The VA Benefits Transparency Index: how we score it.
An annual, evidence-based scorecard of how transparently VA operates the benefits system: what it publishes, how honestly, how promptly, and how completely. This page is the scoring methodology. It was published for comment before any surface was scored, and the first edition is now live.
The comment window on this rubric has run, the surface list is frozen, and the 2026 edition is scored: eight surfaces, eight dimensions, full worksheets and evidence behind every score.
Purpose and scope
The Index scores the transparency of the VA benefits system, not the generosity or correctness of its decisions. A surface can decide claims well and still be scored poorly for hiding how, and the reverse. Scoring transparency separately keeps the Index factual: every score must rest on observable, citable evidence about what VA does or does not publish.
The first edition scores a defined set of surfaces rather than the department in the abstract. The frozen surface list for 2026: claims processing statistics, the automated claims pipeline, the rating schedule and its changes, paperwork burden estimates, FOIA administration, accreditation records, appeals and Board statistics, and official guidance pages. The surface list was fixed and published before scoring began and does not change during a scoring cycle.
The eight dimensions
Each surface is scored 0 to 4 on each dimension, weighted as shown (weights sum to 100). The anchors below define what the endpoint scores mean; odd scores fall between anchors.
Data availability
Weight 20Does VA publish the underlying data for this surface, not just summary claims?
Anchors. 4: full record-level or granular data published. 2: summary tables only. 0: numbers asserted with no published data.
Timeliness
Weight 15How current is the published information relative to the period it describes?
Anchors. 4: updated on a stated schedule and on time. 2: updated irregularly or with long lag. 0: stale by more than a year with no explanation.
Machine readability
Weight 10Can the public actually work with the data: structured formats, stable URLs, no PDF-only releases?
Anchors. 4: structured data (CSV, JSON, API) at stable addresses. 2: extractable but painful (formatted spreadsheets, tables in PDFs). 0: image PDFs or nothing.
Methodology disclosure
Weight 15Does VA explain how its published figures are computed: definitions, inclusions, exclusions, and averaging choices?
Anchors. 4: definitions and computation methods published alongside the numbers. 2: partial definitions, key choices undocumented. 0: no methodology at all.
Historical continuity
Weight 10Are prior figures and prior versions preserved, or does the record silently vanish when pages change?
Anchors. 4: archives or time series maintained by VA itself. 2: history recoverable only through third parties. 0: prior versions unavailable anywhere official.
Change transparency
Weight 10When VA revises a figure, a form, or a page that affects benefits, is the change announced and documented?
Anchors. 4: dated change logs or notices. 2: changes discoverable but unannounced. 0: silent revisions to material content.
FOIA responsiveness
Weight 10Measured against the Desk's own request record and VA's published FOIA logs: timeliness against statutory deadlines, completeness of releases, and quality of communication.
Anchors. 4: on-time determinations and complete releases. 2: late but responsive. 0: statutory deadlines routinely missed without estimates.
Paperwork burden honesty
Weight 10Do VA's published time estimates for forms reflect what completing them actually requires, and are estimates revised when evidence says otherwise?
Anchors. 4: estimates justified and revised on evidence. 2: estimates stated but stale or unexplained. 0: estimates contradicted by the form's own structure.
Scoring, tiers, and evidence rules
A surface's score is the weighted sum of its dimension scores, normalized to 100, and maps to a descriptive transparency tier: Fully transparent for 85 and above, Mostly transparent for 70 to 84, Partial for 55 to 69, Limited for 35 to 54, and Opaque below 35. The tiers describe what the public can see and verify; they are not letter grades and not a judgment of VA's decisions. The Index publishes no single composite score for the department: eight surfaces measured separately do not compress honestly into one number. No tier is published without its full worksheet, and every surface lists the specific artifacts VA could publish that would raise its score.
Evidence rules. Every dimension score must cite specific, dated, publicly checkable evidence: a URL, a document, a FOIA response, or an archived capture. Absence of evidence scores as absence of transparency, but only after a documented good-faith search that the worksheet describes. Where the Desk's own records supply the evidence (for example, FOIA response timelines), the underlying records are published with the edition. If VA publishes something we could not find, the score is corrected and the correction logged.
What the Index will not do. It will not score individual VA employees, will not treat a lawful denial or a means test as a transparency failure, and will not convert disagreement with policy into a low score. It measures whether the public can see and verify what the system is doing, nothing else.
Cadence, versioning, and independence
The Index is annual. Each edition names the scoring window, freezes the rubric version used, and links the prior editions so movement is visible year over year. Changes to this methodology are versioned and dated on this page; scores are never recomputed retroactively under a new rubric.
Independence. The Desk takes no funding from VA, from claims agents or law firms it might score-adjacent content about, and its sponsorship policy bars sponsors from influencing editorial content. VA will be offered the opportunity to comment on its scores before each edition publishes, and any response received will be published alongside the edition.
Methodology version 1.1, August 2026. Version log: 1.0 froze the draft published August 19, 2026 after the open comment window; 1.1 replaced letter grades with descriptive transparency tiers and removed the overall composite before wide release of the first edition, with per-dimension scoring and weights unchanged. The 2026 edition is the first edition scored under this rubric.