We publish what we measured, the uncertainty on it, and what we have not measured.

What this laboratory is

Two ideas underlie everything here. One is a framework for treating belief as a measured quantity; the other is the laboratory that applies it. Neither requires prior familiarity with metrology or with information conflict.

CWIMF - the measurement framework

The Cognitive Warfare Information Metrology Framework treats a claim circulating in the information environment the way a laboratory treats any measurand. A proposition is anchored to a verified truth value; the prevalence of belief in it is measured against that anchor; and the result is reported with a full uncertainty budget in the sense of the Guide to the Expression of Uncertainty in Measurement (GUM). The central quantity is belief displacement: how far a population’s belief sits from what the verified evidence supports. Displacement is reported the way any measurement is, with a stated coverage interval and an explicit account of what has not been evaluated. The framework borrows its discipline directly from formal metrology - the International Vocabulary of Metrology (VIM), the GUM, and the laboratory-competence standard ISO/IEC 17025 - so that a figure about belief carries the same traceability a calibration certificate does.

CWIAL - the laboratory

Perialetheia is a Cognitive Warfare Information Analysis Laboratory: the first working instance (No. 1) that applies the framework to real cases. Where CWIMF is the method written down, CWIAL is the bench on which measurements are actually made - references verified, instruments calibrated, and reports issued clause by clause under an ISO/IEC 17025 structure. Each monthly report re-audits one contested figure from the public record and states what can, and what cannot yet, be measured from it.

The name Perialetheia joins peri‑ (around, concerning) with alētheia (truth, literally un‑concealment). The laboratory measures how belief stands in relation to what is verified, and treats every gap in that relation as something to be declared rather than hidden.

The Brexit £350m claim

One displacement that can be measured, and one comparison, widely repeated, that cannot be made at all.

Why the comparison does not close

Respondents in 2018 who had not heard the claim were not asked whether they believed it. What share of them would have endorsed it on being shown it is unmeasured, and it is the whole of the gap between the two bases. Bounding that share at nothing and at everything gives the widest interval the record permits, and that interval is wider than the change under discussion.

Anchor measurement

The 2018 wave asks belief in F directly of respondents who report awareness of it, so it maps onto the proposition without a bridging assumption, provided the population is stated as what it is: the claim-aware subpopulation, not the electorate. A displacement can be computed from it. A rate of change cannot, for the reason above.

@techreport{urekar2026cwial02,
  author      = {Urekar, Marjan},
  title       = {Belief Displacement in the Brexit {\pounds}350m Claim (Report No. 2)},
  institution = {Perialetheia - Cognitive Warfare Information Analysis Laboratory No.~1},
  year        = {2026},
  number      = {PERIALETHEIA-CWIAL-2026-02},
  note        = {Uncertainty budget after JCGM 100:2008 (GUM); capability after Currie 1968 / ISO 11843-1},
  doi         = {10.5281/zenodo.22303739},
  url         = {https://doi.org/10.5281/zenodo.22303739}
}

TY  - RPRT
AU  - Urekar, Marjan
TI  - Belief Displacement in the Brexit GBP 350m Claim (Report No. 2)
PB  - Perialetheia - Cognitive Warfare Information Analysis Laboratory No. 1
PY  - 2026
M1  - PERIALETHEIA-CWIAL-2026-02
DO  - 10.5281/zenodo.22303739
UR  - https://doi.org/10.5281/zenodo.22303739
N1  - Uncertainty budget after JCGM 100:2008 (GUM); capability after Currie 1968 / ISO 11843-1
ER  -

GM food safety beliefs in Europe

A measurement error in the public record, and one anchor measurement that survives it.

The only pair that is one instrument read twice

Bounding how fast this belief can move organically requires two readings of an instrument that did not change between them. The 2005 and 2010 concern ratings do not qualify: the list grew from 14 items to 17, the wording moved from products to organisms, and the population went from EU-25 to EU-27. A report that treats a change of referent as a change of instrument cannot treat three simultaneous changes as none.

One pair does qualify. The four-judgement battery of 1996 and 1999 - useful, a risk for society, morally acceptable, should be encouraged - was administered unchanged, and its risk for society item rose from 67% to 73% across 37 months of fieldwork. That is 1.95 percentage points a year, disclosed here as 2.0 to stay conservative. Two things must be said alongside it. The direction is upward, against the received narrative of declining concern. And the share expressing no view rose from 12% to 20% over the same interval, which inflates the agreeing share among those who answered.

The 1996 and 1999 marginals are extracted from the GESIS codebooks (ZA2899, ZA3205) and are not available from secondary sources. Anyone checking this result should go to the codebooks, and this laboratory would rather be corrected than believed.

Anchor measurement

One wave asks directly about health safety and therefore maps onto proposition F without a bridging assumption. It is the only wave from which a belief displacement can currently be computed.

@techreport{urekar2026cwial01,
  author      = {Urekar, Marjan},
  title       = {Belief Displacement in European GM-Food Safety Discourse (Report No. 1)},
  institution = {Perialetheia - Cognitive Warfare Information Analysis Laboratory No.~1},
  year        = {2026},
  number      = {PERIALETHEIA-CWIAL-2026-01},
  note        = {Uncertainty budget after JCGM 100:2008 (GUM); detection after Currie 1968 / ISO 11843-1},
  doi         = {10.5281/zenodo.22302286},
  url         = {https://doi.org/10.5281/zenodo.22302286}
}

TY  - RPRT
AU  - Urekar, Marjan
TI  - Belief Displacement in European GM-Food Safety Discourse (Report No. 1)
PB  - Perialetheia - Cognitive Warfare Information Analysis Laboratory No. 1
PY  - 2026
M1  - PERIALETHEIA-CWIAL-2026-01
DO  - 10.5281/zenodo.22302286
UR  - https://doi.org/10.5281/zenodo.22302286
N1  - Uncertainty budget after JCGM 100:2008 (GUM); detection after Currie 1968 / ISO 11843-1
ER  -

How a measurement is made

Five stages, in order. Each one has to complete before the next is meaningful.

1. Anchor the claimMatch the claim to a Verified Factual Reference Base entry: a proposition with a stated truth value, a documented source chain of at least three independent Grade I–II sources, and a re-verification date.
2. Define the population and the instrumentState exactly who is being measured and with what. A survey question format is an instrument; changing it changes the reading. This is where most published trend claims fail.
3. Compute the displacementBelief prevalence minus the prevalence that would align with the verified truth. For a verified-false proposition this is the prevalence itself.
4. Build the uncertainty budgetType A from the sampling. Type B for reference classification risk, instrument bias, temporal mismatch, representativeness, and adversarial adaptation. Combine per GUM.
5. Report, including the gapsAn ISO/IEC 17025 clause 7.8 report: measurand, reference, method, budget, result with coverage factor, detection capability, and an explicit statement of what was not evaluated.

The standards this laboratory works to

Every term used here maps onto an established measurement standard. Nothing in the method is invented where an accepted definition already exists.

Where this laboratory departs from a standard, it says so and why. Where it follows one, the mapping above is the audit trail.

The register and the toolkit

The reference base every measurement here is traceable to, and the software that computes the budgets, are deposited as versioned records with permanent identifiers. Each release is archived separately; the identifiers below resolve to the current version.

Sources in the register support the truth value of a proposition and nothing else. The survey or classifier used to read belief is the measuring instrument and is recorded in the instrument registry, never counted as a source for the reference. Where an entry falls below the three-source working minimum, the shortfall is stated in the entry rather than made up with additional citations.

The monthly reports are deposited as their own records and carry their own identifiers: Report No. 1, doi:10.5281/zenodo.22302286, and Report No. 2, doi:10.5281/zenodo.22303739. Each cites the register version and the software release its figures were computed against.

Development happens in the open: the toolkit repository and this site repository. The archived records are the citable form; the repositories are the working one.

The method, written down

M. Urekar · Faculty of Technical Sciences, University of Novi Sad · ORCID 0000-0002-6089-9148

• IcETRAN 2026 Ten Metrological Principles for Understanding and Countering Cognitive Warfare 13th Int. Conf. on Electrical, Electronic and Computing Engineering (IcETRAN), Silver Lake, Serbia · to be indexed in IEEE Xplore

This is the only paper published to date; the register and the toolkit are archived separately and listed above. Further papers - the framework specification, the laboratory information-integrity study, and the empirical re-audits behind these monthly reports - will be listed here as each is published.

Working with the laboratory

No prior expertise in cognitive warfare is required. The measurement discipline is learnable; the willingness to state what you do not know is the harder part.

The laboratory is not running a public call at this time. Involvement is arranged in person, case by case, with contributors who want to apply the measurement discipline to a documented case, curate and verify references, or help calibrate the instruments the method depends on. If that interests you, get in touch directly.