The measurement of racial discrimination may seem like a purely technical topic, but in reality it comes with a huge moral dilemma. In order to measure racial discrimination, you have to categorize people into different racial groups (usually in your national census). On the basis of this you can then collect social information about those groups, and compare the average outcomes in order to detect large discrepancies between them. For example, do blacks in the US earn less, achieve less in school etc. Only then can you assume that there may be racism or discrimination and can you design policies that deal with it.
Now, categorizing people into different racial groups is not straightforward. You need to do violence to reality. Racial classifications and categorizations are not simply a reflection of factual reality, of “real group identities”. Instead they are social constructions or even fantasies influenced by centuries of prejudice, stereotypes and power relations. If we want to use racial classifications to measure discrimination, then we give people labels that may have little or nothing to do with who or what they are and how they identify themselves. Instead, these labels perpetuate the stereotypes and power relations that were the basis of the racial classifications when they were first conceived centuries ago. For example, “black” or “African-American” is not a simple descriptive label of a well-defined and existing group of people; instead it’s an ideological construction that was once used to segregate certain groups of very different people and subordinate them to a lower station in life. (Evidence for the claim that race is a social construct rather than a natural fact can be found in biology and in the fact that racial classifications differ wildly from one country to another).
In other words, the “statistical representation of diversity is a complex process which reveals the foundations of societies and their political choices” (source). In this particular case, the foundation of society was racism and the political choices were segregation and discrimination. If today we use the same racial and ethnic classifications that were once used to justify segregation and discrimination, then we run the risk of perpetuating racist social constructions. As a result, we may also help to perpetuate stereotypes and discrimination, even as we try to go in the opposite direction. It’s a form of path dependence.
Statistics are not just a reflection of social reality, but also affect this reality. Statistical categories are supposed to describe social groups, but at the same time they may influence people’s attitudes towards those groups because they contain memories of older judgments that were once attached to those groups. The dilemma is the following: the use of racial classifications to measure discrimination means giving people labels that have little or nothing to do with who they are or what they are; but they have something to do with how others treat them. It’s this treatment that we want to measure, and we can’t do so without the use of classifications. Using such classifications, however, can help to perpetuate the treatment we want to measure and avoid.
More posts in this series are here.