An election-data explainer
Michigan’s voting dashboard can lag local reports.
Livingston shows how a county can report ballots before the state catches up. Ann Arbor shows why combining local numbers does not automatically produce a better count.
See the local comparisonsCombined recorded participation
2,313,165
November 3 general election
- Absentee ballots returned
- 1,569,715
- Early in-person ballots cast
- 743,450
01 / Livingston County
A local report can get ahead of the state’s next refresh.
A county may publish later activity before it reaches the statewide release. Livingston offers a concrete example: its early-voting report showed a temporary lead that disappeared when the state caught up.
Early in-person ballots
The county report included 2,767 more early ballots than the available state snapshot. The gap reflected a temporary reporting lead.
View the comparison and source notes
| Source comparison | County | State | Gap |
|---|---|---|---|
| County ahead | 19,454 | 16,687 | 2,767 |
| Counts aligned | 23,057 | 23,057 | 0 |
These are cumulative early-voting totals, not ballots cast during the interval between updates. The county report covers all 18 city and township jurisdictions; shared-site subtotals are excluded to avoid double counting.
The county PDF does not print a precise update time, so the comparison cannot establish how long the lead lasted.
Sources: Livingston County’s early-voting report and the state dashboard. November 1–2, 2026 comparison; source pages may update.
Ann Arbor city
A higher count is not automatically a better count.
Ann Arbor illustrates a less tidy case. Its city page listed one more early ballot than the state, but 560 fewer absentee returns. With no precise cutoff attached to the city’s absentee figures, the difference did not establish which record was newer.
The broader principle is comparability. A report covering an entire county cannot be combined with a city total already inside it. And ballots received, processed and tabulated are different measures, even when the labels sound similar.
Taking the largest available number from each source would create a total whose meaning is uncertain. Each substitution needs a matching geography, measure and reporting period.
Sources: Ann Arbor City Clerk and the state dashboard. November 2, 2026 figures; Ann Arbor city, not Washtenaw County.
| Ballot measure | City | State |
|---|---|---|
| Early in-person | 19,481 | 19,480 |
| Absentee returned | 19,443 | 20,003 |
Different counts require an explanation before they can support a comparison.
The electorate is not yet complete
Pre-election totals leave part of the electorate unobserved.
General absentee-return deadline
Local time
Remaining absentee returns and Election Day voters can change both the size and the composition of the electorate. An early snapshot cannot settle either question.
Most absentee ballots must reach the local clerk by 8 p.m. on Election Day. Eligible mailed military and overseas ballots may arrive within six days afterward if postmarked by Election Day.
Michigan’s official absentee-return rules ↗
Read the voting-deadlines guideThe political interpretation
More early voting need not mean more total voting.
Consider an illustrative electorate that casts 40,000 ballots in each of two elections. If 10,000 voters move from Election Day to advance voting, an early total of 10,000 becomes 20,000. Early participation doubles; final turnout is unchanged.
Timing can change without turnout changing.
A larger advance vote can reflect existing voters changing methods. It does not, by itself, establish higher final turnout.
County geography is not candidate choice.
A county’s past partisan lean does not reveal how its early voters voted. That inference requires additional evidence and assumptions.
Return rates and statewide shares differ.
A county can return a high proportion of its requested absentee ballots while supplying a small share of the statewide total.
Historical comparisons need consistent terms.
Compare the same voting methods at similar points before Election Day. Changes in rules, adoption or reporting can still affect the comparison.
The tracker uses verified state snapshots for all 83 counties. Livingston and Ann Arbor illustrate reporting differences; their local figures have not been substituted into the statewide total. The 40,000-ballot example is hypothetical, not a forecast.
A rise in recorded ballots can reflect more voting, a reporting backlog clearing, or both. The size of an update alone cannot tell us which.
Daily updates.
No promised intraday schedule.
A fresh timestamp does not guarantee equally fresh records.
Michigan describes its dashboard as a daily update. That is a publication cadence, not a guarantee that every jurisdiction’s records are equally current.
Consider a clerk entering a batch of returns just before a refresh. The resulting jump may look like a sudden acceleration in voting, even if the ballots arrived over a longer period. Without the underlying timing, the public increase cannot establish a one-day surge.
The useful distinction is between ballots newly recorded and ballots newly cast. A series of snapshots shows changes in the published record; it cannot, by itself, reconstruct a smooth voting trend.
See the latest verified trackerThe reporting clock
Voting activity and reporting activity follow different timelines.
Receipt precedes public reporting.
A ballot received late in the day may first appear in a subsequent public update.
Administrative records can change.
Clerks log absentee returns in Michigan’s Qualified Voter File. Early-voting records enter through the electronic pollbook. Uploads and corrections can revise earlier totals.
The dashboard captures a snapshot.
A statewide release brings those records together. Its timestamp identifies a publication point, not when every included ballot was cast.
Source: Michigan Department of State’s dashboard methodology.
Reporting order can shape an apparent early lead.
Election results come from tabulated votes, not the voter-history records used in the participation dashboard. Michigan cautions that the two counts will not match.
Suppose an initial release consists largely of absentee votes from one city, while most Election Day precincts remain outstanding. Even a perfectly accurate tally would describe only that portion of the electorate. It would say nothing directly about the choices on unreported ballots.
The important question is what remains: which places, which voting methods and how much of the electorate. A percentage of precincts reporting is not necessarily the percentage of votes counted; precincts need not contain equal numbers of ballots.
The reporting-order scenario is illustrative. For Michigan’s distinction between participation records and results, see the official dashboard methodology. Election-night results remain unofficial until certified.
Find official county election resultsSources and verification
Reviewed November 2, 2026. The statewide snapshot and local comparisons are dated records; the reporting and turnout scenarios are illustrative. For current participation figures, use the tracker and its displayed source date.
- Michigan Department of State: Voting Dashboard and methodology. Participation figures, reporting cadence and the distinction from tabulator results.
- Livingston County: Early Voting, Daily Number of Voters. Countywide early-voting reports used in the November 1–2 comparison.
- Ann Arbor City Clerk: Elections. City participation figures used in the November 2 comparison.
- Michigan Department of State: Absentee voting. Ballot-return deadlines and the military/overseas exception.
- Michigan Department of State: County election-results directory. Routes to official local reporting.
Michigan Republican Primary is the publisher of this explainer. It is not an election authority. Source pages may update after this edition.
Back to the beginning