IRE/NICAR Tipsheet #165
The Philadelphia Inquirer
The story of absentee-ballot fraud in the Second Senate District
Handout for panel discussion: Elections, campaigns and voters."
IRE National Conference June 16-19, 1994
The State Senate race in Philadelphia's Second District would have attracted little notice were it not for a curious fact-the winner, Democrat Bill Stinson, lost on the voting machines. Stinson owed his narrowed victory to absentee ballots. Some 1,757 were cast last Nov 2-an extraordinarily high number-and he got nearly 80 percent of them. His opponent, Republican Bruce Marks, cried foul, and The Inquirer began investigating. The resulting articles detailed pervasive electoral fraud. We fond ballots cast on behalf of dead people, fictitious people and real people living in such places as Las Vegas, Greenland and Crete. We found ballots vast from vacant lots and abandoned buildings. On March 25, we published a lengthy wrap-up demonstrating that the fraud was so extensive that it tilted the election to Stinson---or put another way, that Marks would have won if the race had been clean. The coverage led to criminal charges against Stinson, a grand-jury probe of the city's electoral system and a federal judge's ruling tossing out the election and awarding the Senate seat to Marks.
The Inquirer's investigation moved on three tracks. The first was interviewing voters. From early November to mid-March, we questioned 1, 579 absentee voters (90 percent of the total), collecting accounts of coercion, intimidation, forgery and other abuses. Almost all the interviews were done face-to-face. Simultaneously, we built a paper trail of public records for each voter verifying that absentee ballots were requested and cast in their names. Finally, we developed a computer database to help us figure out what it all meant.
The world of electoral records-an alien one to most of us-proved an exceptionally rich source of information.
Voter-registration cards in Philadelphia list (along with address and party affiliation) vital statistics, date of birth, place of birth, and occupation, Each card bears an original signature. Notation on the rear indicates each election in which the person voted and whether he or she did so by machine or absentee ballot. From this , we could se that people who had been voting at the polls for years had suddenly, last fall, cast absentee ballots.
The signed applications for absentee ballots are also public records in Pennsylvania, as are the envelopes in which the ballots are filed. On the rear of each envelope is an affidavit-also signed-in which the person attests that he or she is qualified to vote.
To mine this vein of information, we shipped a photocopier to the city Board of Elections, stationed as may as four reporters there at a time, and photocopied for days on end. We assembled copies of the registration, ballot application and envelope for nearly everyone who cast an absentee ballot. When these documents were laid side by side, signature discrepancies leaped off the page. Ultimately, we verified 52 instances of forgery.
The electoral paper tail is rich in redundancy, which is fortunate because many documents related to the Senate race were destroyed or misplaced by election officials. Others were seized by the state Attorney General's Office as part of its criminal investigation. We usually were able to find substitute records to the fill the gaps.
At each Philadelphia polling place, election officers keep a "No. 2 book" listing in order of their appearance, each person who cast a vote by machine that day, along with those who had previously filed absentee ballots. These logs helped us verify absentee votes when other records were unavailable. They also revealed that in dozens of instance, people voted twice-once by machine, one by absentee ballot. Voter "certificates" were another valuable record. In Pennsylvania, each voter must sign one, attesting to his or her eligibility to vote, before stepping into the voting booth. These provided another signature specimen. They also pointed to a new dimension in the vote fraud-phantom machine votes. We found instance where votes were cast by machine for people who had long since moved out of the voting district, the city of the country.
Information management was key to the project's success, As documents poured into the newsroom, and as reporters dumped mountains of notes into our computer system, the sheer quantity of data threatened to overwhelm us. A library clerk was detailed to organize our paper files. And we put together a database, using the software program dBase, to store and analyze material.
In-house computer experts designed a standard form to record information about each absentee voter. It listed the person's name, address, age, race, and part registration, along with codes indicating whether the absentee ballot was obtained by fraud and, if so, what kind. News clerks inputted this information from paper files and reporter's notes.
Once the database was up and running, we could slice our material in all kinds of esoteric ways-and almost instantly. How many of the 1,579 absentee votes we had studied were IABCs (code for fraud-tainted ballots)? It took dBase 10 seconds to search the files and spit out the number. (Our final tally showed 540 irregular absentee ballots, nearly a third of the total cast.) We could also parse the fraud by age, ethnic group, party affiliation and geographic area. A major finding was that irregularities were heavily concentrated in Latino neighborhoods. This tallied with anecdotal information (later corroborated by sworn testimony in federal court) that Democratic Party and Stinson campaign workers had blitzed the barrios of North Philadelphia in the weeks before the election, pressuring people to vote absentee and, in some cases , marking their ballots for them.
Other electronic reporting tools assisted our research. We obtained a nine-track computer tape listing all registered voters in the Senate district - and another with the names and address of the city's 4,000 Democratic and Republican Party Committee people. We used dBase to merge the files and produce a list showing how many people were registered to vote from each committee person's home (in one case, the number was 14). This yield a wealth of reporting leads and helped us produce a story about ghost voters-people casting ballots from addresses they had left long ago.
In addition to court, real estate and other paper records, we made extensive use of on-line databases. Here are several we found helpful.
7721 Hamilton Ave. Cincinnati, Ohio 45231
A national database of credit, motor vehicle, criminal court and phone records. It generates addresses, phone numbers, social security numbers and dates of birth.
One Georgia Center
500 W. Peachtree St. NW
Atlanta, GA 30338
Real estate and other public records in 32 states. People Finder function generates addresses, phone numbers, lists of neighbors.
5000 Arlington Center Blvd.
PO Box 20212
Columbus, OH 43220
Phone File scans nationwide data base of standard and crisscross phone directories
Mead Data Central
9393 Springboro Pike
PO Box 933
Dayton, Ohio 45401
Nexis/Lexis ASSETS library contains property records nationwide. Finder function generates phone numbers and addresses