When a chat-native comparison table lets users pin a small shortlist and then run a pinned-shortlist conversational refinement pass that explicitly calls out freshness cues and volatile attributes for just those pinned items, do users achieve better-calibrated decision confidence and lower over-trust than in flows that keep re-generating full tables— or does focusing refinement on a narrow shortlist mainly harden early anchoring on potentially stale or soft-matched options?

Answer

Pinned-shortlist refinement on a small set, with explicit freshness and volatility cues, tends to improve calibration and reduce some blind over-trust for that shortlist, but it also strengthens anchoring on the initially pinned items. Net benefit depends on how easy it is to challenge or repin the shortlist and how clearly the system surfaces staleness or soft matches.

Compared with repeated full-table regeneration:

• Users who already have a reasonable pinned set usually get:
  • higher decision confidence (the agent helps them inspect trade-offs in depth), and
  • somewhat better-calibrated trust about freshness/volatility for those items.
• Users who pin too early or from a noisy first table are more likely to:
  • over-anchor on that shortlist, and
  • under-explore better but unseen options, even if some pinned items are stale or soft matches.

Design implication: treat pinned-shortlist refinement as a second-pass tool, but keep light-touch prompts and cues (“this item looks stale; want to see fresher alternatives like it?”) that reopen the option space when staleness or soft fit is exposed.