What a Polygon Can't Tell You: Shifting from Area-Level to Point-Level Investigation of Residential Segregation Patterns
The study of segregation is essential for understanding how place influences life outcomes. However, traditional segregation indices rely heavily on the use of areal units for calculation, which risks introducing both measurement and interpretation error. Researchers suggest that individual-level data avoids many of the problems facing traditional area-level indices. In this Dissertation, I use the recent release of the complete 1940 Census to investigate the potential problems with measuring segregation with areal units and develop a new method for measuring segregation at the individual level. In Chapter 1, I investigate the potential impact the modifiable areal unit problem (MAUP) may have on accurately measuring segregation when using areal unit indices. In Chapter 2, I develop a new measure of segregation, the Shortest Path Isolation (SPI) index, which captures the degree of racial isolation from the perspective of what an individual would experience. Using the SPI index developed in Chapter 2, Chapter 3 investigates how individual-level racial isolation in 1940 West Philadelphia is associated with access to neighborhood resources by race. Given that our understanding is only as good as our measurement, it is imperative that our measures accurately reflect our perceptions of segregation. ^
Geographic information science and geodesy|Sociology|Demography
Fineman, Ross William, "What a Polygon Can't Tell You: Shifting from Area-Level to Point-Level Investigation of Residential Segregation Patterns" (2017). Dissertations available from ProQuest. AAI10257957.