Is handwriting analysis a valid forensic tool?
April 8, 2011 1:44 PM Subscribe
Is handwriting analysis a valid forensic technique?
I was watching the movie Zodiac for about the fifteenth time the other night, and I have a question about one of the elements in it. In the movie (and presumably, in real life), several people were written off as suspects of being the Zodiac killer because their handwriting didn't match that of the Zodiac killer, according to experts in handwriting analysis.
My question is this: is this really a valid forensic technique?
To be clear: I don't particularly care about the Zodiac case specifically. I'm sure that issue of the case has been debated endlessly by people interested in the case, and I'm not interested in jumping in to that debate. I'm curious about this issue generally. Also, I'm not talking about the "the crossbars of the T's in this persons handwriting has a pronounced downward slope, therefore he likes to strangle people with piano-wire", that is obviously nonsense. I'm asking whether a sufficiently qualified expert can make a definite pronouncement on whether someone has written a particular document based on handwriting alone.
The reason I'm asking is that in the last few years I've been reading shocking articles about people that have been convicted of crimes based solely on the opinions of so-called "experts" in different forensic techniques, when it turns out that those experts have no actual science backing them or their forensic specialty. The most obvious and terrifying example is the story of Cameron Todd Willingham who was convicted (and executed) mostly because of fire investigators who didn't know what the hell they were talking about, but there are many others. Radley Balko at Reason (a person with whom I disagree on just about everything except for his reporting on crime and criminal investigators) has written a lot about incompetent medical examiners and "bite mark analysts" who'll just go along with whatever prosecutors want them to say.
It seemed to me, watching the movie, that handwriting analysis belongs in the same category. At one point, when one expert disagrees with them, they go to another one and badgers him into saying that, yes, if a person wrote with his left hand, he could change his handwriting. That doesn't sound very scientific, and I have a hard time imagining a person analyzing DNA doing something like it. And the premise that a person cannot fake another way of writing also seems suspect to me; I mean, there's a whole profession based on being able to write in many different scripts. More than that, there's lots of handwriting courses available for adults where the very idea is to change how someone writes. I find it very hard to believe that if a person really, really wanted to, he couldn't write a letter in a hand unrecognizable from his own. It'd be probably be difficult, and take many drafts, but surely you could do it? And wouldn't that be something you'd want to do if you're a serial killer intent on sending letters to newspapers?
Anyway, that was a bit longer than I intended, but my question is simple: is there good scientific backing of handwriting analysis? Can an expert really always identify the author of a text based solely on handwriting? Surely there have been studies done on this?
I was watching the movie Zodiac for about the fifteenth time the other night, and I have a question about one of the elements in it. In the movie (and presumably, in real life), several people were written off as suspects of being the Zodiac killer because their handwriting didn't match that of the Zodiac killer, according to experts in handwriting analysis.
My question is this: is this really a valid forensic technique?
To be clear: I don't particularly care about the Zodiac case specifically. I'm sure that issue of the case has been debated endlessly by people interested in the case, and I'm not interested in jumping in to that debate. I'm curious about this issue generally. Also, I'm not talking about the "the crossbars of the T's in this persons handwriting has a pronounced downward slope, therefore he likes to strangle people with piano-wire", that is obviously nonsense. I'm asking whether a sufficiently qualified expert can make a definite pronouncement on whether someone has written a particular document based on handwriting alone.
The reason I'm asking is that in the last few years I've been reading shocking articles about people that have been convicted of crimes based solely on the opinions of so-called "experts" in different forensic techniques, when it turns out that those experts have no actual science backing them or their forensic specialty. The most obvious and terrifying example is the story of Cameron Todd Willingham who was convicted (and executed) mostly because of fire investigators who didn't know what the hell they were talking about, but there are many others. Radley Balko at Reason (a person with whom I disagree on just about everything except for his reporting on crime and criminal investigators) has written a lot about incompetent medical examiners and "bite mark analysts" who'll just go along with whatever prosecutors want them to say.
It seemed to me, watching the movie, that handwriting analysis belongs in the same category. At one point, when one expert disagrees with them, they go to another one and badgers him into saying that, yes, if a person wrote with his left hand, he could change his handwriting. That doesn't sound very scientific, and I have a hard time imagining a person analyzing DNA doing something like it. And the premise that a person cannot fake another way of writing also seems suspect to me; I mean, there's a whole profession based on being able to write in many different scripts. More than that, there's lots of handwriting courses available for adults where the very idea is to change how someone writes. I find it very hard to believe that if a person really, really wanted to, he couldn't write a letter in a hand unrecognizable from his own. It'd be probably be difficult, and take many drafts, but surely you could do it? And wouldn't that be something you'd want to do if you're a serial killer intent on sending letters to newspapers?
Anyway, that was a bit longer than I intended, but my question is simple: is there good scientific backing of handwriting analysis? Can an expert really always identify the author of a text based solely on handwriting? Surely there have been studies done on this?
I am a forensic scientist, though not a handwriting analyst, but did cover some handwriting analysis type stuff durring my schooling and training.
Handwriting analysis is still used and like you suspected can be really subjective in some cases, and an opinion can vary between different examiners. The thing to keep in mind is for ever Zodiac investigation there are 1000 cases where they are looking at a check thats been forged for 100$ or a bank robbery note which are usually cases where the perpetrator has made little to no effort to hide or alter their handwriting. Cases like those are usually a lot less subjective but in the end it still comes down to an experts opinion as to weather the question handwriting matches the suspect sample. Could a determined criminal fake his handwriting to throw off an investigator?...sure....though it would be way less work to just type his letter and print it out at say kinkos.
Because of it's subjective nature handwriting analysis isn't really used ALL that often. These days DNA is really king, and with advancements in low copy type DNA work a profile can be obtained from as little a 5-10 skin cells so using something like handwriting analysis to identify a totally unknown subject or eliminate a pool of suspects is really more of an additional investigative tool or even a last resort sort of thing.
posted by Captain_Science at 2:08 PM on April 8, 2011 [2 favorites]
Handwriting analysis is still used and like you suspected can be really subjective in some cases, and an opinion can vary between different examiners. The thing to keep in mind is for ever Zodiac investigation there are 1000 cases where they are looking at a check thats been forged for 100$ or a bank robbery note which are usually cases where the perpetrator has made little to no effort to hide or alter their handwriting. Cases like those are usually a lot less subjective but in the end it still comes down to an experts opinion as to weather the question handwriting matches the suspect sample. Could a determined criminal fake his handwriting to throw off an investigator?...sure....though it would be way less work to just type his letter and print it out at say kinkos.
Because of it's subjective nature handwriting analysis isn't really used ALL that often. These days DNA is really king, and with advancements in low copy type DNA work a profile can be obtained from as little a 5-10 skin cells so using something like handwriting analysis to identify a totally unknown subject or eliminate a pool of suspects is really more of an additional investigative tool or even a last resort sort of thing.
posted by Captain_Science at 2:08 PM on April 8, 2011 [2 favorites]
This American Life did a story last month about some JFK archives/letters which were thoroughly analyzed and declared to be legitimate. As the collection started to be sold for millions of dollars, other analysis proved it to be false. There are lots of references in that piece about the science of handwriting analysis.
In searching for more details about that story, I found this search page from the Chicago Tribune containing articles about handwriting and I see there are several articles on there about analyzing handwriting for the purposes of determining the author.
posted by CathyG at 2:08 PM on April 8, 2011
In searching for more details about that story, I found this search page from the Chicago Tribune containing articles about handwriting and I see there are several articles on there about analyzing handwriting for the purposes of determining the author.
posted by CathyG at 2:08 PM on April 8, 2011
The short answer to your question is: Yes, handwriting analysis is a valid forensic tool. Forensic document examiners, whose background includes a combination of experience and training, routinely testify in courts as expert witnesses concerning the authenticity of documents and the identity of the drafter.
Is there good scientific backing of handwriting analysis?
Hard to say. Most forensic document examiners rely solely on experience in offering their opinion. In most courts anyone, as long as they can establish familiarity with the handwriting in question (e.g. boss familiar with subordinate's writing), can testify as to the authenticity of writing or identity of drafter. The issue is the weight this testimony has with the finder of fact. A forensic document examiner often has a lot more experience with handwriting analysis and may have accreditation/membership with professional societies (e.g. abfde).
Can an expert really always identify the author of a text based solely on handwriting?
No. Their opinions are fact dependent, taking into consideration such factors as the quality of the questioned document and the quality of the control sample set.
posted by Mr. X at 2:09 PM on April 8, 2011
Is there good scientific backing of handwriting analysis?
Hard to say. Most forensic document examiners rely solely on experience in offering their opinion. In most courts anyone, as long as they can establish familiarity with the handwriting in question (e.g. boss familiar with subordinate's writing), can testify as to the authenticity of writing or identity of drafter. The issue is the weight this testimony has with the finder of fact. A forensic document examiner often has a lot more experience with handwriting analysis and may have accreditation/membership with professional societies (e.g. abfde).
Can an expert really always identify the author of a text based solely on handwriting?
No. Their opinions are fact dependent, taking into consideration such factors as the quality of the questioned document and the quality of the control sample set.
posted by Mr. X at 2:09 PM on April 8, 2011
Prosecutors rarely use it anymore except as "pile on" evidence. Unlike the scientific validity of DNA, a good defense lawyer will simply have an expert "handwriting analyst" spend 30 minutes on the stand telling of all the handwriting experts whose "analysis" was proved to be false time and time and time again. To a jury, handwriting analysis carries almost no weight after defense presentations of how it's simply not scientific. You can't run handwriting through any kind of widely accepted laboratory tests to get what juries fully expect in these days of CSI (DNA, blood, etc.).
posted by Gerard Sorme at 2:20 PM on April 8, 2011
posted by Gerard Sorme at 2:20 PM on April 8, 2011
This isn't exactly what you're asking about, but it's related and fun. And James Randi is adorable.
posted by phunniemee at 2:47 PM on April 8, 2011
posted by phunniemee at 2:47 PM on April 8, 2011
I would think that similarities in handwriting would be on par with images showing someone wearing what appear to be the same clothes. Easily faked, but possibly significant in context.
posted by StickyCarpet at 4:35 PM on April 8, 2011
posted by StickyCarpet at 4:35 PM on April 8, 2011
No, it's absolutely not a science. It's opinion. The methods used by these "experts" have no basis in a peer review process. There are no validation studies. There are no hypotheses. There are no published error rates showing true/false positives/negatives. Certainly none that show that a method is good independent of who wields it.
Is it valid to hear the opinion of someone in court about handwriting? Yes. Does it have evidentiary value? Not much.
You might be asked in court "Is that your mother's handwriting", and in your opinion and experience you can say yes or no. Or a cop can ask you the same question and take some action based on your answer. But that doesn't make you a scientist. Handwriting experts are not scientists either.
See a discussion of US v. Plaza, Acosta, and Rodriguez by Smith & Bace (pg 138) about whether fingerprint analysis (similarly, a pseudo-science) should be allowed in court as expert testimony. In that discussion, they talk about handwriting experts.
posted by about_time at 9:15 PM on April 9, 2011
Is it valid to hear the opinion of someone in court about handwriting? Yes. Does it have evidentiary value? Not much.
You might be asked in court "Is that your mother's handwriting", and in your opinion and experience you can say yes or no. Or a cop can ask you the same question and take some action based on your answer. But that doesn't make you a scientist. Handwriting experts are not scientists either.
See a discussion of US v. Plaza, Acosta, and Rodriguez by Smith & Bace (pg 138) about whether fingerprint analysis (similarly, a pseudo-science) should be allowed in court as expert testimony. In that discussion, they talk about handwriting experts.
posted by about_time at 9:15 PM on April 9, 2011
This thread is closed to new comments.
Yes and no. Can you derive information from handwriting? Sure. Can you tie a specific piece of writing to a specific person? Not so much.
Can an expert really always identify the author of a text based solely on handwriting?
No.
Now, let's talk forensics more generally. Just because there is "good science" behind a lot of forensics work, doesn't mean you get a 100% answer. Take DNA for example - if you put a DNA sample into CODIS, you don't get a CSI-like flashing screen that says, "This is your guy!" You get the ten closest matches, and it's the investigator/scientist's job to narrow from there.
There's very little in each forensics field that's about nailing someone dead to rights. But forensics, taken collectively, is about giving someone scientific evidence from which to identify someone or answer a question. It helps triangulate and narrow things down.
The second part of your question isn't a question about the validity of any forensics field. People like. Some people are crooked. That there are people who abuse the justice system doesn't mean that the way they do so is perminantly tainted, or scientifically invalid, or always wrong. It just means someone lied.
posted by NotMyselfRightNow at 2:07 PM on April 8, 2011 [1 favorite]