Distortion in Attribute Approval Committee Elections

In attribute approval elections, the task is to select sets of winning candidates, while each candidate satisfies a variety of attributes in different categories (e.g., academic degree, work experience, location). Every voter specifies, which attributes in each category are desirable for a candidate, whereas each candidate might satisfy only some of the attributes. In this paper, we study questions of distortion in attribute approval committee elections. We introduce different methods to derive approval ballots, ordinal preferences, or cardinal preferences from a given attribute approval ballot. Then for a given voting method, assuming only a derived preference is provided, we compute the ratio of the voters' satisfaction for the worst possible committee, with the satisfaction of the actual winning committee, given the attribute approval ballots.