Mac Booker's videocam moves in for what is arguably the world's least-dramatic closeup in a Youtube video: (www.youtube.com/watch?v=mRtsKF4Ox2c&feature=channel_page). It shows the top corner of a General Return Sheet, which Allegheny County poll workers must fill out each Election Day after gathering the vote tally from the iVotronic touch-screen voting machines. This particular sheet, from a Wilkinsburg polling place during November's election, records that the "Total Number Of Voters Admitted To Machine" -- the number of voters whose names poll workers wrote down -- was 591, while the "Total iVotronic Voters" -- the number of voters tallied by the machine -- was 601.
The difference in the two counts (so-called "phantom ballots") is a problem, says Booker, field director for the state League of Young Voters group -- especially with the League's get-out-the-vote mission. "If there's no confidence in the actual voting system," he says, then explaining this discrepancy "is something that needs to be done to make everything else we do meaningful."
The League, VoteAllegheny and other local groups say they will join to seek a report from the county about the total number of phantom ballots and any other voting machine anomalies from the most recent election.
Local vote-watcher Richard King, of Squirrel Hill, says he and Booker have made a preliminary examination of a dozen polling places that reported voter or machine problems on Election Day and found one, two or three such phantom ballots noted by poll workers in a third of the places, alongside the 10 at a single Wilkinsburg polling site. On a second check, King found eight more sites with at least one phantom ballot.
If those extra ballots are blank -- that is, if they list the candidates and referenda but none has been checked off as a vote -- there is probably not an issue, King allows. But if the ballots contain legitimate votes that weren't counted, that's a problem, he says. And if they contain votes that were counted but didn't originate with a voter who signed in that day, that's also a problem, he adds.
Booker checked his own East Liberty precinct, just to look at another random site, and found a single phantom ballot there as well.
"That very likely indicates that this is a systemic problem," he says. "But until we do [an] audit, there's no way to know."
King and Booker discount the possibility of experienced poll workers causing the machines to record extra ballots -- forgetting to sign in, and assigning a number to, a voter or three, for instance. They say they simply cannot believe so many workers would make such mistakes. But King says he has seen various explanations written on General Return Sheets by poll workers -- explanations that don't really explain mismatched voter/ballot tallies: "I've seen 'machine malfunction,'" he says. "I've seen 'I don't know.'"
County Elections Director Mark Wolosik discounts the very notion of phantom ballots. He says human error by poll workers is likely to be the cause of the discrepancy between voters signed in and ballots on the machine. He says poll workers are either forgetting to add themselves to the polling place voter list when they vote or allowing voters to mistakenly push the iVotronic button that activates the audio ballot for the hearing impaired. Without earphones attached to the machine, he says, the voter at that point would hear nothing, and the screen would still be blank. A poll worker might then activate the usual, visual ballot, causing the extra ballot to be created.
"If the local board does not put an explanation in their Return Sheet, we call them," Wolosik reports. "If it seems like a satisfactory explanation ... we accept it as true and correct. They take an oath on Election Day ..."
Similar instances would sometimes happen in the days of mechanical, lever-based voting machines, he adds.
"The county has said that ... essentially, they didn't see this as a crisis. I disagree," says VoteAllegheny President Collin Lynch. "It would be just as bad if we were using paper ballots and we opened up a ballot box and there were extras. It's necessary to sit down and figure out how often this occurs. Honestly, no one seems to have conducted a check. It's basic election auditing -- we should have done that with lever machines and we should be doing it with these machines."
Wolosik says the county has no plans to issue a report of voting machine performance, but that the public is welcome to examine individual polling place voting records themselves.
The voting-watch groups are seeking printouts of all ballots in the suspect machines -- ballots the machines keep in randomized order -- or at least an electronic copy, which the machines can also produce. At a Harrisburg meeting Dec. 17 between these groups and Pennsylvania State Department officials, King says the department agreed to come up with a set of election evaluation questions for all counties to answer, including an examination of the phantom ballot count. State Department spokesperson Leslie Amoros confirmed the meeting but could not say whether the department was proceeding with this request.
"It's been up to the citizens ... to figure out how smoothly the election went," King asserts, since the county and state routinely report an essentially flawless process. "I think we're a pain in the butt" to election officials, King adds, "but I think we can help them get better."