Elsevier

Social Science Research

Volume 42, Issue 1, January 2013, Pages 106-119
Social Science Research

Measures of “Race” and the analysis of racial inequality in Brazil

https://doi.org/10.1016/j.ssresearch.2012.06.006Get rights and content

Abstract

Quantitative analyses of racial disparities typically rely on a single categorical measure to operationalize race. We demonstrate the value of an approach that compares results obtained using various measures of race. Using a national probability sample of the Brazilian population that captured race in six formats, we first show how the racial composition of Brazil can shift from majority white to majority black depending on the classification scheme. In addition, using quantile regression, we find that racial disparities are most severe at the upper end of the income distribution; that racial disparities in earnings are larger when race is defined by interviewers rather than self-identified; and that those classified as “black” suffer a greater wage penalty than those classified as “brown.” Our findings extend prior conclusions about racial inequality in Brazil. More generally, our analysis demonstrates that comparison of results across measures represents a neglected source of analytic leverage for advancing empirical knowledge and theoretical understanding of how race, as a multidimensional social construct, contributes to the production of social inequality.

Highlights

► We explore income disparities in Brazil across six racial measurement schemes. ► Estimates of inequality are sensitive to researchers’ choice of race measure. ► Across-measure disparities are greatest at upper ends of the income distribution. ► Disparities are larger when race defined by interviewers rather than self-defined. ► Comparison across race measures can improve understandings of inequality.

Introduction

Social scientific understanding of racial inequality is shaped by the way race is measured and operationalized in quantitative research. This is true for at least three reasons: First, the choice of categories in a given data set conditions what we see in terms of that context’s racial composition. The decision to include any given set of categories is contingent, reflecting judgments about who should be counted and with what terminology (Harris, 2002, Nobles, 2000). Second, the categories we employ determine how we understand the distribution of resources along racial lines. Classification by skin color, for example, can reveal inter- and intra-racial category stratification (Hughes and Hertel, 1990, Keith and Herring, 1991, Gullickson, 2005, Goldsmith et al., 2006). Third, the particular method of classification we choose shapes how we understand racial inequality. For instance, interviewer classification versus self-classification can lead to different conclusions about the magnitude of income inequality between racial groups (Saperstein, 2006, Telles and Lim, 1998).

Although race is typically experienced as a stable and unitary individual trait, social constructionist theory posits that race is actually a contextual and multidimensional social construct. Racial identifications hinge on a constellation of factors, including self-perception, ascription by others, interactional cues, institutional contexts, and prevailing cultural understandings of consequential markers of human difference (Cornell and Hartmann, 1998, Jenkins, 1994, Harris and Sim, 2002, Roth, 2005, Brunsma, 2005, Daynes and Lee, 2008). Since race is socially constructed, there is no a priori reason to assume that any given measure of race is more or less valid than others. Indeed, studies by Telles and Lim, 1998, Saperstein, 2006, Saperstein, 2008, and Campbell (2009) call attention to the fact that various measures that researchers have traditionally treated as alternative proxies for “race” do not necessarily capture the same underlying “thing.” Far from being a liability, the fact that different measures of race may be used to operationalize different dimensions of race provides a valuable and underutilized source of analytic leverage for studies of racial inequality.

In this article, we present an analytical approach that exploits the comparison of results across different methods and formats of racial classification better understand patterns and dynamics of racial inequality. We demonstrate the substantive and theoretical value added by the use of multiple measures through an original analysis of racial inequality in wages in Brazil based on the 2002 Brazilian Social Survey (PESB). The PESB is a national probability sample that captures racial classification through six formats and methods. We first present descriptive statistics to show how the method used to collect and code racial data can dramatically affect starting assumptions about a population’s racial composition. We then demonstrate how the measurement of race can affect the analysis of racial inequality by using quantile regression methods to generate and compare levels of income inequality in Brazil based on different definitions of “race.”1

Results show that differing classification schemes significantly alter the overall picture of the racial composition of the Brazilian population. According to some schemes, Brazil is a predominantly nonwhite country; in others, it becomes majority white. We also find that the magnitude of racial disparities in wages changes depending on how race is defined and according to location along the income distribution. Finally, comparisons of level of inequality across classification schemes provide clues to the underlying mechanisms fueling racial disparities. We observe, for example, that the darkest segment of the nonwhite population is especially vulnerable in Brazil, signaling the nuanced role of skin color in estimates of inequality, a factor that is obscured in schemes that treat all nonwhites as equally disadvantaged.

Our empirical findings build upon and extend recent scholarship on racial stratification in contemporary Brazil. Looking beyond the Brazilian context, our analysis demonstrates more generally how comparison of results across measures can advance understanding of descriptive trends and underlying dynamics that fuel racial inequality.

Section snippets

Race measures and inequality studies in Brazil

The vast majority of research on racial dynamics in Brazil focuses on the black-to-white continuum. Brazil’s large-scale social surveys typically use three racial or color terms to capture the range of identifications on this continuum: white (branco), brown (pardo, or “mixed”), and black (preto).

Data

Our analysis uses data from the Brazilian Social Survey (PESB), conducted in July and August 2002. PESB follows the model of the American General Social Survey (GSS). The data are based on a nationally representative sample covering the five regions of Brazil and all persons aged 18 and over. The complete sample consists of 2364 persons sampled across 102 municipalities. Because we are interested in the black–white continuum, we exclude 34 individuals who self-classified as of Asian or

Methods

We begin by presenting descriptive statistics that show how the racial composition of the Brazilian population shifts when individuals are sorted according to six possible classification schemes. Each scheme captures a different dimension or definition of “race.” Table 1 summarizes the classification schemes used in our analyses.

  • 1.

    Census: Self-classification as black, brown, or white. The Brazilian Census Bureau (IBGE) has employed these three terms for much of the last two centuries. This scheme

Population distribution by classification scheme

Fig. 1 shows the distribution of persons by racial category in each of the six different schemes. Different classification schemes clearly generate very different pictures of Brazil’s racial composition. The two ternary formats look most alike and show that browns make up a significant share of Brazil’s population. In contrast, the “post hoc binary” scheme, which takes the picture generated by the census format and then eliminates the distinction between browns and blacks, portrays a population

Discussion

Our analysis of race-based wage inequality in Brazil under alternative racial classification schemes demonstrates how the way race is measured influences descriptive and analytic conclusions about the nature and extent of racial inequality. In this discussion, we highlight how the use and comparison of multiple measures deepens our understanding of patterns and sources of racial inequality in Brazil.

The choice of classification scheme and method clearly affects how we perceive Brazil’s racial

Conclusion

We began this article with the observation that social scientific understanding of racial inequality is shaped by the way race is measured and operationalized in quantitative research. To conclude, we point to three specific ways in which the adoption of an analytic approach that compares results across multiple measures can advance sociological understanding of racial stratification.

First, comparison across measures provides a means to investigate empirically how race is socially defined in a

Acknowledgments

Earlier versions of this article were presented at the 2009 meetings of the American Sociological Association and of the Associação Nacional de Pós-graduação e Pesquisa em Ciências Sociais (ANPOCS) in Brazil. We thank participants in the sessions for their comments. The authors wish to thank Laura Randall, Sarah Bruch, Andrew Penner, Andreas Wimmer, and Tom DiPrete for comments on earlier drafts.

References (68)

  • K.A. Rockquemore et al.

    Opting for white: choice, fluidity, and racial identity construction in post–civil rights America

    Race and Society

    (2002)
  • Richard Alba

    Bright vs. blurred boundaries: second-generation assimilation and exclusion in France, Germany, and the United States

    Ethnic and Racial Studies

    (2005)
  • J.-L. Arcand et al.

    Racial discrimination in the Brazilian labour market: Wage, employment and segregation effects

    Journal of International Development

    (2004)
  • O. Arias et al.

    Education, family background and racial earnings inequality in Brazil

    International Journal of Manpower

    (2004)
  • S.R. Bailey

    Unmixing for race making in Brazil

    American Journal of Sociology

    (2008)
  • S.R. Bailey et al.

    Multiracialism vs. a collective black: census debates in Brazil

    Ethnicities

    (2006)
  • Barros, R.P., Mendonça, R.S., 1995. Os determinantes da desigualdade no Brasil. Texto para Discussão do IPEA no. 377....
  • F. Blau et al.

    Wage structures and gender earnings differentials: an international comparison

    Economica

    (1996)
  • Bruch, S.K., Loveman, M., 2011. Measuring and modeling race as a multidimensional construct: evidence from research on...
  • D.L. Brunsma

    Interracial families and the racial identification of mixed-race children: evidence from the Early Childhood Longitudinal Study

    Social Forces

    (2005)
  • M.J. Budig et al.

    Differences in disadvantage: variation in the motherhood wage penalty across white women’s earnings distribution

    American Sociological Review

    (2010)
  • B.S. Cade et al.

    A gentle introduction to quantile regression for ecologists

    Frontiers in Ecology and the Environment

    (2003)
  • F.R. Campante et al.

    Desigualdade salarial entre raças no mercado de trabalho urbano brasileiro: aspectos regionais

    Revista Brasileira de Economia

    (2004)
  • M.E. Campbell

    Thinking outside the (black) box: measuring black and multiracial identification on surveys

    Social Science Research

    (2007)
  • M.E. Campbell

    Multiracial groups and educational inequality: a rainbow or a divide?

    Social Problems

    (2009)
  • M.E. Campbell et al.

    The implications of racial misclassification by observers

    American Sociological Review

    (2007)
  • J.A.M. Carvalho et al.

    Estimating the stability of census-based racial/ethnic classifications: the case of Brazil

    Population Studies

    (2004)
  • S. Cornell et al.

    Ethnicity and Race. Making Identities in a Changing World

    (1998)
  • S. Daynes et al.

    Desire for Race

    (2008)
  • C.N. Degler

    Neither Black nor White: Slavery and Race Relations in Brazil and the United States

    (1971)
  • Ferreira, F.H.G., Leite, P.G., Litchfield, J.A., 2006. The rise and fall of Brazilian inequality, 1981–2004. Policy...
  • T. Golash-Boza

    Does whitening happen? Distinguishing between race and color labels in an African-descended community in Peru

    Social Problems

    (2010)
  • A.H. Goldsmith et al.

    Shades of discrimination: skin tone and wages

    American Economic Review

    (2006)
  • E. Grodsky et al.

    The structure of disadvantage: individual and occupational determinants of the black–white wage gap

    American Sociological Review

    (2001)
  • A. Gullickson

    The significance of skin color declines: a re-analysis of skin tone differentials in post-civil rights America

    Social Forces

    (2005)
  • D.R. Harris

    Does it matter how we measure? Racial classification and the characteristics of multiracial youth

  • D.R. Harris et al.

    Who is multiracial? Assessing the complexity of lived race

    American Sociological Review

    (2002)
  • C. Hasenbalg

    Race and socioeconomic inequalities in Brazil

  • S. Hitlin et al.

    Measuring Latinos: racial vs. ethnic classification and self-understandings

    Social Forces

    (2007)
  • M. Hughes et al.

    The significance of color remains: a study of life chances, mate selection, and ethnic consciousness among Black Americans

    Social Forces

    (1990)
  • Ianni, O., 1960. Segunda parte. In: Cardoso, F.H., Ianni, O. (Eds.), Cor e Mobilidade Social em Florianópolis....
  • R. Jenkins

    Rethinking ethnicity: identity, categorization and power

    Ethnic and Racial Studies

    (1994)
  • G. Kao

    Racial identity and academic performance: an examination of biracial Asian and African American youth

    Journal of Asian American Studies

    (1999)
  • V.M. Keith et al.

    Skin tone and stratification in the black community

    American Journal of Sociology

    (1991)
  • Cited by (110)

    • Trends in the academic achievement gap between high and low social class children: The case of Brazil

      2022, International Journal of Educational Development
      Citation Excerpt :

      There are important differences in the history of racial discrimination in Brazil and the U.S. – namely, that legalized race discrimination (Jim Crow) in the U.S. post-Reconstruction was not a feature of late 19th and 20th century Brazil (Marx, 1997). However, if we assume that educational attainment and earnings gaps controlling for attainment are reasonable indicators of racial bias in a society, research on attainment and earnings differences between blacks (pretos and pardos separated or combined) and whites in Brazil show that they are as large in Brazil as in the U.S. (Bayer and Charles, 2018; Firpo et al., 2021), although how race is classified in Brazil appears to make a difference (Bailey et al., 2013). Further, as we show below, in 9th grade, race is significantly correlated with the social class achievement gap.

    View all citing articles on Scopus
    View full text