Every data platform has boundaries. Understanding what this platform cannot tell you is as important as understanding what it can. This reference documents the key limitations to keep in mind.
Data Freshness
Campaign finance has filing periods, not real-time updates.
Committees report contributions and expenditures on fixed schedules. Between filing deadlines, financial activity is not visible. Late contribution reports (Form 497) provide some real-time visibility for large gifts, but routine activity only appears with periodic filings.
Cash on hand is a snapshot, not current balance.
The cash figure on a committee profile reflects the balance at the end of the last filing period—which may be weeks or months old. During active campaigns, this number changes daily.
Election results appear after certification.
Results are loaded after official certification, typically 4-6 weeks post-election. This is not a live election night dashboard.
Census data represents a past time period.
Population counts are from 2020. Socioeconomic data (ACS) represents 2019-2023 averages. Fast-changing areas may look significantly different today.
Coverage Gaps
Not all local jurisdictions are included.
Local campaign finance data covers approximately 185 jurisdictions that file electronically through Netfile. Jurisdictions not on Netfile are not included.
Precinct data isn't available for all races and years.
Precinct-level maps and Election Reports require precinct-level results data, which varies in availability. Recent elections (2018+) generally have better coverage.
Local candidate discovery is incomplete.
State candidates appear when they file official paperwork. Local candidates are discovered through news coverage, announcements, and local filings—a less systematic process that may miss candidates, especially early in the cycle.
Some election results aren't imported.
Unopposed races, races with only top-line results reported, or races from jurisdictions without electronic data may not appear.
Analytical Limitations
Correlation is not causation.
Precinct-level analysis shows patterns (e.g., "Democratic precincts voted Yes") but cannot prove why voters chose as they did. Other factors may drive the relationship.
Ecological fallacy applies to all precinct data.
Aggregate precinct data cannot reveal individual voter behavior. "High-income precincts voted Yes" doesn't mean high-income individuals voted Yes—other characteristics of those precincts might be the actual driver.
Ethnicity is modeled, not reported.
The Secretary of State doesn't collect ethnicity. Ethnic breakdowns are derived from surname-based modeling, which has accuracy limitations—particularly for multiracial individuals and groups with overlapping naming patterns.
Projections are estimates, not predictions.
The Turnout Projection Tool shows what historical patterns imply. It cannot account for unprecedented events, unusual candidates, or conditions that differ from the past.
Small jurisdictions have more uncertainty.
Statistical analyses (correlations, projections, percentile rankings) are less reliable when based on small numbers of precincts or voters.
What the Platform Does Not Include
Individual voter records.
The platform shows aggregate statistics, not individual voter information. You cannot look up how a specific person is registered or whether they voted.
Polling or survey data.
No polling results are incorporated. All data comes from official government records.
News or editorial content.
The platform provides data, not news coverage or analysis of campaigns. Candidate source citations link to external news, but the platform doesn't produce journalism.
Contribution limit compliance.
The platform displays what was reported, not whether contributions comply with legal limits. Compliance analysis requires understanding office-specific rules.
Real-time precinct results on election night.
Only certified post-election results are displayed.
Data Quality Caveats
Donor information is self-reported.
Employer and occupation data comes from donors themselves and varies in quality. Some donors report vague occupations or leave fields incomplete.
Amendments supersede original filings.
The platform uses the most recent amended filing for calculations. If you access an original PDF, you may see different numbers.
Precinct boundaries change.
Precinct lines can change between elections—not just with redistricting. Direct precinct-to-precinct comparisons across elections should be done cautiously.
Registration data is point-in-time.
Voter registration figures reflect the state of the voter file as of each past election. Current registration may differ.
Interpretation Guidelines
Registration doesn't equal votes.
A D+15 registration district doesn't necessarily produce D+15 election results. Turnout patterns, persuasion, and ticket-splitting all affect outcomes.
Past performance doesn't guarantee future results.
Historical patterns inform analysis but don't predict outcomes. Candidates, issues, and turnout conditions change.
Relative standings are comparative, not absolute.
A "Very High" ranking means top 20% among similar jurisdictions—not that the raw number is objectively high. A "High" median income in a rural region might still be modest compared to urban California.
Primary and general elections are different universes.
Turnout, electorate composition, and competitive dynamics differ fundamentally between primary and general elections. Compare like to like.
When to Seek Other Sources
- Legal compliance questions: Consult the FPPC or an election law attorney
- Real-time election night results: Use county or Secretary of State websites
- Individual voter information: Contact county election offices
- Polling and public opinion: Use professional polling sources
- Historical context and news: Use news archives and journalism