A long-standing question within the Consumer Price Index (CPI) program has been how best to determine the contribution of geographic areas to the overall CPI variance using standard statistical inference tools. The CPI is constructed of higher-level AREA-ITEM aggregates that are built up from an initial set of AREA-ITEM cells at the basic Index-Area—Item-Stratum level. The CPI produces summary percent price changes for all of these aggregate levels. By utilizing the basic level price changes and their higher level price changes, we will proceed to construct an ?adaptive? analysis of variance (ANOVA) using these basic level price changes as the initial set of observations. A standard two-way ANOVA with one observation per cell is then applied. The ANOVA results provide F statistics that demonstrate the significance (or not) of AREA and ITEM in the two-way model. For the time periods covered, two out of every three models show AREA to be a significant effect.