Investigative journalist, Greg Palast, winner of six Project Censored Awards originally brought the "caging lists" from Duval County, Florida to widespread attention in his report for an October 26^th, 2004 BBC Newsnight program.

ePluribus Media reporters also conducted an inquiry into DuvalCounty "caging".  Our analysis of the available data seems to point to discrepancies and contradictions with what was reported by Palast. Since any efforts to suppress voters by caging would be serious ^__ and possibly illegal ^__ we believe the reporting on caging should be restricted to the facts.

In his book, Armed Madhouse, and in press releases, Palast cites one "Randall Prausa" as an example of an African American Serviceman disenfranchised by caging. On page 204 of this book he reports:

We checked one list that included 50 black soldiers. We called one, Randall Prausa. His wife indicated that his address had changed because he was shipped overseas. Go to Baghdad, lose your vote. Nice. A black soldier's vote gone. Mission Accomplished.

In addition, in a June 2^nd, 2006 report, by Palast contains this:

Was it deliberate? Oh, my God, yes. I'd like you to take a look at the "caging" lists the Republican National Committee concocted to challenge voters with "suspect" addresses. It included page after page of African-American soldiers, like one Randall Prausa, shipped overseas. Mission accomplished, Mr. President?

The problem with this choice of Randall Prausa to illustrate disenfranchisement of African American servicemen by caging is that ^__ Randall Prausa is not black. [see photo above of Seaman Randall Prausa]

Reached by email and telephone, Mr. Prausa confirms that he previously lived at the address listed on the caging list. He also reports that he was deployed as a sailor, but voted by absentee ballot in 2004. Review of the DuvalCounty election data confirms that Mr. Prausa voted absentee in 2004, and his ballot was counted. However, he also is white, and states that he has never spoken with Palast or granted permission for his name to be published in any books. His wife, Nan, states that she was called by the BBC in October 2004, but that the question of ethnicity never came up. Neither Randall nor Nan Prausa has been contacted by BBC or any other person since October, 2004 ^__ at least, not until we contacted them at the end of May, 2007. Additionally, the 50 "black soldiers" on the caging list are actually sailors, listing "Naval Air Station" as their address.  Of the fifty sailors listed, 38 remained on the Voter Registration roll, with nine Black, four Hispanic , and nineteen White. Nine of them are registered Democrats and twelve are registered Republicans.

Caging Data?

Palast reported that the names of 1886 voters in Jacksonville, Florida had been found on "caging lists" received by Republican political operative Tim Griffin. According to Palast, "Two e-mails, prepared for the executive director of the Bush campaign in Florida and the campaign's national research director in Washington DC, contain a 15-page so-called 'caging list'. It lists 1,886 names and addresses of voters in predominantly black and traditionally Democrat[ic] areas of Jacksonville, Florida". The claims of caging in 2004 have been repeated by Palast in books and interviews since 2004 and often touted as evidence of voter suppression; however, there were no poll challenges in Duval County in 2004, which would be the most likely way of using the "caging" information.

As described in the text, the 25 .csv files attached to one GeorgeWBush.org email do not appear to be caging lists. Furthermore, these lists are sorted by voting precinct, and contain addresses from 23 of Broward County's 816 precincts. In total, Broward County Registered voters are 18.6% Black, 61.7% White and 12.5% Hispanic. Of the 23 selected precincts, only one has a higher percentage of Black registered voters than 18.6%. Indeed, of the selected precincts, the Average Black % is 6.2, White % is 72.9 and Hispanic 13.0. Thus, the .csv files do not look like caging lists, and they target Broward precincts with low Black population percentage. They are more consistent with Republican Party "get out the vote" efforts. Of note, The ZIP code corresponding to Plantation Florida, 33211, is included in the .csv files.

According to his book Armed Madhouse, in October, 2004 Palast received those two emails from John A. Wooden, the owner of the parody websites WhiteHouse.org and GeorgeWBush.org. We contacted John Wooden to help illuminate the contents of the files at his website. According to Wooden, the emails were among those collected by his parody website when people mistakenly sent email to any address ending with "@georgewbush.org". The site's catch-all email address was closed October 29, 2004, ending the collecting process. These emails are believed to be authentic Republican National Committee emails accidentally collected by the parody site when the sender entered dot-org, instead of dot-com.

Wooden states that he sent the emails containing two "caging lists" to Palast in October, 2004. Later, he found it odd that Palast never spoke with him about them. Palast's office did not request additional information from Wooden until March, 2007, when Wooden provided the "entire stash" of GeorgeWBush.org emails (totaling 620 emails). From 2004 through March, 2007, then, Palast's GeorgeWBush.org emails consisted only of what was presented publicly on the GeorgeWBush.org website. After March 2007, Palast did obtain the remainder of the GeorgeWBush.org emails. We have since also obtained the entire cache of Wooden's emails. We quickly determined that, other than what is publicly available, there are no attachments with "caging" in the file name, and no spreadsheets resembling the "caging" lists in the emails that have not been made public.

Palast, on page 200 of Armed Madhouse characterizes the emails:

After our team spent hours going over them, it became clear that virtually every list- GeorgeBush.org captured over 50 of them, with tens of thousands of names- were made up entirely of voters in African-American neighborhoods that we checked against ZIP codes. The lists included the ghettos of Lauderdale, Pompano Beach and a town with the Gone-with-the-Wind name of Plantation, Florida.

On June 16^th, 2006, Greg Palast posted on his website

The BBC obtained several dozen confidential emails sent by the Republican's national Research Director and Deputy Communications chief, Tim Griffin to GOP Florida campaign chairman Brett Doster and other party leaders. Attached were spreadsheets marked, "Caging.xls." Each of these contained several hundred to a few thousand voters and their addresses.

Caging Lists: The Real Numbers

Palast refers to ZIP codes in his description. Although the precinct racial statistics are much more powerful in that they assign addresses to smaller windows of registered voters, and regardless, definitive racial information can be obtained from the actual Voter Registration lists, we also performed ZIP code analysis to evaluate Palast's claims. We evaluated all of the ZIP codes on the comprehensive caging list ... more

We reviewed the entire cache of emails, including the publicly available emails obtained by John Wooden, which are, Palast implies in his book and June 16^th, 2006 press release, the source of much of Palast's analysis. From October, 2004 through March, 2007, only 129 emails were available publicly on John Wooden's website, though these include the emails containing "caging lists." The two emails including attachments containing the word "caging" contain 1833 unique names and addresses. There is one attachment, "Douglas County Voter Fraud Master Spreadsheet.xls" that is a list of 21 names from Douglas County, Nevada, that also appears to be a caging list based on file structure. None of the other attachments to the public or private GeorgeWBush.org emails can reasonably be considered a "caging list".

One attachment, "Lakeland FTM JAX Melbourne POTUS rallies1.xls", contains 178 names from all over Florida, and is apparently a list that pertains to a rally featuring President George W. Bush. It cannot be reasonably considered a "caging list". Another email within the Wooden public files contains 25 spreadsheets in comma-seperated-values (.csv) format. These files (24 of which can be opened) include 3496 names sorted by precincts in and around Ft. Lauderdale, Florida, and includes ZIP codes in Pompano Beach and Plantation, Florida. None of these files include the word "caging" in their filename, and indeed, they include telephone numbers and appear to catalogue responses to surveys, quite unlike the caging list files, which include information about when and why the names and addresses were "caged".

Of the public email attachments, three files list churches or evangelicals in Minnesota, Pennsylvania, and Dallas County, Iowa. One Excel attachment outlines "poll-watching" plans in Washington State. Finally, four files concern contact efforts by GOP "get out the vote" teams in Colorado, Pennsylvania, Broward County and Palm Beach, Florida. These latter files indicate contacts by phone calls or "door knocks" with 68,350 people. In total, there are 72,045 total people identified or indicated as part of a statistic of "being contacted" in the email attachments but just 1833 names appear on the only lists that could reasonably be considered Jacksonville "caging lists". Palast's claim of finding GeorgeWBush.org emails containing more than two Florida caging lists appears inconsistent with the actual GeorgeWBush.org files.

Caging Lists: Suppression of Minority and Democratic Voters

The main concern with the use of "caging lists" is whether they are were used to specifically suppress minority and Democratic votes as Mr. Palast claims in his book and again:

June 16, 2006: A check of the demographics of the addresses on the "caging lists," as the GOP leaders called them indicated that most were in African-American majority ZIP codes.

March 28, 2007: And we dug in, decoding, and mapping the voters on what Griffin called, "Caging" lists, spreadsheets with 70,000 names of voters marked for challenge. Overwhelmingly, these were Black and Hispanic voters from Democratic precincts.

In fact, although our analysis found that the effect of caging identifies addresses of African-American voters more frequently than "white" voters, the acquisition of caged addresses was not restricted to "black" precincts. Of the 90 precincts containing higher than Jacksonville average black population (29%, "High Black"), 84 (93.3%) contained caged addresses. Of the 149 precincts with a higher than Jacksonville average white population (64.5%, "High White"), 111 (74.5%) contained caged addresses. Because of the higher caging rate in "black" precincts, the number of addresses caged in precincts with higher than Jacksonville average black population was greater (938 "High Black" vs 731"High White"), but the data by precinct analysis do not support Palast claims that the caging lists "were made up entirely of voters in African-American neighborhoods.. In fact, when looking at the race of the voters on the list using Duval County voter registration data, out of 1510 caged voters for whom an exact match could be made, 739 were Black, 78 Hispanic and 517 White.

When looking at the party affiliation of the voters on the list using Duval County voter registration data, out of 1510 caged voters for whom an exact match could be made, 795 were registered Democrats and 269 registered Republicans, whereas 446 were not affiliated with these parties. Precinct analysis found higher caging rates and a higher total number of addresses in Democratic-leaning precincts than Republican leaning precincts (1234 addresses in 144 precincts in which registered Democrats outnumber Republicans; 509 addresses in 141 precincts in which registered Republicans outnumber Democrats). However, of the 144 Democratic-leaning precincts, only 69 had a white population of less than 50%, whereas 75 had a white population of greater than 50%. Thus, claims that caging list addresses were overwhelmingly "Black and Hispanic voters from Democratic precincts" are exaggerated.

Caging Lists: Other Reporting by Greg Palast

With respect to the "Homeless Shelter" data, the address cited as a homeless shelter is indeed a homeless shelter. However, searching the "caging1.xls" file shows the address column was sorted independently of the attached data. In other words, the presented image does not reflect correct data. The names shown residing at the homeless shelter (Lewis, Lofton, Luckett, etc) are present sequentially in the unsorted source file and reside at different addresses. The image presentation makes it difficult to determine the number of people at the homeless shelter address. The presentation of only last names beginning with "L" or "M" suggests many more exist at that address. That the bottom of the list is covered with another image also leads one to believe there are more names than shown. In fact, the source file reveals 28 names at this address.

With respect to "Deployed Soldiers" data, how they were identified as "deployed soldiers" is not disclosed. From the published image, the beginning and end of the list is not shown. The source file reveals 50 residents at "Naval Air Station", and again, the address column was sorted independently of the attached data. For example, whereas the image shows five last names beginning with "B" at the Naval Air Station, the source file shows one. The "B" last names listed are present in the unsorted source file, but none are at the Naval Air Station. Again the image is presented without showing beginning or end, which effectively conveys larger number of names than actually present.

In addition to the description of total numbers of caging lists and caged names, characterization of "African American ZIP code" targeting, other errors exist in Palast's book and press releases. One example involves a graphic produced from Caging-1.xls which is found on page 207 of Armed Madhouse and publicly available at Palast's Flickr account. This figure, intended to highlight caging of addresses belonging to people in a "homeless shelter" and "deployed soldiers," was generated by manipulating the data in Caging-1.xls without documentation (see Sidebar "Homeless Shelter" Data). In addition to not acknowledging the manipulations in the book or press releases, the graphic representation of the data creates a visual impression of more caged homeless or servicemen than actually present in the source data. Palast responded to questions about this manipulation on May 29^th, 2007: "The image that appears on flickr was manipulated merely for illustrative purposes and was put together by someone within my publisher's art department".

To summarize, the evidence for caging in the 2004 elections that we have seen and that comes from the GeorgeWBush.org website consists of two files comprising one list with 1833 unique names from Duval County, Florida, and one separate file listing 21 names in Nevada. Other emails from this website that Palast had access to for his book and reporting prior to March, 2007 do not contain caging lists or anything that could be reasonably considered a caging list. Private emails from this website that Palast obtained after March, 2007 do not contain any files that could be reasonably considered caging lists. The process of developing a caging list by sending letters to newly registered voters (as was admitted by the Florida Republican party regarding the Jacksonville caging list) results in the "caging" of addresses at a higher frequency in black and poor ZIP codes, as well as in Democratic-leaning precincts. However, the description of the caging list as containing exclusively or predominantly voters from "Black" ZIP codes or Democratic precincts is not correct. This distinction is critical; if we are to resist biased or illegal voter suppression techniques, the arguments should be based in fact and free from errors.

All of that notwithstanding, Palast deserves tremendous credit for bringing the caging lists to public attention.

Further, Palast's BBC report on the Florida caging list and the RNC knowledge and approval of this list may have played a substantial role in preventing caging lists from being used for widespread voter challenges in the 2004 election. Palast's focus on the issue of caging may help prevent unethical or illegal voter suppression in future elections. Regardless of this potentially heroic impact, investigations into the acquisition and use of caging lists by political parties must be based on careful research and legitimate reporting that includes verifiable sourcing and truthful accounts. The importance of these investigations to voter rights and our Democracy is simply too monumental to allow for mistakes.

ePluribus Media Contributors: standingup, cho, publius revolts, Aaron Barlow, BronxDem