Content uploaded by David A. Harrison
Author content
All content in this area was uploaded by David A. Harrison on Oct 22, 2014
Content may be subject to copyright.
Time, Teams, and Task Performance
In Press - AMJ
1
TIME, TEAMS, AND TASK PERFORMANCE:
CHANGING EFFECTS OF SURFACE- AND DEEP-LEVEL DIVERSITY ON GROUP
FUNCTIONING
DAVID A. HARRISON
Department of Management and Organization
Smeal College of Business Administration, 410 Beam
The Pennsylvania State University
University Park, PA 16802
Tel: (814) 865-1522
Fax: (814) 863-7261
E-mail: dharrison@psu.edu
KENNETH H. PRICE
JOANNE H. GAVIN
ANNA T. FLOREY
Department of Management
College of Business Administration
University of Texas at Arlington
Box 19467
Arlington, TX 76019-0467
Tel: (817) 272-3863
Fax: (817) 272-3122
E-mail: price@uta.edu
Running head: TIME, TEAMS, AND TASK PERFORMANCE
This paper was presented in the Gender and Diversity in Organizations (GDO) Division
program at the 2000 Academy of Management meetings in Toronto, Canada, and was a finalist
for the Dorothy Harlow Best Paper Award. We are grateful to the anonymous reviewers and to
the session participants for their feedback, as well as to Beta Mannix and the AMJ reviewers.
Time, Teams, and Task Performance
In Press - AMJ
2
TIME, TEAMS, AND TASK PERFORMANCE:
CHANGING EFFECTS OF SURFACE- AND DEEP-LEVEL DIVERSITY ON GROUP
FUNCTIONING
ABSTRACT
Time serves as a medium for collaboration in teams, allowing personal and task-related
information to be exchanged between members. We propose that stronger team reward
contingencies will stimulate collaboration; and, as time passes, increasing collaboration will
weaken the effects of surface-level (demographic) diversity, but strengthen the effects of deep-
level (psychological) diversity on team outcomes. We also propose that perceived diversity
transmits the impact of actual demographic and psychological diversity on team social
integration. Social integration affects task performance. These hypotheses were tested by
tracking 144 student project teams in a four-wave design. Results supported the propositions
and suggest a much stronger role for time in research attempting to unpack the "black box" of
diversity effects in work teams.
Time, Teams, and Task Performance
In Press - AMJ
3
Managing diverse work groups is one of the most difficult and pressing challenges in
modern organizations, part of a "grand experiment" that, by many accounts, is "not going
smoothly" (Tsui & Gutek, 1999, p. 1). The conventional focus of diversity research has been
on connecting demographic differences among team members, such as age, sex, or race, to
reactions toward team-level functioning (e.g., team social integration) and performance (Williams
& O'Reilly, 1998). These “surface-level” (Jackson, May, & Whitney, 1995) or "high visibility"
(Pelled, 1996) demographic characteristics are easily observed and measured. They are
presumed to be important because of the underlying differences they are thought to reflect, and
because they can evoke individual prejudices, biases, or stereotypes (Fiske & Neuberg, 1990).
A complementary paradigm has begun to emerge, however, marking the start of a "new time" in
research on work team diversity. It involves the investigation of deep-level (Harrison, Price, &
Bell, 1998) or less readily apparent diversity (Riordan, 2000). This form of diversity is based on
psychological features of work team members and includes individual differences such as
personality traits (Barsade, Ward, Turner, & Sonnenfeld, 2000) and values (Jehn, Chadwick, &
Thatcher, 1997), as well as attitudes, preferences, and beliefs (Harrison et al., 1998).
The present study furthers research on work team diversity by continuing to delve into
its surface- versus deep-level forms, by examining the course of its effects from actual member
differences through perceived differences, and by highlighting the importance of time. We
separate the effects of overt demographic differences from underlying psychological differences
in the current study, predicting that they will have differentially salient consequences for team
social integration over time, as team members collaborate and learn more about each other.
Time, Teams, and Task Performance
In Press - AMJ
4
Another contribution of the current research is examination of a process assumed to occur but
rarely assessed in diversity research (Riordan, 2000): tracking effects of actual team member
differences through the perceptions of such differences. Finally, by examining these issues using
a four-wave design problems of reverse causation and common method variance can be mitigated,
affording stronger conclusions than in earlier cross-sectional studies.
LINKS FROM DIVERSITY TO TEAM SOCIAL INTEGRATION AND PEFORMANCE
As research accumulates, there has been growing recognition that the paths linking work
team diversity to team functioning and performance outcomes are complex. These complexities
are reflected in the theoretical framework guiding our research, presented in Figure 1. In the next
sections we review theory and data supporting the proposed relationships in our model.
--------------------------------
Insert Figure 1 about here
--------------------------------
Multiple Types of Diversity
Surface-Level Diversity
Definition. Surface-level diversity (Harrison et al., 1998) is defined as differences among
team members in overt demographic characteristics (Milliken & Martins, 1996; also, see Riordan,
2000, who describes these as "surface-level, visible dimensions"; p. 162). Such characteristics
include age, sex, and race/ethnicity that are often reflected in physical features. Almost
immediately, individuals can make reasonable estimates of the age, gender or racial/ethnic
background of someone else, and therefore, that person's (dis)similarity to themselves (Jackson et
al., 1995). Surface-level diversity is equivalent to what other researchers label as social category
Time, Teams, and Task Performance
In Press - AMJ
5
diversity (Jehn, Northcraft, & Neale, 1999) or demographic diversity (e.g., O'Reilly, Caldwell, &
Barnett, 1989). Most importantly, it is well established that individuals quickly use these
characteristics to assign themselves and others to social classifications involving ascribed patterns
of thought, attitudes, and behaviors (e.g., Fiske, 2000).
General theories. There are several broad, deductive theories that have been used to
support arguments about the consequences of surface-level diversity. Some researchers (e.g.,
Jackson, Brett, Sessa, Cooper, Julin, & Peyronnin, 1991) posit inverse relationships between
surface-level differences and team functioning or performance outcomes using basic propositions
of social identity (Tajfel, 1978) and self-categorization theories (Turner, 1982). These theories
propose that group members define and differentiate themselves from others based on observable
differences such as age, race, and gender. As individuals are motivated to maintain or enhance
their social identities, they are more likely to positively evaluate and identify with persons and
groups whose members appear to hold the same overt features they do (Tajfel & Turner, 1986).
Another general theoretical premise used by diversity researchers (e.g., Barsade et al.,
2000) is the well-supported similarity-attraction paradigm (Byrne, 1971). People are attracted to
and prefer to be with similar others because they anticipate their own values, attitudes, and
beliefs will be reinforced or upheld. Both categorization and similarity-attraction theories make
the same prediction. Team members tend to have less positive attitudes toward, and form fewer
social attachments with, those who are perceived to be less like themselves.
Specific elements. We chose to continue to study age, sex, and race/ethnicity, as surface-
level diversity variables in the current investigation. Their effects are inconsistent, but they have
Time, Teams, and Task Performance
In Press - AMJ
6
a fairly long history of study in this area (Riordan, 2000). They are also immediately
recognizable and used to assign others to tacit social categories (Tsui & Gutek, 1999).
Although some researchers report null (Bantel & Jackson, 1989; Wiersema & Bird,
1993), or asymmetrical effects (Chattopadhyay, 1999), differences in age are often negatively
related to individual or within-team functioning: social isolation (Kirchmeyer, 1995), reduced
cohesion (O’Reilly et al., 1989), lowered communication (Zenger & Lawrence, 1989), and higher
turnover (Jackson, et al., 1991; O'Reilly et al., 1989; Tsui, Egan, & O'Reilly, 1992). Similar
relationships exist for sex diversity in teams, where null and asymmetrical effects have been
shown (Chattopadhyay, 1999; Riordan & Shore, 1997), but researchers frequently report
negative outcomes such as feelings of isolation, dissatisfaction, and lack of attachment in some
sex- dissimilar situations for females (Konrad, Winter & Gutek, 1992; Pelled & Xin, 1997), and
reduced organizational attachment for some males (e.g., Tsui et al., 1992). Reported
consequences of racial or ethnic diversity are also somewhat inconsistent (Williams & O'Reilly,
1998), and sometimes nonlinear (Chattopadhyay, 1999). However, lower performance ratings
(Kraiger & Ford, 1985), diminished communication (Larkey, 1996), and reduced commitment
(among majority members: Tsui et al., 1992), are related to within-team differences in race.
Marital status was also included as an manifest difference in our study. It is a surface-
level diversity variable that has not been examined in previous research, although it is both an
overt and immediately recognizable demographic characteristic (i.e., from wedding rings). In pilot
interviews this demographic feature was found to be perceptually salient to our current sample of
undergraduate and graduate students working on course projects. Most importantly, marital
Time, Teams, and Task Performance
In Press - AMJ
7
status marks clearly different social categories (Tsui & Gutek, 1999) prompting attributions,
stereotypes, and perhaps even different interpersonal affiliations.
Deep-Level Diversity
Definition. Deep-level diversity refers to differences among team members' psychological
characteristics: personalities, values, and attitudes (Jackson et al., 1995; Harrison et al., 1998).
Clues to these latent individual differences are taken from members' interactions with one another
as they unfold over time. Those clues are expressed in behavior patterns, verbal and non-verbal
communication, and in exchange of personal information.
General theories. Notably, the theories marshaled by authors to support surface-level
diversity effects say as much, if not more, about deep-level effects (Tsui et al., 1992). That is,
presumed underlying differences between people in their attitudes, values, and personality are
the basis of similarity-attraction or fit paradigms, including social psychological theories about
similarity in attitudes (e.g., Newcomb, 1961; Byrne, 1971) and organizational behavior theories
about similarity in values and personality (e.g., Schneider, 1987). The arguments are virtually
identical to those presented above: people find it more pleasurable to interact with others who
have similar psychological characteristics, because it verifies and reinforces their own beliefs,
affect, and expressed behaviors (e.g., Swann, Stein-Seroussi, & Giesler, 1992). This form of
attraction occurs even when attitudes are negative ("we hate this project") or when personality
dimensions are dysphoric (Locke & Horowitz, 1990).
Personality elements. Evidence from other literatures and at other levels of analysis
suggests an impact of personality differences within teams (e.g., Schneider, Goldstein, & Smith,
Time, Teams, and Task Performance
In Press - AMJ
8
1995). There are innumerable personality differences that might make up deep-level diversity.
In task-performing groups researchers have proposed that the task- or job-relatedness of
differences is critical (e.g., Harrison et al. 1998). That is, does the deep-level attribute bear on
fundamental purposes of the team? With that criterion in mind, we adopted conscientiousness as
the deep-level personality difference most likely to be consequential in the teams we studied.
At the individual level, Barrick and Mount (1991) demonstrated that conscientiousness
was the Big Five dimension most consistently and most strongly related to performance across a
variety of task settings. However, other studies have found mixed effects of conscientiousness at
the team level. Barrick, Stewart, Neubert, & Mount (1998) reported higher mean levels of team
conscientiousness were associated with team performance but team member diversity in
conscientiousness was not associated with team viability or social cohesion. LePine, Hollenbeck,
Ilgen, and Hedlund (1997) found that group members low in conscientiousness were ignored by
the rest of the team, while Neuman, Wagner, and Christiansen (1999) did not detect differences in
team performance as a function of mean levels of team conscientiousness. In sum,
conscientiousness is theoretically implicated in team dynamics, but the data are equivocal.
Value elements. There have been few empirical investigations of this aspect of deep-level
diversity in ongoing work teams (e.g., Jehn, Northcraft, & Neale, 1999). Still, extant literature
suggests a negative association between value diversity and outcomes in teams. O'Reilly,
Chatman, and Caldwell (1991) showed that new employees whose individual values differ from
the mean values of the rest of their work groups or small organizations are less satisfied,
demonstrate lower organizational commitment, and are more likely to quit. Jehn and Mannix
Time, Teams, and Task Performance
In Press - AMJ
9
(2001) reported that greater a priori consensus on work values leads to effective patterns of task
conflict and lower levels of relationship conflict over time.
Tasks in our study were not being done for an organization, as in the studies cited above,
but as part of students' class work. Therefore, we chose to study how student team members
saw their educational context as fundamental to their values, using Rokeach's (1973) terminal
values as our probe of value diversity. It has been used with success in predicting various
between-group differences for students (e.g., college major, church attendance; Schwartz &
Bilsky, 1990).
Attitude elements. Attitude similarity is an important predictor of attraction and
friendship (Byrne, 1971; McGrath, 1984; Newcomb, 1961). In work situations, attitude
similarity among team members has been linked to higher levels of team cohesiveness (Harrison et
al., 1998). In keeping with our job-relatedness criterion, we chose to study within-team
differences in two attitudes, task meaningfulness and outcome importance. Task meaningfulness
refers to the personal salience and importance of the team's project. It has been associated with
greater intrinsic motivation and with more pronounced attitudinal and behavioral consequences
(Petty & Cacioppo, 1979). Outcome importance refers to the value of getting a good project
grade. It reflects the valence, which is a central construct in theories of motivation (e.g., Vroom,
1964), of the project grade to the participants.
Actual and Perceived Diversity
Testing the Assumption of Veridical Perceptions
In the preceding literature review measures of diversity, typically relational (Euclidian
Time, Teams, and Task Performance
In Press - AMJ
10
distance) or compositional (standard deviation) measures, are examined in terms of team
functioning and team performance (Williams & O'Reilly, 1998). Results of that approach are
mixed in findings, and generally weak in terms of effect sizes (R2 = .02-.19: Riordan, 2000). One
approach to resolve inconsistencies across studies, to explain small effect sizes, and the impact of
different facets of diversity, is to note that diversity effects rely on perceptions (Lawrence,
1997). Yet, these perceptual mechanisms are rarely studied in diversity research (see Riordan,
2000, pp. 160-161, for a detailed account of theoretical statements and existing data).
Consequently, based on the foundation that if differences are to be meaningful they must be
perceived, our first set of hypotheses, which is also reiterated in Figure 1, is described below.
Hypothesis 1: (a) Actual surface-level diversity will have positive effects on perceived
surface-level diversity; and, (b) actual deep-level diversity will have positive effects on
perceived deep-level diversity.
Perceived Diversity and Team Social Integration
We propose that the impact of reactions to perceived diversity on team functioning
should be felt on team social integration. Our use of this term is similar to O'Reilly et al. (1989),
and Smith, Smith, Olian, Sims, Jr., O'Bannon, & Scully (1994). For these researchers, team social
integration is a multi-faceted construct including elements of cohesiveness, satisfaction with co-
workers, positive social interaction, and enjoyment of team experiences. Elements of team social
integration are the most commonly studied outcomes in diversity research (Tsui & Gutek, 1999).
Consequently, our second set of hypotheses covers the next part of the model (see Figure 1),
based on our assertion that perceived diversity reflects the psychological importance of, and
carries the substantive impact of, actual diversity into team social integration (cf., Ashforth &
Time, Teams, and Task Performance
In Press - AMJ
11
Mael, 1989). Theory and research suggests a negative relationship between both aspects of work
team diversity and team social integration.
Hypothesis 2: (a) Perceived surface-level diversity and (b) perceived deep-level diversity
will have a negative effect on team social integration.
Mediation by Perceptions and the Indirect Nature of Actual Diversity Effects
If the first two hypotheses are taken to be logical premises, they necessitate a third set of
hypotheses as a conclusion. Specifically, if actual diversity affects perceived diversity and
perceived diversity affects team social integration, then perceptions are a mediating (intervening)
construct. Actually diversity has only indirect effects on team social integration (see Figure 1).
Results from a few studies bear indirectly on this conclusion. Cleveland and Shore (1992)
reported that inclusion of perceptual measures of relative age, in additional to chronological age,
increased predictability of individual outcomes such as perceived organizational support. Turban
and Jones (1988) showed that perceptions of (rather than actual) attitudinal similarity between
supervisors and subordinates were uniquely and positively related to subordinates' satisfaction,
performance ratings, and pay ratings. These arguments and data lead to our third proposition.
Hypothesis 3. Perceived (surface- and deep-level) diversity will mediate the negative
impact of actual (surface- and deep-level) diversity on team social integration.
Team Reward Contingency
Auspiciously absent from most diversity research is a consideration of team reward
structures (e.g., Tsui & Gutek, 1999). In the current paper, we define team reward contingency
as the degree to which individual members' outcomes depend on team outcomes. This construct
Time, Teams, and Task Performance
In Press - AMJ
12
is also referred to as shared fate (Pettigrew, 1998) and positive outcome interdependence (Kelley
& Thibaut, 1978; Tjosvold, 1984; Wageman, 1995). As personal costs and benefits become more
contingent on how well the team performs, individuation theory (Fiske, 2000) predicts that team
members will pay more attention to one another's personal (deep-level) features and reduce
stereotypic thoughts and evaluations about them. Moreover, interdependence theory (Kelley &
Thibaut, 1978) and cooperation theory (Tjosvold, 1984) predict that this alignment of individual
and team outcomes motivates members to spend more time and effort interacting with one
another, which is our definition of collaboration.
Hypothesis 4. Strength of team reward contingency will have a positive effect on team
collaboration.
Moderating Effects of Time via Collaboration
Identity theory, categorization theory, and prior evidence all support the notion
that in initial interactions, categorization of other team members is based on surface-level features
(Schneider et al., 1995; Berger, Rosenholtz, & Zelditch, 1980). Our model takes these ideas a
step further and proposes that over time, as team members collaborate, they have more
opportunities for the exchange of personal, idiosyncratic information and larger samples of each
other’s behavior to observe (Gruenfeld, Mannix, Williams, & Neale, 1996). Consequently, over
time, as team members learn more about each other we propose that surface-level diversity
becomes less important and deep-level diversity becomes more important in determining team
social integration.
Support for proposition comes from a series of recent studies. Pelled, Eisenhardt, and
Time, Teams, and Task Performance
In Press - AMJ
13
Xin (1999) reported that the effects of surface-level diversity (age and member tenure) on
emotional conflict diminished as a function of team longevity. Jehn et al., (1999) suggested that
social category (surface-level) diversity may become less relevant over time and other factors,
such as value (deep-level) diversity, may become more salient determinants of team morale.
Finally, using a cross-sectional design, Harrison et al. (1998), reported that the influence of sex
differences on group cohesiveness diminished and the influence of attitude differences increased
as a function of team tenure.
Hypothesis 5: As team members collaborate more (spend more time together) the impact
of (a) perceived surface-level diversity on team social integration will diminish, but the
impact of (b) perceived deep-level diversity on team social integration will intensify.
Team Social Integration and Task Performance
The last part of our model proposes that team social integration is the proximal input to a
more distal outcome, team task performance (see Figure 1). The rationale is that teams with a
higher level of social integration are more willing to subjugate individual interests for team goals
that should direct member resources toward higher team task performance. This proposed role is
consistent with findings that team social integration mediated the effects of surface-level diversity
on tenure (O'Reilly et al., 1989), and mediated the impact of diversity in experience on
performance (Smith et al., 1994. There is also strong meta-analytic evidence that group
cohesiveness (a primary dimension of social integration) facilitates performance with consistent
effects across a wide variety of settings and tasks (Gully, Devine, & Whitney, 1995). This leads
us to our final proposition.
Time, Teams, and Task Performance
In Press - AMJ
14
Hypothesis 6: Team social integration will have a positive effect on task performance.
METHOD
Teams and Tasks
To test our hypotheses, we tracked the development and performance of 113 graduate
and 449 undergraduate business students doing team projects at a large university in the
southwest. Team members were enrolled in 23 upper-division or graduate sections of 13 different
courses. Individual ages of participants ranged from 19 to 55; mean age was 28 (sd = 7). Fifty
percent of the participants were female; 65% percent were Caucasian, 9% African American,
18% Asian, 7% Hispanic, and 1% were Native Americans or Pacific Islanders. Approximately
57% of the participants were not married. Seventy-five percent of the participants currently
worked at least part-time at a paid job.
A total of 144 teams were formed, each with a minimum of 2 to a maximum of 9
members. Mean team size was 4.1 members (median = 4). Only five percent of the teams were
dyads; 80% of the teams had 3-5 members. Team projects lasted from 9 to 14 weeks of the 16-
week semester. Project tasks varied from industry sector analyses (for strategic management
classes) to customer survey development (marketing) to corporate audits (accounting) to web
programming (information systems).
Final assignment of members to teams was determined by the instructor, typically on a
random basis, but placement could have been suggested at the beginning of the semester by
students. All decisions regarding team membership, projects, grading, and so on were made by
each instructor with no input from the researchers. Members of the research team taught none of
the classes in this research, nor were any of the instructors or participants involved in the study
Time, Teams, and Task Performance
In Press - AMJ
15
aware of the hypotheses being tested.
Design and Procedures
Team members completed survey instruments at three different times over the course of
nearly four months. The first survey (Time 1) was completed and returned in the first two
weeks of the semester, before individuals had been assigned to teams or had worked with other
team members. The second survey (Time 2) was completed and returned in the fifth to seventh
week of the semester, after teams were formed and members had worked together for
approximately 1/3 to 1/2 of the total time they needed to complete their project. The third and
last survey (Time 3) was administered in the last two weeks of the semester, when team
members had just completed their assigned project(s), but had not yet received their project grade
or final course grade. Each wave of surveys was distributed at the same (calendar) time to
everyone. Finally, after the end of the semester (Time 4) teacher evaluations of team
performance were collected. To be included, at least 2/3 of team members had to return all three
waves of surveys.
At each administration, course instructors distributed the surveys to the participants.
Each survey was accompanied by a sealable return envelope. Each questionnaire contained a
unique ID number for tracking over time. In most classes (21 out of 23), instructors offered a
nominal amount of extra credit to teams in which all members returned all three waves of surveys.
Measures
Surface-level diversity. The characteristics of team members selected for inclusion were
age, sex, ethnicity and marital status. All surface level demographic measures were collected via
Time, Teams, and Task Performance
In Press - AMJ
16
self-report at Time 1. We used within-group standard deviations (sd) to reflect diversity in the
ages of team members (Bedeian & Mossholder, 2000). Blau's (1977) index was used to indicate
diversity in sex, racial/ethnic background, and marital status (cf., Bantel & Jackson, 1989).
Deep-level diversity. At Time 1, we administered a 24-item conscientiousness measure
(Goldberg, 1992). Coefficient alpha at the individual level was .89. At Time 1, participants also
completed Rokeach's (1973) scale, rating the extent to which their university courses would help
them attain each of the terminal values (e.g., a comfortable life), on a format ranging from not at
all (1) to a great extent (7). Estimated was α = .93. To assess the deep level attitudes, a three-
item scale assessed task meaningfulness, in which participants strongly disagreed (1) to strongly
agreed (7) with statements that they could “learn a lot from the project,” “it was more than busy
work,” and that “doing the project is worthwhile.” Estimated reliability was α = .89. A second
measure of outcome importance consisted of two-items asking team members the extent to which
it was "important to get an 'A' on the project," and how much they "needed to do well."
Responses ranged from not important (1) to extremely important (7). Coefficient alpha was .87.
We used the within-group standard deviation on each variable to index deep-level diversity.
Team reward contingency. Team projects contributed from 10% to 75% of each
student’s final course grade. The mean level of this shared outcome was 34%. The same reward
contingency applied to all students enrolled in the same course, and it was known to students
before their project work started. Two different members of the research team examined the
syllabi and coded the team reward contingency variable. There was 100% agreement.
Perceived surface-level and deep-level diversity. Time 2 indices of perceived diversity
Time, Teams, and Task Performance
In Press - AMJ
17
were adopted from Harrison et al. (1998). Team members rated how members of their project
group were very similar (1) to very different (5) on three surface-level diversity variables (age,
ethnicity, and marital status). As Harrison et al. (1998) cautioned, a question about how much
there were sex differences within the team would have had little face validity because of the
glaring obviousness and dichotomous nature of the sex variable. Seven questions assessed deep
level diversity, using the same response format. They included perceptions of (dis)similarity on
personal values, personalities, priorities, commitment to the project, attitudes about school and
education, and project goals (the latter adapted from Jehn, 1995). Estimated reliabilities were α =
.68 and α = .82 for perceived surface-level and deep-level diversity, respectively.
Collaboration. Participants were asked two open-ended questions to assess the number
of times they met with some or all members of their team. Based on similar constructs discussed
by Hambrick (1994), Campion, Medsker, and Higgs (1993), and Wageman (1995), we also
included five items (scaled from 1= not at all to 5 = all the time) asking team members how often
they engaged in interactions such as reviewing each other’s work and coordinating each other's
activities. The open-ended and scaled items were first standardized and then summed.
Coefficient alpha for this measure was .75.
Team social integration. Following O'Reilly et al. (1989) we administered a number of
different measures to assess the extent of team social integration at Time 3. Based on an existing
measure of cohesiveness, participants indicated the extent they disagreed (1) to agreed (7) with
three items (e. g., "I really like most of the other group members;" Seashore, 1979; α = .82).
Using the same response format, satisfaction with the team was measured using a three-item scale
Time, Teams, and Task Performance
In Press - AMJ
18
(e.g., "my team is doing a good job;" α = .92). A two-item scale measuring fairness of team
practices was adopted from research in procedural justice (e.g., "team decision processes were
fair;" α = 92). Additionally, team members rated their willingness to work with each of the other
team members on a future class project. Ratings for each judge (on a seven-point scale) were
averaged to obtain a measure of attraction toward the other team members. As expected, all these
measures correlated highly with each other (ranging from r = .65 to .75). Therefore, we
standardized each of the four original measures and summed them to create a composite measure
of team social integration. Coefficient alpha for this composite was .91.
Team Performance. Each teacher had a slightly different grading scheme, percentage, or
point total for the team projects. To minimize differences across classes, we created a single
standardized index. This index was a ratio of number of team points earned on the project relative
to the maximum number of points that could have been earned. Theoretically, this index could
have ranged from 0 to 1.00; however, no team scored below .50
Aggregation of Individual Responses to the Group Level
Team reward contingency and team performance were measured directly at the group
level. For measures that were initially taken at the individual level, it is important to show
agreement or consensus among within-team responses before aggregating them to the group-level.
Therefore, we calculated the James, Demaree, and Wolf (1993) within-group agreement (rwg)
index for each measure we wished to aggregate, for each of the 144 teams in our analyses. The
average rwg indices were .82, .86, .92, and .82, respectively, for perceived surface-level diversity,
perceived deep-level diversity, collaboration, and team social integration. Likewise, 88%, 81%,
Time, Teams, and Task Performance
In Press - AMJ
19
97%, and 92% of the teams (respectively) had rwg > .70 for each of the four aggregate measures.
Finally, the group-level component of variance was strong in each aggregate measure, explaining
more than 50% of the variance in all four constructs (F 143,445 = 3.49, 1.83, 4.25, and 4.51, in the
same order as listed above, all p-values < .001). Means, correlations, and standard deviations for
these and all of the other team-level variables are reported in Table 1. Hypothesis tests are
shown in Tables 2-4.
------------------------------------
Insert Tables 1-4 about here
------------------------------------
RESULTS
Control Variables
Team size (e.g., Jackson et al., 1991) and cognitive ability (Barrick et al., 1998) can
influence a variety of processes, outcomes, and diversity measures. Therefore, each of these was
first entered as a control variable in the analyses described below (ability was measured as mean
GPA). Also, group total or average scores on deep-level diversity measures can be confounded
with within-group standard deviations (Bedieian & Mossholder, 2000). Therefore, group means
on conscientiousness, value attainment, attitudes toward the project, and the importance of
grades were also used as control variables in the analyses involving deep-level diversity.
Linkages Between Actual and Perceived Diversity
Hypotheses 1a and 1b predicted that actual differences in diversity variables would be
reflected in perceptions, both surface-level and deep-level variables, respectively. One of the
more interesting features of the correlations in Table 1 is the empirical separation of types of
Time, Teams, and Task Performance
In Press - AMJ
20
actual diversity (surface v. deep), an important condition for the viability of our theoretical
arguments. There are weak but significantly positive correlations within sets of actual surface
level and deep-level variables. But, none of the sixteen correlations between these sets is
significant and the pattern suggests no general positive or negative connection.
Connections between actual and perceived diversity followed our predictions.
Supporting Hypothesis 1a and the implicit assumption of most relational demography research.
A regression test of that hypothesis (Table 2) showed that actual diversity in age, race/ethnicity,
and marital status each had significant regression weights in a model that explained 50% of the
variance in (overall) perceived surface-level differences. Notably, one deep-level diversity
measure, outcome (grade) importance also contributed to the perception of surface differences
suggesting that more than simple outward features might influence surface-level perceptions.
Results for Hypothesis 1b were also supportive. Although actual differences in
conscientiousness and values were not uniquely related to (overall) perceived deep-level
diversity, actual differences in task meaningfulness and outcome importance were, as shown in
the pattern of regression weights. None of the actual surface-level diversity variables helped to
predict this deep-level perception.
Linkages Between Perceived Diversity and Team Social Integration
Hypotheses 2a and 2b predicted effects of perceived surface- and deep-level diversity on
social integration. Strong and supportive results for these Hypotheses are presented in Table 1
(r's = -.41 and -59, respectively; p < .01) and in the second column of Table 3. Perceptions of
both forms of diversity are uniquely related to lower team social integration (adjusted R2 = .35, p
Time, Teams, and Task Performance
In Press - AMJ
21
< .01), although the standardized regression weight for perceived surface-level diversity is less
than half as large as that for perceived deep-level diversity (β's = -.18 and -.42, respectively; p <
.05). In other words, early perceptions of both demographic and psychological differences
among team members have important negative consequences for how well a diverse group gets
along -- months later -- with the latter perceptions being more consequential than the former.
Mediating Effects of Perceived Diversity
Following Baron and Kenny's (1986) suggested procedures, a series of hierarchical
regression analyses were used to test Hypotheses 3a-b, which proposed intervening effects of
perceived diversity. Inspection of Table 1 shows that one of the actual surface-level variables
(ethnicity) and two of the actual deep-level variables (task meaningfulness and outcome
importance) had significant zero-order relationships. A more rigorous test of this first premise of
mediation, that the distal independent variables are related to the dependent variable, is given in
the first column of Table 3. These results show a weak but detectable relationship between the
entire group of actual diversity variables and team social integration (adj. R2 = .06, p < .05). Of
these eight variables, however, only outcome importance had a significant regression weight. The
second and third premises of mediation testing are that distal independent variables are related to
the mediators, and the mediators are in turn related to the dependent variable. Both of these
premises were satisfied through the previous regression tests for Hypotheses 1a-b and 2a-b.
The fourth and final step for mediation testing is shown in the third column of Table 3.
All measures of perceived diversity and actual diversity were entered into the same regression
equation predicting team social integration. In this overall equation, both perceived surface-level
Time, Teams, and Task Performance
In Press - AMJ
22
(β= -.25, p<.05) and perceived deep-level (β= -.38, p < .01) diversity had significant, negative
impacts on team social integration. The overall (as a set) unique contribution of actual diversity
variables to team social integration was non-significant (F8, 123 = 1.00, p > .10). In this last
equation, the lack of significant regression coefficients for actual diversity variables also supports
a statistical conclusion of complete mediation. Yet only one had a significant weight to begin
with, making it difficult to argue the data authorized a conclusion of substantive mediation.
Indeed, the weak and inconsistent relationships between actual diversity and team social
integration was part of the motivation for the current study. Therefore, a more descriptive
conclusion from these analyses might be to regard perceived diversity as intervening variables,
and the effects of actual diversity as indirect (James & Brett, 1984).
Linkage Between Team Reward Contingency and Collaboration
Hypothesis 4 predicted that a stronger team reward contingency would lead to greater
collaboration. It is supported by a positive correlation (r = .38, p < .05). In more rigorous tests
of this linkage, we regressed collaboration on all the other control and actual diversity variables
measured at Time 1. In each case, team reward contingency was the strongest predictor, uniquely
explaining at least ten percent of variance in collaboration (p < .01). Reinforcing its place in our
model, team reward contingency had no direct or moderating relationships with either form of
perceived diversity, or with team social integration (see Table 1 for correlations).
Moderating Effects of Collaboration
Hypotheses 5a-b proposed that as time spent in collaboration increases, the negative
impact of perceived surface-level diversity diminishes and the impact of perceived deep-level
Time, Teams, and Task Performance
In Press - AMJ
23
diversity grows. To test these hypotheses, control variables, measures of perceived diversity,
and collaboration were first entered as main effect predictors of team social integration (see the
second column of Table 3). Next, moderator terms were created by multiplying each of the
perceived diversity measures by collaboration. When the moderator terms were entered into the
equation (see the final column of Table 3), there was a significant increase in R2 (∆R2 =.06, F2, 136
= 7.99, p < .01). The negative effect of perceived surface-level diversity flattened toward zero (β
for perceived surface-level diversity × collaboration =.84, p < .05). In contrast, the negative
effect of perceived deep-level diversity became steeper (β for perceived deep-level diversity ×
collaboration = -1.15, p < .01). Supplementary plots in Figure 2, shown for a three levels of
collaboration, confirm this interpretation.
--------------------------------
Insert Figure 2 about here
--------------------------------
Linkage Between Team Social Integration and Task Performance
To test Hypothesis 6, that team social integration positively affects performance, control
variables were first entered into a regression predicting team performance, followed by team
social integration and collaboration. Consistent with our proposition (left column of Table 4),
social integration had a strong, positive impact on task performance (β = .32, p < .01). Along
with team ability and collaboration, it helped to explain an adjusted R2 = .23 (p < .01) in how
well the teams carried out their projects. The second and third columns of Table 4 also show that
this relationship between team affect and team performance remains strong even when perceived
and actual diversity variables are added as possible predictors. Another noteworthy feature in
Time, Teams, and Task Performance
In Press - AMJ
24
Table 4 is the fact that none of the diversity variables has a unique effect on performance once
team social integration is accounted for.
DISCUSSION
The present study continues a "new look" in diversity research (e.g., Riordan and Shore,
1997), by (a) separately assessing the impact of surface-level and deep-level types of diversity,
(b) specifying and testing the presumed but unexamined mediating role of diversity perceptions,
(c) proposing and testing the moderating influence of temporal constructs (collaboration) on the
consequences of diversity, and (d) showing how diversity perceptions first travels through team
social integration before having an effect on performance.
Distinctions Among Surface- and Deep-Level, Actual and Perceived Forms of Diversity
We tried to include a broad range of potential forms of diversity in this study from
conventional demographic facets such as age, sex, and so on, to forms of deep-level diversity
including personality, values, and attitudes (e.g., Harrison et al. 1998). The association of actual
surface and deep-level measures to their perceptual counterparts attests to the salience of these
actual differences in the shared cognitive framework of team members. Findings indicated that
the actual measures of surface-level diversity did not contribute to perceptions of psychological
diversity, suggesting that categorical or overt differences have little impact on the specific deep-
level inferences we studied (e.g., task meaningfulness).
Two actual, deep-level diversity variables, Rokeachian values and conscientiousness, were
not linked with corresponding perceptual measures of diversity, while differences in deep level
attitudes (task meaningfulness and outcome importance) were strongly related to their
Time, Teams, and Task Performance
In Press - AMJ
25
corresponding perceptual measures. As measures of perceived diversity were assessed early in
the team's development, participants may not have had enough time to learn about these deeper-
level personality or value differences, or they may have suspended their judgments until a larger
sample of other members' behaviors could be obtained. Additional possibilities are participants
may want to mask certain types of deep level differences in order to appear to fit in with the
other members of the team and in the relatively short life of span of our teams, task related
diversity may be more salient than value based diversity. As members of the team continue to
work together, it may be harder to mask these deep level differences, and such differences might
become more salient to the team members. These arguments are consistent with the findings of
personality theorists who report that over longer periods of time there is growth in the variance
of behavior that can be accurately attributed to personality or value differences (Epstein, 1980).
The inclusion of perceptual measures as markers of the salience of actual amounts of
diversity might help researchers in two ways. First, they can assist in determining which
elements of diversity are relevant to participants. For example, task-related attitudes of team
members in other settings may be more homogeneous (e.g., all members strongly committed to
attaining the same goal, reward, or outcome). Consequently, the inclusion of perceptual measures
of diversity would provide a critical check on the salience of this specific deep-level facet.
Second, our results indicate perceptual measures of diversity carry the indirect impact of a-priori
or actual differences among team members. These findings provide support for researchers who
have argued that perceptual measures can enhance our ability to account for diversity related
outcomes, and potentially reconcile inconsistencies across studies (e.g., Lawrence, 1997).
Time, Teams, and Task Performance
In Press - AMJ
26
Differences between surface- and deep-level facets of diversity were also uncovered that
served as something of an assumption check in our model. In supplementary data, ratings of how
much the participants said they knew about other team members increased significantly from
Time 2 to Time 3 -- the only occasions at which they were measured -- for totals of surface- and
deep-level characteristics. More importantly, the increase in knowledge for the deep-level
features was significantly greater than for surface-level features. Based on these findings, time
did serve as medium through which members exchanged personal information and exhibited
behaviors reflecting (at least in team members' minds) more fundamental psychological features.
Time and the Consequences of Getting Together
In one way, our findings support the general assumption of relational demography, that
outward differences in groups are quickly perceived and used to make judgments (Tsui et al.,
1992). Those judgments affect later interactions. On the other hand, our data suggest that getting
together frequently can reduce the impact of demographic differences, as suggested by various
organizational theories (e.g., Elsass & Graves, 1997).
Accounting for collaboration might also help to explain findings of previous research
reporting no impact or a small impact of surface-level diversity on team outcomes (Riordan,
2000). It is possible that as research examines work groups that are psychologically meaningful
to the participants (e.g., top management teams) they may be capturing teams at a stage of their
development when they have collaborated enough to have sufficient information to make
judgments of deep-level diversity. Under these circumstances, our research suggests that
personality conflicts, disagreements about strategic goals, and differential levels of commitment
Time, Teams, and Task Performance
In Press - AMJ
27
to the organization might be more crucial to communication, team cohesion and perhaps, firm
performance (Zenger & Lawrence, 1989).
Our findings do sound an optimistic note. For managers, the effects of demographic
diversity play a less significant role on team outcomes than may have been thought. Our results
do not suggest, however, that the path to team social integration is necessarily an easier one. The
challenge managers have is to find a way to integrate team members who differ in underlying, but
fundamental ways. One step in the development of such programs is to look beyond the
presumed negative impact of surface-level diversity to possible positive effects (e.g., Jehn, 1999),
as well as to assess the deep-level differences that might erupt into negative affect and
relationship conflict. In that sense, managing diverse teams might be akin to doing a tacit job
analysis. Maximizing differences in knowledge, skills and abilities, while minimizing differences
in personalities, attitudes and values, might create especially effective teams.
Another step in programs for managing surface-level diversity might be to structure
rewards that foster greater collaboration. Our findings fit with of research (DeMatteo, Eby, &
Sundstrom, 1998) showing that when members' individual outcomes depend more on team
performance, they collaborate more frequently. This collaboration brings important deep-level
similarities and differences to the foreground and pushes surface-level effects to the background.
Importance of Team Social Integration
Team social integration strongly predicted team performance, even after controlling for
potential confounds. This linkage between social integration and performance converges with
other diversity research (e.g., O'Reilly et al., 1989), and a wealth of studies in the groups
Time, Teams, and Task Performance
In Press - AMJ
28
literature (i.e., the correlation we obtained is within the confidence interval of the meta-analytic ρ
reported by Gully et al., 1995). Team social integration also absorbed the effects of perceived or
actual diversity, which had no unique impact on performance. These findings suggest that the
diversity we studied is mainly important to the extent it plays into team social integration.
Limitations and Research Directions
We designed our study to guard against several common limitations. Given the time
separation between questionnaires, and because common method variance (CMV) is conceived
primarily as a transient, memory-based systematic error, it cannot provide an explanation for our
results. Additionally, because CMV itself is a type of main effect or correlated error, it cannot
explain the moderating effects of collaboration (Harrison, McLaughlin, & Coalter, 1996).
Although our measures of actual diversity, social integration, and collaboration had good
psychometric properties, the perceived diversity measures require further construct validation
work. In exploratory factor analyses, perceived surface- and deep-level diversity did form a
separable factor. However, the deep-level items contain fairly generic perceptions (e.g.,
differences in "priorities", and "personalities") that did not provide the same one-to-one mapping
to actual diversity measures that were in the surface-level items (e.g., differences in "age" and
"ethnic background"). It will be important to see if perceptual measures can be developed that
have the same level of specificity as actual deep-level differences.
The use of student teams working on class projects also raises serious questions about the
external validity of our findings.. On the positive side, the team members were not role-playing;
they were "field" teams in the sense that they would have existed whether or not we were doing
Time, Teams, and Task Performance
In Press - AMJ
29
ongoing research. They would fall under the rubric of task forces in McGrath's (1984) typology,
in that they each had an assigned project with a fixed duration. In addition, team task
performance had meaningful outcomes (10%-75% of course grades) for all of the participants.
On the negative side, the teams we studied were together for only 9-14 weeks. Despite our own
emphasis on the importance of time, the brief lifetime of these teams limits generalizability of our
results to newly formed and short-term task forces, and argues for work on long-term groups,
perhaps with changing tasks or membership. Although we examined time, we did so ordinally,
and in the compressed time of student project teams -- not in the natural or 'system time' of most
organizational teams (Kelly & McGrath, 1988).
Conclusion
We have tested a comprehensive model of the effects of work team diversity, providing
possible explanations for previous inconsistencies and small effect sizes in diversity research.
Consideration of perceived diversity, and the moderating role of collaboration made possible by
the passage of time, are important in understanding which differences make a difference and the
consequences of diversity. We think that bringing time more fully into the open provides a
compelling medium and conceptual lever for theorizing about diversity's effects in teams,
minimizing its potentially negative consequences, and developing mechanisms to capitalize on it.
Time, Teams, and Task Performance 30
REFERENCES
Ashforth, B. E., & Mael, F. A. 1989. Social identity theory and the organization. Academy of
Management Review, 14: 20-39.
Bantel, K. A. & Jackson, S. E. 1989. Top management and innovations in banking: Does
composition of the top team make a difference? Strategic Management Journal, 10: 107-124.
Baron, R. M., & Kenny, D. A. 1986. The mediator - moderator variable distinction in social
psychological research: Conceptual and statistical considerations. Journal of Personality and
Social Psychology, 51: 1173-1182.
Barrick, M. R., & Mount, M. K. 1991. The Big Five personality dimensions and job performance:
A meta-analysis. Personnel Psychology, 44: 1-26.
Barrick, M. R., Stewart, G. L., Neubert, M. J., & Mount, M. K. 1998. Relating member ability and
personality to work-team processes and team effectiveness. Journal of Applied Psychology,
83 (3): 377-391.
Barsade, S. G., Ward, A. J., Turner, J. D. F., & Sonnenfeld, J. (2000). To your heart's content: A
model of affective diversity in top management teams. Administrative Science Quarterly,
45(4): 802-836.
Bedeian, A. G., & Mossholder, K. W. 2000. On the use of the coefficient of variation as a measure
of diversity. Organizational Research Methods, 3(3): 285-297.
Berger, J., Rosenholtz, S. J., & Zelditch, M. 1980. Status organizing processes. In A. Inkeles, N. J.
Smelser, & K. H. Turner (Eds.), Annual Review of Sociology, Vol. 6: 479-508. Palo Alto,
CA: Annual Reviews.
Blau, P. 1977. Inequality and heterogeneity. New York: Free Press.
Time, Teams, and Task Performance 31
Byrne, D. 1971. The attraction paradigm. New York: Academic Press.
Campion, M. A., Medsker, G. J., & Higgs, A. C. 1993. Relations between work group
characteristics and effectiveness: Implications for designing effective work groups. Personnel
Psychology, 46: 823-847.
Chattopadhyay, P. 1999. Beyond direct and symmetrical effects: The influence of demographic
dissimilarity on organizational citizenship behavior. Academy of Management Journal,
42(3): 273-287.
Cleveland, J. N., & Shore, L. M. 1992. Self- and supervisory perspectives on age and work
attitudes and performance. Journal of Applied Psychology, 77: 469-484.
DeMatteo, J. S., Eby, L. T., & Sundstrom, E. 1998. Team-based rewards: Current empirical evidence
and directions for future research. In Staw, B. M.& Cummings, L. L. (Eds.) Research in
Organizational Behavior, Vol. 20, (p. 141-183). Stamford, CT: JAI Press.
Elsass, P. M., & Graves, L. M. 1997. Demographic diversity in decision-making groups: The
experiences of women and people of color. Academy of Management Review, 22: 946-973.
Epstein, S. 1980. The stability of behavior: II. Implications for psychological research. American
Psychologist, 35: 790-806.
Fiske, S. T. 2000. Interdependence and the reduction of prejudice. In Oskamp, S. (Ed.), Reducing
prejudice and discrimination, p. 115-135. Mahwah, NJ: Erlbaum.
Goldberg, L. R. 1992. The development of markers for the big-five factor structure. Psychological
Assessment, 4 (1): 26-42.
Gruenfeld, D. H., Mannix, E. A., Williams, K. Y., & Neale, M. A. 1996. Group composition and
decision-making: How member familiarity and information distribution affect process and
Time, Teams, and Task Performance 32
performance. Organizational Behavior and Human Decision Processes, 67: 1-15.
Gully, S. M., Devine, D. J., & Whitney, D. J. 1995. A meta-analysis of cohesion and performance:
Effects of level of analysis and task interdependence. Small Group Research, 26: 497-520.
Hambrick, D. C. 1994. Top management groups: A conceptual integration and reconsideration of the
“team” label. Research in Organizational Behavior, 16: 171-214.
Harrison, D. A., McLaughlin, M. E., & Coalter, T. M. 1996. Context, cognition, and common
method variance: Psychometric and verbal protocol evidence. Organizational Behavior and
Human Decision Processes, 68: 246-261.
Harrison, D. A., Price, K. H. & Bell, M. P. 1998. Beyond relational demography: Time and the
effects of surface- and deep-level diversity on work group cohesion. Academy of
Management Journal, 41(1): 96-107.
Jackson, S.E. Brett, J.F., Sessa, V.I., Cooper, D.M., Julin, J.A., & Peyronnin, K. 1991. Some
differences make a difference: Interpersonal dissimilarity and group heterogeneity as
correlates of recruitment, promotion, and turnover. Journal of Applied Psychology, 76: 675-
689.
Jackson, S. E., May, K. E. & Whitney, K. 1995. Understanding the dynamics of diversity in
decision-making teams. In R .A. Guzzo & E. Salas (Eds.) Team decision-making
effectiveness in organizations: 204-261. San Francisco: Jossey-Bass.
James, L. R., & Brett, J. M. 1984. Mediators, moderators, and tests for mediation. Journal of
Applied Psychology, 69 (2): 307-321.
James, L. R., Demaree, R. G., Wolfe, G. 1993. Rwg: An assessment of within-group interrater
Time, Teams, and Task Performance 33
agreement. Journal of Applied Psychology, 78: 306-309.
Jehn, K. A. 1995. A multimethod examination of the benefits and detriments of intragroup conflict.
Administrative Science Quarterly, 40: 256-282.
Jehn, K.A., Chadwick, C., & Thatcher, S. M. B. 1997. To agree or not to agree: The effects of
value congruence, individual demographic dissimilarity, and conflict on workgroup outcomes.
International Journal of Conflict Management, 8(4): 287-305.
Jehn, K. A., & Mannix, E. A. 2001. The dynamic nature of conflict: A longitudinal study of
intragroup conflict and group performance. Academy of Management Journal, 44: 238-251.
Jehn, K. A., Northcraft, G.B., & Neale, M.A. 1999. Why differences make a difference: A field
study of diversity, conflict, and performance in workgroups. Administrative Science
Quarterly, 44: 741-763.
Kelley, H. H., & Thibaut, J. W. 1978. Interpersonal relations: A theory of interdependence. New
York: Wiley.
Kelly, J.R., & McGrath, J.E. 1988. On time and method. Newbury Park, CA: Sage.
Kirchmeyer, C. 1995. Demographic similarity to the work group: A longitudinal study of managers
at the early career stage. Journal of Organizational Behavior, 16(1): 67-83.
Konrad, A. M., Winter, S., & Gutek, B. A. 1992. Diversity in work group sex composition:
Implications for majority and minority members. In P. Tolbert & S. B. Bacharach (Eds.),
Research in the Sociology of Organizations, Vol. 10: 115-140. Greenwich, CT: JAI Press.
Kraiger, K., & Ford, J. K. 1985. A meta-analysis of ratee-race effects in performance ratings.
Journal of Applied Psychology, 70: 56-65.
Time, Teams, and Task Performance 34
Larkey, L. K. 1996. Toward a theory of communicative interactions in culturally diverse work
groups. Academy of Management Review, 21: 463-491.
Lawrence, B. S. 1997. The black box of organizational demography. Organization Science, 8: 1-22.
LePine, J. A., Hollenbeck, J. R., Ilgen, K. R., & Hedlund, J. 1997. Effects of individual differences
on the performance of hierarchical decision-making teams: Much more than g. Journal of
Applied Psychology, 82 (5): 803-811.
Locke, K. D., & Horowitz, L. M. 1990 Satisfaction in interpersonal interactions as a function of
similarity in level of dysphoria. Journal of Personality and Social Psychology, 58: 823-831.
McGrath, J. E. 1984. Groups: Interaction and performance. Englewood Cliffs, NJ: Prentice-Hall.
Milliken, F. J., & Martins, L. L. 1996. Searching for common threads: Understanding the multiple
effects of diversity in organizational groups. Academy of Management Review, 21: 402-433.
Neuman, G.A., Wagner, S.H., & Christiansen, N.D. 1999. The relationship between work-team
personality composition and the job performance of teams. Group & Organization
Management, 24: 28-46.
Newcomb, T. M. 1961. The acquaintance process. New York: Holt, Rinehart & Winston.
O’Reilly, C. A., III, Caldwell, D. F., & Barnett, W. P. 1989. Work group demography, social
integration, and turnover. Administrative Science Quarterly, 34: 21-37.
O'Reilly, C. A., Chatman, J., & Caldwell, D. F. 1991. People and organizational culture: A profile
comparison approach to assessing person-organization fit. Academy of Management Journal,
34: 487-516.
Pelled, L. H. 1996. Demographic diversity, conflict, and work group outcomes: An intervening
process theory. Organization Science, 6: 207-229.
Time, Teams, and Task Performance 35
Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. 1999. Exploring the black box: An analysis of work
group diversity, conflict, and performance. Administrative Science Quarterly, 44(1): 1-28.
Pelled, L. H. & Xin, K. R. 1997. Birds of a feather: Leader-member demographic similarity and
organizational attachment in Mexico. Leadership Quarterly, 8(4): 433-450.
Pettigrew, T. F. 1998. Intergroup contact theory. Annual Review of Psychology, 49: 65-85.
Petty, R. E. & Cacioppo, J. T. 1979. Issue involvement can increase or decrease persuasion by
enhancing message-relevant responses. Journal of Personality and Social Psychology, 37:
1915-1926.
Riordan, C. M. 2000. Relational demography within groups: Past developments, contradictions,
and new directions. Research in Personnel and Human Resources Management, 19: 131-173.
Riordan, C. M. & Shore, L. M. 1997. Demographic diversity and employee attitudes: An empirical
examination of relational demography within work units. Journal of Applied Psychology, 82:
342-358.
Rokeach, M. 1973. The nature of human values. New York: The Free Press.
Schneider, B. 1987. The people make the place. Personnel Psychology, 40: 437-453.
Schneider, B. Goldstein, H. W. & Smith, D. B. 1995. The ASA framework: An update. Personnel
Psychology, 48(4): 747-773.
Schwartz, S. H., & Bilsky, W. 1990. Toward a theory of universal content and structure of values:
Extensions and cross-cultural replications. Journal of Personality and Social Psychology, 58
(5): 878-891.
Seashore, S. E. 1979. Group cohesiveness in the industrial work group. Ann Arbor, Michigan:
University of Michigan Press.
Time, Teams, and Task Performance 36
Smith, K.G., Smith, K. A., Olian, J. D., Sims, Jr., H. P., O'Bannon, D. P. & Scully, J. A. 1994. Top
management team demography and process: The role of social integration and communication.
Administrative Science Quarterly, 39: 412-438.
Swann, W. B., Stein-Seroussi, A., & Giesler, B. R. 1992 Why people self-verify. Journal of
Personality and Social Psychology, 62: 392-401.
Tajfel, H. 1978. Differentiation between social groups: Studies in the social psychology of
intergroup relations. New York: Academic Press.
Tajfel, H., & Turner, J. 1986. The social identity theory of intergroup behavior. In S. Worchel &
W. G. Austin (Eds.), The social psychology of intergroup relations (pp. 7-24). Chicago:
Nelson-Hall.
Tjosvold, D. 1984. Cooperation theory and organizations. Human Relations, 37: 743-767.
Tsui, A. S., Eagan, T. D., & O’Reilly, C. A., III. 1992. Being different: Relational demography
and organizational attachment. Administrative Science Quarterly, 37: 549-579.
Tsui, A. S., & Gutek, B. A. 1999. Demographic differences in organizations: Current research and
future directions. Lexington: Lanham, MD.
Turban, D. B. & Jones, A. P. 1988. Supervisor-subordinate similarities: Types, effects, and
mechanisms. Journal of Applied Psychology, 73: 228-234.
Turner, J. C. 1982. Toward a cognitive redefinition of the social group. In Henri Tajfel (ed.), Social
identity and intergroup relations (p 15-40). Cambridge: Cambridge University Press.
Vroom, V. H. 1964. Work and motivation. New York: Wiley.
Wageman, R. 1995. Interdependence and group effectiveness. Administrative Science Quarterly,
40: 145-180.
Time, Teams, and Task Performance 37
Wiersema, M. F., & Bird, A. 1993. Organizational demography in Japanese firms: Group
heterogeneity, individual dissimilarity, and top management team turnover. Academy of
Management Journal, 36: 996-1025.
Williams, K. & O'Reilly, C. 1998. Demography and diversity in organizations: A review of 40
years of research. Research in Organizational Behavior, 20: 77-140.
Zenger, T. R., & Lawrence, B. S. 1989. Organizational demography: The differential effects of age
and tenure distributions on technical communication. Academy of Management Journal, 32:
353-376.
Time, Teams, and Task Performance 38
TABLE 1
Means, Standard Deviations and Correlations Among all Group-Level Variables
Variable Mean S.D. 1 2 345678910111213141516171819
Controls
1. Group Size 4.07 1.38
2. Mean GPA 2.97 0.40 .01
3. Conscientiousness (m) 172.29 11.93 .14 .20* (.89)
4. Values (m) 65.16 10.47 -.10 .17* .02 (.93)
5. Task Meaningfulness 14.29 3.01 .42** .01 -.08 -.03 (.89)
6. Outcome Importance (m) 12.65 1.18 .18* -.03 .07 -.16* .15 (.87)
Surface-Level Diversity
7. Age (sd) 5.06 3.20 .26** .05 .10 .10 .18* .02
8. Sex (b) 0.31 0.20 .08 -.12 -.13 -.04 .02 .03 -.08
9. Race/Ethnicity (b) 0.32 0.24 .22** -.09 .07 -.21* .14 .10 .04 .12
10. Marital Status (b) 0.31 0.22 .41** .13 .10 .08 .08 .13 .19* .18* .06
Deep-Level Diversity
11. Conscientiousness (sd) 17.52 9.33 -.01 .02 -.20* .00 .08 .08 .13 -.10 -.01 -.12
12. Values (sd) 16.29 8.33 .02 -.19* -.04 .04 .02 -.11 -.07 -.03 .10 .03 .23**
13. Task Meaningfulness (sd) 3.75 1.69 -.14 -.18* -.02 .06 -.40** .04 -.12 .01 .02 .04 .02 .13
14. Outcome Importance (sd) 2.07 1.15 -.06 -.09 -.14 .10 -.15 -.78** -.01 .09 .07 -.11 -.01 .17* .06
15. Team Reward Contingency .34 .17 .37** -.17* -.04 .03 .30** .09 .14 .16 .03 .23** .02 .01 -.11 .02
16. Perc. Surface-Level
Diversity
20.96 0.64 .19* .06 -.14 -.04 .12 .10 .30** .17* .52** .37** .02 .03 .09 .12 .12 (.68)
17. Perc. Deep-Level Diversity 47.65 0.54 .05 -.14 -.28** .01 .01 -.08 .00 .18* .15 .09 -.01 .08 .21** .28** .03 .47** (.82)
18. Collaboration .00 1.76 .34** .05 .12 .02 .25 .18* .03 -.02 -.03 .08 -.16 -.07 -.28** -.20* .38** -.12 -.30** (.75)
19. Team Social Integration .00 3.54 -.02 .22** .16* .15 -.05 .01 .01 -.09 -.18* -.04 -.03 -.19* -.14 -.21** -.07 -.41** -.57** .29** (.91)
20. Team Task Performance .87 .08 .13 .25** .16 .20* .00 .12 .04 .00 .03 .11 -.05 -.13 .00 -.13 .07 -.02 -.14 .30** .40**
Note: (b) is Blau's index, (sd) is standard deviation, and (m) is mean. *p < .05; **p < .01; n = 144.
Time, Teams, and Task Performance 39
TABLE 2
Regression Analyses for Testing Effects of
Effects of Actual Diversity (Time 1) on Perceived Diversity (Time 2)
Perceived Surface-Level Perceived Deep-Level
Independent Variables Diversity Diversity
ββ
Controls
Group Size -.16* -.03
Mean GPA .00 .01
Means of Deep-Level Variables
Conscientiousness (m) -.18** -.25**
Values (m) -.02 .00
Task Meaningfulness (m) .06 .11
Outcome Importance (m) .29 .25
Actual Diversity
Surface-Level
Age (sd) .26** .00
Sex (b) .00 .02
Race/Ethnicity (b) .47** .06
Marital Status (b) .37** .08
Deep-Level
Conscientiousness (sd) .00 -.05
Values (sd) -.06 -.02
Task Meaningfulness (sd) .06 .23**
Outcome Importance (sd) .34** .45**
Adjusted R2 .48** .16**
F-tests
All (Actual) Surface-Level 25.32** .41
All (Actual) Deep-Level 2.80* 4.71**
df 4, 129 4, 129
Note: (b) is Blau's index, (sd) is standard deviation, and (m) is mean. *p<.05; **p<.01
Time, Teams, and Task Performance 40
TABLE 3
Regression Analyses for Testing Moderated and Mediating Effects of
Perceived Diversity (Time 2) on Team Social Integration (Time 3)
Distal Main Mediated Moderated
Variables Effects Effects Effects Effects
ββββ
Controls
Group Size .01 -.02 .00 -.09
Mean GPA .13 .15* .07 .16*
Actual Diversity
Surface-Level
Age (sd) .02 .07
Sex (b) .01 .07
Race/Ethnicity (b) -.06 .10
Marital Status (b) -.13 .04
Deep-Level
Conscientiousness (sd) -.02 .03
Values (sd) -.09 -.12
Task Meaningfulness (sd) -.12 -.04
Outcome Importance (sd) -.18* -.18
Perceived Diversity
Surface-Level -.18* -.25* -.15*
Deep-Level -.42** -.38** -.37**
Collaboration .14 .15 -.05
Interaction Terms
Collaboration × Surface-Level .84*
Collaboration × Deep-Level -1.15**
Adjusted R2 .06* .35** .40** .41**
Overall F-test 1.89* 16.67** 5.47** 15.40**
df 10, 133 5, 138 13, 130 7, 136
F-test for ∆ R2 27.10** 27.10** 1.00 7.99**
df 2, 138 2, 138 8, 123 2, 136
Note: (b) is Blau's index and (sd) is standard deviation. *p<.05 **p<.01
Time, Teams, and Task Performance 41
TABLE 4
Regression Analysis for Testing Main and Mediating Effects of
Team Social Integration (Time 3) on Team Task Performance (Time 4)
Potential
Effects of
Potential
Effects of
Main Perceived Actual
Variables Effects Diversity Diversity
ββ β
Controls
Group Size .05 .02 .00
Mean GPA .18* .17* .17*
Team Social Integration .32** .43** .41**
Collaboration .21* .23** .27**
Perceived Diversity
Surface-Level .11 .07
Deep-Level .12 .11
Actual Diversity
Surface-Level
Age (sd) -.05
Sex (b) -.01
Race/Ethnicity (b) .05
Marital Status (b) .02
Deep-Level
Conscientiousness (sd) .06
Values (sd) -.08
Task Meaningfulness (sd) .13
Outcome Importance (sd) .13
Adjusted R2 .23 .25 .22
Overall F-test 11.56** 8.61** 3.21**
df 4, 139 6, 137 14, 129
F-test for ∆R2 11.56** 2.23 .56
df 4, 139 2, 137 8, 129
Time, Teams, and Task Performance 42
Note: (b) is Blau's index and (sd) is standard deviation. *p<.05; **p<.01.
Time, Teams, and Task Performance 43
a
p
riori characteristics
p
rocess variables affective outcomes
p
erformance outcomes
Time 1: Time 2: Time 3: Time 4:
++ Team
Social
Integration
Team Task
Performance
Collaboration
(Time Spent
In Interactions)
Perceived
Surface-Level
(Demographic)
Diversity
Perceived
Deep-Level
(Psychological)
Diversity
Actual
Surface-Level
(Demographic)
Diversity
•Race/Ethnicity
•Sex
•Age
•Marital Status
Actual
Deep-Level
(Psychological)
Diversity
•Personality
•Values
•Attitudes
•Beliefs
_
_
+
+
Intervening and Interactive Temporal Mechanisms
Translating Team Surface - and Deep -Level Diversity
into Social Integration and Performance
FIGURE 1
Team
Reward
Contingency
(Outcome Inter-
dependence)
H1a
H1b
H4
H2a
H2b
H5a
H5b
H6
+
_
a
p
riori characteristics
p
rocess variables affective outcomes
p
erformance outcomes
Time 1: Time 2: Time 3: Time 4:
++++ Team
Social
Integration
Team Task
Performance
Collaboration
(Time Spent
In Interactions)
Perceived
Surface-Level
(Demographic)
Diversity
Perceived
Deep-Level
(Psychological)
Diversity
Actual
Surface-Level
(Demographic)
Diversity
•Race/Ethnicity
•Sex
•Age
•Marital Status
Actual
Deep-Level
(Psychological)
Diversity
•Personality
•Values
•Attitudes
•Beliefs
_
_
+
+
Intervening and Interactive Temporal Mechanisms
Translating Team Surface - and Deep -Level Diversity
into Social Integration and Performance
FIGURE 1
Team
Reward
Contingency
(Outcome Inter-
dependence)
H1a
H1b
H4
H2a
H2b
H5a
H5b
H6
+
_
Time, Teams, and Task Performance 44
-5
-4
-3
-2
-1
0
1
2
3
4
5
19.5 20.0 20.5 21.0 21.5 22.0 22.5
Panel A: Perceived Surface-Level Diversity
Team Social Integration
-5
-4
-3
-2
-1
0
1
2
3
4
5
46.0 46.5 47.0 47.5 48.0 48.5 49.0
Panel B: Perceived Deep-Level Diversity
Team Social Integration
High (+1 s.d.) Collaboration
Low ( -1 s.d.) Collaboration
Mean Collaboration
High (+1 s.d.) Collaboration
Low ( -1 s.d.) Collaboration
Mean Collaboration
FIGURE 2
Differential Interaction Effects of Time Spent in Collaboration and (A) Perceived
Surface-Level Diversity and (B) Perceived Deep-Level Diversity, on Team Social Integration
-5
-4
-3
-2
-1
0
1
2
3
4
5
19.5 20.0 20.5 21.0 21.5 22.0 22.5
Panel A: Perceived Surface-Level Diversity
Team Social Integration
-5
-4
-3
-2
-1
0
1
2
3
4
5
46.0 46.5 47.0 47.5 48.0 48.5 49.0
Panel B: Perceived Deep-Level Diversity
Team Social Integration
High (+1 s.d.) Collaboration
Low ( -1 s.d.) Collaboration
Mean Collaboration
High (+1 s.d.) Collaboration
Low ( -1 s.d.) Collaboration
Mean Collaboration
FIGURE 2
Differential Interaction Effects of Time Spent in Collaboration and (A) Perceived
Surface-Level Diversity and (B) Perceived Deep-Level Diversity, on Team Social Integration
Time, Teams, and Task Performance 45
AUTHOR BIOS
David A. Harrison is a professor of management in the Smeal College of Business Administration
at Penn State University. His current research interests deal with temporal issues in
organizations, work role adjustment, executive decision-making, and measurement. He received
his a Ph.D. in I-O psychology from the University of Illinois, Urbana-Champaign.
Kenneth Price received his Ph.D. degree from Michigan State University and is currently a
professor of management at the University of Texas at Arlington. His current research interests
include diversity, organizational justice, and effective work teams.
Joanne H. Gavin is a Ph.D. student in organizational behavior at the University of Texas at
Arlington. She earned her M.B.A. and B.S. in Business Administration at the University of New
Orleans. Her research interest is in the area of personal integrity and decision making.
Anna T. Florey received her Ph.D. in human resources management and organizational behavior
from the University of Texas at Arlington. She is currently involved in training and evaluation
work in Asheville, North Carolina.
