The empirical literature on price indices consistently finds that aggregation methods have a considerable impact, particularly when scanner data are used. This paper outlines a novel approach to test for the homogeneity of goods and hence for the appropriateness of aggregation. A hedonic regression framework is used to test for item homogeneity across four supermarket chains and across stores within each of these supermarket chains. We find empirical support for the aggregation of prices across stores which belong to the same supermarket chain. Support was also found for the aggregation of prices across three of the four supermarket chains.