If a pollster calls someone who says they voted already, they are automatically passed through the likely-voter screen since they have, after all, voted. The problem with this can be best summed up by Gregory House: “Everybody lies.”
Pollsters can only work with what their respondents tell them, and this is the reason that likely-voter screens can be so tricky, though important, in polling. The preferable response is that you are going to vote or, in the case of Ohio, that you’ve already voted. Many respondents will say they are going to vote (or have voted) when in fact they may not end up doing it (this effect is known as social-desirability bias). For this reason, some likely-voter screens ask about previous elections and general political enthusiasm to gauge the actual likelihood that a voter will end up in the booth on Election Day. But that is where early voting throws the screen out the window — if a voter says they voted, there is nothing a pollster can do to but assume that it’s true.
Enter Ohio, where the current estimates from compiling early in-person and absentee voting shows early turnout to be about 15 percent of voters. But responses in the current polls claim that 23 percent of registered voters have already voted. That means that polls are overstating early voting by eight percentage points on average. This could be in part because some voters have requested an absentee ballot and report that as voting, some have mailed in ballots that haven’t been counted as received yet, but some voters are also just flat out saying they voted when they haven’t. It’s impossible to know the exact reason, but it’s clear that more are claiming to vote than really have.