Supporting Statement Part B Ask US Pilot_2_15_legal

SUPPORTING STATEMENT B

U.S. Department of Commerce

U.S. Census Bureau

Ask U.S. Panel Pilot

OMB Control No. 0607-XXXX

B. Collections of Information Employing Statistical Methods

1. Describe (including a numerical estimate) the potential respondent universe and any sampling or other respondent selection method to be used. Data on the number of entities (e.g., establishments, State and local government units, households, or persons) in the universe covered by the collection and in the corresponding sample are to be provided in tabular form for the universe as a whole and for each of the strata in the proposed sample. Indicate expected response rates for the collection as a whole. If the collection had been conducted previously, include the actual response rate achieved during the last collection.

Target Populations

The Ask U.S. Panel (“the Panel”) Pilot currently includes three target populations: adults in residential households, active-duty military in pay grades E1-E4, and non-military spouses of active-duty military personnel in pay grades E1-E4. Future efforts may focus on the inclusion of other specialized subpopulations. Samples for the Panel will be selected independently from a sampling frame constructed for each target population.

Persons residing in residential households are eligible for recruitment into the Panel given the following criteria:

• Non-institutionalized residential population within the contiguous United States and Washington, DC.

• Fluent in either English or Spanish (assumed eighth grade reading level at minimum)

• Adults aged 18 years of age and older

Sampling is restricted to the contiguous U.S. states (i.e., excludes Hawaii and Alaska).

Additionally, a random sample of active-duty military and non-military spouses of active-duty military members living within the 50 United States and Washington DC will be selected. Additional eligibility criteria include:

• Active-duty military, E1 – E4 only

• Activated for more than 4 months (military service member, member of non-military spouse)

• Not in the Coast Guard / Reservists

• Not members of the general population Ask U.S. Panel

A panelist would be no longer eligible for the panel if:

• Separated from the military after sample selection (military service member, member of non-military spouse)

• Non-military spouse became legally separated, divorced, or widowed from member

Residential Household Population: Three-stage Stratified Probability-based Sampling Design and Sample Size for Pilot

The Pilot sampling design is clustered to enable cost-effective evaluation of an in-person nonresponse follow-up (NRFU) on a subsample of nonresponding households (HHs) and randomly chosen adults within HH. If the NRFU is shown ineffective or cost prohibitive and no other in-person activities are required (e.g., data collection, tablet assistance), then we will evaluate the utility of a single-stage stratified design for a recommendation for the main Panel sample.

Stage 1: Census Block Groups (CBGs) and Clusters of Geographically Small CBGs Within Strata

We will select a stratified probability proportional to population size sample of CBGs where the measure of size is total HH population based on the most recently available American Community Survey (ACS) 5-year estimates. Prior to selection, we will create CBGs of roughly equal size by splitting large CBGs and collapsing small CBGs to equalize workload and to improve efficiency of the estimates. All CBGs will contain at least 350 addresses to support the sampling activity.

We will classify all CBGs into one of four mutually exclusive geographic design strata defined by the following characteristics from the ACS or other sources:

• High versus low internet penetration¹

• High versus low Hispanic/Latino concentration

We will evaluate the inclusion of U.S. region (Northeast, Midwest, South, and West) as an additional stratifier based on the distribution of CBGs within the currently proposed strata.

Table 1 shows the proposed distribution for 20 pilot sample CBGs by the two first-stage stratification variables—internet penetration and concentration of the Hispanic/Latino population. We desire a higher proportion of CBGs with these historically “hard to interview” adults (see, e.g., Perrin and Atske 2021; Kelley et al. 2020; Jang and Vorderstrasse 2019; Brown 2015) for the pilot to test panel recruitment procedures—for example, 80% of the sampled CBGs will be selected from low internet penetration areas.

Table 1. Distribution of Sampled CBGs by First-stage Strata: Ask U.S. Panel, Pilot Recruitment

Internet Penetration	HL Concentration		Total	Pct of CBGs
	Low	High
Low	6	10	16	80.0
High	2	2	4
Total	8	12	20
Pct of CBGs		60.0

CBG = census block group; HL = Hispanic / Latino

Stage 2: Residential Addresses within CBG

We will select a systematic sample of approximately 240 addresses from each Stage-1 CBG frame derived from an address-based sampling frame. We will sort the CBG-specific sampling frame by postal-delivery sequence number prior to selection so that the resulting sample will contain geographic representation from across the CBG.

Table 2 shows the expected response distribution for 4,800 selected pilot residential addresses. We will select almost 1,000 additional addresses to be released only if response rates are lower than anticipated.

Table 2. Expected Distribution Recruitment Protocol and Sample Group: Ask U.S. Panel, Pilot

Recruitment Protocol	Assumed Rates, %	Sampled Addresses
		Main	Hold (20%)a	Total
Phase I: Mail invitation
Sampled Addresses		4,800	960	5,760
Mail delivered	92	4,416	883	5,299
Completes HH roster	19	848	170	1,018
Phase II: FTF follow-up with NRs
Sampled Addresses	50	1,784	357	2,141
Mail delivered	88	1,570	314	1,884
Completes HH roster	52	816	163	979

HH = household

^a Additional addresses selected and released only if needed given results from the main sample recruitment and differential participation by CBG.

Stage 3: At Most Two Adult Residents per Household

A HH respondent is asked to complete a HH roster by web or telephone containing the first name and age of each adult member of the HH. From this roster, the system will randomly select up to two adult residents per HH using a simple random sampling methodology; each selected adult is asked to complete a baseline questionnaire.

Table 3 shows the expected response distribution of adults selected from pilot HHs with a completed screener. We anticipate sampling almost 1,700 adults to yield just approximately 1,500 enrolled panelists from the pilot design.

Table 3. Expected Distribution of Adults by Recruitment Protocol: Ask U.S. Panel, Pilot

Recruitment Protocol	Assumed Rates, %	Number of Adults
Phase I: Mail invitation
Sampled adults		1,416
Responds to baseline, enrolls		679
First sampled person responds	60	509
Second sampled person responds	30	170
Phase II: FTF follow-up with NRs
Sampled adults		1,363
Responds to baseline, enrolls		872
First sampled person responds	80	653
Second sampled person responds	40	219
Total adults recruited to the panel		1,551

FTF = face to face; NR = nonrespondents (households, sampled adults)

Military Populations: Single-stage Stratified Probability-based Sampling Design and Sample Size for Pilot

Independent single-stage stratified Pilot samples of active-duty military in paygrades E1-E4 and non-military spouses of active duty E1-E4 service members will be selected independently from sampling frames derived for each branch of the military (Army, Navy, Marine Corps, and Air Force). Sampling strata within branch will include sex, race/ethnicity, marital status (active-duty military sample only), and having dependents. Table 4 shows the initial sample size given expected response rates for each military sample.

Department of Defense (DoD) personnel will generate frame counts by sampling strata from military and non-military spouse administrative records to determine final sample allocation. DoD personnel will select the samples using a stratified simple random sample and provide identifying information, demographics, etc. via an approved secure transfer protocol based on specifications for Panel recruitment.

Table 4. Expected Distribution of Participants by Sample: Ask U.S. Panel DoD Component, Pilot

Recruitment Protocol	Assumed Rates, %	Count
Active-Duty Military Sample
Sampled personnel, E1-E4		1,756
Responds to baseline, enrolls	6	100
Via initial invitation	80	80
Via nonresponse follow up	20	20
Non-Military Spouse Sample
Sampled spouses		700
Responds to baseline, enrolls	14	100
Via initial invitation	80	80
Via nonresponse follow up	20	20

2. Describe the procedures for the collection of information including:

   • Statistical methodology for stratification and sample selection,

   • Estimation procedure,

   • Degree of accuracy needed for the purpose described in the justification,

   • Unusual problems requiring specialized sampling procedures, and

   • Any use of periodic (less frequent than annual) data collection cycles to reduce burden.

Residential Household Population: Research Questions

Several unknowns at the Panel design stage necessitate the launch of a pilot to inform key design features related to coverage and nonresponse. Specifically, the Pilot will have three research foci:

1. How to optimize panel recruitment

2. How to deal with non-internet users (estimated to be about 7% of the adult U.S. population in 2020, according to Pew Research Center)

3. How to minimize nonresponse bias

To address the first research question, we plan a sponsorship experiment crossed with incentive visibility (see Table 3.1). To address the second research question, we plan on observational studies to examine the success of an internet-enabled tablet offer. Similarly, the third research objective will also be addressed through an observational study that would examine the pool of respondents brought in via computer-assisted telephone and personal interview (CATI) and computer-assisted personal interview (CAPI) and whether weighting adjustments can eliminate potential coverage or nonresponse bias.

Residential Household Population: Recruitment

The experimental component of the Pilot will involve a 2x2 design that focuses on sponsorship (explicit U.S. Census Bureau involvement vs. a general panel name) and visibility of a $5 prepaid incentive mailed with the initial recruitment letter via USPS first class mail (visible through a windowed envelope vs. not visible and mailed in a regular envelope). Table 5 outlines the design and sample size in each cell.

Table 5. Pilot Sample Member Allocation by Experimental Condition from Residential Household Population

Sponsorship	Prepaid $5 Incentive
	Visible	Non-visible
Ask U.S. Panel, U.S. Census Bureau and Research Partners	1,200	1,200
Ask U.S. Panel	1,200	1,200

The three-stage stratified probability sampling design described under response to item 1 above will allow us to detect 6% difference in response rate (about 3% in the marginals) at alpha = 0.05 and 80% power.

Pilot enrollment will consist of two phases and will involve three recruitment modes (web, inbound CATI, and face-to-face):

Phase 1—Initial Recruitment

We will mail a survey invitation with $5 prepaid incentive to all sampled addresses. The letter will contain a unique link to the web survey and a phone number for inbound calling. Respondents who choose web will complete the Household Roster on their computer or smartphone. Those who choose to complete via phone will call into RTI’s 240-station call center which will be staffed Monday - Saturday from 9 AM ET – 11 PM ET and Sundays from 12 PM ET-8 PM ET.

If the roster respondent is selected, they will move directly to the Baseline Questionnaire. If another adult from the HH has been selected, they will be mailed an invitation to complete the Baseline Questionnaire with a web link and an inbound CATI phone number.

One week after the initial survey invitation, nonresponding cases will be mailed a brochure with web link and inbound CATI number. One week later, nonresponding cases will receive a postcard reminder (unique ID; web link to Household Roster and a phone number for inbound CATI). A final mailing of the survey invitation with web link and inbound CATI will be sent a week after the after the postcard reminder.

Phase 2—Face-to-face Nonresponse Follow-up

As a final effort to convert nonrespondents, we will subsample 50% of the remaining cases for face-to-face data collection. We will allocate four visits per address, but allow for four additional visits for a total of eight possible visits for HHs with members in subgroups of interest based on the actualized distribution in Phase 1.

At the HH roster level, a resident will complete the HH roster to list all eligible members. An online algorithm will randomly select at most two adults 18 years of age or older per address, who speak either English or Spanish. Sampled adults will be asked to complete the Baseline Questionnaire via web or inbound CATI. Remaining nonrespondents will be subsampled at 50% for face-to-face follow-up. As part of the observational study on other modes of data collection, we will evaluate the utility and cost-effectiveness of this approach.

We will classify internet access at the person rather than HH level and eligibility for a tablet offer will be determined based on measures of internet affinity, and frequency/consistency of internet access (included in the Baseline Questionnaire). Respondents who complete the panel enrollment via inbound CATI and do not have reliable internet access (i.e., non-internet HH) will be mailed a study tablet with detailed instructions on how to activate it. Those, who are part of face-to-face NRFU and do not have internet access will also be provided study tablets. We will assess the utility of this approach by tracking calls to the help desk, device activation rate, and completion of the first topical surveys. We will model the likelihood of remaining in the panel for tablet respondents as a function of their demographic characteristics, assistance required to set up the tablet and topical survey completion, as part of the evaluation of the feasibility of a tablet offer for the main panel recruitment.

Once recruited for the pilot, panelists will participate via web. We will send each enrolled pilot study panel member an invitation to complete a 15-minute web (topical) survey 2-4 weeks after enrollment. The contacting protocol for the topical surveys is described in Section IV, 1.4).

Military Populations: Research Questions

As with the residential component, several unknowns currently exist for the military populations under study (active duty and non-military spouses) the necessitate a pilot to inform key design features related to coverage and nonresponse. Specifically, the military Pilot will have two primary research questions:

1. How to optimize panel recruitment

2. How to minimize nonresponse bias

Military Populations: Recruitment

The Ask U.S. DoD Panel consists of two independent target populations living in the 50 States and the District of Columbia:²

1. Active-duty military, E1 – E4 only

2. Non-military spouse of active-duty military members (excluding those legally separated but not divorced)

Panel recruitment targets for the Pilot are as follows with anticipated equal distribution by military branch (Army, Navy, Marine Corps, and Air Force):

• ~75 non-married active-duty members

• ~25 married active-duty members

• ~100 non-military spouses

Table 6 displays the proposed pilot study marginal target allocation by stratification variables. Table 7 displays the current marginal sample sizes for selection to accommodate targets.

Table 6. Proposed Pilot Study Recruitment Targets by Population and Stratification Variable

Strataa	US Active Duty, Not Married			US Active Duty, Married			Active Duty Overall	Non-Military Spouses
	pctb	Targetsc	DoD	pctb	Targetsc	DoD		pctb	Targetsc	Overall
Total		17	68		8	32	100		25	100
Sex
Male	0.70	12	48	0.5	4	16	64	0.20	5	20
Female	0.30	5	20	0.5	4	16	36	0.80	20	80
Race/Eth
Hispanic	0.35	6	24	0.40	3	12	36	0.30	8	32
NH Black	0.35	6	24	0.40	3	12	36	0.30	8	32
NH Other	0.30	5	20	0.20	2	8	28	0.40	9	36
Dependents
No	0.90	15	60	0.50	4	16	76	0.50	13	52
Yes	0.10	2	8	0.50	4	16	24	0.50	12	48

^a Strata defined by the intersection of all stratum characteristics.

^b Percent distribution by category.

^c Target number enrolled by service (USA, USN, USMC, USAF) within the 50 states and DC.

Table 7. Proposed Pilot Study Sample Size by Population and Stratification Variable

Strataa	US Active Duty, Not Married			US Active Duty, Married			Active Duty Overall	Non-Military Spouses
	pctb	Targetsc	DoD	pctb	Targetsc	DoD		pctb	Targetsc	Overall
Total		299	1,196		140	560	1,756		175	700
Sex
Male	0.06	211	844	0.06	70	280	1,124	0.15	35	140
Female	0.06	88	352	0.06	70	280	632	0.15	140	560
Race/Eth
Hispanic	0.06	105	420	0.06	53	212	632	0.15	56	224
NH Black	0.06	105	420	0.06	53	212	632	0.15	56	224
NH Other	0.06	88	352	0.06	35	140	492	0.15	63	252
Dependents
No	0.06	263	1,052	0.06	70	280	1,332	0.15	91	364
Yes	0.06	35	140	0.06	70	280	420	0.15	84	336

^a Strata defined by the intersection of all stratum characteristics.

^b Unweighted response rate.

^c Target number enrolled by service (USA, USN, USMC, USAF) within the 50 states and DC.

To join the panel, members must complete the baseline questionnaire and agree to become panel members. The baseline questionnaire is expected to take approximately 20 minutes for sampled respondents to complete. The demographic and background information collected in this questionnaire that is not readily available from administrative data will provide important benchmarks for subsequent analyses, including examination of characteristics of topical survey nonrespondents and panel members who attrit over time. The baseline questionnaire will also collect detailed contact information (e.g., telephone and email for military spouses) and permission to send text messages for survey prompting and nonresponse follow-up.

We will use a two-phase design to recruit participants to the DoD samples:

Phase 1—Initial Recruitment of Active-Duty Military and Non-military Spouse Sample Members (Mail or Email)

We will mail the survey invitation using a 9x12 envelope that contains a $5 prepaid incentive, a letter, brochure, and informed consent document. The letter will contain a unique web link to the baseline questionnaire and will mention the additional $20 incentive for those completing the baseline questionnaire and agreeing to enroll in the panel. After one more week, we will send a self-mailer (made from folded, printed cardstock) containing the unique web link to nonrespondents. After one more week, we will mail a second packet to all nonrespondents. The contents of the second packet will be similar to the initial invitation but will not include the prepaid incentive.

Phase 2—Nonresponse Follow-up of Active-Duty Military Sample Members

Interviewers will call a random subsample of nonresponding military personnel for which we have telephone numbers. We will determine the subsample size by design stratum based on the availability of telephone numbers and likely differential enrollment by the Pilot sample members. Interviewers will make up to five attempts to reach the select sample member.

Phase 2—Nonresponse Follow-up of Non- Military Spouse Sample Members

Owing to limited telephone numbers, the nonresponse follow-up will not include calls from trained interviewers. Instead, priority mailing or equivalent will be used to follow up with a subsample of nonresponding non-military spouses.

Similar to the general population sample for the Ask U.S. Panel, all sampled cases who consent to be enrolled in the panel and complete the baseline questionnaire will receive a $20 incentive. Respondents will have the option to receive a mailed a check, cash, or a gift card (electronic or physical). We expect that these procedures will result in 209 completed baseline questionnaires across the two samples. These cases will be the DoD Pilot Panel members.

Assumptions for Inflation of Targets to Yield Pilot Study Sample Sizes

Sample targets are inflated for sample loss to determine the marginal sample size by target population. The actual pilot rates, in combination with information from other surveys, will inform the inflation rates for future studies. Assumed rates are shown in the tables above.

3. Describe methods to maximize response rates and to deal with issues of non-response. The accuracy and reliability of information collected must be shown to be adequate for intended uses. For collections based on sampling, a special justification must be provided for any collection that will not yield "reliable" data that can be generalized to the universe studied.

The Ask U.S. Panel will include a multimode contacting approach, relying heavily on mail mode contacts for panel recruitment, and adding email and text message contacts for panel maintenance and topical surveys. This approach will maximize our ability to enroll and maintain contact with the panel.

Residential Household Population

The Pilot Panel Recruitment phase will provide an opportunity to test the contacting protocol that will support the recruitment of the Ask U.S. Panel. This phase of the project will begin with a 4-week web data collection period that will launch with the mailing of the survey invitations with a $5 prepaid incentive to sampled addresses. HH residents will be invited to visit the Ask U.S. Panel website to complete the HH roster online. Those who complete the roster within 1 week will receive additional $5 (early bird incentive). Once rostering is completed, the selected adult can complete the baseline survey via web. If a second person from the HH is selected, a second survey invitation will be sent to that person. Rostered adults who complete the baseline survey will be sent $20 via check, cash, or digital gift card. All mailing materials will include a phone number to a toll-free project number that they can call to complete the roster or the baseline questionnaire with a telephone interviewer. Mailings will include a QR code that will take respondents to the Ask U.S. Panel log-in page, where respondents will enter the passcode from their invitation letter.

Once the initial 4-week data collection period is complete, a subsample of nonresponding cases will proceed to field data collection (NRFU).

• Survey Invitation Packet: A survey invitation letter, informed consent, and $5 prepaid incentive mailed via USPS First Class in 6x9 envelope. The survey invitation will include a QR code to the HH roster.

• If a second HH member is selected for participation in the panel, a separate survey invitation packet will be sent to that person.

• Nonresponse Reminder Packet: A nonresponse reminder letter and information brochure sent via USPS First Class in a 6x9 envelope. The nonresponse reminder letter will include a unique link to the HH roster.

• Thank you/Reminder Self-mailer: A self-sealed postcard that will include a unique link to the HH roster.

• Final Reminder Letter: A final reminder letter that will be mailed via USPS First Class in 6x9 envelope. The final reminder letter will include a unique link to the HH roster.

Topical Survey

During the pilot panel recruitment phase, enrolled panel members will be asked to complete one topical survey. The contact for that topical survey will consist of a single self-sealed postcard that will include a link to the panel member’s survey and two e-mail and two text invitations.

The 4-week topical study data collection period will run concurrently with the panel development.

• Week 1—Panel members will receive an email invitation that will include a personalized link to the survey. This will be followed 3 days later by a text message prompt.

• Week 2—Panel members will receive a reminder email that will include a personalized link to the survey, followed 3 days later by a second text message prompt.

• Week 3—Panel members will receive a reminder self-mailer that will include a link to the survey, followed 3 days later by a third text message prompt.

• Week 4—Panel members will receive a final email reminder that will include the personalized link to the survey.

Military Populations

We will use a two-phased design to recruit participants to the DoD oversample:

• Phase 1—Initial Recruitment of Military and non-military spouse Sample Members (Mail or Email)

• Phase 2—Nonresponse Follow-up of Active-Duty Military Sample Members

• Phase 2—Nonresponse Follow-up of Non- Military Spouse Sample Members

Summary information is provided in Table 8 with additional details shown in response to Task B, Item 2.

Table 8: Responsive design approach to DoD oversample participant recruitment:

Phase	Contact	Description
1	1	Mail initial survey invitation with $5 a
	2	Self-mailer and Email (if applicable) reminder
	3	Mail 2nd survey invitation
2	4-8	Up to five call attempts with random subsample of Phase 1 nonrespondents b

^a Request for approval of active-duty member incentives is forthcoming.

^b Those without phone numbers will be classified as nonrespondents.

Sampled persons who complete the baseline questionnaire and consent to enroll in the panel are classified as panel members (see panel member targets).

Topical Survey

RTI will conduct one 15-minute topical survey with members of the DoD Pilot Panel. The topical survey will be administered in the web mode only to reduce measurement error that a multimode approach could introduce. Respondents will receive a $10 incentive upon survey completion as a check, cash, or a gift card.

We will send invitations to the topical survey via an email containing a unique link to the survey. We will send invitations via email and text message spaced approximately 2-3 days apart and sent on days and times targeted for maximum effectiveness (see Exhibit 1). RTI will staff a help desk with trained staff for panel members who have difficulties with or questions about the links, the panel website, or the survey instrument.

Similar to the main panel, the proposed incentive structure for the DoD Pilot includes:

• $5 prepaid incentive in the survey invitation package

• $20 promised baseline incentive for completion of a 20 min baseline survey

• $10 promised topical survey incentive for completion of a 15 min survey

Thus, the total incentive amount an eligible sample member can receive for participation in the DoD Pilot is $35 for a total of 35 min participation time.

Respondents will have the option to receive an electronic gift card, a physical gift card, cash, or a mailed check. Those who opt for the gift card will be able to use the incentive immediately. The mailed incentive options ensure that respondents with limited internet use will be able to use their earned incen­tive.

4. Describe any tests of procedures or methods to be undertaken. Testing is encouraged as an effective means of refining collections of information to minimize burden and improve utility. Tests must be approved if they call for answers to identical questions from 10 or more respondents. A proposed test or set of tests may be submitted for approval separately or in combination with the main collection of information.

This package is for a test of the full Panel development methodologies. Embedded experiments are described above.

5. Provide the name and telephone number of individuals consulted on statistical aspects of the design and the name of the agency unit, contractor(s), grantee(s), or other person(s) who will actually collect and/or analyze the information for the agency.

6. Survey Design:

7. Jason Fields

8. Social Economic and Housing Statistics Division

9. Demographic Programs Directorate

10. [email protected]

12. Jennifer Hunter Childs

13. Center for Behavioral Science Methods

14. Associate Director Research and Methodology

15. [email protected]

16. Statistical Design

17. Jill A. Dever

18. RTI International

19. [email protected]

21. Emilia Peytcheva

22. RTI International

23. [email protected]

25. Stephanie Eckman

26. RTI International

27. [email protected]

29. Data Collection

31. Kim Aspinwall

32. RTI International

33. [email protected]

References

Brown, A. (2015). The Unique Challenges of Surveying U.S. Latinos. Pew Research Center Report, 2015 Nov 12. Accessed 2022 Jan 27 at https://www.pewresearch.org/methods/2015/11/12/the-unique-challenges-of-surveying-u-s-latinos/.

Jang, M. and Vorderstrasse, A. (2019). Socioeconomic Status and Racial or Ethnic Differences in Participation: Web-Based Survey. JMIR Res Protocol 2019;8(4):e11865. doi: 10.2196/11865

Kelley, A., McCoy, T., Fisher, A., Witzel, M., Fatupaito, B. and Restad, D. (2020) "Comparability of Survey Measures in Hard to Reach Populations: Methods and Recommendations," Practical Assessment, Research, and Evaluation: Vol. 25, Article 9. Available at: https://scholarworks.umass.edu/pare/vol25/iss1/9

Perrin, A. and Atske, S. (2021). 7% of Americans don’t use the internet. Who are they? Pew Research Center Report, 2021 Apr 2. Accessed on 2022 Jan 27 at https://www.pewresearch.org/fact-tank/2021/04/02/7-of-americans-dont-use-the-internet-who-are-they/.

1 Sources for internet penetration may include the following:

• Estimates from ACS Table ID: B28002, https://data.census.gov/cedsci/table?q=B28002&tid=ACSDT1Y2019.B28002&hidePreview=false

• Areas with no/slow internet speeds from https://www.ntia.gov/press-release/2021/ntia-creates-first-interactive-map-help-public-see-digital-divide-across-country or https://broadbandusa.ntia.doc.gov/sites/default/files/2021-06/Indicators%20of%20Broadband%20Need%20-%20User%20Guide.pdf

2 Location is determined at the time of sampling. If active-duty personnel are stationed overseas when recruited, they will remain eligible for panel enrollment.