We hypothesize that temporary sales might bias the results of vertical price transmission analysis towards findings of asymmetric price transmission. We test this hypothesis using a scanner dataset of retail butter prices in Germany. To this end we first use filters to identify underlying reference retail prices and thus remove temporary sales prices. We then compare estimates of vertical price transmission from the wholesale to the retail level that are generated with raw and with filtered retail prices. Our results confirm that temporary sales prices increase the speed and asymmetry of vertical price transmission. These results add a potential cause of asymmetry to those, such as market power, that have already been identified in the literature. [EconLit citations: C22; L10; Q11].