Skip to main content
eScholarship
Open Access Publications from the University of California

Scholarly Works (2 results)

  • Article
  • Peer Reviewed
Proceedings of the Annual Meeting of the Cognitive Science Society, Volume 46 (2024)
Content-curating algorithms provide a crucial service for social media users by surfacing relevant content, but they can also bring about harms when their objectives are misaligned with user values and welfare. Yet, potential behavioral consequences of this alignment problem remain understudied in controlled experiments. In a preregistered, two-wave, collaborative filtering experiment, we demonstrate that small changes to the metrics used for sampling and ranking posts affect the beliefs people form. Our results show observable differences in two types of outcomes within statisticized groups: belief accuracy and consensus. We find partial support for hypotheses that the recently proposed approaches of "bridging-based ranking" and "intelligence-based ranking" promote consensus and belief accuracy, respectively. We also find that while personalized, engagement-based ranking promotes posts that participants perceive favorably, it simultaneously leads those participants to form more polarized and less accurate beliefs than any of the other algorithms considered.
    Cover page: Simple changes to content curation algorithms affect the beliefs people form in a collaborative filtering experimentCreative Commons 'BY' version 4.0 license
    • Article
    • Peer Reviewed
    Proceedings of the Annual Meeting of the Cognitive Science Society, Volume 37 (2015)

    From choosing a book to picking a restaurant, most choices people encounter are about “matters of taste” and thus no universal, objective criterion about the options’ quality exists. Tapping into the knowledge of individuals with similar tastes who have already experienced and evaluated options—as harnessed by recommender system algorithms—helps people select options that they will enjoy. Although recommender systems are available in some domains, for most everyday decisions there is neither an algorithm nor “big data” at hand. We mapped recommender system algorithms to models of human judgment and decision making about “matters of fact” and then recast the latter as social recommender strategies for “matters of taste”. This allowed us to investigate how people can leverage the experiences of other individuals to make better decisions when no machine recommender systems are available. Using computer simulations on a widely used data set from the recommender systems literature, we show that experienced individuals can benefit from relying on only the opinions of seemingly similar people. Inexperienced individuals, in contrast, are often well-advised to pick the mainstream option (i.e., the one with the highest average evaluation) even if there are interindividual differences in taste; this is because reliable estimation of similarity requires considerable experience

      Cover page: You're special, but it doesn't matter if you're a greenhorn: Social recommender strategies for mere mortals
      Top