Approaches to Enhance Racial and Ethnic Representation in Secondary Data Analysis

Researchers, advocates, and policymakers are currently discussing how to revise racial and ethnic categories in the U.S. census and other federal government surveys. These discussions are particularly focused on better representing Latinos and people of Middle Eastern and North African descent. At the same time, social work scholars are developing, adapting, and using various methods to improve the representation of specific populations in existing datasets. Their goal is to emphasize intragroup heterogeneity, such as distinguishing between people of Mexican or Salvadoran descent, instead of grouping people of diverse Hispanics backgrounds together. Yet, investigating differences across race, ethnicity, language, and culture is challenging, especially when minority populations are underrepresented in existing studies. This situation underscores the need for researchers to share innovative methodologies to address the problem of small sample sizes and make the most of available datasets. Enhancing the representation of these groups is crucial for highlighting both disparities and strengths within subgroups. This detailed information is invaluable to social workers and communities advocating for policy and program changes.

In this roundtable, we will discuss four examples from recent studies that address the following question: How can social workers conduct high-quality secondary data analysis with limited sample sizes or leverage underutilized variables in existing datasets to document the diversity among their populations of interest?

First, we will hear from two presenters regarding techniques employed with data from the Current Population Survey, jointly sponsored by the U.S. Census Bureau and the Bureau of Labor Statistics. They will describe strategies to enhance the samples among various racial and ethnic groups, including Asian, Hawaiian and Pacific Islanders; American Indian or Alaskan Natives; and Hispanic subgroups such as Mexicans, Dominicans, Salvadorans, Cubans, and Puerto Ricans.

Next, another speaker will share her experiences using data from the National Social Life, Health, and Aging Project (NSHAP) to understand Latino caregiving. NSHAP is longitudinal study that focuses on the physical and mental health, well-being, social networks, and sexual health of older Americans. Piedra will emphasize the importance of the language variable in datasets that offer limited background information on Hispanic participants and discuss the value of adopting flexible interpretive strategies for data analysis. Additionally, she will address how differentiating between surveys completed in Spanish versus English can reveal significant differences in nativity or educational attainment among Latinos, even though many researchers currently aggregate all Hispanic or Latino respondents to increase sample sizes. Piedra will also explore strategies to rectify this practice, thereby improving the accuracy and usefulness of research outcomes.

Lastly, a presenter will discuss the importance of skin tone and how exploring it can add rich intra-racial/ethnic nuance to analyses. Several frequently used secondary datasets (e.g., Add Health, CARDIA, National Survey of American Life) include variables for skin tone. Adding skin tone variables and stratifying analyses by race/ethnicity can reveal intragroup skin tone differences in outcomes and potentially unmask inequalities within specific groups.

The overarching goal is to stimulate conversations and share methodologies to reflect and celebrate the growing diversity across the nation.