Florida 2000 Presidential Vote

L. David Roper (roperld@vt.edu)

(Back to Politics)
If George W. Bush wins the 2000 election for President, Florida will get its just rewards for making it happen: Bush will aid global warming and Florida will be the only state to become totally underwater because of global warming.

Famous Butterfly Ballot:

This is the famous butterfly ballot:

The point of confusion is that Al Gore is the second name on the ballot but to vote for him one had to punch the third punch hole on the ballot. If one just punched the second hole, thinking that it was for Al Gore, he/she would have voted instead for Pat Buchanan. Someone with poor eyesight or who was in a hurry to vote could have made that mistake. Apparently many thousands, perhaps over 3,000, of the 450,000+ votes in Palm Beach county did make that mistake.

This shows the Buchanan vote for Florida counties that reported an overvote, sorted by Buchanan vote. Note that Palm Beach county has almost three times the vote of the next highest county, Pinellas.
This shows the Buchanan vote for Florida counties that reported an overvote, sorted by % Gore had of the vote. The last eight are the counties that voted a majority for Gore. Note that Palm Beach county has five times the vote of the next highest Democratic county, Broward.
This shows the % Buchanan vote for Florida counties that reported an overvote, sorted by % Gore had of the vote. The last eight are the counties that voted a majority for Gore. Note that Palm Beach county has more than twice the vote of the next highest Democratic county, Flagler.

It appears that over 19,000 Palm Beach voters saw that they had made that mistake and then punched the correct third hole for Al Gore, also. These ballots were rejected by the vote counting machine as double-voted ballots.

A way to determine the likelihood of such a mistake is to examine all of the double-voted ballots in Palm Beach county to see if they were predominantly votes for Gore and Buchanan, instead of, say, for Gore and Bush or any other two-candidate combination. I have not been able to find such data on the Internet.

Another indication of the likelihood of such confusion in Palm Beach would be to compare the fraction of double-voted ballots in all counties of Florida to see if the Palm Beach fraction is much larger than all other counties, remembering that Palm Beach county is the only county that used this butterfly ballot. I have not been able to find such data on the Internet.

Some Republicans stated that the same butterfly ballot was used in Chicago, but the Chicago ballot they displayed in the background, such that it could not be examined, on television was quite different in two ways:

This display and mention of the Chicago ballot was very dishonest.

Some Republicans said that the voters had a sample ballot that they could have studied to eliminate the confusion before voting. But the sample ballot was:

Note that it does not show the single column of holes in the middle that the real ballot had. Also, it is not immediately obvious that the two sides of the ballot are the same race; in the usual butterfly ballot the two sides are separate races and each side has its own column of holes to punch. It is not obvious from this sample ballot that it does not have two columns of holes to punch.

Punch Ballots and Optical Ballots:

There are two types of ballots in the state of Florida. About twenty of the sixty-seven counties use antiquated punch-card ballots. An example is given above for Palm Beach county. The rest of the counties use optical-scan ballots, which are more accurately read. The following graph shows the undervotes for the fifty-four counties which reported undervotes, sorted by undervote and separated according to whether they used punch cards or optical-scan ballots:

Note that the punch-ballot counties have an average undervote of (1.52±0.54)%, whereas the optical-ballot counties have an average undervote of (0.62±0.77)%. Thus some citizens of the antiquated punch-ballot counties are being deprived of their vote being counted compared to optical-ballot counties. (The1.52±0.54 indicates that roughly 2/3 of the values lie within the limits of 1-0.54=0.46 and 1+0.54=1.54.)

Thus, the citizens of counties with punch-card ballots have been denied "equal protection", which is forbidden in Amendment XIV, Section I of the United States Constitution.

This analysis indicates that all counties with undervotes greater than 1% should have counted the undervote ballots by hand. Those twenty-two counties are:

Counties Over 1% Undervote
County % undervote Ballot Type Undervote Bush Vote Gore Vote County
Punch: Pinellas 1.04% 0 4,226 184,825 200,630 Pinellas
Lee 1.07% 0 2,017 106,141 73,560 Lee
Sarasota 1.10% 0 1,809 83,100 72,853 Sarasota
Osceola 1.12% 0 642 26,212 28,181 Osceola
Baker 1.13% 0 94 5,610 2,392 Baker
Broward 1.13% 0 6,686 177,902 387,703 Broward
Pasco 1.21% 0 1,776 68,582 69,564 Pasco
Hardee 1.28% 0 85 3,765 2,339 Hardee
Highlands 1.35% 0 489 20,206 14,167 Highlands
Hillsborough 1.50% 0 5,531 180,760 169,557 Hillsborough
Miami-Dade 1.64% 0 10,750 289,533 328,808 Miami-Dade
Duval 1.70% 0 4,967 152,098 107,864 Duval
Indian River 2.05% 0 1,058 28,635 19,768 Indian River
Collier 2.18% 0 2,082 60,450 29,921 Collier
Palm Beach 2.29% 0 10,582 152,951 269,732 Palm Beach
Marion 2.30% 0 2,445 55,141 44,665 Marion
Sumter 2.58% 0 593 12,127 9,637 Sumter
Optical: Taylor 1.11% 1 82 4,056 2,649 Taylor
Liberty 1.12% 1 29 1,317 1,017 Liberty
Franklin 1.38% 1 70 2,454 2,046 Franklin
Calhoun 1.48% 1 78 2,873 2,155 Calhoun
Columbia 3.21% 1 617 10,964 7,047 Columbia
Washington 3.49% 1 292 4,994 2,798 Washington
22 counties Votes: 57,000 1,634,696 1,849,053
%: 1.64% 43.9% 53.1%

Note that the vote for these twenty-two counties is overwhelmingly for Gore. Thus a hand count of the undervotes for these counties is likely to swing the Florida vote to Gore. If only half of these undervotes are actually votes and they have the same percentage for Gore as the counted votes did, then Gore would gain 1,754 votes. If the undervote for the punch ballots is reduced to the average for the optical-scan ballots (0.62%), then the Gore gain would be 1,329 votes. Thus, Judge Saul was incorrect in stating that there is no creditable evidence that counting the undervotes would change the result of the election.

An execellent article in The Washington Post for 9 December 2000 argued very convincingly that the undervote in the Republican counties was most likely Democratic because:

Counties requested by Gore for undervote to be hand counted:
County % undervote Ballot Type Undervote Bush Vote Gore Vote % Gore
Broward 1.13% 0 6,686 177,902 387,703 68.5%
Miami-Dade 1.64% 0 10,750 289,533 328,808 53.2%
Palm Beach 2.29% 0 10,582 152,951 269,732 63.8%
Votes: 28,018 620,386 986,243
%: 1.74% 38.6% 61.4%
% of vote: 46.7%
% of undervote: 50.8%

Note that these three counties have a considerably larger (0.1% more) undervote than the twenty-two over-1%-undervote counties. Although they represent 46.7% of the vote in the twenty-two over-1%-undervote counties, they represent 50.8% of the undervote in the twenty-two over-1%-undervote counties. Thus hand counting only these three counties would well represent the over-1%-undervote counties. Palm Beach county, being one of the largest vote and undervote counties, definitely should have its undervotes counted by hand. A similar analysis for Miami-Dade county yields smaller numbers because of the smaller Gore percentage for that county.

If only half of the undervotes for these three counties are actually votes and they have the same percentage for Gore as the counted votes did, then Gore would gain 3,043 votes. Palm Beach county alone would yield a 1,462 vote gain for Gore. In order for Gore to gain 600 votes over Bush in Palm Beach county the percent of undervotes to be actual votes needs to be 20.5%.

In the following Florida map the counties are shown with punch ballots (blue) and with optical-scan ballots (yellow) that had undervotes greater than 1%:

There appears to be some problem with the optical-scan readers in the six counties in the panhandle of Florida or the panhandle population was more ambiguous about for whom to vote than was the rest of the state. It would be interesting to compare these counties exit polls to the rest of the county.


When the antiquated punch-card ballot is punched by the voter, it apparently often occurs that the punching mechanism does not completely dislodge the small piece of paper called the "chad". (Anyone who has ever used a paper punch extensively has observed this phenomenon.) The optical readers that read the punched ballots to tally the votes apparently will not let enough light through the hole to record a vote if the chad is still attached. The card readers reject such cards without counting them; thus such votes have not been "counted" by the card-reading machines. For an accurate tally of the votes the ballots with partially dislodged chads must be scanned by human eye, i.e. by a "hand" count, because the reader will not read such ballots. Such counting is not a "recount" since the machines rejected those cards without actually counting them.

Hand Counts of Punch Ballots:

Much of the controversy in the Florida hand counts has been over the criteria used in determining whether a ballot was intended to be punched at a particular location where a chad remains attached. Some of the criteria discussed were:

It appears to me that the first three "hanging chads" criteria are sufficient to indicate that a location was punched. If a dimple is square in the middle of the location specified for a voting area on the ballot and no other dimple occurs in the opponents' voting areas on the ballot, it is a strong argument that the voter intended to vote for that candidate rather than some other candidate. It is unlikely that such a dimple could have been accidentally caused at that exact location by handling the ballots after the voting.

Some Republican complained that chads were lying on the floor of the counting rooms during the hand recounting in Palm Beach, Florida. That is nothing to be concerned about, since chads that are hanging by one, two or three corners can be expected to drop off when the ballots are handled. Some such chads undoubtedly dropped off in the optical readers before the hand count, also. It is unlikely that normal handling of the ballots would dislodge a chad which was attached at all four corners.

Some Republicans claim that machine optical readers are more accurate than people ballot counters. Whether an optical reader reads punched areas which have hanging chads depends on how sensitive to light the reader is adjusted and whether the chad is pushed back into place by the next ballot card in the stack of cards. It could be that at different polling locations the light-sensitive settings for the optical readers are different. Only human counters can be sure to account for all types of hanging chads. Machine optical readers will not count any dimpled chads.

In fact, the Republicans requested and got hand counts of ballots in New Mexico, claiming the the hand count is more accurate than the machine count. Thus, the Republicans are being very dishonest saying that hand counts are not valid in Florida.

Integrity of Punch Ballots:

As someone who has used computer punched cards for many years, I am not concerned about the punch-card ballots being damaged by a hand count. Just the fact that some voters find it difficult to completely punch out a chad on a card indicates that accidental punches during hand counts are unlikely. Intentional fraudulent punches during hand counts are unlikely because observers of both parties are watching every move.

Absentee Ballots:

Most political pundits say that absentee ballots favor Republican candidates over Democrats, probably because the well-to-do are the ones who can afford to go overseas and because the military overseas who vote are weighted toward the officers rather than the enlisted personnel. However, it is not clear to me whether the absentee ballots will favor Bush or Gore. There are two major components:

So it may be that the absentee ballots will favor Gore rather than Bush, contrary to most predictions. For more details about the absentee ballots, see http://cbsnews.com/now/story/0,1597,248594-412,00.shtml.

However, it turned out that the late-arriving absentee ballots gave Bush an extra vote margin of ~600.

A Republican election commissioner in Seminole county violated the law by writing in the voter numbers for Republican voters on absentee ballot requests. Republicans have claimed that it is ok to change the rules of the election after the election is over and allow these ballots to be counted. In fact, they counted these ballots.

Banana Republic?:

I got the following e-mail:

A history professor from Uppsala University in Sweden read an article in which a Zimbabwe politician was quoted as saying that children should study what is going on in the U.S. elections closely for it shows that election fraud is not only a third world phenomenon.

  1. Imagine that we read of an election occurring anywhere in the third world in which the self declared winner was the son of the former prime minister and that former prime minister was himself the former head of that nation's secret police (CIA).
  2. Imagine that the self declared winner lost the popular vote but won based on some old colonial holdover (electoral college) from the nation's pre-democracy past.
  3. Imagine that the self declared winner's 'victory' turned on disputed votes cast in a province governed by his brother.
  4. Imagine that the poorly drafted ballots of one district, a district heavily favoring the self declared winner's opponent, led thousands of voters to vote for the wrong candidate.
  5. Imagine that that members of that nation's most despised caste, fearing for their lives/livelihoods, turned out in record numbers to vote in near universal opposition to the self declared winner's candidacy.
  6. Imagine that hundreds of members of that most despised caste were intercepted on their way to the polls by state police operating under the authority of the self declared winner's brother.
  7. Imagine that six million people voted in the disputed province and that the self declared winner's 'lead' was only 327 votes. Fewer, certainly, than the vote counting machines' margin of error.
  8. Imagine that the self declared winner and his political party opposed a more careful by hand inspection and re-counting of the ballots in the disputed province or in its most hotly disputed district.
  9. Imagine that the self declared winner, himself a governor of a major province, had the worst human rights record of any province in his nation and actually led the nation in executions.
  10. Imagine that a major campaign promise of the self declared winner was to appoint like-minded human rights violators to lifetime positions on the high court of that nation

Good URLs about the Election:


Cartoonists View of the Electoral Circus:


Canadian 2000 Election:


Voting Machines:

It was shocking that Palm Beach, Florida, and apparently many other places in the U.S. use such antiquated balloting methods as punched cards. My locality of Blacksburg, Virginia uses a better, but still antiquated, method of pulling levers, which prevents double-voting. So the Palm-Beach voters do not have the same opportunity as I have when voting of being assured that they will not accidently double vote.

The federal government needs to supply funds to all localities in the U.S. to modernize the voting methods.

http://alalm.org/Technology in the Voting Process.html

Internet Voting:

Would voting by Internet increase the accuracy and speed up the vote counting? See: http://www.cnn.com/2000/TECH/computing/11/08/e.voting.no.gamble.idg/


"When I was coming up, it was a dangerous world, and you knew exactly who they were," he said. "It was us vs. them, and it was clear who them was. Today, we are not so sure who the they are, but we know they're there."—George W. Bush at Iowa Western Community College, Jan 21, 2000

Yeah, Dubya, they are people who want to count the votes.

(Back to Politics)