A Classification Approach to Uncertainty in Macroscale Measurement of Gentrification: A Case Study

Keywords: Gentrification, Uncertainty, Consistency Index, Prominence Index
Abstract Type: Paper Abstract

Abstract

Gentrification persistently reshapes its share of neighborhoods and transforms the urban landscape. For over a half-century, numerous efforts have been invested to understand the causal forces of gentrification and its impacts on neighborhoods. Varying theoretical frameworks and operationalization of gentrification among scholars, however, have made room for uncertainty in gentrification measurement. This paper seeks to explore uncertainty in the macroscale quantitative measurement of gentrification by integrating outputs of multiple measurement methods. We adopted four indicators commonly used in gentrification studies -median household income, education, occupation, and proportion of older housing units- in 2009-2017 in Harris County, Texas, to produce respective gentrified areas. Then, we created Consistency Index (CI) and Prominence Index (PI) to measure the inconsistency level of and severity of gentrification. CI/PI cross-classification categorized gentrified areas into 16 uncertainty scenarios. We then used county records of residential property renovations with the associated change in property values, to evaluate the uncertainty of lower CI/PI to high CI/PI classes. We found that gentrified block groups by the four distinct variables are inconsistent, overlapping only 36.7-62.5%. The verification with residential property renovation and changes in property values confirmed that lower CI/PI classes have a higher level of uncertainty in their definition, and the highest CI/PI class is the most certain, outstandingly, to contain gentrified areas. We suggest the need for caution when accepting gentrification definitions with the hope that insights into the levels of uncertainty within the measurements can be developed.