The belief that handwriting is a sign of the inner personality is very old. The first serious attempt to analyze handwriting seems to have been that of Camillo Baldi, an Italian scholar, who published a book on the subject in 1622. As literacy spread, handwriting analysis became popular, being practiced as an art form by such literary figures as Goethe, Poe, the Brownings and Dickens. Jean Hippolyte Michon coined the term "graphology" in 1875. Michon systematized handwriting analysis by associating hundreds of graphic signs with specific personality traits.
Around the turn of the century, the French psychologist Alfred Binet performed several experiments with handwriting analysis as a device for testing personality. Binet claimed that handwriting experts could distinguish successful from unsuccessful persons with high accuracy. The German school of handwriting analysis, led by Ludwig Klages, developed a subjective and esoteric approach to graphology, and apparently never even attempted experimental verification of its claims.
There is today no single theory or method that dominates graphology. The French school concentrated on isolated signs as specific indicators of personality, and the Germans sought to make subjective interpretations based on a total impression of a person's handwriting. In 1929 M. N. Bunker founded "graphoanalysis" as a compromise between these two extreme positions The language and techniques of graphoanalysis seem to be more or less the common graphological practice in the United States today. (Bunker 1971). Bunker founded the International Graphoanalysis Society, which now offers an 18-month correspondence course for analysts. The society is based in Chicago and claims 10,000 active members. The Institute of Graphological Science in Dallas also offers courses and accreditation in graphology, but it is not affiliated with the Graphoanalysis Society. Unless indicated otherwise, I will use the term "graphologist" to refer to a practitioner of any school of handwriting analysis for personality assessment.
Even though there is no canonical school of graphology, some discussion of the practice may be helpful. Remember that the basic assumption underlying graphology is that handwriting is an expression of the personality; hence, a systematic analysis of the way a person forms words and letters will reveal traits of personality. The graphologists are fond of repeating, "Handwriting is brain writing."
Graphologists look for such features as the slant of characters, the size of individual letters, angularity and curvature, and such non-graphic features as the pressure of upward and downward strokes. In most systems, the slant of the letters is very important. A right slant generally correlates with extroversion, and a left slant with introversion. The shape of the letter "t" seems important to all systems. Bunker's book contains a dictionary of specific signs and their correlates with personality, such as "pride: tall d-stems, t-stems not vertical." (Bunker 1971). Another system (Rosen 1965) defines sixteen factors, including graphic factors such as slant, spacing and letter size, as well as global features such as "rhythm" and "tempo." Sheila Kurtz uses handwriting features such as slant, pressure and t-formation to create a subject's "graphoprofile," which reveals, among other things, his thinking pattern, goal orientation, fear traits, defenses, integrity traits and social traits. (Kurtz and Marilyn 1983). Graphologists prefer handwriting samples that are spontaneous, and not given for the express purpose of being analyzed. They prefer a text of some length, written with a tool sensitive to pressure and speed. Graphologists want a text with biographical material if possible, and they want to know the age and sex of the writer. The output of the analysis is a free-form personality description, perhaps systematized in some way (Kurtz, et al. 1983).
The claims of graphology are bold: "The analyst can, with surprising accuracy, predict what the writer will do and how he will react under certain conditions" (Bunker 1971). "Analysis of a handwriting specimen becomes a behavioral indicator and a remarkably accurate gauge of the writer's character. Graphology can even "...foretell a person's future actions and future intellectual performance" (Rosen 1965). "By focusing attention on [a child's] drawings, you will discover the attitudes and reactions responsible for shaping his personality" (Solomon 1978). Some even claim that changing one's handwriting can change one's personality (Kurtz 1983).
The references cited are only a few of the many books by graphologists. The Internet bookstore, Amazon Books, offers 170 works on the subject. Titles include: Change Your Handwriting, Change Your Life, by Vimala Rodgers; Lovescript: What Handwriting Analysis Reveals About Love and Romance, by Michael Watts; Write What's Wrong: Using Handwriting to Reveal Hidden Health Problems, by Claude Santoy. These titles certainly suggest graphology is a powerful tool for understanding and even changing personality.
Like everyone else nowadays, graphologists offer their services on the Internet. Mark Hopper of Handwriting Research Corporation in Phoenix claims to have developed "...the world's first and only scientifically validated handwriting system." (Hopper 1998). Hooper did not respond to my request for information supporting this claim. Sheila Lowe offers the Handwriting Analyzer software, which she claims is "...the world's most advanced handwriting analyzer; ideal for business." Lowe claims the reports prepared by the program can tell a person's strengths and weaknesses "in certain areas" as well as point out characteristics indicating a potential problem for an employer. (Lowe 1998). One of the handwriting organizations has a home page with information and pointers to graphological material on the Internet. (American Society of Professional Graphologists 1998).
Is graphology really a valid method of assessing personality? Psychological tests that attempt to judge personality, as opposed to ability, are notoriously difficult to validate. The difficulty would seem even greater when we are talking about such complex traits as honesty or integrity. Psychologists freely admit that they do not have any paper-and-pencil tests that give even remotely reliable information about these traits. It seems only fair to ask the graphologists to come up with some evidence to support their claims.
The evidence for graphological claims
Unfortunately for the graphologists, scientific research has found no clear correlation between handwriting behavior and basic personality patterns. Although some psychologists believe that the subject is worth further study, the existing research is almost uniformly inconclusive or negative.
The article by Bowman discusses the difficulties psychologists have in attempting to measure personality and predict behavior, and contrasts this with the glib claims of graphologists to do the same thing. (Bowman 1992). Bowman cites studies showing a person's handwriting is reasonably stable across time. This stability makes forensic analysis of handwriting possible. Indeed, some studies show good reliability in the degree to which a judge of handwriting makes similar judgments of the same handwriting sample at different times. The degree to which different judges make similar judgments of script characteristics has also been studied, and here, researchers have found lower reliability. However, reliability alone is not enough to make and assessment procedure useful; the technique must also be valid. That is, it must measure what it purports to measure. We generally measure validity by showing there is a consistent relationship between test scores and some independent measure of the trait in question. Scientists have studied the validity of handwriting analysis, and the results are not encouraging for graphologists.
For example, one study examined the ability of graphologists to predict sales success, in the sense that graphologists' evaluations would concur with supervisor's ratings. The researchers found no evidence of validity (Rafaeli and Klimoski 1983). In a later survey of research on handwriting analysis for personnel selection, Klimoski discusses the standards for evaluating personnel selection methods, and concludes the "credible, empirical evidence" does not support the claims of graphology as applied to personnel selection. (Klimoski 1992). A study for the Netherlands Society of Industrial Psychology considered 2250 judgments by graphologists and 6000 by a control group. The Dutch researchers found that graphological judgments, for large numbers, were slightly more often correct than incorrect, but they concluded that for judging the individual, "...graphology is a diagnostic method of highly questionable and in all probability minimal, practical value." (Jansen 1973). A 1989 analysis of 17 earlier studies of the validity of graphology as a personnel selection device found that even the minimal correlations sometimes obtained by graphologists could be accounted for by content-laden scripts. (Neter and Ben-Shakhar 1989). Summaries of the research literature can be found in Beyerstein, ed. (1992); Nevo ed. (1986); and Hines (1988). None of the many graphology books I examined cited any validating research. The attitude of the practicing graphologists seems to be represented by this quote: "Moreover, graphology is a self-validating science. This means that the credibility of the method can be proven immediately by the feedback you get from the graphologist." (Kurtz 1983). This comment gives us a hint why graphology seems work for its practitioners and their clients.
Persons who have some experience with graphological analysis are usually positively impressed. That is, they personally validate what the graphologist is telling them about themselves (psychologists sometimes express this idea by saying a technique has "face validity"). Personal validation gives a powerful subjective impression of truth, but unfortunately, the impression can be created by methods that have nothing to do with objective validity. A character reading containing statements are vague or general enough to be true about almost anyone tends to be accepted as unique by the client. If the reader has real information about the client from other sources, as graphologists usually do, he can tailor his stock reading to fit more closely the client's case, thus increasing the subjective feeling of validity. A handwriting sample with autobiographical material is obviously helpful to the analyst, as is the age and sex of the subject, or the job he is applying for. Hyman describes the process of character reading in more detail (Hyman 1976). A corporate user of graphological judgments may be in no better position to evaluate their validity. How does the employer really know that the persons rejected on graphological advice were likely to be incompetent or dishonest? The personnel officer may get a sense of personal validation simply because the graphologist can prepare a character sketch that sounds detailed and credible. Beyerstein relates the psychological process of personal validation to pre-scientific beliefs of sympathetic magic and finds parallels with graphological interpretation. (Beyerstein 1992).
In summary, then, it seems that graphology as currently practiced is a typical pseudoscience and has no place in character assessment or employment practice. There is no good scientific evidence to justify its use, and the graphologists do not seem about to come up with any. Next, we look into the use, if any, of graphology by employers in the north Texas area, and consider some of the legal issues raised by the use of pseudoscientific employment tests or courtroom testimony.
In late 1988 I undertook to estimate the extent to which graphology is used in personnel practice in the Dallas-Fort Worth area. I contacted local graphologists, employment agencies and personnel consultants, a firm of industrial psychologists, and a selection of large corporate employers. The latter included the airlines, real estate, electronics, security service, and stock brokerage. I also called the personnel offices of the cities of Dallas and Fort Worth and the counselling and testing office at the University of Texas at Arlington. The uniform response was "we don't use it and we don't know anyone who does." In 1998 I again attempted to find large companies using graphology in personnel management, and again, without success. Apparently, press claims about the wide use of graphology in business are exaggerated, at least in this area.
One of the graphologists I spoke with was the late Dr. Mary Lynn Bryden, head of The Institute of Graphological Science in Dallas (TIGS). Dr. Bryden declined to give me the names of her clients, save one, a real estate agency. That agency has recently changed ownership, and the current personnel director told me that she was unaware of any use of graphology. From my conversations with Dr. Bryden and a local graphoanalytic practitioner, I gathered the impression that most users of graphological services are individuals or small businesses. Obviously, someone is using it, or the several local graphologists would not be in business.
Graphological testing and the Civil Rights Act
There is a good reason why large corporate employers and local governments do not use graphology: the use of unvalidated pre-employment tests is unlawful if the tests have any discriminatory impact. A full review of employment discrimination law is beyond the scope of this article. I will concentrate on Title VII of the Civil Rights Act of 1964, the primary federal law relating to employment discrimination. 42 U.S.C. § 2000e-2(k)(1)(A)(i). Title VII applies to most employers, employment agencies and unions. It prohibits discrimination on the basis of race, color, religion, sex, national origin, or reprisal. If a test is not job-related and in practice excludes a protected group, then the test, whether or not designed or intended to discriminate, was indeed "used" to discriminate. The U.S. Supreme Court so interpreted the Act in Griggs v. Duke Power Co., 401 U.S. 424 (1971). In 1975, the Supreme Court held that an employment test that excludes one group may be used only if it is demonstrated to be job related through "professionally acceptable methods." Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975).
The Equal Employment Opportunity Commission (EEOC), jointly with the Department of Labor, the Office of Personnel Management, and the Department of Justice, has issued the Uniform Guidelines on Employee Selection Procedures, embodied in federal regulations at 29 C.F.R. Part 1607. The EEOC guidelines are based on the standards of the American Psychological Association (APA) and require "criteria-related validity." (APA 1985). That is, validity must be demonstrated by comparing test scores with some external variable that provides a direct measure of the characteristic in question. Put another way, test scores must correlate with performance on the job, and this must be established in a statistically proper way.
Graphological assessments are usually free-form personality descriptions rather than scored tests, but this does not exempt them from Title VII's coverage. The APA standards require that when an interviewer uses any kind of test to aid his assessment, the resulting assessment should be validated like other psychometric measures.
If graphology is a pseudoscience, as seems highly probable, then graphological evaluations are nothing more than character readings combined with a large chance element. The biases of the graphologist will consciously or unconsciously skew the results. If the result is an assessment that results in discrimination on one of the grounds recognized by Title VII, then users of graphological tests bear a heavy burden of showing job-related validity for them. Spohn critically examines the legal issues raised by the use of graphology in personnel practice (Spohn 1997). She also discusses other legal concerns arising the use of graphological tests, such as violation of the Americans With Disability Act (42 U.S.C. §§ 12101-12213), defamation, and invasion of privacy. Since the current state of scientific knowledge does not support the validity of graphological judgments, employers using such assessments are in a very risky legal position.
Graphology in court
Most graphologists boast of providing evidence as expert witnesses in court. I was unable to locate a single reported case in which graphological testimony for the purpose of personality assessment has been admitted in evidence. A few courts have considered whether or not a handwriting analyst can give testimony about a person's physical or mental condition, and all of them have held such testimony inadmissible. See, Warren v. Hartnett, 561 S.W.2d 860, 863 (Tex. Civ. App., Dallas, 1977 — ref. n.r.e.) (court rejects handwriting analyst's opinion on decedent's mental condition: "...we are aware of no recognized field of scientific inquiry which permits divination of mental capacity by persons whose expertise is limited to handwriting analysis"); Sosa v. State, 841 S.W.2d 912, 916-17 (Tex. App.—Houston [1st Dist.] 1992, no pet.) (proponent of graphoanalyst' character assessment failed to prove that the underlying scientific theory was valid; that the technique applying the theory was valid; or that the technique was properly applied on the occasion in question); Cameron v. Knapp, 520 N.Y.S.2d 917 (Sup. Ct. N.Y. County 1987) (court holds testimony of handwriting analyst not admissible in medical malpractice action to to show doctor's mental and physical condition; proponent could not show such procedures were generally accepted as reliable in the scientific community); Daniels v. Cummins, 321 N.Y.S.2d 1009 (Sup. Ct. Westchester County 1971) (handwriting expert held competent to testify that signature on deed was genuine, but not competent to testify as graphologist that signature indicated grantor was not of sound mind. "In pronouncing such grautitous conclusions she [the analyst] leaped into the occult, esoteric, pseudo-scientific pursuit known as graphology, venturing far beyond the province of a handwriting expert"); State v. Davis, 742 P.2d 1356 (Ariz. App. 1987) (graphologist not allowed to testify in support of defendant's insanity defense; state presented evidence that "...graphology is not a science because its results are neither verifiable nor repeatable"); State v. Anderson, 379 N.W.2d 70 (Minn. 1985) (court finds graphology is accorded a low measure of scientific reliability in predicting character or state of mind and is not generally accepted in the scientific fields of psychology and psychiatry); People v. Hester, 237 N.E.2d 466 (Ill. 1968) (Trial court properly excluded testimony of handwriting analyst that victim's handwriting showed her fear at time she writing).
If graphological evidence is offered, it must meet the test for scientific evidence required in most jurisdictions. That is, there must be a general consensus in the appropriate scientific field that the theory behind the technique is sound, and that it has been reliably reduced to practice. This is the Frye rule, from Frye v. United States, 293 F. 1013 (D.C. Cir. 1923). All the cases just cited, except for Warren v. Hartnett, and Sosa v. State, relied on Frye. The Warren court simply found the handwriting analyst's testimony useless, since she had no training or experience in evaluating mental disorders, and had made no personal observations of the decedent which would qualify her to testify as a lay witness. The Sosa court relied on Kelly v. State, 824 S.W.2d 568, 572 (Tex. Crim. App. 1992). In Kelly, the Texas Court of Criminal Appeals held the Frye general-acceptance test was not, alone, sufficient for the admissibility of novel scientific evidence. Instead, the Sosa court held, the 1986 amendments to Rule 702 of the Texas Rules of Criminal Evidence require a court to find such evidence to also be valid and reliable.
The U.S. Supreme Court recently changed the rules for admisibility of scientific evidence in federal courts in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S 579 (1993). The Daubert test was adopted by the Texas Supreme Court in E.I. duPont de Nemours & Co. v. Robinson, 923 S.W.2d 549 (Tex. 1995). It appears graphological testimony would not be admissible under the Frye standard, and it is even less likely to be admissible if challenged under the Daubert/Robinson analysis. It could be argued that Kelly v. State, above, anticipated the Daubert/Robinson decisions.
The Supreme Court in Daubert rejected an "anything goes" approach to expert scientific testimony. Rule 104 of the Federal Rules of Evidence requires the trial judge to determine at the outset whether the expert is proposing to testify to (1) scientific knowledge that (2) will assist the trier of fact to "...understand or determine a fact in issue," as required by Rule 702 of the Federal Rules of Evidence. The Court wrote: "This entails a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue." Justice Blackmun set out some "general observations" to assist trial judges in determining what is reliable scientific knowledge:
Valid knowledge consists of facts, as well as ideas inferred from facts "on good grounds." 509 U.S. at 590. If an expert's conclusion is not supported by valid reasoning, it fails the knowledge test and must be excluded. Graphological rationales cannot pass the valid knowledge test when they rely on magical thinking or illogical associations. Also, graphological judgments are subjective interpretations, and there is no evidence these interpretations measure what they purport to measure.
Reliability means that valid reasoning results in conclusions sufficiently reliable to be helpful for legal purposes. If graphological predictions were validly shown to be right twenty per cent of the time, they would still not be reliable. This low level of certainty would be not be helpful to a jury.
Even if valid and reliable, an expert's testimony might still be irrelevant. If a graphologist could validly and reliably testify that a job applicant possessed certain personality traits, for example, the evidence would not be relevant if those traits were not logically relevant to job performance.
So far, there have been no reported cases of graphological testimony decided under the Daubert/Robinson line of cases. We can be confident such cases will hold the use of graphology for personality assessment does not produce valid, reliable, and relevant evidence.
Where do the graphologists' claims of courtroom expertise come from? Many graphologists today attempt to qualify themselves as questioned-document examiners. Forensic handwriting examination for the purpose of detecting altered or forged writings has long been recognized by the courts as admissible. It appears that graphologists are attempting to bootstrap themselves into respectability by combining forensic handwriting analysis with their questionable personality-assessment techniques. In some cases, testimony of handwriting experts who were also graphologists has been admitted, where the evidence went only to quesitions of document authenticity. See, Daniels v. Cummins, supra; Hooten v. State, 492 So.2d 948 (Miss. 1986) (strong dissent by three judges taking view witness' education as a graphoanalyst did not even qualify her to express opinions on comparison of handwriting samples); 1st Coppell Bank v. Smith, 742 S.W.2d 454 (Tex. App., Dallas, 1987 — no hist.) (witness claimed to be trained as a graphoanalyst and also as a questioned documents examiner). Other courts have held the qualifications of graphologists to be questioned document examiners insufficient. See, State v. Livanos, 725 P.2d 505 (Ariz. App. 1986) (court distinguishes between graphology as the study of handwriting to determine personality traits and forensic document analysis); People v. Tidwell, 706 P.2d 438 (Colo. App. 1985); Carroll v. State, 634 S.W.2d 99 (Ark. 1982) (witness had only taken graphoanalysis training course).
Educational certification for graphology
Some other legal aspects of graphology are worth mentioning. TIGS appears to have been the only graphology school licensed under the Texas Proprietary School Act (Tex. Education Code, Ch. 132). A call to the Texas Workforce Commission, which administers the Act, reveals TIGS is no longer licensed, although it is still in business. The Act requires applicants for a proprietary school certificate to show that the courses, curriculum and instruction are of such a quality, content and length as to reasonably achieve the stated objectives for which they are offered. Tex. Education Code, § 132.055(a). Apparently the Act's administrator is not required to go behind the course outline and judge the validity of what is being taught. He is only required to find that the school will adequately meet its teaching objectives. Officials in Austin confirmed this interpretation with me in a telephone conversation. I posed the question of how an application for a course in astrology would be handled. I was told (after some hesitation) that the statute would probably not prevent the approval of such an application. It seems to me that a proper interpretation of the statute would hold "objectives" to mean the objectives of assessing personality and predicting behavior, not merely the objectives of teaching graphological techniques. Anyway, the use of state licensing laws by pseudoscience practitioners to achieve credibility and legal protection is a potentially serious problem. Consider the success of the chiropractors, who years ago embraced medical licensing laws instead of fighting them.
Finally, there is no state certification or licensing of graphologists in Texas. Statements to that effect in graphologists' promotional literature can only refer to "certification" by a graphological school or association.
American Psychological Association, Inc.
Standards for educational and psychological tests, Washington, D.C. 1985.
American Society of Professional Graphologists:
Beyerstein, Barry L., "The Origins of Graphology in Sympathetic Magic", in The Write Stuff: Evaluations of Graphology, the Study of Handwriting Analysis 163.
Beyerstein, Barry L. and Beyerstein, Dale F., Ed. (1992). in The Write Stuff: Evaluations of Graphology, the Study of Handwriting Analysis. Prometheus Books.
Bowman, Marilyn L., "Difficulties in Assessing Personality and Predicting Behavior: Psychological Tests and Handwriting Analyses Contrasted", in The Write Stuff: Evaluations of Graphology, the Study of Handwriting Analysis 203.
Bunker, M. N. (1970).
The Science and Art of Reading Character by Grapho Analysis. Nelson-Hall Co.
Hopper, Mark, in "Handwriting Resource Corporation": http://www.handwriting.com/.
Hines, Terence (1988).
Pseudoscience and the Paranormal. Prometheus Books.
Hyman, Ray (1976). "'Cold Reading': How to Convince Strangers that You Know All About Them."
Zetetic, 1 (No. 2), 18-37.
Jansen, Abraham (1973). "Validation of Graphological Judgments: An Experimental Study", Mouton Publishing.
Klimoski, Richard, "Graphology and Personnel Selection", in The Write Stuff: Evaluations of Graphology, the Study of Handwriting Analysis 232.
Kurtz, Sheila and Lester, Marilyn (1983).
Graphotypes. Crown Publishers.
Lowe, Sheila in "Welcome to Sheila Lowe's Handwriting Analyzer": http://www.writinganalysis.com/. (Lowe is also author of The Complete Idiot's Guide to Handwriting Analysis published by Alpha Books).
Neter, Efrat and Ben-Shakhar, Gershon (1989). "The Predictive Validity of Graphological Inferences: A Meta-Analytic Approach." Personality and Individual Differences, 10(7), 737-745.
Nevo, Baruch, Ed. (1986).
Scientific Aspects of Graphology. Charles S. Thomas, Publisher.
Rafaeli, Anat and Klimoski, Richard J. (1983). "Predicting Sales Success Through Handwriting Analysis: An Evaluation of the Effects of Training and Handwriting Sample Content." Journal of Applied Psychology 68, 212-217.
Rosen, Billie Pesin (1965).
The Science of Handwriting Analysis. Crown Publishers, 1965.
Solomon, Shirl (1978).
Knowing Your Child Through His Handwriting and Drawings. Crown Publishers. Spohn, Julie A., The Legal Implications of Graphology, 75 WASHINGTON UNIVERSITY LAW QUARTERLY, No. 3 (Fall 1997); also
This fact sheet was written by John A. Thomas. Portions previously appeared in The Skeptic. Copyright 2002 by the North Texas Skeptics.