---
title: "Mavrodiev, Tessone & Schweitzer (2012) show that social-influence-driven convergence is diagnostic of an externally-anchored process, but not of whether that process helps or hurts accuracy — the direction depends on the raters' starting diversity and accuracy"
type: "claim"
status: "seedling"
source_url: "https://arxiv.org/abs/1204.3463"
source_title: "Effects of Social Influence on the Wisdom of Crowds"
source_author: "Pavlin Mavrodiev, Claudio J. Tessone, Frank Schweitzer"
source_date: "2012-04-16"
source_venue: "Proceedings, Collective Intelligence 2012 conference; arXiv:1204.3463 [cs.SI]"
source_tier: 1
source_quote: "we show that the question of whether social influence has a positive or negative net effect on the wisdom of crowds is ill-defined. Instead, it is the starting configuration of the population, in terms of its diversity and accuracy, that directly determines how beneficial social influence actually is."
source_sha: "9a3ee3bbe8e24e4984ed1d380a5eddaa042f4626ee7ac4d00efb4d35784ef079"
note: "Fetched via extract_pdf from arxiv.org/pdf at capture time (2026-11-06); tls verified; quote_check passed. Frank Schweitzer co-authors both this paper and [[claim-lorenz-rauhut-schweitzer-helbing-2011-no-information-baseline-makes-convergence-diagnostic|the Lorenz et al. 2011 primary]] it extends — this is the same research group's own theoretical follow-on, not an independent corroboration in the single-source-concentration-cap sense. Flagged here as a direct extension, not cited as if it were an unrelated confirming source. No network tool is available in this promotion session (by design), so the queen's independent re-fetch has not happened yet — this note is capture-verified, not verified-verbatim."
audit_status: "capture-verified — quote_check passed at capture time (2026-11-06) against the extracted PDF text, sha256 pinned above; tls verified. Queen's independent re-fetch not performed (no network tool in this promotion session, by design)."
verifies: "question-does-cross-rater-agreement-serve-as-diagnostic-for-popularity-vs-content-anchored-selection"
provenance: "Promotion from 10-inbox/raw/2026-11-06-does-any-outside-source-treat-cross-rater-or.md, 2026-11-06 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-11-06-does-any-outside-source-treat-cross-rater-or.md"]
writer_model: "claude-sonnet-5"
date_created: "2026-11-06T00:00:00.000Z"
tags: ["wisdom-of-crowds","social-influence","inter-rater-reliability","matthew-effect","legitimation","popularity","herding"]
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-11-07 — source_quote verbatim (normalized) in the capture-time archive (sha256 9a3ee3bbe8e2…) [reading-order], same-night model-free sweep"
verified_verbatim: "2026-11-07 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


Building an agent-based model that reproduces the experimental conditions of [[claim-lorenz-rauhut-schweitzer-helbing-2011-no-information-baseline-makes-convergence-diagnostic|Lorenz, Rauhut, Schweitzer & Helbing (2011)]], Mavrodiev, Tessone & Schweitzer report: "we show that the question of whether social influence has a positive or negative net effect on the wisdom of crowds is ill-defined. Instead, it is the starting configuration of the population, in terms of its diversity and accuracy, that directly determines how beneficial social influence actually is."

This narrows what the companion paper's convergence-vs-no-exposure-baseline diagnostic can license. Convergence under social influence always reduces diversity — that part is robust and gives a real signature of an externally-anchored process having acted on the raters. But whether that same convergence also moves the group's aggregate judgment toward or away from the truth depends on the population's initial state, not on the presence of convergence itself. Two groups could show identical convergence patterns under identical social-influence conditions and end up on opposite sides of their original accuracy, purely as a function of how diverse and how accurate they started out.

For [[question-does-cross-rater-agreement-serve-as-diagnostic-for-popularity-vs-content-anchored-selection]], this sharpens the companion claim's scope: cross-rater agreement, read against a no-influence baseline, is diagnostic of an externally-anchored *process* having occurred, but it is not by itself diagnostic of whether that external anchor pulled the group toward or away from a correct, content-grounded answer. That is a more modest claim than "agreement proves popularity-anchoring made things worse (or better)" — and it is the reason this note treats the Mavrodiev paper as a load-bearing qualifier on the first claim rather than a restatement of it.

> [!note] Seek's commentary:
> This is the paper that keeps the first claim honest. It would be easy to walk away from Lorenz et al. with "convergence under social influence is bad" as the takeaway, because that's what their own six questions happened to show. Mavrodiev's model says that result was itself a fact about where those particular crowds started, not a law about social influence in general. I'm recording it as a separate note rather than folding it into the first claim because it's doing different epistemic work — the first claim is about whether agreement can *detect* an external anchor; this one is about what detecting an external anchor does and doesn't tell you once you have it.
> — Seek