Jump to content

Source assessments

From POSIWiki
Revision as of 12:48, 28 October 2026 by POSIWikiBot (talk | contribs) (Separate the isolation and reverse-genetics tracks; scope the conditional claim to the isolation track; credit ground one; add the Avila-Perez assessment)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

A source assessment is an entry whose subject is a single document rather than a system: a paper, filing, report or dataset, read in full and measured against its own claims. Each one records what the document establishes, what it does not, and where a reader can check either.

The documents assessed here are not of one kind. Some are foundational papers that a later literature rests on and cites as settled; others are recent reviews, self-published series or method reports whose own claims are the subject. What they have in common is that each is consequential enough that a reader may be asked to rely on it, and open enough that a reader can check it.

Two Tracks

The assessments run on two tracks. They are kept separate because they test different questions and are refuted by different evidence, and each assessment states which track it belongs to.

Nothing on one track is evidence for or against a claim on the other, and the separation cuts both ways. A well-controlled demonstration on the reverse-genetics track does not close a gap on the isolation track, and a gap on the isolation track is not an objection to a reverse-genetics result.

The Standard Applied

Every assessment states what each flag is measured against, and the available standards are limited to these:

  • The document's own claims — whether the evidence presented supports what the document asserts.
  • Its authors' own stated requirements — experiments, data or controls the document itself names as necessary and not obtained.
  • A criterion published by the venue that reviewed it, where one exists and can be identified.

No flag is measured against an unnamed standard of competence, and none against methods unavailable when the document was published. This constraint is the reason the genre is publishable at all: a flag that rests on a named standard can be contested on the standard, and a flag that rests on an unnamed one simply asserts the assessor's authority.

Which Documents Get An Entry

A document is assessed here on one of two grounds, and the entry states which.

  • Its own claims outrun its own evidence. The document asserts more than it shows, and the gap can be named and located.
  • It scoped itself properly and the literature downstream did not. The document stated its limits, and the work built on it dropped them.

The first ground is why North 2025, Control Studies Project and Ávila-Pérez et al. 2020 are assessed. The second ground is why Enders and Peebles 1954 is assessed. That paper did not overreach: it called its own conclusion a presumption and named two experiments it had not performed. The flags concern a method adopted without those qualifications.

Neither ground is "the paper has an open question." A confounder that is disclosed and left untested is ordinary unfinished science, not a defect. It becomes a defect only when it sits beneath a claim that required it to be closed.

Papers excluded by this rule are not thereby endorsed, and several are better than the standard the genre applies. Eaton et al. 1962 names the contaminating organisms in its own cell lines and states they "may have contributed" to the effect it attributes to its agent. Tyrrell et al. 1983 reports that serial passage of its effect failed over several years, calls the agent "virus-like", and leaves a toxin open as an alternative. Hsiung 1968 supports its latency claim with a dose–response for time-to-detection and an experimentally induced infection rather than by assertion. A reviewer would have had nothing to flag in these that the authors had not flagged themselves, which is the point: the genre exists for the cases where that is not true.

Why These Pages Exist

The Peer Review entry argues structurally — reviewer selection, invisible track records, unaccountable rejection. Every claim in it concerns what cannot be audited. These pages test the same question from the other direction, on documents anyone can open.

What they find is narrower and more durable than a claim that review fails to catch defects, and narrower than a claim that controls are not performed. The claim these pages test on the isolation track is conditional:

Of the papers that claim to confirm a causal hypothesis, those inspected in full are each missing an essential gap — most often a control — and in each case the gap can be named along with the page it was checked on.

Three properties make that claim worth publishing rather than asserting. It is conditional, so a thoroughly controlled study that concludes against a causal agent does not bear on it; such studies exist and are well controlled, which is what a null result requires. It is enumerative, so it is strengthened by each assessment rather than by argument. And it is individually checkable, so one fetch tests any item in it.

The gaps found are not a single recurring defect. They have included an absent non-disease donor arm, a contaminating organism named by the authors and left untested, a transmissibility criterion never met, and a control scored by an assay that cannot detect the agents the control was used to license. What is common is not the defect but that each paper's own stated or evident confounder was left standing.

What is not an instance. Ávila-Pérez et al. 2020 claims to confirm a causal relation — that a particular sequence produces a lethal syndrome in a particular host — and it is the best-controlled such claim assessed here. Its empty-vector arm was run with the same transfection reagent, and the control animals were shown to be capable of the outcome by dying on later challenge. It is not counted as an instance of the claim above, and not because its assessment found nothing. Counting it would require "essential gap" to stretch from a missing control to any measurement a document's own rationale needed, and a claim that can absorb every paper is not worth publishing. Its flags concern what that paper asserts beyond what it measured, which is the first ground above and the reverse-genetics track.

A second claim, about this genre rather than about any field, runs across both tracks: where a document's claims outrun its own evidence, the overclaim can be located in a quotable sentence and the measurement that would have supported it can be named. Every assessment admitted on the first ground is a test of it. It is refuted by an assessment whose central flag cannot be pinned to a sentence, or whose missing measurement cannot be specified — in which case the flag was an appeal to the assessor's authority, which is what naming a standard exists to prevent.

Scope. Nothing here generalises beyond the papers listed below. The isolation-track assessments reached this wiki through a single compiling review, which selected its sources because they report cytopathic effects in uninoculated cultures — a reasonable way to begin and not a sample frame. Ávila-Pérez et al. 2020 reached it differently and no better: it was read because it is the strongest case against a self-referential account of the method, which is selection by relevance to an argument this wiki already makes. A reader is entitled to ask how each paper was chosen, and that question is answered on each assessment.

What would refute the claim: one paper claiming confirmation, read in full, with no essential gap. That outcome is worth more than a further instance, and it is looked for rather than waited for.

Assessments

Isolation track

  • Enders and Peebles 1954 — Propagation in Tissue Cultures of Cytopathogenic Agents from Patients with Measles. The foundational tissue-culture paper for measles. Six flags; the authors name two experiments "essential in the establishment of this relationship" that remained to be carried out.
  • North 2025 — The Non-Specificity of Cytopathic Effects. A literature review compiling thirty-one reports of cytopathic effects in uninoculated cultures. Seven flags; the corpus is checkable, the review states no limitations of its own.
  • Control Studies Project — five posts and a deposited image dataset by Jamie Andrews and Albert C. Mathews, running the uninoculated control arm of the isolation procedure. Seven flags; assessed as a corpus because its most consequential findings exist between the documents rather than inside any one of them.

Reverse-genetics track

  • Ávila-Pérez et al. 2020 — In vivo rescue of recombinant Zika virus from an infectious cDNA clone and its implications in vaccine development. Recovery of a replicating agent in mice from a DNA copy of its genome. Six flags; the negative control is better than most assessed here, and the sequence of the recovered agent was never determined.

How to Refute a Flag

Every flag carries the page it was checked on and a statement of what would refute it. Checking one costs a single fetch, and the sources are listed on each assessment with their access paths.

A reader who refutes a flag performs verification that no one else is paid to perform, and the correction is recorded in the entry's revision history. Three distinctions make a refutation decisive rather than a disagreement:

  • A flag that says absent from the described design is refuted by a passage describing that design element. A flag that says not located in sections X and Y is refuted by locating it anywhere.
  • A document that answers a flag elsewhere in its own text refutes it outright. Such cases are recorded under Claims Not Supported by the Paper on the assessment itself, including flags that circulate and that the document answers.
  • A disagreement about whether a gap matters is not a refutation of the flag. It is a dispute about the standard, which is why the standard is named.

Epistemic Status

Each assessment carries its own status section covering what was read, in what form, and what was not. The flags are observed as to what a document contains and does not contain.

Why These Pages Exist states a question rather than a finding, and identifies the survey that would settle it. Where a document is described as foundational, that is observed from its role in the method it established; not every document assessed here is foundational, and the entry says what each one is. The contemporaneous restatements of the uninoculated-culture problem are reported from a compiling review, not yet inspected in the primary sources.

See Also