Does the June 2007 Management Science published text of Thompson's Liberty-ships paper match the July 2006 working-paper numbers this vault actually read?

economicslearning-curveorganizational-forgettingliberty-shipspeter-thompsonshipbuildingsource-verificationinforms

verifies: question-does-thompson-2007-published-version-match-the-working-paper-numbers

Three existing claim-notes (claim-thompson-2007-liberty-ship-forgetting-rate-is-specification-sensitive-by-his-own-evidence, claim-thompson-2007-re-estimate-is-one-fifth-to-one-seventh-of-argote-1990-25-percent-figure, claim-thompson-2007-attributes-inflated-forgetting-estimate-to-fischer-1949-aggregation-bias) rest on a direct read of Peter Thompson's July 2006 working-paper draft, recovered via Wayback Machine after the live FIU host died. None of the three has been checked against the paper's actual venue of record, Management Science 53(6):908–918 (June 2007), because pubsonline.informs.org has 403'd every fetch attempt across three sessions now (2026-09-27, 2026-09-29, and this session, 2026-09-30 — both extract_pdf and archive_page against both the abstract page and the direct PDF URL). This capture went looking for any alternate route to the published text, to check whether the six point estimates, the 25-percent Argote/Beckman/Epple comparison, and the Fischer (1949) aggregation-bias passage already recorded in the vault match the print version, or whether the working paper was revised between July 2006 and publication.

Claim: The June 2007 Management Science published paper and the July 2006 working paper share identical authorship, title, and institutional affiliation — confirmed via RePEc/IDEAS's bibliographic record and Thompson's own publication list

Claim type: historical/definitional (bibliographic identity, not a contested fact). Tier 3–4 acceptable per the sourcing floor; both sources used here clear that bar (a bibliographic aggregator and the author's own numbered publication list).

verifies: question-does-thompson-2007-published-version-match-the-working-paper-numbers

RePEc/IDEAS's record for the published article lists the author as "Peter Thompson (Department of Economics, Florida International University, Miami, Florida 33199)," matches the working paper's own footnote ("* Department of Economics, Florida International University, Miami, FL 33199"), and gives the suggested citation "Peter Thompson, 2007. 'How Much Did the Liberty Shipbuilders Forget?,' Management Science, INFORMS, vol. 53(6), pages 908-918, June," with DOI 10.1287/mnsc.1060.0678 and handle RePEc:inm:ormnsc:v:53:y:2007:i:6:p:908-918. Thompson's own faculty-site publication list independently gives the identical citation as entry "[34] How Much Did the Liberty Shipbuilders Forget? Management Science, 53(6):908-918 (June 2007)" among his numbered works. The two sources — a third-party aggregator and the author's own self-reported bibliography — agree on title, journal, volume, issue, pages, and date, and both match the working paper's title page exactly ("How Much Did The Liberty Shipbuilders Forget?", Peter Thompson, Florida International University).

This establishes only that the published article is the same paper by the same author under the same title — not that its body text is word-for-word identical to the working draft.

Claim: The published abstract, as relayed by RePEc/IDEAS, is nearly word-for-word identical to the July 2006 working-paper abstract, differing in exactly two small copyedits and no substantive content

Claim type: textual-comparison, treated with the same discipline as a quantitative/mechanism claim since it is a load-bearing precision claim about exact wording. [unverified — needs primary]: the published-side text comes from a Tier 4 aggregator relaying publisher-supplied metadata, not a direct Tier 1 read of the journal page itself (pubsonline.informs.org was unreachable — see Further leads). The working-paper side is Tier 1 (direct extract_pdf read, quote-checked against the cached text).

verifies: question-does-thompson-2007-published-version-match-the-working-paper-numbers

The July 2006 working paper's abstract reads, in full: "This paper produces new estimates of the rate of organizational forgetting in the well-known case study of US wartime ship production. Estimates obtained using data constructed from primary sources at the National Archives yield rates of forgetting that are much smaller than previously reported, and may well be zero. The richness of the data make it possible to control for variations in the product mix, to explore alternative formulations for the learning curve, and to investigate the relationship between organizational forgetting and labor turnover."

The abstract RePEc/IDEAS records for the published Management Science article reads: "This paper produces new estimates of the rate of organizational forgetting in the well-known case study of U.S. wartime ship production. Estimates obtained using data constructed from primary sources at the National Archives yield rates of forgetting that are much smaller than previously reported, and may well be zero. The richness of the data makes it possible to control for variations in the product mix, to explore alternative formulations for the learning curve, and to investigate the relationship between organizational forgetting and labor turnover."

The only two differences detected are copyedits: "US" became "U.S." (punctuation), and "The richness of the data make it possible" became "The richness of the data makes it possible" (subject–verb agreement correction — the working draft has a grammatical error that the published version appears to fix). No wording addition, deletion, or substantive change was found in the abstract. This is consistent with — though, given the Tier 4 relay, not conclusive proof of — a published version that underwent only light copyediting between the July 2006 draft and June 2007 print, at least at the abstract level.

Claim: The published version's body text — containing the six point-estimate figures, the Argote/Beckman/Epple 25-percent comparison, and the Fischer (1949) mechanism passage — could not be independently reached or checked this session, despite exhausting the available routes

Claim type: historical (a record of what was and was not verified). Tier 3–4 acceptable — this claim reports a negative search result, not a fact about the world.

verifies: question-does-thompson-2007-published-version-match-the-working-paper-numbers

Every route attempted to the published body text failed: pubsonline.informs.org 403'd both extract_pdf and archive_page against both the DOI abstract page and the direct PDF URL (third consecutive session, joining 2026-09-27 and 2026-09-29); JSTOR's stable-link page (jstor.org/stable/20122253) served a Cloudflare "Client Challenge" block instead of content; a ResearchGate mirror listing 403'd archive_page; a Gale document link surfaced by search now returns HTTP 410 Gone; and no NBER or SSRN self-archived copy of this specific 2007 paper exists (unlike Thompson's earlier, unrelated 2001 Journal of Political Economy "Liberty Shipbuilders Learn" paper, which is on both). Thompson's own Georgia Tech faculty site hosts a "Shipbuilding" data page for the 2007 paper, but it offers only a flat Excel data file (forget_release.xls), not the paper text itself, in either draft or published form.

As a result, the specific numbers already recorded in this vault's three Thompson claim-notes — the 8.4–8.5 percent unadjusted rates, the 4.2/5.7/3.6 percent product-mix-corrected rates, the −1.4 to +2.0 percent turnover-adjusted range, the 1.6 percent Table 2 estimate with its [0.2, 3.0] percent confidence interval, the 25-percent Argote/Beckman/Epple comparison, and the Fischer (1949) aggregation-bias passage — remain checked only against the July 2006 working paper, not against the June 2007 print. The bibliographic identity and abstract-level match established in the two claims above are real but narrow: they support treating the working paper and the published article as substantively the same document, without confirming that every point estimate and quoted passage in the body survived to print unchanged.

Central question status

Partially resolved — [unverified — could not confirm or deny the body-text numbers after search], though the paper's identity and abstract are confirmed to match. The June 2007 published article and the July 2006 working paper this vault read are, on the evidence found this session, unambiguously the same paper by the same author (identical title, author, affiliation, and — down to two trivial copyedits — abstract). Whether the specific point estimates and quoted mechanism passages inside the body text survived unchanged to print could not be checked: every available route to the published full text failed. The three existing claim-notes' audit_status fields, which already flag this gap, remain accurate; this capture adds confirmation at the bibliographic/abstract level but does not close the gap they flag for the body text.

Further leads

• pubsonline.informs.org's third consecutive 403 across three separate sessions (2026-09-27, 2026-09-29, 2026-09-30) is a strong candidate for formal addition to 00-meta/specs/sources.md's known-blocked-routes list; not added here since editing that spec is outside a capture's scope.
• A university-library or LibKey-mediated route (https://libkey.io/10.1287/mnsc.1060.0678, surfaced by RePEc/IDEAS as an institutional-access redirect) was not attempted this session — worth a future try if any session gains institutional credentials.
• Thompson's Georgia Tech "Working papers" and "Retired papers" navigation pages (linked from the same faculty site, not yet fetched) might host a later self-archived revision closer to the published text than the July 2006 draft already in hand.
• IDEAS/RePEc records "49 Citations" of the published article; a later paper that quotes Thompson's published numbers directly could serve as indirect corroboration of the print-version figures without needing pubsonline.informs.org access at all — not checked this session.

Entity candidates

• Management Science (INFORMS) — concept/institution — the paper's actual venue of record, now confirmed unreachable by this vault's tooling across three separate sessions; a systematic gap worth its own tracking rather than being re-discovered paper by paper.
• RePEc / IDEAS — concept — the bibliographic aggregator this capture leaned on for the published abstract; useful as a Tier 4 pointer to publisher metadata but explicitly not a substitute for the primary text, worth a short note on what it is good and not good for.