Racial Inequality in Education in Brazil: A Twins Fixed-Effects Approach

Citation: Letícia J. Marteleto, Molly Dondero (2016) Racial Inequality in Education in Brazil: A Twins Fixed-Effects Approach. Demography (RSS)
DOI (original publisher): 10.1007/s13524-016-0484-8
Semantic Scholar (metadata): 10.1007/s13524-016-0484-8
Sci-Hub (fulltext): 10.1007/s13524-016-0484-8
Internet Archive Scholar (search for fulltext): Racial Inequality in Education in Brazil: A Twins Fixed-Effects Approach
Tagged: Brazil (RSS), Race (RSS), Twins (RSS)

Summary

Background (quote):

The conventional approach to examining racial inequality in education is to model variance between individuals in different families, controlling for a number of key social origin and demographic variables — such as parents’ education and child’s sex and age — with a focus on the marginal effect of family and demographic characteristics Racial Inequality in Education in Brazil on educational outcomes. Although critical, this approach does not allow for separating the two main mechanisms that generally explain race disparities in opportunities and outcomes: (1) structural differences due to exposure to poverty and accumulated disadvantages, and (2) the social effects of race in itself.

A few previous researchers have compared siblings with different skin color and found lighter skinned siblings obtained better outcomes. This research improves by using twins with different skin color, removing additional unrelated sources of variation.

Used data from 1982, 1987-2009 nationally representative survey, identifying twins as "as children living in the same household who are classified as child of the head of the household and who share the same month and year of birth" and using black, pardo, and white racial categorizations in the survey as independent variable (treated as white and non-white; white/pardo twin sets large majority in study). Authors "compare educational disparities [completed years of education at time of survey is the outcome variable] across three analytic subsamples of 12- to 18-year-olds: singletons, same-sex twins in all families, and same-sex twins in multiracial families (families with parents of two races)".

Among all families with twins, the years of schooling are 4.81 for non-whites, 5.99 for whites (1.18 year difference). For twins within multiracial families, 5.14 and 5.61 (.47 year).

Non-white males are the most disadvantaged group within multiracial twins (0.9 year difference).

Possibility that parents racially categorize children based on outcomes, but this potential confounder would still indicate a form of racism.

Theoretical and Practical Relevance

Does not seem to attempt to determine proportion of differences in outcomes due to accumulated structural differences and present differences in treatment due to race. Does .47/1.18 = 40% attributable to present racism?