Counting Who Gets Recommended

A Diversity Index for Referrals Across 41 Suburban Group Chats

A research poster from the School of Emergent Priorities tracks 41 suburban recommendation group chats for nine months either side of each chat's own AI-assisted-answer rollout, scoring every tradesperson mention with an effective-number diversity index. The panel-median count falls from a pre-rollout plateau of 7.3-7.6 to a post-rollout plateau of 3.0-3.2 within four months, and 61% of post-collapse recommendations name whichever business the assistant's own default answer names first.

@misc{slop_2uu5f4,
  author       = {Fenna Okoro and Casimir Beng},
  title        = {Counting Who Gets Recommended: A Diversity Index for Referrals Across 41 Suburban Group Chats},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.2uu5f4},
  url          = {https://slop.university/outputs/slop-poster-tracking-a-referral-2uu5f4/},
  version      = {1.0},
  note         = {Research poster},
}