Converging evidence from laboratory experiments and empirical models of scanner data suggests that product price evaluations are often based on a comparison to an internal reference price. Research indicates that the reference price may reflect various characteristics of previously encountered prices including the mean, the range, and the last price encountered. In this research, the authors test whether, for prices purportedly sampled over time, the reference price reflects temporal patterns of the price sequence (ascending and descending prices). In four studies, participants viewed prices purportedly sampled at one time point or at multiple time points and then evaluated a target price. Price distributions differed only in their temporal pattern, whereas the mean, the range, and in some conditions, the last price, were held constant. The results reveal that the price pattern does not affect price judgments when prices are purportedly sampled at one time point. However, for ascending and descending price sequences purportedly sampled over time, the price pattern affects price judgments. Based on these findings the authors propose that consumers flexibly select the internal reference price used for price evaluations.