United States · Try it · turnout

If everyone voted

What would have happened at every presidential election since 2000 if different people had voted? This page takes the real result in each state and changes one thing: how many people in each group turned out. Everyone else votes exactly as they did.

Pick an election from 2000 to 2024.
Pick a replay: Black, Hispanic or Asian American voters turning out at the same rate as white voters, men voting as often as women, everyone voting, or only one group voting. Or move each group's slider yourself.
See the result: the electoral votes, who wins and the states that change hands, on the map.

Why turnout decides elections

In 2000 the presidency came down to 537 votes in Florida. Raise any one group's turnout to the rate at which white citizens voted that year, whether Black, Hispanic or Asian American, and Al Gore carries Florida and the White House.

The gap is wide. In 2024, 71% of white citizens voted, against 60% of Black citizens, 51% of Hispanic citizens and 56% of Asian American citizens. It has not always been this way: in 2012 Black turnout was higher than white turnout, 64% to 62%. By 2016 it had fallen to 59%. Replay 2016 with Black voters turning out at the white rate and Hillary Clinton wins 287 to 251, carrying Florida, Michigan and Wisconsin.

But turnout is not a lever for one party, because the people who stay home don't vote like the people who show up. In 2024, Black non-voters leaned Democratic by 43 points, against 67 among Black voters, and Hispanic non-voters were almost evenly split while Hispanic voters leaned Democratic by 16 points. Bring every 2020 non-voter to the polls and Donald Trump, not Joe Biden, wins: Georgia, New Hampshire, Pennsylvania and Wisconsin change hands.

Women have outvoted men at every election here: 67% to 64% in 2024. Had only women voted, the Democrat would have won every election since 2000. Had only men voted, the Republican would have won every one except 2008.

Pick any election below and change who votes. The replays hold everything else fixed, so read them as the size of what turnout can do, not as what would have happened. Every figure comes from the Census Bureau's turnout survey and the Cooperative Election Study, about 60,000 people a year whose voting is checked against the voter files.

The replay

Change who votes, and watch the map

Pick an election, then a replay, or set each group's turnout yourself. Turnout is the share of adult citizens who voted.

Break down by

White voters 71%actual 71%

Black voters 60%actual 60%

Hispanic voters 51%actual 51%

Asian American voters 56%actual 56%

Everyone else 58%actual 58%

Extra voters vote like the group's

People who don't vote often lean differently from people who do: in 2024, Black and Hispanic non-voters were far less Democratic than Black and Hispanic voters. "Non-voters" uses what the group's non-voters told the Cooperative Election Study before the election.

226Kamala Harris

Donald Trump wins

312Donald Trump

Popular vote R+1.5Voters 155.2m

Loading the map…

Every election, every replay

Who would have won

Electoral votes in each replay, with extra voters voting like the group's non-voters. A ring marks a different winner. Tap a cell to open it above.

Replay	2000Gore v Bush	2004Kerry v Bush	2008Obama v McCain	2012Obama v Romney	2016Clinton v Trump	2020Biden v Trump	2024Harris v Trump
What happened
Black turnout at the white rate
Hispanic turnout at the white rate
Asian American turnout at the white rate
Every group at the white rate
Every group at the highest rate
Everyone votes
Only white voters
Only Black voters
Only Hispanic voters
Only Asian American voters
Only voters of color
Men and women
Men at women's turnout
Only women
Only men

Different winners

When turnout alone changes the president

The turnout gap

Who votes, election by election

Share of adult citizens who voted, by group. From the Census survey, scaled in each state to the votes actually counted.

Who stays home

Voters and non-voters don't lean the same way

Democratic lead (or Republican lead) among each group's voters, and among its non-voters, as they told the Cooperative Election Study. Turning out more of a group helps its non-voters' side, not its voters'.

White VotersR+15.3 Non-votersR+20.7

Black VotersD+67.0 Non-votersD+42.9

Hispanic VotersD+15.5 Non-votersD+1.5

Asian American VotersD+34.1 Non-votersD+15.8

Men VotersR+10.5 Non-votersR+9.5

Women VotersD+4.7 Non-votersD+0.7

How the replays work

Turnout. The Census Bureau's November survey gives citizens of voting age and voters by state and race. People over-report voting, so in each state we scale every group's figure by the same amount until the voters add up to the votes actually counted.
How each group voted. From 2008, the Cooperative Election Study, about 60,000 people a year whose voting is checked against the voter files: each state's figure for a group, pulled toward the group's national figure where the state sample is small. For 2000 and 2004, the national exit polls.
Matching the real result. In each state, one shift applied to every group makes the starting point reproduce the actual result exactly. A replay then changes only who votes.
The extra voters. By default they vote the way the group's non-voters said they would before the election; the switch makes them vote like the group's voters. Voters removed from a replay take their group's split.
Men and women. The same method with the Census figures by sex. The gap between men's and women's votes is much the same from state to state, so each state starts from the national gap and is then shifted to its own result. "Everyone votes" is shown by race and ethnicity only: the two breakdowns use different surveys of non-voters, so they give slightly different answers.
Electoral votes. Each year's apportionment. Maine and Nebraska's split district votes stay split while the state's winner is unchanged. Faithless electors are left out, so 2000 shows Gore with 267 and 2016 Clinton with 232.
Limits. A replay holds everything else fixed: candidates, campaigns and how everyone else voted. Small groups in small states rest more on the national figure for the group. Read the replays as the size of what turnout can do, not as what would have happened.

Sources: Census Bureau, Current Population Survey November supplements 2000-2024; Cooperative Election Study cumulative file (Kuriwaki, Harvard Dataverse, CC0); national exit polls 2000 and 2004; certified state results.