UC Merced CAPE Election Lab · Weather & turnout← Back to the big screen

Is the weather keeping people home?

Live National Weather Service alerts laid over the counties in the 53 House races the lab is tracking, with each county's turnout against what was expected, and the state rules that decide whether bad weather on Election Day matters at all.

Jump to the hypothesis tests ↓Live radar map

Live radar over the district map

Loading the map…
How this works
tracked race (bold line, dark label; click for its page)Democrat aheadRepublican ahead (every district, once it has votes) radar: rain/snow intensityLabels appear when a district is big enough on screen; zoom in for the rest.

Weather right now, where our races are

Loading alerts…
How this works

Turnout vs. expectations: alert counties against the rest

Decision Desk HQ projects how many votes each county will cast. Before polls close that projection is the pre-election estimate; once votes are counted it moves with the count. The gap between the two is the turnout surprise. Weather would show up as a bigger shortfall in alert counties than elsewhere.

How this works

Who loses more?

The same shortfall split by how the county voted in 2024, each compared with no-alert counties of the same lean. If Trump counties under alerts fall further below their expected vote than no-alert Trump counties do, and Harris counties don't, the missing votes are in Republican-leaning places; that is where, not whose.

Same weather, different rules

Where a storm sits across a state line, the two sides are a natural experiment: same rain, different rules on mail ballots and early voting. States where most votes were already in before Tuesday should barely notice; states that vote mostly on Election Day should feel it.

Hypothesis testing in real time

Five claims from the turnout literature, re-estimated every five minutes as the count comes in: the counties in all 435 House races that are at least half counted, today's precipitation at each county's center (coverage shown below), and each state's mail and early-voting rules. The second test is a border-pair comparison: counties on either side of a state line where the rules change. Verdicts are deliberately soft; this is a live, self-selected sample re-tested all night, not a paper.

Loading the latest run…
How this works