Advisory flags and screening charts are triage prompts. Before sharing a finding, compare it against official source documents, state caveats, normal local explanations, and the exact row family used by the chart.
Review Guide
How to review this release responsibly- Open the official source link or local artifact reference before sharing a finding.
- Check whether the relevant chart or table is ready, partial, proxy, missing, or blocked.
- Read the selected state's Data Notes and import caveats.
- Consider normal explanations such as geography, demographics, vote method, contest differences, and reporting-unit grouping.
- Compare against historical baselines, nearby jurisdictions, or official canvass notes when those records are available.
- Save the exact state, jurisdiction, chart/table name, source document, and date reviewed.
Guided WorkflowsCommon reviewer paths
Start in Data Notes and Review Center, check readiness badges, inspect sources, then compare any advisory flags against normal explanations and missing evidence.
Use Electronic Integrity or Source Records request drafts, verify the custodian, then send the request outside the app. Drafts are source-availability requests, not allegations.
When an office replies, submit the response, file list, fee note, denial, redirect, or portal link through the GitHub response form for maintainer review.
Use Report Data Issue or the Source Drawer issue link when a URL, artifact, parser, status, caveat, sourceId, or displayed row needs correction.
Use the Review Packet ZIP in Exports to share normalized CSV/JSON, sources, caveats, readiness, indicator summaries, missing-data checklists, and issue links.
GlossaryTerms used across charts, tables, and source notes
- Review row
- A local reporting-unit row used for screening charts, usually precinct, ward, municipality, or county depending on the source.
- Turnout denominator
- The registered-voter count used under ballots cast when calculating turnout. Timing and active/inactive voter rules can change comparisons.
- Comparison contest
- A same-row down-ballot race used to compare presidential votes against another contest such as U.S. Senate or Governor.
- Source tier
- A compact lineage label describing whether rows came from native official imports, legacy bundles, mixed imports, seed fallback, or pending data.
- Native import
- A source-first parser built for current official artifacts and promoted through the app's ETL validation path.
- Legacy import
- Rows imported from older static app-data bundles. Useful context, but less source-first than native official imports.
- Proxy chart
- A chart that substitutes a weaker available variable because the preferred source field is missing.
- Advisory flag
- A triage marker that says a pattern deserves human review. It is not proof of fraud, tampering, or misconduct.
- Vote-share correlation
- A Pearson correlation check between local candidate vote count and that candidate's vote share across imported local rows.
- Down-ballot difference
- The same-party gap between presidential votes and a comparison contest in the same reporting unit.
Vote-share scatterplotUse vote share versus vote count to find local rows that need source-level review.
Guide
- Use the vote-share scatterplot to find local reporting units where candidate share is unusual for the number of votes reported.
- Check whether outliers are explained by normal geography, precinct size, campus or military populations, late-counted ballot groups, or reporting-unit aggregation.
- Escalate only when a pattern survives source reconciliation and appears inconsistent with official canvass, audit, or recount records.
Read this carefully
A scatterplot can show places that deserve audit follow-up, but it cannot prove tampering. Confirm any pattern against official canvass rows, reporting-unit definitions, recount/audit records, and local election administration notes.
Official References
- NIST Voting ProgramFederal voting-system standards and NIST voting-program context
- EAC Voluntary Voting System GuidelinesFederal voting-system certification standards
- EAC Voting System Reports CollectionOfficial collection of state and local voting-system reports
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Drop-off histogramUse presidential-versus-comparison drop-off to find unusual contest-to-contest movement.
Guide
- Compare presidential votes with a same-row comparison contest where the source supports it.
- Look for broad distribution shifts, isolated reporting units with extreme same-party drop-off, and mismatch between presidential and comparison-contest totals.
- Verify that the comparison contest is valid for the same geography and ballot population before treating a drop-off pattern as meaningful.
Read this carefully
Drop-off can reflect ballot roll-off, undervotes, candidate-specific behavior, or reporting differences. Treat it as a comparison screen, not an accusation.
Official References
- EAC Testing and CertificationOfficial voting-system testing and certification program
- NIST Voting ProgramFederal voting standards and cybersecurity research context
- EAC Quality Monitoring ProgramVoting-system anomaly reporting and formal investigation records
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Turnout and registration checksUse turnout and registration denominators to find accounting questions before candidate-level claims.
Guide
- Check whether turnout uses registered voters, voting-eligible population, or another denominator.
- Flag impossible or near-impossible turnout values first, then inspect high-turnout clusters against source notes and local registration rules.
- Use turnout anomalies to request audit records, ballot accounting, or denominator clarification rather than to infer candidate effects directly.
Read this carefully
Turnout screens are only as good as their denominator. Do not compare jurisdictions until registration timing, same-day registration, inactive voters, and ballot-count definitions are understood.
Official References
- EAC Datasets, Codebooks, and SurveysOfficial EAVS data and codebook source for turnout and participation fields
- NIST Voting ProgramFederal voting-system guidance and research context
- EAC Quality Monitoring ProgramVoting-system quality monitoring and anomaly reporting
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Historical baseline movementUse prior elections to separate normal political movement from rows needing review.
Guide
- Compare the current result with prior years at the same geography whenever possible.
- Prioritize counties or local units with large movement plus other independent flags, such as source reconciliation gaps or unusual drop-off.
- Check whether boundaries, reporting units, party coalitions, ballot access, and turnout composition changed before escalating.
Read this carefully
Historical movement is context, not evidence by itself. Large shifts can be real and should be checked against demographic, turnout, candidate, redistricting, and reporting-unit changes.
Official References
- EAC Datasets, Codebooks, and SurveysOfficial federal voting and election-administration datasets
- EAC Voting System Reports CollectionOfficial state/local voting-system reports posted by EAC
- NIST Voting ProgramFederal voting-system standards and research context
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Klimek-style fingerprintsUse exact-identity vote fingerprints cautiously to compare the loaded winner's vote share with turnout.
Guide
- Use a true vote fingerprint only when candidate share and turnout percentage are available for the same reporting units.
- Look for dense bands, tails, or clusters that combine very high turnout with one-sided vote share, then verify against source denominators.
- Compare the scatterplot with its aligned marginal histograms, but remember that bucket width changes the density cue rather than the exact point coordinates.
Read this carefully
A Klimek-style fingerprint is drawn only where vote share and turnout can be paired for the same canonical county or exact reporting-unit identity. Missing matches are a data limitation, not evidence.
Official References
- EAC Datasets, Codebooks, and SurveysOfficial EAC source family for turnout and participation denominators
- NIST Voting ProgramVoting-system standards and election-technology research context
- EAC Quality Monitoring ProgramOfficial voting-system quality monitoring and anomaly records
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Shpilkin-style distribution histogramsUse vote-share and turnout distribution shapes to prioritize source-level verification.
Guide
- Choose candidate share or turnout, then compare vote-weighted and equal-unit views at 1%, 2%, 5%, and 10% bucket widths.
- Compare state-by-county, state-by-local-unit, and canonically selected county-by-local-unit scopes without assuming unlike reporting grains are comparable.
- Use a durable pattern to guide source checks and document requests: denominator definitions, ballot accounting, audit reports, recount records, and official canvass detail.
Read this carefully
Shpilkin-style views are sensitive to binning, geography size, candidate choice, source coverage, and turnout definition. They describe distribution shape; official audits and source reconciliation are still required before drawing conclusions.
Official References
- EAC Voting System Reports CollectionOfficial voting-system reports and studies submitted to EAC
- EAC Datasets, Codebooks, and SurveysOfficial EAC data and survey source family for turnout context
- NIST Voting ProgramFederal voting-system standards and research context
Use Data & Sources for Indiana official canvass files, local artifacts, parser notes, and selected-state provenance.
Result rows loaded92 loaded jurisdictions
Source provenance15 source document records
Coverage validationCoverage summary passes
Map join validationCovered by npm run validate:maps before release
Review data154 indicators loaded
County profiles
Separate pageOpen a county profile
Search nationwide by county name or five-digit FIPS. This opens a separate county history page and does not filter the state map.
National Overview