Circumplex Models and empirically.

However, rather than simply

offering a convenient picture for a group of
Michael B. Gurtman interrelated concepts, the circle instead implies
University of WisconsinParkside, U.S.A. a rigorous and testable set of prescriptive crite-
ria for the structure of a particular domain. If
Circumplex, or circular, models in clinical a certain domain (or associated measure) has
psychology are theoretical models intended circumplex properties, then it should meet the
to capture the postulated circular structure three structural criteria for a circle.
of particular domains of interest. The defin- To start, a circular model would imply a
ing feature of such models is the assumption two-dimensional structure. That is, differences
that the variables comprising the domain between the variables in the domain (such as
are best represented by a circlevariables the interpersonal or the affective) should be
arranged in a circular order. Circumplex largely reducible to differences in two dimen-
models are therefore similar to other struc- sions (a plane). The dimensions themselves are
tural models in clinical psychology, such as elemental and become the axesthe coordi-
the five-factor model, in that they provide a nate systemby which the variables can be
framework for understanding how a certain signified. For example, as shown in Figure 1, the
substantive domain, usually related to person- two dimensions that form the interpersonal
ality or individual differences, is structured space are traditionally defined as Domi-
or organized internally. Circumplex models nance (vs. Submissiveness) and Nurturance
have been constructed to represent a number (vs. Hostility or Detachment). Consequently,
of domains relevant to clinical psychology, each variable in the interpersonal circum-
notably interpersonal, affect and emotion, plex can also be regarded, at a higher level of
vocational interests, and human values. The abstraction, as a particular blend of Dominance
interpersonal circumplex (also referred to as (positive or negative) and Nurturance (positive
the interpersonal circle) has received by far the or negative). For example, many concepts of
most attention and development. An example Dependency are forms of FriendlySubmis-
of an interpersonal circumplex is provided in siveness, blending (negative) Dominance and
Figure 1. Consequently, this entry will focus (positive) Nurturance. Vindictiveness is a form
mainly on the interpersonal circumplex model, of HostileDominance, combining (positive)
and its implications and applications to clin- Dominance with (negative) Nurturance. The
ical and abnormal psychology. However, the particular amalgam of Dominance and Nur-
principles of analysis apply equally to the other turance provides interpersonal nuance to the
circumplex models. variable, and allows conceptually similar inter-
personal concepts (e.g., Gregariousness and
Sociability) to be distinguished.
What is a Circumplex?
Other substantive domains are defined by
Fundamentally, a circumplex is a two- different two-dimensional schemes. Circum-
dimensional representation of a particular plex models of affect are typically defined by
domain in which the relevant set of variables the bipolar dimensions of Pleasure (Pleas-
are arranged as a circlea continuous order ant vs. Unpleasant) and Arousal (Activated
with no beginning or end. The geometry of vs. Deactivated); thus affective experiences
the circle becomes a concise summary for (e.g., ashamed, peaceful) vary on these two
how variables are interrelated, theoretically dimensions. Theoretically, vocational interests

The Encyclopedia of Clinical Psychology, First Edition. Edited by Robin L. Cautin and Scott O. Lilienfeld.
2015 John Wiley & Sons, Inc. Published 2015 by John Wiley & Sons, Inc.
DOI: 10.1002/9781118625392.wbecp107
2 CIRCUMPLEX MODELS

Domineering
90

Vindictive Intrusive
135 45

DOMINANCE
NURTURANCE
Cold 180 0 Overly nurturant

Socially avoidant 225 315 Exploitable

270
Nonassertive

Figure 1 A circumplex of interpersonal problems, with dimensions, variables, and polar coordinates
indicated.

are defined at a higher level by the dimensions radius would suggest that the factors of Dom-
of Data vs. Ideas and People vs. Things. For inance and Nurturance are equally relevant
universal human values, Self-Transcendence overall for describing the set of interpersonal
vs. Self-Enhancement and Openness to Change variations.
vs. Conservation are generally regarded as pri- Circumplex models are often set apart
mary dimensions. The Abridged Big Five from other two-dimensional models of per-
Circumplex (see Hofstee, de Raad, & Gold- sonality and emotion by the third defining
berg, 1992) uses pairs of Big Five factors to propertythe continuous distribution of
define various domains of personality beyond variables. Theoretically, variables are uniformly
the interpersonal. distributed around the circumference of a cir-
Circular models also have the property of cle, with no major gaps or discontinuities.
constant radius: by definition, each variable Thus, at any point on the circular continuum,
in a circle is equidistant from the fixed point a meaningful concept or variation should be
of the origin. When differences between vari- potentially specifiable, even if not directly mea-
ables in a given domain are modeled in a sured. In contrast, univocal (simple structure)
two-dimensional space, each variable should models of personality and emotion, such as
show an equal projectiondistance from the most versions of the five factor model, assume
originin that space. This is evident in the that relevant variables cluster at or near the
projections of variables depicted in Figure 1. poles of the defining principal dimensions,
Departures from this ideal may be random with parts of the factor space empty.
or lawful. For example, one alternative to a Although the circumplex is theoretically
circular model is an elliptical one; although the continuous, personality measures based on
arrangement of variables also follows a closed the circumplex will typically sample at equal
curve, the distances from the origin are not intervals along the circular continuum, yield-
equal due to one factor generally accounting ing scales having the property of equal spacing,
for more variability than the other. For an inter- For example, in Figure 1, the interpersonal
personal circumplex, the property of constant circumplex is divided into eight octants that
 CIRCUMPLEX MODELS 3

are equally spaced around the circle. Each Taking the cosine of the angular discrepancy
octant is, in effect, a bipolar dimension, with yields a useful correlation statistic, which, like
its opposite directly across the circle. Other a correlation, ranges from + 1 (a discrepancy
circumplex models divide the space into as few of 0 ) through 0 (90 apart, or orthogonal) to
as 6 or as many as 16 equally spaced sectors for 1 (180 apart, opposing).
purposes of measuring relevant concepts.
Brief History of the Circumplex
Circumplex as a Polar Coordinate Concept
System
The term circumplex was first proposed by
Because the circumplex is a circle, circumplex Louis Guttman (1954), who described it as a
models typically use a polar coordinate system system of variables which has a circular law of
to locate variables and people on the circular order. (p. 325). It was proposed as one kind of
continuum. Advantageously, this means that a structural model (another being a rank order
variable (or persons) position on the circle can model, the simplex) intended to explain and
be specified by a single numberby the angle predict the observed relations among partic-
of displacement (in degrees) from an arbitrary ular sets of tests of mental abilities. Guttman
starting point. Figure 1 also shows the polar demonstrated that, under certain ideal con-
coordinates (angular positions) for the vari- ditions, the correlations among these tests
ables of the interpersonal circumplex. By con- would yield a correlation matrix (described as
vention, 0 is at the three-oclock position and a circulant, p. 328), which would have a kind of
displacement is measured in degrees counter- circular form. This circulant matrix is charac-
clockwise from that point on the circle. Based terized by recurring values that are largest near
on Figure 1, it can be deduced that variables the main diagonal, tapering off and reaching a
(and people) between 0 and 90 involve dif- minimum, and then increasing once again in
ferent forms of FriendlyDominance; between proximity to the main diagonal. Correlations
90 and 180 , HostileDominance; between at equal distances from the main diagonal
180 and 270 , HostileSubmissiveness; and would be equal. A formal version of this model
between 270 and 0 , FriendlySubmissiveness. is presented in Table 1. Note that the pattern
The similarity of two variables (or again, peo- of the correlations, but not the values them-
ple) would also be a function of their angular selves, are specifiedconsequently, within
discrepancy on the circle, with smaller sepa- the constraints of the model, many different
ration (in degrees) denoting greater similarity. instantiations of the circumplex are possible.

Table 1 Guttmans circulant matrix.

Variable V1 V2 V3 V4 V5 V6 V7 V8

V1 1
V2 1 1
V3 2 1 1
V4 3 2 1 1
V5 4 3 2 1 1
V6 3 4 3 2 1 1
V7 2 3 4 3 2 1 1
V8 1 2 3 4 3 2 1 1

where 1 >2 >3 >4.

4 CIRCUMPLEX MODELS

Finally, Guttman defined a quasi-circumplex these empirically in the interpersonal space.

(p. 329) as a circumplex plus deviations. His work also stimulated the development
The second milestone in the history of the of additional measures designed to assess a
circular model was the development, by the variety of interpersonal characteristics in line
Kaiser Foundation Research group, of the with the circumplex model. The works of other
interpersonal circle, which began in the late pioneers in the development of the interper-
1940s and culminated with the publication sonal circumplex and its applications are in the
of Learys (1957) Interpersonal Diagnosis of Reference and Further Reading sections of this
Personality. As a tool for the interpersonal entry.
assessment of personality at different levels, Finally, the circumplex model has also
the interpersonal circle was a circular model been extended to other domains relevant to
that divided the interpersonal space into 16 assessment of personality and adjustment (see
segments, each of which could be assessed Plutchik and Contes (1997) edited volume on
at multiple levels and be applicable to both circumplex models for examples). Of historical
normal and abnormal adjustments. Variables note: Beginning in the 1950s, a number of
were arranged on a circular continuum on the investigators have proposed circular models
basis of their presumed interrelatedness. Con- of emotions, mood, or affect. For example,
sistent with modern circumplex models, each Plutchiks psychoevolutionary theory of emo-
of the variables (or interpersonal mechanisms) tions posits eight bipolar, basic emotions
in the system was regarded as either pure arranged in a circle. More recently, Russell and
forms or a blend of two factors: Dominance his colleagues have argued for and validated
vs. Submission and Affiliation (or Love) vs. a circumplex model of core affect. Tracey,
Hostility (or Hate). The Kaiser group also pio- Rounds, and their colleagues have provided a
neered methods for doing structural analyses circumplex reconceptualization of Hollands
of interpersonal behavior that capitalized on RIASEC (realistic, investigative, artistic, social,
the circular form of the model. enterprising, and conventional) model of the
Although the interpersonal circle was devel- vocational interests, which has also led to new
oped independently of Guttmans work on the measures based on circular structure. Schwartz
circumplex, the connection between the two and his colleagues (e.g., Schwartz & Bilksy,
was apparently made by a number of investiga- 1990) have also presented a quasi-circumplex
tors in the 1960s and 1970s. Perhaps the most model of universal human values that has
important contribution, however, was made been central to their research on value priori-
by Wiggins (1979), who developed what was ties, and has yielded a large body of work on
arguably the first interpersonal circumplex, cross-cultural differences.
using the circumplex model to guide develop-
ment of a taxonomy of interpersonal traits. The Interpersonal Circumplex Measures
Interpersonal Adjective Scales (IAS), a 64-item
of Personality and Adjustment
set of items (personality adjectives) assessing
eight octants of a circumplex space, were the The Kaiser Research group in the 1950s was
direct consequence. Wiggins and subsequent instrumental in forging links between the
analyses of the IAS indicated that this measure circular model and the assessment and diag-
of interpersonal traits had close fit to the cir- nostic activities of the clinician. Their most
cumplex model and could serve as a taxonomy important tool toward that end was the Inter-
of the interpersonal domain of personality. In personal Checklist (ICL), which consisted of
later work, Wiggins and his colleagues also 144 personality-descriptive words or phrases
showed how the circumplex could be used to grouped according to the 16 sectors of the
discern the interpersonal content of various interpersonal circle. In each sector, words
measures and their constructs, by locating or phrases ranged from moderate intensity
 CIRCUMPLEX MODELS 5

(able to give orders) to extreme (dicta- 1980s by Horowitz and his colleagues, the IIP
torial), radiating from the center to the is a self-report instrument that assesses sources
periphery of the circle. Combining adjacent of interpersonal difficulty and distress. Items
sectors across levels yielded eight octant scores originated from an analysis of the presenting
(e.g., Managerial-Autocratic), which provided complaints of individuals seeking psychother-
the basic data of the assessment (see Leary, apy. In its current form, the IIP consists of
1957, for a summary). The ICL, although 64 items, each a short statement expressing a
pioneering and innovative, also suffered from particular interpersonal complaint or difficulty.
psychometric and conceptual problems that Items are grouped, with the first set of items
limited its usefulness, but ultimately stimulated expressing things a person finds hard to do
development of new, refined measures. (e.g., join in on groups, say no to other
Table 2 provides a list of several contem- people), and the second set, what a person
porary measures of interpersonal personality may do too much of (e.g., keep people at a
and adjustment, each constructed on the basis distance, put other peoples needs before my
of an eight-octant circumplex model. (Refer- own). Each problem statement is rated on a
ences to most appear in Lockes chapter in the 04 scale of distress. As with the IAS and the
recent volume edited by Horowitz & Strack, other measures in Table 2, the IIP is scored for
2011.) These address a variety of interper- the eight octants of the interpersonal circum-
sonal constructs (e.g., problems, traits, values) plex, as well as the various summary indices
that would be relevant to an interpersonally described later. Several short forms, appropri-
focused assessment of individual differences ate for quick clinical screenings or assessments,
and personality functioning. Not on the list, have also been developed. Horowitz (2004)
but worthy of mention and further explo- provides an excellent overview of the IIP and
ration, is Benjamins (1996) Structural Analysis its role in interpersonal assessment.
of Social Behavior (SASB) system, which is a
multifaceted, multidimensional clinical tool Evaluating the Fit of a Circumplex
with clear links to the circumplex tradition.
Measure to the Model
Of the circumplex measures cited in Table 2,
the Inventory of Interpersonal Problems (IIP), Circumplex measures, such as those in Table 2,
offers perhaps the most clinically relevant tool are constructed on the basis of both substan-
for interpersonal assessment from the circum- tive and structural criteria. At the same time,
plex perspective. Developed in the 1970s and methods of evaluating the fit of the circumplex

Table 2 Interpersonal circumplex measures and their constructs.

Interpersonal circumplex measures Interpersonal construct

Checklist of Interpersonal Transactions (CLOIT) Behaviors

Circumplex Scales of Interpersonal Efficacy (CSIE) Self-efficacy
Circumplex Scales of Interpersonal Values (CSIV) Values or motives
Impact Message Inventory (IMI) Impacts or evoked reactions
Interpersonal Adjective Scales (IAS) Traits
Interpersonal Grid Perceptions of behavior
Interpersonal Sensitivities Circumplex (ISC) Irritants
Inventory of Interpersonal Goals Interaction goals
Inventory of Interpersonal Problems Circumplex (IIP) Problems
Inventory of Interpersonal Strengths (IIS) Strengths, competencies
Social Actions Scale Circumplex (SAS-C) Social supports
6 CIRCUMPLEX MODELS

model to these measures are critical for both examines the order relations in the obtained
validating those efforts, and for ensuring that correlation matrix and compares these to the
the profile analysis techniques described later hypothesized relations in the ideal circulant.
are appropriate and meaningful. Evaluating fit Order relations are predicted on the basis of a
is also important for testing theory about the variables theoretical location on the circum-
hypothesized structure of particular domains, plex; pairs of variables closer together should be
such as the interpersonal and vocational. more highly intercorrelated than those further
Tracey (2000) has reviewed and demon- apart. RANDALL computes a correspondence
strated the common statistical approaches index (CI) that indicates the proportion of
for fitting the circumplex model to measures. confirmed predictions, and also calculates a
For additional detail, the interested reader is bootstrap probability for that result. CI values
referred to his chapter. near 1 are desirable and indicate a close fit to
Exploratory methods of analysis, such as the model.
factor analysis and multidimensional scaling,
are valuable as heuristics, providing spatial rep- Analyzing and Interpreting
resentations of the distribution of variables (or Circumplex Profiles
scales) in a two-dimensional space. The start-
ing point is typically the correlation matrix, When personality data are collected using
which, as indicated earlier, will conform in a circumplex measures, the resultant profiles
general way to the circulant of Table 1 if fit is will have unique structural features that follow
good. As one example of this method, Wiggins from the circular structure of those measures.
(1979) subjected the correlation matrix of the These circular profiles are often displayed to
eight IAS to a principal components analysis. best effect when plotted in a circular (polar
He found that a two-factor solution accounted coordinate) space and when analyzed using
for a large proportion of the covariance in IAS; circular statistics.
that the two factors (identifiable as Dominance Figure 2 shows an example of a circular pro-
and Nurturance) were of nearly equal mag- file. The profile shows the pattern of scores of a
nitude, and that the eight scales had roughly hypothetical individual on the eight octants of
equal projections and were equally spaced in a circumplex measure, in this case, the Inven-
the two-dimensional space. Guttmans Smallest tory of Interpersonal Problems. Note that the
Space Analysis, a form of multidimensional scale scores themselves are presented in a polar
scaling, has also been used with similar effect, coordinate space, each score radiating from the
especially in validating the structure of human center. Scales are ordered and referenced by
values measures. their angular positions on the theoretical cir-
Confirmatory methods of analysis provide cle; neighboring scales are more closely related
formal tests of model fit, and are also neces- (conceptually and empirically) than those fur-
sary for comparing competing models. Two ther apart on the circle. Each of the eight scales
approaches have been widely used in the samples systematically from a particular region
circumplex literature. Browne and his col- of the circumplex space (a 45 span), and all
leagues (e.g., Browne, 1992) have developed regions of that theoretical space (e.g., of inter-
CIRCUM, which implements a covariance personal problems) are equally represented.
structure model specifically for circumplex Given these features, circular profiles have a
models. In addition to providing indices of characteristic shape that facilitates interpreta-
model fit, CIRCUM also offers estimates for tion. Specifically, because they are drawn from
various key model parameters, including the circularly structured measures, they often have
polar angles (locations) of the variables in a a sinusoidal form, with scores focused around
two-dimensional space. RANDALL, developed a single peak value and then declining away
by Tracey and his colleagues (see Tracey, 2000), from that point. As shown in Figure 2, the
 CIRCUMPLEX MODELS 7

90

135 45

180 0
2 1 0 1 2

225 315

270

Figure 2 An example of a circular profile.

entire profile is, in effect, shifted away from than an individual reporting few problems. In
the origin toward that point, which signifies this case, expanse is related to general interper-
the predominant trend in the profile. For sonal distress or maladjustment (as measured
Figure 2 the profile is clearly displaced toward by the Inventory of Interpersonal Problems).
the FriendlySubmissive quadrant of the circle Second, some profiles may be more clearly
(about 315 ), suggesting that for this indi- articulated or defined than others, following
vidual, themes related to FriendlySubmission the prototypic shape just described; some pro-
(e.g., dependency, deference, exploitability) are files present a more complex picture requiring
central. Although the spatial orientation of the a more individualized interpretation strategy.
profile, including its peak, is the most salient Circular profiles are typically subjected to
feature, circular profiles also differ structurally summary analyses that reduce the pattern of
in two other meaningful ways. For one, the scores to essential features. Assume that each
expanse of the profile (its interior area) may score (as in Figure 2) is a vector with a cer-
have interpretive value for some measures, tain magnitude (or length) and direction. The
including those that relate to maladjustment length of the vector is given by the score itself,
or psychopathology. An individual reporting and the direction is given by the scales location
a high number of interpersonal problems, (or angular displacement) on the circumplex.
generally, will have a more expansive profile Using simple vector arithmetica method
8 CIRCUMPLEX MODELS

proposed and first demonstrated in the 1950s The structural summary method (see
by the Kaiser Research groupit is possible Gurtman & Pincus, 2003) provides a recent
to calculate a resultant vector that summarizes alternative to the standard approach of cal-
the overall trend. The resultants angle (vector culating vectors. The structural summary
angle, or ) is the circular mean; it indicates method involves fitting a cosine curve to the
the mean direction of the profiles vectors. It profile data. The curve has four parameters,
thus signifies the overall theme of the profile. each having certain interpretive significance:
For the Figure 2 example, the resultant vector elevation, amplitude, angular displacement,
has an angle, , of 320 , which references the and goodness-of-fit. Angular displacement
interpersonal theme of FriendlySubmission. indicates the point, on the circle, where the
The length of the resultant vector (vector curve achieves its maximum value, and thus
length, VL) is related to the circular variance; signifies the predominant interpersonal theme
it indicates the dispersion around the mean. A of the profile. Amplitude indicates the distance
value close to 0 would occur if scores are gen- between the mean level and the peak, and
erally close together, with little differentiation is thus related to the structured patterning
in the profile; higher values occur when scores of the curve. Elevation is the mean level of
are distributed around the peak value, con- the curve; its interpretation depends on the
forming, generally, to the prototypic pattern potential meaning of a general factor in the
of the sinusoidal curve. VL for the example is scales. Goodness-of-fit, or R2 , is a summary
1.19, which, if compared to normative data, statistic that indicates the fit of the model to
would indicate a relatively differentiated and the data, and is now viewed as a measure of
structured profile. the profiles prototypicality. As one application
The implications of vector angle and length of the method, Pincus and colleagues (e.g.,
for interpersonal assessment and diagnosis Przeworski et al., 2011) have used structural
have been developed by a number of theorists summary analysis to identify subgroups of
beginning with those in the Kaiser Research individuals within diagnostic categories having
group. Wiggins, among others, maintained that similar interpersonal features. The diversity of
vector angle is a key variable for interpersonal those derived features suggests interpersonal
diagnosis; that is, for classifying individuals pathoplasticity in how maladjustments are
on the basis of their interpersonal tendencies expressed interpersonally.
into one of eight octants of the circumplex.
These placements suggest certain generalized From Individual to Group
tendencies, which can be the starting point
for creating the more detailed and nuanced Assessments involving the circumplex can be
descriptions of personality functioning asso- extended beyond the individual to situations
ciated with interpersonal assessment. Vector in which group tendencies are of interest.
length, as a measure of profile variability, is Rather than characterizing a single case, the
probably best viewed as a kind of moderator. goal generally is to summarize or represent a
For individuals with high VL, profiles are more group of individuals, for example, a sample of
differentiated around the central tendency; outpatients who meet the diagnostic criteria
hence, summary statements based on the inter- for a certain personality disorder.
personal classification are likely to be more Profile averaging is a simple approach that
fitting of the individual (more prototypic). involves combining individual data into a
Regardless, there is little evidence to support group composite. The circular profiles of indi-
early views that VL is related to interpersonal vidual cases are averaged at each measurement
deviance or maladjustment, or more specif- point (e.g., the eight octant scores) to produce
ically to an interpersonal style that is rigid, a mean profile for the group. Because individ-
intense, or extreme. ual differences are smoothed over, the group
 CIRCUMPLEX MODELS 9

profiles tend to have better fit to the ideal of the As with any frequency distribution, the central
sinusoidal curve, and thus can be meaningfully tendency and dispersion are of general interest.
analyzed for their structural features. As with Using circular statistics appropriate for direc-
individual profiles, the circular mean (angle) tional data (where the values are angles), the
and variability (vector length) are computed to circular mean and circular standard deviation
reveal the interpersonal tendencies associated are calculated; these indicate the center point
with the group; a more elaborate analysis is of the distribution of angles and the degree
possible using the structural summary method of dispersion around that point, respectively.
of modeling the profile. One disadvantage of Often, cases are concentrated and dispersed
profile averaging is that the distinctive fea- around a single point, which simplifies inter-
tures of individual cases are lost through the pretation of these statistics. In the example
averaging process; indeed, conceivably, the of Figure 3, the circular mean is 138 , and
group average may not fit any particular case the circular standard deviation is 19 , which
very well. suggests that most cases fall within a span of
As a graphical technique, circular plots (or 119157 . (Formulas appear in Gurtman &
circular distributions) are also effective ways Pincus, 2003.) It is also possible to calculate
of revealing group tendencies, and have the confidence intervals around the mean, which
advantage that the individual data points are is useful for comparing the distributions of
preserved. To create a circular plot, each per- multiple groups.
sons set of scores is reduced to a summary point
(an angle), which is then plotted on the circle.
Current and Future Directions
(This point is given by the angle of the persons
vector.) Figure 3 provides an example of a The circumplex model continues to be the
circular plot based on 50 hypothetical cases. centerpiece for new and emerging research,

90

135 45

180 0

225 315

270

Figure 3 An example of a circular plot.

10 CIRCUMPLEX MODELS

especially in the interpersonal area. At the same have introduced the concept of interpersonal
time, theoretical refinements and extensions of pathoplasticity to suggest that many forms of
the model have been proposed. psychopathology, including some personality
Wiggins and his colleagues were the first to disorders, are heterogeneous in their interper-
use the circumplex as a kind of factor space sonal expression; rather than being associated
for elucidating the interpersonal contents with a single and possibly unique interpersonal
of various personality constructs and their style, certain diagnostic categories may divide
measures; by projecting a variable of interest into different interpersonal subtypes. They
onto the circle, they showed it was possible showed, for example, that generalized anxiety
to objectively identify a variables interper- disorder is associated with four distinct and
sonal features. Since their pioneering work, clinically meaningful interpersonal clusters
many variables (e.g., autonomy, dependency, on the circumplex. Analysis of other tradi-
narcissism, social anxiety) have been sub- tional diagnostic categories, such as major
jected to this kind of interpersonal analysis. depression, may also yield lawful interper-
As a recent example, Timothy Smith and his sonal subtyping. Moscowitz, Zuroff and their
colleagues have shown that personality risk colleagues have highlighted links between per-
factors associated with coronary heart disease sonality, psychopathology, and intraindividual
(CHD) were located in the HostileDominant variabilitywhat they refer to as flux,pulse, and
and HostileSubmissive quadrants of the circle; spin in an individuals circumplex placements
their work, more generally, reveals that the cir- over time. The tendency to be interpersonally
cumplex can help guide our understanding of variable rather than consistent over time and
how personality factors relate to CHD and its situations may be related to certain features of
precursors. Paulhus and his colleagues showed psychopathology, as they have demonstrated
that the dark triad of narcissism, psychopathy, in their analysis of borderline personality dis-
and Machiavellianism occupy a similar space order (see Horowitz & Strack, 2011, for their
in the HostileDominance quadrant, but are respective chapters).
differentiated by other factors that are orthog- The idea that the circumplex could serve as
onal to the interpersonal plane (see Horowitz a common space for charting the relational
& Strack, 2011, for their respective chapters). patterns and dynamics of paired individuals
Researchers continue to use the circumplex has a long history, dating back to the work
as a base for developing new measures of of the Kaiser Research group. This led to
personality and adjustment, many designed the modern concept of interpersonal comple-
to expand assessment via the model into new mentarity. Kiesler (1983) among others has
areas of interest. For example, as suggested argued that interpersonal behaviors tend to
by Table 2, interpersonal assessment has now pull or invite certain classes of interper-
been broadened by the development of a host sonal responses: correspondence (or similarity)
of new measures that tap into critical aspects of along the HostileFriendly dimension of
interpersonal functioning, both in adults and the circumplex and reciprocity (or opposites)
children The circumplex has also guided recent on the DominanceSubmission dimension.
construction of new measures of core affect Research, especially in the last 10 years, has
and of vocational interests, and may potentially provided general but not unequivocal sup-
be extended to assessment of goal contents and port for complementarity at the level both
values based on recent circumplex models of of individual behavior (interactions) and of
these domains. interpersonal styles. New methodologies and
Applications of the model to our understand- technologies have recently been introduced
ing the relationship between personality and that promise to provide better and more refined
psychopathology continue to evolve and be tests of complementarity theory, and its appli-
refined. For example, Pincus and his colleagues cations to social behavior and to relationship
 CIRCUMPLEX MODELS 11

quality (e.g., Sadler, Ethier, Gunn, Duong, & Horowitz, L. M., & Strack, S. (2011). Handbook of
Woody, 2009). interpersonal psychology: Theory, research,
Although the interpersonal circumplex assessment, and therapeutic interventions.
model has remained fundamentally unchanged Hoboken, NJ: John Wiley & Sons.
since its inception, a recent trend Hofstee, W. K., de Raad, B., & Goldberg, L. R.
(e.g., Pincus, Lukowitsky, & Wright, 2010) (1992). Integration of the Big Five and
circumplex approaches to trait structure. Journal
has been to reconceptualize its two core
of Personality and Social Psychology, 63, 146163.
dimensions as Agency and Communion, terms
Kiesler, D. J. (1983). The 1982 interpersonal circle:
taken from Bakans well-known work on A taxonomy for complementarity in human
the dual modalities of existence. Broader transactions. Psychological Review, 90, 185214.
concepts than simply Dominance and Nur- Leary, T. (1957). Interpersonal diagnosis of
turance, Agency connotes themes of power, personality; a functional theory and methodology
control, expansion, and individuation, whereas for personality evaluation. Oxford, U.K.: Ronald
Communion suggests connections to others, Press.
including intimacy, love, and friendship. This Pincus, A. L., Lukowitsky, M. R., & Wright, A. G. C.
change puts the circumplex model within a (2010). The interpersonal nexus of personality
broader context in which interpersonal behav- and psychopathology. In T. Millon, R. F. Krueger,
ior is aligned to the major tasks, goals, and & E. Simonsen (Eds.), Contemporary directions in
purposes of human life. psychopathology: Scientific foundations of the
DSM-V and ICD-11 (pp. 523552). New York:
SEE ALSO: Classification of Disorders versus Guilford Press.
Individuals; Construct Validity; Definitions of Plutchik, R., & Conte, H. R. (1997). Circumplex
Abnormality; Dimensional versus Categorical models of personality and emotions. Washington,
Models of Psychopathology; Factor Analysis; DC: American Psychological Association.
Interpersonal Circumplex; Pathoplasticity; Per- Przeworski, A., Newman, M. G., Pincus, A. L.,
sonality Disorders; Rational/Theoretical Approach Kasoff, M. B., Yamasaki, A. S., Castonguay, L. G.,
to Test Construction; Self-Report Questionnaires; & Berlin, K. S. (2011). Interpersonal
Structural Equation Modeling; Wiggins, Jerry S. pathoplasticity in individuals with generalized
(19312006) anxiety disorder. Journal of Abnormal Psychology,
120, 286298.
References Sadler, P., Ethier, N., Gunn, G. R., Duong, D., &
Woody, E. (2009). Are we on the same
Benjamin, L. S. (1996). Interpersonal diagnosis and
wavelength? Interpersonal complementarity as
treatment of personality disorders (2nd ed.). New
York: Guilford Press. shared cyclical patterns during
Gurtman, M. B., & Pincus, A. L. (2003). The interactions. Journal of Personality And Social
circumplex model: Methods and research Psychology, 97, 10051020.
applications. In I. B. Weiner, J. A. Schinka, & V. F. Schwartz, S. H., & Bilsky, W. (1990). Toward a
Velicer, V. F. (Eds.), Handbook of psychology: theory of the universal content and structure of
Research methods in psychology (Vol. 2, pp. values: Extensions and cross-cultural
407428). Hoboken, NJ: John Wiley & Sons. replications. Journal of Personality and Social
Browne, M. W. (1992). Circumplex models for Psychology, 58, 878891.
correlation matrices. Psychometrika, 57, Tracey, T. J. G. (2000). Analysis of circumplex
469497. models. In H. E. A. Tinsley & S. D. Brown (Eds.),
Guttman, L. (1954). A new approach to factor Handbook of applied multivariate statistics and
analysis: the Radex. In P. F. Lazarsfeld (Ed.), mathematical modeling (pp. 641664). San
Mathematical thinking in the social sciences (pp. Diego, CA: Academic Press.
258348). New York: Free Press. Wiggins, J. S. (1979). A psychological taxonomy of
Horowitz, L. M. (2004). Interpersonal foundations of trait-descriptive terms: The interpersonal
psychopathology. Washington, DC: American domain. Journal of Personality and Social
Psychological Association. Psychology, 37, 395412.
12 CIRCUMPLEX MODELS

Further Reading and measurement of interpersonal behavior. In

W. M. Grove & D. Cichetti (Eds.), Thinking
Kiesler, D. J. (1996). Contemporary interpersonal
clearly about psychology: Essays in honor of Paul
theory and research: Personality, psychopathology,
E. Meehl, Vol. 1: Matters of public interest; Vol. 2:
and psychotherapy. Oxford, U.K.: John Wiley &
Personality and psychopathology (pp. 89113).
Sons.
Minneapolis, MN: University of Minnesota Press.
Wiggins, J. S. (1991). Agency and communion as
conceptual coordinates for the understanding