SMALL GROUP RESEARCH / October 1999Henry et al. / GROUP IDENTIFICATION
A TRIPARTITE MODEL
OF GROUP IDENTIFICATION
Theory and Measurement
KELLY BOUAS HENRY
University of Oklahoma
HOLLY ARROW
BARBARA CARINI
University of Oregon
Group identification is defined as member identification with an interacting group and is dis-
tinguished conceptually from social identity, cohesion, and common fate. Group identifica-
tion is proposed to have three sources: cognitive (social categorization), affective (interper-
sonal attraction), and behavioral (interdependence). Inconsistent use of the term and
problematic measurement mar existing literature on group identity and group identification.
A new group identification scale, composed of three subscales that match the tripartite
model for the cognitive, affective, and behavioral sources, is presented and its psychometric
properties described.
Group identification (also called group identity, in-group identity,
and intragroup identification) has appeared as a concept in a wide
range of research over the past few decades (Bat-Chava, 1994;
Brewer, 1979; Brewer & Kramer, 1986; Duck, Terry, & Hogg, 1998;
558
AUTHORS’ NOTE: We are grateful to Lewis Goldberg for assistance with factor analyses,
Matt O’Laughlin for writing trial items on the Group Identification Scale, and numerous
members of the University of Oregon Small Groups Laboratory in 1997-1999 for help with
data collection and entry. We also thank two anonymous reviewers for helpful comments on
an earlier draft. An earlier version of this article was presented at the 1998 Meeting of the
Society for Industrial and Organizational Psychology, Dallas, Texas. Correspondence con-
cerning this article should be addressed to Kelly Bouas Henry, now at Department of Psy-
chology, 4525 Downs Drive, Missouri Western State College, St. Joseph, MO 64507, or to
Holly Arrow, Department of Psychology, 1227 University of Oregon, Eugene, OR
97403-1227. Questions about the scale should be addressed to Holly Arrow. Electronic mail
may be sent via the internet to henry@griffon.mwsc.edu or harrow@darkwing.uoregon.edu.
SMALL GROUP RESEARCH, Vol. 30 No. 5, October 1999 558-581
© 1999 Sage Publications, Inc.
Ellemers, Spears, & Doosje, 1997; Ellemers, Van Rijswijk, Roefs,
& Simons, 1997; Terry & Hogg, 1996). A close reading of that lit-
erature, however, reveals general confusion about what, exactly,
group identification is. In some contexts, it means social identity
(Conover & Feldman, 1984; Miller, Gurin, Gurin, & Malanchuk,
1981); in others, it means cohesion (Duckitt & Mphuthing, 1998).
It has been measured with many scales (e.g., Bouas & Arrow,
1996; Brown, Condor, Mathews, Wade, & Williams, 1986; Duck,
Hogg, & Terry, 1995; Ethier & Deaux, 1994; Hinkle, Taylor, Fox-
Cardamone, & Crook, 1989; Riordan & Weatherly, in press; Spears,
Doosje, & Ellemers, 1997) and has also been invoked as an expla-
nation with no attempt at measurement (e.g., Brewer, Weber, &
Carini, 1995; Dawes, McTavish, & Shaklee, 1977).
Such differences in usage and measurement make it difficult to
know what group identification means. Is it unitary or multidimen-
sional? Is it a group-level construct, like cohesion? Is it an
individual-level construct, like social identity? This article
addresses such questions by integrating several bodies of literature
and identifying members of interacting small groups as the locus of
group identification. We note both the similarities and differences
between group identification and related concepts and propose that
group identification has three sources: cognitive, affective, and
behavioral. We close by introducing a new group identification
scale matching our theoretical view.
TRIPARTITE VIEW OF GROUP IDENTIFICATION
The first explicitly tripartite view of intragroup group identifica-
tion is presented in Hinkle et al. (1989). This view, which equates
group and social identification, was based on Tajfel’s (1978) early
social identity discussion. Hinkle et al. developed a scale to meas-
ure three proposed aspects of group identification: cognitive,
evaluative, and affective. Factor analysis yielded three factors,
which they labeled affective, cognitive, and correspondence
between group and individual interests. Unfortunately, items on
Henry et al. / GROUP IDENTIFICATION 559
these scales do not map well to the labels. The items on the
group/individual correspondence factor—“I feel uneasy with the
members of this group” and “I feel held back by this group”—both
seem affective and are the only reverse-scored items. Thus, we
interpret this as a negative affect factor. One of two items loading on
the cognitive factor, “I feel strong ties to this group,” also empha-
sizes emotional connections over cognition.
Bouas and Arrow (1996) agreed that group identity includes
cognitive and affective components but differed on the third ele-
ment in their model. They defined group identity as “awareness of
and attraction toward an interacting group of interdependent mem-
bers, by self-identified members of that group” (pp. 155-156) and
then decomposed this definition into cognitive, affective, and
behavioral components. Social identity literature emphasizes the
cognitive aspect—awareness of a group and self-categorization of
oneself as a member (Tajfel, Billig, Bundy, & Flament, 1971). The
cohesion literature emphasizes the affective aspect, focusing on
interpersonal attraction (e.g., Festinger, Schachter, & Back, 1950;
Piper, Marrache, Lacroix, Richardsen, & Jones, 1983; Turner,
Hogg, Turner, & Smith, 1984). The common fate literature empha-
sizes the behavioral aspect by pointing to the importance of interde-
pendence (Brewer & Kramer, 1986; Chen, 1996).
Using each proposed component to generate predictions about
the development of group identity, Bouas and Arrow (1996) tested
their tripartite view in a longitudinal study of groups. They also
examined individual differences in group identity conceived as an
individual variable. Some support was found for patterns based on a
changing contribution of common fate and interpersonal attraction
over time; little support was found for predictions based on social
identity. The strongest finding was evidence of individual differ-
ences among members in their propensity for high or low group
identity. An individual’s group identity score in one group corre-
lated significantly (r = .42, p < .0001) with that same individual’s
group identity score in a second group. This suggests that although
it may be useful to aggregate member scores to create an overall
group identity score, identification varies among group members,
or members assess group identity differently, or both.
560 SMALL GROUP RESEARCH / October 1999
Building on Bouas and Arrow’s (1996) findings, we have modi-
fied the tripartite view to construe cognitive, affective, and behav-
ioral aspects not as components of a multidimensional construct
but as multiple sources of group identification. We define group
identification as member identification with an interacting group.
Scholars sometimes conflate two levels of group identity (see
Bouas & Arrow, 1996; Worchel, 1996, for examples), using the
same term to refer to the individual-level process of identification
and to group-level identity that is collectively established and can
be perceived by members and nonmembers. To retain these two
senses but distinguish them linguistically, we propose that the term
group identification be used to refer to the individual-level process,
and the term group identity be used to refer to the distinctive iden-
tity of the group as a collective, analogous to the term corporate
identity. We follow this convention in the rest of this article.
Although we view group identification as occurring at the indi-
vidual level, the sources of group identification vary in level. The
cognitive source taps how social identity and social categorization—
aspects of individual cognition and the self—influence group iden-
tification. The affective source focuses on the contribution of inter-
personal attraction (interpersonal level), and the behavioral source
focuses on the group-level construct of cooperative interdepend-
ence. The sources are discussed below.
SOCIAL IDENTITY: THE COGNITIVE SOURCE
We see group identification as conceptually distinct from social
identity, which refers primarily to people’s identification with
broad social categories such as race or gender but is also applied to
interpersonal roles such as lover or husband (Deaux, Reid, Mizrahi, &
Ethier, 1995). The social identity approach views group member-
ship as a psychological state that leads to a collective representation
of who one is and how one should behave (Hogg & Abrams, 1988).
Although abundant research attests to the importance of this cogni-
tive process of self-categorization, we view it as different from
group identification for three reasons.
Henry et al. / GROUP IDENTIFICATION 561
Distinguishing group identification from social identity. First,
according to our definition, group identification occurs in interact-
ing groups. The social identity approach omits any requirement for
group interaction (Lau, 1989). Some theorists prefer not to call
categories of unconnected individuals groups (e.g., Merton, 1949).
The emphasis of the social identity literature on the “group in the
individual’s head” is quite different from the grounded experience
of an individual interacting in an actual small group.
A second reason we see social identity as distinct from group
identification focuses on the intergroup versus intragroup distinc-
tion. Social identity emphasizes self-categorization: Based on
shared attributes with others, one defines oneself as a member of a
social category (Flippen, Hornstein, Siegal, & Weitzman, 1996;
Tajfel et al., 1971). A need for self-esteem motivates one to attain
positive distinctiveness for that social category, producing conse-
quences such as in-group bias (Brewer, 1979). Categorization cre-
ates an in- versus out-group distinction. Minimal group paradigms
that rely on this effect (Jackson & Sullivan, 1987; Tajfel et al.,
1971) focus on social identity’s impact on intergroup relations.
Group identification, however, relates more to intragroup
processes: how members identify with each other and with their
own group without regard to out-group members. Out-groups may
influence group identification but are not necessary to it. A singer
may identify with a choir she has joined, for example, without clas-
sifying nonmembers (the rest of the congregation or choirs at other
churches) as out-group members.
A third reason to view group and social identification as opera-
tively different focuses on the historical construal of the concepts.
In its early conception, social identity was seen as a dichotomy:
One was either a category member or not (Turner, 1982). In recent
years, the term group identification has been used almost inter-
changeably with social identity (e.g., Ellemers, Spears, et al., 1997;
Spears et al., 1997). However, studies that consistently use the term
social identity tend to refer to category membership (e.g., Jackson,
Sullivan, Harnish, & Hodge, 1996), whereas group identification is
most frequently conceptualized and measured as a continuous vari-
able (e.g., Brown et al., 1986; Hinkle et al., 1989; Riordan &
562 SMALL GROUP RESEARCH / October 1999
Weatherly, in press; Spears et al., 1997). Historically, the literature
has treated social identity as a matter of membership and its conse-
quences and group identification as a matter of attachment to a
group, which may vary in strength. In this matter, we again find
group and social identification to be conceptually distinct.
Cognitive sources of group identification. By distinguishing
social and group identification, we do not intend to imply the con-
structs are unrelated. Rather, we view the cognitive process of self-
categorization that is central to social identity theory as a source of
group identification. Categorizing oneself as a member of a group
is an important cognitive source of group identification.
According to the social identity literature, one’s social identity
depends on membership in social categories and evaluations of
these memberships (Deaux et al., 1995). Deaux and colleagues
identified five identity categories, which have different evaluative
parameters. Group members bring social identity categories and
their evaluative parameters to the group context. One’s social iden-
tity can either cross-cut (differ from) or overlap with the social
identities of other members (Brewer, 1995). To the extent that the
social categories being cross-cut are salient to members, cross-
cutting could create cognitive conflict, making identification
among a group of dissimilar members difficult.
Although Bouas and Arrow (1996) found little support for the
effect of cross-cutting versus overlapping social identities on group
identification, other studies suggest that group identification is
weaker if members’ social identities differ. Studies using arbitrary
(e.g., overestimators versus underestimators) rather than demo-
graphic categories found that cross-cutting social identity within a
group inhibits group identification (Gaertner, Mann, Murrell, &
Dovidio, 1989; Marcus-Newhall, Miller, Holtz, & Brewer, 1993).
Ethier and Deaux (1994) studied Hispanic students moving from
home to university. They found ethnic involvement at home tended
to lead to students “re-mooring” their social identity at university
by getting involved in campus friendships consistent with a His-
panic identity. This shows that previous social identities can influ-
ence group identification in a new context. Taken together, these
Henry et al. / GROUP IDENTIFICATION 563
findings suggest that the cognitive process of self-categorization
can either facilitate or hinder the emergence of group identification.
Hogg and Hains (1996) have shown that social identity affects
the type of attraction toward a group and argue that social identifi-
cation yields depersonalized social attraction based on prototypi-
cality. Depersonalized social attraction is similar to Brewer and
Gardner’s (1996) group level of self-representation, which leads to
an intergroup frame of reference and is likely in common identity
groups, in which attraction is to the group rather than its members
(Prentice, Miller, & Lightdale, 1994). Hogg and Hains argue that
interindividual similarity yields interpersonal social attraction.
Interpersonal attraction is similar to Brewer and Gardner’s (1996)
interpersonal self-representation level, which leads to appropriate
role behavior and occurs more in common bond groups, in which
attraction is to members rather than the group itself (Prentice et al.,
1994).
Depersonalized attraction is an affective consequence of cogni-
tive judgments about joint membership and evokes an intergroup
reference frame. Because group identification is an intragroup
process, interpersonal attraction is just as important in promoting
it. We turn next to this affective source.
COHESION AND INTERPERSONAL
ATTRACTION: THE AFFECTIVE SOURCE
Cohesion has been conceptualized in many ways (Drescher,
Burlingame, & Fuhriman, 1985; Evans & Jarvis, 1980; Hogg,
1987; Mudrack, 1989; Stokes, 1983), but nearly all definitions
include interpersonal attraction (Cartwright, 1968). Some have
used cohesion interchangeably with group identification (Duckitt &
Mphuthing, 1998). We agree that cohesion is a source of group
identification but argue the two constructs are not interchangeable.
Defining cohesion. Perhaps the most well-known definition of
cohesion is Festinger et al.’s (1950): “the total field of forces that act
on members to remain in the group” (p. 164). This force field has
two sources: (a) the group’s attractiveness and (b) the group’s
564 SMALL GROUP RESEARCH / October 1999
ability to help members achieve their goals. Scholars who take a
therapeutic perspective have defined cohesion as connectedness
evidenced by working together toward a common therapeutic goal
(Budman, Soldz, Demby, Davis, & Merry, 1993). The industrial-
organizational literature includes definitions such as commitment
to a group task (Goodman, Ravlin, & Schminke, 1987), whereas
sports psychologists favor definitions that include quality of team-
work, attraction to group, unity of purpose, and valued roles
(Yukelson, Weinberg, & Jackson, 1984).
In reviewing the cohesion literature, Cota, Evans, Dion, Kilik,
and Longman (1995) classify conceptions of cohesion as either
unidimensional or multidimensional. Unidimensional approaches
delineate cohesion’s parameters but may underrepresent it. Schol-
ars such as Hogg (1992) and Zaccaro and Lowe (1988) criticize
studies that equate cohesion with interpersonal attraction, for
example, for their inappropriately narrow conception of cohesion.
Multidimensional approaches provide a richer representation but
may contain extraneous aspects. Cota et al. (1995) offer a two-
dimensional view, distinguishing among cohesion’s (a) primary
aspects that apply to all groups and (b) secondary aspects that apply
to specific types of groups. Primary aspects of cohesion are inter-
personal attraction that yields resistance to disruption and consen-
sus among members about behavioral rules and values.
Distinguishing group identification from cohesion. Group iden-
tification differs from cohesion in that it involves individual cogni-
tion and is meaningful at the individual level (Bouas & Arrow,
1996), whereas cohesion exists at the group and interpersonal lev-
els. This is true whether cohesion is defined as a force field or as an
attraction network that resists disruption. Secondary aspects have
even less overlap with group identification. Risk-taking, for exam-
ple, which Cota et al. identify as a secondary aspect in therapy
groups, might affect group identification but is not an aspect of it.
Relating cohesion and interpersonal attraction to group identifi-
cation. We view interpersonal attraction as a source of group identi-
fication. Kerr and Kaufman-Gilliland (1994) reason that group
Henry et al. / GROUP IDENTIFICATION 565
identification develops as a result of affective bonds among group
members. Viewing these bonds as a source of group identification
is consistent with our view that the development of cohesion and
group identification should overlap.
How does the affective component contribute to group identifi-
cation? If members are attracted to one another, they may prefer to
spend more time together as interaction leads to goal attainment
(Sherif, 1967). If they interact cooperatively to attain shared goals,
then this collection of individuals who feel mutual attraction has
become a group of interdependent members. Flippen et al. (1996)
have shown that interdependence has an even greater causal link to
in-group formation than similarity does. Similarity may be respon-
sible for the initial attraction, which leads to repeated positive inter-
action resulting in interdependence. With interdependence estab-
lished, an in-group is formed. Thus, similarity yields interpersonal
attraction, which in turn results in an interacting, interdependent
group. Interaction then strengthens self-identification as a member.
The importance of perceived similarity, a cognitive judgment, in
promoting interpersonal attraction shows how closely cognition
and affect are linked. Together, they can promote identification
with social collectives in the absence of interaction. They can also
promote identification with collectives to which one does not
belong—fan identification with sports teams is one example. How-
ever, we view group identification as an intragroup process influ-
enced by a third group-level source, interdependence.
INTERDEPENDENCE: THE BEHAVIORAL SOURCE
Interest in group identification as an antecedent of cooperation
has focused on common fate (Brewer & Kramer, 1986; Kramer &
Brewer, 1984). Researchers propose that interdependent out-
comes will evoke group identification, an idea supported by some
research on social dilemmas (Brewer & Kramer, 1986; Chen, 1996;
Kramer & Brewer, 1984). Shared outcomes are not the only form of
interdependence, however, and may in fact be the weakest form
566 SMALL GROUP RESEARCH / October 1999
(Brown, 1988). Although Bouas and Arrow (1996) found some
support for the impact of common fate on group identification,
group members in their study shared many kinds of interdepend-
ence. Thus, we broaden the behavioral source of group identifi-
cation to include outcome and behavioral interdependence—the
need to coordinate actions among members in pursuit of group
objectives.
Distinguishing group identification from interdependence.
Behavioral interdependence (we use the term to describe the process
as well as any shared outcome dependent on the process) is a source
of group identification, not group identification itself. Much
research supports this assertion.
Flippen et al. (1996) found interdependence to be a stronger
basis for in-group formation than similarity. Because group forma-
tion is a necessary prerequisite for group identification, this result
implies that interdependence is an antecedent of group identifica-
tion. Pettigrew (1997) looked at interdependence as part of the
contact hypothesis (Allport, 1954). He found contact involving
intergroup friendships—not coworkers or neighbors—reduced
prejudice toward the groups involved in the friendships and other
out-groups. Friendship produces interdependent contact without
competition, a feature that may be absent in neighborhood or organi-
zation contacts.
Sherif (1967) studied how interdependence and shared goals
affect group process and found that cooperative interdependence in
pursuit of shared goals yields a well-defined group structure. Sherif
and Sherif (1969) also argued that interdependent interaction is
crucial to intragroup attraction. This work emphasizes the interde-
pendent process of achieving shared goals (Hogg & Abrams,
1988). Duck (1977) found interdependence and shared experiences
lead to long-term liking. The experience of interdependence in
attaining outcomes leads to these consequences. These studies
together indicate that interdependence causes in-group formation,
a necessary backdrop for attraction to a group and categorization of
oneself as a member. In this way, the three sources work together to
promote member identification with an interacting group.
Henry et al. / GROUP IDENTIFICATION 567
Relating group identification and behavioral interdependence.
Because past studies lack operational agreement on group identifi-
cation, assessing the effects of behavioral interdependence is diffi-
cult. Below, we discuss relevant literature, noting how group identi-
fication was measured.
Cooperation may result from group identification (Dawes et al.,
1977), and common fate manipulations have increased cooperation
in social dilemmas (Brewer & Kramer, 1986; Kramer & Brewer,
1984). The effects of interdependence on cooperation are also
shown by evidence that pledging increases contributions to public
goods (Chen & Komorita, 1994), but only if the pledges promote
interdependence by having participants contribute the mean or
minimum amount pledged (Chen, 1996). Chen suggests that pledg-
ing elicits cooperation because it promotes group identification, as
measured by the Hinkle et al. (1989) scale.
Other research also supports behavioral interdependence as a
source of group identification. Lott and Lott (1965) found interac-
tion that mediates goal achievement produces interpersonal attrac-
tion, a common construal of cohesion (Hogg & Abrams, 1988).
Berkowitz and Walster (1976) argue a group is defined by equitable
interdependence among individuals. The link between interde-
pendence and cohesion, group formation, and group identification
is found in many theories, ranging from learning theory (Lott &
Lott, 1965) to equity theory (Berkowitz & Walster, 1976) to social
identity theory (Flippen et al., 1996). The support found for this
notion in the literature lends credibility to behavioral interdepend-
ence as a source of group identification.
MEASURING GROUP IDENTIFICATION
To resolve the conceptual confusion about group identification,
we have distinguished group identification from related constructs
and proposed an integrative view of group identification as member
identification with small interacting groups, which results from but
is not identical to affective ties, cognitive categorization processes,
and interdependent behavior and outcomes. We have also proposed
568 SMALL GROUP RESEARCH / October 1999
an explicit distinction between the individual-level process of
group identification (the focus of this article) and a group-level
construct of group identity as something that the group as a whole
establishes. Our tripartite conception of group identification, how-
ever, does not match existing scales used to measure the construct.
The items that make up the three factors in the Hinkle et al. (1989)
scale match neither our definition nor Hinkle et al.’s own tripartite
conceptualization.
The unidimensional scales we are aware of have other problems.
A measure of group identification in one study (Spears et al., 1997),
for example, consisted of virtually the same items, with minor adjust-
ments, as those used to measure group commitment in another
study by the same authors (Ellemers, Spears, et al., 1997). The lat-
ter study reported that group identification promotes group com-
mitment. Perhaps, group identification and commitment are
related. Perhaps, they are the same thing. But, our data will be of lit-
tle use in illuminating psychological processes if there is no consis-
tency between the words we use and the instruments that measure
the constructs to which the words refer.
Another unidimensional scale has been developed by Riordan
and Weatherly (in press). They define group identification as a cog-
nitive connection between an individual and a work group, consis-
tent with our view of group identification as an intragroup process.
Their scale successfully discriminates this cognitive conception of
identity from the related constructs of group cohesion and group
communication. Unfortunately, the items in their group identifica-
tion scale do not match their definition very well. Three of the five
items measure the importance of how outsiders view and respond to
the group, and a fourth measures the importance of group success—
which, in work groups, is often defined by out-group members
as well.
We are sympathetic to the measurement problem, having our-
selves used measures of group identification that were a poor fit
with our stated conceptualization (Bouas & Arrow, 1996; Bouas &
Komorita, 1996). To address this problem, Arrow and Carini have
developed a scale that corresponds to our theoretical definition of
Henry et al. / GROUP IDENTIFICATION 569
the construct and that factors into meaningful subscales that reflect
the proposed sources of group identification.
PARTICIPANTS AND GROUPS USED
FOR INITIAL SCALE DEVELOPMENT
Across five rounds of data collection, 965 college students at the
University of Oregon completed five different sets of trial items in
sessions for which they received credit in an introductory psychol-
ogy class. Consistent with our tripartite conception of group identi-
fication, we wrote items to tap affective, behavioral, and cognitive
sources of group identification. After factor-analyzing items from
each round of data collection, we deleted items that did not load
together with other items designated as belonging to the same
subscale, and we added new trial items. Students were directed to
complete the scale for a group they belonged to that had 3 to 25
members; they rated each statement on a 7-point scale, with 1 =
strongly disagree and 7 = strongly agree. Out of 56 items tested, we
selected 13 that had face validity and together factored into a struc-
ture consistent with our tripartite model. This scale included 5
affective items, 5 behavioral items, and 3 cognitive items (Henry,
Arrow, & Carini, 1998).
CONSTRUCT VALIDITY
To check for construct validity, the 290 students in the fifth round
of data collection completed the scale twice, once for a group that
was highly important to them and once for a group that they
belonged to but did not consider very important. Order was coun-
terbalanced, and students filled out several other, unrelated ques-
tionnaires in between the two group identification scales. Respon-
dents were 70% female and 30% male, 76% Caucasian and 14%
Asian, with the remaining 10% split among Hispanic, African
American, and Native American.
The groups that students identified as important groups were
typically athletic teams, religious groups such as Bible study groups,
groups of friends, and academic groups such as study groups.
570 SMALL GROUP RESEARCH / October 1999
Together, these accounted for 70% of the groups. Most common for
the unimportant groups were academic groups, groups at work, and
housemates, which together accounted for 75% of groups.
Comparisons between scores for important and unimportant
groups indicated that the scale discriminated well between those
groups that we would expect people to identify with strongly and
those that they should identify with less strongly. The mean for the
important groups was 5.93 (SD = .93); the mean for the unimpor-
tant groups was 3.89 (SD = 1.07). This difference was reliable at the
p < .0001 level, t(289) = 24.9.
PROBLEMS WITH THE SCALE
In the next round of data collection, in which 179 students com-
pleted the scale, the factor structure of the initial 13-item scale did
not replicate satisfactorily. Although the behavioral items sepa-
rated out as a distinct factor, the affective and cognitive items
loaded together on a single factor. When a three-factor solution was
forced, the third factor was not conceptually meaningful. The scale
included only 3 reverse-scored items, another weakness, and the
cognitive scale had only 3 items. Two of the cognitive items—
“Others think of me as a member of this group” and “Others think
of us as a real group”—also seemed problematic because they
emphasized categorization by nonmembers (a failing we criticized
in the Riordan & Weatherly [in press] scale) rather than the
thoughts of the member completing the scale.
REVISED SCALE
To address these problems, we revised the scale to balance the
reverse- and direct-scored items for each subscale and substitute
some new cognitive items. Students from the University of Okla-
homa completed the original scale plus reverse-worded versions of
the direct items and a few new cognitive items. The Hinkle et al.
(1989) scale was also included, so that we could check for conver-
gent validity. From the items completed, we selected 4 each (2
direct, 2 reverse) for each subscale, for a total of 12 items. Nine of
Henry et al. / GROUP IDENTIFICATION 571
the items from the original scale were included (with 3 changed to
reverse wording), and 3 of the cognitive items were new.
The resulting Group Identification Scale 2.0 is reproduced in
Table 1. To check for stability of factors, we also had University of
Oregon students complete the new scale and the Hinkle et al. (1989)
scale. Factor loadings for the Oregon data collection are shown in
parentheses in Table 1. For ease of inspection, we have grouped
items from the three subscales together, although they are inter-
spersed in the instrument. The proportion of variance explained by
each rotated factor (ignoring the other factors, which all have some
572 SMALL GROUP RESEARCH / October 1999
TABLE 1: Factor Pattern for Arrow-Carini Group Identification Scale 2.0,
Oklahoma and Oregon Data
Factor
Subscale Affective Behavioral Cognitive
Affective
1. I would prefer to be in a different group (R) .73 (.25) (.52)
4. Members of this group like one another .82 (.73)
7. I enjoy interacting with the members of
this group .78 (.63)
10. I don’t like many of the other people in
this group (R) .89 (.89)
Behavioral
2. In this group, members don’t have to rely
on one another (R) .84 (.79)
5. All members need to contribute to achieve
the group’s goals .86 (.83)
8. This group accomplishes things that no
single member could achieve .71 (.80)
11. In this group, members do not need to
cooperate to complete group tasks (R) .73 (.81)
Cognitive
3. I think of this group as part of who I am .74 (.92)
6. I see myself as quite different from other
members of the group (R) .77 (.37)
9. I don’t think of this group as part of who
I am (R) .77 (.77)
12. I see myself as quite similar to other
members of the group .75 (.92)
Variance explained by each factor, ignoring
other factors 3.61 (3.21) 3.06 (3.08) 3.41 (3.53)
NOTE: (R) indicates a reverse-scored item. Numbers in parentheses refer to Oregon data set.
shared variance) ranged between 26% and 30% (see Table 1). The
eigenvalues of the first three orthogonal principal components
(unrotated) explained 64% of the variance in each sample.
PARTICIPANTS
The 420 University of Oklahoma students were 51% female and
49% male, with mean age 19.7 years, ranging from 17 to 49. Eth-
nicity was 77% European American, 7% African American, and
4% Asian American, with the remaining 12% divided among His-
panic, Native American, mixed ethnicity, foreign nationals, and not
specified. The 320 University of Oregon students were 65% female
and 35% male, with mean age 19.3 years. Ethnicity was 81% Euro-
pean American, 11% Asian-Pacific, and 3% African American,
with the remaining 5% divided among Hispanic, Native American,
and multiple or not specified.
DISCRIMINATION AMONG THE SUBSCALES
The three subscales in the group identification scale are not
orthogonal. Correlations between the subscales (see Table 2) indi-
cate that the affective and cognitive sources have high overlap,
whereas the behavioral source is more distinct. Because of the cor-
relation among subscales, we chose oblique rotation, using the SAS
Promax program, to examine the factor structure of the scale. The
rotated factor pattern, with three factors specified, is shown in
Table 1 for the Oklahoma and Oregon groups. Loadings not shown
were below .30, Gorsuch’s (1983) minimum. Correlations between
the factors are shown in Table 2.
With two exceptions, the scale items passed the following three
criteria for both the Oklahoma and Oregon data sets: (a) the item
should load on the same factor as the other items in the same
subscale; (b) loading on this factor should be at least .40 (using the
cutoff Gorsuch suggests as a more conservative minimum); and (c)
if the item also loads on a second factor, there should be at least a
.10 difference in the strength of the loading. The two problematic
items were Item 1, keyed as an affective item, which loaded with
Henry et al. / GROUP IDENTIFICATION 573
TABLE 2: Correlations and Internal Consistency of Subscales and Correlations Among Factors for Oklahoma and Oregon Data Sets
Oklahoma Data Set (N = 420) Oregon Data Set (N = 320)
Affective Behavioral Cognitive Total Affective Behavioral Cognitive Total
Correlations and internal consistency for subscales
Affective (.84) (.79)
Behavioral .31 (.80) .31 (.83)
Cognitive .55 .33 (.78) .66 .33 (.76)
Total .78 .70 .84 (.89) .80 .71 .85 (.85)
Correlations among rotated factors
Behavioral .30 .26
Cognitive .50 .29 .45 .25
NOTE: Figures in parentheses give internal consistency as measured by Cronbach’s alpha for standardized variables.
574
the cognitive items in the Oregon data set, and Cognitive item 6,
which dropped below the .40 minimum in the Oregon data set. Both
items are reverse-scored, and in our experience, it is especially dif-
ficult to write reverse-scored items that load consistently with the
direct-scored items in the same scale. Because the affective and
cognitive scales are highly correlated, it is not surprising that main-
taining differentiation between items in the two scales is difficult.
Although the two sources of group identification are conceptually
distinct, they are closely linked in the pattern of responses when
measuring group identification. Specification of two factors indi-
cated that a combined affective-cognitive factor with a second
behavioral factor was a better solution for the Oregon data than a
three-factor solution.
RELIABILITY
The internal consistency of the subscales was assessed using
Cronbach’s alpha. Results are shown in Table 2 for both Oklahoma
and Oregon groups.
CONVERGENT VALIDITY
The correlation between scores on the group identification scale
and the Hinkle et al. (1989) scale was .83 for the Oklahoma data and
.85 for the Oregon data. We interpreted the Hinkle et al. items as
primarily measuring the affective source of group identification,
and in line with this interpretation, scores on this scale correlated
more highly with our affective subscale (.83/.84 for the Oklahoma
and Oregon data sets) than with our behavioral (.41 for both data
sets) and cognitive (.69/.79 for Oklahoma and Oregon) subscales.
USEFULNESS OF THE SCALE ACROSS POPULATIONS
We have demonstrated reasonable stability of the factor struc-
ture across two college student populations in different regions
of the United States. Respondents used the scale to report on the
sources of group identification for a variety of different small groups.
Henry et al. / GROUP IDENTIFICATION 575
For researchers, one practical limitation of the scale is that it pre-
sumes the existence of a real group that is perceived to exist not
only by a researcher but also by its members. For this reason, it is
not well-suited for assessing group identification in the ad hoc,
minimal groups used in many experiments. When a colleague tried
the scale in this setting, participants were prone to leave questions
blank or protest that the questions did not make sense. For these
minimal groups, constructs such as social identity, perceived simi-
larity, or self-categorization are more appropriate. In our view,
group identification emerges only when people perceive them-
selves to be members of an actual group and should not be applied
to collections of people who lack behavioral interdependence or
any concrete connection to one another.
The usefulness of the scale for group members in broader com-
munity samples and in organizational settings, both within the
United States and in other English-speaking communities, is yet to
be determined. We hope that colleagues who use the scale with dif-
ferent populations will report on its usefulness with respondents
other than college students.
CONCLUSIONS
Although group identification has been a popular construct in
recent years, research on group identification has suffered from
conceptual confusion and poor measurement. Researchers from
different theoretical traditions have used the term to mean different
things, and even researchers within the same tradition do not
always use the term consistently. Group identification has been
evoked as an explanatory factor with no attempt to measure it (e.g.,
Brewer et al., 1995), and when group identification is measured,
researchers frequently use ad hoc scales that do not reflect their
theoretical definitions.
We have attempted to clarify these matters by developing both
an integrative theoretical conception of group identification and a
measurement instrument that matches that conception. Our tripar-
tite model integrates three ostensibly separate bodies of literature
576 SMALL GROUP RESEARCH / October 1999
that have used the term group identification. The social identity lit-
erature has focused on identification as self-categorization. The
cohesion literature has focused on interpersonal attraction. The
common fate literature emphasizes interdependence among group
members. We propose that each perspective describes a conceptu-
ally distinct source of group identification, and our scale measures
the strength of these three sources.
The tripartite view can be used as a guide to generate predictions
for future research. In particular, we are interested in the conse-
quences of the cognitive, affective, and behavioral sources for
group functioning. Both the cognitive and affective sources could
promote adherence to group norms (Hogg & Hains, 1996; Sea-
shore, 1954) and in-group bias (Duck et al., 1998; Tajfel et al.,
1971). The behavioral source may be the most likely to yield bene-
fits in group performance, as outcome interdependence has been
associated with cooperation (Dawes et al., 1977) in previous
research. In groups with high demographic diversity, the cognitive
source of group identification may be weak. The key to promoting
group identification in such groups is, we believe, an emphasis on
the behavioral sources of group identification and a de-emphasis on
demographic characteristics on which members differ.
Worchel’s (1994) model of group development proposes that
groups cycle through distinct phases, including a group identifica-
tion phase. Characterizations of the different phases suggest that
during this phase, affective and cognitive sources should be strong,
but behavioral sources may be weak. The next phase, group pro-
ductivity, should be characterized by increased contribution of the
behavioral source of group identification, while the cognitive
source wanes throughout this phase and the next phase of indi-
viduation. Our group identification scale would be useful in testing
these predictions. We also expect that the development of group
identification will differ in groups that meet primarily face to face
compared to distributed groups whose communication is
computer-mediated, as indicated by prior work (Bouas & Arrow,
1996). Underlying these differences may be differential contribu-
tions of the three sources of group identification for groups using
different communication media.
Henry et al. / GROUP IDENTIFICATION 577
It is our hope that the development of an explicit theory of group
identification will stimulate more systematic thinking and research
on the topic. In addition, more attention to measurement issues
across research studies and programs will help us to build a litera-
ture that provides clearer and more cumulative findings about the
sources, the development, and the consequences of group
identification.
REFERENCES
Allport, G. W. (1954). The nature of prejudice. Reading, MA: Addison-Wesley.
Bat-Chava, Y. (1994). Group identification and self-esteem in deaf adults. Personality and
Social Psychology Bulletin, 20, 494-502.
Berkowitz, L., & Walster, E. (Eds.). (1976). Equity theory: Toward a general theory of social
interaction (Vol. 9 of Advances in Experimental and Social Psychology). New York:
Academic Press.
Bouas, K. S., & Arrow, H. (1996). The development of group identity in computer and face-
to-face groups with membership change. Computer Supported Cooperative Work, 4,
153-178.
Bouas, K. S., & Komorita, S. S. (1996). Group discussion and cooperation in social dilem-
mas. Personality and Social Psychology Bulletin, 22, 1144-1150.
Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitive-
motivational analysis. Psychological Bulletin, 86, 307-324.
Brewer, M. B. (1995). Managing diversity: The role of social identities. In S. E. Jackson &
M. N. Ruderman (Eds.), Diversity in work teams: Research paradigms for a changing
workplace (pp. 47-68). Washington, DC: American Psychological Association.
Brewer, M. B., & Gardner, W. (1996). Who is this “we”? Levels of collective identity and self
representations. Journal of Personality and Social Psychology, 71, 83-93.
Brewer, M. B., & Kramer, R. M. (1986). Choice behavior in social dilemmas: Effects of
social identity, group size, and decision framing. Journal of Personality and Social Psy-
chology, 50, 543-549.
Brewer, M. B., Weber, J. G., & Carini, B. (1995). Person memory in intergroup contexts:
Categorization versus individuation. Journal of Personality and Social Psychology, 69,
29-40.
Brown, R. (1988). Group processes: Dynamics within and between groups. Oxford, UK:
Basil Blackwell.
Brown, R. J., Condor, S., Mathews, A., Wade, G., & Williams, J. A. (1986). Explaining inter-
group differentiation in an industrial organization. Journal of Occupational Psychology,
59, 273-286.
Budman, S. H., Soldz, S., Demby, A., Davis, M., & Merry, J. (1993). What is cohesiveness?
An empirical investigation. Small Group Research, 24, 199-216.
Cartwright, D. (1968). The nature of group cohesiveness. In D. Cartwright & A. Zander
(Eds.), Group dynamics: Research and theory (3rd ed.). New York: Harper & Row.
578 SMALL GROUP RESEARCH / October 1999
Chen, X. P. (1996). The group-based binding pledge as a solution to public goods problems.
Organizational Behavior and Human Decision Processes, 66, 192-202.
Chen, X. P., & Komorita, S. S. (1994). The effects of communication and commitment in a
public goods social dilemma. Organizational Behavior and Human Decision Processes,
60, 367-386.
Conover, P. J., & Feldman, S. (1984). Group identification, values, and the nature of political
beliefs. American Politics Quarterly, 12, 151-175.
Cota, A. A., Evans, C. R., Dion, K. L., Kilik, L., & Longman, R. S. (1995). The structure of
group cohesion. Personality and Social Psychology Bulletin, 21, 572-580.
Dawes, R. M., McTavish, J., & Shaklee, H. (1977). Behavior, communication, and assump-
tions about other people’s behavior in a common dilemma situation. Journal of Personal-
ity and Social Psychology, 35, 1-11.
Deaux, K., Reid, A., Mizrahi, K., & Ethier, K. A. (1995). Parameters of social identity. Jour-
nal of Personality and Social Psychology, 68, 280-291.
Drescher, S., Burlingame, G., & Fuhriman, A. (1985). Cohesion: An odyssey in empirical
understanding. Small Group Behavior, 16, 3-30.
Duck, J. M., Hogg, M. A., & Terry, D. J. (1995). Me, us, and them: Political identification
and the third-person effect in the 1993 Australian federal election. European Journal of
Social Psychology, 25, 195-215.
Duck, J. M., Terry, D. J., & Hogg, M. A. (1998). Perceptions of a media campaign: The role
of social identity and the changing intergroup context. Personality and Social Psychol-
ogy Bulletin, 24, 3-16.
Duck, S. W. (1977). The study of acquaintance. Farnborough, Hants: Saxon House.
Duckitt, J., & Mphuthing, T. (1998). Group identification and intergroup attitudes: A longi-
tudinal analysis in South Africa. Journal of Personality and Social Psychology, 74, 80-85.
Ellemers, N., Spears, R., & Doosje, B. (1997). Sticking together or falling apart: In-group
identification as a psychological determinant of group commitment versus individual
mobility. Journal of Personality and Social Psychology, 72, 617-626.
Ellemers, N., Van Rijswijk, W., Roefs, M., & Simons, C. (1997). Bias in intergroup percep-
tions: Balancing group identity with social reality. Personality and Social Psychology
Bulletin, 23, 186-198.
Ethier, K. A., & Deaux, K. (1994). Negotiating social identity when contexts change: Main-
taining identification and responding to threat. Journal of Personality and Social Psy-
chology, 67, 243-251.
Evans, N. J., & Jarvis, P. A. (1980). Group cohesion: A review and reevaluation. Small Group
Behavior, 11, 359-370.
Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in informal groups. New
York: Harper & Row.
Flippen, A. R., Hornstein, H. A., Siegal, W. E., & Weitzman, E. A. (1996). A comparison of
similarity and interdependence as triggers for in-group formation. Personality and
Social Psychology Bulletin, 22, 882-893.
Gaertner, S. L., Mann, J., Murrell, A., & Dovidio, J. F. (1989). Reducing intergroup bias: The
benefits of recategorization. Journal of Personality and Social Psychology, 57, 239-249.
Goodman, P. S., Ravlin, E., & Schminke, M. (1987). Understanding groups in organizations.
Research in Organizational Behavior, 9, 121-173.
Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Henry, K. B., Arrow, H., & Carini, B. (1998, April). Theoretical extensions and organiza-
tional applications of the tripartite model of group identification. In K. B. Henry (Chair),
Henry et al. / GROUP IDENTIFICATION 579
Group identity, cohesion, and diversity in work teams. Symposium conducted at the
meeting of the Society for Industrial and Organizational Psychology, Dallas, TX.
Hinkle, S., Taylor, L. A., Fox-Cardamone, D. L., & Crook, K. F. (1989). Intragroup identifi-
cation and intergroup differentiation: A multicomponent approach. British Journal of
Social Psychology, 28, 305-317.
Hogg, M. A. (1987). Social identity and group cohesiveness. In J. C. Turner, M. A. Hogg, P. J.
Oakes, S. D. Reicher, & M. Wetherell (Eds.), Rediscovering the social group: A self-
categorization theory. Oxford, UK: Basil Blackwell.
Hogg, M. A. (1992). The social psychology of group cohesiveness: From attraction to social
identity. New York: New York University Press.
Hogg, M. A., & Abrams, D. (1988). Social identifications: A social psychology of intergroup
relations and group processes. New York: Routledge, Chapman, & Hall.
Hogg, M. A., & Hains, S. C. (1996). Intergroup relations and group solidarity: Effects of
group identification and social beliefs on depersonalized social attraction. Journal of
Personality and Social Psychology, 70, 295-309.
Jackson, L. A., & Sullivan, L. (1987). The ingroup favorability bias in the minimal groups
situation. Journal of Social Psychology, 127, 461-472.
Jackson, L. A., Sullivan, L. A., Harnish, R., & Hodge, C. (1996). Achieving positive social
identity: Social mobility, social creativity, and permeability of group boundaries. Jour-
nal of Personality and Social Psychology, 70, 241-254.
Kerr, N. L., & Kaufman-Gilliland, C. M. (1994). Communication, commitment, and coop-
eration in social dilemmas. Journal of Personality and Social Psychology, 66, 513-529.
Kramer, R. M., & Brewer, M. B. (1984). Effects of group identity on resource use in a simu-
lated commons dilemma. Journal of Personality and Social Psychology, 46, 1044-1057.
Lau, R. R. (1989). Individual and contextual influences on group identification. Social Psy-
chology Quarterly, 52, 220-231.
Lott, A. J., & Lott, B. E. (1965). Group cohesiveness as interpersonal attraction. Psychologi-
cal Bulletin, 55, 259-309.
Marcus-Newhall, A., Miller, N., Holtz, R., & Brewer, M. B. (1993). Cross-cutting category
membership with role assignment: A means of reducing intergroup bias. British Journal
of Social Psychology, 32, 125-146.
Merton, R. K. (1949). Social theory and social structure (rev. ed.). New York: Free Press.
Miller, A. H., Gurin, P., Gurin, G., & Malanchuk, O. (1981). Group consciousness and politi-
cal participation. American Journal of Political Science, 25, 494-511.
Mudrack, P. E. (1989). Defining group cohesiveness: A legacy of confusion. Small Group
Behavior, 20, 37-49.
Pettigrew, T. F. (1997). Generalized intergroup contact effects on prejudice. Personality and
Social Psychology Bulletin, 23, 173-185.
Piper, W. E., Marrache, M., Lacroix, R., Richardsen, A. M., & Jones, B. D. (1983). Cohesion
as a basic bond in groups. Human Relations, 36, 93-108.
Prentice, D. A., Miller, D. T., & Lightdale, J. R. (1994). Asymmetries in attachments to
groups and to their members: Distinguishing between common-identity and common-
bond groups. Personality and Social Psychology Bulletin, 20, 484-493.
Riordan, C. M., & Weatherly, E. W. (in press). Defining and measuring employee’s identifi-
cation with their work groups. Educational and Psychological Measurement.
Seashore, S. E. (1954). Group cohesiveness in the industrial work group. Ann Arbor: Uni-
versity of Michigan, Institute for Social Research.
Sherif, M. (1967). Group conflict and cooperation. London: Routledge & Kegan Paul.
580 SMALL GROUP RESEARCH / October 1999
Sherif, M., & Sherif, C. W. (1969). Social psychology. New York: Harper & Row.
Spears, R., Doosje, B., & Ellemers, N. (1997). Self-stereotyping in the face of threats to
group status and distinctiveness: The role of group identification. Personality and Social
Psychology Bulletin, 23, 538-553.
Stokes, J. P. (1983). Components of group cohesion: Intermember attraction, instrumental
value, and risk taking. Small Group Behavior, 14, 163-173.
Tajfel, H. (1978). Social categorization, social identity, and social comparison. In H. Tajfel
(Ed.), Differentiation between social groups: Studies in the social psychology of inter-
group relations. London: Academic Press.
Tajfel, H., Billig, M., Bundy, R., & Flament, C. (1971). Social categorization and intergroup
behavior. European Journal of Social Psychology, 1, 149-177.
Terry, D. J., & Hogg, M. A. (1996). Group norms and the attitude behavior relationship: A
role for group identification. Personality and Social Psychology Bulletin, 22, 776-793.
Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.),
Social identity and intergroup relations. Cambridge, UK: Cambridge University Press.
Turner, J. C., Hogg, M. A., Turner, P. J., & Smith, P. M. (1984). Failure and defeat as determi-
nants of group cohesiveness. British Journal of Social Psychology, 23, 97-111.
Worchel, S. (1994). You can go home again: Returning group research to the group context
with an eye on developmental issues. Small Group Research, 25, 205-223.
Worchel, S. (1996). Emphasizing the social nature of groups in a developmental framework.
In J. L. Nye & A. M. Brower (Eds.), What’s social about social cognition? (pp. 261-282).
Thousand Oaks, CA: Sage.
Yukelson, D., Weinberg, R., & Jackson, A. (1984). A multidimensional group cohesion
instrument for intercollegiate basketball teams. Journal of Sport Psychology, 6, 103-117.
Zaccaro, S. J., & Lowe, C. A. (1988). Cohesiveness and performance on an additive task:
Evidence for multidimensionality. The Journal of Social Psychology, 128, 547-558.
Kelly Bouas Henry joined the psychology faculty of Missouri Western State College
in Fall 1999. She received her Ph.D. in social and organizational psychology from
the University of Illinois, Urbana-Champaign, in 1997 and was a member of the psy-
chology faculty at the University of Oklahoma from 1997 to 1999. Her research inter-
ests include the antecedents and consequences of group identification, various
aspects of group structure, and the development of cooperation in social dilemmas.
Holly Arrow, a member of the Institute for Cognitive and Decision Sciences and the
psychology faculty at the University of Oregon, received her Ph.D. in social and
organizational psychology from the University of Illinois, Urbana-Champaign, in
1996. She studies issues in group formation and development, including the emer-
gence and transformation of group structure, membership dynamics, and the experi-
ence of newcomers.
Barbara Carini is a graduate student in social psychology at the University of Ore-
gon. Her research interests include social categorization, intergroup relations, and
small-group processes.
Henry et al. / GROUP IDENTIFICATION 581