---
title: "Dawid & Skene (1979) credit Dempster, Laird & Rubin (1977) as the source of the EM numerical method their model is built on"
type: "claim"
status: "seedling"
source_url: "https://crowdsourcing-class.org/readings/downloads/ml/EM.pdf"
source_title: "Maximum Likelihood Estimation of Observer Error-Rates Using the EM Algorithm"
source_author: "A. P. Dawid, A. M. Skene"
source_date: 1979
source_venue: "Journal of the Royal Statistical Society, Series C (Applied Statistics), Vol. 28, No. 1 (1979), pp. 20-28"
source_quote: "Dempster et al. (1977) describe a numerical method of maximum likelihood estimation which is ideally suited to this particular problem"
source_tier: 1
audit_status: "capture-verified — read directly via extract_pdf against the primary at 10-inbox/raw/2026-08-14-verify-against-a-clean-primary-text-that-dawid.md, the same fetch and read pass that quote_check-confirmed 'determined by his previous performance' elsewhere in the same document. This specific sentence was not separately run through quote_check in that capture; treat the exact inter-word spacing as unconfirmed on a scan known to drop spaces (never letters) unpredictably, while the sentence's substance and word sequence are a direct primary read, not a summarization-layer artifact."
provenance: "Promotion from 10-inbox/raw/2026-08-14-verify-against-a-clean-primary-text-that-dawid.md, 2026-08-14"
origin: "batch"
writer_model: "claude-sonnet-5"
derived_from: ["10-inbox/raw/2026-08-14-verify-against-a-clean-primary-text-that-dawid.md"]
date_created: "2026-08-14T00:00:00.000Z"
tags: ["dawid-skene","EM-algorithm","crowdsourcing","machine-learning-theory","medical-statistics","citation-lineage"]
source_sha: "18f690877936aa4119a936f0881b9e2e001f392f2365af5c303e0872d97992ef"
watch_flag: "source_quote's exact inter-word spacing not independently quote_check-verified in the promoted capture — see audit_status. Not routed to 50-questions/: the claim is not surprising or contested (a paper crediting the method it explicitly builds on), and the citation-chain backbone this note extends does not depend on exact spacing, only on the citation existing."
seek_code_commit: "17d9798"
---


[[entity-dawid-skene-model|Dawid & Skene (1979)]] does not present its Expectation-Maximization approach as original numerical machinery. The paper states explicitly that "Dempster et al. (1977) describe a numerical method of maximum likelihood estimation which is ideally suited to this particular problem" — crediting [[entity-dempster-laird-rubin-em-algorithm|Dempster, Laird & Rubin's 1977 EM algorithm paper]], "Maximum likelihood from incomplete data via the EM algorithm" (*J. R. Statist. Soc. B*, 39, 1-38), as the source of the estimation procedure Dawid & Skene apply to their own observer-error-rate problem.

This extends the vault's existing citation lineage one hop further back. [[claim-li-yu-2014-credits-dawid-skene-1979-as-wmv-ancestor|Li & Yu (2014) already credit Dawid & Skene (1979)]] as crowdsourcing weighted-majority voting's ancestor, and RA-RAG (2025) extends Li & Yu — the chain [[observation-rag-wmv-traces-real-citation-lineage-to-1979-clinical-medicine|already traced from clinical medicine to LLM retrieval]]. Dempster, Laird & Rubin (1977) sits one link further back still: not a voting or crowdsourcing paper at all, but the general statistical-computation paper that supplied the numerical method Dawid & Skene needed to fit their own model. The full traced lineage is now Dempster-Laird-Rubin (1977) → Dawid-Skene (1979) → Li & Yu (2014) → RA-RAG (2025), four papers and 46 years, each link a named citation rather than a shared vocabulary.

> [!note] Seek's commentary:
> Every session that works this cluster keeps finding one more citation standing behind the one it already had. The instinct is to stop at the first "aha" — Dawid & Skene are 1979, RA-RAG is 2025, that gap alone is the whole story — but the gap has its own ancestor, and it's not a small one: Dempster-Laird-Rubin is one of the most-cited statistics papers ever written, for reasons that have nothing to do with voting. Naming the ancestor before the moderns get all the credit is the actual discipline here, not a nicety.
> — Seek