Diversity in U.S. Military Families: An Environmental Scan of the Peer-Reviewed Literature on Race and Ethnic Variation for Select Well-Being Outcomes

Key Findings

• Existing articles on service member well-being lack a focus on R/E in core research questions. Although researchers often recognize that R/E could affect the outcome of interest, the most common approach to addressing this impact is to control for the variation associated with R/E instead of specifically exploring it. This leads to fragmented development of knowledge, which makes it difficult for researchers to consolidate and use that knowledge to develop policies and programs.
• The existing body of research includes a limited variety of analytic methods. Specifically, the work does not include qualitative approaches that could tell us why R/E outcome differences occur. An investment in understanding the why behind R/E variation in outcomes is essential to supporting a diverse group of service members and their families and developing effective evidence-based policies and programs.
• A substantial number of articles relied on convenience samples to answer research questions. These articles lacked generalizability to broader service member and family populations and were unable to sufficiently answer questions about R/E. For research designed to address R/E well-being questions, ensuring that quantitative analyses have a broad and representative sample across R/E groups is critical.
• The existing literature on R/E and well-being lacks attention to the spouses and children of service members. Only a handful of articles included children and spouses of service members, and many of those did not separate their findings across these unique groups. This leaves the reader with no appreciation for how R/E might differentially affect children, service members, and spouses.

Military leadership has increasingly recognized the importance of supporting a diverse workforce and the deficits in efforts to support service members of color (Burns and Baldor, 2021). In a recent memo, former U.S. Secretary of Defense Mark Esper outlined a set of immediate actions to stop discrimination, prejudice, and bias in the ranks of the U.S. armed forces (Esper, 2020). For example, the memo directed the services to remove photographs from consideration by promotion boards, add bias awareness and bystander intervention training, and develop educational requirements about unintentional bias. The memo reflects a growing commitment to diversifying the force and encouraging service members of color and women to pursue positions in higher ranks, which remain dominated by white men.

These efforts are important steps toward supporting diverse service members, but more work is needed to ensure those supports extend to military families, including spouses and children. A service member's family plays a critical role in their ability to serve (Le Menestrel and Kizer, 2019). Family members provide support to services members while they serve and often care for service members as they recover from physical and mental health injuries (Institute of Medicine, 2013). Issues within a family can also cause stress for service members and might interfere with their readiness (Keller et al., 2018; Meyers, 2018; Schneider and Martin, 1994; Sims et al., 2017; Strong et al., 2021). Moreover, family well-being is often a consideration when service members face retention decisions (Keller et al., 2018; Meyers, 2018). Therefore, ensuring the health and well-being of military families must be a priority for the Department of Defense (DoD).

While research on the relationship between race, ethnicity, and family well-being is well developed in the civilian literature (e.g., mental and physical health: Luo and Waite, 2005; Williams et al., 1997; marital satisfaction: Dillaway and Broman, 2001; family violence: Cho, 2012; financial stress: Al-Bahrani, Weathers, and Patel, 2019), there is a gap in our knowledge about the role of race and ethnicity (R/E) in well-being outcomes among service members and their families. Going forward, military leaders and policymakers need evidence to inform their efforts to support a diverse population of military families. This study begins to fill this gap. Our research objective was to broadly review and synthesize existing literature to better understand what researchers have found about how one specific aspect of diversity—R/E—is related to military family well-being. Specifically, we were interested in (1) whether researchers focused on R/E differences in outcomes as a driving research question (not just as a control variable); (2) the variables used to capture R/E; and (3) the quality of research in terms of design, data, and analysis.

To accomplish this goal, we conducted an environmental scan of the literature across five key components of well-being: mental health, behavioral health, family violence, marital satisfaction, and financial stress.¹ The five domains represent a diverse set of outcomes that have been examined in the peer-reviewed literature on military family well-being. Working with a librarian, we developed a search strategy to facilitate our scan. The review was limited to English-language, peer-reviewed articles and reports published between 2000 and early 2021. The search itself occurred between February and June 2021.

Summary of Findings

We identified 110 unique articles that met the inclusion criteria for our well-being outcome domains. We identified 75 articles that were included in the mental health domain, 54 articles that were included in the behavioral health domain, seven articles that were included in the family violence domain, four articles that were included in the marital satisfaction domain, and one article that was included in the financial stress domain.² The remainder of this section describes these articles’ most common attributes, including designs, population, outcomes, and investigation of differences by R/E across service members and their families. Table 1 provides a summary of the most common features of these articles, although the examples are not exhaustive.

Data Sources

The included articles used a variety of data sources. Most commonly, articles drew from large secondary data sets familiar to the fields of military behavioral health research, such as the Millennium Cohort Article (e.g., Ursano et al., 2016), Army Study to Assess Risk and Resilience in Servicemembers (e.g., Ursano et al., 2016, 2018; Ursano, Kessler, Heeringa, et al., 2015), the American Community Survey (e.g., London and Heflin, 2015), and medical or insurance records based on TRICARE usage and diagnoses (e.g., Schmied, Highfill-McRoy, and Larson, 2012). The samples and populations from these sources were often representative of the military population (or statistical methods were used to make them representative). However, many articles also used data from convenience samples. For example, an article by Steele, Germain, and Campbell (2017) used anonymous self-report surveys collected by a U.S. Navy Mobile Care team. These convenience samples were not representative of the broader population of service members and their families.

Sample Populations

Sample populations were largely service members themselves. Six articles focused on children of service members and five articles included spouses. Articles often included individuals from a variety of R/E backgrounds. However, it was not possible to compare across articles because researchers used different levels of specificity when it came to R/E to describe their sample. For example, some researchers provided information only on the percentage of white service members in their population, considering the remaining population “other race,” while other articles included a variety of categories for R/E (e.g., Gibbs, Clinton-Sherrod, and Johnson, 2012).

Analytic Approach

Most articles used regression analysis (e.g., logistic, multivariate) and descriptive statistics (e.g., chi-square tests, t-tests) to answer their research questions. This is consistent with researchers’ focus on large secondary data sources. The one exception was systematic reviews, which used a systematized process for selecting and abstracting data from existing research (e.g., O'Keefe and Reger, 2017; Jones et al., 2020).

Measuring Race and Ethnicity

Measures of R/E lacked consistency and varied widely across articles. The least detailed definition of race was simply minority versus nonminority (or white versus non-white) while the most detailed measure included a “select all that apply” race item (e.g., Black/African American, white/Caucasian, Asian, Pacific Islander, Native American) and a separate item for Hispanic ethnicity, usually indicated by a yes or no. This more-detailed measure of race allowed authors to include multiracial and “other” as derived categories. Although rare, some articles measured only race and did not measure Hispanic ethnicity. Generally, when “Hispanic” was included in analyses, the other racial groups were defined as “non-Hispanic.” That is, hierarchically, Hispanic ethnicity took precedence over all other racial categories. The most commonly used R/E measures in analyses were Black/white/Hispanic/other and Black/white/Hispanic/Asian and Pacific Islander/other. The ability to disaggregate the “other” race category was largely a function of sample size, but in most cases, authors did not explicitly explain why they collapsed certain groups.

Outcomes

Researchers focused on a variety of outcomes within each domain; however, some outcomes were more common than others. The most common outcomes studied across domains were PTSD and suicidality, including both risk and ideation. Other outcomes included general measures of mental health (e.g., Waitzkin et al., 2018), interpersonal conflict between service members and spouses (e.g., Gibbs, Clinton-Sherrod, and Johnson, 2012), substance use (e.g., Chaudhary et al., 2019), and sleep (e.g., Hill et al., 2020). Some articles also included measures of military readiness, such as unit cohesion (Dickstein et al., 2010), resilience (Vyas et al., 2016), and retention (Taylor, Hernández, and Clinton-Sherrod, 2021).