personality Types or personality Traits: what … the difference between a trait view of personality and a type ... leadership ability, ... behavior such as compulsiveness. Type theory

personality Types or personality Traits:what Difference Does It Make?by Naomi L. Quenk

I asked Naomi Quenk to write thisarticle for the Bulletin because I believeunderstanding the difference between atrait view of personality and a typeview of personality is crucial to ourethical and wise use of psychologicaltype. Psychology in the United States isdominated by trait approaches, and theyare so much a part of our picture ofhow human beings function that we usethe assumptions and explanations oftrait theory without thinking aboutthem. They frequently creep into howwe think and talk about type, distortingthe meaning of type preferences andwhat we can attribute to psychologicaltype preferences. My hope is that all ofus who use psychological type willstudy this issue and use the under-standing to evaluate the way we presentand use type. - Linda Kirby

Perhaps the greatest obstacle toeffective applications of type theoryand the MBTI is that the majority oftype practitioners are largely una-ware of the critical differencesbetween typological and trait psy-chology approaches to personality.This article explains the importantdifferences in these two ways ofunderstanding personality, andillustrates the negative effects ofconfusing them in applying [ung'stheory and the MBTI.

Trait psychology and type psy-chology embody two very differentapproaches to personality. Both arereasonable explanatory systems, andwhether one is "better" than theother is irrelevant to this discussion.What is relevant, however, are thedetrimental consequences of confus-ing one with the other, in treating atype theory as though it were a traittheory or a trait theory as though itwere a type theory. My focus hereis on the former error, since muchof the misunderstanding and conse-quent misuse of type theory and

the Myers-Briggs Type Indicator isdirectly related to this misin-terpretation.

Predominance of trait approachesPersonality theorists, particularly inthe United States, favor traitapproaches for at least two reasons.First, they often associate type the-ories with attempts to explain perso-nality by body-type (Kretschmer,1961; Sheldon, 1942), and thenreject type as a static system thatdoes not reflect the richn~ss anddiversity of human perso~ality.

They also prefer trait approachesbecause traits can be studied with awide range of statisticaltechniques - all those methodsbased on the assumption of a nor-mal distribution and continuousscores (means, standard deviations,etc.). Many fewer methods are

If we measured height, weight, age,and IQ for a large number of peo-ple, we would find that few peopleare very tall or very short, veryheavy or very light, very old or veryyoung, very intelligent or very dull;most people would be average onthese traits, obtaining some "middle"score on what is being measured.

We can also describe peoples' per-sonalities in terms of traits. We maysay, "She is a very optimistic per-son;' or "He is quite domineering;'or "Most of the time he is rigid andanxious:' If we wanted to measuresuch traits objectively, we mightdevise a set of questions that wouldreflect the traits of Optimism,Dominance, Rigidity, and Anxiety.

If we created these trait scales andadministered them to large numbersof people, we would find that mostpeople have pretty moderate amounts

Perhaps the greatest obstacle to effective applicationsof type theory and the MBT! is that the majority oftype practitioners are largely unaware of the criticaldifferences between typological and trait psychologyapproaches to personality.

appropriate for doing statisticalresearch with type categories andthey often require much larger num-bers of subjects than do traitstudies. This means that typeresearch is more time consumingand more difficult.

Trait-based theories andinstrumentsHuman characteristics like height,weight, age, and intelligence (IQ)are traits. Everyone has a height,weighs a certain amount, is a partic-ular age, obtains a particular scoreon an IQ test. We merely differ inhow much of a trait each of us has.

of these traits, with a relativelysmall number reporting a great dealof Optimism, Dominance, Rigidity,and Anxiety, or very little of thesequalities. Thus personality traits,like other trait measures, tend to benormally distributed in the popula-tion. Statistical research on normaldistributions indicates that about68% of the people tested score inthe middle range, with a decreasingpercent scoring by the extremes.

Inferences promoted by traitapproachesWe tend to notice people who areextreme on something, whereas

Bulletin of Psychological Type 16:2 Spring 1993

The extreme person is different and therefore"outside of the norm," not "normal"; so it iseasy to make a value judgement about peoplewho are not average on a trait.

"average" people are unlikely toattract our attention. The extremeperson is different and therefore"outside of the norm;' not "nor-mal'; so it is easy to make a valuejudgment about people who are notaverage on a trait. They may beseen as abnormal, and their abnor-mality may be either positive ornegative. There are advantages ordisadvantages to being very tall orvery short, very fat or very thin,older or younger. When it comes toIQ however, one extreme may bevery much preferred. It is clearlybetter to be more rather than lessintelligent!

Similar qualitative comments maybe made for personality traits. Wecan see something wrong withbeing very optimistic (perhaps theperson is not accepting reality) orwith having too little optimism andtherefore being pessimistic anddepressed. We may want to avoidvery domineering people who areoverbearing and intrusive, as wellas people with minimal dominance.

qualities as creativity, leadershipability, and insightfulness, may beamong these. So we see thatimplicit in a trait approach are 1)universality of the attribute witheveryone having at least some of it,and 2) often an explicit or implicitvalue attached to differing amountsof the trait.

Another important feature of traitapproaches lies in the view thattraits cause behavior. We act in cer-tain ways because we have someunderlying trait that activates a par-ticular response or level ofresponse. This is similar to Freud'scausal approach where symptomsare caused by an early event result-ing in a character trait (e.g., "analretentiveness") that in turn causes abehavior such as compulsiveness.

Type theory and the MBTIType theories characterize peopleaccording to certain qualitatively dis-tinctive categories. In common lan-guage, we may describe someone asan intellectual type or a cooperative

... implicit in a trait approach are 1) universality ofthe attribute with everyone having at least some ofit, and 2) often an explicit or implicit value attachedto differeing amounts of the trait.

They may be ineffectual. Very rigidpeople are likely to have limitedperspectives on life, but people atthe opposite extreme may be toounstructured and thereforeirresponsible and undependable.Extreme anxiety can immobilizesomeone, but people with littleanxiety may be unmotivated andwithout direction.

Sometimes, however, it is betterto be at one extreme rather than theother on a personality trait. Such

type. Where a polar oppositecategory exists, we don't expect aperson to be appropriatelydescribed by both of the two oppo-site poles. Thus a cooperative typewould not likely also be called anantagonistic type.

Type dimensions such as thoseon the MBTI are not assumed tocharacterize all people withindividuals differing merely in theamount of "type" they possess.Each dimension has two opposite

poles - a person characterized byone pole is qualitatively differentfrom a person characterized by theopposite pole. An apple and anorange are qualitatively differentfruit. They do not merely differalong some dimension of "fruiti-ness:' Similarly, you are either anExtravert or and Introvert, a Sensingtype or an Intuitive type.

Unlike the fruit analogy, however,type theory allows for the use of aless preferred pole by the personpreferring its opposite. For unlikethe static, relatively immutablecharacter of the "fruit typology;'psychological typology is a dynamicsystem embodying complex move-ment and interaction among ele-ments. An apple cannot suspend its"appleness" in order to temporarilytaste like a peach! But a personwho prefers Thinking can suspendit and make a Feeling decision. Thisfeature of typology is in accord withthe notion of qualitative differencesamong types that are by naturecomplex and dynamic wholes.

How types are qualitativelydifferentIn what sense, then, are typepreferences "qualitatively differentcategories"? First, they are polaropposite mental activities, whichmeans that we cannot do one at thesame time that we do the other. Wecannot focus on the outer environ-ment and our inner reflections, orattend to concrete reality and futurepossibilities, simultaneously. But wecan do these opposites consecu-tively, usually our preferred onefirst and the less preferred second.

Over and beyond this technicaldefinition of polar opposites is theissue of how these qualitativelydifferent categories emerge anddevelop within us. For we do notneed to discriminate between sim-ple categories like apples versus

ElO Bulletin of Psychological Type 16:2 Spring 1993

... personality traits will naturally form as theresult of the habitual exercise of particular type

preferences and/or one's whole type.

peaches, but rather explain somekind of consistency and predictabil-ity that result in highly complex anddistinct human beings.

Perhaps a conception of thehuman psyche that views typologyas a template upon and withinwhich other relevant factors interactcan be applied here. Thus, innatetype as a dynamic whole and theindividual preferences themselvesmay serve as a central core ororganizing principle for significantlife experience.

Within such a structure, personalitytraits will naturally form as theresult of the habitual exercise of par-ticular type preferences and/or one'swhole type. The traits involved arelikely to be more or less related toeach other since the underlyingpreference or interaction of prefer-ences will serve as the "glue" thatmakes them cohere.

It is commonplace to identify typedifferences by describing consistentsets of traits associated with them.We do this in our everyday observa-tions of type and in the multitudeof research studies available. Often,however, research examines onetype dimension at a time ratherthan exploring the character ofwhole types. This is partiallybecause of the difficulties men-tioned above in doing statisticalresearch with type. In order to doresearch using standard trait-basedstatistical methods, researchers sus-pend the basic premise of indepen-dent type categories and convertMBTIdata into continuous scores.Those who are aware of the basicassumptions of type theory are cau-tious in interpreting the results ofthis kind of research and take careto interpret within the context oftype theory. Those who faiJ torecognize the theoretical context oftype theory interpret their researchresults in trait terms and thereby

come to erroneous conclusions.But even research performed in

accord with type theory largelyignores the dynamic character oftypology in both the design ofstudies and testing and interpreta-tion of results. The consequence is afailure to make the most of theexplanatory ability of the theory. Anotable exception to this is therecent research compendium byThorne and Gough (1991), whichlooked at personality descriptors forwhole four-letter types. (See myreview of their book in this Bulletinissue.)

The normal distribution and normsdo not applyThe important point to make here isthat, unlike traits (induding thoseassociated with type preferences),

indicate a preference. So rather thanmeasuring the amounts appropriateto traits, scores on the MBTIareused as an estimate of the accuracy ofour sorting procedure. Given manyopportunities to indicate a prefer-ence, the respondent to the MBTIindicated a very clear, clear, moderate,or slight preference for one or theother pole of a dimension. Thus incontrast to scores on trait measures,MBTIscores do not mean the per-son has more or less of the dimen-sion under study. The scores merelyreflect how confident we can be thata preference has been correctlyreported.

Behavior as an expression of typeIn a type approach, behavior is notseen as caused by one's type. Rather,behavior (often through the vehicle

. .. type is not nonnative: there is no "normal" or"best" score to obtain or type to be.

type dimensions are not normallydistributed in the population. Forexample, when preference scores (orcorresponding continuous scores)on the Extraversion-Introversiondimension are arrayed, theyapprox-imate a bimodal distribution. Thisallows interpretation of two dis-tinctly different kinds of peoplecharacterized by each pole of adimension. We assume that mostpeople will score at or near eitherpole (end) of the distribution, notthat the majority will score in themid-range of low Preference Scores.

Another important distinction isthat type is not normative: there isno "normal" or "best" score toobtain or type to be. The scale foreach dimension has many questionsin order to provide a respondentwith numerous opportunities to

of traits) is an expression of type. Incontrast to the Freudian reduc-tive/causal approach, it is consistentwith Jung's prospective, purposiveapproach. The purpose or goal oftypological expression relates to ourinstinctual push toward completion,which is the final stage of theIndividuation process. Althoughtrait approaches are not necessarilyinconsistent with a prospective, pur-posive approach to personality, theyare not commonly viewed from thatperspective but rather within acausal framework.

Type dimensions such asExtraversion, however, are oftentreated as traits within a trait the-ory. In fact, every major personalityinstrument in use includes a meas-ure of extraversion/introversion. It istypically treated as a trait. One

Bulletin of Psychologiall Type 16:2 Spring 1993

In a type approach, behavior is not seenas caused by one's type. Rather, behavior(often through the vehicle of traits) isan expression of type.

example of such a trait-based instru-ment is the Cattell 16-PF (1970).Intrait approaches, a continuum isassumed such that, for example, aperson with a low score would beseen as '1ess extraverted" than aperson with a higher score. In sucha system, we might explain some-one's behavior as a function of theirdegree of Extraversion. We might say,"You like having many peoplearound you because you are high onthe trait of Extraversion:'

In type theory, a statement ascrib-ing cause would not be appropriate.The correct statement would be,"Your enjoyment in having manypeople around you is an expressionof your preference for Extraversion;it is one way in which you demon-strate your Extraverted attitude:' Orperhaps more theoretically, "yourlife experience has been filteredthrough an Extraverted templatethat has encouraged development ofmany traits associated withExtraversion, your enjoyment ofmany people being one of them:'

What happens when people treattype dimensions as traits?The misuse of typology and theMBTI follow from the differencesoutlined above.

Effects of assuming a normaldistributionWhen we assume that type dimen-sions are normally distributed, wecan readily infer that there is a nor-mal or "good" score to obtain. Wethus give meaning to the numericalscores whose primary purpose is tosort people into equally desirablecategories. Erroneously assuming anormal distribution also encouragesus to speak of "amounts" of a typedimension rather than qualitativelydifferent type preferences. Here isan example to illustrate the effectsof this kind of error.

If an Extraversion-Introversiondimension is treated as a trait, peo-ple merely differ in how muchextraversion (or introversion,depending on how the scale is con-structed) they demonstrate. Intro-

version could then be defined as adiminished amount or an absence ofExtraversion (or vice versa). However,if Extraversion and Introversion arequalitatively different modes ofbeing, we cannot define one asmerely less or the absence of theother. Rather, we are dealing with acoherent set of qualities that defineExtraversion, and a different set ofqualities that define Introversion.Our descriptors will likely be men-tal and behavioral traits that resultfrom habitually having one's energyflow outward (Extraversion) orinward (Introversion). But we willnot speak of a "degree" of outwardand inward flow or a degree ofExtraversion or Introversion.

Inferences about skills and pathologiesIf I am "more Intuitive" than you,are my intuitions better than yours?Could you do something to raiseyour Intuition to the level of mine?If you get a higher preference scorefor Intuition the next time you takethe Indicator, does that mean you

The following summarizes the essential differences between trait and type theories:

TRAIT THEORIES

I Universals differing only in amount possessed.

Involves measuring amounts.

Normally distributed.

Extreme scores important for discrimination.

Scores show amount of trait possessed.

Behavior is caused by traits; a reductive approach.

Too much or too little is often negative ordiagnostic.

TYPE THEORIES

Qualitatively distinct inborn preferences.

Involves sorting into categories.

Bimodal! skewed distributions.

Midpoint separating categories important fordiscrimination.

Scores show confidence in the sorting procedures.

Behavior is an expression of type; a purposiveapproach.

"Too much" or "too little" is irrelevant.

Bulletin of Psychological Type 16:2 Spring 1993

What is typical or "normal' for one typemay be quite out of character for another.

are more intuitive than you werebefore?Would an employer whoruns an intuitive company be betteroff hiring me than you because I am"more intuitive" than you?

When we speak of "amounts" inrelation to type dimensions, wereadily fall into erroneous interpre-tations of differential skill levels andother "surplus meanings:' If you are"very introverted;' you probablylack social skills, dislike people, andare inhibited, as compared to some-one who is only slightly introvertedand probably has fewer deficienciesin these areas. Is someone with avery clear Preference for Judgingover Perceiving "very judgmental;'and should I rather favor the personwith a clear score on Perceptionbecause she must be "very percep-tive"? Should I be wary of peoplewith low preference scores becausethey may lack self-awareness or con-fidence in themselves? Or should Irather regard them highly as "flexi-ble and able to go both ways"?

As you can see, treating typedimensions and scores on the Indi-cator as traits leads us to incorrectattributions of skill and permits usto pathologize or overvalue one orthe other extreme score or the mid-dle score, depending on what othererroneous notions we hold abouttype theory and the MBTI.We maythink it is best to be in the middle ifwe are looking for the "normality"associated with trait measures; orwe may opt for one of the extremesif we assume more of less of sometype dimension is better than "aver-age:' Sixteen "norms" versus onenonn. Perhaps the most negativeeffect of interpreting type from atrait perspective is the lost opportu-nity to describe, appreciate, andassess individuals from the perspec-tive of their own character structure;for attitudes, motives, and behaviorsare seen in entirely different ways

within the context of one type ascompared to another. What is typi-cal 01' "normal" for one type may bequite out of character for another.

It is type-consonant for anIntroverted child to spend much ofhis recess alone reading a book;when an Extraverted child spendsmuch of his free school time thatway, one might be wise to pay spe-cial attention to his state of mind.But from a single-norm perspective,both children might be seen as shy,unsociable, and maladjusted. Simi-larly, taking control and organizingpeople is expected as an expressionof type for Extraverted Thinkers,regardless of gender. But in ournormative culture, women whodemonstrate such behavior may beviewed as inappropriately aggressiveand over-controlling.

Myriad examples of the effects ofa single-norm approach on differenttypes can be cited. Some clues tothe negative consequences from thisnormative approach are shown insome of the correlational datareported in the MBTIManual (1985),such as the consistent associationbetween Introversion and variousmeasures of neuroticism andmaladaptation. A more recent reportis available in the Thome andGough book previously cited. Inthat research, MBTItypes weredescribed by themselves as well asobservers using assessment instru-ments based on the single-normapproach. It was inevitable, there-fore, that some types weredescribed rather negatively overall,while others were portrayed as well-adjusted and psychologicallyhealthy.

Many people who use the TypeIndicator are aware that it differsmarkedly from other personalityapproaches - but most practitionersare not clear about the profoundand critical nature of those differ-

ences. The goal of the foregoing dis-cussion was to provide a usefulexplanation of type theory's essen-tial characteristics and therebyencourage accurate understandingand more effective applications.

ReferencesCattell, R.B., Eber, H.W, & Tatsuoka, M.M.

(1970). Handbook for the Sixteen PersorllliityFactor Questionrlllire (16-PF). Champaign,IL: Institute for Personality and AbilityTesting.

Kretschmer, E. (1%1). The Psychology of men ofgenius. Trans. by R.B. Cattell. London:Routledge & I<egan Paul.

Myers, LB., & McCaulley, M.H. (1985).Manual: A guide to the development and useof the Myers-Briggs Type Indicator. PaloAlto: Consulting Psychologists Press.

Sheldon, WH., & Stevens, S.S. (1942). Thevarieties of temperament: A psychology of con-stitutional differences. N.Y.: Harper.

Thome, A. and Gough, H. (1991). Portraits oftype: An MIrrI research compendium. PaloAlto: Consulting Psychological Press.

Bulletin of Psychological Type 16:2 Spring 1993