D’Ignazio and Klein offer examples from Facebook and Carl Linneaus to explore examples of sexism and racism in systems of classification.
Those in power typically have the resources to do such classification, counting, and analysis so it is sometimes easy to just dismiss classification all together. As D’Ignazio and Klein note, the flaw in that thinking is that to use data at all is to classify: “by the time information becomes data, it’s already been classified in some way. Data, after all, is information made tractable, to borrow a term from computer science.” To be able to analyze information, it needs to be come data. That takes classification.
However, beyond the easier to understand issues of ways in which classification can do harm (e.g., the scientific racism of the eighteenth century that set up hierarchies among different people), even in more well-intentioned efforts there are drawbacks that can be hard to see.
Sometimes those who would benefit from classification might also be harmed by that same classification, as D’Ignazio and Klein point about about census data collected on documented, undocumented, and nationalized immigrants (e.g., documented immigrants with undocumented family members might not participate in the census which would then undercount that population and diminish resources for that population). They call this the “paradox of exposure”.
Much of these issues go back to issues from chapter 1: who benefits from this data collection? Who is potentially harmed? what goals seem to be prioritized?