Rethinking Reference and Racial Categories: Implications and Insights for Researchers

• In Event: Lightning Session: Social Equity and Race

Abstract

Racial and ethnic categories are socially constructed and fluid, reflecting past cultural practices and assumptions (Cabrera, 2018; Parker, 1998). Racial categories are also often conceptualized and used in models as relative to white individuals (i.e., “people of color” or “non-white”). We build on the robust literature across disciplines, particularly in anthropology, public health, and sociology on the use of racial categories in research (Covington, 1995; Elliott et al., 2022) and employ Quantitative Critical Race Theory (QuantCrit) to consider the implications of collapsing ethnic categories and reference group category selection. 

We ask, “How does math interest vary among racially minoritized students?” Typically, this question would be modeled with a categorical variable for race/ethnicity where one category is excluded (generally white). We build on work from prior researchers (Arellano, 2022) who suggest several alternatives to this approach: 1) use the average of all students, 2) use the average of all students excluding white students, 3) use a universal target, and 4) analyze within-group differences.  We utilize the publicly available Early Childhood Longitudinal Studies (ECLS-K) dataset, which tracks one cohort of kindergarteners through fifth grade and provides student-level information on demographics, including race and ethnicity, as well as academic outcomes. Our sample consists of 12,283 students. 

Our results demonstrate how interpretation may vary across approaches (the relationships themselves remain unchanged). One clear result here is that, counter to many deficit narratives around students of color, white students score lower on average in their interest in math than Asian, Black, Latino, Native Hawaiian/Pacific Islander, and students of two or more races. Interpretations change for Asian students. Column two’s interpretation, the typical choice in education research, reads that being an Asian student compared to being a white student is related to an increase in math interest of .17 points (1-4 scale) on average at a .01 significance level. This interpretation differs from the grand mean comparison, which indicates that being an Asian student compared to the average is associated with a 0.02-point increase in math interest (on a 1-4 scale).

We conclude that thoughtful choices around reference category selection and disaggregation of ethnic groups can reveal patterns that would otherwise be missed in traditional approaches.  We offer a framework for researchers interested in thinking through the implications of racial group choices. Firstly, we encourage researchers to ask whether they are centering racism in their research questions. If so, we ask whether they are studying racial/ethnic disparities. If the answer is no, it is still important to consider whether categories unfold or, in other words, are nested within a larger group. In the context of centering race, for example, is there a larger Black category, with subgroups such as Black Caribbean and/or African American? This prompting question can help researchers choose racial categories and further understand within-group differences. However, this nuanced data might not be available. Instead, researchers can compare respondents from each racial category to the overall average (with and without white individuals) and/or a universal target.

Authors

• Wendy Castillo, Montclair State University
  Presenting Author

• Hannah Kistler, University at Albany
  Non-Presenting Co-Author