How Many Americans Change Their Racial Identification over Time?

Abstract

Despite long-standing assumptions about racial identity as a fixed characteristic, social scientists have increasingly recognized its fluidity and examined the origins of micro-level race change. However, knowledge regarding the prevalence of race change is limited. This data visualization fills this descriptive gap by providing a comprehensive account of recent levels of racial self-identification change among Americans. The author uses five high-quality panel surveys in which race is asked of the same nationally representative adult samples several years apart. Among all respondents, race change rates range from 5 percent to 12 percent across surveys, averaging 8 percent. Original white identifiers (4 percent on average) are much less prone to change than initial nonwhites collectively (20 percent). Blacks have as stable identities as whites, while mixed-race (52 percent) and “other”-race (73 percent) Americans undergo substantial identity shifts over time. Results further cement a perspective of race as flexible for some in the United States.

In the United States, racial identity is widely regarded as a fundamental determinant of various social outcomes. Despite ingrained assumptions about this characteristic as fixed and exogenous, scholars across the social sciences have increasingly appreciated the notion of racial fluidity: that race is a flexible and impermanent trait (Davenport 2020). For one prominent type of fluidity, micro-level race change over time, work has already uncovered its social (Saperstein and Penner 2012), economic (Antman and Duncan 2015), and political (Agadjanian and Lacy 2021) foundations. However, this body of research lacks something more basic: comprehensive and recent descriptive knowledge of how many Americans change their racial identification and what racial categories undergo race change most. Work linking census responses over time has been groundbreaking (Liebler et al. 2017), but because single informants record race for all their household members, instability cannot be fully ascribed to change in racial self-identification.

The aim of this data visualization is to paint a fuller, up-to-date picture of race change in America by assessing micro-level rates of change among all Americans and among initial racial subgroup categories. I use five high-quality panel surveys—interviews of the same individuals at two points in time—over the past two decades. Taking this comprehensive approach is important: although the surveys vary in methods, none is perfect, but if common patterns emerge, we can more confidently trust results and dispel concerns that they might be idiosyncratic had they come from a single source. Five panels provide nationally representative adult samples in which race is asked of the same respondents in two waves: the American National Election Studies from 2000 to 2004 and 2016 to 2020, the Cooperative Congressional Election Study from 2010 to 2014, the General Social Survey from 2010 to 2014, and the Democracy Fund Voter Study Group from 2011 to 2019. The Supplementary Information contains full data set names, race question wordings, and more data and method details.

Figure 1 plots race change rates among all adults in the top left panel and rates by initial racial subgroup category in other panels. Across the surveys, between 5 percent and 12 percent of Americans change their race over time, averaging 8 percent. Overall rates mask considerable subgroup variation. Although typically 4 percent of individuals who initially self-categorized as white change their race, about 20 percent of Americans originally in nonwhite categories do so, a disparity that holds across surveys. Among specific nonwhite groups, those initially in the Black racial category have the least fluid identities (a roughly 4 percent change rate). Original Hispanics (20 percent average), and to a lesser degree Asians (12 percent), change their race at higher rates. For those initially in the Hispanic category, one survey (the General Social Survey) reveals much higher change than other sources do. With a weighted sample of 43 respondents, this outlier likely just reflects a noisy estimate. Other rates are more meaningful here, and for a summary metric, the median of all five race change rates (11 percent) is a better representation than the outlier-sensitive average. Notably, a comparison across categories using medians still points to initial Hispanics shifting their racial identities more than initial whites, Blacks, and Asians do.

Figure 1.

Rates of racial self-identification change over time across five panel surveys, overall and by original racial subgroup categories.

Consistently large subgroup identity change does appear in other cases: Americans who initially self-classify as mixed race or specifically select the “other” racial category exhibit far and away the highest rates of change. Across-survey averages for these categories stand at 52 percent and 73 percent, respectively. Given the rapid growth of intermarriage rates and of individuals of mixed parentage in the United States, the fact that about half of those originally expressing multiracial identities sort into monoracial identities over time is especially meaningful. These higher race change rates hold across surveys in which mixed race is defined by either multiple race selection or single-response “mixed” identification but are higher for the former than the latter.

This visualization provides a valuable descriptive basis for our understanding of micro-level race change in the United States. Overall rates of change remain small but not insignificant, and for initial nonwhite categories, racial identity is disproportionately in flux. Future work should continue to examine the nature and implications of racial fluidity. With respect to panel survey design, results here highlight the importance of asking racial identity in every survey wave and further support abandoning the notion of race as a static individual level trait.

Supplemental Material

sj-docx-1-srd-10.1177_23780231221098547 – Supplemental material for How Many Americans Change Their Racial Identification over Time?

Supplemental material, sj-docx-1-srd-10.1177_23780231221098547 for How Many Americans Change Their Racial Identification over Time? by Alexander Agadjanian in Socius