This article was downloaded by: [University Library Utrecht]

On: 20 September 2013, At: 11:14

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK

Multivariate Behavioral
Research
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/hmbr20

Psychometric Properties of the

HEXACO Personality Inventory
Kibeom Lee & Michael C. Ashton
Published online: 10 Jun 2010.

To cite this article: Kibeom Lee & Michael C. Ashton (2004) Psychometric Properties
of the HEXACO Personality Inventory, Multivariate Behavioral Research, 39:2, 329-358,
DOI: 10.1207/s15327906mbr3902_8

To link to this article: http://dx.doi.org/10.1207/s15327906mbr3902_8

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
 expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions
Downloaded by [University Library Utrecht] at 11:14 20 September 2013
 Multivariate Behavioral Research, 39 (2), 329-358
Copyright © 2004, Lawrence Erlbaum Associates, Inc.

Psychometric Properties of the

HEXACO Personality Inventory
Kibeom Lee
University of Calgary

Michael C. Ashton
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Brock University

We introduce a personality inventory designed to measure six major dimensions of personality

derived from lexical studies of personality structure. The HEXACO Personality Inventory
(HEXACO-PI) consists of 24 facet-level personality trait scales that define the six personality
factors named Honesty-Humility (H), Emotionality (E), Extraversion (X), Agreeableness (A),
Conscientiousness (C), and Openness to Experience (O). In this validation study involving a
sample of over 400 respondents, all HEXACO-PI scales showed high internal consistency
reliabilities, conformed to the hypothesized six-factor structure, and showed adequate
convergent validities with external variables. The HEXACO factor space, and the rotations of
factors within that space, are discussed with reference to J. S. Wiggins’ work on the circumplex.

Introduction

In this article, we will introduce the HEXACO Personality Inventory

(HEXACO-PI), a new questionnaire that is intended to measure six dimensions
of personality variation that have been recovered in lexical studies of
personality structure in several languages. After reporting the psychometric
properties of the HEXACO-PI, we will discuss the HEXACO factor axes in
relation to the circumplex, with reference to the work of Jerry Wiggins.

Searching for the Structure of Personality Characteristics

One of the fundamental problems of personality psychology has been to

discover the structure of human personality characteristics. Without an
accurate structural model of personality, researchers who hope to understand

Address correspondence to Kibeom Lee, Department of Psychology, The University of

Calgary, Calgary, AB T2N 1N4 Canada (e-mail: kibeom@ucalgary.ca) or to Michael C.
Ashton, Department of Psychology, Brock University, St. Catharines, ON L2S 3A1 Canada (e-
mail: mashton@brocku.ca).

MULTIVARIATE BEHAVIORAL RESEARCH 329

 K. Lee and M. Ashton

the cause and function of personality variation can only approach these
subjects in a haphazard and piecemeal way. But by finding the basic
dimensions of personality, researchers will be able to investigate the origins
and consequences of personality variation much more systematically.
Throughout most of the 20th century, there was no consensus regarding
the structure of personality. By the 1980s, however, many researchers
began to agree that personality variation was best summarized in terms of
five broad, roughly independent dimensions. These five factors were
discovered in the English language, first in investigations involving a small
number of English personality descriptive adjectives (e.g., Digman &
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Takemoto-Chock, 1981; Norman, 1963; Tupes & Christal, 1961, 1992) and
later in investigations involving a more comprehensive set of personality
adjectives (Goldberg, 1990; Saucier & Goldberg, 1996). Subsequently, these
five factors were collectively named the “Big Five” factors (Goldberg, 1990,
1993), which include (a) Extraversion (or Surgency), (b) Agreeableness, (c)
Conscientiousness, (d) Emotional Stability (versus Neuroticism), and (e)
Intellect/Imagination. These lexical dimensions were incorporated into
personality questionnaire research by Costa and McCrae (1985; McCrae &
Costa, 1985), who proposed a very similar, although not quite identical,
structure known as the Five-Factor Model. This structure contains
dimensions named Extraversion, Agreeableness, Conscien-tiousness,
Neuroticism, and Openness to Experience (see Wiggins, 1996, for a
discussion of various theoretical perspectives on the Five-Factor Model and
the Big Five).
Personality research has benefited enormously during the past decade
from the development of brief but reliable and valid markers of the Big Five
and the Five-Factor Model. Goldberg (1992) introduced adjective scale
markers for the Big Five, and later developed questionnaire scale markers of
the same structure, using his International Personality Item Pool (IPIP;
Goldberg, 1999). Costa and McCrae (1992) introduced a short and a long set
of marker scales for the dimensions of the Five-Factor Model — namely, the
NEO Five-Factor Inventory (NEO-FFI) and NEO Personality Inventory-
Revised (NEO-PI-R) — and both of these tests have become extremely
widely used. Goldberg (1999) has since produced another set of scales
measuring the Five-Factor Model constructs, and these scales are essentially
parallel forms to the NEO scales.
Despite the familiarity and the psychometric quality of these instruments,
however, we believe that there are some limitations associated with five-
dimensional models of personality structure. Recent lexical investigations in
diverse languages have revealed that, when representative sets of personality
variables are factor-analyzed, there are six factors — not just five factors —

330 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

that repeatedly emerge (see Ashton et al., 2004). Moreover, the rotational
positions of some of these factors are somewhat different from the traditional
axis locations prescribed by the Big Five and Five-Factor Model, and we
believe that these alternative rotations allow for a simpler theoretical
interpretation of the factors (Ashton & Lee, 2001). Therefore, we believe
that a “re-organization” (Goldberg, 2001) of the Big Five and Five-Factor
Model is warranted, in order to accommodate these recurring results.
As discussed below, our proposed model of personality structure contains
six dimensions, whose number and names suggest the acronym HEXACO:
Honesty-Humility (H), Emotionality (E), eXtraversion (X), Agreeableness
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

(A), Conscientiousness (C), and Openness to Experience (O). Hereafter,

we will refer to our proposed model of personality structure as the HEXACO
model, and we will refer to the questionnaire based on that model as the
HEXACO Personality Inventory, or HEXACO-PI. In the next section, we
discuss the empirical data that underlie the proposed six-dimensional
structural model of personality upon which the HEXACO-PI is based.

Empirical Findings from Lexical Studies of Personality Structure and

the Domains and Facets of HEXACO-PI

Ashton et al. (2004) recently compared six-factor solutions observed in

eight independent investigations involving seven different languages —
Dutch, French, German, Hungarian, Italian, Korean, and Polish — and found
that a similar structure emerged from each language. We refer to these
investigations as standard lexical studies of personality structure, which
involve factor analyses of ratings on a large number of adjectives that
describe variation in normal personality traits.1 As such, the HEXACO-PI
was developed with the aim of operationalizing the six factors that have been
repeatedly found across various languages. In the sections below, we
describe the content of the six factors as obtained in lexical studies of
personality structure, and we identify the four facets of each factor as
operationalized in the HEXACO-PI.

1
We should note that standard lexical studies in two other languages — English (Saucier &
Goldberg, 1996) and Czech (Hrebickova, 1995) — did not recover the same set of six factors
(see Ashton et al., 2004, for a discussion). In addition, this six-factor solution did not emerge
from rigorous lexical studies conducted in the Turkish (Goldberg & Somer, 2000; Somer &
Goldberg, 1999) and Filipino/Tagalog (e.g., Church, Katigbak, & Reyes, 1998) languages. These
latter two studies, however, differed from what we have called standard lexical studies in terms
of their variable selection, which included purely evaluative terms and/or terms describing
physical attractiveness. Elsewhere, we have argued that these terms should be excluded from
lexical studies (see Ashton & Lee, 2001, for this argument and for a detailed discussion of
findings from the Turkish and Filipino/Tagalog studies).

MULTIVARIATE BEHAVIORAL RESEARCH 331

 K. Lee and M. Ashton

Honesty-Humility and its Facets

A factor that is typically defined by honesty, fairness, sincerity, modesty,

and lack of greed has been observed as either the fifth or the sixth largest
factor in several lexical studies of personality structure (Ashton et al., 2004).
Ashton, Lee, and Son (2000) previously suggested Honesty as a name for
this dimension, but this name may not completely capture the broad content
domain covered by this factor. Therefore, Ashton et al. suggested a new
name, Honesty-Humility, to better reflect this breadth of content. The
addition of the Honesty-Humility dimension is one of the most important
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

characteristics of the HEXACO model, and represents a major departure

from the Big Five or Five-Factor Model.
Based on the common content of this factor as found in various lexical
studies of personality structure, we conceptualized the HEXACO-PI
Honesty-Humility dimension in terms of four distinct facets: Sincerity,
Fairness, Greed Avoidance, and Modesty. Brief definitions of the scales are
provided in Table 1.2

Emotionality and its Facets

Lexical studies of personality structure have repeatedly recovered a

factor that is defined by such characteristics as anxiety, fearfulness,
sentimentality, dependence, and emotional reactivity versus self-assurance,
toughness, and bravery. Although this factor has often been interpreted as
the Big Five Emotional Stability factor by the authors of the lexical studies in
which the factor has emerged, this factor differs from the traditional
Emotional Stability dimension in some important ways. Most notably, this
factor does not include the irritability and temperamentalness content that is
an important element of the low pole of the traditional Emotional Stability
factor (see Costa & McCrae, 1992; Saucier & Goldberg, 1996). Second,
unlike the traditional (low) Emotional Stability dimension, this factor typically
includes sentimentality and sensitivity content at its positive pole and bravery
and toughness content at its negative pole. Given this combination of
characteristics, we have suggested elsewhere that the name of Emotionality
can better describe this factor than can Emotional Stability (Ashton et al.,

2
The facets of the Honesty-Humility factor — unlike the other five HEXACO dimensions —
do not contain any items that resemble adjective self-ratings (e.g., “I am an honest person.”).
We avoided Honesty-Humility items of this kind because of concern that such items might have
somewhat higher means and more skewed response distributions, than would be the case for the
other factors.

332 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton
3
2004). Based on the common content described above, we included
Fearfulness, Anxiety, Dependence, and Sentimentality as the four facets of
the HEXACO-PI Emotionality domain. The definitions of these four scales
can be found in Table 1.

Extraversion and its Facets

A factor interpretable as Extraversion has typically been found as the

largest or the second largest varimax-rotated factor in most lexical studies of
personality structure. Content related to talkativeness, sociability, and
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

cheerfulness versus shyness, passivity, and quietness has consistently defined

this dimension, which is therefore very similar to the Big Five Extraversion
factor. (It should be noted, however, that such traits as bravery, self-
assurance, and toughness, which are sometimes viewed as part of the Big
Five Extraversion factor, usually do not define this factor; instead, such traits
are assimilated within the Emotionality dimension in the HEXACO model.)
Accordingly, the HEXACO-PI Extraversion dimension was conceptualized
to subsume facets that we have called Expressiveness, Social Boldness,
Sociability, and Liveliness. Definitions of these four scales are shown in
Table 1.

Agreeableness and its Facets

The common content of this dimension as obtained in various lexical

studies of personality structure includes good-naturedness, tolerance, and
agreeableness versus temperamentalness, irritability, argumentativeness, and
criticalness. We believe that the term Agreeableness can summarize the
nature of this factor fairly well. It is important to note, however, that this
dimension is somewhat different from Big Five Agreeableness in its content.
Most notably, the HEXACO version of Agreeableness includes the
temperamentalness and irritability content that is an element of low Emotional
Stability in the Big Five (see above). Based on the common content of this
dimension as described above, we included four facets within the HEXACO
Agreeableness dimension: Forgiveness, Gentleness, Flexibility, and Patience.
Brief definitions of these scales are provided in Table 1.

3
The choice of Emotionality as the name of this factor is consistent with the behavior of the
adjective emotional in lexical studies of personality structure. Apparently, people’s implicit
understanding of this term primarily involves sentimentality rather than, say, irritability or
joyfulness. Nevertheless, some readers might reasonably feel that Emotionality is an
ambiguous name for this factor, and we suggest that they understand this label to represent
Emotional Vulnerability or Emotional Sensitivity.

MULTIVARIATE BEHAVIORAL RESEARCH 333

 K. Lee and M. Ashton

Table 1
Definitions of the HEXACO-PI Scales

Scale Definition
Honesty-Humility Domain
Sincerity Assesses a tendency to be genuine in interpersonal relations.
Low scorers will flatter others or pretend to like them in order
to obtain favors, whereas high scorers are unwilling to
manipulate others.
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Fairness Assesses a tendency to avoid fraud and corruption. Low

scorers are willing to gain by cheating or stealing, whereas
high scorers are unwilling to take advantage of other
individuals or of society at large.
Greed Assesses a tendency to be uninterested in possessing
Avoidance lavish wealth, luxury goods, and signs of high social status.
Low scorers want to enjoy and to display wealth and
privilege, whereas high scorers are not especially motivated
by monetary or social-status considerations.
Modesty Assesses a tendency to be modest and unassuming. Low
scorers consider themselves as superior and as entitled to
privileges that others do not have, whereas high scorers view
themselves as ordinary people without any claim to special
treatment.

Emotionality Domain
Fearfulness Assesses a tendency to experience fear. Low scorers feel little
fear of injury and are relatively tough, brave, and insensitive
to physical pain, whereas high scorers are strongly inclined to
avoid physical harm.
Anxiety Assesses a tendency to worry in a variety of contexts. Low
scorers feel little stress in response to difficulties, whereas
high scorers tend to become preoccupied even by relatively
minor problems.
Dependence Assesses one’s need for emotional support from others. Low
scorers feel self-assured and able to deal with problems
without any help or advice, whereas high scorers want to
share their difficulties with those who will provide
encouragement and comfort.
Sentimentality Assesses a tendency to feel strong emotional bonds with
others. Low scorers feel little emotion when saying good-bye
or in reaction to the concerns of others, whereas high scorers
feel strong emotional attachments and an empathic sensitivity
to the feelings of others.
334 MULTIVARIATE BEHAVIORAL RESEARCH
 K. Lee and M. Ashton

Table 1 (cont'd.)

Scale Definition
Extraversion Domain
Expressiveness Assesses a tendency to be excitable and dramatic in one’s
interpersonal style. Low scorers tend not to speak in an
excited or animated way, whereas high scorers tend to do
most of the talking and have a dramatic style of speaking.
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Social Boldness Assesses one’s comfort or confidence within a variety of

social situations. Low scorers feel shy or awkward in
positions of leadership or when speaking in public, whereas
high scorers are willing to approach strangers and are willing
to speak up within group settings.
Sociability Assesses a tendency to enjoy conversation, social
interaction, and parties. Low scorers generally prefer solitary
activities and do not seek out conversation, whereas high
scorers enjoy talking, visiting, and celebrating with others.
Liveliness Assesses one’s typical enthusiasm and energy. Low scorers
tend not to feel especially cheerful or dynamic, whereas high
scorers usually experience a sense of optimism and high
spirits.

Agreeableness Domain
Forgiveness Assesses one’s willingness to feel trust and liking toward
those who may have caused one harm. Low scorers tend
“hold a grudge” against those who have offended them,
whereas high scorers are usually ready to trust others again
and to re-establish friendly relations after having been treated
badly.
Gentleness Assesses a tendency to be mild and lenient in dealings with
other people. Low scorers tend to be critical in their
evaluations of others, whereas high scorers are reluctant to
judge others harshly.
Flexibility Assesses one’s willingness to compromise and cooperate
with others. Low scorers are seen as stubborn and are willing
to argue, whereas high scorers avoid arguments and
accommodate others’ suggestions, even when these may be
unreasonable.
Patience Assesses a tendency to remain calm rather than to become
angry. Low scorers tend to lose their tempers quickly,
whereas high scorers have a high threshold for feeling or
expressing anger.
MULTIVARIATE BEHAVIORAL RESEARCH 335
 K. Lee and M. Ashton
Table 1 (cont'd.)

Scale Definition
Conscientiousness Domain
Organization Assesses a tendency to seek order, particularly in one’s
physical surroundings. Low scorers tend to be sloppy and
haphazard, whereas high scorers keep things tidy and prefer a
structured approach to tasks.
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Diligence Assesses a tendency to work hard. Low scorers have little

self-discipline and are not strongly motivated to achieve,
whereas high scorers have a strong “work ethic” and are
willing to exert themselves.
Perfectionism Assesses a tendency to be thorough and concerned with
details. Low scorers tolerate some errors in their work and
tend to neglect details, whereas high scorers check carefully
for mistakes and potential improvements.
Prudence Assesses a tendency to deliberate carefully and to inhibit
impulses. Low scorers act on impulse and tend not to
consider consequences, whereas high scorers consider their
options carefully and tend to be cautious and self-controlled.

Openness to Experience Domain

Aesthetic Assesses one’s enjoyment of beauty in art and in nature. Low
Appreciation scorers tend not to become absorbed in works of art or in
natural wonders, whereas high scorers have a strong
appreciation of various art forms and of natural beauty.
Inquisitiveness Assesses a tendency to seek information about, and
experience with, the natural and human world. Low scorers
have little curiosity about the natural or social sciences,
whereas high scorers read widely and are interested in travel.
Creativity Assesses one’s preference for innovation and experiment.
Low scorers have little inclination for original thought,
whereas high scorers actively seek new solutions to problems
and express themselves in art.
Unconventionality Assesses a tendency to accept the unusual. Low scorers
avoid eccentric or nonconforming persons, whereas high
scorers are receptive to ideas that might seem strange or
radical.

336 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

Conscientiousness and its Facets

Across lexical studies of personality structure, this factor has been

consistently defined by such content as organization, hard work, carefulness,
and thoroughness. Therefore, this factor is almost identical to the Big Five
Conscientiousness dimension. (Note that this Conscientiousness factor is
usually not defined by terms suggesting a specifically moral conscience, such
as honest or sincere; these terms generally load on the Honesty-Humility
factor.) The HEXACO Conscientiousness dimension was conceptualized as
having four facets named Organization, Diligence, Perfectionism, and
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Prudence. Definitions of the scales are provided in Table 1.

Openness to Experience and its Facets

There is one last factor — namely, Intellect/Imagination — that is

perhaps most controversial in terms of the nature of its common content
across various lexical studies. The Imagination aspect, subsuming traits such
as originality and creativity, appears to be a robust common element of this
factor in previous lexical studies. Other elements, however, have not
uniformly defined this factor. For example, the Intellect aspect (e.g.,
intelligent, smart) was one of the primary elements in this factor for some
languages (e.g., German, Polish). In other languages (e.g., Dutch, Italian),
content related to unconventionality and rebelliousness has been a central
element of this factor.
In conceptualizing the HEXACO Openness to Experience dimension, we
did not include Intellect content, in the sense of intelligence or mental ability.
This reflects our belief that much variance in overall intelligence is likely to
be due to a non-personality construct of fluid intelligence or “fluid g,” in
combination with the personality dimension of Intellect, Imagination, or
Openness to Experience (Ashton, Lee, Vernon, & Jang, 2000). However,
we did include Intellect content in the sense of intellectual curiosity or
inquisitiveness, which, unlike intelligence per se, suggests a typical behavioral
tendency rather than raw cognitive ability.4
Within the HEXACO framework, we decided to use the name Openness
to Experience for this factor instead of Intellect/Imagination, because of the
greater familiarity of the former name. HEXACO Openness to Experience
is conceptualized to subsume the following four facets: Aesthetic Appreciation,

4
Interestingly, intelligence-related terms have separated from imagination- or unconventionality-
related terms within the seven- or eight-factor solutions of some lexical studies of personality
structure, including those conducted in German, Italian, and Czech.

MULTIVARIATE BEHAVIORAL RESEARCH 337

 K. Lee and M. Ashton

Inquisitiveness, Creativity, and Unconventionality. Definitions of these scales

are provided in Table 1.5

Expected Relations with Other Variables

Extraversion, Conscientiousness, and Openness to Experience.

These three dimensions of the HEXACO-PI were expected to show quite
strong correlations with the corresponding factors in the Big Five framework.
In particular, Extraversion and Conscientiousness in the HEXACO-PI are
quite similar to their Big Five counterparts, albeit not perfectly isomorphic
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

with them (e.g., as noted above, HEXACO Extraversion does not contain
such content as self-assurance and bravery). The level of the convergence
between HEXACO Openness to Experience and Big Five Intellect/
Imagination is expected to be somewhat lower than the levels observed for
Extraversion and Conscientiousness. This expectation is attributable to our a
priori decision that intelligence-related content should be excluded from the
HEXACO Openness to Experience dimension, even though this content is
sometimes represented within Big Five Intellect/Imagination.

Emotionality and Agreeableness. Above, we have noted some

differences between the HEXACO and Big Five frameworks with respect to
Emotionality and Agreeableness. Our re-conceptualization of these two
factors can be understood as a re-rotation of the axis locations of the two
corresponding factors in the Big Five framework. Therefore, we expected
that the two HEXACO scales would show nearly isomorphic relations with
marker scales of rotational variants of Agreeableness and Emotional Stability
in the Big Five.

Honesty-Humility. Previously, Ashton et al. (2000) reported that some

existing personality traits did not correlate strongly with any of the lexical Big
Five, but did correlate significantly with Honesty-Humility. Specifically,
Honesty-Humility substantially increased the explained variance in such traits
as Primary Psychopathy (Levenson, Kiehl, & Fitzpatrick, 1995),
Machiavellianism (Christie & Geis, 1970), and Social Adroitness (Jackson,
1994) beyond the level that the Big Five factors could achieve. Although that
study was conducted in a Korean context using Korean adjective marker
scales for Honesty-Humility, a similar pattern of relationships is expected to
5
Although the Unconventionality facet is defined by nonconformity of opinions, the items of
this facet do not refer to specific political attitudes or religious beliefs. We believe that these
important variables are likely to be strongly influenced by personality, but that they are not
themselves personality characteristics.

338 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

be observed in the present study. Therefore, we expected a strong negative

correlation between HEXACO-PI Honesty-Humility and the Primary
Psychopathy scale.

Gender Differences. Gender differences in the Big Five factors have

generally been found in Agreeableness and Emotional Stability. Specifically,
women describe themselves as being more agreeable and less emotionally
stable than men describe themselves as being (Costa & McCrae, 1992, p.
55). We therefore expected substantial gender differences in the HEXACO-
PI Emotionality dimension, which represents the axis corresponding to high
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Agreeableness and low Emotional Stability in the Big Five space. Moreover,
we expected that the gender differences in Emotionality would exceed those
found in any other HEXACO-PI dimension.
One other noteworthy gender difference was also expected, however.
Some evidence suggests that women tend to score higher than do men on
those scales whose content is highly saturated with Honesty-Humility (e.g., low
Primary Psychopathy; Levenson et al., 1995). We expected, therefore, that
there would be a significant gender difference on the HEXACO Honesty-
Humility dimension. For both the Emotionality and the Honesty-Humility
factors, the expected gender differences as described above are consistent
with the theoretical interpretations suggested by Ashton and Lee (2001).

Method

Participants

409 undergraduate students in two Canadian universities participated in

this study (50% women, mean age 22.3 years with SD = 6.3). The students
received course credits or financial reimbursement for their participation.

Measures

The response format for all the measures described below was a five-
point scale (1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree,
and 5 = strongly agree).

HEXACO-PI. This personality inventory was developed to measure the

six major dimensions that have been found in several previous lexical studies of
personality structure, as described in the Introduction (see also Ashton et al.,
2004). The HEXACO-PI consists of 24 facet-level traits that are subsumed
within the six higher-order personality dimensions. Each scale consists of eight

MULTIVARIATE BEHAVIORAL RESEARCH 339

 K. Lee and M. Ashton

items, thus constituting a questionnaire of 192 items.6 To control the effects of

acquiescence, we tried to include a roughly equal number of positively- and
negatively-keyed items within each scale. As a result, all of the scales have a
minimum of two and a maximum of six negatively-keyed items.
The facet scales representing the lower-level factors associated with
each of the six broad HEXACO factors were derived by the authors with
direct reference to the various results of previous lexical studies (see Ashton
et al., 2004). We selected the final set of 192 items from over 600 items that
we had initially written, and that we had administered to various samples
involving more than 1000 participants in various stages. The process of test
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

construction was based largely on the strategy described by Jackson (1971).

Other Measures for Tests of Convergent Validity. We selected six

scales that were hypothesized to show strong correlations with the
HEXACO-PI scales. Five scales were selected from those included in
Goldberg’s (1999) International Personality Item Pool (IPIP). First, 10-item
markers of three of the lexical Big Five personality factors, Extraversion,
Conscientiousness, and Intellect/Imagination, were included. As noted in the
Introduction, these are expected to show strong correlations with HEXACO-
PI Extraversion, Conscientiousness, and Openness to Experience,
respectively. Second, the IPIP also provides scales that measure the 45
facets of the Abridged Big Five Dimensional Circumplex (AB5C), which
includes the personality traits that are located interstitially between each pair
of the Big Five factors (Hofstee, De Raad, & Goldberg, 1992). Among these
AB5C facets, we included the IPIP scales labeled as Pleasantness and
Imperturbability in the present research. The former represents a personality
trait contrasting high Agreeableness/high Emotional Stability with low
Agreeableness/low Emotional Stability (II+/IV+ versus II-/IV-), and the
latter represents a personality trait contrasting high Emotional Stability/low
Agreeableness with low Emotional Stability/high Agreeableness (IV+II-
versus IV-II+). Thus, these scales correspond to variants of Agreeableness
and Emotional Stability that have been rotated 30 degrees from their usual
axis locations. Therefore, these two variables correspond very closely to our
proposed axis locations for Agreeableness and (low) Emotionality. The
Imperturbability and Pleasantness scales have 9 and 11 items, respectively.

6
In some previous studies, we have used a 108-item version of the HEXACO-PI, for which
only domain scores are reported. This earlier version of the HEXACO-PI has shown quite
satisfactory psychometric properties both in its original, English-language version, and in its
Dutch, French, Italian, and Korean translations. We are currently revising this shorter version
of the test, to further improve its psychometric performance and to make it somewhat shorter
still.

340 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

Finally, the 16-item Primary Psychopathy scale (Levenson et al., 1995)

was selected because of its strong (negative) correlation with a lexical
Honesty-Humility factor, as measured by a Korean adjective marker scale
(Ashton et al., 2000).
These measures were administered to a subset (N = 250) of the total
sample. Internal-consistency reliabilities were .90 for IPIP Extraversion, .82
for IPIP Conscientiousness, .84 for IPIP Intellect, .90 for IPIP
Imperturbability, .80 for IPIP Pleasantness, .86 for Primary Psychopathy.

Results and Discussion

Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Internal-Consistency Reliabilities and Descriptive Statistics of the

HEXACO-PI Scales

Table 2 shows that the internal-consistency reliabilities (coefficient alpha)

for the HEXACO-PI scales, at both factor and facet levels, were high.
Reliabilities ranged from .89 (Conscientiousness) to .92 (Honesty-Humility)

Table 2
Internal Consistency Reliabilities of the HEXACO-PI Scales (Coefficient
alpha)
Scale Reliability Scale Reliability
Honesty-Humility .92 Agreeableness .89
Sincerity .79 Forgiveness .88
Fairness .85 Gentleness .77
Greed Avoidance .87 Flexibility .75
Modesty .83 Patience .80
Emotionality .90 Conscientiousness .89
Fearfulness .84 Organization .85
Anxiety .84 Diligence .79
Dependence .85 Perfectionism .79
Sentimentality .81 Prudence .78
Extraversion .92 Openness to Experience .90
Expressiveness .84 Aesthetic Appreciation .86
Social Boldness .86 Inquisitiveness .81
Sociability .79 Creativity .79
Liveliness .85 Unconventionality .80

Note. N = 409. Each facet-level scale has 8 items, and each of factor-level scale has 32 items.

MULTIVARIATE BEHAVIORAL RESEARCH 341

 K. Lee and M. Ashton

at the factor level, and from .75 (Flexibility) to .88 (Forgiveness) at the facet
level. The levels of reliability observed in the present research compare
favorably with those of other existing personality inventories.
Table 3 provides means and standard deviations of the scales for the
entire sample, and also for men and women separately. It was observed that
there are significant gender differences in some facet-level traits.
Interestingly, however, only those facets belonging to Emotionality and
Honesty-Humility showed a consistent pattern of sizeable gender differences
in the same direction. Consequently, it was these two factor-level traits that
showed the largest gender differences (d = 1.08 for Emotionality and d = .59
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

for Honesty-Humility). This is consistent with our theoretical expectation

elaborated elsewhere (see Ashton & Lee, 2001). Somewhat weak but
statistically significant gender differences were also found for other factor-
level traits. Specifically, women averaged slightly higher than men in
Conscientiousness (d = .26, p < .01), and women averaged slightly lower
than men in Openness to Experience (d = –.22, p < .05).

Factor Analyses

The 24 facet scales in the HEXACO-PI were submitted to a principal

axis factor analysis.7 The first ten eigenvalues were 3.5, 3.4, 2.7, 2.5, 1.7,
1.3, 0.8, 0.8, 0.7, and 0.6; the last important decrease in eigenvalues occurred
between the sixth and seventh factors. The first six factors collectively
explained 63.2 % of the total variance, and when these factors were rotated
to a varimax solution, a clear simple factor structure emerged (see Table 4).
All of the scales showed their highest loadings on the designated factors, and
the sizes of the corresponding factor loadings were large, ranging from .54 to
.73. The factor loadings of the scales on non-corresponding factors were
generally small; none of the secondary loadings exceeded an absolute value
of .30 in this sample.

Specificity of Facets. We estimated an index of specificity of the facet

scales, to quantify the amount of the specific variance in each facet that is
not explained by the six higher-order factors. We used the following
procedures to obtain the index of specificity for each facet measure (see

7
We report principal axis factor analysis here because our subsequent analyses are intended to
identify the amount of specificity in each trait, which would be slightly underestimated by
principal component analysis. Principal components analysis produced results nearly identical
to those reported here. In addition, the six varimax-rotated factors were almost identical to the
six obliquely-rotated factors: all corresponding factor score correlations were above .98 (for
oblimin-rotated factors) or above .97 (for promax-rotated factors).

342 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

Table 3
Means and Standard Deviations of the HEXACO-PI Scales
Total Women Men d
Scale Mean SD Mean SD Mean SD (Women - Men)
Honesty-Humility 3.36 .60 3.53 .53 3.19 .62 .59**
Sincerity 3.39 .67 3.48 .64 3.30 .69 .28**
Fairness 3.44 .85 3.69 .74 3.19 .89 .62**
Greed Avoidance 3.04 .85 3.13 .82 2.94 .86 .23*
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Modesty 3.58 .72 3.83 .60 3.35 .75 .71**

Emotionality 3.20 .55 3.46 .50 2.93 .48 1.08**
Fearfulness 2.76 .77 3.03 .77 2.50 .68 .73**
Anxiety 3.33 .77 3.58 .73 3.08 .74 .68**
Dependence 3.09 .77 3.32 .78 2.87 .69 .61**
Sentimentality 3.59 .69 3.91 .58 3.28 .65 1.01**
Extraversion 3.29 .57 3.33 .59 3.25 .55 .13
Expressiveness 3.19 .73 3.30 .72 3.08 .73 .30**
Social Boldness 3.11 .80 2.99 .86 3.23 .73 –.30**
Sociability 3.43 .69 3.54 .68 3.33 .68 .31**
Liveliness 3.42 .71 3.48 .71 3.37 .71 .15
Agreeableness 2.94 .49 2.91 .47 2.97 .50 –.11
Forgiveness 2.85 .74 2.82 .72 2.89 .76 –.09
Gentleness 2.96 .62 2.96 .61 2.96 .63 .00
Flexibility 2.73 .61 2.76 .62 2.71 .59 .08
Patience 3.22 .66 3.12 .61 3.32 .69 –.30**
Conscientiousness 3.32 .51 3.38 .51 3.25 .49 .26**
Organization 3.12 .83 3.20 .83 3.04 .83 .19
Diligence 3.40 .63 3.44 .65 3.35 .62 .14
Perfectionism 3.51 .70 3.60 .71 3.43 .68 .25*
Prudence 3.24 .64 3.29 .63 3.20 .64 .15
Openness to Experience 3.37 .57 3.30 .56 3.43 .58 –.22*
Aesthetic
Appreciation 3.42 .87 3.59 .79 3.25 .91 .39**
Inquisitiveness 3.31 .79 3.07 .77 3.54 .74 –.62**
Creativity 3.23 .73 3.14 .74 3.32 .71 –.25*
Unconventionality 3.51 .64 3.41 .61 3.60 .65 –.31**
Note. N = 203 women, 206 men.
*
p < .05 ** p < .01.

MULTIVARIATE BEHAVIORAL RESEARCH 343

 K. Lee and M. Ashton

Table 4
Factor Analysis of the 24 HEXACO-PI Facet Scales

Factor
Scale 1 2 3 4 5 6 s
Liveliness .73 –.03 .11 .01 .26 .11 .47
Expressiveness .72 .09 .19 –.07 –.17 –.13 .47
Sociability .68 .20 –.09 –.13 .05 –.07 .50
Social Boldness .66 –.24 .24 –.04 –.02 .03 .55
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Sentimentality .18 .70 .09 .21 .08 .05 .47

Anxiety –.26 .67 –.09 –.11 –.25 .07 .48
Dependence .27 .61 –.11 –.05 .00 –.03 .62
Fearfulness –.18 .57 –.28 .10 –.02 .06 .62
Unconventionality .13 –.15 .70 .07 –.09 –.10 .49
Aesthetic Appreciation .06 .24 .70 .18 .07 .11 .52
Creativity .25 –.14 .65 .07 –.11 .03 .51
Inquisitiveness –.05 –.18 .57 –.03 .10 .14 .65
Greed Avoidance –.09 –.08 .19 .72 .19 .01 .51
Sincerity –.05 –.08 .18 .70 –.03 .19 .48
Modesty –.03 .17 –.11 .68 .12 .06 .56
Fairness –.05 .14 .02 .58 .12 .29 .62
Patience .07 –.27 .05 .04 .68 .01 .50
Flexibility –.09 .08 –.08 .08 .66 –.05 .53
Gentleness –.07 .14 –.03 .21 .62 .07 .56
Forgiveness .23 –.11 .04 .04 .57 –.11 .69
Diligence .21 –.04 .10 .09 –.06 .72 .45
Perfectionism .00 .16 .11 .06 –.10 .59 .62
Organization –.05 .02 –.01 .11 –.01 .56 .72
Prudence –.28 –.07 –.07 .18 .12 .54 .59
Note. N = 409. s = index of specificity. Absolute loadings greater than .30 are typed in bold.

Costa & McCrae, 1995; Harman, 1976). First, the specific variance of each
facet scale was estimated by subtracting its communality from its reliability.
Second, we took the square root of this specific variance to make this index
directly comparable to a factor loading. That is, this index can be thought of
as a factor loading of each facet scale on its own specific “factor”. The
index of specificity is shown in the last column in the Table 4.

344 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

The results indicated that all facets contained substantial specificity, with
a mean index of .55 and a range from .45 to .69. This level of specificity
slightly exceeds that of the NEO-PI-R facets (with a mean of .44 and a
range from .29 to .68) (Costa & McCrae, 1995). Taken together, the factor
analytic results indicate that the HEXACO-PI facets contain adequate
amounts of both theoretically relevant common variance and (potentially)
practically useful unique variance.

Inter-Correlations among the Factor-Level Scales

Downloaded by [University Library Utrecht] at 11:14 20 September 2013

An important question involves the extent to which the six factor-level

scales are intercorrelated. As shown in Table 5, the correlations among the
six factor-level scales were fairly low, indicating that those constructs are
roughly independent. The highest correlation (r = .28) was found between
Honesty-Humility and Conscientiousness. Honesty-Humility also correlated
modestly with Agreeableness (r = .21).8 The relatively weak correlations of
Honesty-Humility with the other factors indicate that Honesty-Humility
variance is not well represented by the other HEXACO factors.9
It is important to note that the correlations among the six HEXACO-PI
domain scales observed in the present research are actually somewhat lower
than those observed among the scales of five-dimensional measures such as
the NEO-PI-R (Costa & McCrae, 1992), the NEO-FFI (see Egan, Deary, &
Austin, 2001) or the IPIP-NEO and IPIP-Big Five scales (Goldberg, 1999).
For example, Costa and McCrae (1992, p. 100) reported that NEO-PI-R
Neuroticism correlated –.53 with NEO-PI-R Conscientiousness and that
NEO-PI-R Openness to Experience correlated .40 with NEO-PI-R
Extraversion.

8
One might be surprised by the weak correlation observed between HEXACO Agreeableness
and Honesty-Humility, given that Five-Factor Model Agreeableness (as assessed by the NEO-
PI-R) contains facets that are closely related to Honesty-Humility. As discussed in the
Introduction, however, one should note that HEXACO Agreeableness, with its emphasis on
lack of anger and of hostility, was conceptualized differently from Five-Factor Model
Agreeableness. Interestingly, two facet measures in the NEO-PI-R that seem roughly parallel
to aspects of HEXACO Honesty-Humility and (low) Agreeableness are Straightforwardness
(A2) and Anger/Hostility (N2), respectively. The correlation between these two NEO-PI-R
facets (r = .29, from Costa & McCrae, 1992, p.101) is comparable to that observed between
HEXACO Honesty-Humility and Agreeableness in the present study.
9
To further evaluate the extent to which the six factor-level scales are independent from each
other, we computed the multiple correlations for each of the six factor-level scales using the
other five factor scales as predictors. In general, the multiple correlations were found to be
modest: .42 for Honesty-Humility, .24 for Emotionality, .30 for Extraversion, .28 for
Agreeableness, .31 for Conscientiousness, and .36 for Openness to Experience.

MULTIVARIATE BEHAVIORAL RESEARCH 345

 K. Lee and M. Ashton

Table 5
Intercorrelations Among the Six Factor-level Scales

Scale 1 2 3 4 5

1. Honesty-Humility
2. Emotionality .07
3. Extraversion –.11* –.04
4. Agreeableness .21** –.11* .07
5. Conscientiousness .28** .07 –.06 –.03
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

6. Openness to Experience .16** –.16** .23** .01 .10*

Note. N = 409.
*
p < .05, ** p < .01.

Convergent Validities of HEXACO-PI Scales

Recall that we included six marker scales of constructs that are expected
to be strongly related to our HEXACO-PI scales. As discussed in the
Introduction, HEXACO Extraversion, Conscientiousness, Openness to
Experience are almost parallel to their lexical Big Five counterparts.
Therefore, we included three IPIP scales representing the lexical
Extraversion, Conscientiousness, and Intellect/Imagination factors. As also
discussed in the Introduction, HEXACO Agreeableness and Emotionality can
be interpreted as rotated variants of Big Five Agreeableness and Emotional
Stability. Two scales selected from the Abridged Big Five Dimensional
Circumplex (AB5C) (named Imperturbability and Pleasantness) correspond
to variants of Agreeableness and Emotional Stability that have been rotated
30 degrees from their usual axis locations. Therefore, we hypothesized that
HEXACO Agreeableness and Emotionality should show a nearly one-to-one
correspondence with these two IPIP scales. Finally, as an external marker
variable of HEXACO Honesty-Humility, we included Primary Psychopathy
(Levenson et al., 1995), which was previously found to correlate primarily
with lexical Honesty-Humility, and to be weakly correlated with the lexical
Big Five (Ashton et al., 2000).
Table 6 provides correlations of all HEXACO-PI scales, at both factor-
and facet levels, with the above-mentioned marker scales. All of the factor-
level traits and their associated facets showed strong correlations with their
corresponding marker variables. The highest convergent validity was found for
Extraversion (r = .86), followed by Conscientiousness (r = .83). The lowest
convergent validity was found between IPIP Intellect/Imagination and

346 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

Table 6
Correlations of the HEXACO-PI scales with marker variables of the six factors

Marker Variable

Primary IPIP IPIP IPIP IPIP IPIP

Psycho- Imperturb- Extra- Pleasant- Conscien- Intellect/
Scale pathy ability version ness tiousness Imagination

Honesty-Humility –0.75 –0.10 –0.04 0.31 0.23 0.11

Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Sincerity –0.54 0.03 –0.02 0.17 0.23 0.20

Fairness –0.69 –0.11 –0.02 0.21 0.24 0.08
Greed Avoidance –0.60 –0.08 –0.06 0.27 0.11 0.11
Modesty –0.50 –0.14 –0.01 0.30 0.15 –0.04
Emotionality –0.18 –0.74 –0.06 –0.01 0.04 –0.16
Fearfulness –0.14 –0.52 –0.19 –0.10 0.05 –0.21
Anxiety 0.08 –0.54 –0.27 –0.17 0.02 –0.21
Dependence –0.14 –0.48 0.19 0.04 –0.05 –0.07
Sentimentality –0.39 –0.68 0.11 0.25 0.11 0.02
Extraversion –0.07 0.05 0.86 0.19 –0.01 0.35
Expressiveness –0.08 –0.17 0.67 0.02 –0.08 0.30
Social Boldness –0.02 0.24 0.70 0.04 0.02 0.34
Sociability 0.02 –0.07 0.71 0.18 –0.04 0.14
Liveliness –0.14 0.12 0.57 0.38 0.07 0.28
Agreeableness –0.19 0.16 0.01 0.72 –0.05 –0.08
Forgiveness –0.10 0.11 0.08 0.53 –0.08 0.02
Gentleness –0.20 –0.01 –0.07 0.66 –0.01 –0.14
Flexibility –0.11 0.00 –0.04 0.47 –0.04 –0.23
Patience –0.16 0.34 0.04 0.51 –0.03 0.06
Conscientiousness –0.18 0.00 0.01 0.09 0.83 0.18
Organization –0.10 0.00 0.03 0.08 0.81 0.12
Diligence –0.14 0.01 0.18 0.10 0.55 0.24
Perfectionism –0.11 –0.08 0.01 –0.06 0.51 0.16
Prudence –0.16 0.09 –0.22 0.15 0.46 –0.02
Openness to Experience –0.30 0.04 0.28 0.04 0.09 0.68
Aesthetic Appreciation –0.32 –0.22 0.12 0.07 0.17 0.41
Inquisitiveness –0.14 0.27 0.13 0.03 0.09 0.44
Creativity –0.23 0.03 0.30 0.03 0.03 0.72
Unconventionality –0.20 0.07 0.31 0.00 –0.06 0.53

Note. N = 250. Absolute values greater than .40 are typed in bold.

MULTIVARIATE BEHAVIORAL RESEARCH 347

 K. Lee and M. Ashton

HEXACO-PI Openness to Experience (r = .68). As noted in the Introduction,

this reflects our decision to exclude items involving self-ratings of intelligence,
which were quite heavily represented in the IPIP Intellect/Imagination scale.
(Nevertheless, the Creativity facet of HEXACO-PI Openness to Experience
did correlate strongly, r = .72, with Intellect/Imagination.)
The pattern of convergent and discriminant correlations shown by the
facet-level scales was generally quite similar to that shown by the higher-
order factor scales. It should be noted, however, that the facet correlations
are not so high as to suggest that they are redundant. This, along with the
considerable unique variance associated with each facet, suggests that
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

measuring traits at the facet level would potentially be useful in predicting

various important criteria.

Summary

The psychometric properties of the scales in the newly developed

HEXACO-PI appear to be satisfactory. All the scales were demonstrated to
possess high levels of internal consistency reliabilities. In addition, the
hypothesized factor structure of the HEXACO-PI scales was clearly
identified. All aspects of the results of the factor analyses, including the
scree plot and the patterns of primary and secondary loadings, were
consistent with expectations.
The six factor-level scales were found to show relatively little overlap
with each other, suggesting that these six lexical-derived factors were
roughly orthogonal. The HEXACO-PI scales also showed theoretically
meaningful relations with external variables included in the present research.
Moreover, the HEXACO-PI scales, at both factor and facet levels, uniformly
showed acceptable convergent validities with the six external marker
variables that were hypothesized to have nearly isomorphic relationships with
the HEXACO-PI scales. These six marker scales did not correlate strongly
with any of the HEXACO-PI scales other than their conceptually similar
scales, thus providing evidence of discriminant validity. Also, as expected,
the HEXACO-PI Emotionality and Honesty-Humility scales showed the
largest gender differences (see Ashton & Lee, 2001, for a theoretical
interpretation of this phenomenon).
In conclusion, the HEXACO-PI appears to be a psychometrically sound
measure of the six major personality dimensions that have been found in
lexical studies of personality structure across several diverse languages (see
Ashton et al., 2004).

348 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

Circumplexes and the HEXACO Structure

As reported in the previous section, the facet scales of the HEXACO-PI

show a reasonably simple structure. Nevertheless, several HEXACO-PI
scales exhibit secondary loadings that are both appreciable in size and
theoretically meaningful, and this result is a reflection of the inherently
circumplexical nature of several of the planes defined by pairs of HEXACO
factors. In the next section, we first briefly describe the notion of the
circumplex — and the contributions of Jerry Wiggins to the development of
that concept — and we then discuss the relevance of the circumplexical
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

approach to the HEXACO structure, in the context of the problem of

deciding on the optimal rotation of factor axes.

Circumplexes and Personality Structure: Historical Background

Guttman (1954) used the term “circumplex” to describe a circular

arrangement of variables, and this term was soon applied in the context of
the so-called interpersonal circle, a two-dimensional classification of
interpersonal behaviors (e.g., Freedman, Leary, Ossorio, & Coffey, 1951;
Leary, 1957; see McCormick & Goldberg, 1997, for a historical overview).
The interpersonal circle was later refined and extended by Wiggins (1979;
Wiggins, Trapnell, & Phillips, 1988), who developed a set of adjective
markers of various constructs that fall within that two-dimensional space.
Trapnell and Wiggins (1990) integrated the interpersonal circumplex, which
corresponds roughly to the Extraversion/Agreeableness plane of the Big Five,
with the remaining three dimensions of the Big Five factor structure.
Meanwhile, other researchers also began to study circumplexes beyond the
traditional interpersonal plane. For example, Saucier (1992) explored
circumplexes involving Emotional Stability and each of the two interpersonal
axes, and Hofstee et al. (1992) described the Abridged Big Five Circumplex
(AB5C), which contains all 10 of the planes formed by the Big Five factors.
We believe that studies of the interpersonal circle, and of personality
circumplexes within other planes, have had an important influence on the
thinking of personality psychologists. The work of Wiggins, and later of
Hofstee et al. and of Saucier, has reminded researchers that much of the
personality space is not characterized by simple structure, but by a rather
continuous arrangement of variables between factor axes. Therefore, within
at least some of the planes defined by pairs of personality dimensions, a
variety of equally plausible reference vectors could be selected. This is
especially true of the regions defined by the Big Five axes of Extraversion,
Emotional Stability, and Agreeableness; within this “affective-interpersonal

MULTIVARIATE BEHAVIORAL RESEARCH 349

 K. Lee and M. Ashton

sphere” (Saucier, 1992), factor axis locations may fluctuate across samples
and across variable sets, without any obviously ideal location. For example,
one researcher might prefer three factors that are defined most strongly by
shyness, irritability, and sentimentality, respectively, whereas another might
favor a mathematically equivalent space whose axes are instead located
through friendliness, anxiety, and bossiness.

The HEXACO Factor Rotation: Empirical Basis

At this point, it might be useful to explain the origins of the factor axis
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

locations that characterize the six HEXACO dimensions. Such a discussion

may be especially useful given that the HEXACO Agreeableness and
Emotionality dimensions represent a re-rotation of the Big Five
Agreeableness and Emotional Stability axes, and may therefore be a source
of some discomfort for those readers who view the traditional axis locations
as the “correct” orientation. Below, we first recall the lexical origins of the
HEXACO structure, and we then consider the theoretical considerations that
also contribute to our preference for the HEXACO axis locations.
As described in the Introduction to this article, lexical studies of
personality structure in several languages have produced a similar six-factor
structure. Within this structure, two of the varimax-rotated factor axes
correspond to rotated variants of Agreeableness and Emotional Stability, with
perhaps some additional tilting of this plane with respect to the Extraversion
axis. One of these axes tends to be defined by content related to patience,
tolerance, and good-naturedness versus quick temper, criticalness, and
stubbornness, and the other tends to be defined by content related to anxiety,
vulnerability, and sentimentality versus toughness, bravery, and
independence. Such a result might seem unsatisfying to those readers who
view the traditional Agreeableness and Emotional Stability vectors as the
natural axis locations; indeed, those readers might wonder why the
“interpersonal” dimension of Agreeableness should be mixed with the
“affective” dimension of Emotional Stability. But it must be remembered
that, even though this solution may seem new and unusual, it is nonetheless
the result that nature tends to favor, at least in the context of lexical studies
of personality structure from diverse languages. Thus, on the basis of these
results, the HEXACO factor rotation has an empirical basis that is every bit
as plausible as that of the traditional Big Five axes.
We have noted above that in several languages the results of lexical
studies of personality structure — the investigations that have the best claim
to a representative sampling of personality variables — tend to favor the
HEXACO factor axis locations. But given the highly circumplexical nature

350 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

of much of the personality space, it seems somehow unsatisfying to decide

on the ideal reference vectors simply by counting the number of languages in
which competing rotational positions have emerged: if, say, the HEXACO
solution emerges in two-thirds of the cases, and a competing orientation in
one-third of the cases, this would not be a particularly convincing basis for
the selection of the former structure as the ideal. Instead, it would be better
to have a more compelling theoretical rationale for the selection of a given
set of axes. Previously, we have tried to provide such a rationale.

The HEXACO Factor Rotation: Theoretical Basis

Downloaded by [University Library Utrecht] at 11:14 20 September 2013

In an earlier article (Ashton & Lee, 2001), we explained factors

corresponding to HEXACO Emotionality and Agreeableness, and also
Honesty-Humility, in terms of three dimensions that are relevant to altruistic
versus antagonistic behavior. According to this explanation, Honesty-
Humility and Agreeableness are the basis of individual differences in
reciprocally altruistic behavior. Specifically, high levels of Honesty-Humility
are associated with a tendency to cooperate even when one could exploit
another individual who seems relatively unlikely to retaliate, whereas high
levels of Agreeableness are associated with a tendency to cooperate even
when one could be exploited by another individual who seems relatively
unlikely to reciprocate.10
The role of Emotionality within this theory is to govern individual
differences in behaviors relevant to kin altruism and inclusive fitness. In our
earlier article (Ashton & Lee, 2001) we emphasized certain aspects of
Emotionality — specifically, empathic concern and emotional attachment —
that correspond chiefly to the Sentimentality facet of the HEXACO-PI
Emotionality domain. The link between sentimentality and kin-altruistic
tendencies is straightforward, but the other aspects of Emotionality are also
relevant to individual differences in kin altruism, or in inclusive fitness more
generally. For example, Campbell (1999) has suggested that sex differences
in fear and anxiety can be understood in terms of parental investment:
because women have a greater biological investment in their children than do
men, personal survival has been more strongly associated with reproductive
success in women than in men; consequently, the optimal thresholds for
feelings of physical fear and of anxiety have tended to be somewhat lower

10
The common link to reciprocity might explain the tendency for Honesty-Humility and
Agreeableness to form a single dimension within solutions involving fewer than five or six
factors, and for many traits to show same-signed loadings on these two factors. Future research
should investigate the conditions that influence the extent to which these two factors approach
orthogonality.

MULTIVARIATE BEHAVIORAL RESEARCH 351

 K. Lee and M. Ashton

for women than for men. Similarly, gender differences in the help-seeking
tendencies that characterize the dependence aspect of Emotionality might
also be explained in terms of kin altruism and inclusive fitness: Taylor et al.
(2000, p. 412) have suggested that women are more likely than men to
“selectively affiliate in response to stress, which maximizes the likelihood that
multiple group members will protect both them and their offspring.”
Thus, we believe that the HEXACO Agreeableness and Emotionality axes
possess, in addition to their widespread empirical prevalence, a strong
theoretical basis. Of course, this is not to suggest that other rotations cannot be
defended. For example, consider an interesting result from the Italian lexical
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

studies conducted in Trieste (Di Blas, 2003; Di Blas & Forzi, 1998; 1999; see
also Di Blas & Perugini, 2002). Within these studies, the varimax-rotated six-
factor solution produced a factor defined by content related to sentimentality
and aesthetic appreciation (corresponding to high Emotionality and high
Openness to Experience) and another factor defined by content related to
fearlessness and creativity (corresponding to low Emotionality and high
Openness to Experience). Presumably, one could imagine plausible adaptive
trade-offs that could underlie individual differences in these two dimensions.
But on the other hand, it might be difficult to integrate those interpretations with
those of the other four factors, in such a way as to give a parsimonious
explanation of the entire six-dimensional framework, as we have attempted to
provide (Ashton & Lee, 2001). Thus, a re-rotation of these two Italian
(Trieste) factors to align more closely with the more familiar axis locations
(Ashton et al., 2004; Di Blas, 2003) might be preferred.
The need for a parsimonious explanation of the entire personality space
might provide an argument against certain rotations of the major personality
factors. For example, suppose that one wanted to view “self-esteem” or
“self-efficacy” as a major dimension of personality. It might not be difficult
to imagine an adaptive trade-off between high or low levels of this factor,
and thereby to generate hypotheses regarding the function of this dimension.
However, because self-esteem and self-efficacy are apparently related to
the socially desirable poles of several of the traditional Big Five factors —
particularly Emotional Stability, Extraversion, and Conscientiousness — the
adoption of self-esteem/self-efficacy as a basic factor would then require a
re-rotation of several axes within the personality space, in order to maintain
roughly orthogonal factors. The new vectors obtained in this way would
likely be quite complex in terms of their content, representing desirability-
neutral combinations of several of the traditional factors; consequently, it
might be rather difficult to interpret simultaneously the psychological meaning
and adaptive function of all of these dimensions, whose content would be
much different from that of any simple-structure dimensions.

352 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

The HEXACO Model and the Interpersonal Circle

We should conclude this section by considering the implications of the

HEXACO structure for the interpersonal circle. On one hand, we believe that
the interpersonal circumplex has been superseded as a (partial) representation
of personality structure; instead of just two factors that are directly relevant to
interpersonal behavior, several lexical studies have found that there are four —
Extraversion, Honesty-Humility, Agreeableness, and Emotionality.
Nonetheless, we believe that the interpersonal circle has been — and will
remain — a very useful framework, for two reasons. First, the interpersonal
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

circle manages to capture most of the variance associated with interpersonal

interactions within a plane that represents the intensity and the valence of
those interactions (Ashton & Lee, 2001); this simple framework thereby
provides a very efficient, if somewhat rough, description of interpersonal
behavior. But perhaps even more importantly, the crucial contribution of
Wiggins’s work on the interpersonal circle has been to remind us that
meaningful factor axis locations can often be found in many regions of a given
personality space, and that the problem of optimal factor rotations is a
challenging one that personality researchers must confront with an open mind.

General Discussion

In this article we have described the HEXACO Personality Inventory, a

new measure that is intended to assess six factors that have been found in
lexical studies of personality structure conducted in several languages. An
initial psychometric examination of the HEXACO-PI appears to be
promising. All of the factor- and facet level scales in the HEXACO-PI were
found to be highly reliable and factorially valid. In addition, the six HEXACO
factors were roughly orthogonal to each other. With respect to convergent
validities, the HEXACO scales showed strong correlations with the external
measures in the predicted directions.
We view the psychometric results described in the previous section, as
good reasons to consider the HEXACO-PI as a useful operationalization of
the six-dimensional model of personality structure that we have proposed.
However, we should emphasize that these results are not the reasons why
we believe that this six-dimensional model should be adopted. It is possible
that other psychometricians could construct a set of six broad personality
scales — perhaps by subdividing one or more of the Big Five or Five-Factor
Model factors — that would exhibit reasonable levels of internal-consistency
reliability, of mutual independence, and of convergent and discriminant
validity. Instead, we advocate the HEXACO model on the basis of empirical

MULTIVARIATE BEHAVIORAL RESEARCH 353

 K. Lee and M. Ashton

findings from lexical studies in multiple languages (see Ashton et al., 2004),
and also of theoretical considerations (see Ashton & Lee, 2001).

Potential Predictive Advantages of the HEXACO Model

There remains the important question of what predictive advantages

might be provided by the HEXACO-PI in comparison with currently popular
inventories representing the Five-Factor Model, such as the NEO-PI-R.
Below, we suggest some of these potential predictive advantages.
We believe that the HEXACO-PI is likely to possess some predictive
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

advantages over other inventories, due to its inclusion of a separate Honesty-

Humility factor that is defined by several facets. Although the NEO-PI-R
clearly contains some Honesty-Humility-related content within its Agreeableness
domain, there are nevertheless two HEXACO-PI Honesty-Humility facets —
Fairness and Greed-Avoidance — that have no analogue within the NEO-PI-R.
We believe that these facets are likely to be strongly correlated with certain
important constructs that would probably show only moderate correlations with
the NEO-PI-R variables. For example, scores on overt integrity tests and
similar variables would likely be more strongly correlated with the HEXACO-PI
Fairness facet than with any NEO-PI-R facet. Similarly, variables related to
materialism or status seeking would also be likely to correlate more strongly with
the HEXACO-PI Greed Avoidance facet than with any NEO-PI-R facet.
There are also likely to exist some personality-related variables that can be
predicted fairly well by the broad HEXACO-PI factors, but that are somewhat
poorly predicted by the factor-level variables of alternative personality models
such as the Five-Factor Model. For example, because the HEXACO model is
equipped with the Honesty-Humility domain, it is likely to explain more
variance in various exploitative and deceptive behaviors than the Five-Factor
Model can. The initial evidence seems generally consistent with this
speculation. In recent studies, it was found that the Honesty-Humility factor
was more strongly related than was any of the Big Five factors to workplace
deviance (Lee, Ashton, & Shin, in press) and to sexual harassment proclivities
(Lee, Gizzarone, & Ashton, 2003). In addition, Lee, Ashton, and De Vries
(2003) recently found that the HEXACO six factors significantly better
predicted a workplace delinquency measure than did the Five-Factor Model as
measured by the NEO-FFI (Costa & McCrae, 1992) and the IPIP versions of
the NEO variables (Goldberg, 1999). This predictive gain was primarily due to
the inclusion of the Honesty-Humility factor.
Although we expect that the HEXACO-PI factors will show good levels of
criterion validity, we should point out that the optimal prediction of many criteria
will be achieved with the use of narrower trait measures selected according to

354 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton

their substantive relations to the criteria in question. Many recent studies (e.g.,
Ashton, 1998; Paunonen & Ashton, 2001) have shown that predictive validity is
improved significantly when “facets” of the major personality factors, chosen
according to their substantive relevance to the targeted criteria, are used
instead of the broader, more heterogeneous factor-level measures. Because
the HEXACO-PI is hierarchically organized and provides the measurement of
personality traits at both factor and facet levels, the predictive validity of the
HEXACO-PI is likely to be improved when prediction is made on the basis of
combinations of rationally selected narrow facet measures. Furthermore, we
should note that facet scales in the HEXACO-PI were found to have
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

substantial amounts of variance that cannot be explained by their higher-order

factors, and that it is precisely this specific variance that is responsible for the
increased predictive validity associated with narrow trait measures. Therefore,
the facet-level measurement provided by the HEXACO-PI is likely to increase
prediction significantly beyond the level that is achieved by the factor-level
personality traits alone.

Is the Honesty-Humility Scale Merely Measuring Response Style Variance?

As mentioned in the Introduction, we constructed four Honesty-Humility

facet scales based on the results of several lexical studies of personality
structure. High levels of these traits — sincerity, fairness, modesty, and
greed avoidance — are likely to be desirable in many social and occupational
contexts. Therefore, it would be interesting to investigate the performance
of the Honesty-Humility scales in situations in which respondents are
motivated to fake good — for example, when applying for a job.
Interestingly, some existing measures of intentional distortion are heavily
represented by honesty-related behavioral descriptions, and this suggests that
Honesty-Humility scores would likely increase when self-reports are
provided for purposes of selection. However, in situations in which
respondents have no particular motivation to make a good impression, the
Honesty-Humility scale appears to capture true substantive personality
variance rather than merely response style variance associated with making
a good impression.
For example, in one of our previous studies (Lee et al., 2003), we
collected data from 150 male participants who provided both self- and peer
ratings on the earlier version of 18-item Honesty-Humility scale and on the
IPIP-Big Five scales. In this data set, the correlation between self- and peer
ratings on Honesty-Humility was .43, a value that falls within the typical
range of cross-source correlations of personality traits (see Watson,
Hubbard, & Wiese, 2000). The self/peer correlations for the Big Five

MULTIVARIATE BEHAVIORAL RESEARCH 355

 K. Lee and M. Ashton

factors in the same data set ranged from .27 (Emotional Stability) to .48
(Extraversion). In fact, in terms of cross-source correlations, HEXACO
Honesty-Humility was the second highest, surpassed only by IPIP-Big Five
Extraversion. Moreover, Lee et al. further found that peer rated Honesty
was the strongest peer rating predictor of self-reported sexual harassment
proclivities, surpassing the validity of peer ratings on the Big Five scales in
predicting this criterion. These results are inconsistent with the argument
that variance in self-reports on the HEXACO Honesty-Humility scale is
largely due to response styles.
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Conclusion

In this article, we have introduced a new inventory — the HEXACO-PI

— that is designed to measure six major dimensions of personality and their
constituent facets. On the basis of the psychometric analyses reported here,
we believe that this new inventory will be useful to researchers who wish to
measure the major dimensions of personality, for theoretical and for applied
purposes.

References

Ashton, M. C. (1998). Personality and job performance: The importance of narrow traits.
Journal of Organizational Behavior, 19, 289-303.
Ashton, M. C. & Lee, K. (2001). A theoretical basis for the major dimensions of personality.
European Journal of Personality, 15, 327-353.
Ashton, M. C., Lee, K., Perugini, M., Szarota, P., De Vries, R. E., Di Blas, L., Boies, K., & De
Raad, B. (2004). A six-factor structure of personality-descriptive adjectives: Solutions
from psycholexical studies in seven languages. Journal of Personality and Social
Psychology, 86, 356-366.
Ashton, M. C., Lee, K., & Son, C. (2000). Honesty as the sixth factor of personality:
Correlations with Machiavellianism, primary psychopathy, and social adroitness.
European Journal of Personality, 14, 359-368.
Ashton, M. C., Lee, K., Vernon, P. A., & Jang, K. L. (2000). Fluid intelligence, crystallized
intelligence, and the openness/intellect factor. Journal of Research in Personality, 34, 198-207.
Campbell, A. (1999). Staying alive: Evolution, culture, and women’s intrasexual aggression.
Behavioral and Brain Sciences, 22, 223-252.
Christie, R. & Geis, F. L. (1970). Studies in machiavellianism. Academic Press: New York.
Church, A. T., Katigbak, M. S., & Reyes, J. A. S. (1998). Further exploration of Filipino
personality structure using the lexical approach: Do the big-five or big seven dimensions
emerge? European Journal of Personality, 12, 249-269.
Costa, P. T., Jr. & McCrae, R. R. (1985). The NEO personality inventory manual. Psychological
Assessment Resources: Odessa, FL.
Costa, P. T., Jr. & McCrae, R. R. (1992). NEO Personality Inventory–Revised (NEO-PI-R) and
NEO Five-Factor Inventory (NEO-FFI) professional manual. Psychological Assessment
Resources: Odessa, FL.

356 MULTIVARIATE BEHAVIORAL RESEARCH

 K. Lee and M. Ashton
Costa, P. T., Jr. & McCrae, R. R. (1995). Domains and facets: Hierarchical personality assessment
using the revised NEO Personality Inventory. Journal of Personality Assessment, 64, 21-50.
Di Blas, L. (2003). Personality-relevant attribute-nouns: A taxonomic study in the Italian
language. Manuscript submitted for publication.
Di Blas, L. & Forzi, M. (1998). An alternative taxonomic study of personality-descriptive
adjectives in the Italian language. European Journal of Personality, 12, 75-101.
Di Blas, L. & Forzi, M. (1999). Refining a descriptive structure of personality attributes in the
Italian language. Journal of Personality and Social Psychology, 76, 451-481.
Di Blas, L. & Perugini, M. (2002). L’approccio psicolessicale nella lingua italiana: Due studi
tassonomici a confronto. Giornale Italiano di Psicologia, 29, 67-93 (in Italian).
Digman, J. M. & Takemoto-Chock, N. K. (1981). Factors in the natural language of personality:
Re-analysis, comparison, and interpretation of six major studies. Multivariate Behavioral
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Research, 16, 149-170.

Egan, V., Deary, I., & Austin, E. (2001). The NEO-FFI: Emerging British norms and an item-
level analysis suggest N, A and C are more reliable than O and E. Personality and
Individual Differences, 29, 907-920.
Freedman, M. B., Leary, T., Ossorio, A. G., & Coffey, H. S. (1951). The interpersonal
dimension of personality. Journal of Personality, 20, 143-161.
Goldberg, L. R. (1990). An alternative “description of personality”: The Big-Five factor
structure. Journal of Personality and Social Psychology, 59, 1216-1229.
Goldberg, L. R. (1992). The development of markers for the Big-Five factor structure.
Psychological Assessment, 4, 26-42.
Goldberg, L. R. (1993). The structure of phenotypic personality traits. American
Psychologist, 48, 26-34.
Goldberg, L. R. (1999). A broad-bandwidth, public-domain, personality inventory measuring
the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt,
& F. Ostendorf (Eds.), Personality psychology in Europe, Vol. 7 (pp. 7-28). Tilburg
University Press: The Netherlands.
Goldberg, L. R. (2001). Analyses of Digman’s child-personality data: Derivation of Big-Five
factor scores from each of six samples. Journal of Personality, 69, 709-743.
Goldberg, L. R. & Somer, O. (2000). The hierarchical structure of common Turkish person-
descriptive adjectives. European Journal of Personality, 14, 497-531.
Guttman, L. (1954). A new approach to factor analysis: The Radex. In P. R. Lazarsfeld (Ed.),
Mathematical thinking in the social sciences (pp. 258-348). Glencoe, IL: Free Press.
Harman, H. H. (1976). Modern factor analysis. Chicago: University of Chicago Press.
Hofstee, W. K. B., De Raad, B., & Goldberg, L. R. (1992). Integration of the Big Five and
circumplex approaches to trait structure. Journal of Personality and Social Psychology,
63, 146-163.
Hrebickova, M. (1995). The structural model of personality based on the lexical analysis: A
Czech replication study of the five-factor model based on a comprehensive taxonomy of
personality-descriptive adjectives. Unpublished manuscript, Institute of Psychology,
Academy of Sciences of the Czech Republic, Brno.
Jackson, D. N. (1971). The dynamics of structured personality tests: 1971. Psychological
Review, 78, 229-248.
Jackson, D. N. (1994). Jackson personality inventory — Revised manual. Port Huron, MI:
Sigma Assessment Systems.
Leary, T. (1957). Interpersonal diagnosis of personality. New York: Ronald Press.
Lee, K., Ashton, M. C., & De Vries, R. E. (2003, August). Predicting workplace delinquency
using personality: Findings from three countries. Paper presented at the meeting of
Academy of Management, Seattle, WA.

MULTIVARIATE BEHAVIORAL RESEARCH 357

 K. Lee and M. Ashton
Lee, K., Ashton, M. C., & Shin, K.-H. (in press). Personality correlates of workplace anti-
social behavior. Applied psychology: An international review.
Lee, K., Gizzarone, M., & Ashton, M. C. (2003). Personality and the likelihood to sexually
harass. Sex Roles, 49.
Levenson, M. R., Kiehl, K. A., & Fitzpatrick, C. M. (1995). Assessing psychopathic attributes in
a noninstitutionalized population. Journal of Personality and Social Psychology, 68, 151-158.
McCormick, C. C., & Goldberg, L. R. (1997). Two at a time is better than one at a time:
Exploiting the horizontal aspects of factor representations. In R. Plutchik & H. Conte
(Eds.), Circumplex models of personality and emotions (pp. 103-132). Washington, DC:
American Psychological Association.
McCrae, R. R. & Costa, P. T., Jr. (1985). Updating Norman’s “adequate taxonomy”:
Intelligence and personality dimensions in the natural language and in questionnaires.
Downloaded by [University Library Utrecht] at 11:14 20 September 2013

Journal of Personality and Social Psychology, 49, 710-721.

Norman, W. (1963). Toward an adequate taxonomy of personality attributes: Replicated
factor structure in peer nomination personality ratings. Journal of Abnormal and Social
Psychology, 66, 574-583.
Paunonen, S. V. & Ashton, M. C. (2001). Big Five factors and facets and the prediction of
behavior. Journal of Personality and Social Psychology, 81, 524-539.
Saucier, G. (1992). Benchmarks: Integrating affective and interpersonal circles with the Big-
Five personality factors. Journal of Personality and Social Psychology, 62, 1025-1035.
Saucier, G. & Goldberg, L. R. (1996). Evidence for the Big Five in analyses of familiar English
personality adjectives. European Journal of Personality, 10, 61-77.
Somer, O. & Goldberg, L. R. (1999). The structure of Turkish trait-descriptive adjectives.
Journal of Personality and Social Psychology, 76, 431-450.
Taylor, S. E., Klein, L. C., Lewis, B. P., Gruenewald, T. L., Gurung, R. A. R., & Updegraff, J. A.
(2000). Biobehavioral responses to stress in females: Tend-and-befriend, not fight-or-
flight. Psychological Review, 107, 411-429.
Trapnell, P. D. & Wiggins, J. S. (1990). Extension of the Interpersonal Adjective Scales to
include the Big Five dimensions of personality. Journal of Personality and Social
Psychology, 59, 781-790.
Tupes, E. C. & Christal, R. E. (1992). Recurrent personality factors based on trait ratings.
Journal of Personality, 60, 225-251. (Original published in 1961, USAF Tech. Rep. No.
61-97. U.S. Air Force: Lackland Air Force Base, TX.)
Watson, D., Hubbard, B., & Wiese, D. (2000). Self-other agreement in personality and affectivity:
The role of acquaintanceship, trait visibility, and assumed similarity. Journal of Personality
and Social Psychology, 78, 546-558.
Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptive terms: The interpersonal
domain. Journal of Personality and Social Psychology, 59, 781-790.
Wiggins, J. S. (1996). The five-factor model of personality: Theoretical perspectives. New York:
The Guilford Press.
Wiggins, J. S., Trapnell, P., & Phillips, N. (1988). Psychometric and geometric characteristics
of the Revised Interpersonal Adjective Scales (IAS–R). Multivariate Behavioral
Research, 23, 517-530.

358 MULTIVARIATE BEHAVIORAL RESEARCH