“Humans and non-human animals frequently violate principles of economic rationality, such as transitivity, independence of irrelevant alternatives, and regularity. The conditions that lead to uneconomical choices are not completely understood, but violations predominantly occur when options differ along more than one dimension. Thus, we need to understand how decision makers integrate information from more than one reward dimensions, in order to expose potential limitations of the decision-making mechanisms. I will present two studies on mice tested in automated home-cage setups using rewards of drinking water, as well as results from decision-making models. In the first study, rewards differed on only one of two dimensions, volume or probability, and mouse choice conformed to the principles of economic rationality. However, despite equivalence in return rates, mice responded more strongly to differences in probability than to differences in volume. In the second study the choice options differed along one or both of the reward dimensions (volume and probability) and mouse choice in three different experiments could be best explained by a reliance on the full information from both reward dimensions. However, the results also suggest that over time a bias for one of the two dimensions may have developed, because mouse choice in the last experiment of a within-subject series of four (same as first experiment but seven weeks later) could be best explained by a lexicographic rule (take the best heuristic).”