Interactive online decision aids often employ user‐decision aid dialogues as forms of user‐system interaction to help construct and elicit users' attribute preferences about a product type. This study extends prior research on online decision aids by investigating the effects of a decision aid's user‐system interaction mode (USIM), which can be either user‐guided or system‐controlled, on users' effort‐related (number of iterations of using the aid and perceived cognitive effort expended in using it) and quality‐related (perceived quality of the aid and acceptance of the product advice it provides) assessments. A contingency approach with two moderating factors is employed. One factor is the decision strategy (additive‐compensatory or elimination) employed by the aid, and the other is the users' product knowledge (high or low).A laboratory experiment was conducted to compare online decision aids with different USIMs. Although the results largely confirm that users assess the user‐guided USIM more positively than the system‐controlled USIM, the effects of USIM are stronger in two settings: for the elimination‐based aid than for the additive‐compensatory‐based aid and for users with low product knowledge than for those with high product knowledge, especially in terms of effort assessments. This research advances the theoretical understanding of the effects of interaction between two critical components of online decision aids (USIMs and decision strategies) and the moderating role of user characteristics (product knowledge) in affecting users' evaluations. It also provides practitioners with design advice for developing these aids.
