Lena Maria Wollschläger

• The 2N-ary choice tree model, a computational cognitive process model of multi-alternative multi-attribute preferential choice is proposed, revised, tested for its ability to simulate three benchmark context effects and interactions between them, and compared with earlier and more recent theories. The 2N-ary choice tree model assumes that the decision maker, given a set of N choice alternatives that are described by the same attributes, repeatedly compares pairs of attribute values and counts how often each alternative wins and loses a comparison. The number of favorable and unfavorable comparisons is stored in two separate counters per alternative and the difference of the counter states forms the preference state for the respective alternative. If the preference state for an alternative hits a negative threshold, this alternative is eliminated from the choice set and the comparison process continues without it. On the other hand, if the preference state for an alternative hits a positive threshold, this alternative is chosen and the whole process stops. The simple choice tree model, a revised version of the 2N-ary choice tree model, introduces an additional parameter for regulating the focus on the winning or losing alternative in a comparison, which has an effect on the proportion of choices and eliminations that take place. The 2N-ary choice tree model and the simple choice tree model are both able to explain similarity, attraction, and compromise effects, three context effects that have been observed after adding a third option to a set of two choice alternatives. With its additional parameter, the simple choice tree model beyond that accounts for recently observed correlations between these effects. It is the only model that accounts for the whole range of related findings, including negative similarity, attraction, and compromise effects.