Understanding the Sunshine Count: Who is Watching
the Watchers?
by Alan Balch, Doctoral Candidate
In this article:
A Brief Review
Understanding
the "Who Would Have Won Florida" Stories
Judicial
Watch Equivocates
Debunking
the Heart of the Judicial Watch Report
Judging Judicial
Watch by Their Own Standards
The
Washington Times Distorts Judicial Watch's Findings
For a group that claims the moral high ground, Judicial Watch's announcement that "Bush is still the winner" following their partial "sunshine count" of the undervotes was contrived and misleading at best. Moreover, the reports in the media regarding this analysis are similarly spurious.
Pressed for time? Start with this section Debunking the Heart of the Judicial Watch Report or click here to link to a condensed version of this article
The whole focus on counting the undervotes was born out of the State
Supreme Court's decision to require 64 of Florida's 67 counties to manually
canvass all their undervotes (roughly 45,000 total ballots). Such
a count would not have included Palm Beach, Volusia, or Broward Counties;
the latter two because their undervotes were included in the certified
total prior to the Court's Decision, and Palm Beach because the Court's
decision included an order to add to the state's certified vote tally the
results of their completed but tardy manual recount (a net gain of 176
votes for Gore). In addition, the court decision required Dade's
partial manual recount (a net gain of 168 votes for Gore) included in the
certified total. These additions shrunk Bush's 537 vote lead down
to 193 votes. The undervote count was to include the overvotes
from all the remaining counties including Dade, which still had roughly
9,000 undervotes yet to be reviewed. As we all now know, the US Supreme
Court halted this count and closed the door on the Florida election.
Enter the "sunshine count" -- a review of the ballots by various organizations
made possible by Florida's open access laws to public records.
Understanding the "Who Would Have Won Florida" Stories
Those reporting the results of the sunshine reviews thus far seem bent on using them to play the "who really won Florida" game (in my opinion, they should be reporting what was found rather than declaring a winner). If such speculation is going to be promoted, then it seems only fair to try and place such predictions in the context of plausible election events with the potential ti change the election's outcome. At least three possible scenarios exist for trying to determine a "winner" based on the sunshine review (yes, I know there are more, but this list should suffice for the purposes of this article):
1) Who would have won if the State Supreme Court's final decision was left standing by the US Supreme Court? To explore this scenario, a complete analysis of all the undervotes from across the state (excluding Broward, Palm Beach, and Volusia) must be compared against a 193 vote Bush lead.
2) Who would have won if the manual recounts requested by the Gore campaign in 4 South Florida Counties were all completed and included in the state totals? Recall that everyone finished their manual recounts except for Dade (Palm Beach finished, but missed the court imposed deadline by a few hours). To explore this scenario, the results of a sunshine review of both the undervotes and overvotes in Dade County should be compared to roughly a 360 vote Bush following the manual recounts in Broward, Volusia, and Palm Beach.
3) Who would have won in a manual review of all the machine unreadable ballots from across the state (including undervotes and overvotes)? This scenario is what many critics and pundits were howling for (myself included), but it is the highly removed from actual events in Florida (a statewide manual recount of undervotes and overvotes was never on the table at any point).
Several media organizations have weighed in on the "who would have won Florida" debate. Click here for a brief synopsis of recent entries. The latest installment comes from a group called Judicial Watch with strong right wing ties, which recently reported the results of its partial sunshine count (partial meaning six counties). They assert "that a statewide recount of the Florida undervotes would not have changed the outcome of the presidential election." To summarize, Judicial Watch determined who would have won Florida if the State Supreme Court's statewide undervote count order had been followed based on the following: out of 45,000 from 64 counties slated to be counted, they counted 25,000 of them in six counties and reported a net gain of 106 to 117 Bush votes. They then speculate that a statewide recount would not have been enough to erase Bush's 193 vote lead resulting from the State Supreme Court's decision (of course they leave out the latter information).
They also "inspected" the ballots from Palm Beach and Broward, but undervotes from these counties would not have been counted during the statewide undervote count. Palm Beach and Broward conducted full manual recounts of all their ballots including overvotes and undervotes prior to the State Supreme Court Decision. The results of Broward's manual count were already included in the state's totals and the Supreme Court decision ordering the statewide undervote count also added Palm Beach's manual recount results to the official tally. The purpose of Judicial Watch's analysis of these counties was to double check the work of the canvassing boards and election officials, not to shed light on the "what if" scenario surrounding the undervote count. More details about Judicial Watch's analysis can be found here. There own press release is available here.
Thus, Judicial Watch essentially did two different types of analyses: one was the six-county undervote count to determine what would have happened if the undervote order was executed in only six counties and the other was the double check of what actually did happen during the manual recounts in Palm Beach and Broward counties. These two inspections are not commensurable for the purposes of determining who would have won Florida during a statewide undervote count. However, Judicial Watch combines the two studies into one to make it appear as if they studied the bulk of the undervotes that would have been investigated during the undervote count. This implication is false. Moreover, Judicial Watch's analysis is incomplete in two important but undisclosed ways: 1) their undervote count does not include roughly 45% of the undervotes spread unevenly across 58 counties; 2) they only "double checked" the entire canvass of ballots in two counties to assess what actually happened, leaving 65 counties unscrutinized and thereby failing to "judiciously watch" the questionable behavior and decisions of numerous other canvassing boards like those in Gadsden, Lake, Duval, Seminole, Orange, etc, etc.
Unpacking
the Key Sentence in the Judicial Watch Report
Here is the crucial excerpt from their analysis (download
entire report here).
"In total, we have inspected ballots representing 42,724 (68%) of the 62,605 ballots reported as undervotes in the certified State tally. Based upon the results of this inspection, the results of the Florida election would not have changed with a manual recount of these counties. Further, based upon the scope of our inspection and the analysis of the results thereof, it is our view that the results of the Florida election would not have changed with a manual recount of all the counties in the State."
Commentary: Technically, these sentences are accurate by themselves. The first is a statement of fact, and the second one is an opinion, albeit one not backed up with any evidence. However, these two sentences put together are highly misleading because they provide the appearance that their review was sufficient to determine who would have won Florida following the statewide undervote count -- an appearance dutifully reported by some media outlets as fact. If their goal was determining who would have won the statewide undervote count, then their analysis is significantly less extensive than they make it sound on paper. Relevant to the statewide undervote count, they studied approximately 25,000 of the 45,000 undervotes statewide that would have been canvassed pursuant to the State Supreme Court's decision. They studied ballots from six of the 64 counties covered by the Court's recount order. While their results from these six counties show that Bush's 193 vote lead would expand if only these six counties conducted undervote counts, they tell us nothing about the remaining 20,000 undervotes from 58 counties that would also have examined their undervotes along with these six counties.
Their 42,724 is derived by adding the Palm Beach and Broward undervotes, which they analyzed as part of their review of the ballot counting methods in these counties. However, the undervotes from these counties would not have been included in the statewide undervote count because they would have already been included in the state's certified totals prior to the undervote count. The 62,605 number they use includes all the undervotes from across the state including those from Broward, Palm Beach, and Volusia. Once again, the ballots from these counties would be irrelevant to determining the outcome of the statewide undervote canvass because these ballots would not have been included in that effort. Judicial Watch notes that these 62,000 ballots represent the total reported to the state; what they neglect to tell you is that this figure is not the number of ballots slated to be canvassed in the undervote count.
They apparently rely upon the 68% figure quoted above to make their study seem more expansive than it actually is in terms of determining the outcome of the statewide undervote count. They make it clear that their conclusion is only an opinion (i.e., "their view"), but their report offers no evidence or validation for this opinion. In other words, they did not show that Bush would have won in a statewide undervote count; they only speculate that he would have based on the results from the six counties they actually analyzed. Why should we assume that a review of the remaining 20,000 undervotes would not turn up a net gain of 300 to 500 or even 1,000 Gore votes? Their own analysis shows incredible variations across counties in the number of possible uncounted votes and the distribution of those votes between Bush and Gore. Click here for an example. No discernible pattern exists in the six counties they studied, so how can they use the results from those counties to predict the results of the undervote count from the remaining 58 counties and 20,000 undervotes? They only rationale they provide for such a conclusion is that their examination of a little more than half of the undervote ballots is sufficient in scope to justify such a conclusion. Their lack of empirical or logical evidence for this conclusion suggests an ulterior motive: a "Bush Wins Statewide Recount" headline. A more accurate and honest press release would have simply stated that Bush netted 116 votes in undervote count from six large counties.
Judging Judicial
Watch by Their Own Standards
Judicial Watch claims to be "an ethical and legal "watchdog" over
our government" in an effort to "promote a return to ethics and
morality in our nation's public life." Based on the way they
unethically distorted and misrepresented their own supposedly bipartisan
and unbiased study, it seem that this pledge does not apply to their own
efforts. In a clear effort to stretch the truth until they could
arrive at their desired conclusion (i.e., that Bush would have won Florida
during an statewide undervote count) and provide the media the necessary
fodder they need to spin the same story, Judicial Watch displays a flagrant
disregard for the type of ethics and morality so desperately needed in
our nation's political discourse.
In other words, they reported just enough information and blew the
little information they did gather out of proportion in order to reach
their politically loaded and partisan conclusions. As is true
of most politicians and political groups across the political spectrum,
ethical and moral behavior conveniently become irrelevant or excusable
in the expedient pursuant of one's own partisan agenda.
The Washington Times Distorts Judicial Watch's Findings
At least two major media outlets ran the Judicial Watch story: USA Today (click here for article) and The Washington Times (click here for article). The USA Today's coverage pretty much reported the key excerpt from Judicial Watch, which was examined already in detail. Thus, USA Today's piece was no more misleading than the Judicial Watch's Press Release and for the same reasons already covered up to this point. However, the Washington Times took the Judicial Watch Report and distorted it. Written by Audrey Hudson, the article was titled "Bush is victor again in latest Florida tally" and ran on Page A9 of the March 23, 2001 edition. Here are two key excerpts from that piece that show a flagrant disregard for accuracy.
The first sentence for the article was as follows:
"An independent audit of the Florida election process shows that
a statewide recount of the undervotes would not have changed the outcome
of the presidential election."
The audit was not of the Florida election process. It was a canvass of the undervotes in six of the 64 counties that would have conducted such counts pursuant to the State Supreme Court decision. The audit did include a review of the canvassing process, but only in 2 out of 68 counties. Such a limited review of ballots hardly constitutes an audit of Florida's election process. Moreover, the audit in no way shows that a statewide recount of the undervotes would not have changed the election's outcome as the Times asserts. It only shows that a manual canvass of the undervotes in six counties would not have changed the outcome, but it does not demonstrate the same for entire state. Judicial Watch makes it clear that their assertion that a statewide recount would not have changed the election outcome is an opinion based on the six counties they analyzed, not a demonstrated fact based on analysis of all the undervotes from across the state. In other words, it was the opinion of the independent auditors that a statewide undervote count would not have changed the election outcome; they did not, however, "show" this to be the case.
"Applying a strict standard of clear punches, hanging or swinging chads Mr. Bush's net increase was 116 votes. Mr Bush also received an additional 116 votes from dimpled chads, according to the audit."
Actually, Judicial Watch's "strict" standard included all types of chads (hanging, swinging, and tri chads), plus those with other holes – "A pinhole was punched through some portion of the candidate box but not necessarily the chad" and other marks - "the presidential chad had a distinguishing mark from a writing instrument, such as a circle or an “X”, which clearly showed intent to caste a vote in that instance." The second sentence in this excerpt is the especially misleading because it indicates that Bush would have picked an additional 116 votes if dimpled chads were counted on top of the 116 gained from those ballots with clear punches and dislodged chad. This statement is false. According to Judicial Watch's Analysis, Bush votes were found on 3,098 dimpled ballots in the six counties and Gore votes were found on 3,155 dimpled ballots. In fact, Gore gained 57 votes in the dimpled category, which would reduce Mr. Bush's net gain to 59 if the results from the dimpled ballots are added to the ones with clear punches, and hanging or swinging chads.
The Statewide Undervote Count
If those conducting the sunshine review are going to play this game
about "who really won Florida" (in my opinion, they should be reporting
what was found rather than declaring a winner), then the State Supreme
Court's ruling in its entirety is the logical starting point for such an
analysis. With Bush's lead shaved to 154 votes by the Palm Beach's
full manual recount and Dade's partial recount, it was the count of all
undervotes statewide that would have determined who actually won Florida
(assuming the Florida legislature and/or the Congress respected the results).
The Knight Ridder, USA Today, Miami Herald Consortium is conducting an
analysis to determine what would have happened had this final order been
carried out. Judicial Watch and the press have mistakenly placed
their results in a similar context. The only way the undervotes would
have been counted in Florida was via the State Supreme Court decision.
Those groups studying only undervotes seeking to declare a "true
winner" must base such an assessment on a full analysis of undervotes statewide
(e.g., not just the undervotes in Dade County or a handful of other counties)
and compare those results to a 154 vote Bush lead. To do anything
other ignores the State Supreme Court's decision that empowered the undervote
count in the first place and requires progressive steps backwards in the
recount Florida debacle.
Focus on Dade County
What would have happened if the manual recounts begun in 4 counties
pursuant to the protest guidelines had been completed and included in the
state certified total? In this scenario, Gore presumably would not
have contested the election so the State Supreme Court would not have ordered
the statewide undervote count. Recall that Broward and Volusia completed
their manual count and the new tallies were certified (Bush's 537 vote
lead reflects these manual counts), while Palm Beach finished a few hours
past the court imposed deadline (Bush's lead shrinks to roughly 350 with
their manual count). Dade's manual count barely got off the ground
and thus becomes the deciding factor in a "manual recount only" world wherein
all 4 manual recounts were completed and the statewide undervote count
order never existed. Would a full manual recount of all Dade's ballots
(including overvotes and undervotes) provided enough additional Gore votes
to overcome a 350 vote Bush margin? The Palm Beach Post and the Miami
Herald both ran stories (as did numerous major media outlets) suggesting
that this would not have been the case; they speciously anointed Bush the
"winner" once more. However, their reviews only considered undervotes
rather than a full manual recount. Predicting a winner based only
on Dade County logically requires a complete manual recount of all Dade's
ballots, not just the undervotes. The only way Dade would have
counted only its undervotes is pursuant to the State Supreme Court order.
However, that order also included similar counts in every county across
the state except for Palm Beach, Volusia, and Broward. For the purposes
of declaring a "winner," Dade's undervote count is meaningless until
the results of the statewide undervote count are reported and compared
to a 154 vote Bush lead. The results of a full manual count of all
Dade's over and under votes has yet to be reported, but should be out soon.
A Statewide Manual Recount
What would have happened if Bush had accepted Gore's challenge to a
full manual recount of the entire state? If all the undervotes and
overvotes were counted from across the state, who would have won?
The other major media consortium (including the Washington Post and the
New York Times) is taking this approach. They are conducting a full
manual recount of every machine unreadable ballot in every county.
This count will probably yield the most interesting results, but it is
also the farthest removed from the realities of the Florida vote count.
Such a count was never a possibility, unlike the Dade manual recount or
the statewide undervote count.
Summary
of What Judicial Watch Actually Found (as opposed to what they claim to
have found)
After reading Judicial Watch's official report and analysis (a curiously
large and disheveled piece available
for download at http://www.judicialwatch.org/pressrelease/floridarecount/recountreport.htm),
it is clear that their announcement of a "winner" is at best premature
and misleading . They completed a manual recount of the roughly 25,000
undervotes in six counties containing roughly 55% of all the undervotes
statewide (roughly 45,000 undervotes would have been included in the statewide
overvote count). These counties include Collier (Bush won 67% of
the Bush+Gore ballots), Hillsborough (52% Bush of Bush+Gore total), Indian
River (60% Bush of Bush+Gore total), Dade (47% Bush), Pinellas (48% Bush)
and Sarasota (53% Bush). The report contains no data regarding the
undervote counts from roughly 60 other counties that would have been included
in the undervote count, which collectively contain another 20,000 or so
undervotes. The auditing firm hired by Judicial Watch found that
using strict standards, Bush would have netted 116 votes from undervote
counts in these six counties. Their definition of strict includes
counting any variation of chad ballots (one corner, two corner, three corner),
ballots with marks other than punches that indicate the will of the voter
(such as Xs over the punchhole), and unpunched ballots with holes in the
chad. Even using loser standards, they report net gains for Bush
between 107 and 116 votes. What does this all mean? They counted
undervotes only, which means their analysis is grounded in the State Supreme
Court's last standing order. Based on the State Supreme Court's decision,
Bush's lead was down to 154 votes including the Palm Beach manual recount
and Dade's partial recount. What they really discovered was that
the results from these six counties would have increased Bush's lead to
roughly 270 votes (all of this is left out of their analysis). However,
the State Supreme Court undervote count covered the entire state, so their
analysis is missing 58 counties and 20,000 undervotes. Despite these
glaring gaps in their analysis, Judicial Watch declares that Bush would
have won Florida even after a statewide count of all the undervotes.
This conclusion assumes that the results from the remaining 58 counties
will track with the results from these six. The accuracy of such
an assumption will be revealed soon enough as a full report on all the
state's undervotes is due out next week.
Other "Who Won Florida Stories"
The Miami Herald ran a story (as did numerous major media outlets)
speciously anointing Bush the "winner" once more based on reviews of Dade's
undervotes. Predicting a winner based only on Dade County logically
requires a complete manual recount of all Dade's ballots, not just the
undervotes (See #2 above). The only way Dade would have counted
only its undervotes is pursuant to the State Supreme Court order.
However, that order also included similar counts in every county across
the state except for Palm Beach, Volusia, and Broward. The Palm Beach
reported that Gore probably would have won if not for the emphasis Butterfly
Ballot. While they presented some reasonable evidence, their results
are purely speculative. Moreover, such analyses beg other questions
like "Who would have won if the media had not declared Gore the winner
prior to the poles closing in Florida Panhandle?"
Example of Judicial Watch Results
They found 1,407 ballots with discernible voter preferences for Bush
or Gore in Collier County, which contained more than 90,000 certified votes
for Bush and Gore. A similar number of discernible markings were
identified in Dade (1,109) and Pinellas Counties (1,525), but these counties
contained more than 618,000 and 385,000 certified Bush and Gore votes respectively.
Furthermore, applying their liberal to moderate standards, two counties
generated large net gains for Bush (Collier and Indian River), one turned
up large net gains for Gore (Hillsborough), while the remaining three spread
their potential votes out fairly evenly.
