# Validity of Graphology

**URL:** <https://boards.straightdope.com/t/validity-of-graphology/169710>\
**Category:** Cecil's Columns/Staff Reports\
**Created:** [April 19, 2003, 10:30pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710 "2003-04-19T22:30:47Z")\
**Posts on this page:** 20\
**Page:** 6

<div class="post-metadata">

**Author:** ![Achernar](https://avatars.discourse-cdn.com/v4/letter/a/e274bd/32.png) [@Achernar](https://boards.straightdope.com/u/Achernar)\
**Post date:** [April 25, 2003, 2:34am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/101 "2003-04-25T02:34:21Z")

</div>

> [@](#):
>
> \*Originally posted by Jennifer McIlwee Myers \*  
> \*What is the terminology here for going off-topic with a vengeance? I’m going to do it.

We call it a hijack. 🙂

---

<div class="post-metadata">

**Author:** ![Irishman](https://avatars.discourse-cdn.com/v4/letter/i/b487fb/32.png) [@Irishman](https://boards.straightdope.com/u/Irishman)\
**Post date:** [April 25, 2003, 6:29am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/102 "2003-04-25T06:29:15Z")

</div>

**zut** , I read your response regarding the non-graphologist control. The analogy certainly helps explain. I see now how it isn’t much help. On the other hand, you list a lot of other factors that could conceivably be used instead of personality for sorting. That was my intention with the control. We’re not trying to compare using personality versus using job description or using gender + handedness + size of the hand. I suppose the method of control we have is making all the communications open after the test.

Of course, the weakness with the entire test seems to be the lack of ability to define personality traits and quantify them.

**Bill H.** , I’m sorry you don’t feel like continuing. I was curious by your strong assertion that you can identify people to their handwriting samples by personality. I wonder if you’ve ever instituted any type of controls on your “party trick” version. I was hoping we could devise a test that would apply the right controls to weed out exactly the kind of carnival psychic nebulous personality statements that Dex was thinking of. That’s why I was striving for the “match the samples to the writer” over just a list of the descriptions. However, the list of descriptions on your part would be a good follow-up for your train of thought. Unfortunately, this thread doesn’t seem to be headed in that direction. For my own part, I’ve not had any experience in trying to devise tests, and I hoped this would prove a learning experience for me in that manner. In a way it has, though not in the way I had hoped.

> [@](#):
>
> If I have a question on an article in the San Jose Mercury News (as an example), I can and do write to the author and they’ll typically respond to me. In the case of the Straight Dope, this is the appointed channel by the Chicago Reader to get responses. No response? That’s fine too; **Cecil** doesn’t owe me crap. But I’d appreciate it, and this is my request for it.

I wouldn’t go so far as to say this is _the_ approved channel. It is certainly one channel that The Chicago Reader has made available, but I don’t expect this precludes the old fashioned paper mail letter to Cecil Adams c/o The Chicago Reader. Given that Cecil himself has not always paid close attention to the boards, perhaps you might want to try that route.

> [@](#):
>
> But more to the point about posting the bibliography of a cite: giving a list of studies does not make them relevant or accurate. Especially as they are completely unreviewed. And the fact that they are extensive, and not one of them has been reviewed makes presenting them a filibuster at best, and a diversion at worst.

I think I see your point about how nobody here reviewing them to see if they actually apply, but I’m still confused. You asked Cecil for a list of cites, now you’re really asking for an in depth analysis of 200 cites as to how they apply to the issue of using graphology to assess personalities? Or is it just that you expect the person who lists them to have read them to verify that they are on topic?

**Greg Charles** said:

> [@](#):
>
> The part of the article that seems strangest to me is that forensic handwriting experts … the ones that determine if two samples of writing came from the same person … have not been subjected to scientific rigor. Can that be possible? How do they set themselves up as experts if they don’t have to pass some kind of test?

There’s another thread on the topic in this forum that addresses this question. Essentially, there have been further tests to prove scientifically that human identification works, and the courts have now upheld forensic analysis as legitimate.

---

<div class="post-metadata">

**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\
**Post date:** [April 25, 2003, 7:38am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/103 "2003-04-25T07:38:09Z")

</div>

As for the question of whether psychology is a science: While it’s certainly in its infancy, it seems to me that psychology is more of a science, even, than is astronomy (and remember, I’m speaking as an astronomer here). While we understand much more about astronomy than about psychology, psychologists adhere much more closely to the standard picture of “The Scientific Method”. Astronomers, unfortunately, can only observe the situation as it is: We cannot peform real experiments on the subjects of our study. Psychologist can (and do), even if in a very limited way.

And there are well-accepted tests which measure qualities of a person which may as well be considered “personality”. If we’re worried about other factors confounding an identification test, then we might (in a later test) look for things like correlations between graphological assessment and a Briggs-Meyers personality test.

Finally, **Bill H.** , you mention that there are certain features which you know you’re likely to find in physicists’ handwriting. Out of pure selfish curiousity, would it be possible to elucidate what those might be?

---

<div class="post-metadata">

**Author:** ![Hilarity\_N.Suze](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/hilarity_n.suze/32/6223_2.png) [@Hilarity\_N.Suze](https://boards.straightdope.com/u/Hilarity_N.Suze)\
**Post date:** [April 25, 2003, 8:24am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/104 "2003-04-25T08:24:02Z")

</div>

> [@](#):
>
> \*Originally posted by Greg Charles \*  
> \*\*/…The part of the article that seems strangest to me is that forensic handwriting experts … the ones that determine if two samples of writing came from the same person … have not been subjected to scientific rigor. Can that be possible? How do they set themselves up as experts if they don’t have to pass some kind of test? \*\*

YES. Forensic graphologists have to have a specific course of training and documented skills. These classes are available at, for instance, state-supported universities. Of course these universities may also offer classes in phrenology and tarot reading.

But, forensic graphologists don’t go to court and say “There is a 98% certainty that you wrote that because somebody with XXX and YYY traits wrote it and you have XXX and YYY traits.” They say, ". . . you wrote that because it matches the sample you submitted on DD/MM/YY in the following points. . . " and after that it kind of reads like the results of fingerprinting.

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 25, 2003, 5:46pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/105 "2003-04-25T17:46:24Z")

</div>

**Jennifer McIlwee Myers** wrote

> [@](#):
>
> Either than or maybe you are just kinda naive about what the Internet actually has to offer us.
> 
> Contrary to popular belief, the Internet is not the repository of all human knowledge. It is the repostitory of two things: (1) whatever someone feels like putting on a web site and (2) lots and lots o’ p\*orn.

If you wish to discuss the viability of the Internet as a research vehicle (and it could be an interesting discussion), please start a thread elsewhere and I’ll be happy to contribute. I didn’t mandate that **Cecil** ’s cites be online, though I suspect most of his research is. That’s not an insult to him by the way. It’s my position that your position is incorrect, that there is plenty of valuable information on the Internet, and that in many key ways it’s a far better research device then a library. But like I say, please don’t hijack my thread to discuss that.

> [@](#):
>
> And I still feel it is on-topic here to discuss what level of evidence is likely to be actually effective. However, if someone actually _wants_ to discuss the stuff I brought up outside of the issue of graphology (not bloody likely, eh?), I’d be glad to start a new thread wherever it is most appropriate to do so.

It is absolutely off-topic. The topic is the research that **Cecil** did. It’s not on topic **where** he did the research, and it’s not on topic **whether the research will be accepted** , and it’s certainly not on topic **what color pants he was wearing when he reserached it**. You could literally claim that “effectiveness of evidence” is relevant to **every other** thread on this board and be equally off-topic. Now, effectiveness of evidence **is** an interesting topic, and one that I personally would be interested in discussing. But elsewhere. Please.

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 25, 2003, 6:13pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/106 "2003-04-25T18:13:28Z")

</div>

**Irishman** wrote

> [@](#):
>
> Bill H., I’m sorry you don’t feel like continuing. I was curious by your strong assertion that you can identify people to their handwriting samples by personality. I wonder if you’ve ever instituted any type of controls on your “party trick” version.

No, and that’s why I stated the OP. I’ve never seen a scientific study one way or the other. I’ve seen plenty of “studies”. But they were either testing something I didn’t think was reasonable to claim “like success of a salesperson based on handwriting”, or completely unscientific. (As you might imagine, the first set of studies were conducted by anti-graphologists, and the latter by pro-graphologists). I would enjoy analyzing an objective study, and **Cecil** has a bunch of them. Doesn’t seem like an unreasonable request.

> [@](#):
>
> I think I see your point about how nobody here reviewing them to see if they actually apply, but I’m still confused. You asked Cecil for a list of cites, now you’re really asking for an in depth analysis of 200 cites as to how they apply to the issue of using graphology to assess personalities? Or is it just that you expect the person who lists them to have read them to verify that they are on topic?

**Cecil** claimed that the cites were a) “objective,” b) “scientific,” c) analyzed “graphology as a predictor of personality,” and d) concluded it was “worthless.” That implies that he’s either analyzed them, or scanned them, or at least read or talked to someone else who analyzed them. As an example: Perhaps **Cecil** ’s research was reading an article in an established psychiatry magazine titled “Graphology is Junk!,” wherein the author ripped apart handwriting analysis as a predictor, gave several quotes from studies proving it, and then listed a bibliography of 200 cites. That would make **Cecil** ’s claim reasonable. Someone somewhere reviewed and listed 200 cites claiming this, and **Cecil** reviewed that someone’s review.  
But posting 200 cites backing an irrelevant cite makes them suspect at best. And posting 200 random cites is worse.

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 25, 2003, 6:22pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/107 "2003-04-25T18:22:15Z")

</div>

**Chronos** wrote

> [@](#):
>
> Bill H., you mention that there are certain features which you know you’re likely to find in physicists’ handwriting. Out of pure selfish curiousity, would it be possible to elucidate what those might be?

In this entire discussion of a test (neglecting the first post, but we weren’t discussing a test then), I’ve been conciously very careful to not mention any specifics of what I’ll test. Frankly, because I was worried that if I did, that someone(s) would send me faked tests with those things I mention altered. And I’m not certain a test won’t come out of this at some point, perhaps even a different test then originally proposed, so I’m hesitant to start now.

But… I can perhaps do better. I can mention a trait or two that I suspect (almost) every poster on this thread has, and likely most of the regulars in CCC or GQ has. (These are traits I’d expect a physicist would have, but also most smart, inquisitive, detail-oriented people as well). I can either tell you, or perhaps you’d like to formulate that claim into some sort of test. It would certainly take a lot less of my time (as I wouldn’t do the actual analysis), and wouldn’t require physical paper to be passed around, as the traits don’t involve pressure or margins.

---

<div class="post-metadata">

**Author:** ![mok](https://avatars.discourse-cdn.com/v4/letter/m/edb3f5/32.png) [@mok](https://boards.straightdope.com/u/mok)\
**Post date:** [April 25, 2003, 6:50pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/108 "2003-04-25T18:50:07Z")

</div>

All

Here is a situation that would be very interesting to see how Graphology handles:

I have a family member with Dysgraphia ( [look it up](http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=Dysgraphia) ). This disorder does not otherwise affect his personality, and his language skills are otherwise above average. He can type well.

The disorder does however have the effect that his handwriting is severely affected, is occasionally illegible, and is notably inconsistent.

**Do proponents of graphology make the claim that _his_ handwriting will be predictive of his personality traits?**

**If so** : Propose a mechanism explaining how his personality traits “make it through” his dysgraphia. For example, given that his handwriting is inconsistent, this “e” may not look like the next “e”. His writing may slant to the right this time, to the left next time. He may dot his “i” with a circle this time, and forget to dot it at all the next time. As I understand graphology, the claim is made that the letter forms, slant, idiosyncratic shapes, etc. are what are used to determine personality. If the forms, slant, shapes, etc. are never consistent – how could graphology tell anything about this person?

**If not** : Why not? Where then is the predictive value of graphology? Consider that given a random human being you otherwise know nothing about except for a sample of their handwriting. You have no way of knowing for example if they have a minor form of dysgraphia, or if they have injured their hand, or if they are visually impaired, or if they were raised in a foreign country, etc. All of these could obviously affect their handwriting, but are in no way predictive of their personality.

-mok

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 25, 2003, 9:02pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/109 "2003-04-25T21:02:43Z")

</div>

This proponent has no idea.

---

<div class="post-metadata">

**Author:** ![Jennifer\_McIlwee\_Myers](https://avatars.discourse-cdn.com/v4/letter/j/54ee81/32.png) [@Jennifer\_McIlwee\_Myers](https://boards.straightdope.com/u/Jennifer_McIlwee_Myers)\
**Post date:** [April 25, 2003, 9:21pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/110 "2003-04-25T21:21:46Z")

</div>

* * *

**mok** brings up an interesting issue. I have a young, healthy, highly intelligent friend whose handwriting looks like that of a person who is either ill and elderly or barely literate. This is due to a hereditary palsy. I wonder what a graphologist would make of her handwriting – and more important, whether graphologists are able to tell which handwriting is _not_ representative of the “whole person.”

* * *

**Bill H.** You seem to feel my discussion of how viable the Internet is as a way to find such things is completely and utterly unrelated to your original question. Actually, what I was thinking of was the following from your original post:

> [@](#):
>
> I’d like to know about those 200 objective scientific studies, and a web search brought up nothing. Cite, please.

I merely felt it was important (in the light of that quote) to point out that the lack of results from your web search does not in any way reflect on the number or quality of studies on the (in)effectiveness of graphology as a personality predictor.

The issue of belief perseverance is related to your request “Cite, please,” and your subsequent detailing of your interest in graphology. Given a basic knowledge of the nature of belief perserverance, one can hardly blame anyone, much less Mr. Adams, for not bothering to cough up cites. Your initial email makes it plain that you are heavily invested in your belief, and cites would make no difference. Thus, my discussion of belief perseverance was written as a very direct response to your original post as well.

HOWEVER, in both of these cases, as well as the below, I have to admit I was wrong in assuming the connection would seem obvious to everyone reading the thread. I should have used both quotes and explanations, and I’m sorry I failed to do so. Mea culpa, mea culpa, mea maxima culpa.  
**Bill H.** also wrote:

> [@](#):
>
> Now, effectiveness of evidence is an interesting topic, and one that I personally would be interested in discussing. But elsewhere. Please.

Hey, if you want evidence (as per your original request), I’m going to want to have a sense that you’ll respond in some meaningful way to evidence before I’m going to drag my nasty-head-cold-infested body down to the University library and hijack a librarian.

And, heck yes, I’m interested in going elsewhere to pursue the related topics I’ve brought up, but there hasn’t been any feedback as to precisely where it is appropriate to post them on this site. BION, I have no idea whether those issues belong in “General Questions,” “Great Debates,” MPSIMS, or ???. Feel free to say that this means I am incredibly thick or slow – I’ve been called worse – but I simply haven’t the foggiest idea where to start such a thread.

* * *

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 26, 2003, 12:50am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/111 "2003-04-26T00:50:37Z")

</div>

General Questions

---

<div class="post-metadata">

**Author:** ![Jennifer\_McIlwee\_Myers](https://avatars.discourse-cdn.com/v4/letter/j/54ee81/32.png) [@Jennifer\_McIlwee\_Myers](https://boards.straightdope.com/u/Jennifer_McIlwee_Myers)\
**Post date:** [April 26, 2003, 4:43pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/112 "2003-04-26T16:43:37Z")

</div>

Hello, again, **Bill H.**!

The brevity of your recent replies suggests you are short on time. I hope nonetheless that you will soon have enough time to be able to supply a bit more info on (1) your parameters for what graphology is successful in doing and (2) what you want to get out of this thread, if anything, if Mr. Adams is not available to post the now-legendary 200 cites.

I realize that you have gone a long way towards answering (1) in your original post and subsequent posts. The reasons I’m asking for clarification are two-fold: first, if we’re looking into research on graphology, I would prefer to pre-eliminate research that covers areas where you think graphology is not effective; second, I’m trying to figure out if the areas where you think graphology is effective are reasonably quantifiable for the purposes of the kind of studies we’d be looking for.

So I guess my ideal answer to question (1) would involve your idea of what specific qualities of graphology are truly testable.

BTW, I still suspect that some of the articles cited by **Duck Duck Goose** at the start of this might be quite applicable to the question of the validity of graphology, but again, finding out if this is the case would require one or more persons to go to visit the dead-tree repository.  
As for question (2), I’m just curious.

* * *

And now for something slighly different (but still obviously on topic, I promise, **Bill**!):

An interesting test of graphology would be to get decent photos of handwriting from people of famous or notorious characteristics, hand them unlabeled to the graphologist, and let the fun begin. If, for example, the best they could come up with for Hitler’s writing was, “Sometimes you feel more confident than others,” you’d think that maybe their ability to detect traits was minimal at best. Just a suggestion. And yes, **Bill** , I know you are unlikely to find such a study do-able under these circumstances, but thought I’d toss it off anyway.

[Note: This mention of Hitler fulfills the weekend requirement for mentioning Hitler and/or Nazi-ism during an Internet discussion. Any further mentions before Monday are superfluous.]

* * *

---

<div class="post-metadata">

**Author:** ![Bill\_H](https://avatars.discourse-cdn.com/v4/letter/b/a5b964/32.png) [@Bill\_H](https://boards.straightdope.com/u/Bill_H)\
**Post date:** [April 26, 2003, 6:00pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/113 "2003-04-26T18:00:12Z")

</div>

**Jennifer McIlwee Myers** wrote

> [@](#):
>
> I realize that you have gone a long way towards answering (1) in your original post and subsequent posts. (…) I’m trying to figure out if the areas where you think graphology is effective are reasonably quantifiable for the purposes of the kind of studies we’d be looking for. (…) So I guess my ideal answer to question (1) would involve your idea of what specific qualities of graphology are truly testable.

Honestly, I’ve talked so much in this thread, I feel I have nothing to add as far as what I’m looking for and what I think is testable. Please re-read my posts, including the very first one. I’m not sure how this could be in question. But if in fact I’ve not been clear, please make your question more specific.

> [@](#):
>
> BTW, I still suspect that some of the articles cited by Duck Duck Goose at the start of this might be quite applicable to the question of the validity of graphology, but again, finding out if this is the case would require one or more persons to go to visit the dead-tree repository.

I absolutely agree. But “might be” is not enough. To be frank, if in fact I did take on a serious mission of investigating this, I wouldn’t start with that list, as it’s off to a bad start, i.e. used to support a proposition which is not relevant to this discussion.

But more importantly, I’m not interested in undertaking a serious investigation. Well, I am interested, but time just ain’t available. So, when I read that **Cecil** in fact had made a serious investigation, I very much wanted to examine his research. And here we are.

> [@](#):
>
> An interesting test of graphology would be to get decent photos of handwriting from people of famous or notorious characteristics, hand them unlabeled to the graphologist, and let the fun begin.

In fact, there are tons of examples of this done on a non-scientific basis. Google on “celebrity graphology” and you’ll find very many. It probably wouldn’t surprise you to learn that most handwriting analysts are familiar with many celbrities handwriting, and Hitler in particular is one that 99% (including me) have seen and are familiar with.

Which makes the test you propose very difficult to do scientifically as you can’t give me Hitler’s handwriting “blind”; I’ll recognize it immediately. But if you’re interested from a curiousity perspective, you can find what other handwriting analysts see in various celebrities writing.

---

<div class="post-metadata">

**Author:** ![Jennifer\_McIlwee\_Myers](https://avatars.discourse-cdn.com/v4/letter/j/54ee81/32.png) [@Jennifer\_McIlwee\_Myers](https://boards.straightdope.com/u/Jennifer_McIlwee_Myers)\
**Post date:** [April 26, 2003, 7:13pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/114 "2003-04-26T19:13:09Z")

</div>

**Bill H.** wrote

> [@](#):
>
> Honestly, I’ve talked so much in this thread, I feel I have nothing to add as far as what I’m looking for and what I think is testable. Please re-read my posts, including the very first one. I’m not sure how this could be in question. But if in fact I’ve not been clear, please make your question more specific.

A more than reasonable request. What I am hoping for is one or more specifically testable statements about graphology, e.g. “Graphology can determine whether a person is creative in one of the fine arts or not.”

I’m looking for something testable because, as has been mentioned, a lot of traits that are supposed to show up in handwriting would have to be much better defined to be testable. Traits like “tends to think more quickly than average” or “tends to be more passionate than average” just aren’t useful in terms of experimental testing, especially if the person tested is the main source of confirmation.  
**Bill H.** again:

> [@](#):
>
> But more importantly, I’m not interested in undertaking a serious investigation. Well, I am interested, but time just ain’t available. So, when I read that Cecil in fact had made a serious investigation, I very much wanted to examine his research. And here we are.

Mr. Adams has already stated the results of his research, and you disagree.

I ask again, as I really, truly am interested in knowing: what do you want from this thread (if anything) now that it seems so unlikely that Mr. Adams will show up with the cites? Would you prefer no discussion at all unless it is Mr. A or his representatives posting cites? Do you want to discuss the possible validity of graphology? After all, if you wanted to just walk away and ignore the thread except to check once in a while for the presence of cites, that’s a valid response.

And finally, **Bill H.** mentions:

> [@](#):
>
> Which makes the test you propose very difficult to do scientifically as you can’t give me Hitler’s handwriting “blind”; I’ll recognize it immediately.

Well, I didn’t say it would be an easy test to do, I said it would be interesting. 😉

---

<div class="post-metadata">

**Author:** ![Irishman](https://avatars.discourse-cdn.com/v4/letter/i/b487fb/32.png) [@Irishman](https://boards.straightdope.com/u/Irishman)\
**Post date:** [April 27, 2003, 10:37am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/115 "2003-04-27T10:37:00Z")

</div>

You could put the effectiveness of evidence discussion in Great Debates. It could go in IMHO. Depends on how you word the question.

People hold onto their beliefs despite any evidence. - that’s GD

How open to evidence are your beliefs? - that’s IMHO

Nobody Frickin’ Cares What the Evidence Says! - that’s the Pit.

So really you have to consider just what your intent for the thread is. You can either formulate the question and then ask in ATMB which forum it should go, or post to General Questions and then the moderators will move it and chasten you for putting it in the wrong forum. 😉

**Bill H.** , now I see what you’re saying about the relevancy of the cites. Because the argument they were being used to back was not related, you don’t know if they cites themselves are related or not. You’d rather not wade through 20 unrelated cites that happen to talk about graphology, especially if it means going to the library and expending extensive work to even get them in hand.

---

<div class="post-metadata">

**Author:** ![refusal](https://avatars.discourse-cdn.com/v4/letter/r/ac91a4/32.png) [@refusal](https://boards.straightdope.com/u/refusal)\
**Post date:** [April 28, 2003, 2:21pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/116 "2003-04-28T14:21:07Z")

</div>

I’m interested in how quickly Bill H can reject all the cites given. A number were provided by Duck Duck Goose which show that graphology cannot predict those elements of personality which affect performance at work. So what elements of personality can it predict? Does Bill H have a separate personality for work and one for leisure?

The studies evaluating graphology as a measure of performance in the workplace seem an excellent example of trying to construct objective studies on the validity of graphology. It is easier and more objective to measure the performance of a salesman than to judge, for example, if someone is creative or has other “personality traits” (since we don’t even have a definition of personality traits).

Even if they are not directly related to Bill H’s claims (which seem to vary somewhat), the cites provide evidence that graphology is not effective in certain areas. It is impossible to prove that in every circumstance graphology is worthless, but studies such as the ones referred to must be attacks on the value of graphology in other circumstances.

---

<div class="post-metadata">

**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\
**Post date:** [April 28, 2003, 3:18pm UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/117 "2003-04-28T15:18:21Z")

</div>

**Bill H.** , you haven’t been inspired enough, it seems, to look up the non-online refs supplied by **DDG,** and I’ll admit that could be a lot of legwork. So here is one, 515-page, comprehensive book on the subject, available thru Amazon, deliverable to your front door. Although the list price is US$30, Amazon will sell it used for US$4.89. If that fits your budget, I highly recommend it for your reading pleasure:

[Beyerstein, “The Write Stuff - Evaluations of Graphology”](http://www.amazon.com/exec/obidos/tg/detail/-/0879756136/qid=1051542557/sr=1-1/ref=sr_1_1/002-1605467-4184867?v=glance&s=books)

It may also be available in your local library. It is at mine. [sup][sub]But perhaps that’s because I donated a copy.[/sub][/sup] 🙂

---

<div class="post-metadata">

**Author:** ![Jennifer\_McIlwee\_Myers](https://avatars.discourse-cdn.com/v4/letter/j/54ee81/32.png) [@Jennifer\_McIlwee\_Myers](https://boards.straightdope.com/u/Jennifer_McIlwee_Myers)\
**Post date:** [April 29, 2003, 4:21am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/118 "2003-04-29T04:21:54Z")

</div>

**zut** stated:

> [@](#):
>
> Additionally, I still disagree with you about the “control” test. If Bill H. performs, it’s based on some level of skill (possibly, but not necessarily, in personality analysis), and that can’t be quantified, so the proposed control really is not a control. Consider this rather silly analogy: Bill H. claims to have the ability to, more often than not, sink a basket from the free throw line. Being skeptical, we arrange to test this. As a control, we pick a random high school kid to undertake the same test. How is this an adequate control over a relatively short-lived  
> test? If hitting a free throw really is a skill (as opposed to just random luck), we don’t know if the kid has some moderate skill, or is pretty coordinated and a quick learner. We don’t know how difficult this skill may be to pick up. We don’t know how much knowledge the kid may get by observing Bill H. warm up. We don’t know how making a basket is accomplished. So, when all’s said and done, what does the addition of the control add to the test?

The “free throw” test analogy is seriously flawed. The reason that the analogy seems to work on the surface is that we all have a rough idea of that, “Whoo, golly, it sure is tuff to sink them free-throws!” In other words, we have (or think we have) a working knowledge of how difficult the feat actually is.

Of course the lone high-school kid would be useless – one would absolutely need _multiple_ controls to determine the statistical likelihood of making, say, 70% of 20 free throws. Mind you, given the availability of sporting statistics, it could be possible to put together statistics on human free throw abilities pretty easily.

Of course, if one was testing a specific new free-throw technique, one would never allow the controls to see the main testee perform.

If we were running the proposed test on graphology, the point of using multiple (_not_ just one) non-graphologist controls would be to determine whether the graphologist is doing something unusual or not. In other words, we can’t just _imagine_ how likely it is for a person to match 10 handwriting samples to ten people s/he has been allowed to interview; nor can a statistician come up with reasonable likelihoods without any data (unless all you want is to test against random chance).

Since we don’t know that matching handwriting requires special skills, and in fact one of the things we are testing for is whether graphologists have special skills, it would be necessary to have non-graphologists attempt the same feat in order to determine the normal rate of success for an untrained human.

Interestingly, I found a website that refers to studies in which tests of graphologists included non-graphologist controls, and found that if the handwriting sample included content created by the individual writer, non-graphologists did better. Unfortunately, the citation for the study is a link that now leads to a dead web page, so I have to find where the darn study was moved to.

* * *

Separately, in response to the continuing discussion of the value or lack thereof in the cites given early on by **Duck Duck Goose** : I would say the original article with regards to “The use of graphology as a tool for employee hiring and evaluation” has some value outside of it’s seemingly narrow topic. This is due to the fact that it discusses research dealing the ways in which graphology can seem to work without doing so – for one, the “fallacy of personal validation.”

If one isn’t able to identify and rule out those things most likely to lead to false positives, how can one tell what the heck kinda results one is actually getting? For anyone to rate their own success as a graphologist (or at using other forms of personality analysis), they would need to have a working knowledge of the phenomenon of personal validation. The fact is that it is possible to come up with seemingly highly individualized graphological analyses that get strong positive feedback from the people who are presented with them as “your own personality analysis.” This is not necessarily the case with every graphologist, but is absolutely germaine to the topic of how well graphology works in general.

Yet another interesting test (and perhaps a better one) would be to have a graphologist volunteer from the board (there must be someone other than **Bill H.** around) who would be willing to do personality analyses for a set of uncommitted subjects (persons who are willing to state that they have no ax to grind). Half the subjects get their own analyses, half get analyses randomly chosen from those given. Then the ratings the subjects give to their analyses would be compared.

* * *

---

<div class="post-metadata">

**Author:** ![Hilarity\_N.Suze](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/hilarity_n.suze/32/6223_2.png) [@Hilarity\_N.Suze](https://boards.straightdope.com/u/Hilarity_N.Suze)\
**Post date:** [April 29, 2003, 6:10am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/119 "2003-04-29T06:10:48Z")

</div>

Back to the question of the 200 sources used by Cecil. Duck Duck Goose provided these and a link to an article that soundly refuted the use of graphology not only as a hiring tool (which was the main focus of the article) but as a science in general, categorizing it as a “pseudoscience.” While I have not researched any of these cites, I did read the whole article.

But some of these cites are automatically suspect because many of them are generated from one organization, the British Columbia skeptics association, which is dedicated to debunking. In other words, I’d hesitate about calling these “objective.” It looks to me like the research was slanted against graphology as a science. “Objective” to me means no axe to grind, no preconceived notions, and strict application of the scientific method (which, as a happy convert to liberal arts early in my college days as a consequence of a course called ‘Psychological Tests and MEasurement,’ I know almost nothing about).

Looking at these 23 sources, it looks to me like the writer of the article in question was not particularly objective, not that anybody is, really. My criteria: If you were an objective person just looking for the truth of the matter and merely seeing the title of these, would it look like they would shed light? Anyway, as an objective person here’s what I’d think of these sources, _without having read them_:

> [@](#):
>
> ALLPORT, Gordon & P.E. VERNON (1933): Studies in Expressive Movement, New York, Macmillan.

Okay, sounds objective enough. That’s **ONE**.

> [@](#):
>
> BEYERSTEIN, B. (1987): “Graphology: Write or Wrong?”, the Rational Enquirer (newsletter of the BC Skeptics), Vol. 1, No. 1 (June), p. 1.  
> BEYERSTEIN, D. (1988) “B.C. Skeptics Cooperate with BCCLA to End Graphology”, the Rational Enquirer (newsletter of the BC Skeptics), Vol. 1, No. 3 (March), pp. 1-2.  
> BCCLA (1987): Psychological Testing and Privacy. Available from the Association.  
> B.C. SKEPTICS (1988): “Graphology Revisited”, the Rational Enquirer (newsletter of the BC Skeptics), Vol. 1, No. 3, (January), p. 6.

These I’d wonder about. I figure the BC skeptics have a certain agenda. In the interests of fairness and objectivity I would certainly accept ONE such article, but here are five. So, accepting one of the above–don’t know which one–that’s **TWO**.

> [@](#):
>
> BEN-SHAKHAR, Gershon, M. BAR-HILLEL, Y. BILU, E. BEN-ABBA & A. FLUG (1986): “Can Graphology Predict Occupational Success? Two Empirical Studies and Some Methodological Ruminations”, Journal of Applied Psychology, Vol. 71, pp. 645-53.

Well, in the course of finding out whether graphology can predict occupational success it’s realistic to first find out whether graphology can, in fact, predict anything. So this one seems pertinent. **THREE**

> [@](#):
>
> CROWN, D. (1987): “A Review of Scientific Aspects of Graphology: A Handbook”, Journal of Forensic Sciences, Vol II, No. 1, pp. 287-8.

this sounds both relevant and objective. **FOUR**

> [@](#):
>
> CRUMBAUGH, J.C. & E. STOCKHOLM (1977): “Validation of Graphoanalysis by global or holistic method”, Perceptual and Motor Skills, Vol. 44, pp. 403-10.

I don’t know what they mean by global or holistic method, but with the word “validation” in there it seems like it could be pertinent. **FIVE**.

> [@](#):
>
> FESTINGER, L. REIKEN, H.W. & SCHACTER, S. (1956): When Prophesy Fails, Minneapolis, University of Minnesota Press.

Now here’s a suspect title. If you were doing an objective study about graphology, why would this one come up? Doesn’t mention graphology. Maybe something in the index?

> [@](#):
>
> FORER, B.R. (1949): “The Fallacy of Personal Validation: A Classroom Demonstration of Gullibility”, Journal of Abnormal Psychology, Vol. 44, pp. 118-23.

I don’t question this source, but it doesn’t seem like this issue would come up in an objective study, either. At least not on the first pass.

> [@](#):
>
> GOLDBERG, L. (1987): in Nevo, op. cit.

Not a separate source.

> [@](#):
>
> HYMAN, Ray (1981a): “The Psychic Reading”, Annals of the New York Academy of Sciences, Vol. 364, pp. 169-81.

Wouldn’t count this one at all–who said anything about psychic?

> [@](#):
>
> HYMAN, Ray (1981b): “Cold Reading: How to Convince Strangers that You Know All About Them”, in Kendrick Frazier (ed): Paranormal Borderlands of Science, Buffalo, NY, Prometheus Books.

Good grief! Objective? No way.

> [@](#):
>
> JANSEN, A. (1973): Validation of Graphological Judgments: An Experimental Study, The Hague, Mouten.

Okay, sounds promising. This is, what, **SIX**?

> [@](#):
>
> KARNES, Edward (1986): “Graphoanalytic and Psychometric Personality Profiles: Validity and Barnum Effects”, in press.

I’ll allow this one even with the addition of “Barnum Effects” because I’m feeling expansive. **SEVEN**.

> [@](#):
>
> KLIMOSKY, Richard & Anat RAFAELI (1983): “Inferring Personal Qualities Through Handwriting Analysis”, Journal of Occupational Psychology, Vol. 56, pp. 191-202.

The fact that this one’s in the J. of Occ. Psych. kinda makes me wonder.

> [@](#):
>
> MARKS, David & Richard KAMMANN (1980): The Psychology of the Psychic, Buffalo, Prometheus Books.

Gimme a break. More psychic stuff.

> [@](#):
>
> MOORE, Michael (1985): “About the Sad State of Scientific Graphology”, Psychological Documents, Vol. 15, No. 2 (MS #2676).

Okay, I’d take a look at this one. **EIGHT**.

> [@](#):
>
> NEVO, B. (ed.) (1986): Scientific Aspects of Graphology A Handbook, Springfield, Ill., Charles C. Thomas.

Definitely. **NINE**.

> [@](#):
>
> PATTERSON, J. (1976): Interpreting Handwriting, New York, McKay Press.

**TEN**

> [@](#):
>
> RAFAELI, A. & R.J. KLIMOSKI (1983): “Predicting Sales Through Handwriting Analysis”, Journal of Applied Psychology, Vol. 68, pp. 212-17.

I want to give this one a half, because while it does mention handwriting analysis, it also throws in the bit about predicting sales. So, I just decided–no halves.

> [@](#):
>
> SNYDER, C.R., R.J. SHENEKL & C.R. LOWERY (1977): “Acceptance of Personality Intepretations: The ’Barnum Effect’ and Beyond”, Journal of Consulting and Clinical Psychology, Vol. 45, pp. 118-23.

Anything with “the Barnum effect” in it reeks of non-objectivity and I’ve already allowed one.

> [@](#):
>
> SUNDBERG, N.D. (1955): “The Acceptability of ’Fake’ versus ’Bona Fide’ Personality Test Interpretations”, Journal of Abnormal Social Psychology, Vol. 50, pp. 145-7.

This looks like it was cited just because it debunked handwriting analysis.

SO, out of 23 cites, I’ve selected only TEN that look reasonably objective based solely on their titles. This is a far cry from 200 objective sources. But the article has given me an idea of what an objective study of handwriting analysis might look like.

---

<div class="post-metadata">

**Author:** ![Hilarity\_N.Suze](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/hilarity_n.suze/32/6223_2.png) [@Hilarity\_N.Suze](https://boards.straightdope.com/u/Hilarity_N.Suze)\
**Post date:** [April 29, 2003, 6:24am UTC](https://boards.straightdope.com/t/validity-of-graphology/169710/120 "2003-04-29T06:24:12Z")

</div>

Okay, for those of you who made it through my last post, here’s my idea of how to do an objective test.  
Let Bill H. or the next volunteer who knows graphology select four or five traits that can be reliably predicted from handwriting. These should be things that an individual can easily assess in himself in terms of being high or low. For example, well organized/ poorly organized; highly social/antisocial, accepting/rebellious  
Each test participant would write the same thing (the first couple of paragraphs of Cecil’s response to the question, for instance), using their choice of non-personalized paper and their preferred writing instrument, for the handwriting analysis. On a separate sheet they would rate themselves on the characteristics chosen–maybe with three gradations? (high, moderate, low) They would mail both sheets to Volunteer #1, who would assign some kind of a number sequence to each one and put that number sequence on each sheet of paper. The handwriting samples only would be mailed to Volunteer #2, the person with skill in handwriting analysis. On a preselected day, Volunteers #1 and #2 would post their findings here, so all could see if Sample 342108, self-assessed as highly organized, moderately antisocial, and intensely rebellious, was assessed in the same way by the analyst.  
Is this simple, clear, and reasonable, or what? (But still–a lot of work.)  
Maybe one of the traits should be honesty–if that’s something that can be culled from a handwriting sample.

[Previous page](https://boards.straightdope.com/t/validity-of-graphology/169710.md?page=5)

[Next page](https://boards.straightdope.com/t/validity-of-graphology/169710.md?page=7)
