software the VSPs will use to enter their data. Furthermore, telephone interviewers will receive extensive training specific to the project to ensure high response rates and data quality. First, trainees will complete self-paced modules to become familiar with the study background, purpose, and the nature of the content of the survey. In the second phase, trainees will attend a trainer-led session that focuses on gaining cooperation, addressing respondent questions and concerns, and the unique details of the survey. The third phase will consist of role-play exercises where one trainee administers the survey to another trainee over the telephone as if conducting a live interview. ICF supervisors will monitor trainees to evaluate their readiness for live data collection. Once approved, trainees will be assigned as interviewers for the project.
Over the course of data collection, ICF will provide BJS access to a secure portal with real-time unit response rates by mode and VSP type. ICF will also send biweekly production reports to BJS via email that include additional paradata and cost monitoring and highlight the current collection phase and number of recent contacts, item response rates, data quality assessments, and follow-ups. Monitoring response rate and performance paradata will also allow the team to quickly identify problems with participation, and if needed, implement an adaptive design to ensure a representative distribution of VSPs at the national level (e.g., targeting outreach toward specific groups that are not responding at high rates). 
Promoting Awareness of the NSVSP in the VSP Field
Prior to survey launch and throughout the data collection period, outreach efforts will raise awareness of the survey and promote enthusiasm for its success. Project staff will offer fact sheets, office hours (30-minute group sessions with ICF staff, grouped by organizations that use similar case management systems), and one-on-one guidance to VSPs who have been sampled for the NSVSP. All VSPs on the 2023 NCVSP frame will be invited to attend a webinar in late 2026 or early 2027 that will raise awareness of the VSSRP data collections and encourage participation in the upcoming NSVSP. In addition to advertising these resources on the project website and through the administration materials, project staff will ask external stakeholders, including OVC, OVW, and the NSVSP’s expert panel members, to share these resources with the field when posting messages that encourage VSPs to participate in the survey if they were selected. This outreach is expected to be particularly helpful in legitimizing the 2026 NSVSP. 
Addressing Nonresponse and Weighting Adjustments
The anticipated unit response rate for the 2026 collection is 80%. Once the data collection has launched, the approach that will be used for nonresponse follow-up is consistent with best practices in establishment survey research. For example, the phone-based follow-up efforts aim to reduce both unit nonresponse (by encouraging nonresponding VSPs to complete their survey) and item nonresponse (by contacting VSPs with missing data from submitted web surveys). See above in #2 (Procedures for Collecting Information) for the full schedule and description of the NSVSP’s nonresponse follow-up protocols.  
BJS will require responses to the following items to be considered a completed survey response for the purposes of calculating the response rate: A1, A2, A3, A4-A4b, A5-6, A7, A8, C1, E1-E9, E15-E17, and G3-G6. These items, which represent nearly 50% of the full questionnaire, were selected because of their particular importance to the goals of the NSVSP. In addition to being needed to establish eligibility (per BJS’s definition of a VSP), these items will enable comparisons to key items from the 2019 NSVSP and will enable BJS to report out estimates related to mass violence. However, if unit response rates fall below 80%, a nonresponse bias analysis will be conducted to assess the extent to which nonresponding VSPs differ from responders on key characteristics available on the NSVSP frame. Nonresponse bias analyses will start with bivariate comparisons between non-respondents and respondents using characteristics (or variables) that are available on the frame. These variables include, but are not limited to, organization size, type, and location (e.g., urbanicity, region). The bivariate analysis will be followed by multivariate analyses that allow for the identification of independent effects on the response propensity of predictors found to be significant in the bivariate analyses. Machine learning models will be used to more fully account for interactions between the model predictors. The models will fit the probabilities of being eligible as well as being a survey respondent. The weighting process will be informed by the results of the nonresponse bias analysis, and weighting class adjustments will minimize the potential for nonresponse bias. 
The NSVSP weighting process will begin with the computation of sampling weights as the reciprocal of the selection probability for each selected VSP. These sampling weights will then be adjusted, if needed, to minimize nonresponse bias. The adjustments will capitalize on the thorough nonresponse bias analysis described above. Nonresponse adjustment cells will be developed using those variables that are found to be associated with potential bias. Sampling weights will be adjusted for nonresponse so that the sum of the adjusted weights over all responding (eligible) institutions equals the sum of initial weights over all selected institutions, in each adjustment cell and overall. The weight adjustment cells will be created using the variables found to significantly predict response propensity in the models.
In addition to these adjustments, a final post-stratification adjustment will be considered to ensure that final weights sum to known (the 2023 NCVSP frame) totals for key variables such as organization type. The number of nonresponse adjustment variables and post-stratification cells will be limited to control the possible variance inflation that goes along with the bias reduction induced by these adjustments. 
Item nonresponse rates will be analyzed after the end of data collection to inform whether and how to impute. The choice of imputation method will depend on the degree of item-level missingness. If an item nonresponse rate is less than 5%, imputation will not be conducted on the item. If item-level missing is between 5-20%, hot-deck imputation will be used. If item-level missing is greater than 20% but less than 60%, multiple imputation will be conducted. If item-level missing is greater than 60%, no imputation would be conducted and the item would not be analyzed. 
4. Testing of Procedures
ICF conducted two rounds of cognitive testing of the 2026 NSVSP instrument (OMB No. 1121-0339; exp. 01/31/2028) to finalize the survey content and questions (for example, assessing how best to word questions, testing response options, and understanding the feasibility of reporting certain data). One round of usability testing (described in more detail below) will be completed before the full 2026 NSVSP administration to ensure the presentation and programming of the instrument is user friendly. 
Cognitive testing
As part of readministering the NSVSP and prior to cognitive testing, ICF’s survey methodologists reviewed the 2019 NSVSP instrument and its resulting data, alongside the 2017 and 2023 NCVSP instruments and their resulting data, to identify areas for improvement. ICF first conducted a qualitative review of the 2019 instrument to assess its clarity, flow, efficiency, and burden. This examination was followed by a quantitative review of the collection’s resulting data and paradata, focusing on areas like item nonresponse rates and response validity. Results from these reviews informed a draft of what would become the 2026 NSVSP instrument, which was shared with an advisory board of subject matter experts in May 2025 for their review and feedback (see Supporting Statement A #8 for a list of individuals included in the expert panel). Key feedback from this virtual meeting included suggestions for how to measure staffing levels and the most important topics to capture in a new mass violence module. Feedback from the subject matter experts was incorporated into a revised draft of the instrument that was then cognitively tested during the fall and winter of 2025-2026. 
A total of two rounds of cognitive testing began in September 2025 and in January 2026, respectively. Round 1 of testing focused on several key items to assess the completeness of the information collected (e.g., were respondents able to furnish the requested information?) and the uniformity of understanding the survey from one respondent to the next (e.g., did each respondent define mental health services in the same way?). Interviewers discussed selected items with participants but did not ask the participant to actually complete the survey in its entirety. The feedback from the cognitive interviews resulted in edits to the wording of questions and adjustments to particular response options to enhance clarity. During this round of cognitive testing, many cognitive interview participants reported that they would likely participate in the NSVSP if invited. 
Additional revisions to those questions were then tested again in Round 2. In Round 2, cognitive interviewers focused on whether revisions made during Round 1 improved information provision and understanding. The cognitive testing aided the team in refining the questionnaire to reduce burden on the respondent, increase readability, and improve understanding of key terms. Recommendations following cognitive testing were incorporated into the final 2026 NSVSP instrument.
Thirty VSPs participated in Round 1 of cognitive testing of the NSVSP instrument. The 30 participants represented a range of VSPs, including campus-based providers, prosecutor-based providers, sheriff-based providers, hospital-based programs, sexual assault and domestic violence providers, among others. A total of thirteen VSPs participated in Round 2. Between Rounds 1 and 2, discussions were held between BJS and ICF regarding feedback captured from VSPs and what recommendations to implement for Round 2 testing. In Round 2, consensus was achieved in most areas and survey items were adjusted for consistency in references to dates, stem wording, and alignment with federal definitions. New and revised questions were tested in each section of the survey, and key changes within each section of the 2026 instrument that were implemented by BJS are detailed below. A full accounting of which items were tested and the results of the cognitive testing can be found in the 2026 NSVSP Cognitive Testing Report (see Attachment 4).  
About Your Organization (Section A). Section A of the 2026 NSVSP has a series of questions designed to identify the structure (i.e., standalone or embedded; main or satellite location) and organization type (e.g., for-profit, government) of VSPs. Compared to this section’s counterpart in the 2019 instrument, major changes were made to this section for 2026, which draw on revisions that were made to a similar section on the 2023 NCVSP, with additional tweaks based on expert review and cognitive testing. New for 2026, respondents will now be asked to identify whether they are a standalone or embedded VSP at the outset of the survey (item A1), and their response will inform piping in almost all subsequent items on the survey to instruct respondents on how to report (if standalone, “organization” will be piped; if embedded, “program, division, or department” will be piped). Also new for 2026, the definition for “direct services” (item A4) was revised to better align with the 2023 NCVSP, subject matter expert feedback, and how VSPs typically describe their work. Items that were added to 2023 to identify relationships between VSP main and satellite office locations, as well as whether the VSP’s services were fully virtual, were added to 2026 as well, with minimal changes (see items A7-A9). New instructions explicitly clarify what should and should not be counted as a site (e.g., excluding staff home offices or unofficial locations). 
Victims Served (Section B). Section B captures information on how many hotline, helpline, or crisis line calls the VSP received during calendar year 2026, as well as how many victims the VSP served during the same time period (excluding hotline calls). Outside of removing a few items from this section that were no longer needed (e.g., whether the VSP provided email or mail notification services and whether the VSP collected demographic information about the victims), this section remains similar to 2019. BJS did modify the item that asks about the number of victims who received direct services (B3) to no longer ask about unique victims served, as was done in 2019 (item 3a). Results from cognitive testing indicated that many VSPs did not track the number of unique victims served. 
Crime and Abuse Types (Section C). Section C captures information on the crime and abuse types VSPs provided services for during calendar year 2026. The major change implemented in this section for 2026 was to introduce “yes/no” check boxes for each crime/abuse type to replace the write-in responses that asked VSPs to provide a count of the number of unique victims they served for each crime type in 2019. These items had high levels of missing data for 2019, likely due to their high levels of burden. This change to the question structure substantially decreases the reporting burden on respondents. Smaller changes to this section include consolidating all crime/abuse types into a single grid, adding some crime types (e.g., technology-facilitated abuse), and aligning some crime types (e.g., Kidnapping [Custodial] and Kidnapping [Noncustodial]) with OVC’s Performance Management Tool (PMT) to enhance reporting capabilities. 
Victim Characteristics (Section D). Section D captures information on the sociodemographic characteristics of the victims that VSPs served during calendar year 2026. The major change implemented in this section for 2026 was to introduce a “select all that apply” question structure, where respondents check a box if they provided a service to a victim of that demographic, instead of asking VSPs to enter a count of the unique number of victims of that demographic they served, as was done in 2019. This section had high item and section nonresponse in 2019, likely due to its high levels of burden. This change to the question structure substantially decreases the reporting burden on respondents. Most of the same demographic characteristics are measured in 2026 as were measured in 2019.
Services for Victims (Section E). Section E captures information on the specific services VSPs provided to victims during calendar year 2026. This section retained the same overarching structure it had in 2019, where respondents were asked to indicate, by checking yes or no, whether their VSP provided the specified service during the time period. Item nonresponse in this section in 2019 was low. Changes that were made for 2026 include editing some services to align with the 2023 NCVSP and the PMT, consolidating items (e.g., collapsing “Court accompaniment – civil court” and “Court accompaniment – criminal court” into “Court accompaniment – civil or criminal court” in item E7), and removing or modifying examples that respondents struggled with in cognitive testing (e.g., specifying “nonmedical” in front of “forensic interviews” in item E8, where there was confusion about which type BJS was asking about). Many items, including, but not limited to, the five most common services provided (item E9), locations where off-site services are provided (item E12), 24/7 response coverage availability (item E14), and top referral sources (item E15), were retained for 2026, with minor wording modifications to improve readability. New items include whether the VSP is part of a multidisciplinary team (item E10) and whether services are offered remotely or virtually (item E13). 
Staffing and Resources (Section F). Section F captures information on VSPs’ overall staffing levels, staff training requirements, and high-level information about funding levels. Major changes to this section include removing all items that asked respondents to provide counts of staff in specific roles (e.g., administrative positions) by full- and part-time status, as well as removing the series of questions that asked about the job characteristics of the highest management and most recent direct hire positions. These items had very high levels of missing in 2019 and were very burdensome for VSPs to answer. Instead, in 2026, these items were replaced with a total of four items that ask VSPs to provide counts of the number of paid full-time (item F1a), part-time (item F1b), unpaid volunteers/interns (item F1c), and subcontractors (item F1d) working for the VSP in broad categories only (e.g., all paid staff working full time, all paid staff working part time). Feedback from the subject matter experts and cognitive testing indicated that VSPs could provide this information, especially when provided with the instruction that “estimates are acceptable.” To minimize respondent burden, only two high-level items on funding were added to 2026 – one that asks respondents whether their 2026 victim services funding levels increased, decreased, or remained the same compared to 2025 funding levels (item F4), and the respondent’s level of concern about the amount of victim services funding they expect to be available for their VSP in the future (item F5). 
Mass Violence Incidents (Section G). All items in this section are new for 2026 and replace the 2019 section on Sex Trafficking Victim Characteristics, a special topic for 2019. Feedback from the expert panel indicated that having nationally representative data on VSPs’ mass violence preparedness would be a major benefit to the field. Cognitive testing of these items revealed that not all organizations were prepared to provide victim services in response to mass violence incidents (item G3) and that different organizations used different criteria to define mass violence incidents (item G4), so this section was restructured to ask a few items about the provision of trauma-informed care training to external entities (items G1 and G2) before limiting the rest of Section G to only those VSPs that indicated they are prepared to provide victim services in response to mass violence incidents should they occur in their jurisdiction. 
Usability testing
In December 2026, before launching the NSVSP, the project team will prepare materials for and conduct a small (10 VSPs) usability test to ensure the online and CATI instruments function as intended and are clear to VSPs. This testing was previously approved by OMB under BJS’s generic clearance agreement (OMB No. 1121-0339; exp. 01/31/2028). Specifically, ten VSPs will be asked to complete the web instrument through a guided virtual interview to share their experiences specific to the presentation and programming of the survey. Questions will focus on user perceptions concerning the presentation of questions and response options, where the instrument could be made more user friendly, and other issues related to the general ease of navigating through the survey. This final test will ensure the survey is easy to access and navigate. In addition, the research team will examine paradata for usability indicators such as the overall time spent across all sessions, errors thrown, and help documents accessed. Results from usability testing will inform improvements to the visual design and layout of the survey instrument. The usability testing will not result in any changes to the content of the instrument but will provide an opportunity to modify the design and presentation of the survey to be more user friendly, if needed, protecting against break offs or item skips. 
5. Consultants on the Statistical Aspects of the Design
The Victimization Statistics Unit of BJS takes responsibility for the overall design and management of the activities described in this submission, including data collection procedures, development of the questionnaire, and analysis of the data. Dr. Rachel Morgan is the Victimization Statistics Unit Chief.
ICF is responsible for the collection of all data through a cooperative agreement with BJS (Award #: 15PBJS-24-GK-02863-BJSB). ICF Principal Investigators Samantha Lowry and Kathryn Harper manage and coordinate the NSVSP at ICF. BJS and ICF staff contacts include: 
BJS Staff: 
all staff located at-
999 N. Capitol Street, NE
Washington, DC 20531
202-307-0765
ICF Staff:
all staff located at-
1902 Reston Metro Plaza
Reston, VA 20190
703-251-0368
Kevin M. Scott, Ph.D.
Acting Director
Samantha Lowry
Principal Investigator
Shelley S. Hyland, Ph.D.
Senior Statistical Advisor
Kathryn Harper
Principal Investigator
Heather Brotsos
Deputy Director
Statistical Programs Division
Erica Fissel, Ph.D.
Project Manager
Rachel Morgan, Ph.D.
Chief
Victimization Statistics Unit
Alana Henninger, Ph.D.
Cognitive and Usability Testing Lead and Outreach Lead
Rebecca Bielamowicz, Ph.D. 
Statistician
Victimization Statistics Unit
Ronaldo Iachan, Ph.D.
Sampling Lead