A User Study on Explanations with Different Levels of Detail in Recommender Systems

Explanations are a useful feature to improve the transparency and reliability of recommender systems. However, the amount of information presented to the user regarding a recommendation is an important parameter that is not usually considered. On the one hand, large explanations may carry more information than the users would like, which could be a potential case of the information overload problem. Similarly, if short explanations are provided, some users may find insufficient information to make decisions. In this context, our main objective is to analyze perceptions of users that were presented to explanations with varying level of detail. In our validation process, we conducted user experiments and identified a correlation between users reporting that the ability to choose the level of detail in the explanation aids in finding a recommendation that aligns with their interests, and their subsequent acceptance of the recommendation.