ChapterPDF Available

Social Intelligence--A Review and Critical Discussion of Measurement Concepts.

Authors:

Abstract

This chapter provides a description of theoretical and empirical approaches to sketch the nature and scope of the social intelligence construct. Detailed attention is given to the empirical investigations of the structure and validity of this construct. Research designs and outcomes of these studies are described along a classification of the applied measurement procedures that affect the validity of the studies. Our considerations support the assumption that method-related variance can explain a substantial part of the results. Therefore, we suggest applying multitrait-multimethod designs to control for this bias. In addition, past theoretical and empirical accounts are integrated into a performance model of social intelligence with the main focus on the cognitive facets of the construct: social understanding, social memory, social perception, and social creativity. Some empirical data is provided that supports this model. The chapter concludes by discussing important conceptual and measurement issues for future research: the importance of thoroughly specifying the intended measurement construct and the corresponding task requirements, the construction of tests that reflect the real-life significance of the construct, and a well-considered validation strategy (construct and predictive validity) that also takes related constructs like emotional intelligence into account. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Emotiona I
I
ntel
Iigence
An
lnternational
Handbook
Edited by
Ralf Schulze
Richard
D.
Roberts
HOGREFE
Library
of Congress
Cataloging-in-Publication
Data
is available
via
the
Library
of Congress
Marc
Database
under
the
LC
Control Number
.2004114391
Library and
Archives
Canada
Cataloguing
in Publication
Emotional intelligence : an international
handbook / Ralf Schulze,
Richard D. Roberts
(eds.)
Includes bibliographical
references.
ISBN 0-88937-283-7
1. Emotional intelligence.
I. Schulze, Ralf II.
Roberts,
Richard
D.
t
O 2005
by
Hogrefe &
Huber
Publishers
PUBLISHING OFFICES
USA: Hogrefe & Huber
Publishers, 875 Massachus€tts Avenue, 7th Floor,
Cambridge,
MA 02139
Phone
(866)
823-4726,
Fax
(617)
354-6815; E-mail info@hhpub.com
EUROPE:
Hogrefe & Huber Publishers,
Rohnsweg 25,37085
Göttingen,
Germany
Phone
+49
551
49609-0, Fax
+49
551 49609-88, E-mail
hh@hhpub.com
SALES
& DISTRIBUTION
USA: Hogrefe & Huber Publishers, Customer Services Department,
30 Amberwood Parkway,
Ashland,
OH
44805
Phone
(800)
228-3749, Fax
(419)
281-6883,
E-mail custserv@hhpub.com
EUROPE: Hogrefe & Huber Publishers, Robnsweg 25,37085
Göttingen, Germany
Phone
+49
55i
49609-0, Fax
+49
551 49609-88, E-mail hh@hhpub.com
CORPORAIE OFFICE
Hogrefe
& Huber
Publishers, Inc., 218 Main Street, Suite
485,
Kirkland WA
98033
OTHER OFFICES
CANADA:
Hogrefe & Huber Pubiishers, 1543 Bayview Avenue,
Toronto,
Ontario M4G
3B5
SWITZERLAND: Hogrefe
& Huber Publishers, Länggass-Strasse
76,
CH-3000 Bem 9
Hogrefe & Huber Publishers
Incorporated
and registered in the State of Washington,
USA, and in
Göttingen, Lower
Saxony,
Germany
No
part
of
this
book
may
be
reproduced, stored in a retrieval
system
or
transmitted,
in any form
or by
any means, electronic, mechanical,
photocopying,
microfrlming,
recording
or otherwise,
without
written permission from the
publisher.
Printed
and bound in Germany
ISBN 0-88937-283-7
wffi
Social Intelligence-A
Review
and
Critical
Discussion
of Measurement
Concepts,
Susanne Weis
Heinz-Martin
Süß
Otto-von-Guericke
University Magdeburg,
Germany
Summary
This
chapter
provides a
description
of
theoretical
and
empirical approach-
es to
sketch
the nature and
scope
of
the
social intelligence
construct.
De-
tailed attention
is given to the
empirical investigations
of the structure and
validity
of this
construct. Research
designs and
outcomes
of these
studies
are
described
along a
classification
of
the applied
measurement
proce-
dures
that affect
the validity
of the
studies. Our
considerations
support
the assumption
that method-related
variance
can
explain a substantial
part of the results.
Therefore,
we
suggest applying multitrait-multimethod
designs to
control for
this bias. In
addition,
past theoretical and
empir-
ical
accounts
are integrated
into a
performance
model
of social intelli-
gence with
the main
focus
on
the
cognitive facets
of the
construct:
so-
cial understanding,
social memory
social
perception, and
social creativ-
ity. Some
empirical
data is
provided that
supports this
model. The
chap-
ter concludes
by discussing
important
conceptual and
measurement
is-
sues for future
research:
ihe importance
of thoroughly
specifying
the in-
tended measurement
construct and
the corresponding
task
requirements,
the construction
of tests that reflect
the real-life
significance of the
con-
struct, and
a well-considered
validation
strategy
(construct
and predictive
validity)
that also
takes related
constructs like
emotional
intelligence
into
account.
204
Sociallntelligence
10.1
INTRODUCTION
Psychological
research
has been concerned with
the
study of
human intelli-
gence
for
over
a century.
From its inception
on up
through the
present time,
academic
(i.e.,
abstract
or
general)
intelligence represents
the
most examined
and
clearly
defined
construct
investigated as
part
of this
scientific enterprise.
Recently,
however,
the
concept
of human intelligence
has
been expanded
with
the
introduction
of
so-called
new intelligences,
that
is,
social, emotional,
and
practical.
The
chapters
contained in this
book
give detailed
attention
to the-
oretical
and
measurement
issues
of emotional
intelligence.
Nevertheless,
a
comprehensive
scientific
treatment
of emotional
intelligence
canlot ignore
the
appareltly
related
concept
of social
intelligence.
Apart
from
comrnon accep-
tance
that
both concepts
receive in
all
parts of
contemporary
society
(see
e.[.,
Gardner, 1983;
Goleman,7995;
Matthews,
Zeidner,
& Roberts, 2002),
substan-
tial
overlap
can also
be
perceived
in theoretical
definitions
and
measurement
approaches.
Relying
on the
long
research tradition
of social intelligence,
fu-
ture
research
in
both fields
may
benefit
from
past lessons
learned. In
return,
it
seems
essential
for research
on social intelligence
to
profit from the
scientific
interest
and
concerted
endeavors
that are
currently
concentrated,
to
a large
measure/
on emotional
intelligence.
Despite
the long
tradition
of research
on
social intelligence,
both
theory
and measurement
issues
remain unresolved
at
a
(fairly)
low
level
of sophistication.
Further
examination
is also
indispensable
to
eventually
identifying
a viable
and discriminable
domain
of social intelii-
gence.
The
purpose
of
this
chapter
is to
provide
an
overview
of scientific
investi-
gations
in the
domain
of
social intelligence.
Comparable
approaches
date back
many
years
(see
Orlik, 1978;
Waiker
&
Foley,
1973,
for reviews).
Since then, the
research
landscape
has
changed
with respect
to
some aspects.
For
example,
empirical
studies
have
begun
to make
use of multitrait-multimethod
designs,
used
structural
equation
modeling
for data
analysis,
and various
situational
judgment
testing
paradigms
to assess the
so-called "construct
space". ln this
chapter,
we review
the
literature
on social intelligence,
including
findings
that
have
been
obtained
recently
with these newer
approaches.
TO.2
THEORIES
AND
DEFINITIONS
OF
SOCIAL
INTELLIGENCE
Thorndike (1'920)
postulated
a framework
of human
intelligence
differentiat.
ing between
ideas,
objects, and
people as the
contents
that human
intellect
has
to deal
with.
In
other words,
he discriminated
between
academic,
mechani-
cal, and
social
intelligence.
In this
framework,
Thorndike (1920)
defined
the
latter
as
"the
ability to
understand
and
manage
men and
women,
boys and
girls, and to
act
wisely in
human
relations"
(p.
z2B).
Thorndike's
idea
of
social
intelligence
is
still fundamental
to,
and
even more
extensive
than,
any
other
given definition.
Indeed,
most
contemporary
research
efforts appear
to
cite
THEORIES
AND
DEFIMTIONS
OF SOCIAL
INTELLIGENCE
Operations
Contents
Products
Figure
10.1
Structural Model
of
Human krtellect
(Guilford,
7967); the
domain
o{
social intelligence
ftehavior)
is
highlighted in bold letters and
the lines.
(and
subsequently rely)
on
this
definition when
examining the
concept
of so-
cial intelligence.
Notably, his distinction between
cognitive
(i.e.,
understand
other people) and
behavioral
(i.e.,
to act wisely
in human relations)
compo-
nents has been
specified in
only one other definition
of social intelligence.
Thus,
Vernon
(1933)
defined social intelligence as
"knowledge
of
social mat-
ters and insight
into the moods
or personality traits
of strangers"
(cognition)
and
as the ability
to
"get
along with
others and ease in
society"
(behavior) (p.
44).
Other definitions
focus either
on cognitive or behavioral
aspects.
Some
of
these
definitions,
along with
their chief
protagonists, are listed as
follows:
"the
abiLity
to
get
along
with others"
(Moss
&
Hunt,
7927,p.108);
"judge
correctly
the feelings,
moods,
and motivation
of
individuals"
(Wedeck,
1947,
p.133);
"abllity
to
judge
people
with
respect to feelings,
motives, thoughts,
intentions,
attitudes,
etc."
(O'Sullivan,
Guilford, &
deMille,
7965,p.6);
"individuals
fund
of knowledge
about the
social world"
(Cantor
&
Kihlstrom,1987).
Indeed,
the
establishment and
subsequent empirical
application
of broad
theoretical
frameworks
of social intelligence appear
scant in the
literature.
The
most
prominent
and broadest
conceptualization was
introduced
by
Guilford
(7967).
In his
Structural Model
of Human lntellect, the
three dimensions
of
operations,
contents,
and products are
pivotal. The dimension
of operations
de-
scribes
the
cognitive requirements
participants
need
to accomplish
a task and
contains
five
elements. The content dimension,
with
four elements,
refers to
the
properties
of task
material. Finally, tlne product
dimension
comprises
six el-
ements,
each describing
a type of outcome associated
with a
certain task.
The
model, which
relies
on a
complex interaction between
these three
dimensions.
is
depicted in
Figure
10.1.
Cuilford's
conceptualization
resulted in
120 factors that
described
distinct
human
iltellectual
abilities.
For
Guilford,
the behavioral
content facet, along
with
its
cross-classification
in terms of both operations
and
products, repre-
sented the
domain
of social intelligence, thus
comprising
30
(=
5
x
6) distinct
abilities
as
demarcated in
Fieure 10.1. Guilford and
his colleasues
(Hendricks.
Cognition
Memory
Divergent
Production
Convergent
Production
Evaluation
Units
Relations
Systems
Trarsformatlons
Implications
Figural
Symbolic
Semmtic
Behavioral
206
Socialfntelligence
Guil{ord,
& Hoepfner,7969;
O'Sullivan
et al.,
1965)
focused
on
the operational
domains
of
cognition
and
diaergent
production
to
construct
possible measures
of
social
intelligence. O'Sullivan
and Guilford's
efforts
in the domain
of behav-
ioral
cognition
resulted
in two test publications:
the
Factor Test
(O'sullivan
& Guil{ord,
7966) and
the
Four
Factor Test
(O'Sullivan
&
Guilford, 1976)
of So-
cial
Intelligence. The task
material
consisted,
above all,
of
pictures.
Only a few
purely
verbal
measures
were
constructed.
These test
batteries
of the cognitiae
behavioral
domain received wide interest in
the research
community.
At
about
the
same
time, Hendricks et al.
(7969)
specified the
domain
of diaergent
or
"cre-
ative"
production
of
behavioral
contents according
to the
six
possible products.
Thus,
they
postulated
the
following
constructs: the
ability to
engage in
behav-
ioral
acts that
communicate internal
mental
states
(units),
the ability
to
create
recognizable
categories
of behavioral acts
(classes),
the ability
to
perform
an
act that
has a bearing
on what another
person
is
doing
(relations),
the ability
to
maintain
a sequence
of
interactions
with
another
person
(systems),
the ability
to alter
an expression
or a
sequence
of expressions
(transformations),
and
the
ability
to
predict many
possible outcomes
of a
setting
(implications).
Although
'
ttüs domain
of
"creative"
social intelligence
appeared
to
be a meaningful
facet
of social intelligence
performance, further
reaching
investigations
relying
on
these
types
of operationalizations
have not
been forthcoming
in
the literature.
Both Thorndike (1920)
and
Guilford
(7967),
in their
theoretical
frameworks,
,
located
the domain
of social intelligence
as equal and
discrirninable
on
one
level with
the traditional
domain
of
academic
inteliigence.
However,
empir-
ical
results
suggesting the
autonomy
of social
intelligence
from academic
in-
tellectual
abilities
are equivocal,
seemingly
dependent
on
the
measurement
procedures
adopted. Indeed,
empirical evidence
for the
relation
of social to
emotional
(or
practical)
intelligence
barely
exist.
Instead,
the relation
of so-
cial to
emotional
intelligence
has largely been
examined
using rather
specific
measures
of
social and emotional
skills. For
example,
Davies,
Stankov,
and
Roberts
(1998)
operationalized
social intelLigence
with
the
hrterpersonal
Per-
ception
Task-15
(IPT-15
Costanzo
& Archer,
7993), a
performance
measure
of
social
perception
presented on
videotape.
Additionally,
they
employed
a
per-
formance
measure
of emotional intelligence,
that
is, the
Emotion
Perception
in Faces
Test
(Mayer,
DiPaolo,
& Salovey,
1990).
However,
the two
measures
correlated
r
:
-.09.
A
subsequent factor
analysis
showed
that these
measures
had
bipolar loadings
on one factor.
Given this
constitutes
one of the few
pub-
lished
studies of its kind,
we
contend that
empirical
approaches
investigäting
the
relation
of the
two constructs,
particularly
those
that rely
on
performance
data,
are not
readily apparent
in the literature.
10.3
THE
ASSESSMENT
OF
SOCIAL
INTELLIGENCE
10.3.1
Multiple
Test Batteries
of Social
lntelligence
Two broad
attempts
to assess
social
intelligence
using
comprehensive
test bat-
teries
are
discussed in
the
passages that
follow.
These
represent
approaches
THE ASSESSMENT
OF SOCIAL INTELLIGENCE
207
that seem to address problematic
features put
forward by various critiques of
social
intelligence
(Orlik,
7978;
Walker
&
Foley,
1973).
These
attempts aside,
we note from
the outset
that
the
idea
of
developing
tests
of social
intelligence
from
an a priori theoretical
framework
remains
an
outstanding problem
in the
long research tradition on social
intelligence.
One of the first broad
measures of social inteiligence
constructed
is
the
George
Washington Social
Intelligence Test
(GWSIT
Moss, Hunt, Omwake,
&
Woodward, 1955).
This test
is based on
the authors'
definition of social in-
telligence
as
"the
ability
to
get along
with
others"
(p.
108). A revised
(short)
form of
the test,
containing
five
subscales,
comprises the following abilities:
fudgment
in Social Situations: Find
possible
solutions for a
social problem.
Memory
for Names and Faces: Recognize target
photographs previously
studied
and
presented
later among a larger
group of
photographs.
Observation of Human
Behavior:
Answer questions about human function-
ing
on
a true-false basis.
Recognition of the
Mental
States
Behind Words:
Choose the correct mental
state or emotion, among four, reflected in a given
statement.
Sense
of Humor:
Select
the best ending to a
joke.
Despite the authors'claim,
performance
in these
subscales appears
to
be
less
dependent on socially intelligent behavior
and
more on understanding the irri-
portance of certain social
milieu. Moreover, as
Orlik
(7978)
points out, several
validation
studies show
that variance in performance data may be
explained,
to a large extent, by verbal measures of
academic
intelligence. \Atrhereas
per-
formance in the
GWSIT
has been
shown
to correlate up
to
.70 with
academic
(verbal)
intelligence, correlations with other social intelligence indicators
show
no evidence for convergent validity.
A
second
major approach to the assessment
of socia-l intelligence was in-
troduced by Guilford and colleagues/ under the framework
provided
by the
previously elucidated Structural
Model of Human Intellect
(Guilford,
7967).
The following list briefly describes some examples of tasks
out
of the and
the Four Factor Test
of Social
lntelligence
(O'Sullivan
& Guilford, 7966,1976)
and their classification in that model:
Expression
Grouping
(Classes):
Participants find
one
facial
expression, out of
four alternatives, which best fit a group of three
other
facial
expressions.
Missing Pictures
(Systems):
Participants are
presented with a sequence of
events,
pictured
in
photographs,
and have to
complete the sequence by
choosing the correct
last
photograph.
Missing Cartoons
(Systems):
Participants
are
required to fill-in a blank, in a
sequence
of
cartoons,
by
selecting
the
correct cartoon out of four choice
alternatives.
Picture Exchange
(Transformations):
A
sequence of photographs is
presented
that tells a
story.
Participants are required to replace
one
marked
photo-
graph of
this
sequence,
with
one of four
alternatives,
in
order
to
give
the
story a different meaning.
208
Social lntelligence
Social
Tianslations
(Transformations):
This
test
is
the
only
verbal
social intelligence
in
this
battery.
Participants
ut"
drr"^
u.
o
toon series.
As O'sullivan
et
al.
(1965)
recognized,
these
tests
were
measuröi
nitive rather
than
behavioral
skills.
The authors
themselves
reporterJ-;
stantial
correlations
with
general
intellectual
abilities
(O'sullivan
&
1966).
More
recent
studies
have
focused on
the
construct
validity
of
batteries.
For
example,
Probst
(1982)
applied the
Six
Factor
tesi
in
sive
study
of social
intelligence,
finding
empirical
support
for
an
ind
ability
construct.
However,
factor analysis
did
not
yield
a commonj
telligence
factor
comprising
different
types
of assessment
methods.
Ir
study,
Riggio,
Messamer,
and
Throckmorton
(1991)
neither
found
evii
convergent
nor
for
discriminant
construct
validity.
They
applied
thö'i
tor
Test of Social
Intelligence,
along
with a measure
of
academic
ihi
that is,
the
Wechsler
Adult
Intelligence
Scale-Revised
Edition
(W
cabulary
Subscale
(Wechsler,
1981). Furthermore,
the
Social
skills In
(SSI;
Riggio
,1989)
was
administered
as a measure
of
self-reported soü
ln an exploratory
factor
analysis,
the
subscales
of
the
Four Factor Te
on
one
fäctor
wlih the
WAIS-R,
showing
near to
zero
correlations wii
Thus,
neither
convergent
nor
discrirninant
construct
validity were
in this
investigation.
1:0.3.2
Individual
Tests
of
Social
Intelligence
Empirical
approaches
that occurred
after
Walker
and Foley's
(797,3)
tik's
(1978)
summarizing
works, somewhat
surprisingly,
appear
lesl
guided
than
those
discussed
in the
aforementioned
passages.
More
üut seemingly
related
concepts
like social
skills,
nonverbal
decodin-g
nonverbal
comrnunication
skills
have subsequently
been
operationalii
dicators of
social
intelligence
(Barnes
& Sternberg,7989; Feldman,
Tol
Coats,
1999;
Riggio
,7986;
Sternberg
& Smith,
1985).
With
every
cons
of
the contributive
value of
these
investigations,
it
seems
somehow
ment made
between
a pair of people,
in a
defined
Jocial
rela
have
to choose
one pair
of people out of
three
altematives,
for:w
given
statement
has a
different meaning.
(i.e.,
performance
vs" self-report
data).
tn the
following
subsections,
vide a critical
analysis
of
tests represented
by this
classification
Cartoon
prediction
(Implications):
participants
are
required
to
seleciu
toon, out
bf
three
alternatives,
that most appropriately
completr
cult
to
find appropriate
indicators
of
these concepts, as
task
requirer
more
explicit
and
less complex.
It
appears
necessary
to classify
these
t
.o*
upprouches
in order
to facilitate
the
interpretation of
research
r'
this
puipose,
operationalizations
can be
cross-classified
along
two
dj
One
is defined
by
the content
under examination
(i.e., cognitive
vs'
social
skills),
while
the other
dimension
describes
the
method
of
THE
ASSESSMENT
OF
SOCIAL
INTELLIGENCE
that the
treatment
of self-report
measures
wiil
include
attempts
to assess
both
the
cognitive
and
behavioral
dimensions
of social
intelligence.
Cognitiae
performance
measures.
Keating
(1978)
employed
three
verbal
in-
dicators of social
iltelligence performance
designed
previously
(e.g.,
Chapin
Social
Insight
Test,
Chapin,7967;
Gough,
1968), as well as three measures
of
academic intelligence
(both
verbal
and
nonverbal
material). Neither correla-
tional
nor
factor analytic results supported
construct validity. Within-domain
coruelations
did not
exceed
across-domain correlations, and no
coherent fac-
tor
structure was
observed.
Furthermore,
the
social intelligence performance
measures
did
not predict
effective social functioning
(assessed
by
peer-reports)
to a larger extent than academic intelligence. Sternberg and colleagues
(Barnes
& Sternberg,7989; Sternberg
& Smith, 1,985) operationalized the
concept of
nonverbal decoding skills
as an indicator
of social intelligence. Th"y devel-
oped
two
tasks
relying
on similar principles.
One was the
so-called
"Couples"
Test, which
contained
photographs of
heterosexual
couples
that were
either in
a close relationship
or
were strangers. Participants had to
judge
each
photb-
graph
for
the kind of
relation depicted
(i.e.,
close
relationship or strangers).
The
second
task
consisted
of
photographs
of a
supervisor
and his
or
her
su-
pervisee. In this instance, participants
had to
judge
who the supervisor was.
Barnes and
Sternberg
(1989)
used
self-report
inventories
of social
competence,
as well as
performance
measures
of
academic intelligence,
to
ascertain
con-
struct
validity.
Correlational
analyses showed an unequivocal
pattern with
only
significant
convergent and non-significant discriminant validity
coeffi-
cients.
Along
with studies
by Riggio
et
al.
(1991)
as well as
O'Sullivan
and
GuilJord
(1966),
these
results
alone only allow ambiguous conclusions about the valid-
ity of
social
intelligence based on performance
measurement.
At first blush, it
seems
that applying verbal
performance
measures results in
substantial over-
lap between
social
intelligence and
academic
(especially
verbal)
abilities
(Keat-
tng, 1978). Thus, investigations using nonverbal measures as indicators
of so-
cially intelligent
performance
succeed somewhat
better in identi{ying
a con-
ceptually
coherent
domain of social intelligence
(O'Sullivan
&
Guilford,
1966;
Barnes
& Sternberg,7989). However,
this result
is
not
always demonstrated
(Sternberg
& Smith, 1985;
Riggio
et
al., 1991).
The difficulties
of
both verbal and nonverbal cognitive
performance mea-
sures in defining an unequivocal social
intelligence
construct could be attri-
buted to a methodological
problem.
According to
Schneider,
Ackerman, and
Kanfer
(1996),
certain
characteristics of social cognitive tasks increase the
over-
lap with
academic intellectual abilities
by matching
their typical measurement
features.
These characteristics
include: when
participants encounter
social
stimuli that are inconsistent with their expectancies, when
participants
are
faced with novel stimuli, and
when
participants
are faced with highly
struc-
tured
tasks
(Schneider
et
aI.,7996,
p.
a69). Among
cognitive performance mea-
sures
of social
intelligence, above all, those relying on verbal material
seem
to
meet
all three criteria.
The
sequential
type
of presentation inherent to writ-
209
210 SociallntelliSence
tenlanguagedoesnotSeemtobeanadequateoperationalizationofmoreor
less
coäplä* social
stimuli'
Instead,
writfen
language
liPeals
to be
distinct
from
socially
relevant
stimuii
found
in
real-life
seitings.
Plausibly,
this
type of
presentation
conJronts
the
participants
with novel
(or thus
far
unexperienced)
ätimull
that implicate
cognitive
functions
that parallel
those
necessary
for the
accomplishme.tt
of
academic
intelligence
tests'
B'ehaaioral
performance
measures'
Ford
and
Tisak
(1983)
app-lied.a.perfor-
mance
measure
of
socially
intelligent
behavior
as
an
indicator
of
social
intelli-
sence.
Participants'behavior
in in
interview
situation
was
rated
along
certain
E
ir"Ctl.g.,
tie
ability
to speak
effectively,
to be
appropriately
responsive
to
the
interview"r',
qn"riio.rs,
to
displuy
appropriate
nonverbal,behaviors)'
Ad-
Jitionally,
the
authärs
assessed
s"ü-
ut abiftei-reported
socialüehavioral
skills
and
academic
inteliigence.
Correlational,
as
well
as
factor
analytic,
results
sug-
gested
a
distinct
,oJiul
ittt"lligence
construct.
Within-domain
correlations
ex-
ceeded
across-domain
correläons,
and
social
intelligence
measures
loaded
on
a separate,
interPretable
factor'
A
comparable
study
was
conducted
by
Frederiksen,
Carlson,
and
Ward
tfq84l
Again,
perforÄut
ce
in
an
interview
setting
lerv-ed
as
an indicator
of
,o.iui
it
titig"n"".
Participants
had
to
take
the
role of
a doctor
who
w.as,inter-
viewing
his/her
patient.
Additlonaily,
Frederiksen
et
al.
(1984)
applied
vari-
ous
measures
of
academic
intelligence
and
problem-solving
abilities'
Results
st
o*"d
only
a
few
substantiai
ärrelationJ
between
interview
performance
and
academic
intelligence
measures.
These
correlations
were
partly
negative
in
sign,
suggesting
ttat
high
acadernic
intelligence
was
accompanied
by
low
social
behavioral
skills.
Finally,
stricker
and
Rock
(1990) applied
a
techniqrre
sirnilar
to
the
inter-
view
settings
d.escribed
thus
fär. Striätcer
developed
his
own
measure
of so-
.iuffy
it
t"ffigent
behavior,
the
Interpersonal
Competence
Inventory
(ICI)'
The
ICI
was
based
on
video
scenes
"ot
tuit
ing
an interview
situation
between
a
subordinate
and
his
superior.
In
the
Replies
section,
participants
had
to
re-
spo.,a
orally
to
the subärdinate
in
place'bf
a superior.
Answers
were
judged
i^
t"r^,
of
äffectiveness
and
originätity.
In
the
|udgment
section,
participants
had
to
write
down
their
descripüon
of
ihe situation
and
its important
features'
The
performance
criterion
was
accuracy.
Conceptually,
the
Replies
section
op-
erationalized
socially
intelligent
behavior,
whereas
the
]udgment
section
as-
sessed,
for
the
mosi
part,
cögnitive
skills.
Along
with
the-ICI,
Stricker
and
Rock
(1990)
ussess"d'nor,-rr"öd
social
skills,
acaäemic
intelligence,
and.self-
,"po.t"d
social
intelligence.
Results
from
correlational
and
multidimensional
r.är*g
analyses
show"ed
no
coherent
structure
either
within
the
domain
of
so-
cial
in"teliigence
or
concerning
the
relation
of
social
intelLigence
measules
to
academic
ilt"llig"r,.".
Social
ät"llig"t.u
performance
measures
correlated
in-
consistently
witl
each
other
(betwe€il.
/
:
-.08
and
.37)
and
the
judgment
section
of
the
ICi
correlated
substantially
with
the
verbal
measule
of
academic
intelligence
(r
:
.30).
THE
ASSESSMENT
OF SOCIAL
INTELLIGENCE 211
The interpretation
of results in
the aforementioned
studies
allows us to
draw
some
tentative
conclusions
concernjng
the validity
of
social
intelligence.
Both
Ford and
Tisak
(1983)
as well
as
Frederiksen et al.
(1984)
succeeded
in
sepa-
rating
social
from
academic
intelligence
thus
proving discriminant
construct
validity.
However,
Frederiksen
et al.'s
(1984)
findings-partly
negative
corre-
lations
with
academic
performance-raise doubts about
the nature of the
per-
formance construct of social intelligence. It should be
expected that a
so-called
"intelligence"
construct would
at least be slightly
positively correlated with
traditional measures
of academic intelligence. Furthermore,
a
strict
account
would
note that the
generalization of
the findings
is restricted to
a rather
spe-
cific
(albeit
practically
meaningful)
instantiation
of
the
social context in
which
humans interact:
interview settings. Finally,
it must be
stated
that
the conver-
gent construct validity was not
convincingly
proven
in
these
studies,
neither
for the restricted interview
settings
as indicators
of social inteiligence
nor for
a
possibly more
general social
intelligence
construct.
Self-reported
social intelligence. Numerous
studies have
applied self-repört
inventories
as
measrues of social intelligence. In
several
of
these investiga-
tions
self-reported
social skills serve as
psychologically meaningful validation
criteria
(Barnes
&
Sternberg,I9B9;
Ford
&
Tisak,
1983; Frederiksen
et a1.,79,84;
Riggo
et al., 1991). However,
there are a large number
of
studies
that
rely
only
on self-reported social
skills
as indicators
of
social
intelligence
(Brown
&
Anthony, 1990; Marlow
e, 7986; Riggio, 1986).
We
have already described
Riggio et al.'s
(1991)
study in the context
of
the
O'Sullivan and Guilford
(1976)
test battery. ln
this investigation,
the subscales
of the Four Factor
Test of Social Intelligence loaded
on one factor together
with
academic
intellectual abilities, whereas
the subscales
of the Social
Skills
In-
ventory
(SSI;
Riggio, 1.989) loaded
on
a
separate factor.
Self-reported social
skills and
performance measures of social intelligence
did not
correlate sub-
stantially and
only one correlation
(viz.,
the
Social Translation
Subtest with
the
SSi) reached
significance.
FIowever,
other
studies employing both
self-
reported
social skills and social intelligence
performance tests report
evidence
of convergent validity
(Barnes
& Sternberg,l9B9;
Ford & Tisak, 1983).
Riggio
(7986)
validated the
SSI
using the
traditional
personality
scales of
the
16 Personality
Factor
Questionnaire
(16PF).
The
SSI contained
six sub-
facets
that resulted from a
cross-classification
of contents
(viz.,
social
vs.
emo-
tional
contents) and
postulated ski11s
(viz.,
sensitivity,
expressivity, and
con-
trol).
Sumrnarizing
these results, the
SSI subfacets correlated
substaltially
with various
personality
traits
(e.g.,
social expressivity:
outgoing,
happy-go-
lucky, venturesome,
group dependent;
social sensitivity:
affected by feelings,
shy,
astute,
apprehensive,
conservative, tense, undisciplined).
Moreover,
par-
ticipants
scoring high
on
the different
SSI
subfacets could be described
by a dif-
fering
personality
structure.
According
to Riggio
(7986),
these results
proved
the convergent validity
of SSI as a measure
of
nonverbal
social skills. His con-
clusion
was also
supported
by further validity
evidence: high
scorers on the
SSI tended
to report more
socially effective
behavior
and
richer
social contacts
212 Sociallntelligence
(Riggio,
1986).
The
aforementioned
findings
of
Riggio
et
al.
(1991)
put
some
ambiguity
into
this
interpretation.
In terms
of Schneider
et
al.'s
(1996)
crit-
icismi, the
subscales
of
the
Four
Factor Test
of
Social
Intelligence embodied
operationalizations
that
were
conceptually
too close
to
academic
intelligence
performance
measures.
Thus,
they represent
no valid operationalization
of the
iocial
intelligence
ability
construct.
From
another
viewpoint,
it is also possible
that
these
results
could
be
attributed
to
common
method-related
variance
in
'self-report
and perfotmance
data.
To examile
these propositions
more closely,
two other
studies
are worth
noting.
Marlowe
(1986)
operationalized
social
intelligence
via a
self-report
in-
strument.
He intended
to demonstrate
that social
intelligence
would show
independence
from academic
intelligence.
Secondly,
Marlowe postulated
the
multidimensionality
of
social
intelligence.
He extracted
four dimensions
from
the empirical
literature.
Along
these
dimensions,
social
intelligence
includes
social
interest, social self-efficacy,
empathy skills,
and
social performance
skills.
Factor
analytic results of
the social
intelligence
measures yielded
five separpte
factors
labeled pro-social
attitudes,
social skills,
empathy
skills, emotionality,
and
social
anxiety.
The
postulated
dimensions
could
thus not be instantiated,
though
there
was clear evidence
for
the multidimensionality
of social intel-
ligence. Correlational
analyses
suggested
construct
independence,
showing
near
to zero correlations
with
acadernic
intellectual
abilities assessed
by
per-
formance
data. Anyway,
evidence
for the convergent
constluct
validity
was
again
missing.
Subsequently,
Brown &
Anthony
(1990)
found similar
results.
They as-
sessed
self-
and peer
ratings
of
both social
behavior
and
personality
traits,
along
with general
intellectual
performance.
A factor
analysis
resulted in
a
clearly
defined
factor structure.
The
three factors
were identified
as:
(a)
acad-
emic
intelligence,
(b)
peer
ratings
of
both social
behavior
and
personality,
and
(c)
self-reported
social
behavior
and personality.
However, it seems plausi-
ble
that these
findings
point
to meaningful
method-related
variance,
which is
idrerent
to diJferent measurement
approaches.
10.3.3 Recent
MTMM Studies
Most of
the
aforementioned
approaches
did
not clari{y
the role
of
the intended
measurement
constructs
in a putative
higher-order
framework of social
intel-
ligence.
During
the
past
decade,
however,
attempts have been
made to
ap-
ply
multitrait-multimethod
(MTMM) designs for a better
understanding of
the
structure
and construct
validity of social
intelligence
(Jones
&
Day, 1997;Lee,
Day, Meara, &
Maxwell,
2002;
Lee, Wong,
Day, Maxwell, &
Thorpe,
2000;
Wong,
Day,
Maxwell,
&
Meara,7995).
All these investigations
have assessed
verbal
and nonverbal performance
measures,
as
well
as self-
and sometimes
other-report
data,
of
the
respective
trait-facets.
Furthermore, the
use
of
con-
firmatory
factor
analysis in
these studies
allowed the separation of
trait-
and
method-related
variance to
derive an empirically
defensible
structural
model
of
social
intelligence.
THE ASSESSMENT
OF
SOCIAL
INTELLIGENCE
In
the
first
of
these studies,
Wong
et al.
(1995,
Study
1)
set out to
measure
academic
intelligence,
social perception as
a
cognitive facet
of social
intelli-
gence/
and
socially intelligent behavior
(operationalized
as effective
hetero-
sexual
interaction).
The
latter
include.d ratings
of both
verbal
and
nonverbal
behavior
in
a first
encounter between
a male
and a female
(recorded
on
video-
tape).
verbal
social
perception
was operationalized
by
a
subtest of the
George
Washington
Social
Intelligence Test
(i.e.,
recognition
of the mental
state
behind
words,
see above).
The
Expression
Grouping
subtest of the
Four Factor
Test
of social
Intelligence (o'sullivan
&
Guilford, 1976)
was also
used
as a
mea-
sure
of nonverbal
social
perception.
Results
yielded a
model with
four
uncor-
related
method-factors
(viz.,
verbal,
nonverbal,
self-report,
and
other-report)
and three
correlated
trait-factors
(viz.,
academic
intelligence,
social
perception,
and
effective
heterosexual
interaction).
However,
both
zero-order
correlätions
as well
as trait-factor
intercorrelations
pointed to
substantial
overlap between
social
perception
and academic
intellectual
abilities
(r
:
.67),
a value that
ex-
ceeded
the intercorrelation
between
social
perception and
effective
heterosex-
ual
interaction
(r
:
.54).
I
In
the
second
of these
studies, Wong
et al.
(1995,
Study
2)
postulated three
facets
of social
perception,
social insight,
and
social knowledge.
In the
verbal
measures
of social
knowledge,
participants
had to identify
the
best
solution
for
a
social
problem. The
nonverbal
measure
demanded
the identification
of
etiquette mistakes,
pictured in drawings.
verbal
social
perception was
opera-
tionalized
by the
Social Translation
Test
of the Four
Factor
Test
of Social lntelli-
gence
(o'sullivan
&
Guilford, 7976),
while
the nonverbal
measure
of this facet
was
again
the Expression
Grouping
subtest
of the
same
test
battery.
The verbal
measure
of social insight
was the
Judgment
in
Social
Situations
subtest
of
the
GWSir
(Moss
et al., 1955).
The
nonverbal
measure was
the
Cartoon
prediction
subtest
of o'Sullivan
and
Guilford
(1976).
The
authors
successfully
identified
the
cognitive facets
of social insight
and
social knowledge
as trait-factors
sep-
arable
from,
but
positively
related to,
academic
intelligence.
social
perception
could not
be
separated
from
social insight.
In
yet
another
study,
Jones
and
Day
(7997)
applied
Cattell's distinction
of
fluid versus
crystallized
intelligence
on the
social intelligence
construct ald
thus
operationalized
verbal
and
nonverbal
social
cognitive
flexibility
(fluid
intelLigence)
and verbal
and
nonverbal
social knowledge
(crystallized
intel-
ligence).
The nonverbal
measure
of
social cognitive
flexibility
contained
short
video
clips of ambiguous
social situations.
Participants
had
to list all
possible
interpretations
of
each scene.
The verbal
task
of this facet
included
written
descriptions
of
ambiguous
social situations.
Participants
had again
to list
all
possible interpretations.
The Expression
Grouping
subtest
of o'sullivan and
Guilford
(1976)
represented
the
nonverbal
measure
of social
knowledge.
The
social rranslation
subtest
(o'sullivan
& Guilford,
r976)
was
used as the verbal
measure
of social
knowledge.
|ones
and Day
(7997)
could
show
a
trait-factor
of social cognitive
flexibility
again
separable
from, but
positively related
to,
academic
problem
solving, whereas
social knowledge
could not
be
separated
from
academic
problem
solving.
21,4
Sociallntelligence
Extending
on these findings, Lee
et al.
(2000)
operationalized
both
fluid
and crystallized
social and'academic intelligence.
The authors
specified
fluid
and
crystallized
social intelligence
as
social inJerence
and
social knowledge,
respectively.
Results
showed
that all four
postulated trait-factors
were
dis-
criminable from each
other. Lee
et ai.
(2002)
diverged
from the
just
described
approaches
by using tasks
with
open-ended
questions
to
operationalize
social
knowledge
and the flexible
application
of it. Thus,
they
rather
represented
the
ideas
of Cantor and Kihlstrorn (7987),
who claimed
that
open-ended
questions
would
be
more indicative
of real-life
social
problems
than tasks
with
just
one
correct answer.
The verbal
measure
of social knowledge
was the Role
iategory
Questionnaire
(see
Lee
et
aL.,2002).
Participants
had
to write
detailed
desirip-
tions
of
persons fitting into
a
certain kind
of
social role
(e.g.,
liked
rurn"
,ä"
friend).
ln the nonverbal
measure
of social knowledge,
participants
had to
de-
scribe as fully
as they
could well-known
target
persons
(e.g.,
oprah winfrey),
whose
photos were
presented
on
screens. Answers
were
scored in terms
of
the number
of different
personality
and behavioral
characteristics
identified.
The
verbal
and nonverbal
measure
of social
cognitive
flexibility
representdd
the
same as applied
in the
study
of
Jones
and
Day
(1997).
Results
of this
study
showed
separable
social intelligence
trait-factors
distinct
from,
but
positively
correlated
with,
(general)
creativity.
ln
summary, these
MTMM-studies provide
clear
evidence
for the
murti-
dimensionality
of
social intelligence.
Ho*ever,
although
the
method-related
variance
of
self- and
other-report data
was
controlled
by
the introduction
of
method-factors
or correlations
among
the respective
measures,
trait-factor
Ioadings
vary
strikingly
between
performance
measures
and
self- and
other-
report
data.
Moreover,
the different
measurement
procedures
exhibit no
co-
herent
loading
pattern on one trait-factor.
Consequently,
it remains
uncertain
what
influence
the inclusion
of self- and
other-report
data has
on the
identi-
fied
trait-structure.
Particularly,
no further
(convergent)
validity
evidence
was
available
since self-report
data were
already
included
in
the
social intelligence
models.
1.0.3.4
Summary
In
spite
of the
early,
extensive work
of Guilford
and
his
colleagues
on
so-
cial intelligence
and
their attempts
to
establish
a theoretical
framework,
not
many
comparable
systematic approaches
may be
found
in the
literature.
Most
empirical
studies
focus
on
a
single,
very
specificcognitiae
aspect
of social in-
telligence.
These
operational definitions
seldom
clarify the
role
of measure-
ment
constructs
within the
context
of a higher-order
framework.
hr
addition,
MTMM
approaches
do not
conceptually
integrate
lower-order
facets
of
social
intelligence
(and
their
concomitant
cognitive
determinants)
into
a comprehen-
sive model
of social intelligence.
Since
there is
clear
evidence
for
the
multi-
dimensionality
of social intelligence (Lee
et aL.,2000,2002;
wong
et al. 1995),
it
seems
important
for
future
studies
to locate
constructs
within
a
coherent,
taxonomic
model
of
social intelligence.
The
same kind
of critique
might
be
FACETS
OF SOCIAL INTELLIGENCE 21.5
addressed
to
approaches
focusing on
the
measurement
of social
behaaioral
ef-
fectiueness.
Neither
the role of
effective social
behavior in a framework
of
social
intelligence
nor
an
internal
classification of relevant
social settings appears
to
underlie
extant approaches.
Empirical
evidence for
the construct validity
of social intelligence varies
strikingly across
the measurement procedures that have been adopted.
For
example,
self-report inventories
and behavioral
effectiveness
criteria
suggest
a distinct domain
of social
intelligence.
Approaches relying
on
verbal
(and
sometimes also
nonverbal)
tasks
fail to provide incontrovertible
evidence of a
discriminable
performance construct. However,
the problems associated with
the various
types
of
measurement
procedures remajns an
empirical
issue.
It is
relatively
self-evident that self-report data
better serve as measures
of
typical
social
intellectual
performance in comparison to measures
of
a
performance
construct that is
based
on
the idea
of
maximal
performance. Approaches re-
lying
on pure
performance
measurement
should carefully
consider
the nature
of the task material,
both with respect to the
selection of convergent validity
criteria
and
to real-life
congruence
(Schneider
et a1.,7996).
10.4 FACETS OF SOCIAL INTELLIGENCE
ln this section, we will
attempt
to
integrate
past
theoretical and
empirical
work
into a
performance model of
social
intelligence.
This model is based
on
the
idea
of
a faceted
intelligence model as a framework
for the description and
classification
of
variables
or
tests
(Süß
& Beauducel, 2005), the main focus
of
which will
be various cognitiae facets. It
does
not
lay claim
on
completeness
or
conclusiveness
and will need to be
supported
by
empirical data. In any
event, considering the diversity
of
past
empirical approaches,
it seems
neces-
sary
to
classify the theoretical and
operational definitions
of social
intelligence
in
a unified framework.
The model is based on five
cognitive facets: social ul-
derstanding,
social memory,
social perception, social creativity
(or
flexibility),
and
social knowledge. After a description
of this model, we will
provide some
preliminary results
of
a
study
based
on
this
performance model.
10.4.1 A Taxonomy
of Cognitive Facets of
Social
Intelligence
The facet
of
social understanding
(or
insight) was included
in a large number of
theoretical
and operational definitions,
given different labels
but comprising
similar requirements.
It can be
perceived
as the
pivotal facet
of social
intelli-
gence
in
past
investigations.
Thus,
several
definitions
of social intelligence re-
ferred to in
Section 70.2, narnely, the
ability to understand
people
(Thorndike,
1920), the ability
to define a
given situation in terms
of
the behavior imputed
to others
present
(Chapin,
1942), and to
judge
correctly the feelings, moods,
and motivations
of individuals
(Wedeck,
7947)
cowld
all
be subsumed
under
the facet
of social understanding
or social
insight.
Additionally, both broad
and
specific
operationalizations
of socially
inteliigent
cognition may be classi-
21.6
Sociailntelligence
fied
under this facet:
the
GWSIT
(Moss et
al., 1955),
the Chapin
Social
lnsight
Test
(Chapin,1967;
Gough,
7968),
the
broad
test batteries
of
O'Sullivan
and
Guilford
itOOO,7976),
no--nverbal
decoding
skills
(Barnes
& Sternberg,
T9B9),
and so
forth. Social
understanding
abilities
thus
require
individuals
to inter-
pret
or
understand
given
social
stimuli,
which
may
vary
according
to their
iomplexity,
in termsäf
the
implications
for
the situation
and
their
underlying
features.
the
point
is
well
ifustrated
by
a
sample
test requirement:
ulder-
ständ
correctty
wtiat
a
person
wants
to express
via verbal
or
nonverbal
means
of
communication.
The
facet
of
social
memory
(i.e., behavioral
memory)
was included
in
the
structural
Model
of
Human
Intellect
(Guitford,
1967).
Kosmitzki
and
]ohn
(1993)
also
discovered
a social
memoly
factor
in laypersons'
implic-it
theories
aboui social
intelligence,
that
is,
memory
for
names
and
!,'9s.
One
docu-
mented.
operational*ization
is
that provided
by
Mosset
al.
(1955) in
the GWSIT
(see
also
Probst,
7982).
They
operationalized
social
memoly
as
memory
for
,ru^"s,
and
faces.
The
facet
of social
memory
requires
the
intentional
storing
and
recall of
both
episodic
and semantic
memory
contents.
Correspondingly,
social
memo.y
p"tiot*ance
is
determined
by
the conscious
recall
of objec-
tively
and. expliiitly
given
social
injormation
that can
vary
along
a continuum
of
complexity.
So
far,
the facet
of
social
perception
has
not
been
reflected
in
theoretical
ac-
counts
of social
intelligence.
Nevertheless,
according
to
our
view,
the
ability
to perceive
socially
relävant
information
should
play
a role
in
a performance
modet
of
social
intäligence.
The
ability
to
(quickly)
perceive
social
information
in
a
given
situation
"o=.tta
deterrnine
further
inJormation
processing
that
isrel-
evan"t
for the exhibition
of socially
intelligent
behavior.
Only
Wong
et
al'
(1995)
attempted
to
operationalize
sociäl
perception.
However,
they
did
not succeed
in sepärating
social
perception
from social
understanding
abilities.
These
re-
sults
could
üe attriblted
io
the requirements
of
the selected
tasks.
The
tasks
also included
interpretational
demänds
that,
in our
view,
cannot
be
subsumed
under
the
facet of
pure
perceptual
abilities.
To meet
these conceptual
require-
ments,
social
perception
can6e
specified
as social
perceptual
speed,
analogous
to the
idea oi
peräptual
speed
in models
of
academic
intelligence
(Carroll,
1993;
Thurstone,
1938).
Social
creatiaity
(or flexibil-ity)
was
conceptualized
in
Guilford's
Structural
Model of
Human
Intellect
(Guilford,
7967)
as
divergent production
of
behav-
ioral
contents.
Recent
empirical
work
(Jones
& Day,
7997;
Lee
et
al.,
2002)
operationalizes
social
cognitive
flexibility
as the
flu9nt production
of possi-
bie
interpretations
of,
orlolutions
for,
a given
social
situation.
Importantly,
participäts'performalce
is
not
based
on one
corlect
answer
but on
the num-
L"t u.,ä
diveisity
of
ideas.
The
measures
used
by
both
Jones
and
Day
(7997)
and
Lee et
al.
(Zö02) to define
this
construct
were partly
in line
with Guilford's
early propositions.
Note
that
these
authors
were
capable of
successfully
distin-
guishing^the
domain
of
social
cognitive
flexibility
from
academic
intellectual
abilities.
l'
FACETS
OF SOCIAL
INTELLIGENCE
217
Social
knowledgehasbeen
given
credence
in
the
definitions of
Vernon
(1933,
viz.,
knowledge
of social
matters)
and
Cantor and Kihlstrorn
(1987,
viz., in-
dividual's
fund
of knowledge
about the
social
world). The
concept of social
knowledge
also
plays
a
substantial
role in recent
conceptualizations of practi-
cal
intelligence and
the
related concept of wisdom
(Baltes,
Staudinger, Maer-
cker,
& Smith, 1995; Staudinger,
Lopez, & Baltes, 1997;Sternberg,7998;
Stern-
berg et al.,
2000).
So
far,
social
knowledge has mostly been
operationalized by
measures relying on
knowledge of good etiquette
(Lee
et a1., 2000; Wong
et
al.
1995). Contrary
to these
operationalizations,
Kjhlstrom and
Cantor
(2000)
differentiated
between procedural
(or
so-called
tacit knowledge)
and declara-
tive
social
knowledge.
They
postulated
that
procedural
knowledge could
not
be
taught
or
recalled explicitly,
in
contrast
to declarative
knowledge and the
corresponding
memory components
of episodic and
semantic
memory. With
respect
to these
considerations,
social
knowledge
can be specified as contents
stored in the
procedural
memory component that cannot be taught
or
recalled
explicitly.
6
Given
these constraints, social
knowledge becomes
conceptually
distinct
from
social
memory. However, social
knowledge, as
specified in
these
consid-
erations,
is dependent on
the influence of the cultural
environment
in
general
or
the
specialty of
the
situation
(Weber
& Westmeyer,2001). The assessment of
social
knowledge,
thus,
would require a comprehensive
classification of pos-
sible social situations. Still,
any assessment
would be
subject to respective bul-
tural values and
standards.
From ttris description,
it is not
possible
to conclude how these facets inter-
act to enable people
to exhibit socially intelligent behavior, in
general.
These
cognitive
determinants need
not necessarily
stand on one and
the
same level
and, thus, contribute to
higher-order
performance
to the
same extent.
Figure
10.2 portrays
the
proposed
model of social intelligence including the cognitive
facets, their
possible
interactions
with
each other,
and
with social
behavior as
the
outcome of social
cognitive
intelligence.
IrL this illustration, the
just
described facets constitute the
social
(cognitive)
intelligence construct.
F{owever, the facet of social knowledge, as depicted in
Figure 10.2, does not
play
the
same role
as the other
four facets.
It
seems rea-
sonable to assume
that
social
knowledge
(as
a kind
of
meta-concept) might
also irL{luence the
performance,
for example, in social understanding
or social
perception
abilities. Furthermore,
it is
questionable
whether
social knowledge
only
contains cognitive
requirements, following the aforementioned
consider-
ations
of
Weber and Westmeyer
(2001).
Altogether, the
social cognitive facets
surely
determine social
behavior
performance
to an important
degree. How-
ever,
the
extent of this determination
and, hence, the final
exhibition of socially
intelligent behavior is also influenced by some other, at this
point
indetermi-
nate, array
of person
and environmental
variables
(i.e.,
situational demands,
moods, personality, aims, etc.),
as indicated in the figure.
21,8
Sociallntelligence
o
Situationai
demands
o
Values
/
Norms
o
Personality
o
Moods
r
fnterests
o
Aims
o
Experience
*fillill;t.,
A
possible
performance
moder
of
social
intelligence,
incruding
five
cog-
70.4.2
A
Preliminary
Test
of
the
Model
The
focus
of
the
p-reselt.rlvestigation
(weis
&
snß,2004)was
to
assess
three
cognitive
facets
of
social
intellilence:
social
"r,a"r.tu.,äir,g,'socrat
memory,
and
social
knowredge,
based
ori
performance
measures.
contror
possibre
effects
of
task
mater-ial,
we
used.
ir"Jui
iurt
",
pictures,
and
videos.
The
ver_
bal
measures
of
social
äg::rtqdTg
yere.t1"
Chapin
Social
L_rsight
Test (SIT;
chapin,
7967;
Goush,
196g)
and
the"social
tansla^tio*;;t*
(o,Surivan
&
Guilford,
1976).
T]^e
pictoriar
measure
of
sociar
understanding
was
the
Faces
Te-st (Mayer,
salovey,
Caruso,
and
sitarenios,
2002),
while
tire
video-based
measure
was
the
Interpersonal
perception
Task-15
liortu"ro
äAr.h"r,
rg93).
The
tasks
for
the
sociär
memory
facät
were
art
newly
constructed.
The
Tacit
9:1t"0g,"
tnventory
for
Managers
(TKIM;
wug^e,
d
S,;;Jä
,
rsgt)served
as
the
verbal
measure.
of
sociaf
knowredge.
Ä
confirmatory
iactor
analysis
supported
the
post'lated
trait
structure
#itni"
th"
;"1il;ämg"^."
perfor_
mance
measures/
when
variance
due
to
verbal
content
*u,
lo.,irolled.
The
fit
indices
for
moder
with
best
data
fit
were
as
fonows:
cpi
r
.öo+
;
v21267
:
30'277,
p
:
'256;
RMSEA
:
.037
with
a
90%
confidence
interval
of
[.000,.0g5];
SRMR
:
.056.
The
moder
is
depicted
in
Figure
10.3.
It
postulated
three
corre_
lated
trait-factors
corresponding
to
the
a"rYg.,
of
the
strlJy (rir.,
,o.iur
under-
standing,
memory,
and
knowleäge)
and
a värbal
facto,
with
buäirrg,
from
all
measures
based
on
verbal
materiäI.
FACETS
OF SOCIAL
INTELLIGENCE 219
Figure 10.3
Structural model of social
intelligence
(standardized
solution; ML). SK
=
Social Knowledge;
SM
=
Social Memory; SU
=
Social Understanding; TKIM P1-3
=
Tacit
Knowledge Inventory for Managers Parcel 1-3;
SIT
=
Chapin
Social
Insight Test;
ST
=
Social Translation Test; Faces
=
Faces Test; IPT-15
=
Interpersonal Perception
Test-15.
The
social
knowledge
factor correlated significantly
with the
social memory
and
the
social understanding factor
(.42
and
.50,
respectively). The
social
mem-
ory and
social
understanding
factors
also
correlated significantly
(i.e.,
.45). The
factor loadings
of
the manifest variables
on
the respective trait-factors
showed
a
coherent pattern. The loadings on
the verbal
method factor
were
heteroge-
neous, but all verbal indicators loaded
positively on
this factor.
We further investigated whether
social
intelligence was
separable
from aca-
demic
intelligence, as specified by the
Berlin
Intelligence Structure Test
(BIS-
Test;
]äger,
Süß, &
Beauducel,
7997). Corcelational
and
multiple regression
analysis
showed domain-specific overlap of the social intelligence trait-factors
with
specific domains of the BIS
(Weis
& Süß,
2004).
Results from
confirmatory
factor analysis
suggested still separable
trait-factors
of social and academic in-
telligence.
Additionally,
several social
intelligence
self-report inventories and
scales of Extraversion,
Openness,
and Agreeableness were assessed.
However,
just
as in
past studies, results did not show any evidence for the
convergent
construct validity
of performance
based
social
intelligence with
self-reported
social skills. Furthermore, self-report
data
on
social
intelligence
could
be ex-
plained,
to
large measure,
by the
personality
traits
that we assessed.
'ij-
220
Sociallntelligence
1.0.5
SOCIAL
INTELLIGENCE:
CURRENT
AND FUTURE
PERSPECTIVES
Despite arguably
uncritical
acceptance
of social
or emotional
intelligence
as
relevant
individual differences
constructs,
the introduction
of new statistical
methods
(e.g.,
structural equation modeling) provides opportunities for
clari-
fying formerly unresolved
problems.
Furthermore,
recent
advances in technol-
ogy, including digital means
of
stimulus
recording,
preparation, and
presen-
tation, allow the application
of
task material
that
is closer to real-life
scenarios
than
paper-and-pencil
drawings
or
black and white copies
of photographs.
However, future
research on social intelligence
is
still faced with
overcoming
the
failures and difficulties
of past
research and, thus, the
challenge of
prov-
ing the nature,
structu{e, convergent
and
discriminant
validity,
and
predictive
value
of
the
social intelligence construct.
L0.5.1
Importance
of Resolving Conceptual Issues
b
Ford
(1994)
claimed that social
intelligence
could not be
specified
as
a
pure
ability construct. According to Ford, individual differences
in
socially
intelli-
gent performance
should
not be
specified
without considering
situational
de-
mands,
social
values, and
personal
airns. Weinstein
(1969)
also related
socially
intelligent
behavior to its underlying intentions. For Weinstein,
one aspect
of
social intelligence is
the
ability to
manipulate
the
responses
of others. As a
matter of course, the
eventual exhibition of social
behavior
cannot be
specified
without
considerilg, and
perhaps specifying
in advance, the
relevant äelirnit-
ing conditions in
which social intelligence operates. Nevertheless,
it is neces-
sary to differentiate between the fundamental
cognitive ability
structure and
the
conditions that allow
or
influence the final performance
of social behavior.
If not, this criticism
might
justifiably
be
applied
to the construct
of
academic
intelligence.
Of course, intelligent
performance
in real-Iife
situations
certainly
depends on
present
moods
or
motivation, and/or
on peer group values
(Steele,
1997).
So
do
socially relevant
personality
traits
(e.g.,
Agreeableness,
Extraver-
sion), while
social
interests
clearly
influence
socially intelligent behavior
in
everyday-life.
Even
so,
certain cognitive determinants
of socially intelligent
behavior are
necessary requirements for the accomplishment
of social tasks
and
need to be identified by
empirical
research.
Consequently,
concepts like
social
engagement,
social
interests, or Machiavellian world
views
should not
be
con-founded with a
pure social
(cognitive)
intelligence construct.
No
matter whether
future
studies
rely
on
broad
measurement approaches
or rather
focus
investigation
on specific
domains
of social intelligence, the
con-
ceptualization of the
design demands a thorough specification
of the intended
measurement
constructs and
the corresponding
task
requirements.
This ap-
proach has
proven useful in the academic intelligence
domain and
we argue
that
it is equally important
when considering social intelligence.
Thus,
even
when the
focus is
on
a
narrow constructs,
which
claim to measure
social intel-
ligence in terms
of specific
social skills,
there will be a need
to
place
these
con-
SOCIAL
INTELLIGENCE:
CURRENT
AND FUTURE
PERSPECTIVES 221
stlucts
within a higher-order
framework.
Analogous
to Carroll's
idea
to sys-
tematically integrate
the
various
specific and more
general
constructs
of
acad-
emic intelligence
in
his
hierarchical
Three-Stratum-Theory
(1993),
it might be
possible to estäblish
a framework
of social
intelligence with comparable char-
acteristics. In his late
work,
Guilford
(1981,
1985) already
recognized the pos-
sibility of several
higher-order
factors within his
Structure
of Lrtellect Model.
Anyway,
empirical
studies
and
theoretical
accounts
are
far away from solving
these questions
that are
thus still
subject
to
speculations.
10.5.2 Resolving
Measurement
Issues
Design
issues.
Besides these
fundamental conceptual concerns, past
research
has
clearly
demonstrated
that the application of
MTMM designs is inevitable
for avoiding
any effects of
task material on
the research results.
Additionally,
the construction
of
new
tests appears necessary
for all different facets of social
intelligence.
The
latest technical developments,
(i.e.,
DVD,
digital cameras,
web-based
test delivery,
and
so
forth) ailow
the development of task
materlals
that are closer
to real-life scenarios
than only verbal performance
measures.
Relying
on spoken
language
(auditory
stimuli) seems
just
one
way
to realize
the assessment
of socially
relevant attributes.
Furthermore, new
tests
should
take
into account
the topicality of
the
social
milieu
(just
as intelligence tests
need to be
modified
to take into account emerging
historical events,
technblo-
gies,
and the like).
Validqtion.
According
to the
postulated
performance
model of social
intel-
ligence,
social
behavior
appears to be an adequate criterion
to validate social
cognitive
intelligence.
However, considering
the aforementioned criticisms,
the conceptualization
of appropriate
indicators
of social
behavior seems
to be
a
difficult
obstacle
to traverse.
The exhibition of intelligent social
behavior
is
certainly
irLfluenced
by the
social
environment, present
moods, prevailing
so-
cial norms,
values, and so
forth. In this
respect, it appears difficult,
if at all
possible,
to
assess
performance
in real-life contexts
under
the
control of
all rel-
evant boundary
conditions constituting
the social
world. This
point
notwith-
standing,
it appears
important not
to
lose sight of
the need to specify
limiting
conditions
in
advance. Consequently,
future
studies
need to
establish
a more
comprehensive
classification of social
settings
and
both
universally and specif-
ically
valid criteria
for the
judgment
of socialiy
intelligent behavior.
Investigating
the construct
validity
of social
intelligence
also needs to match
the latest state-of-the-art
in
scientific
research
in terms of the selection
and
specification
of
validation criteria. Certainly,
the replication of past
findings
by applying similar or
the same
measures of academic
intellectual
skills
is
valuable.
However, in order
to
gain
further
information about construct
va-
lidity,
validating social
intelligence
performance
with what are now thought
of as
relatively obsolete
indicators of
academic intelligence
(e.g.,
simple grade
point
average) or apparently
deficient operationalizations
of
g
appears inade-
quate.
In any case,
it
should
be stated clearly,
in correspondence
with under-
222 Sociallntelligence
lying theory,
what type
of
academic
intelligence
is purportedly
assessed
(i.e.,
g, ciystalliied
inte[ifence,
reasoning
abilities,
or
some other
constellation
of
measurement
constnicts)
and
the
strata
upon
which
the construct
resides
(see
Carroll,7993).
Building
the bridge
to
the
main
topic
of
this
book,
certainly,
empirical
inves-
tigations
aie
req,rirä
that
allow
conilusions
about
the overlap
of
social intelli-
gänce
to the puiportedly
related
concept
of
emotional
intelligence
(Matthews
ät
ul.,
ZOOZ;.
As
mentioned
at
the
beginning
of
this chapter,
at an
empirical
level,
few
data
is available
that
provides
evidence
for the
relation
of
the
two
constructs
(Davies et al.,
1998).
At
a theoretical
level, some
cofiunentators
see
the
constructs
as positively
interrelated
(Salovey &
Mal'eL
1990;
Sternberg
et
al.,
2000).
More
specifically,
salovey
and
Mayer
(1990) defined
"emotional in-
telligence
as
a sJbset
of social
intelligence"
(p.
189).
At
the same
time,
emo-
tionäl
intelligence
was concePtualized
as
a kind
of
metacognitive
ability
(Gole-
man,
1995)
äith
effects
or,
äU
kinds of
cognitive
tasks,
including
tasks
from
the
domain
of social
intelligence.
Anyway,
the
absence
of a
common
model
of
social
intelligence
and
tie
elusiveness
äf
emotional
inteiligence
(Zeidher,
Matthews,
&
Röberts,
2001)
inhibits
a
more
detailed
theoretical
description
of
construct
overlaP.
Consequently.
any statement
about
the
relation of
these
two
constructs.can,
at present,
only
be derived
from
a comparison
of
operationalizations.
For ex-
ample,
the
Mayer-salovey-Caruso
Emotional
Inteiligence
Test
(MSCEIT;
Mayer
et
il.,Z00Z1is
based
on
a performance
model
of emotional
intelligence
(Mayer
& Salovey,
1997)
containing
four
branches
of
different
classes
of
abilities:
per-
ception
of
emotion
(Branch
1),
emotional
facilitation
of
thought
(Branch
2),
understanding
emotions
(Branch 3),
and
managing
emotions
(Branch
4)'
A
detailed
descrlption
of
the
model
and
the
four
branches
can
be found
in Chap-
ter
2by Neubauer
and
Freudenthaler
in
this
volume.
Table
10.1
contrasts
the
operationalizations
of
the
MSCEIT
with
some
traditional
operationalizations
of
social
intelligence.
tra
a
q)
'l
ö
U
.o
..1.t
o
.F
G9
pi
Yoü
::-X
I
J-d
*pa
6:=
;U^
-.!u
horbO
:l*=
s03
*E*
c..i
*'
SOCIAL
INTELLIGENCE:
CURRENT AND
FUTURE
PERSpscTryEs
va
üd
vv
m
3öF
üR
ü
ä
g..r'n;!DoE:
F"i,X EOg8; ** *
T ä e fr E":XE
*
q-8
e: ärH.ä+i5 tFE
,IüEö€aLs
üs.i
d\v
JnF'H
C)
gd
I Pq€.^
tE
5
-9t' s v
EE -E
AFeFE
E4
ä;
€E€Eb
Rs
*E
HEF"
äöo
osl
;ril?E
;.= Yo
üE'O*o
Es
ör F.EE*fr
Hor .('<
.9F'ü,e
t
.öE ,; EöFEE
€B
äE
gs3$H
aE .q.'i
'-*
(
c *
g;
us
SeqF:
eU H *o ä E i
u
a
üäi
.:qrEch*ic4E
sä$i
B€EE^!8+Q
ä
be!äbt'Fi&
f
se
a
Er ir&= *
?'EE
H
Y€EgS:öFs=."öo
EaEE:.äEUEi"cF
J E
g
Ä
ä;
E
E I8 E ;
!O
64.
oo
=a
(!Oo
.Föo
"-
'-
0J
:HO
CJd>
J4E
o
G.- O
*E ä.
.ci0X
'o
o€
>\(6
d'^
-5.-
!?o
-u<
Ht
-Vd
o
c,)
(Eul
tU
qD
a'i' d
kx,^
*cB
r9g
;i0)x
F
o.qr
I
I
I
I
I
I
I
I
l*
8l&
cil
bol
=l
EI
qt
'-l
?l
'ul
ol
ol
qtl
trl
ol
ft
GI
NI
qt
öl
FI
b 1.9
l3
l0r
l'o
t>,
lr
lc
IE
Itr
IX
lEl
I
I
I
I
I
Frl
-t
r!l
UI
cnl
>l
c.r
I
.qt
#l
ol
ol
trt
ol
Fl
Slc
.i
t-
dto
klcJ
aJ
t^
o-lH
-rl
I
l.
IV
lo
l.(g
IF
e.
OJ
oo
o
cn
*
o
o
.0,,|
-
öo
d
a
t
k
U
c.l
öl
H
,
U
vl
OJ
d
H
6
e
N
e
(g
H
Fi
Fl
F
q
+
U
-
bo
=
e.
Äa'" ö
0Jq
^.
--P
';
'.i
+
'F1
tr 9.,|
YHV
9
.qtr
'-
i^
.i
':.=
äigFs
Y AAU
!!^
P-!:
esI9g
.:':
'H
st
ü
-JVyr
P5
T
F€
ö
6HIi
>-
(üotrtr9
cJ+j(d!?o
lUdv0i
aü9w\)
eg"csF
ht I
o
ä
^.-:#!
Y'-H
tr I
'-
>.,-
rii X
C.o
'c
il !.1'! x
o'
rx(!ö0r
-äiN@r
rrln- Y
u
ä'ö.x >-k
trr6.;i.i
ä'>-h ö
tr
bo5Xt9
F 3
ö-P
H
A-!
q2
ft
!ruöJd
*
oG
orä
x; ä >-s,
v+x!H
EJ ä=J
u
o
E'A
o
H+.:(u+
F.l
c,, i
a-, o-,
,,-9la
il----
-L--
ä
"
*",
u
o;
,;
F-t
ü^
E
9.^-'6.
H
S
3&:eE s
.E :,d-i3 R
+to +
go>
E
s9EoVQ
UJTUUHTID
OG
v!
Hq
qJ
d -- 9**4
-q Ä ,'; h
o
x-
+-=
ö.9oJF
+..i';ictotrtr
g*
4öHfrp
3+gHä_P
EüEäES
H
L
|
9.9,q I )
.99
öä.EtEE
+E
Ehe;H:
-rn+ö:l
Il: orhi9?EE9J
: e i: E
I
o4
.
!J@j!iq7f(u(6
5rtr-O^Y^^n'!
:cö.rEr:1 äüo!
# SE
Ef
g#E€
F
A=
tr.g.:i
"r
s
öE
ui
-
gr
o.! e&
o
ü I
s
ü-e*
QE
oq
q
9
or
=.:.F.EsEE<lf
'jä8t;;Fn;E?
Ei El P:ilHg frä
d.: orü.I.I-CP äi
o
fiP:VHITg#H
\-./:
boF
öFr
ö
o.ö ö
RA
-
ti
oE
9 > :
_.ä5!Orrq
P ü.!
p
.q.E
;
-rrul:nl.n
I
ö
c{ * E t
.if
a,.i
)
: ä ö
F;*-90
l;;t
H.!-Vd-:/
^#iLi-LH
Y
F fi
o
6.9'J
Lly*^fe
n,-HrUUr
YV-9pd=
Eää.E
frEE
.!'jrE
c V I
P
HIYF.IQ
Gö9
-6aL.rqil
-c
c
boo-d
ö
e
öo
s;;
ö
ä
lg
9u.yfi;
qo.9a.
r#tF
F
r'94
8
ö
3
ö
t
ö
<r+
d
qu
g> gH
m rrl m rrr
Social
Intelligence
L a^^ n
I c',:
:.?
(d
16;
* CN.A
E-c'; Y,
a
H
-.i\Oco
c
u
F t trH d
r_rro\\O
U.X
öF-
.9 I S F*o'
:^trr.L - /'E !
>i-1
\H
j.orf-
x x tr.9=\J.E --
!
i- Y Y
u'E'=..
^*rö
frE:::äEöOf
Pi
=
ü s ü
I
!
ö05
E
-
ai EY
rs
.E
E&
ÄX Gc
or-
I8 Fy
ir.g
^LV.i
;4.
=bo
liF
e"q
5
P
o
o-
-3
3
r€
j
I
ö
"
A ä'!. Y a
I
u
gö.E:6
ög
o
BHESTN+E
ü€
E#
ä fE;
PEPqilog6;
ä€rx-i?ü:H
ElEiäs#[ü
iE;!!€..'ä8"
d;i^F!-td4
tr.Y-:.a+?t-.5(g
:.gEsPF'tnx'Q
ö:; ü
.9
{o.E ;
;
Er'igEEgilF
u;
öor6-EU'9'do
q3
-.-L
wH-
x "x
n -i c
E
pH
H !
j-,
a
E!:EE
gF
+
u:.! CJ
H.=
o.iX"tr
h:
\
! Y- ^ V
v
=H!19v!
90(strtraH
i .r i A
H
Y
v
: äü
E
E
*.F
c!# c F
d
*j1
g
:r'H
ö
.9 :"^ ö
c.,ötr.! äu
.q
o.ö
bog
.i-c
-
rnu E ! Y.:
@
=/= !u
g.H-!Pt
ei
dFä.9ä
pE
3 ä I fr
e-
J';
doäJ;i ar(ü
H
O.:9
F
L
5
:ö858
UI
co
,.Q a
u
.r{
gq
caj
o
-o
'.)-1
öO
h*
cou
bo
fiH
OJJ
^L
!o
cJ*
JYa'n
ouX
9oY
J
X a.
oeY
Y -:Z
F>;
'^'n
ö'i
a
(g
9--
I
I
I
I
I
I
I
I
rl-
ut@
c
t,q
bol
-=l
aJl
tt
trl
:l
.gl
:l
at
sl
ol
ol
EI
GI
NI
=l
cl
ol
!l!
!ulo
öl
o.
-li
tq
l:
Its
l6
lE
t-
laJ
IX
lF.l
I
I
I
I
I
.l
r!l
UI
(nl
>l
qil
rrl
^l
;l
trl
ol
5l
6l
Nl-
t.;i
Elh
o.lt-..1
ol
I
I
lo
t-v
lo
t.q
t-
qJ
qJ
CJ
-
U
r/)
q
q
-
bo
e
a
0i
k
k
U
öl
öt
:.
o)
k
0)
L
U
v)
e
q
e
ac
N
e.
d
H
11
F
ri
0)
r':
-
SOCIAL
INTELLIGENCE: CURRENT
AND
FUTURE PERSPECTIVES
Without
going
too
much into detail
of the
single
operationalizations,
most
of
those
belonging to
the
social
intelligence
domain
were
included
in
the
test
batteries
of
O'Sullivan
and
Guilford
(7966,1976)
and have already
been
de-
scribed
in the
first
part
of this chapter.
Some aspects
of
the
overview
shown
in Table
10.1
need to be comrnented
on. Two tests
of the
MSCEIT
(Pictures
and
Sensations)
do not find any equivalent
operationalizations
in
the domain
of social intelligence.
From
our
viewpoint,
it is
not conceivable
to construct
equivalent measuräs
for the assessment
of social intelligence,
as can
be done
for
other tests
of
the
MSCEIT.
For
two
tests
(Emotion
Management
[EM]
and
Emotions
in Relationships
[EiR]),
the TKIM
(Wagner
& Sternberg,7997)
repre-
sents a test
with rather
equivalent
cognitive requirements,
orLly differing
with
respect
to
the range
of contents
of
the
predeterrnined aim
of actions
(EM:
reg-
ulate
one's own
emotionsi EiR: achieving
an
outcome involving
other
people;
TKIM:
the combination
of both for the
solution
of
a
given
problem). Compara-
bly,
the
Chapin
SIT
(Chapin,7967;
Gough, 1968)
asks the test-taker
to
identify
the
most logical
or
intelligent
solution
or explanation
for a
given social
prob-
lem
and
just
omits the
effectiveness
ratings for the
alternatives.
Furthermor"e,
for most
of
the
MSCEIT tasks
the
purported task requirements
contribute
to
the
accomplishment
of
several social intelligence
tests. For
example, the
abil-
ity to
identify
how a
person feels based
upon their
facial
expression
(Faces,
MSCEIT)
contributes
to the
performance
in the
Faces Test
of
O'sullivan et;a1.
(1965)
(choose
one
of
four
photographs
of men's faces
that
expresses the
sarne
feeling
as
that
of
a
woman's
face). Moreover,
the ability
to
perceive emotions
as
specified
in the Faces
Test
of
the
MSCEIT
surely contributes
to the
accom-
plishment
of
the
Couples
Test
(Barnes
& Sternberg,
7989) where
test-takers
have to
decide whether
a
pictured couple
represents a
real
or
a faked
couple.
Furthermore,
the knowiedge
of
how
moods interact
(Facilitation
Task)
surely
contributes
to the ability
to choose the
one of four
verbal
statements that
de-
scribes
what
precedes,
or
will follow
a cartoon
situation
(Cartoon
Implications;
O'Sullivan
et al., 1965).
The knowledge
of experiencilg
possibly
con{licting
emotions in
certain
situations and
understanding
emotional
chains
(Changes
Task)
might also
help test-takers
to accomplish
the
Chapin SIT where
solu-
tions
or explanations
for
given problems have
to be identified.
At a
scale level,
the
SIT intends
to measure
the ability to
evaluate
others, to
foretell what
may
occur
in interpersonal
and
social
situations, and the
ability to
rectify disturb-
ing
tensions
or conflicts.
Conceptually, this
definition
certairLly
contains the
requirements
of the
scale definiiion
oi Branch
3
(Understanding
Emotions),
that is, the
ability
to understand
emotional
information,
how
emotions
com-
bine
and
progress through
relationship
transitions,
and to
reason
about
such
emotional
meanings.
Obviously,
several abilities
belonging
to the emotional
intelligence
construct
form
a
subset
of those abilities
belonging to the
domain
of social intelligence.
This
supports
the
early conceptualization
of Salovey and
Mayer
(1990).
As a
matter
of course,
the last
considerations
only
represent
statements
about the
face
validity
of
the
compared tests
and
only with respect
to the
constructs
as
assessed
by the
given operationalizations.
Any further
going
conclusions must
ry
l
226 Sociallntelligence
be
subject
to
empirical investigations.
Flence, at present,
many questions con-
cerning the
overlap of social
and emotional
intelligence
remain unanswered.
For
example,
do the tasks
with no corresponding social intelligence
test
assess
a facet
of
emotional
intelligence
independent
from
social intelligence? That is,
is
emotional
intelligence
not only
a
subset
of social
intelligence,
but contains
distinct abilities? Or is
it possible
to regard social
and
emotional intelligence as
constructs
comprising
the same cognitive
requirements based
on
two dif{erent
kinds of contents
(social
vs. emotional contents)?
10.6 CONCLUDINGREMARKS
To
conclude
the chapter,
we would like to provide some surunarizing
state-
ments on
the assessment of social
intelligence, the associated
problems and
expected
future challenges.
Without repeating
in
detail future requirements
as already outlined
in this chapter, elements
that appear most
important for
successful
studies are
(a)
a theory-guided
approach to
the conceptualizaticin
of
the
construct
with respect
to higher- and lower-order facets and necessary
task requirements,
@)
the control
of method-related
variance
(".9.,
by MTMM-
designs), and
(c)
the application of
nonverbal and auditory
task material to
enhance
real{ife equivalence.
\Arhen the construct can be
conceptually
,!le-
lineated
and adequately operationalized,
examining the
construct and
finally
predictive
validity must be
the focus of research. It appears inevitable for the
conduct of
useful studies
to
provide
evidence for
the
convergent validity as
an essential step
for the establishment of
a new ability
construct
(Süß,
2001).
Moreover, it should be noted
that the
subfacets
of a new hierarchical
concept
are in need of support
by evidence of convergent
validity.
Last but not least,
a discussion
about
the position
of social
and
emotional intelligence within the
field
of
individual differences
research appears indispensable.
Lr our view, it
does not appear convincing
to generate practical relevance
only
by the lexi-
cal introduction of new ability
constructs. The apparent
gold
rush associated
with the
introduction and exploration of emotional
intelligence
might
easily
seduce
researchers to adopt
its irnportance from lalpersons'
theories without
supporting
the
relevance
by meaningful empirical evidence,
especially
for the
convergent
and incremental validity
(Sriß,2001).
In this
respect,
Schaie
(2001)
elaborates
necessary steps towards
the
establishment of the emotional intel-
ligence construct like
a
comprehensive
convergent and discriminant valida-
tion, a well-founded
selection
of
the validation
sample, and the application
of
multivariate statistics for data
analysis. Besides the already
mentioned further
duties
as elaborated
in this
chapter,
these methodological
challenges might in-
spire
researchers to
come
up
with-as a
seemingly overdue
step-a book on
social intelligence in the near
future.
REFERENCES 227
REFERENCES
Baltes, P. 8.,
Staudinger,
U.
M., Maercket
A., &
Smith,
j.
(1995).
People
nominated
as wise:
A
comparative study
of wisdom-related knowledge. Psychology
and
Aging,
10,155-166.
Barnes, M. L., &
Sternberg,
R.
I.
(1989).
Social intelLigence and decoding of nonverbal
n,,oc I u t ot I io"^""
13, 263_287.
Biown,
L. T., & Anthony, R.
G.
(1990).
Continuing
the
search
for
social
intelligence.
P er s on
ality
and India idual D if er enc es, L L, 463 47 0.
Cantor, N., & Kihlstrom,
|.
F.
(1987).
Personality
and
social intelligence. Englewood
Clif{s, NJ: Prentice Hall.
Carroll,
I.
B.
(1993).
Human cognitiae abilities: A suraey of
factor-analytic
studies. New
York:
Cambridge
University Press.
Chapin, F. S.
(1942).
Preliminary standardization of a social insight scale. American
Sociolo gical Reaieu;,
7, 21,4--225.
Chapin,
F.
S.
(1967).
The Social
Insight
Test. Palo Alio, CA: Consulting Psychologists
Press.
Costanzo,
M., & Archer, D.
(1993).
The Interpersonal PerceptionTask-l1
(IPT-1,5)
[Video-
tape]. Berkeley, CA: University of California Extension Media Center.
Davies, M.,
Stankov,
L., & Roberts, R. D.
(1998).
Emotional intelligence: Lr search of
an elusive construct.
lournal
of Personality and Soci,alPsychology,75,989-101,5.
'
Feldman,
R.
S.,
Tomasian,J.C., &
Coats,
E.
I.
(1999).
Nonverbal deception abilities and
adolescents
social competence:
Adolescents with higher
social skills
are better
liar s.
I
ournal of N ona
erb al B ehaoior, 23, 237
-249
.
Ford, M.
E.
(1994).
A new conceptualization of social intelligence.
In R.
].
Sternberg
(Ed.),
Encyclopedia of human
intelligence
(pp.
974-978). New York: Macmillan
Publishing
Company.
Ford, M. E.,
&
Tisak, M.
S.
(1983).
A further search for
social
intelligence.
lournal
of
Educational P
sychology,
7 5, 1.96-206.
Frederiksen, N., Carlson, S., & Ward, W. C.
(1984).
The
place
of
social
intelligence in a
taxonomy
of
cognitive abilities. Intelligence
,
I
,
3L5-337 .
Gardner, H.
(1983).
Frames of mind: The theory of multiple intelligences. New York: Basic
Books.
Goleman,
D.
(1995).
Emotional
intelligence:
INhy
it can matter more than
IQ. New York:
Bantam
Books.
Gough, H. G.
(1968).
Manual
for
the
Chapin Social Insight Tesf. Palo Alto, CA: Consult-
ing Psychologists Press.
Guilford.,
I.P.
(1967).
The nature of human intelligence. New York: McGraw-Hill.
Guilford,
J.
P.
(1981).
Higher-order
structure-of-intellect
abilities.
Multioari.ate Behaa-
ioral
Research, 16,
411-435.
Guilford,
I.
P.
(1985).
The structure-of-intellect model. In B. B. Wolmn
(Ed.),
Handbook
of intelligence. Theories, measurements, and applications
(pp.225-266).
New
York:
Tohn
Wilev &
Sons.
i
228
SociallntelLigence
Hendricks,
M., Guilford,
J.P.,
& Hoepfner,
R.
(1969).
Measuring creatiae
social
intel-
ligence
(Psychological
Laboratory
Reports
No. 42). Los Angeles:
University of
Southern
CaLifornia.
Jäger,
A. O.,
Süß,
H.-M.,
& Beauducel
,
A.
(1997).
Berliner Intelligenzstruktur-Tesf
[Berlin
lnteliigence
Structure
Test].
Göttingen,
Germany:
Hogrefe.
Jon-es,
K., &
Day,
I.D.
(1,997).
Discriminations
of two
aspects
of
cognitive-social
irrtelli-
gence
from
academic
htelligence.
lournal
of Educational Psychology
,
89,
486497 .
Keating,
D. P.
(1978).
A search
for social
intelligence.
lournal
of Educational
Psychology,
70,2L8-223.
Ki.hlstrom,
I.
F.,
& Cantor,
N.
(2000).
Social
intelligence.
In R.
j.
Sternberg
(Ed.),
Hand-
book of
intelligence
(pp.359179). New York: Cambridge
University
Press.
Kosmitzki,
C.,
&
John,
o. P.
(1993).
The
impJicit use of
explicit conceptions
of social
intelligence.
P
ersonality
and Indiaidual
Dffir
ences, 1' 5, 11'-23
-
Lee,
J.-E.,
Duy,
J.D.,
Meara,
N.
M., & Maxwell,
S. E.
(2002).
Discrimination
of social
knowledge
and
its flexible
application
from creativity:
A multitrait-multimethod
approach.
Personality
and Indioidual
Dffitences,
32,913-928.
4
Lee,J.-E.,
Wong, C.-M.
T.,
Duy,I. D.,
Maxwell, S.
E., & Thorpe,
P.
(2000).
Social
and
academic
intelligences:
A multitrait-multimethod
study
of
their crystallized
and
fluid characteristics.
Personality
and Indiaidual
Differences,
29, 539-553.
Marlowe,
H. A.
(1936).
Social
intelligence:
Evidence
for multidimensionality
and con-
struct
independence.
I
o ur n
aI
of
E ducatio
nal P sy cholo gy, 7 8, 52-58
Matthews, G.,
Zeidner,M.,
&
Roberts,
R. D.
(2002).
Emotional
intelligence: science
and
myth. Carnbridge,
MA:
MIT Press.
Mayer,
J.
D., DiPaoto,
M., & Salovey,
P.
(1990).
Perceiving
affective content
in ambigu-
ous
visual stimuli:
A component
of
emotional
intelligence
.
lournal
of Personality
Assessment
,
54,
772-781..
Mayer,
J.
D., & Salovey,
P.
(1997).
\Alhat is emotional
intelligence?
Lr P. Salovey
&
D.
J.
Sluyter
(Eds.),
Emotional
deoelopment
and emotional
intelligence: Educational
implications
(pp.
3-31).
New
York: Basic
Books.
Mayer,
J.
D., Salovey,
P., Caruso,
D. R., & Sitarenios,
G.
(2002).
The Mayer, Salouey,
and Caruso
Emotional
Intelligence
Test:
Technical manual
Toronto, Canada:
Multi-
Health Systems.
Moss,
F. A., & Hunt,
T.
(1'927).
Are you
socially
intelligent? Scientific
American,
137,
108-110.
Moss, F.
A., Hunt, T., Omwake,
K. T., &
Woodward,
L.
G.
(1955).
Manualfor
the George
Washington
lJnioersity
Series
Social
Intelligence Tesf
.
Washington,
DC: The Center
for Psychological
Service.
Orlik,
P.
(1978).
Soziale
Intelligenz
[Social
intelligence].
In K.].
Klauer
(Ed'),
Handbuch
d er P
ädago gis chen
D iagno s
t ik
(pp.
347154).
Düsseldorf
,
Germany:
Schwann.
O'Sullivan,M.,&Guilford,
j.P.
(1966).
SixFactorTestof Sociallntelligence:
Manualof
in-
structions
and interpretations.
BeverTy
Hills, CA: Sheridan
Psychological Services.
O'sullivan,
M., & Guilford,
J.
P.
(1976).
Four
Factor Tests of Social
Intelligence: Manual
of
instructions and
interpretations.
Orange, CA:
Sheridan Psychological
Services.
O'Sullivan,
M., Guilford
,1.P.,
&
deMille, R.
(1965).
The measurement
of social
intelligence
(Psychological Laboratory
Reports No. 34).
Los Angeles: University
of Southern
California.
REFERENCES
))q
Probst, P.
(1982).
Empirische
untersuchung zurn
Konstrukt
der "sozialen
Intelligenz"
[Empirical
investigation
on
the
'social
iltelligence'
construct]. In K.
päwhk
(Ed.),
Multioariate
Persönlichkeitsforschung
(pp.
201-226).
Bern,
Switzerland:
Hans Huber.
Riggio, R. E.
(1986).
Assessment
of basic
social skills.
lournal
of Personality
and
Social
P sy chology,
51,
649-660.
Riggio,
R. E.
(i989).
Manualfor
the
Social SkiIIs
Inaentory. PaIo
Alto,
CA: Consulting
.
Psychologists
Press.
'
Riggio,
R. E.,
Messamer,
J.,
& Throckmorton,
B.
(1,991).
Social and
academic
intelli-
gence:
Conceptually
distinct but
overlapping
constructs.
Personality
and Indiaid-
ual
Differences,
12,
695-7 02.
Salovey,
P., &
Mayer,
I.
D.
(1990).
Emotional
intelligence.
Imagination,
Cognition and
P
ersonality,
9, 185-277.
Schaie,
K. W.
(2001).
Emotional
intelligence:
Psychometric
status and
developmental
characteristics-Comment
on Roberts,
Zeidner,
& Maithews (2001).
Emotion,
1,
243-248.
Schneider, R.
j.,
Ackerman,
P. L.,
& KarLfer, R.
(i996).
To
"act
wisely
in human
relj-
tions":
Exploring
the dimensions
of social
competence.
Personality and
Indiaidual
Differences,
4, 46948L
Süß,
H.-M.
(2001).
Prädiktive
Validität
der
Intelligenz im
schulischen und
außer-
schulischen
Bereich
[The
predictive validity
of intelligence
in the
scholastic and
non-scholastic
domain].
In E.
stern
&
|.
Guthke
(Eds.),
Perspektioen
der
InteIIi-
genzforschtmg
(pp.
109-136). Lengerich,
Germany: Pabst
Science
publisher.
Süß, H.-M.,
& Beauducel,
A.
(2005).
Faceted
models
of intelligence.
In
O. Wilhelm
& R.
Engie
(Eds.),
Handbook
of understanding
and measuring
intelligence
(pp.
313-
332). Thousand
Oaks, CA: Sage.
Staudinger,
U.
M., Lopez,
D. F., & Baltes,
P. B.
(1997).
The
psychometric
location
of
wisdom-related
performance:
Intelligence,
personality or more?
Personality
and
Social P sy
chology Bulletin,
23, 1200-121.4.
Steele,
C.M.
(1997).
A threat
in the
air: How
stereotypes
shape the
intellectual
identi-
ties
and
performance
of women and African-Americans.
American
Psychologist,
52,6rH29.
sternberg,
R.
I.
(1998).
A balance theory
of wisdom. Reoiew
of
General
psychology,
2,
347-365.
Sternberg,
R.
J.,
Forsythe,
G.
8.,
Hedlund,
j.,
Horvath,
J.
A., Wagner,
R.
K., Williams,
W. M.,
et al.
(2000).
Practical
intelligence
in eoeryday
lr/e.
New
york:
Cambridge
University
Press.
Sternberg,
R.
I.,
& Smith,
C.
(1985).
Social inteiligence
and decoding
skills in nonverbal
communicatron.
S ocial Co gnition,
3, 168-792.
stricker,
L.
I.,
& Rock, D.
A.
(1990).
lnterpersonai
competence,
social intelligence,
and
general ability.
P
ersonality and
Indiuidual
D
ffirences,
71,
833-839.
Thorndike,
E.
L.
(1920).
Intelligence
and
its use. Harper's
Magazine,740,227-235.
Thurstone,
L. L.
(1938).
Primary
mental
abilities.
Chicago:
University
of
Chicago Press.
vernon, P.
E.
(1933).
some characieristics
of
the
good
judge
of
personality.
lournal
of
Social P sy chology,
4, 42-57 .
':a.E
'I
I
230
Sociallntelligence
Wagner,
R. K.,
& Sternberg, R.
l.
(1991).
Tacit knowledge
inuentory
for
managers.
San
Antonio,
TX:
The Psychological
Corporation Harcourt Brace & Company.
Walker,
R.
E.,
& Foley,
].
M.
(1973).
Social
intelligence:
Its history and measurement.
P sy
cholo
gical Rep or ts,
3 3, 839-864.
Weber,
H.,
& Westmeyer,
H.
(2001).
Die Inflation
der Intelligenzen
[The
inflation
of
intelligencesl.
ln E. Stern &
].
Guthke
(Eds.),
Perspektiaen der
Intelligenzforschung
\pp.
257-266).
Lengerich,
Germany:
Pabst
Science Publisher.
Wechsler, D.
(1981).
Wechsler Adult
Intelligence
Scale-Reaised. New York: Psychologi-
cal Corporation.
Wedeck,
I. $947).
The relationship
between
personality
and
'psychological
ability'.
British
lournal
of Psychology,
37
,
133-151.
Weinstein,
E. A.
(1969).
The development
of interpersonal
competence. tn D. A.
Goslin
(Ed.),
Handbook of socialization theory
and research
(pp.753-775).
Chicago: Rand
McNaliy.
Weis,
S., & Süß, H.-M.
(2004).
Social intelligence-Inoestigating its
structure and construct
oalidity in a
multitrait-mtLltimethod
design. Manuscript
submitted for
publicat{on.
Wong,
C.-M. T., Duy,
1.D.,
Maxwell, S. E.,
& Meara, N. M.
(1995).
A multitrait-
multimethod
study of
academic