Graphology-L FAQ | Glossary | Bibliography | Newsfeeds
Section Index | Professional Services | Copyright | ToC

Graphology: A Legal Point Of View

The author of this piece is not a lawyer. This article is not meant as legal advice. I wrote it to get an idea of some of the potential legal problems in the field.

If you have any questions, or want advice, talk to a practicing lawyer who specializes in employment legislation.

Abstract

This paper describes describes some of the legal threats to handwriting analysis. Whilst an adequate defense might exist, the lack of sound research greatly inhibits any defense that could be raised.

The use of handwriting analysis, as a tool in the hiring process, is akin to walking through a mine field without a mine detector. The legal department would view the use of graphology in the light of Spohn's (1997) The Legal Implications of Graphology. The public relations department would view it in the light of Barry Beyerstein & Dale Beyerstein's (1992) The Write Stuff: Evaluations of Graphology - The Study of Handwriting Analysis. The human resources department would view it through the observations of Stagner's (1955) The Gullibility of Personnel Managers and Steiner & Gilliland's (1996) Fairness Reactions to Personnel Selection Techniques in France and the United States.

None of those papers are favorable to the use of handwriting analysis. They all ask hard questions about whether or not handwriting analysis should be used under any circumstances, let alone as a tool for evaluating potential employees.

Scientific Status of Handwriting Analysis.

As a general rule of thumb, skeptics claim that there is no validity to handwriting analysis, whilst proponents argue that handwriting analysis is valid.

I estimate that 2 000 papers on handwriting analysis have been published in peer reviewed journals. Blake (2000 b) outlines problems with those papers, noting how and where they fail to meet the minimum acceptable standards of scientific research.

Furnham & Gunter (1987) point out that their research was just one in a long string of studies that demonstrate that handwriting analysis has no validity, further suggesting that no further research is needed.

Scientific Methodology of Handwriting Analysis.

Graphoanalysis (IGAS 1975) claims to be the scientific way to analyze handwriting. The Müller-Enskat Protokol (Müller & Enskat 1973) and Wittlich Character Diagram (Wittlich 1973) claim a scientific methodology. Cole (1961-1968) claims:

The Psychogram offers a proven method of teaching scientific quantitative handwriting analysis.

Edson (1971) describes the differences between Graphoanalysis and other systems of handwriting analysis. In this respect, Carroll v State (634 SW 2d 99 1982) is interesting, in as much as the Graphoanalyst denounces the graphologist, for practicing a pseudo-science.

There are at least twenty different systems of handwriting analysis. They share a common methodology, in that they require that specific features of handwriting be measured, collated, then interpreted. All claim some degree of scientific research behind them. None of them have independently replicated research behind them.

Terminology of Handwriting Analysis

The field of handwriting analysis has a technical jargon, Despite numerous (IGAS 1964; Whiting 1991 a; Bradley 1999, Vels 1972) attempts to provide a comprehensive dictionary of technical jargon, handwriting analysts still use terms that other analysts neither understand nor comprehend. Sometimes, they use the same word, but mean entirely different things. Other times they use different words to mean the same thing. Examples:

Block Print
Unizone print used by architects, draft persons, commercial artists and similar professions. (Whiting 1982)

Calligraphy
Art form of penmanship generally done with nibs that create thick/thin strokes. (Whiting 1982)

Cursive Writing
Writing in which letters are joined in each word as prescribed by most standard Models. (Whiting 1982)

Italic Writing
[ definitions to be added ]

Manuscript Writing
A Disconnected script using all three zones. (Whiting 1982)

PrintScript
  1. Similar to manuscript. Taught with simple "circles and sticks" to first and second grade pupils. (Whiting 1982)
  2. Letters devoid of style or character are plain and straight forward, built up from straight lines, circles and parts of circles. (Piggott:1958:33)

Printed Writing
Letter Style usually taught in the US prior to learning cursive writing. (Whiting 1982)

Rhythm
  1. Primary Trait:
    When the movement of the pen is characterized by basically regular recurrence of beat, the written strokes illustrated in Specimen 11 return to the base line with an even spacing between stroke combinations.
    (IGAS 1964)
  2. Noun:
    Flow of writing responding to natural cadence in pattern of tension and release. May be found in a consistent return of downstrokes to the baseline, especially in letters 'm', 'n' and 'h'. Reflects harmony of arrangement, form and movement.
    (Whiting 1991)

Speed
  1. Number of characters written per minute.
  2. Number of words written per minute.

Specific Objections to Handwriting Analysis.

The general objection to handwriting analysis is that it is not valid. That issue has been dealt with elsewhere. (Blake 2000 b)

Barnum Effect

This is a problem of all methods of personality testing. The higher the face validity of the test, the greater the Barnum Effect is. One of the few ways to minimize this problem, is to make a serious and consistent effort to individualize each and every report. Tallent (1950) discusses this problem for the general clinical report.

The Green Sheet, used by Graphoanalysts, is one example of a systematic tool promoted by an organization of handwriting analysts, to individualize graphological reports.

Erika Karoh's Step By Step System of Handwriting Analysis can produce reports that are individualized. The use of the same description for each aspect of personality does lend itself to dis-individualizing of reports.

Dr Fox Effect

In reading the promotional material that handwriting analysts put out, one reads things like 100% accuracy in detecting dishonesty (Vanguard: 11 August 2000); 80% accuracy in determining your personality (Bart Baggett: 2 August 2000).

In neither instance was supporting research cited. In a week of follow up messages, the individual who wrote the first message made the observation that she could not detect honesty in the script. How then can the original claim be made, or supported?

People accept figures like that because they sound impressive, and the person who makes the claim implies that they know what they are knowing about. This is the Dr Fox Effect in action.

Aunt Fanny Effect

In the graphology field, this is best viewed at meetings of IGAS, or chapters of IGAS. The group think is one of Bunker said something so it must be valid. Thus, devotees of Graphoanalysis spread their enthusiasm, and disregard the facts.

Effect Size

This is Dean's (1992) only substantial argument against the use of graphology. The effect size is too small. In other words, for a specific individual, it possibly is not useful. For a large group, it can be useful. A survey of the studies he cites, shows glaring problems in research design, methodology and assumptions that the experimenters made about handwriting analysis. These are further discussed in Blake (2000b).

Validity

Despite numerous claims that handwriting analysis is valid, there is nothing that actually substantiates that claim. Fluckiger, Tripp & Weinberg (1961), Lockowandte (1976), Nevo (1986a) and Moore (1985) survey research in the field of handwriting analysis. Their basic conclusion is that whilst more research is needed, the published research is of an acceptable quality, to minimally support the validity of handwriting analysis. Beyerstein & Beyerstein (1992), using the same basic data, demonstrate that the published research is not acceptable.

One major objection to all of the published research, is that none of the studies

  1. use random subject selection
  2. are double [ or triple ] blind

Concurrent Validity

This is the most common type of validation study that handwriting analysts do. At best, these can be described as interesting. At worst, they simply demonstrate that handwriting analysis is not valid. The single biggest problem with this type of validation study, is the validity of the test that it is being measured against. For every study which has found support for concurrent validity for handwriting analysis using psychological tests such as MMPI, Rorchach, TAT (LaMonica; Secord 1949), Draw A Person Test, House Tree Person Test, 16PF Scale (Naegler 1960), there is a study which indicates otherwise.

Content Validity

This is a subjective measure of validity. The content validity of a report is as high as the degree to which the report covers what the client expects it to cover. CF Forer (1949) for a discussion of problems with self-validation.

Face Validity

This measures validity subjectively. Briefly, if something makes sense, then it has high face validity. If it does not make sense, it has low face validity.

The literature of the field is replete with quotes by famous psychologists stating that there ought to be something to the field of handwriting analysis. This indicates that handwriting analysis has high to very high face validity.

Global Validity

Crumbaugh & Stockholm (1977) found that Graphoanalysis is valid. Vestewig, Santee & Moss (1976) in a similar study, found that graphology is not valid. Crumbaugh (1977 a) is a criticism of that study. Vestewig & Moss (1976) is their rejoinder to Crumbaugh's criticism.

Predictive Validity

Can one make a prediction from handwriting analysis? Looking at the claims made in the promotional literature, one would expect so.

One of the earlier studies along these lines is Lewinson & Zubin (1941).

More daring, in both what it predicts, and how that prediction is made, is Iannetta, Craine & McLaughlin (1996). They used handwriting to determine whether or not individuals who are criminally insane, will also be violent. Regrettably, this study has not been independently replicated.

Meta-Analysis

Rosenthal (1984) found that meta-analysis is a usable technique. Since then, meta-analysis studies have been done by Nevo (1986) and Dean (1992). Nevo found that handwriting is valid. Dean found that it might be valid, but that the effect size was negligible.

Reliability:

Test Retest Reliability

For handwriting analysis, this can be split into two different things.

  1. Will the same handwriting analyst score the same sample of handwriting the same way, six or more months apart?

    I have not found any literature in the field, that addresses this question. The closest I have come, is inter-rater reliability testing. At least one handwriting analyst has objected to this type of research, on the grounds that since the analyst has gained more experience, the resulting report will obviously be different.

  2. Will the same handwriting analyst score samples of handwriting taken from the same individual, at different points in time, the same?

    The short answer to this question is it depends. Handwriting does change over a period of time. Foley & Miller (1979) report on the short term effects of marijuana and alcohol on handwriting. Beck (1985) and Bosch et al (1974) discuss the long term effects of alcohol on handwriting. In 1973, the House of Seagram distributed a poster showing the effects of alcohol after a number of drinks.

    From a psychological point of view, Epstein (1979; 1980) demonstrates the stability of behavior, and makes the argument that more observations made over time are more reliable than one or two observations made over a short period of time. Supporting this position is Wolff (1948) when he demonstrates that the fundamental ratios of an individual's writing remain constant over time, although the superficial appearance does change.

    Anecdotal evidence implies that the reports will be vastly different, if the subject of the report changed his/her handwriting, either out of a desire to change their penmanship (make it more legible, more aesthetic, etc) or as a result of Graphotherapy, if the analyst is not told prior to writing the reports, that the samples are from the same individual. One can't but help wonder if the graphologists are getting caught at looking at the superficial appearance of the writing, and ignoring the underlying structure.

Alternate Form Reliability

Do different systems of handwriting analysis come to similar conclusions?

This has not been systematically studied. Anecdotal evidence implies that they would.

Alternate Observer Reliability

If two handwriting analysts score the same sample of writing, will they come to the same conclusions?

Galbraith & Wilson (1964) found an inter-rater correlation of 0.78. Impressive, until one realizes that there were just three handwriting analysts, and less than ten samples were done. Vestewig, Santee & Moss (1976) implies that inter-rater correlation is low, but provides no specific figures. Peeples (1980) found that inter-rater scores are similar, for analysts with similar experience and education.

Normative Scores

Published Norms

Piggot (1958) was simply a survey of what the handwriting of the England looked at. Nonetheless, it does make some interesting observations.

Livingstone (1963) attempted to come to definate conclusions as to what constituted an American norm, partially to help in his forensic work. The net result of his paper is to show how the local population of arrestees deviates from Zaner-Bloser and Palmer scripts.

Institut für Webepsychologie und Markterkundung (1966)

Stockholm (1980 a: 1980 b) published the result of The IGAS Trait Norm Project. Whilst this could have been a tremendous advance in the field, the best one can say for it, is that it is a practical demonstration of how not to do research. Belling et al (1977) describes numerous problems with this study. Thus far, no replication studies have been done.

Anthony (1963) boldly declares that the writing sample is to be scored against The American Writing Style, with no further indication as to exactly what is meant. Cole (1961-1968) states that normative scores are not needed, as the Psychogram is normed for Palmer Script. What he omits to state is

  1. why that is the case?
  2. how to correct for individuals who did not have Palmer as their copybook.
Bowers (1999) declared that a normative study was not necessary, and that the Psychogram is normed for Zaner-Bloser, not Palmer Script. Oddly enough, in Cole claims that the Psychogram is normed acording to a different copybook.

The Wittlich Character Diagram might not require norms, as it is based upon the deviation from the copybook that the individual was taught at school. If an individual was taught two or more copybooks, or less commonly, not taught to write at all during their education, no help is provided on how to accurately use this system.

Standard Deviation

Stockholm (1983) reports the standard deviation for each of the traits of Graphoanalysis. This study complements the IGAS Trait Norm Project (Stockholm 1980a; 1980b). None of the other systems of handwriting analysis provide norms, with the attendant standard deviation scores. This implies that the scientific method has yet to be applied to the field.

Falsification

Is there an alternative explanation for positive results found in graphological research?

Virtually none of the research done by graphologers addresses this question. Research by non-graphologers usually has negative results. Beyerstein & Beyerstein (1992) do address this issue, when they rebut the most common criticisms graphological research that handwriting analysts analysis, when results are unfavorable for graphologists.

Peer Review

Impact Magazine was unique, in that all submissions were subject to a blind peer review, prior to publication. After it folded in 1982, both The Journal of Handwriting Psychology and The Journal of Integrative Graphology peer reviewed all submissions. They folded in 1984. Currently, The Journal of the American Graphological Society claims to be peer reviewed. None of the other journals in the field are peer reviewed.

Equally distressing, when flaws in research design or methodology are reported, the reporter often is subject to extremely hostile behaviour.

Publication of Results

Handwriting analysts usually self publish their papers, and books. Vanity publishing appears to be a fall back position. In both instances, peer review is omitted. It should not be a surprise that graphology books contain numerous examples of material that is demonstrably false.

As one example, a review of Andrea McNichol's (1991) Handwriting Analysis: Putting it to Work For You, reveals an average one error per page. Sheila Lowe's (1999) An Idiot's Guide To Handwriting Analysis has an average of one major error per chapter, in addition to the minor error per page. Despite that ratio of errors, both of these books have received high praise by other handwriting analysts. Equally surprizing, these are considered by many to be "professional quality" textbooks.

One might argue that since the research in the field is so sloppy, their material might not be error. If either author provided any explanation or support for their claims, other than this is what I have seen, it would be a valid claim. Let me take one simple example. Both books claim that gender can not be determined from handwriting. Neither of them provide any explanation for the results of any of the 29 studies cited in the section on gender in this paper.

Court Acceptance

Contrary to what numerous handwriting analysts claim, handwriting analysis, for the determination of personality has not been accepted by Appellate courts.

Gold Nibs published a list of court cases, where graphology allegedly was accepted. None of the cited cases were at the Appellate level. The vast majority are for QDE cases. « None of the cases used Blue Book guidelines for legal citations »

Handwriting Consultants of San Diego (1983) make the claim that handwriting analysis has been used in courts throughout the United States as far back as 1881.... (103 ALR 900). A review of that brief reveals that it is a discussion of questioned document examination, not handwriting analysis.

Frye

Frye v United States (54 App DC 46; 293 F 1013 (1923)) was the foundation upon which scientific evidence was admissible as testimony or not, from 1923 to 1995.

Cameron v Knapp (520 NY 2nd 917, 917-18: New York Superior Court 1987) is the first case I've found, where handwriting analysis was taken to an Appellate Court. The court ruled that handwriting analysis was not admissible evidence, pointing out that no other court had accepted graphological testimony.

Daubert

Daubert v Merrel Dow (113 5 CT 2786; 509 US 579 (1993)) radically altered what could, and could not be acceptable as scientific testimony. In brief, it requires that the following questions be answered affirmatively:

  1. Is the science valid?
  2. Is there a scientific methodology to the science?
  3. Is that methodology valid?
  4. Is it reliable?
  5. Is it falsifiable?
  6. Has it been subject to peer review?
  7. Is the supporting research published?

The brief answer to all of those questions is no. A longer answer would be "it depends" and supporting evidence for both "yes" and "no" for each answer. Thus, I would conclude that handwriting analysis does not meet the requirements of a science that Daubert sets forth --- especially since handwriting analysis did not meet the far less stringent requirements of Frye.

More to the point, when graphologists have testified in courts, the courts have ruled it to be junk science. (Cameron v Knapp) For a practical demonstration of the unscientific nature of handwriting analysis, Carroll v State stands out as a striking example of each side's expert witness discrediting the expert witness of the other side.

Mis Diagnosis

This is covered by Türkel (1932, 1933) and Jacoby (1930). The only counter to the errors described in those papers, is continual education and experience in analyzing handwriting. In part to combat this, several organizations of handwriting analysts, have, as part of their code of ethics, a clause such as the following:

Members must recognize the boundaries of their competence
and the limitations of their techniques, and only provide services,
use techniques and offer opinions as professionals that meet
recognized standards.

Third Party Analysis

Reviewing the promotional literature for HuVista, Inc, and Handwriting Consultants, Inc, consent for third party analysis is seldom, if ever required. A review of the code of ethics of AAHA, AHAF, IGAS and Vanguard, shows that they allow third party analysis without the consent of the subject of the report.

AAHA's Code of Ethics states:

Any third party must have a legitimate reason for the analysis.

Any third party shall agree, preferably in writing, to keep the information confidential.

(AAHA 1995)

Whilst that does not prohibit the third party from providing a copy of the report to the subject of the report, it does imply that the report is not to be made available to the subject of the report.

The Personal Worth Course (Handwriting Consultants of San Diego 1983) states that:

Before giving out information, determine the purpose behind the request.
Pure nosiness is not a legitimate purpose.
The following is a short list of normal and legitimate reasons
for requesting the analysis of a third party:
  1. Proposed business relationship.
  2. Pre-Marital analysis.
  3. Compatibility analysis; business, family or friend.
  4. Family members in trouble.
  5. Potential roommate.
  6. Suspicion of illegal activity which affects the client.

Several professional books indicate that the subject may not receive, review or have access to the graphological report. Reading , in the last paragraph, the analyst states that reports are to be delivered over the phone, as a means of avoiding lawsuits.

It should also be pointed that the the FTC has radically changed the defintions found in the FCRA. A private opinion has been provided, implying that graphologists do not currently fall under it, that does not mean that at a later date graphological reports would fall under that statute.

Privacy

Despite the claim of Handwriting Consultants of San Diego (1983) and others, that courts have ruled that handwriting is public behavior, the plain fact is that that is not the case. The following cases are usually cited in support of that thesis:

US v Mara aka Marasovich (410 US 19 (1973)) would be a more appropriate case to cite, as it revolves around whether or not a grand jury can compel an individual to provide a writing sample. In all three instances, the question before the court was about whether or not a defendant could be compelled to produce a sample, to determine whether or not an exhibit was the defendant's or not. In the first instance, it was a voice exemplar, in the other two, it was a writing exemplar. None of these dealt with the question of whether or not the psychological aspects of writing, as determined from handwriting analysis, was an invasion of privacy, or not.

In a similar vein, courts have ruled that a urine test, to determine drug usage, is acceptable, if done with the parameters of the law. Using that same urine, to test if a female applicant is pregnant, is a violation of one's privacy. .

Handwriting Analyst as Charlatan

This is the major underlying theme of Beyerstein & Beyerstein (1992). The field has provided very little to demonstrate otherwise, and much ammunition to support this position. Amend & Ruiz (1992), Lowe (1999) and McNichol (1989) all make numerous statements that either is not backed by research, or else flatly contradicts the results of published research. In no instance to they provide a rational for why they make the claims that they do make.

They also implore the analyst to use his/her intuition, when doing an analysis, thus demonstrating that, in the end, handwriting analysis is nothing more than a sophisticated application of cold reading.

Legislation:

Is handwriting analysis legal to use? When Julie Spohn (1997) addressed this issue she concluded that whilst it was legal, doing so is irresponsible, and it ought to be banned. Whilst her paper is extensive, it omits several instances where the practice of handwriting analysis is currently illegal, either directly, or as a side effect of other legislation.

There have been direct attempts by several state legislatures to ban the use of handwriting analysis as an employment screening tool. (Rhode Island, 1994; Idaho; Iowa; Oregon, 1993.) Those attempts were squashed. The indirect attempts may have succeeded. ( Rhode Island 1996.)

In various jurisdictions, such as Greenwood County, South Carolina, Handwriting Analysis is explicitly prohibited, under the general fortune telling statute. No doubt, if challenged, the result would be much the same as astrology did, when challenged under the New York Statute.

Equal Employment Opportunity Commission Rulings.

The International Graphoanalysis Society (IGAS 2000) claims:

It cannot distinguish sex, race, religious affiliation or sexual preference. This allows the Graphoanalyst ® to provide a detailed and unbiased assessment.

Similar statements have been made by other organizations of handwriting analysts.

Barrow & Scott's (1984) Validation of a Personnel Selection System to meet EEOC Guidelines supposedly demonstrates that graphology meets the 1978 guidelines. It does not appear to meet either the 1984 (3 CFR 402) or 1999 (29 USC 158) guidelines.

The three major problems with that paper are:

  1. It was designed for guidelines that are out of date;
  2. It uses a proprietary system of handwriting analysis;
  3. It makes some demonstrably false assumptions about graphology.

This paper is discussed in more detail in Blake (2000 b). US V Hazelwood School District or Griggs v Duke Power Company are often trotted out as cases which state that:

Handwriting analysis is not precluded in hiring if it is related to the job. Because the analyst never sees the applicant, and because handwriting cannot tell race or whether a writer is male or female, old or young, right or left handed, it obviously non- discriminating. Judgments must be made solely on the writing. Personal bias is kept to a minimum. (Sassi & Whitting 1983)

Hazelwood is about hiring techniques, or the lack thereof. Griggs v Duke Power is about what makes job requirements, and disparate impact. Neither mention handwriting analysis.

Age:

IGAS, in their General Course of Graphoanalysis, Lesson 1, Page 5, states that:

Graphoanalysis does not take into account the physical age of the writer, but confines itself to the mental capacity, maturity or potentialities of the writer as shown in the handwriting.

The Personal Worth Chart Intermediate Course in Handwriting Analysis (Handwriting Consultants of San Diego 1983) states on page 4 that:

Age, sex and handedness are biological factors that cannot be determined from the handwriting alone. Be sure to know these things about your client.

Betty Semler (nd) on page three states that:

Knowing age is critical for an authentic comprehension of the writer. For instance, if traces of tremor are found in the writing of an elderly person, symptoms of aging may be acknowledged, but if the same tremors appear in the writing of youth, weakness in the nervous system, caused by either an illness or recent trauma should be considered. Hans J Jacoby writes, age, according to the calender can be significant for the construction of a correct frame of reference for the psychological evaluation of the handwriting as it bears on qualities such as immaturity or aging.

Similar quotes may be find in most books and courses on handwriting analysis. Under EEOC Guidelines, it is a violation to ask an applicant their age, or anything that would give an indication of how old the job applicant is.

Research by Binet (1904), Binet (1904a), Bryan (1966), Crepieux- Jamin (1924), Epstein, Hartford & Tumarkin (1961), Galdo (1933), Jores (1930), Lëgrun (1931), Middleton (1941a), Paul-Mengelberg (1965) and Tang (1933) imply that age both affects how the handwriting appears, and may influence the the final analysis. Whilst none of the research papers I have studied have conclusively demonstrated that physical age may be determined from handwriting, that has been implied in some of them.

Birren & Botwinick (1950) and Birren (1951) demonstrate how speed in writing differs between "young" populations, "elderly" populations and "elderly senile" populations.

Sassoon (nd) points out a cluster of writing features common to females between the ages of 11 and 22, and very uncommon in other population subgroups. Masson (1985) picked up this same cluster, explicitly naming it The Classic Style. This cluster appears to be cross-cultural.

Finally, one can point to Erika Karoh's How to Tell Life Expectancy From Handwriting for information on the upper limit of a person's age.

Gender

Semler (nd) on page three states that:

Again, quoting Jacoby, The sex of the writer should be known so that the graphologist can determine to what extent the writer is in agreement with his, or her, actual sex. There is a general tendency towards contraction in the writing of men. Women are more likely to accentuate the release impulse when writing.

Both Bunker(1947: :6; 1955:5) and The International Graphoanalysis Society, in their General Course of Graphoanalysis, Lesson 1, (IGAS: 1975: 6) states that:

Handwriting is sexless....Fifty years ago it existed because women were, in many instances, thinking in a restricted, clinging-vine sort of fashion. ... The matter of sex is secondary as far as mental expression of identification is concerned.

This is the closest that any organization or course comes to admitting that gender can determined from handwriting. This admission takes on greater significance when one looks at cultures which still expect males and females to behave differently.

In The Encyclopedic Dictionary for Graphoanalysis, First Edition (IGAS, 1964) we find the following for the definition of sex:

Whether or not a writer is male or female can not be definitely determined by handwriting.

Subsequent editions of the Encyclopedic Dictionary, (IGAS, 1978; IGAS 1984) make no mention of age, sex or gender. The Grapho Analysis Dictionary (Bunker 1938) has the following listed under sexual:

Handwriting does not identify sex, but sexual appetites when highly developed, or out of the ordinary.

Binet (1903), Binet (1904), Broom et al (1929), Castelnuovo- Tedesco (1946), Castelnuovo-Tedesco (1948), Crépieux-Jamin (1924 b), Downey (1910), Eisenberg (1938), Galdo (1933), Gesell (1906), Goodenough (1945), Hardyck & Goldman (1975), Kaplan (1957), Kinder (1926), Lester & McLaughlin (1976), Merllie (1990), Middleton (1938), Newhall (1926), Tenwolde (1934), Ungar (1940), Wagner (1934), Wallner (1961), Wittlich (1927) and Zazzo (1948) imply that gender can be determined from handwriting by both lay-people and professional handwriting analysts.

In a different study, Castelnuovo-Tedesco (1946) found that both intelligence and gender could be determined from handwriting at a statistically significant level.

Zweigenhaft & Marlowe (1973) found a significant correlation between self esteem, signature size and status in society.

Abel Wrifford taught ladies Italian Style as it is suitable for those of limited capacity, whilst his Mercantile Penmanship was for men who pursued a business career.

French (1845, 1846) demonstrates two different copybooks, one designed for woman, and the other designed for men. Indeed, Nash (1969: 52) quotes a flyer offering to teach new and elegant epistolary handwriting which has been so generally adopted by the most fashionable ladies all over the United States.

The Marion Richardson System of Penmanship was taught at girl's schools in England during the fifties. A clear case of where if the copybook can be correctly identified, the gender of the writer can be correctly guessed at.

National Origin

Klara Roman (1952), amongst others, point out that the copybook that one uses affects the way that one writes. That copybook can also be used to determine the national origin of the writer. The US Post Office ( Bureau of the Chief Post Office Inspector: 1947) published a handbook of systems of handwriting that were used in the United States. Stephes & Somerford (1953) provided an addendum to that publication, by including foreign copybooks. Beacon (1965) provides an update by including systems that either first saw use between 1847 and 1965, or were unintentionally omitted from the 1947 publication. Blumenthal (1957) published a book of copybooks of countries which used the Latin alphabet, as a reference work for graphological analysis. Steer (1993) published a book with the alleged current copybook for many countries in the world. Klimonski & Rafaeli (1983) point out that copybook correlates with where and how one is taught to write. Thus, by a close analysis of an individual's script, one could, in theory, determine the copybook that the individual was taught, and hence their national origin.

Will the difference in copybooks affect the interpretation of the handwriting? Perhaps. Perhaps not.

Cole (1961-1968: 201) writes:

Copy Book is scored from 1.0 to 10.0. If the writing in all aspects compares favorably with Copy Book, the score will be 10. Deduct for variations in size of the three zones, formations of all letters, slant, starting and ending strokes, pressure, the capital "I", punctuation, i dots and t crossings; Variations in any of these from Copy Book detract from the score...Had the person learned to write using a different style of alphabet, we still score according to our Copy Book. The differential is made up in other areas of the Psychogram.

In Condensed Psychogram Scoring Guide (Cole 1961-1968: 110-113) we find a score in all forty criteria for copybook writing. If one's copybook is neither Zaner-Bloser, nor Palmer Script, then the score for one's writing suffers accordingly -- if the sample can be correctly analyzed.

For example, if the writing sample is in Turkish, and the analyst doesn't know Turkish, the analyst could very easily incorrectly score various items, because of the apparent inconsistent correct placement of various diacritic marks.

In 1920, some parts of Scotland began teaching Nelson Handwriting. Later, most of the rest of Scotland adopted the Alfred Fairbank Model. In 1976, Glasgow adopted the Gourdie Model of Italic Handwriting. In 1993 most of the rest of Scotland switched to this model. Fife, however, uses New Nelson Handwriting.

Circa 1935, London county schools started to teach The Marion Richardson System of Penmanship. Shortly thereafter it was introduced into provincial schools. However, by the fifties, it was taught almost exclusive at public schools for girls.

The American Association of Handwriting Analysts has stated that:

Printed writing is persona writing, and as such, can not be correctly analyzed. [ AAHAOffice@aol.com, 1998 ]

On a related note, it should be pointed out that scripts based upon Devanagari appear to "break" the Wittlich Character Diagram. Pictoral Moso, as a writing system, is one that can not be analyzed by graphologists. [ I will conceed that there is a question about whether or not it is an actual writing system. ] Ideographic Moso, can, in theory, be analyzed by handwriting analysts. However, it does break some systems of handwriting analysis

Korean writing also presents a problem for handwriting analysis, as the characters are compressed into a square, due to the influence of Chinese Ideograms.

Ethnicity

Garth, Mitchel & Anthony (1939) found no difference between the handwriting of Negroes, and that of Whites, using the Thorndyke handwriting scales. Weisser (1932), and Garth (1931) found no difference between the handwriting of Indians and that of Whites, using the Thorndyke handwriting scales. A careful review of those studies demonstrates some serious flaws. These flaws are discussed in Blake (2000 b).

Krieger (1935)

Krieger (1937)

Caille (1951)

Anecdotal evidence indicates that there are a few subtle, but significant differences between the handwriting of "mainstream USA" and that of inner city ghetto black males under the age of forty. I am not aware of any published studies on this difference. If this difference does occur, then the result will be a disparate impact against inner city black males under the age of forty.

Religion

In theory, religion can not be determined from handwriting. The practice might be slightly different.

Klages "discovered" the "Religious Curve" which reflects the spirituality one has. Whilst it does not point directly to one's religious affiliation, it implies the degree to which one is committed to one's religious beliefs. One does wonder how it would be reflected in the script of deeply committed Discordian. Roman (1952) points out that national origin can be determined from the copybook. This simple principle can be applied elsewhere. For example, those who went to Catholic school usually learned to write . This is not usually taught at non- Catholic schools. In a similar situation, children who learned to write at schools run by ISKCON, learned . This copybook is not usually taught at secular schools. Deseret Script is taught at some schools run by the Church of Jesus Christ of Latter Day Saints, and Bob Jones University Script, which is mainly taught to children of fundamental christians who home school their children.

Additionally, various groups in the religious fringe require their adherents to use a specific copybook. For example, The Order of St Michael uses a copybook based upon cuneiform. Several Norse based groups use Futhark as their copybook. Masons, from the mid- eighteenth century to the early-twentieth century, used their own alphabet for most, if not all of their internal communications. Koehn (nd) points out that one's religious beliefs do correlate, even if only in a subtle way, with one's personality. As such, the handwriting analysis can contribute to disparate impact against members of some religions.

Complicating the picture, are the Keresans, who believe that their language is sacred, and must never be written. This is an interesting switch from the common primitive belief that all writing is sacred, and may not be destroyed.

Illiteracy

The illiterate, by definition, are not able to write. [ There might be a question as to how one defines illiteracy. England, in 1934, for example, defined illiteracy as signing a marriage register using a mark. India, by contrast, in 1921, defined illiteracy as ten years old, or older, and unable to read and write a short letter. Japan, in 1925, defined illiteracy as being unable to write and to figure. ] There is a legal doctrine that all job applicants are to be subject to the same hiring process. Thus, if any applicants are illiterate, the use of handwriting analysis is questionable.

The American's With Disabilities Act.

The ADA (42 USC 1211) made sweeping changes in the US. The most visible to most people was the installation of ramps for wheelchairs in public buildings, and curb cuts. The least visible are conditions under which one may hire, or terminate employees. Litigation as to what is, or is not required under the ADA is under continual review in the judicial system.

Mary Benbow's Loops and Other Groups copybook is almost exclusively taught to individuals who are in special education classes. The exceptions are confined mainly to students in Hingham, MA --- which brings up geographical identification.

Blind Individuals

Heiss (1954)
Masson (1988)
Mathies (1930)
Weiss & Weiss (1978)

Physical Data points

Research has shown a wide range of physical conditions to be detectable from handwriting. Menstruation (Mauthner 1949), cancer (Kanfer & Casten 1958; Vèrtesi 1939; Victor 1959), cerebral palsy (Kalichnyuk 1970); arthritus (Booth 1937; Lewinson 1938), asthma (Pontius 1953), aphasia (Boone & Friedman 1976); Roch et al 1972; Sasanuma & Fujimura 1972; Schwartz, Nereuff & Reiss 1974), echographia (Pick 1900; Pick 1924) and Encephalitis ( Langer 1972) are just some examples of diseases diagnosable in handwriting. Hilton (1969a) demonstrates how health conditions do affect handwriting, when attempting to authenticate handwritten documents. Rider (1982) describes the difference in handwriting between individuals with cancer and individuals with heart disease.

Knopp et al (1970), Lemos-Magelhaes (1927), McLennon et al (1972) and Schwab (1968) demonstrate how easily Parkinson's Disease, and other, similar diseases can be tracked, by the effects that they leave on one's handwriting. Gross (1975) reports on the effect that various drugs used to treat diseases, have on handwriting. Anecdotal evidence indicates that items such as heart disease (Olyanova 1969: 263), tuberculosis (Olyanova 1969:265), menopause (de st Columbe 1966), Obesity (Amend & Ruiz 1992), Thyroid problems (de st Columbe 1966: 311), Kidney Disease (de st Colombe 1966: 314) can be detected from handwriting.

Mental Problems

Hilton (1962) documents how to identify writing of individuals that are suffering from a mental disorder. Unger (1953), Mühl (1950) and Lewinson (1940) describes graphological guidelines for the diagnosis of schizophrenia, paranoia and similar conditions. Mucher, Grünewald & Fuchs (1964) describe the graphomotor effects of claustrophobia. Anecdotal evidence indicates that problems such as hysteria (Olyanova 1969: 259), petit mal (Olyanova 1969: 259), senile dementia (Olyanova 1969: 260), schizophrenia (de st Colombe 1966: 256), locomotor ataxia (de st Colombe 1966: 313), neurasthenia (de st Columbe 1966: 317) can be determined from handwriting.

Psychiatric Problems

With the advent of DSM IV, handwriting is used a criteria for some pathologies. Goodman, Downing & Rickels (1964) demonstrate how the handwriting of a group of depressed patients changes, during the course of their hospitalization and treatment.

Senility (Kapp 1937), Psychosis (Lëgrun 1939), Schizophrenia (Suchenwirth 1967; Schorch 1939) Anxiety (Sultzer 1949), Psychopaths (Breil 1967; Breil 1967a ), Sociopaths (Gerstner 1931) and similar pathologies have been extensively documented in the medical literature. Mühl (1950a) and Pittrich (1957) document how handwriting changes during the course of psychotherapy.

Suicide

This is discussed by Lëgrun (1930) and Vadus (1984). In brief, it appears that graphologers can determine whether or not a suicide note is genuine or not, at a statistically significant level. King (1970) correlated various graphological indicators, with Matza's factors for suicide lethality.

Criminally Insane

Ianneta (2000) describes her research with the criminally insane, and whether or not they are predisposed to violence. Her research has not been replicated. Thus far, it appears that it might be valid for the population group she studied. That has not prevented other handwriting analysts from using her material, and applying the results to the "normal, real world". Such a use, is, at best, unethical.

Addictions and Handwriting

Nicotine
Opium
Vitamins
Minerals

Purtell (1965) reports on the short term effects of drugs on handwriting. Dhawan, Bapat & Savan (1969) demonstrate how a different set of drugs affect handwriting.

Fischer et al (1969, 1970) describe the short term influence of psilocybin on handwriting.

Brun & Reisby (1971), DeTrey (1954), Hilton (1962), Kirchner & Lemke (1973), Mayer (1901), Puri (1965), Rabin & Blair (1953), Reisby & Theilgaard (1969), Tripp, Fluckiger & Weinberg (1959) describe the short term and long term effects of alcohol on handwriting. Lëgrun (1936) looked at the handwriting of children of alcoholics, noting that their handwriting was similar to that of alcoholics.

Hirsch et al (1956) reported that handwriting was uniformly smaller under the influence of Ergovine, but for LSD-25, LAS-32, BOL-148, alcohol, methamphetamine and scopolamine, the writing did not have a uniform change of direction.

Anecdotal evidence implies that long term use of substances such as alcohol (de st Colombe 1966: 315; Friedman 1973; Heiss 1950) and drugs (McNichol 1989) can be detected in handwriting.

At least one handwriting analyst has described The Addictive Personality and how it is found in handwriting. Nothing is given as to what the person is addicted to, merely that they will be addicted to something. Whilst such knowledge might be useful to the potential employer, it is legally questionable to include such information in a profile report for employment purposes, especially since the ADA recognizes recovering alcoholics and drug users as a protected class.

Sexual Behavior

Sexual preference is not protected under federal statutes. Nonetheless, Some states have adopted legislation prohibiting discrimination due to one's sexual preference. There are claims that practices such as transvestism (Olyanova 1969), sadism (de st Colombe 1966: 318), homosexuality (de st Colombe 1966: 318) and lesbianism can be detected from handwriting. The analyst could easily include such data to their reports, as an added service. In so doing, both the analyst and the potential employer risk violating the ADA.

Fair Credit Reporting Act Compliance.

Gary Brown first brought this to the attention of the readers of Handwriting-L and GraphoDigest.English in 1999. In early 2001, he announced that the FTC did not currently consider that graphological reports would fall under the new definitions and proposals.

In theory, the only changes in his/her practice an ethical analyst will have to make, is to ensure that his/her clients provide job applicants with his/her corporate name and address and that the consent form was understood by the applicant, when it was signed.

The specific part of the law that handwriting analysts might fall under, is in the definition of Credit Reporting Agency, and of Consumer Investigative Report.

Reviewing the current websites of the alleged two largest firms of the country -- HuVista, Inc and Handwriting Research. Inc (http://www.handwriting.com), it appears that third party analysis, without consent of the subject is an acceptable practice to them. When the FTC next expands the scope of the FCRA, such practices will most likely be prohibited.

A review of the websites of AAHA, AHAF, IGAS, ASPG, AQG, SHA and COGS reveals nothing about the FCRA. Regardless of whether or not the FCRA does apply to handwriting analysts, other laws such as which require an employer to provide a job applicant to have copies of all data collected about them, as part of the hiring process.

Polygraph and Honesty Testing

The Employee Polygraph Protection Act specifically excluded handwriting analysis, when it prohibited polygraph testing in the hiring process. It does, however, allow states to enact more restrictive statutes.

These are the killer statutes, in terms of the legality of handwriting analysis. Rhode Island effectively banned the use of handwriting analysis in 1995, when it revised General Laws 28-6.1. In specific section 28.6.1-4:

"Lie Detector Test" Defined . -- As used in this chapter the term lie detector test shall mean any test utilizing a polygraph or any other device, mechanism, instrument, or written examination which is operated or the results of which are used or interpreted by an examiner for the purpose of purporting to assist in or enable the detection of deception, the verification of truthfulness, or the rendering of a diagnostic opinion regarding the honesty of an individual.

At least one non-lawyer, non-graphologist has interpreted this to mean that handwriting analysis is not legal to use, as part of the hiring process, in the state of Rhode Island.

During 1992 to 1994, Oregon was the focus of an attempt to ban handwriting analysis. Currently it is not explicitly prohibited. The actual ORS statute reads:

659.227 Requiring breathalyzer, polygraph, psychological stress or brain-wave test or genetic test prohibited; exceptions.
... (3) as used in this section: ...
(c) Polygraph examination or psychological stress test means a test to detect deception or to verify the truth of statements through the use of instrumentation or mechanical devices.

As it is, the ORS does not define brain-wave test. Combining this lack of a definition, with the claim that handwriting is brainwriting (Matousek 1987:1), one could claim that handwriting analysis is a brain-wave test, and, as such, prohibited under this statute. Supporting evidence would be Pophal's (1939a, 1939b, 1940, 1966, 1968) research on handwriting and the brain. This legislation is an implicit prohibition of handwriting analysis. With the increasing use of computers by handwriting analysts, and the marketing of handwriting analysis software to HR departments, it is entirely possible that handwriting analysis will be construed as a mechanical lie detector test and challenged accordingly.

In fairness, it must be pointed that Taylor & Sackheim (1988) does advise against providing any services that might be construed as honesty testing.

Criminal Behavior

There has been a long history of looking at handwriting to determine criminal behavior. Adler (1929), Weiser (1930), Weiser (1933), Weiser (1952), Unger (1931), Unger (1936), Spanjaard (1931), Schrijver (1932) and Schrijver (1933) are just a few early citations. Gabriel-Polsterer (1959) demonstrates how inaccurate handwriting analysts are, in differentiating between criminals, and mentally disturbed individuals.

Cole (1961-1968: 186) writes:

Knowing whether the person whose writing you are analyzing is honest or dishonest helps greatly in evaluating the total personality. it is also invaluable to your client who have you do personnel analyses.

Cole (1961-1968: 186) states:

Bear in mind, honesty and dishonesty are not determined by a simple glance at the script, nor by finding one certain stroke to denote whether or not those qualities exist in the character of the scriptor. It is only through evaluating a number of indicators that integrity of character can be determined .... it requires several indices in a given writing before we can say that a person in lacking in integrity. Emphasis is in the original.

Cole (1961-1968: 188) further writes:

If the Slow Writing Score is low, you will have to shift to Spontaneous Dishonesty as shown in Fast Writing, which is covered in Lecture 20. It is also necessary to have a Yes Score in at least four indices before we can diagnose the writing as indicating dishonesty. Each indicator is scored YES if above average, and NO if below average. Emphasis is in the original.

Cole (1961-1968: 188) further writes:

Open Bases of a's and o's

This is the one indicator that is all that you need to determine a person lacking in moral values. It requires slowness to produce this type of letter formation.Emphasis is in the original

Points to question:
Cole (1961-1968) has yet another quote on page 188, that more than one indicator is needed to score dishonesty. Nonetheless, he proceeds to claim that this single indicator is enough to determine a person lacking in moral values.
There is not a single study that demonstrates that an open base of any letter indicates dishonesty. More than one person has commented that despite extensive work with prison populations, that have not seen this formation.

This is an example of how little respect handwriting analysts have for the scientific method, their craft, or the subject of their report.

Business Necessity

One defense to using a tool that discriminates, is that there is a business necessity. For example, Hooters can hire females as wait persons, and not hire males, due to their business model. But, Hooters has to ensure that both males and females have an equal chance to be promoted. This is not a situation that handwriting analysts are likely to be found in.

In passing, Gilliland (1981: 1B) states that first step for a successful evaluation is to:

Obtain age, sex, handedness, nationality, occupation (present or desired), extra training or education, where educated (public or private school and country), physical history (present health, medications, eye glasses, disabilities, surgery, etc), socio-economic status, marital status, children.

Under the EEOC Guidelines, one may ask what the desired occupation is. Asking about the other datapoints can be construed as intent to violate the EEOC.

With just that information, cold reading describes itself as being "hot reading". A good demographer can make accurate statements about what one eats, drinks, watches on TV, or drives, using less information.

Other Tools

From a historical perspective, legislation has consistently been passed, which prohibited a company from using non-job related tests, as part of the hiring process. Thus, instead of using handwriting analysis, a reading test might be more appropriate for telemarketers. Finding a justification for handwriting analysis, under these circumstances, is, at best, non-trivial.

Adverse Publicity

One of the final questions that should be asked, is whether the risk of adverse publicity both acceptable, and worth the potential gain by using handwriting analysis. This is something that can only be answered by the board of directors of the corporation. A legitimate question could be why will the use of handwriting analysis result in adverse publicity? The simple answer is that in contemporary American culture, handwriting analysis has no legitimacy, and there does not appear to be any attempt on the part of any of the organizations of handwriting analysts to address the reasons why handwriting analysis is considered to be "junk science" practiced by con artists and swindlers.

Shareholders

I'm not aware of any reports on shareholder response to the use of handwriting analysis. In passing, it should be noted that many handwriting analysts do not provide the names of their clients, claiming client confidentiality.

Mark Hopper, of Handwriting Research Consultants claims that his clients have said that their use of his services is something that they consider to be a strategic, competitive advantage.

Job Candidates

Studies in France, and the United States, have indicated that job applicants neither trust handwriting analysis, nor believe them to be accurate. (Steiner & Gilliland: 1996; Bruchon- Schweitzer: 1987 - 1988; Bruchon-Schweitzer: 1990; Bruchon-Schweitzer & Ferrieux: 1991)

Anecdotal evidence indicates that job candidates are less hostile to graphological analysis, if they are given a copy of their report, regardless of whether or not they are hired. This practice would appear to have benefits for all parties in the hiring process.

Labor Unions

The only thing to go on, is the AFL-CIO Statement that handwriting analysis is intrusive, invalid and ought to be banned.

The Media

An old news adage is: if it bleeds, it leads. Applying that to handwriting analysis, the conclusion is that position the media will take, depends upon which one will increase their sales.

Man In The Street

I haven't seen any recent surveys on the topic. A perusal of UseNet implies that handwriting analysis is not favorably thought of. Browsing through the archives of mailing lists on the internet, on calligraphy and related topics, reinforces this position.

Conclusion

There is no clear cut evidence that handwriting analysis is valid. There are several things that indicate that the use of handwriting analysis is questionable, for the use of personnel selection. The author of this paper is not an attorney. Do consult with a competent attorney for legal advice. Do note that both the legislation, and the interpretation of that legislation varies from legal jurisdiction to legal jurisdiction. Furthermore interpretations can, and do change from week to week.

If one does use graphology in an employment setting, having weekly sessions with an experienced attorney, who specializes in Employment Legislation is an absolute must. Even then, one must be aware that sometimes errors will be made, in how the law does affect one's practice.

 Previous Page  Up One Level  Next Page
Navigation | Site Wide Table of Contents
Graphology-L FAQ | Graphological Glossary
Graphological Bibliography | Newsfeeds
Section Index | Legal Issues
Professional Services | Copyright Info

Copyright © 1995 - 2001 Jonathon Blake
Distributed under the GNU Free Documentation License.