0704-RPSS_SSB_combined_final_RR

SUPPORTING STATEMENT – PART B

Recruiting, Retaining, and Promoting Servicemembers who Identify as Hispanic or Latino/Latina

Focus Groups and Survey – OMB Number: 0704-RPSS

B.  COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS

1.  Description of the Activity

This is a new data collection intended to identify and address gaps in DoD’s understanding of the recruitment, retention, and career progression of Hispanic Servicemembers; evaluate past efforts and initiatives by DoD and the Services aimed at increasing Hispanic representation and the challenges associated with those efforts and initiatives; and identify key lessons learned or best practices that future initiatives should follow. As part of this study, CNA will conduct Service-specific in-person and virtual focus groups with Active Duty Servicemembers in the Marine Corps, Army, Navy, Air Force, and Space Force and will work with a subcontractor to conduct a survey among eligible members of the general Hispanic population who are registered to vote in the United States.

Focus groups: There will be no sampling methods employed in the collection of focus group data in this study. Those who are eligible to participate in these focus groups are Active Duty Servicemembers from selected units who fall under the following categories within the Air Force, Army, Navy, Marine Corps, and Space Force:

• Junior enlisted (E1-E2s): approximately 40 per Service

• First term enlisted approaching reenlistment decision (E3-E4s): approximately 30 per Service

• Recruiters (E4-E7s): approximately 10-20 per Service, depending on the Service (see table on next page)

• Junior officers (O1-O3s): approximately 40 per Service

For in-person focus groups, participating installations will be selected by the CNA study team in collaboration with the study sponsor and study POCs. Eligible focus group participants from the participating installations will be identified through the study’s designated installation POCs (POCs will send the CNA study team a list of eligible Servicemembers from their installations through secure means, e.g., DoD SAFE). CNA will email with eligible focus group participants to recruit those who are interested in voluntarily participating and assign them to specific in-person focus groups (see Supporting Statement A for additional detail about the recruitment process).

For virtual focus groups, the participating Services will appoint Service-level POCs who will help identify eligible Servicemembers to participate in virtual focus groups. The Service-level POCs will send a heads-up email to eligible participants, which will request that interested Servicemembers email the CNA study team. The CNA study team will follow up with eligible and interested Servicemembers to recruit and assign them to specific virtual focus groups (see Supporting Statement A for additional detail about the recruitment process).

The following table shows the expected number of focus group participants from each of the categories listed above across the Services:

	Army	Air Force	Navy	Marine Corps	Space Force	TOTAL
Junior enlisted (E1-E2s)	40	40	40	40	40	200
First term enlisted approaching reenlistment decision (E3-E4s)	30	30	30	30	30	150
Recruiters (E4-E7s)	20	20	20	20	101	90
Junior officers (O1-O3s)	40	40	40	40	40	200
TOTAL	130	130	130	130	120	640

Survey: The potential respondent universe for this study comprises adult Hispanics/Latino over 18 years of age who are identified through publicly available U.S. voter registration records. The population of Hispanic/Latino eligible voters in the United States is approximately 36.2 million, of which approximately 12.3 million (33.4%) are registered to vote.

The sampling method to be used is a stratified random sample considering age, gender, national origin, and place of residence to ensure representation across various demographic factors within the U.S. Hispanic/Latino population. Given our experience with FIU’s National Hispanic Survey, we anticipate a response rate of 30 percent (1,500 participants).

State	Hispanic/Latino Pop. eligible to vote	% of Hispanic/Latino eligible voter pop.	% of total sample	Sample (MOE) 2.53 at 95%	MOE-state at 90%
California	8,500,000	33%	31.29	468	4.53
Texas	6,465,000	32%	23.8	357	5.19
Florida	3,515,000	22%	12.94	194	5.19
New York	2,170,000	15%	7.98	119	8
Arizona	1,350,000	25%	4.97	75	10
Illinois	1,195,000	13%	4.39	66	10
New Jersey	1,060,000	16%	3.9	59	11
Colorado	755,000	17%	2.77	42	13
New Mexico	695,000	45%	2.5	38	13
Georgia	485,000	6%	1.78	27	16
Nevada	485,000	22%	1.78	27	16
Massachusetts	485,000	9%	1.78	28	16
Total Pop.	27,160,000		99.88	1500
To achieve a margin of error (MOE) of not less than 5.5% for each state with a 95% confidence level, each state would require a sample size of approximately 317 respondents, regardless of the population size. The margin of error formula, with the finite population correction applied, tends to converge to a similar sample size when the population is large, and the desired margin of error is relatively large as well. ​​We would need to draw a sample size of 3,804. If needed, we can adjust the sample size per state so that the samples are not as small, and MOE is reduced to more acceptable levels.

2.  Procedures for the Collection of Information

1. Statistical methodologies for stratification and sample selection;

Focus groups: We are not stratifying the sample for collection of information.

Survey: In the collection of information about Hispanic/Latino adults in the United States, stratification and sample selection techniques will be employed to ensure that the sample accurately represents the diverse demographics of Hispanic/Latino adults in the United States across different regions, age groups, genders, socioeconomic statuses, national origin, and other relevant factors. Stratification involves dividing the population into homogeneous subgroups (strata) based on certain characteristics, and then sampling from each stratum proportionally to ensure adequate representation. This will help to minimize sampling bias and improve the generalizability of the survey results to the broader population of Hispanic/Latino adults in the United States.

2. Estimation procedures;

Focus groups: We do not plan on employing estimation procedures.

Survey: In this particular study we will use standard estimation procedures. We will examine sample proportions as estimators of specific variables. We will also use confidence intervals (in this study 95% with an MOE of 2.53 for the entire sample and a 90% confidence interval with an MOE ranging upwards depending on the state and the sample size).

3. Degree of accuracy needed for the Purpose discussed in the justification;

Focus groups: No estimation procedures are being used.

Survey: Determining the degree of accuracy involves considering primarily the sample size, margin of error and confidence level. These factors help quantify the survey's precision and how well it represents the views of the broader Hispanic/Latino population in the context of our survey’s objectives. We are calculating our degree of accuracy based on a sample size of 1500, a confidence level of 95% and a margin of error of 2.53%. We are also assuming that margins of error for subgroups will be larger than for the total sample. (Please refer to table provided above.)

4. Unusual problems requiring specialized sampling procedures; and

Focus groups: There are no unusual problems requiring specialized sampling procedures.

Survey: We do not anticipate unusual problems that will require specialized sampling procedures. However, if certain segments of the U.S. Hispanic/Latino adult population are difficult to reach or underrepresented in traditional sampling frames (such individuals with limited English proficiency), we will employ innovative sampling methods, such as respondent-driven sampling, to ensure adequate representation. We do not anticipate this situation to occur. Additionally, as with any survey, concerns about non-response bias or sample coverage may require us to implement weighting adjustments or post-stratification techniques to address these issues.

5. Use of periodic or cyclical data collections to reduce respondent burden.

Focus groups: The focus groups for this study will be conducted one time with each of the participating Services.

Survey: The survey for this study will be conducted one time.

3.  Maximization of Response Rates, Non-response, and Reliability

Focus groups: CNA will rely on the study’s designated installation and Service-level POCs to identify eligible focus group participants, and then CNA will conduct the recruitment efforts to help maintain participants’ confidentiality and to help prevent any Servicemembers from feeling coerced to participate in focus groups.

To help focus group participants remember to attend their specific focus group sessions, the CNA study team will email calendar invites to in-person and virtual focus group participants (CNA will ask focus group participants to confirm the correct email address to use for the calendar invite to ensure that the participants receive it at the email address they prefer for the calendar invite).

We plan to supplement the qualitative data we collect through our focus groups with quantitative data from other sources, such as JAMRS. Recognizing that our focus group sample will not be sufficient to make our findings generalizable to the total population, we will use quantitative sources (such as JAMRS) to supplement our findings. The focus group inputs will provide additional information that can only be collected using qualitative means; they will provide context for many JAMRS findings. In other words, the use of the quantitative data (JAMRS) along with the qualitative data (focus groups) will be complimentary rather than duplicative: the quantitative data provide findings on the “what” and the focus groups will provide qualitative data that we are hoping will help explain the “why” behind those responses. Additionally, we may consult the JAMRS data after collecting focus group data to identify and analyze any themes that emerge consistently from both data sources.

Survey: To maximize response rates and address instances of non-response in a survey of 1,500 eligible Hispanic/Latino adults in the United States, we will employ the following methods:

• Clear and Concise Communication: We will ensure that survey invitations and questions are clear, concise, and culturally sensitive to the Hispanic/Latino population. We will use bilingual materials to accommodate respondents who may prefer Spanish.

• Multiple Modes of Data Collection: To gather data for our survey, we will implement a multifaceted approach to ensure a diverse range of responses from the U.S. Hispanic/Latino adult demographic. Initially, we will employ ASISA's Computer Assisted Telephone Interview (CATI) system to reach out to a randomly selected sample of adults who identify as Hispanic/Latino in publicly available U.S. voter registration records for the phone-based, CATI version of the survey (please see the Supporting Statement A for more detail on the CATI procedures). Should the initial call go unanswered, we will attempt a second call and leave a voicemail providing a callback number for the recipient's convenience.

If ASISA is unable to recruit the full survey sample using the CATI methodology, FIU will work through the online survey platform Qualtrics to field the survey online. Through this method, Qualtrics will field the online survey through survey panels made up of eligible individuals who provide their email addresses in the publicly available U.S. voter registration records. The online Qualtrics sample will be a new sample that is different from the sample used through the CATI system. This new Qualtrics sample will not include those who failed to respond to through the CATI system. Upon accessing the online survey through the provided URL, eligible individuals will be provided with information about the study and their participation in the survey as part of the study and will be required to provide their informed consent prior to beginning the survey. Each eligible individual will receive a unique URL that will allow them to each complete and submit the survey once.

Should these measures still fall short of our sample size goal, we will proceed to draw an additional random sample of eligible participants from the publicly available U.S. Hispanic/Latino voter registration records and reiterate the outlined process. By offering a combination of CATI and online surveys, we aim to cater to different preferences and levels of access to technology, thereby maximizing response rates and ensuring a comprehensive dataset.

• Incentives: we will not provide monetary or other incentives.

• Targeted Recruitment Materials: For the Qualtrics online version of the survey, we will email separate survey invitations and follow-up reminders to each eligible individual (i.e., not mass emails) to make respondents feel valued and more likely to participate.

• Follow-up: We will implement a systematic follow-up procedure for non-respondents, which may include: one phone call and a voice mail message for the CATI version of the survey and a follow-up email invitation for the Qualtrics version of the survey.

• Establishing Trust: We will clearly communicate the purpose of the survey, how the data will be used, and assure respondents of confidentiality. Building trust encourages participation.

To ensure the accuracy and reliability of responses, we will employ the following methods:

• Random Sampling: We will ensure that the sample of 1,500 Hispanic/Latino adults is randomly selected from the population of interest. Random sampling helps to reduce bias and allows for generalizability of findings to the broader Hispanic/Latino adult population in the United States.

• Filter questions: Before receiving the first survey question, each respondent will be asked two filter questions to ensure that only those who are eligible can participate in the survey (the filter questions ask whether respondents identify as Hispanic/Latino and whether they are at least 18 years of age).

• Targeted Recruitment Materials: As mentioned above, each eligible individual included in the sample(s) for the Qualtrics online survey will receive a unique URL that will allow them to each complete and submit the survey once. Use of unique URLs sent through targeted recruitment emails also ensures that only those who are eligible can access the survey.

• Data Cleaning: We will implement procedures to clean and verify the data collected, such as removing incomplete responses (defined as those completing less than 75 percent of the survey) that may skew results.

• Statistical Analysis: We will utilize appropriate statistical techniques to analyze the data, ensuring that conclusions drawn from the survey results are robust and reliable. Specifically, we will stratify our sample by gender, state, age group, and income level, to ensure it is representative of the U.S. population as represented in the American Community Survey. After cleaning the final dataset of respondents who completed at least 75 percent of the survey, we will run frequencies, generate cross tabs, conduct t-tests (with associated margins of error), and conduct regression analysis.

• Expert Review: We have consulted with FIU experts in survey methodology and Hispanic/Latino demographics who reviewed the survey instrument and methodology to ensure validity and reliability. The survey methodology, procedures, and materials will also go through DoD expert review as part of the PRA/OMB review process.

4.  Tests of Procedures

Focus groups: Focus groups questions were developed by CNA and reviewed by the sponsor to ensure that all questions are appropriate and useful for their intended purposes.

Survey: To ensure the effectiveness and efficiency of data collection procedures for the random sample of 1,500 eligible Hispanic/Latino adults in the United States, we will undertake the following tests of procedures and methods:

• Usability Testing: We will assess the usability of the data collection tools being used. We work closely with our vendor to ensure that our needs are met. We will ensure that the platform is user-friendly and accessible to respondents.

• Response Time Analysis: We will monitor the time taken by respondents to complete the survey.

• Interviewer Training: We will provide comprehensive training to ensure consistency in administration and data collection procedures.

By conducting these tests of procedures and methods, we will reduce respondent burden, and improve the overall utility and effectiveness of data collection for the sample of 1,500 Hispanic/Latino adults in the United States.

5.  Statistical Consultation and Information Analysis

a. Provide names and telephone number of individual(s) consulted on statistical aspects of the design.

Focus groups: There was no consultation on the study design outside of the CNA team and the sponsor.

Survey:

Eduardo A. Gamarra, Professor, Department of Politics and International Relations and Director of the Latino Public Opinion Forum, FIU (305) 348-1718

Jose Miguel Cruz, Director of Research, Latin American and Caribbean Center, FIU (305) 348-2894

Dr. Guillermo Grenier, Professor, Department of Global & Sociocultural Studies, Florida International University (305) 348-3217

Qing Lai, PhD, Associate Professor of Demography and Sociology, Graduate Program Director, Department of Global & Sociocultural Studies, Florida International University, SIPA 313 (305) 348-6663

b. Provide name and organization of person(s) who will actually collect and analyze the collected information.

Focus groups: Lauren Malone, CNA, will be the lead analyst responsible for collecting and analyzing focus group data.

Survey:

ASISA Research Group https://asisaresearch.com/about-us/?lang=en 1-305-850-6755

Overall supervision by Eduardo A. Gamarra, Professor, FIU

1 Due to the small number of recruiters within the Space Force population, we will be conducting focus groups with a smaller number of Space Force recruiters (10 rather than 20), to help protect confidentiality and to reduce the overall respondent burden on that population. Please note that focus groups with recruiters will be conducted virtually, so we will not know their locations (i.e., it will not be possible to identify which recruiters participated).

5