I did browse one that someone gave me a few years back. Seemed utterly flakey.
But, I suppose these could be failed experiments (we are betting on the null hypothesis, when we conclude graphology doesn't work). What if in theory, handwriting anlalyses did predict better than chance one's personality, but the graphologists-- not being scientifically trained-- just don't know how to validate the idea.
This whole issue is motivating me to collect data on the topic, but doing it by "bootstrapping".
Of course this assumes no one's done the experiment before, but would you be willing to bet one million dollars this experiment would fail:
1) Give lots of subjects the NEO (a perfectly valid measure of the big 5 personality traits-- neuroticism, openness, conscientiousness, agreeableness and extraversion).
2) Collect a paragraph (or however long writing sample, on whatever topic graphologists typically use).
3) Code each person's writing samples on objective, reliable things that might vary across people (e.g., mean letter size, mean spacing between letters, degree of slant from zero, proportion of letters that are left open, etc)
4) Conduct a regression analyses on the handwriting variables to see if they predict scores on the five personality factors.
Would it be paranormal / radical / paradigm shifting to test hypotheses like:
highly conscientious people might have some writing characteristics that differ significantly from people low in conscientiousness?
Or that, the highly neurotic people might differ in how they write compared with totally non-neurotics.
These seem like perfectly reasonable hypothesis to me-- certainly not earth shattering like finding people can bend spoons with their mind, or that magnets cure AIDS.
In fact, the two specific hypotheses above seem so reasonable to me, that unless someone's done it before, I'm going to try this fall semester. It'd be real easy to get the data.
I think the difference here is I would be using dustbowl empiricism (the power of regression) to test (versus predefine) what if anything about peoples' writing styles varies with their personalities.
So, the graphologist says "oooo, look at the big fat O he wrote, that means he is a very openminded person.
The above experiement would go in the reverse direction:
The regression shows that conscientious people use more (or less) distance between letters (or even the variability in distance from letter to letter is lower) than do non-C people. Hmmm, how could we explain such a result without redefining physics chemistry and biology (note sarcasm).
The Woo in graphology-- potentially-- comes from making the rules up as you go. The regression equation would go in the reverse direction: given that people do have different writing styles, does their personality explain any of that variance.
Just in 5 minutes of thinking, I've generated (at least imo) two hypothesis that seem reasonable; non controversial even. So much so that if the experiment worked, and I applied for the prize, Randi would not test me.
All assuming it works / hasn't been tried before, but even though I teach MBA students that graphology is utterly worthless and should never be used (for employment decisions), I'd be reluctant to bet 1 million that the above would fail.
At any rate, graphology is not a paranormal claim.
Jekyll said:
Have you read any graphology texts?
A lot of what they use to determine the characteristics of the individual are artifacts of schools of handwritting.
The most obvious example I can think of is that forward slanted handwritting implies a driven and forward thinking individual. From the ages of 10 to 12 I and my classmates of the time were taught italic handwritting. Unsuprisingly this did nothing to improve our motivation.
Similarly how you dot your i s and cross your t s is a product of huge range of things from how you hold a pen, to the quality of your hand eye co-ordination to the work of others you have been influenced by.
Graphology isn't about detecting parkinson's by looking for hand tremors. It's a claim that almost all personality traits manifest themselves detectably through handwriting.
I think graphology is directly comparible to phrenology, in both its claims and underlying methodology. Whilest my big nose doesnt mean that I lie all the time, it wouldnt break the laws of physics if truth telling and nose size were linked through brain function in some way.
On the other hand, lie detectors can be compaired to identifying a suspect entirely by eye colour. There is clearly understood mechanism, it does work to a limited degree and returns better results than just guessing. I'd just rather not rely on it.
.
