Data Weaknesses
A spatial analysis was not completed, but there is surely spatial correlation between evictions and litigations in zip codes adjacent to one another. In addition, measuring variables at the zip code level is less desirable than at the PUMA or census block level, but eviction data were only available at the zip code level. Using percentage of non-white population as a diversity measure is limited, since there are many factors that can contribute to diversity: education level, language spoken at home, religious affiliations, etc. Finally, in the creation of two of my variables, eviction percentage and litigation percentage, I took these values as a percentage of total units in each zip code. I could have computed these values as a percentage of total population.