The study seeks to ascertain whether the operational definition of “gentrification” has an impact on the apparent extent, location and causal factors associated with the phenomenon. Four alternative definitional criteria are specified, based on areal changes in: proportion black, proportion college-educated, real incomes and real property values. The stringency of the given change needed to qualify as gentrification is also varied. Census tract changes from 1970–80 in Philadelphia are analyzed. Results indicate great sensitivity in the number and location of “gentrified” tracts to the definition chosen and stringency applied. Even more importantly, the 1970 characteristics of tracts which statistically explain their subsequent gentrication vary dramatically across these definitions.