属性への注意のまとめ方で選択行動の違いを説明する

We study an attribute-based model of stochastic choice in which attention to each attribute is limited, and characterize two choice rules within it. The Multiplicative Attention Rule (MAR) rewards an alternative only for the attributes on which it ranks first, combining attention to these multiplicatively across attributes; it captures choice deferral driven by the cognitive load of attending to `too many' attributes. The Additive Attention Rule (AAR) instead averages attention across attributes and credits an alternative by its full ordinal rank within the menu. This difference in aggregation is what separates the two rules behaviourally: AAR reproduces the Attraction Effect while holding the relevant attributes fixed across menus, something MAR can do only when the attribute set itself changes, though AAR, in turn, cannot accommodate the Compromise Effect. Both rules are characterized by axioms on observable choice data, and in both cases the underlying attention parameters and attribute-based preferences are uniquely identified from that data.