Exit polls are surveys conducted with voters as they leave the polling place, asking how they voted along with basic demographic information. Unlike pre-election polls, which try to predict what will happen, exit polls attempt to measure what already did happen — making them, in principle, a useful independent check on officially reported results.
In practice, exit polling is a difficult science. Response rates vary by demographic group, precinct selection has to be carefully designed to represent the broader electorate, and the growing share of early and mail voting means a meaningful portion of the electorate is never present at a polling place to be surveyed at all. These factors mean exit polls will rarely match a final vote count exactly, and a modest gap between the two is expected rather than alarming.
Where exit polls remain genuinely useful is as a rough, real-time sanity check and as a research tool for understanding voter behavior — who voted for whom, and why, broken down by issue, demographic, and region. Treated with appropriate statistical caution, they add a valuable independent layer to how elections are understood and analyzed after the fact.