The objective of the Consumer Price Index (CPI) is to measure the change in cost of living experienced by the average urban consumer residing in the United States. Currently, the outlet sample in this measurement is selected via probability sampling proportional to average daily expenditure on items within a core-based-statistical-area (CBSA). This process yields a large variety of outlets in the CPI sample, but the outlets selected are based exclusively on the expenditures reported and household sampling weights in CPI’s Telephone-Point-of-Purchase Survey (TPOPS) survey. Using data collected from GasBuddy.com1, we attempt to model the explanatory variables of price change in order to identify possible stratification variables in outlet selection. Focusing on the Washington-Arlington-Alexandria, DC-VA-MD-WV CBSA, we calculate one-month price changes for each gas station in the GasBuddy sample. We then apply various statistical techniques to assess the significance of a variety of independent variables constructed using driving distances between stations, population density, income, and housing prices while controlling for various geographic flags. Finally, we construct indexes from the GasBuddy sample using a proposed stratified methodology, and compare them with indexes constructed using the traditional CPI methodology.