Telling Lies In America—The Miami Herald Overvote Analysis
Part One--Herald Cooks The Numbers
Paul Lukasiak firstname.lastname@example.org
Note: this is a draft - more details to follow!
The Miami Herald did far more than simply 'spin' its results from its examination of the Florida overvote---it completely lied about the numbers, double counting some votes, and completely ignoring hundreds of others.
Despite its claim that this was a review of all the overvotes, half the overvoted ballots were never examined for indications of voter intent.
- In Broward County, The Herald ignored both the results of the certified manual recount and its own undervote review conducted by BDO Seidman under fairly rigorous conditions, substituting instead the judgement of a single reporter examining less than one tenth of the ballots reported as undervoted in the machine count.
- The Herald ignored the certified recount from Volusia County, yet examined less than 30% of the ballots reported as overvoted when counted by machine.
- The Herald included numbers for a new 'count' of undervoted ballots in Orange County without ever examining those ballots, and the number of 'new' votes found is completely inconsistent with the number of undervotes reported by Orange County.
- The Herald treated over 10,000 undervoted and overvoted absentee ballots from optically scanned counties as if they were machine countable votes.
Based on these completely erroneous numbers, the Herald reached conclusions about who received the most "legal" votes in Florida without regard to Florida law, Florida court precedent, or the decision regarding equal protection in vote counting by the United States Supreme Court.
And in order to hide the truth, the Herald has removed from its web site the data from its undervote count, on which its overvote review conclusions are based.
METHODOLOGY--THE INCOMPLETE 'COMPLETE' REVIEW
Despite the impression given by Martin Merzer in his Herald article on the Florida overvote review that the Herald's conclusions are based on looking at each and every ballot cast that was not counted as a vote by a counting machine, nothing could be further from the truth. In fact, according to the Herald itself, over half the overvoted ballots were never looked at by a human being, instead relying upon log tapes provided by the counties that were made (supposedly) during the machine recount of the statewide vote. In the same article, the Herald further claims that , "reporters physically examined a small number of actual ballots that had markings that could not be read by counting machines." Clearly, since these ballots were never reviewed by hand, there is no way that the Herald could have accomplished what it claims to have done—review ballots with markings that were not read by the machines themselves.
Nevertheless, The Miami Herald (a Knight-Ridder newspaper), in its on-line edition, has billed its overvote analysis as the "first statewide review of overvotes" . Given the nature of the 'review' itself, that statement is highly deceptive. However, and more to the point, is that the same writer who wrote those words for the Miami Herald's on-line edition flat out lied about the methodology used in the story provided to other members of the Knight-Ridder syndicate, and it this version of the story which was apparently the basis for the way the Herald's overvote analysis by the rest of the national media.
In the version of the story provided to Knight Ridder, Merzer wrote "[a]nd now, after an exhaustive hand recount of every Florida ballot…" Clearly, this statement is absolutely false, even when allowing for the fact that 97% of the ballots in Florida (those that were counted by machine as valid votes) were never examined by hand by the Herald.
And although the Herald's undervote analysis was conducted by an independent accounting firm, and two people examined each ballot, the Herald's overvote analysis is based on the opinion of a single reporter in each county. No effort was made to confirm that the judgement of any of the various reporters that looked at these ballots was impartial. Given the obvious pro-Bush spin that Merzer himself has consistently demonstrated in his articles on the recount, no assumptions should be made about the reliability and impartiality of the decisions made by these reporters.
(The Herald's methodology for the overvote review stands in stark contrast to both the forthcoming NORC (National Opinion Research Center) Consortium review of all of Florida's undervoted and overvoted ballots, and that of BDO Seidman, which conducted the review of the ballots for the Herald's undervote analysis. Although the Herald's analysis of the BDO Seidman undervote data was seriously flaws and biased, there is at this point no reason to question the accuracy of the data provided by BDO Seidman.)
COMPARING ORANGE, AND ROTTEN ORANGES—ALTERING THE BASELINE CRITERIA
If you add the totals from the Herald's own analysis of the undervote to the totals from the Herald's overvote analysis, only one conclusion can be drawn—regardless of the criteria used, Al Gore received far more votes than George W. Bush in Florida in the 2000 presidential election. Nevertheless, the Herald is claiming that using two criteria, Bush would have won in Florida. In order to accomplish this rather extraordinary conclusion, it was necessary for the Herald to establish a new, and completely erroneous baseline criteria.
The Herald used as its baseline in its undervote analysis the 537 vote lead certified by the state of Florida for George W. Bush. It then added in the votes that the Florida Supreme Court had ordered included in the final tallies from the Palm Beach County and (incomplete) Dade County manual recount, examined the ballots that remained undervoted after certification and the inclusion of the Palm Beach and Dade votes, and added the results to that 537 vote lead.
For its overvote analysis, the Herald created an entirely different baseline. The Herald claims that this baseline represents the results of the statewide machine recount with no hand recounted ballots. This criteria completely excluded the results of the certified Volusia County full manual recount, but included hand-counted ballots from at least seven counties which engaged in selective hand counting of ballots and included their results in the totals submitted with the "machine recount" numbers. It also included the votes from over 10,000 ballots that were not machine readable, but for which duplicate ballots were created using a 'voter intent' standard that were then counted by machine. This selective inclusion and exclusion of certain ballots and votes in the "baseline" was instrumental in the creation of the analyses which resulted in ostensible Bush 'victories' in Florida.
FUZZY MATH---THE PRE-OVERVOTE CALCULATIONS
Having decided to completely change its baseline criteria, the Herald then had to account for votes that were not part of that criteria. First the Herald added in the certified "late absentee" ballots, raising the Bush lead to over 1000 votes. Many of the absentee ballots were in fact not readable by machine, and were counted using a "voter intent" standard. It is not the least bit unlikely that overvoted ballots from this pool of votes were included in the Bush overvote totals as well. Certainly, the Herald makes no mention of excluding such ballots from its examination of the overvotes it did examine.
The Herald then adds in another 508 votes merely based on the fact that they wound up as part of the certified totals, from which Bush gains another net advantage of 130 votes. Included in these votes are absentee ballots that were excluded during the initial canvass of returns because they did not meet the statutorily established criteria for absentee votes, including at least one vote for Bush that was received via fax machine, and other votes that were illegally delivered, were undated, unpostmarked, or had non-matching signatures.
Also included in these votes are the 197 votes that Nassau County claimed were mistakenly excluded from the machine count. Nassau County decided to use its original machine count numbers without ever making an effort to figure out what the correct total was. When the Nassau ballots were subjected to media review, Nassau County officials could not account for all of those 197 ballots. Nevertheless, and despite the questionable nature of these votes, the Herald included them solely because they were certified.
The Herald also included a small number of votes found in other counties after the initial certification process.
It is essential to note at this point that all of these ballots were included in the Herald's totals SOLELY because they were part of the certified totals, but that the Herald decided to exclude votes that were certified from Volusia and Broward Counties. Whether the votes that were included were machine readable, whether they met the legal standard of a vote in the state of Florida, indeed whether they actually existed or not, was not considered relevant.
Then, the Herald does something that works to Gore's advantage, but demostrates once again how little regard the Herald had for coming up with vote totals that bore some relation to the actual votes cast. Orange County reported that it had a total of 966 undervoted ballots after the machine recount. But when the Herald initially examined the undervoted ballots, Orange County election officials could only account for 691 undervoted ballots, a difference of 275 ballots. Sometime after the Herald reviewed the undervotes, Orange County told the Herald that it had found that 433 additional undervotes had apparently been incorrectly identified as counted by machine during the separation process. That means that Orange County is claiming that there were 158 MORE undervotes than it originally reported. The Herald never bothered to inspect these ballots, and appears entirely unconcerned with the discrepancy.
The net result of all of this 'fuzzy math' was the pre-ballot-review baseline, which gave George W. Bush a net advantage of 1133 votes, supposedly excluding undervoted and overvoted ballots. In theory, this number represents all of the ballots that were not subject to the Herald undervote and overvote review, and the Herald examined all of the remaining under- and overvoted ballots. But the Herald's own numbers tell a completely different story.
Confused? You should be, and the confusion seems to be deliberate on the Herald's part.
DISSAPEARING VOTES IN VOLUTIA COUNTY
Prior to doing its full manual recount, Volusia County recorded 488 overvotes and 155 undervotes. As a result of the recount, Al Gore received an additional 241 votes, and Bush got an additional 143 votes, for a net gain of 98 votes for Gore. However, not all of those votes were from the pool of undervoted and overvoted ballots. 320 ballots were completely uncounted during the machine count of Volusia County that the Herald used as its baseline number. Those 320 ballots are completely unaccounted for in the Miami Herald overvote analysis because they were not part of the machine count totals.
The Herald review of the undervoted ballots, which probably did not (but may have) include undervoted ballots recovered during the recount, gave Gore a net gain of 33 votes. The net gain of votes for gore from the overvote count was 12, for a total net gain for Gore of 45 votes. By excluding those 320 votes, the Herald underreported Gore's net gain by at least 53 votes, and possibly by as many as 86 votes.
But, as noted above, 488 overvotes were originally reported by Volusia County, yet the Herald based its overvote totals on an examination of only 154 ballots. The Herald makes no attempt to explain this discrepancy.
DISSAPEARING UNDERVOTES IN BROWARD COUNTY
The Herald chose to ignore the final certified totals from Broward County, and instead substitute the original machine count totals, combined with the results of its analysis of the overvoted and undervoted ballots from Broward. Well, almost.
It seems that, according to the Herald, the totals from their undervote count "were based on ballots reviewed by the Broward canvassing board and included more than 1,000 ballots that contained clean punches. Broward officials were uncertain that those ballots hadn't also been included in the machine counts." The Herald, in reaching its conclusions, changed this "uncertainly" into a definite fact, excluding all the cleanly punched ballots it counted as part of its undervote study, and decreasing Gore's net votes by 531. That 531 is 120 votes higher than the lead given to Bush under which the Herald found that George W. Bush did best (the 'two corner chad' crioteria), and is more than triple Bush's projected victory using the "clean punch' standard. In other words, the Herald presented as "certain" Bush victories that were based solely on the "uncertain" report of Broward County officials that these votes may have been included in the machine count.
The data suggests that, in fact, most of these votes were included in the machine recount, and were not part of the undervote. The original undervote count by the Herald found 779 more undervotes than were reported as a result of the machine recount. But the total number of cleanly punched chads assigned to candidates found in that review was 1052, leaving 273 ballots completely unaccounted for in Broward County. Gore's 3 to 1 edge in Broward County would mean that he would have received an additional net gain of around 140 votes from these ballots. (the number 273 could well be much higher, inasmuch as 434 Broward ballots that were reviewed by the Herald in its undervote count were evaluated as "overvotes" by the Herald.)
It is essential to keep in mind that machines identified numerous cleanly punched ballots as undervotes in virtually every Florida punch card county, and that in addition to being counted by machine twice, and counted manually, these ballots were once again separated by machine, so it is highly likely that these 273 ballots did, in fact, represent legal votes that had been counted in the machine recount as undervotes. Nevertheless the Herald decided to exclude ALL the cleanly punched ballots it found in its undervote totals from its analyses of the overvoted and undervoted ballots.
THE DISAPPEARING BLACK VOTE—GADSEN COUNTY
Perhaps the single most bizarre bit of data concerns the only majority black county in Florida, Gadsen County. According to the totals used by the Herald, Gadsen County's 1844 optically scanned overvotes did not yield a single vote for either Bush or Gore. Zero. Zip. Nada.
What is perhaps even more amazing is that the Orlando Sentinel found 142 ballots that would have fallen into the "corrected" ("Bush or Gore oval marked and ovals for other candidates marked and then crossed out") category of valid overvotes used by the Herald in its overvote analysis. These were votes the Sentinel defined as ballots showing "multiple errors on rejected ballots with 'Bush' or 'Gore' written in", with the example provided by the Sentinel being fully consistent with the Herald's definition.
In sum, the reporter sent to Gadsen County to check their overvote was unable to find a single vote in a centrally scanned county with the highest percentage of overvoted ballots in the state regardless of voting method, even though other reporters found such votes. Despite the remarkable nature of this complete lack of potentially recoverable votes, the Herald does not find it worthy of any notice whatsoever.
LEE COUNTY—IGNORING EVIDENCE OF LEGALLY PRESCRIBED VOTES
The 'recoverable' votes found by the Herald fell into one of three categories. The first was "Write-in / Name = Bush or Gore oval marked and Write-in oval marked with same candidate written in." The second was "Write-in / Blank = Bush or Gore oval marked and Write-in oval marked with no candidate written in." The third was the "corrected" category noted above. It should be noted that the votes falling into the first two categories are specifically prescribed as legal votes under the Statutes and Codes of the State of Florida, and are legally NOT overvotes. The fact that they were counted as "overvotes" is a direct result of the nonfeasance, misfeasance, or malfeasance of Florida state and county election officials.
I spoke with the Supervisor of Elections of Lee County, Philinda Young, in conjuction with an article I was writing about the fact that hundreds of these fully legal votes were never included in the certified totals, and that had they been included, Al Gore would have won the Florida election. After initially denying that it was even possible that such legal votes had been excluded, I was able to convince Ms Young that it was indeed possible. Ms Young was unaware that, for the first time ever, Lee County had permitted voters to create a hole in a chad next to the write-in instructions, and then the tabulating machine had been set to count that as a vote, and render as overvoted any ballot with holes both next to a candidate and next to the write-in instructions. After checking with her assistant who had designed the ballot and programmed the tabulating equipment, Ms Young immediately inspected each of the ballots that had been separated as overvotes for examination by the media. She found seven such ballots with legal votes.
Nevertheless, the Herald claims that there were no ballots in either the "write-in/name" or "write in/blank" categories.
Seven votes are not that many. However, the Herald also failed to find a single such vote in other punch card counties that made it possible to punch a hole adjacent to the write in instruction (not all Florida counties did so, because the hole serves absolutely no useful purpose when it comes to write in votes.) Included among those counties in which such miscounted votes are possible is Duval County. According to the Herald, 84% of the nearly 22,000 overvoted ballots in Duval were overvoted on because of votes "cast" on both the first and second page of Duval's two page presidential ballot. The names of Al Gore and George W. Bush appeared on the first page of that ballot. The "write-in" hole appeared on the second page. There is no way to know how many such ballots exist in Duval County, but it simply defies belief that not one single vote met that criteria.
DOUBLE COUNTING AND MISLABELLING THE BUSH OVERVOTES
As noted above, the Herald claims that its baseline numbers do not include any hand recounted ballots. This statement is demonstrably false, as well as wholly deceptive.
Based on reporting by the Orlando Sentinel, the numbers used in the Herald's baseline as well as in the votes that were added from late absentees, include votes on well over 10,000 ballots that were rejected by tabulating machines, evaluated by county canvassing boards, and included in the totals.. Some of them were votes that were 'eaten" by the voting machines, and others were unreadable because of the way that the ballot had been folded.
Some of these ballots, however, were unreadable because of voter error. Sometimes the voter used the wrong pencil, and had their votes initially counted as undervoted as a result. Other ballots were from voters who did not follow instructions, and filled in the ballot wrong, resulting in either overvotes or undervotes.
Most of these ballots were added in Republican dominated counties. Most of them were from absentee ballots where George W. Bush had a statewide 2 to 1 advantage. In some counties, only votes that met the legal definition of a vote outside the "voter intent" standard were counted. In others, county officials used voter intent as its criteria for deciding whether or not to count a vote.
After having determined what these rejected ballots contained, the officials in these counties then created duplicates of these ballots that were counted by machine. Unfortunately, the Sentinel article does not give a clear indication of what the original 10.000 plus ballots showed in the Presidential race. Some of these ballots were doubtless duplicates of ballots with votes for president that would have been read by the machine had it not mangled the ballot. Others may have been duplicated because they were unreadable in races other than for President.
What is clear, however, is that the Herald completely misrepresents its data when it states that its baseline numbers do not include any ballots that were hand recounted. And this is especially significant, because one of the two criteria used by the Herald that gave Bush a victory was the "strict standard" criteria. These votes were included in those totals as if they met that criteria, and it is abundantly clear that many, if not most, of them did not. It is also important because it is highly likely that thanks to the hand recounting and duplication of these otherwise unreadable ballots, George W. Bush led the vote totals after the first machine recount and was declared the winner by 537 votes.
Although the Sentinel does not provide a full breakdown of these ballots, it does include intriguing hints as to how election officials in Republican dominated counties manipulated the vote to maximize the Bush advantage. Escambia County, which went for Bush by a margin of 64 to 35% (and the absentee margin was even greater, 72 to 27%) is a precinct scanned county. Nevertheless, Escambia County officials chose to not take advantage of the protection from overvoting that precinct based scanning can provide. As a result, Escambia's overvote rate was 10 times that of counties that consistently provided its voters with overvote protection.
And although Escambia County did not provide any protection from overvoting to its citizens at the polls, Escambia County went to great lengths to make sure that the voter intent of every person who voted absentee was reflected in the County totals. Over 2400 absentee ballots were duplicated in Escambia County because their voters were not machine readable. Given the almost 3 to 1 margin for Bush in Escambia's absentee ballots, it is easy to see how the decision to protect absentee voters in Escambia County resulted in an advantage for Bush of hundreds of votes that would not have been realized had they been treated the same way as votes cast as the polls.
Conversely, had voters at the polls been given the same treated as absentee voters, the Herald recount totals indicate that Al Gore would have received a net gain of 157 votes, a two to one margin in overvotes for Al Gore. And, inasmuch as this loss of votes for gore reflects less than 12% of the overvoted ballots in Escambia County, one can easily see how the decision to not provide the citizens who voted at the polls with the overvote protection that the machines were designed to provide could, in and of itself, have cost Al Gore the election.
None of this, of course, is mentioned in the Herald analysis of the overvote, which concentrates solely on voter error as the cause of Gore's defeat, and completely ignores the decisions made by county and state officials.
The Herald's review of the Florida overvote is so egregiously flawed that it is completely unreliable. More importantly, however, is that there is a marked bias in that unreliability that provides a significant perceived advantage to George W. Bush. The Herald consistently misrepresented its own data, and just as consistently skewed the numbers to make it look as if Bush got more votes than Gore in Florida.
Part II will examine the Herald's narrative, and the way in which its presentation was biased toward creating a Bush victory when, in fact, all the evidence points to Gore getting the most legal votes in Florida.