Is a Signature enough? Ridiculous!

Signatures have always been used for two purposes: to show our authorization of a transaction, and as a means of identification. That second function — identification — is now an unnecessary fudge-factor — at least with regard to voting.

Prior to the creation of Social Security, there was no national registry of citizens. Signatures were all that we had and all that we needed for voter identification. After all, we almost always voted in person, right in our own neighborhood.

We now have unique ID numbers — courtesy of Social Security and other systems such as driver licensing bureaus and state identification systems. We no longer need to rely on signatures for identification. Furthermore, signatures don’t work in this era of mail-in ballots.

This is not a bi-partisan issue. Republicans generally favor a voter ID requirement, while Democrats do not. Republicans will never achieve the political power they want and deserve until they relentlessly insist on real ID requirements instead of “signature matching.”

If you Google the topic of “voter signature matching,” almost every article you find will bemoan the terrible way signature matching is used to suppress the vote of our marginalized voters. Oddly, however, you won’t hear any of these social crusaders advocating for the replacement of signature matching with any other system (except, perhaps, the “honor system,” which is sure to have no honor at all). Many politicians understand that signature matching is the key to their survival.

Research shows very clearly that non-forensic document examiners, such as the nonprofessionals who work in the local election office, are not even slightly effective when it comes to spotting phony signatures.

In the Journal of Forensic Science, three scholars (Jodi Sita, Bryan Found, and Douglas Rogers) wrote: “To our knowledge, there are only a limited number of published studies that compare [forensic document examiners’] ability in handwriting identification with that of the average person.” In their study, published in 2002, they had 69 forensic document examiners [FDEs] and 50 non-professionals scrutinize 150 signature specimens. Only 3.4% of the opinions of formally trained forensic document examiners were erroneous, but almost 20% of the conclusions of the amateurs was erroneous. If those findings concern you, read on.

There are a few more studies involving non-experts, and in those studies the results are even worse. In a scholarly study (p. 485) by Oliver, Craig, and Nanette Galbraith, a group of 65 nonprofessionals wrongly asserted that signatures did not match 7.7% of the time. That error rate is not too bad.  However, 33% of the time, the nonprofessionals falsely asserted that signatures from different people were from the very same person. In the context of an election, that means that the fraudster’s ballot will be accepted as real.

There is another study with results that are even more troubling. In the early 1990s, Dr. Moshe Kam, a computer scientist, was asked by the FBI to research the effectiveness of forensic document examiners as compared to ordinary people. In one of his peer reviewed studies, Kam and his associates discovered that the amateurs did poorly, but mostly in one direction. They were almost as accurate as the experts when it came to identifying real matches (where one person actually wrote both signatures). However, when it came to distinguishing writing from different people, the amateurs had an abysmal failure rate of 38%. In other words, they thought the signatures matched when they did not match.

Based on the research just identified, how successful would a typical election worker be in spotting a fraudulent signature on a ballot envelope?

Some people may claim that this is just a matter of training; however, those people may be underestimating the skill and guidance required to become a competent FDE. The America Academy of Forensic Sciences requires “successful completion of a minimum 24-month structured training program under the direct and constant supervision of a qualified primary training officer.” That level of training will never take place in our 3,000 county election departments.

The solution is obvious: we need voter ID. Unfortunately, the difficulties are also obvious. Politicians (mostly Democrats) will fight with every insult, smear, threat, and lawsuit in their arsenal to prevent voter ID from becoming a reality in any state they control. To survive, Republicans will have to fight harder.