A year ago, I discussed looming issues with comparing racial and Hispanic origin data between the 1990 and 2000 Census. No solutions were initially offered, but now that the problem is here, solutions must be found. The 2000 Census allowed respondents to choose multiple racial categories, ballooning the categories from 10 in 1990 to 126 in 2000. This presents issues in making valid comparisons for questions related to racial fairness, such as education and health access.
Solutions vary from ignoring multi-racial categories where they are small, to creating a unified 'bridge' method for analysis, which distributes multiracial individuals proportionally across traditional racial categories. Alternatively, combining them through sophisticated dataset manipulation could provide a refined comparison to the simpler 1990 categories. While more solutions like definitive bridge categories from Census Bureau data might be years away, these suggested methodologies can offer interim plans for handling these disparities.