Data On Associated Press Tally Of "New" Votes
The Nashua Advocate
By ADVOCATE STAFF
In an article slated for release at 12:01 A.M. on Saturday, December 18th, the Associated Press will report that, with 74% of votes recounted from the November 2nd, 2004 election, 949 new net votes have been uncovered.
While it is impossible to calculate what percentage of these new votes were uncovered via manual recounts, and what percentage from machine recounts -- the AP claims, in general terms, that the new votes were engendered by "hanging chads that came loose when punch-card ballots were handled again or rerun through tallying machines" -- if these 949 votes came primarily from the 3% hand-recounts, a statewide manual recount would be expected to accumulate approximately 42,320 new net votes. While this total could be much lower (if machine-recounts, not hand-recounts, also loosened many chads) or much higher (if ballot-prepping and precinct-targeting make the 3%-of-ballots recounts conducted thus far unrepresentative of the state as a whole) in the judgment of The Advocate this estimate seems to be a well-grounded, middle-of-the-road figure.
Of course, this figure refers only to new net votes; The Advocate has already posited, in this space, that new gross votes would, of necessity, be much greater in number, perhaps -- and this estimate still seems feasible, albeit slightly more attentuated with current data -- as high as 100,000 new ballots counted.
Those wondering how two major-party candidates could "lose" votes in a tally, thereby allowing "gross" votes added overall to be much higher than "net" votes reported need simply consider that A) manual recounts can show that a vote presumed to be for a candidate was in fact an overvote (e.g., through double-punching, write-in votes, or other markings on the ballot; B) manual recounts can show that a vote presumed to be for a major-party candidate was in fact for a minor-party candidate; C) manual recounts can show that a vote tallied for one candidate was in fact an undervote, either because of machine error designating the ballot incorrectly, or because a visual inspection of the ballot suggests that the ballot does not meet Ohio's "two-corner" rule for hanging chads, or because a manual recount uncovers disparities in poll-books or polling-guides which invalidate altogether individual votes for one candidate or another; D) manual recounts can show that a ballot was illegally filed as either a provisional or absentee ballot, thus striking altogether a given vote for a candidate; and E) manual recounts can find new undervotes or count as valid existing undervotes.
In short, it cannot be presumed that every vote gained for Kerry was lost to Bush, or vice versa, and that therefore "net votes gained for each candidate equals gross votes gained for the two candidates together." Those with much greater faith in ballot-counting than our Staff may wish to reduce The Advocate's estimated statistical disparity between net and gross votes and allow, therefore, that a statewide manual recount would "only" result in 60,000 or so new gross votes -- not enough to swing the election in itself, but, when combined with other coting irregularities and an existing crop of 92,000 undervotes, perhaps still enough to do so.