“’Black and Some Other Race?’ Examining Black Latino Undercounted in Census Bureau’s Modified Race Question”

• In Event: Visualizing AfroLatinos/Black Latinos in Education Research: Illuminating Complex Inequities for Policy and Practice

Abstract

1. Objectives or purposes: This study leverages sample data from the American Community Survey (ACS) to assess how changes to the 2020 race question influenced the odds of multi-race classification for Black Latinos. In fact, we find substantial drops in single-race Black Latino origin group counts between samples, which coincides with a relative surge in multi-race classification. 2. Perspective(s) or theoretical framework: We examine two empirical hypotheses in this study. First, we anticipate that the odds of multiracial classification for Black Latinos will be higher in the post-2020 race question than in the pre-2020 race question. We specifically posit that the odds of multiracial classification will be most prevalent among Latino origin groups that have relatively significant Black population sizes. Second, we hypothesize that the provision of two distinct origin responses, which we consider approximate to multiracial responses, will yield higher levels of multiracial classification in the pre-2020 race question than in the post-2020 race question among Black Latinos. This presumption is rooted in our speculation that the post-2020 race and ethnicity questions may be overestimating multiracial respondents. 3. Methods, techniques, or modes of inquiry: We perform a survey-weighted binary logistic regression model to assess the odds of Black Latino respondents being classified as “Black and Some Other Race” (Black and SOR) as opposed to Black alone in the 2019 and 2021 ACS. To compare results, we evaluate these odds separately since respondents are not linked across survey periods. Our empirical analysis constitutes two stages of modeling whereby the odds of Black and SOR classification are regressed on various Hispanic or Latino origin responses across both samples and then observed in a second model alongside demographic characteristics. 4. Data sources, evidence, objects, or materials: We use sample data from the 2019 and 2021 American Community Survey (ACS) 1-year microdata files from the Integrated Public Use Microdata Series (IPUMS-USA). 5. Results and/or substantiated conclusions or warrants for arguments/point of view: Our results suggest that most Black Latino origin groups exhibit higher odds of multi-race classification in the modified race question compared to the pre-2020 race question. The reclassification of single-race Black Latinos as multiracial is particularly notable for Dominicans and Hondurans. Our findings also reveal that Black Latinos who provided multiple origin responses that are ethno-racially distinct (i.e. Nigerian and Panamanian) are nearly 21 percent less likely to be classified as multiracial in the modified race question. This finding suggests that the coding schemes used in the modified race question may be over-classifying single-race Black Latinos as multiracial. 6. Scientific or scholarly significance of the study or work: This study underscores the ethical consequences of misclassifying single-race Black Latino responses as multiracial while also emphasizing the broader methodological implications in how official changes to question wording and coding schemes lead to artificial population shifts that are not attributed to natural demographic change.

Author

• Ricardo H. Lowe, University of Texas at Austin