Context-dependence of race self-classification: Results from a highly mixed and unequal middle-income country

Ethnic-racial classification criteria are widely recognized to vary according to historical, cultural and political contexts. In Brazil, the strong influence of individual socio-economic factors on race/colour self-classification is well known. With the expansion of genomic technologies, the use of genomic ancestry has been suggested as a substitute for classification procedures such as self-declaring race, as if they represented the same concept. We investigated the association between genomic ancestry, the racial composition of census tracts and individual socioeconomic factors and self-declared race/colour in a cohort of 15,105 Brazilians. Results show that the probability of self-declaring as black or brown increases according to the proportion of African ancestry and varies widely among cities. In Porto Alegre, where most of the population is white, with every 10% increase in the proportion of African ancestry, the odds of self-declaring as black increased 14 times (95%CI 6.08–32.81). In Salvador, where most of the population is black or brown, that increase was of 3.98 times (95%CI 2.96–5.35). The racial composition of the area of residence was also associated with the probability of self-declaring as black or brown. Every 10% increase in the proportion of black and brown inhabitants in the residential census tract increased the odds of self-declaring as black by 1.33 times (95%CI 1.24–1.42). Ancestry alone does not explain self-declared race/colour. An emphasis on multiple situational contexts (both individual and collective) provides a more comprehensive framework for the study of the predictors of self-declared race/colour, a highly relevant construct in many different scenarios, such as public policy, sociology and medicine.种族-族群分类标准因历史、文化与政治语境的差异而各不相同，这已是学界普遍共识。在巴西，个体社会经济因素对种族/肤色自我归类的强烈影响早已广为人知。随着基因组技术的推广应用，有观点提出可采用基因组祖先（genomic ancestry）信息作为自我申报种族等分类方式的替代方案，仿佛二者属于同一概念范畴。本研究针对15105名巴西受试者组成的队列（cohort），分析了基因组祖先比例、人口普查街区（census tract）的种族构成、个体社会经济因素与自我申报种族/肤色之间的关联。研究结果显示，自我申报为黑人或棕种人的概率随非洲血统比例升高而上升，且在不同城市间存在显著差异。在人口以白人为主的阿雷格里港，非洲血统比例每升高10%，受试者自我申报为黑人的优势比升高14倍（95%置信区间（95% Confidence Interval, CI）：6.08~32.81）；在人口以黑人和棕种人为主的萨尔瓦多市，该优势比升高3.98倍（95%置信区间：2.96~5.35）。居住街区的种族构成同样与受试者自我申报为黑人或棕种人的概率存在关联：居住普查街区内黑人和棕种人居民比例每提升10%，受试者自我申报为黑人的优势比升高1.33倍（95%置信区间：1.24~1.42）。仅依靠血统比例无法完全解释自我申报的种族/肤色分类结果。若兼顾个体与集体层面的多重情境因素，可为自我申报种族/肤色这一构念（construct）的预测因素研究提供更全面的分析框架，而该构念在公共政策、社会学与医学等诸多领域均具有重要研究价值。