By Bruce Bower

The landscape of an individual’s personality has much in common with one of those 19th-century impressionist paintings made up of thousands of colored dots. From across the room, an art lover gazing at such a framed creation sees sunbathers lolling by a lake, or perhaps a circus scene. Personality features pondered from a suitable distance also seem to coalesce into familiar forms. For example, a coworker comes across as friendly and fun loving, while an office supervisor always seems grouchy and distracted.

From up close, however, so-called pointillist paintings drive observers dotty. Beautiful scenes crumble into a crazy quilt of tiny tinted flecks. Close-up scrutiny of someone’s personality can prove just as disorienting. Solid-looking dispositions dissolve into pools of often-contradictory desires, feelings, and habits.

Consider that hale and hearty coworker. He may cozy up to lots of folks because he loves social contact and craves his peers’ approval. Or perhaps his chummy behavior masks discomfort around others and a deep-seated need to manipulate them for his own ends. If the latter proves true, is he more shy than gregarious, or vice versa?

Drew Westen, a psychotherapist at Harvard Medical School in Boston, takes the close-up perspective in treating people whose personalities have gone awry. Like most psychotherapists, Westen relies on clinical experience and intuitive guesswork to do his job. Current methods of classifying personality and its disturbances offer clinicians little practical help, he contends.

So, Westen and Aspen, Colo., psychotherapist Jonathan Shedler, both of whom are also research psychologists, have developed their own personality-assessment tool.

They hope their technique will inspire psychiatrists to revise and refine official diagnoses of personality disorders. Westen and Shedler also would like to break psychologists from the habit of using volunteers’ questionnaire responses to map out personality characteristics. This popular approach has yielded a handful of traits that scientists are trying to link to various genes. Its proponents, however, have failed to generate any theories of how personality develops and what purposes it serves, Westen argues.

In his view, questionnaire responses tap into a person’s self-concept and social reputation without addressing the deeper organizing principles of his or her personality.

Not surprisingly, some personality researchers disagree. Westen and Shedler, however, are targeting their method to psychotherapists. These practitioners could use a tool that would act as a scalpel to cut through the self’s outer hide to explore personality’s inner workings, Westen and Shedler say.

Many clinicians find fault with commonly used systems for personality classification. Psychotherapists disparage as laundry lists of symptoms the personality disorders described in psychiatry’s official diagnostic manual. These categories mainly exist for the convenience of insurers that cover mental health care, they say. Moreover, clinicians frequently denigrate the handful of personality traits studied by psychologists as statistical entities that only skim personality’s surface.

Behavioral researchers return the favor by tending to brand psychotherapists as hopelessly subjective, mistake-prone in their judgments about clients, and willfully ignorant of scientific advances.

Westen hopes his new personality measure, which combines clinical experience with hard-nosed statistical analysis, will ease tensions between mental-health practitioners and scientists. However, he notes, it could just as easily alienate both camps.

"It’s hard to sustain the view that fundamental aspects of personality can be found by asking people direct questions about themselves and others on questionnaires," Westen says. "We need to analyze in a scientific way the observations of personality experts, such as experienced clinicians, who listen to the stories people tell about their lives and watch their actual interactions."

It’s time for some old ideas about personality to step boldly into the scientific spotlight, say Westen and Shedler. At the beginning of this century, theorists such as Sigmund Freud and George Herbert Mead argued that personality—enduring ways of behaving, perceiving, and thinking about oneself and others—arises through social interactions, beginning with the mother-child relationship. Encounters with others always carry underlying meanings that occur largely outside of awareness, in this view.

Mental life consists of warring impulses, ambivalent feelings, and a generous capacity for self-deception, these thinkers maintained. It thus takes hard work and courage to confront the nitty-gritty of one’s own personality or that of someone else.

Researchers began to turn away from this Freud-inspired perspective by mid-century. Psychologists embraced a statistical method known as factor analysis, which allowed them to group thousands of adjectives deemed descriptive of personality into coherent clusters or traits. Factor-analysis proponents assume that languages build up reservoirs of adjectives related to important personality traits, so these traits should pop into relief when scientists give the welter of words a vigorous statistical massage.

Factor analysis ran into early problems, however, as different investigators generated different numbers and types of personality traits. In the past decade, however, an influential group has championed what psychologists call the five-factor model.

As research psychologists had earlier rejected the Freudian perspective, psychiatrists in 1980 turned away from early approaches to personality. The American Psychiatric Association sanctioned an extensive revision of its official diagnostic manual of mental diseases. For the first time, it included separate diagnoses of personality disorders. The diagnoses rest on sets of specific, clinically observed symptoms.

The new symptom clusters then qualified as medical conditions, distinct both from other psychiatric disorders and from what are considered normal personalities. The current diagnostic manual describes 10 personality disorders and mentions 2 others deemed worthy of further study.

For instance, the manual portrays paranoid personality disorder as an ingrained distrust and suspiciousness of others’ motives, while the symptoms of histrionic personality disorder include excessive emotional displays and constant attention seeking. Several influential researchers have proposed that, in its next version, the manual should use traits from the five-factor model to classify personality disorders (SN: 3/5/94, p. 152).

Despite its statistical rigor, the five-factor model operates on the dubious assumption that armed with enough adjectives, anyone can decode his or her own personality or that of a friend or family member, Westen asserts.

Psychologist Jack Block of the University of California, Berkeley agrees with Westen. He maintains that factor analysis can’t make sense of the disparate assumptions that typical volunteers have about adjectives in personality questionnaires. For instance, people may disagree whether the term disagreeable refers to being hostile, appearing self-absorbed, or tending to pry into other people’s business.

Of more practical concern to clinicians, knowing a patient’s five-factor status—such as being rated low on agreeableness, relatively closed to new experiences, and high on neuroticism—provides little guidance about how to structure treatment, Westen maintains.

Westen and Shedler presented their alternative classification system in the February American Journal of Psychiatry. Over the past 7 years, they have refined it for use as a clinical tool to describe patients’ personalities in a quantifiable form. Although the two clinicians probe personality from a psychodynamic, or Freudian-inspired, perspective, psychotherapists of any theoretical persuasion can use the new technique, they add.

Their system covers three major realms of individual personality. First, it explores the mental resources at a person’s disposal. These include the ability to express and to rein in emotions and the tendency to deploy unconscious strategies for self-protection, such as perceiving one’s own unsavory traits in others. Second, it addresses a person’s desires, fears, and values and whether these motives are consciously appreciated and compatible with one another. Third, it examines a person’s view of self and others and his or her ability to form fulfilling intimate relationships.

At the heart of this procedure lie 200 personality-related statements printed on cards. Dipping into his or her experience with a patient, a clinician ranks each statement from 0 to 7, signifying the degree to which it describes a particular patient, and then places the card in one of eight piles. Depictions include "tends to feel empty or bored," "tries to manipulate others’ emotions to get what he/she wants," "is able to understand self and others in subtle and sophisticated ways," and "tends to be overly sexually seductive or provocative, whether consciously or unconsciously."

Each of 530 randomly selected psychiatrists and psychologists from throughout the United States—who had an average of 18 years’ experience as psychotherapists—used Westen and Shedler’s instrument to describe a current patient who met diagnostic-manual criteria for a personality disorder and whom they had seen five times or more.

A statistical sorting of these patients based on the similarity of their personality profiles yielded 12 personality dimensions that differ in many ways from the personality disturbances emphasized in the diagnostic manual, the investigators contend.

Moreover, these dimensions delve into corners of personality left unexplored by the five-factor model, according to a statistical comparison of the two approaches conducted by Westen and Shedler.

Examination of these dimensions promises to illuminate the many ways in which personality develops and the means by which this process can foster psychological ailments, Westen holds.

For instance, in an unpublished study, he and a colleague used the new personality instrument to tease out three strands of disturbance in 104 young women suffering from eating disorders. One group of impulsive, highly emotional, and openly distressed women mainly exhibited bulimia’s binges and purges or veered back and forth from bulimia to anorexia. A second group kept a lid on feelings, sexual urges, and any behaviors that might lead to pleasure, avoided other people, and usually clung to the self-starvation typical of anorexia. The third group functioned well at work and with others despite bouts of bulimia or anorexia and high levels of self-criticism.

"Knowing that a patient has bulimia or anorexia may be much less important for designing treatment than understanding how these symptoms function in the context of an individual’s personality," Westen contends.

His new assessment method may also shed light on the aftermath of severe childhood maltreatment. By young adulthood, victims of physical abuse, sexual abuse, and neglect suffer from greatly elevated rates of personality disorders, according to research directed by psychologist Jeffrey G. Johnson of the New York State Psychiatric Institute in New York City. Johnson’s team presents its findings in the July Archives of General Psychiatry. Westen suspects that his clinical aid will generate a deeper understanding of personality problems related to child abuse.

Five-factor advocates, such as psychologist Thomas A. Widiger of the University of Kentucky in Lexington, shrug off Westen and Shedler’s criticisms.

"They want to classify personality by relying on clinicians’ perspectives and returning to a psychodynamic approach," Widiger says. "I see that as a step backwards."

He views as more promising the ongoing five-factor research that’s examining that model’s ability to illuminate individual personality differences and to improve the diagnosis of personality disorders.

In contrast, Block welcomes Westen and Shedler’s approach. Block published a critique of the five-factor model in the March 1995 Psychological Bulletin. Researchers have often used factor analysis in an arbitrary way, he argues. He holds that helpful personality measures will emerge by examining the insights of experienced clinicians and other seasoned personality observers, as Westen and Shedler do.

"Westen and Shedler’s personality model is a breath of fresh air, but it needs to be evaluated against other approaches," remarks psychiatrist Robert L. Spitzer of the New York State Psychiatric Institute, who directed the 1980 revision of the manual of psychiatric diagnoses.

Spitzer is currently coordinating a project aimed at comparing the clinical usefulness of Westen and Shedler’s method with that of the five-factor model and of two other personality measures—one containing four temperament and three character dimensions, the other consisting of 15 personality traits. Clinicians will use the competing instruments to formulate and carry out treatment plans for patients diagnosed with personality disorders and then chart their progress.

For now, nearly everyone agrees that personality disorders as currently defined are "ripe for an overhaul," says psychiatrist John G. Gunderson of McLean Hospital in Belmont, Mass. "It’s just not clear whether one or another approach to personality assessment should be dominant."

May the most conscientious, agreeable, and best-adjusted system win.

References & sources for this article

From Science News, Vol. 156, No. 6, August 7, 1999, p. 88. Copyright © 1999, Science Service.