On the Science of Changing Sex

Personality, My Dear…

Posted in Science Criticism by Kay Brown on October 6, 2016

phrenology…I Don’t Give a Damn!

or What is the Rate of Co-morbid Personality Disorders in Transsexuals?

I recently came upon a blog that made much of a single paper out of Iran which purported to show that about 80% of transsexuals had a serious personality disorder.  I was shocked at such a result and had to dig deeper.  Well… this paper seems to be an outlier and its use by this blogger (who fits the profile of an angry ex-wife of an autogynephilic transgender-woman turned transphobic crusader) is an example of cherry picking.  Another paper I found was only 24% of the transsexuals had “subthreshold” potential personality disorders compared to 17% of the controls.  (Note, “subthreshold” does NOT mean that they actually had the disorder.)  But why did this first paper get this outrageous percentage of actual disorders when other studies didn’t?

Because they did NOT diagnose anyone, period.  Instead Meybodi used the Millon Clinical Multiaxial Inventory II (MCMI- II), a self-scored inventory and assumed that if they got a high score on any particular scale that they must then have that clinical diagnoses.  This alone is a gross misuse of any instrument; a test score alone is NOT a diagnoses of a disorder.  From the paper, all we know is that their subjects scored higher than a non-reported cut-off, likely the one used by the publishers to indicate the “presence” of a trait.  The researchers fully admit that they did NOT actually interview and determine if the score had any bearing on actual dysfunction.  It is important to note that to have a diagnoses of a disorder, the personality trait must cause significant dysfunction to their lives.  This study failed to evaluate this dimension.  The most common of the purported personality disorders was Narcissistic Personality Disorder at nearly 60%, which given that this paper was from Iran, surprised me greatly.

As Lawrence has shown, the percentage of non-androphilic MTF transfolk is correlated with a given society’s Hofstede Individualism Index, which for Iran is 41, and thus we would expect a very low percentage.  And indeed, anecdotal reports regarding Iran’s MTF population agree.  But we know that from a number of studies and clinical surveys, that exclusively androphilic MTF transsexuals have a LOWER co-morbidity rate.   So what gives?

The answer seems to be that this study failed several basic tenets of science.  First, they failed to provide controls, which had they done so, might have flagged another issue with their methodology, namely that the MCMI-II was written and only validated in ENGLISH and is only valid for those who have at least a 5th Grade literacy level IN ENGLISH.  One assumes that the researchers simply translated the inventory items from English to Farsi and did not conduct a proper re-validation study given the very divergent cultural meanings potentially introduced by this translation?  If so, as Rogers points out, this is a gross abuse,

“Multiscale inventories can be translated into different languages with relatively little effort. The critical issue is that linguistic equivalence (i.e., similar sentences) cannot be equated to clinical equivalence (i.e., similar diagnostic relevance).  Clinical equivalence cannot be assumed, but must be objectively tested. Simple comparisons of vocabulary and syntax (e.g., from English to Spanish) are insufficient to establish clinical validity for translated versions.  An approach to translation validation in which a mere lack of significant group differences between two language or ethnic groups is assumed to mean the tests “work the same way” makes little sense. With depressed patients, for example, the clinician needs to know whether or not depressed persons of different cultures and languages have the appropriate elevations on multiscale inventories. Given our dearth of knowledge regarding translated versions and their cultural differences, psychiatrists and their consultants should be very cautious about using and interpreting translated tests.”

Even if the translation into Farsi was validated, there is the problem of interpretation; a high score on a given scale does NOT necessarily mean that one has a personality disorder.  In fact, emotionally healthy people often have high scores on scales that correlate to Narrissistic Personality Disorder as Stephen Strack explains in his book, Essentials of the Millon Inventories,

“Scale 5 has a research base that suggests that elevated scores indicate either a clinical personality disorder or a healthy adaptional personality style associated with with nonclinical people.  In factor analysis studies, Scale 5 loads positively on items dealing with extroverted traits and behaviors and negatively on items pertaining to maladjustment. … Elevations on Scale 5 are rare in psychiatric samples.  Many nonclinical populations attain elevated scores on Scale 5 including air force pilots in basic training. … Thus the clinical task is to determine whether clinically elevated scales represent a Narcissist Personality Disorder or a narcissistic personality style.  … versions of this scale have not correlated well with structured psychiatric interview schedules.”

Anyone who is familiar with either the literature on, or knows “early onset” / transkids in person knows that they can be quite extroverted, even flamboyantly so, without developing Narcissistic Personality Disorder.  Finally, a study that used structured interviews showed significantly lower co-morbid issues than those studies that used translated personality inventories, from the abstract of the Haraldsen paper,

“Transsexual patients scored significantly lower than Personality Disordered patients on the Global Symptom Index and all SCL-90 subscales. Although the transsexual group generally scored slightly higher than the healthy control group, all scores were within the normal range.  Transsexual patients selected for sex reassignment showed a relatively low level of self-rated psychopathology before and after treatment. This finding casts doubt on the view that transsexualism is a severe mental disorder.”

The conclusion here can only be that we must evaluate the literature on transsexuals and co-morbidity very carefully and critically, not accepting them at face value unless we can determine that they have been conducted with proper methodologies, including proper interview based psychiatric diagnoses, compared against valid controls from both clinically relevant disordered and healthy populations, and shown to be reproducible.  Anything else is just junk science.  {And cherry-picking the worst data you can find to defame transfolk is despicable.}


Maybodi, et Al., “The Frequency of Personality Disorders in Patients with Gender Identity Disorder”

Bodlund, et Al., “Personality Traits and Disorders in Transsexual” (1993)

Rogers, R., “Forensic Use and Abuse of Psychological Tests: Multiscale Inventories”

Haraldsen, et. Al., “Symptom profiles of gender dysphoric patients of transsexual type compared to patients with personality disorders and healthy adults”

Comments Off on Personality, My Dear…